Previous Page
448 RISK ANALYSIS OF BURIED WASTES FROM ELECTRICITY GENERATION
Case Study 2
In this case, the resistivity method has been used to map
the extent of contamination induced by an open waste
disposal site and to determine where to locate future
monitoring wells. The site is located SE of the city Isparta,
Turkey, on a nearly 100 m thick Quaternary alluvial
deposit of gravel, sand, and clay. The Schlumberger
sounding resistivity method was used to map resistivities
from 0.5 down to 25 meters in depth. Figure 11 shows the
apparent resistivity maps for different electrode spacings.
As expected, leachate from the waste disposal site has
decreased groundwater resistivity. The deeper layers are
characterized by low resistivities, and the shallower ones
by high resistivities.
A resistivity pseudosection along measurement line 9
is performed, and the result is shown in Fig. 12. As can
be seen, the low resistivity front from the waste disposal
site base is characterized by its low resistivity and is
developing northward. So, the extent of groundwater
pollution can easily be manifested using geoelectrical
mapping (8).
BIBLIOGRAPHY
1. Rao, T.G. (Ed.). (1992). Groundwater Exploration Techniques.
National Geophysical Research Institute, Hyderabad, India.
2. Jahani, H.R. (2005). Applied Geophysical Methods for Ground-
water Investigations. Iranian Hydraulic Association, Tehran,
Iran.
3. Griffiths, D.H. and King, R.F. (1988). Applied Geophysics for
Geologists and Engineers. Pergamon Press, Oxford.
4. U.S. Soil Conservation Service. (1971). Seismic and Resistivity
Methods of Geophysical Exploration. Water Resources Publi-
cations, United States Soil Conservation Service, Washington,
DC.
5. Zohdy, A.A.R. et al. (1990). Application of Surface Geo-
physics to Groundwater Investigations. U.S. Geological Survey,
Denver, CO.
6. Iranian National Karst Research Center. (1997). Resistivity
Methods of Geophysical Explorations in Karstic Areas in West
Azerbaijan Province. Shiraz.
7. Hotzl, H. and Werner, A. (Eds.). (1992). Tracer Hydrology. A.A.
Balkema, Rotterdam, The Netherlands.
8. Karlýk, G. and Ali Kaya, M. (2001). Investigation of groundwa-
ter contamination using electric and electromagnetic methods
at an open waste-disposal site: A case study from Isparta,
Turkey. Environ. Geol. 40(6): 725–731.
RISK ANALYSIS OF BURIED WASTES FROM
ELECTRICITY GENERATION
BERNARD L. COHEN
University of Pittsburgh
Pittsburgh, Pennsylvania
An important aspect of generating electricity from nuclear
power is the disposal of the high level radioactive waste.
The most widely anticipated plan is to convert it into a
rock-like material and bury it deep underground, typically
600 meters below the surface. In evaluating the safety of
this procedure, a key element is to estimate the probability
per year, P, for an atom of this buried waste to be dissolved
by groundwater and eventually enter a human stomach.
Obviously, this probability is site dependent and depends
on present underground characteristics and future events
in the region, many of them unpredictable, including
climate changes, geological land uplifting causing rivers to
change their courses, volcanoes, earthquakes, and human
intrusion.
INTRODUCTION
A first estimate of P can be obtained by considering that
all properties of the region are the U.S. average (lower
48 states); the purpose of this article is to obtain such an
estimate. Most relevant characteristics and future changes
in them are occurring somewhere in the United States
and are hence taken into account with or including some
estimate of their probability of occurrence. The value of
P we obtain can be interpreted as the average value of
P for a large number of randomly chosen sites. It seems
reasonable to assume that the very elaborate site selection
efforts by geological and hydrologic experts should choose
a site at least as secure as a randomly chosen site.
We assume that the waste converted to a rock-like
material behaves like average rock—differences between
them will be considered later—and calculate P for average
rock. P can be calculated as the dissolution rate, R, the
probability per year for an atom of the rock to be dissolved
in groundwater, times the probability, p, for an atom once
dissolved into groundwater to enter a human stomach:
P=R×p (1)
Dissolution Rate for Rock, R
We begin by calculating R. From the rate at which rivers
carry dissolved and suspended material into the oceans, it
is estimated (1) that the surface of the continent is eroding
at an average rate of 5 × 10−5 (hereafter, we use the
notation 5 E−5) meters per year (hereafter, m/y). About
28% of this material is in solution, corresponding to 1.4
E−5 m/y removed by chemical dissolution. It is estimated
(2,3) that 15% of the water flow in rivers is derived
from groundwater (aquifers), and the rest comes from
surface runoff. If the concentrations of dissolved materials
were the same in rivers as in aquifers, this would mean
that the latter dissolve (0.15 × 1.4 E−5) = 2.1 E−6 m/y
of rock thickness. The average concentrations of a few
dissolved materials in aquifers (4) and in river water (5)
are listed in Table 1. Silica and calcium are such important
Table 1. Average Concentrations in mg/kg of Dissolved
Materials in Aquifersa and in River Waterb
Material Silica Calcium Magnesium Potassium Iron Uranium
Aquifers 20 30 3 2 0.3 3 E−4
Rivers 7 15 4 2.2 0.5 4 E−4
a
Ref. 4.
b
Ref. 5.
 RISK ANALYSIS OF BURIED WASTES FROM ELECTRICITY GENERATION 449

components of rock that we interpret these data to indicate Table 2. Calculation of Fractional Removal per year of
that aquifers carry about twice the concentration of Rock Materials by a Typical Aquifer. Columns (2) and (3)
dissolved material in rivers, which leads to the conclusion refer to a 1 Square Meter Cross Section of the Aquifer.
that [30/(30 + 85)] = 26% of the dissolved material in See Discussion in Text
rivers, (0.26 × 1.4 E−5) = 3.6 E−6 m/y of rock depth, is (3) (4)
contributed by aquifer dissolution of rock. (1) (2) Kg in Rock Fraction Removed
The next problem is to estimate what fraction of this Material Kg/y into River (× 1 E6) (× 1 E−8)
material, f , is removed per meter of depth at 600 m; let us Ca 0.3 15 2
say from between 599 m and 600 m. As a crude gross over- Mg 0.03 3 1
estimate, we might assume that rock erosion is constant K 0.02 3 0.7
with depth down to 600 m and zero below 600 m, in which Fe 0.003 9 0.03
case f = 1/600 = 16 E−4. A more reasonable approach is U 3 E−6 8 E−4 0.3
as follows: The annual circulation of groundwater, defined Silica 0.2 150 0.13
as the quantity per year entering or leaving aquifers at Carbonate 1.3 18 8
the stated depth, has been given (2,3) as

shallow (<800 m) : 310 E9 cubic meters per year
deep (>800 m) : 6.2 E9 cubic meters per year
If we assume that the flow decreases exponentially with
depth and that the rate of rock dissolution is proportional
to this groundwater flow, determining the constants in the
relationship from these data, we find (6) that f = 2.6 E−4,
one-sixth of our crude gross overestimate.
Using this gives the quantity of rock eroded per
year per meter of depth at 600 m below the surface
as (3.6 E−6 × 2.6 E−4) = 1.0 E−9 m/y. If 1.0 E−9 m/y of
depth is dissolved from one meter of rock depth, the
probability per year for an atom of that rock to be dissolved
must be 1.0 E−9. Thus,
R = 1.0 E−9 (2)
This result plays a key role in our risk analysis,
so it is interesting to seek alternative approaches to
calculating it. One such approach follows: A typical aquifer
reaching to the waste burial depth of 600 m may be
about 100 km long and have a flow velocity of 100 m/y
through rock of 10% porosity, discharging into a river
The water discharged from it annually per square meter
of cross-sectional area is then 10% of the volume of
a column of water 100 m long and one square meter
in cross section, which is 10 cubic meters, or 10,000
liters (L).
Chemical analyses of groundwater (4) indicate that it
typically contains 30 mg/L or 30 E−6 kg/L (milligrams
per liter, or kilograms per liter) of Calcium (Ca), so it
discharges (30 E−6 kg/L × 10,000L =) 0.3 kg of Ca into
the river each year. This is the first entry in Column (2)
of Table 2. This calcium was derived by dissolution from
the rock through which the aquifer had passed, 100 km,
or 1E 5 m, long, one square meter in cross section, and
a specific gravity of about 3.0, which gives it a mass of
(3000 × 1 E5) = 3 E8 Kg. Typical rock contains 5% Ca (1),
so the Ca contained in this source rock is (0.05 × 3 E8) =
15 E6 kg; this is the first entry in Column (3) of Table 2.
If 0.3 kg of Ca per year is dissolved from 15 E6 kg of
Ca in the rock, the probability for a Ca atom to be
dissolved must be (0.3/15 E6) = 2 E−8 per year; this is
the first entry in Column (4) of Table 2. Table 2 includes
data for other materials, obtained analogously to that
for Ca.
Of the materials listed in Table 2, the materials in the
radioactive waste are most similar to Fe and U and not
at all similar to carbonate ions. Thus, the numbers in
Column (4) of Table 2 might give an estimate of about
R = 7 E−9, seven times higher than Eq. 2. But the aquifer
we have assumed in this calculation is substantially more
robust than the groundwater encountered by average rock
at 600 m depth, so it seems reasonable to conclude that
Eq. 2 is roughly verified.
Probability of Transport from Groundwater to Human
Stomachs, p
Next we turn to the problem of estimating the probability
of transferring an atom of material dissolved in groundwa-
ter into human stomachs, p in Eq. 1. We assume that once
the material is dissolved in groundwater, it moves with
groundwater, eventually reaching shallow aquifers that
feed into rivers. Contributions to p derive from our use
of rivers and wells drilled into aquifers for potable water,
from fish removed from rivers and used for food, and from
use of well and river water for irrigating food crops. The
potable water path is the simplest to calculate and, it
turns out, the most important. The average person ingests
2 L/day of potable water which corresponds to ingestion
by the U.S. population of 2 × 365 × 2.8 E8 = 2.0 E11 L/y
(liters per year). Of this, 45% is derived from wells, and
55% comes from rivers (7), corresponding to 9 E10 and
1.1 E11 L/y respectively entering human stomachs. The
estimated water flow in U.S. rivers is 1.7 E15 L/y (1) and
1.9 E15 L/y (2,3); we use 1.8 E15 L/y. The estimated water
flow in aquifers is 16% of the flow in rivers (3), or 2.9 E14
L/y The contributions to p from the use of rivers and wells
for potable water are then
p(rivers) = (1.1 E11/1.8 E15) = 6.1 E−5
p(well water) = (9 E10/2.9 E14) = 3.1 E−4
Note that we have ignored removal of material by filtration
processes, making these conservative estimates, that is,
more likely to be high than low.
Analyses of the fish and irrigation contributions
to p will be sketched below. They give results that
450 RISK ANALYSIS OF BURIED WASTES FROM ELECTRICITY GENERATION

vary considerably from element to element because of H

variations in bioaccumulative factors (8). For the fish Human oral
pathway, p(fish) is rarely larger than 1 E−5, and for most intake
important elements in the waste, it is much smaller. For
0.9 0.1
the irrigation pathway, p(irrigation) is rarely more than 0.02 F W 0.45
1 E−4 and somewhat less for the most important elements Food Drinking
in the waste. We therefore take as conservative estimates water
0.88 0.1 0.55
p(fish) = 1 E−5
G E
D
p(irrigation) = 1 E−4 Sea-
Soil Irrigation
food water 0.37

Summing these four contributions gives p = 5 E−4. 1.0 0.63

Using this and Eq. 2 in Eq. 1 then gives the result we 0.6
are seeking for P: B C
Rivers Wells
P = (1 E−9 × 5 E−4) = 5 E−13 (3)
0.4 1.0
Alternative Approach for Transfer from Rock to Human
A
Stomachs Rock
We now present an essentially independent method (9)
Figure 1. Pathway diagram for transmission of materials from
for calculating P, based on our knowledge of quantities of
rock and soil into food and water ingested by people. Numbers
elements entering human stomachs each year (10,11) and attached to each line show the fraction of the input to the target
the quantities of these materials in rock (5), supplemented box that is derived from that line.
by an estimation of the fraction of these that derive from
soil rather than from rock.
We begin with the very crude assumption that all their fractional contributions to human intake are then
material entering human stomachs derives from the top calculated as
600 m of rock. We know how much of each chemical
element is contained in this rock (5), and we know how ABWH 0.4 × 0.55 × 0.1 = 0.022;
much of each chemical element enters human stomachs ACWH 1.0 × 0.45 × 0.1 = 0.045;
each year (10,11), so the ratio of these gives a crude ABEFH 0.4 × 0.63 × 0.1 × 0.9 = 0.023;
estimate of P for each element. An average of these, ACEFH 1.0 × 0.37 × 0.1 × 0.9 = 0.033;
weighting for the importance of each element in the ABGFH 0.4 × 1.0 × 0.02 × 0.9 = 0.007;
waste, gives TOTAL = 0.13
P(very crude) = 1 E−11
Our final value for P is then
Some of the crudeness of this estimate may be reduced
by dropping the implication that all depths of rock down to P(final) = (0.13 × 1.6 E−12) = 2 E−13 (4)
600 m are equally likely to contribute and assuming that
the contribution from various depths is proportional to the This result, Eq. 4, is in substantial agreement with
groundwater flow at that depth. Using the exponential the estimate from Equation 3 derived from essentially
decrease in this flow alluded to before, it turns out that independent approaches.
the flow at 600 m depth is 16% of the average for the top
600 m, giving an improved, but still rough estimate:
HIGH LEVEL WASTE VERSUS AVERAGE ROCK
P(rough) = (0.16 × 1 E−11) = 1.6 E−12
This result applies to average rock, but our interest
This result is still an overestimate because most of is in applying it to the rock-like material, probably
the material entering human stomachs derives from a borosilicate glass, into which the waste has been
the top layers of soil rather than from the underlying converted. The principal difference is in leachability
rock. To analyze this problem, a pathway analysis was by groundwater. Leach rates by distilled water for
developed (9) as shown in Fig. 1. The numbers attached various rocks and glasses have been measured, and
to the arrows are the fraction of the input into the the results were correlated with the Gibbs free energy
compartment derived from that arrow. For example, 90% of hydration (12). It was concluded that the leach rate
of human oral intake is from food, and 10% is from for borosilicate glass is similar to that for basalt and
drinking water; 2% of the minerals in food comes from about three times more rapid than the leach rate for
seafood (derived from rivers), 88% comes directly from average rock. Applying this to the results in Eqs. 3 and 4
soil, and 10% comes from soil via irrigation water (which is gives the rate of transfer for an atom of buried high
derived 63% from rivers and 37% from wells). Using Fig. 1, level waste into a human stomach as about 1.0 E−12
the various pathways from rock to human stomachs and per year.
 GROUNDWATER CONTAMINATION FROM RUNOFF 451

Concerns have been raised that the high level waste GROUNDWATER CONTAMINATION FROM
glass may be much less secure than average rock because it RUNOFF
is not in chemical equilibrium with the local rock–ground
water regime. However, chemical equilibrium is a
NITISH PRIYADARSHI
surface phenomenon (13,14). Groundwater is saturated
Ranchi University
with dissolved silica. If a foreign silica-based material
Ranchi, Jharkhand, India
such as glass is emplaced, there is initially rapid
dissolution, removing a very thin surface layer. But this
brings more silica into the water which then becomes
INTRODUCTION
supersaturated with silica, resulting in precipitation of
silica from the water onto the surface—this would be
Groundwater is water that lies below the soil surface
the most insoluble component of the silica originally in
and fills the pore spaces in and around rock, sand,
the water. The result is a buildup of highly insoluble
gravel, and other materials. Groundwater, under most
silica on the surface. Waste can then be dissolved only
conditions, is safer and more reliable for use than surface
by diffusion through this surface layer, a very slow
water because surface water is more readily exposed to
process. Moreover, this leads to further thickening of
pollutants from factories, for example, than groundwater.
the surface layer, which further retards the diffusion
This by no means says that groundwater is invulnerable
process. Consequently, the rate of dissolution decreases
to contamination. Once groundwater is contaminated,
exponentially, only a tiny fraction of the waste becomes
it is extremely costly to remove the contaminant. Any
involved in the process, and dissolution proceeds only as
chemicals that are easily soluble and penetrate the soil
for the surrounding rock.
are prime candidates as groundwater pollutants.

BIBLIOGRAPHY
HOW MUCH DO WE DEPEND ON GROUNDWATER?
1. Garrels, R.M. and MacKenzie, F.T. (1971). Evolution of
Sedimentary Rocks. W.W. Norton Co., New York. According to 1985 U.S. figures, groundwater provides
2. Todd, D.K. (1980). Groundwater Hydrology. John Wiley & an estimated
Sons, New York.
3. Ad Hoc Panel on Hydrology. (1962). Scientific Hydrology, • 22% of all freshwater withdrawals
Federal Council for Science and Technology. Washington,
• 53% of drinking water for the total population and
DC.
97% of drinking water for the rural population
4. White, D.E., Hein, J.D., and Waring, G.A. (1963). Chemical
Composition of Sub-Surface Waters, Chap. F in Data of • 40% of public water supply withdrawals
Geochemistry, 6th Edn. M. Fleischer (Ed.). U.S. Government • 46% of domestic and commercial use
Printing Office, Washington, DC. • 24% of industrial and mining use
5. Bowen, H.J.M. (1979). Environmental Chemistry of the
• 34% of agricultural use (mostly for irrigation)
Elements. Academic Press, New York.
6. Cohen, B.L. (1985). A simple probabilistic risk analysis
for high level waste repositories. Nuclear Technology 68: CAUSES OF URBAN RUNOFF
73–76.
7. Byrne, F. (1974). Earth and Man. Wm. C. Brown Co., Urban runoff can be attributed to many things, including
Dubuque, IA.
the amount of rainfall, the soil conditions, and the degree
8. Cohen, B.L. (1984). Probability for human intake of an atom of urbanization. Rainwater can go in many directions
randomly released into ground, rivers, oceans, and air. Health
once it has reached the earth’s surface. Rainwater can
Physics 47: 281–292.
be absorbed by the soil on the land surface; absorbed
9. Cohen, B.L. (1986). A generic probabilistic risk analysis for a
by surrounding vegetation; directly deposited into oceans,
high level waste repository. Health Physics 51: 519–528.
streams, and rivers; and infiltrate through the surface and
10. International Commission on Radiological Protection. (1975).
subsurface soils into the groundwater.
Report of the Task Group on Reference Man. ICRP Publication
23, Pergamon Press, New York.
Urban runoff can happen anytime of the year when
excessive water use from irrigation, car washing, and other
11. Pennington, J.A.T. et al. (1984). Selected minerals in food
survey, 1974–1982. Jour. Am. Dietetic Assn. 84: 771ff.
sources carries litter, lawn clippings, and other urban
pollutants into storm drains.
12. Plodinek, M.J., Jantzen, C.M., and Wicks, G.C.. (1983). Sta-
bility of Radioactive Waste Glasses Assessed from Hydration
The majority of people live in cities that are becoming
Thermodynamics. Savannah River Laboratory Reports DP- larger and larger and are sprawling into vast suburban
MS-83-61, DP-MS-83-21, and DP-MS-83-66. neighborhoods. Roofs, parking lots, streets, and other
13. Cohen, B.L. (1980). Analysis, critique, and evaluation of impervious surfaces of an urban environment cause
high level waste water intrusion scenario studies. Nuclear rainwater to collect and be forced out through a storm
Technology 48: 63ff. drain system. If the drainage system does not connect to
14. Cohen, B.L. (1985). Critique of the National Academy of a wastewater treatment facility, then the rainwater and
Sciences study of the isolation system for geologic disposal of everything that is carried with it travels into groundwater,
radioactive waste. Nuclear Technology 20: 433ff. local streams, and rivers.
452 GROUNDWATER CONTAMINATION FROM RUNOFF

GROUNDWATER CONTAMINATION FROM URBAN
RUNOFF
Not too long ago, urban runoff was considered an insignif-
icant contributor to groundwater contamination. Urban
runoff is now recognized as a significant source of contam-
ination in water. Effluents from urban areas contain large
concentration of oils, greases, nutrients, heavy metals, and
detergents. Detergents that are soluble can pass through
the soil and pollute groundwater. Raw sewage dumped in
shallow soakpits and seepage from polluted lakes, ponds,
and streams also pollute groundwater. Rainfall could pick
up substantial contaminants from dust and air and join
the aquifer below. The infiltration of liquids containing
toxic pollutants may cause pollution in sandy soils and
well waters.
Some of the contaminants expected in urban runoff are
shown in Table 1. Groundwater moves through rocks and
subsurface soil, so it has lots of opportunity to dissolve
substances as it moves. For that reason, groundwater
often has more dissolved substances than surface water.
Even though the ground is an excellent mechanism
for filtering out particulate matter, such as leaves,
soil, and bugs, dissolved chemicals and gases can still
occur in large enough concentrations in groundwater to
cause problems. Underground water can be contaminated
through urban runoff containing industrial, domestic, and
agricultural chemical wastes, which includes chemicals
such as pesticides and herbicides that many homeowners
apply to their lawns. Contamination of groundwater by
road salt is of major concern in northern areas of the
United States. Salt is spread on the roads to melt ice, and
salt is so soluble in water that excess sodium and chloride
is transported through runoff after ice melts and finally
contaminates subsurface groundwater.
In the United States today, there may be more than
20,000 known abandoned and uncontrolled hazardous
waste sites, and numbers grow every year. If there is
a leak, these contaminants may be carried away with
runoff and finally make their way down through the soil
and into the groundwater.
Landfills are another major source of contamination.
Landfills are the places where our garbage is taken to be
buried. Landfills are supposed to have a protective bottom
layer to prevent contaminants from getting into the water.
After rainfall, contaminants from the landfill (car battery
acid, paint, household cleaners, etc.) are washed away
with surface runoff and make their way down into the
groundwater. The problem of pollution from landfills is
greatest where high rainfall and shallow water tables
occur. Important pollutants frequently found in leachate
include BOD, COD, iron, manganese, chloride, nitrate,
hardness, and trace elements. Hardness, alkalinity, and
total dissolved solids are often increased.
Chemicals include products used on lawns and farm
fields to fertilize plants. The primary fertilizers are
compounds of nitrogen, phosphorus, and potassium.
Phosphate and potassium fertilizers are readily adsorbed
on soil particles and seldom constitute a pollution problem.
But nitrogen in solution is only partially used by plants or
adsorbed by the soils, and it is the primary fertilizer
pollutant (1). When the rain comes, this chemical is
washed away with the surface runoff and eventually goes
into the groundwater.
Animal wastes are confined within a limited area, in
beef or milk production, so large amounts of waste are
deposited on the ground. Thus, for the 120 to 150 days that
a beef animal remains in a feedlot, it produces more than a
half ton of manure on a dry weight basis. When thousands
of animals are in a single feedlot, the natural assimilative
capacity of the soil becomes overtaxed. Surface runoff
in contact with the manure carries highly concentrated
pollutants to subsurface waters. Animal wastes may
transport salts, organic loads, and bacteria into the water
and soil. Nitrate-nitrogen is the most important persistent
pollutant that may reach the groundwater (2).
Sanitary systems have a significant impact on shallow
aquifers in urban areas. Field observations revealed that
contamination of groundwater is caused by infiltration of
surface runoff polluted by municipal waste and sewage
and/or leakage of sanitary effluents. It has increased
the salt content of groundwater, particularly nitrates

Table 1. The Contaminants that may be Expected in Runoffa

1. Asbestos from brake and clutch linings
2. Bacteria and viruses from animals and birds, soils, litter, livestock hauling, livestock waste hauling and on-site sewage tanks and
fields
3. Bromide from auto exhaust
4. Cadmium from tire fillers and insecticides
5. Chromium from moving engine parts and brake linings
6. Copper from bearing and bushing wear, moving engine parts, brake linings, and radiator repair
7. Cyanide and chloride from deicing road salts
8. Pesticides (fungicides, herbicides, and insecticides) from roadside maintenance
9. Iron from auto bodies, moving engine parts, bridges, guardrails, overpasses, lamp standards, and other structures
10. Manganese from moving engine parts and gasoline additives
11. Nickel from diesel fuel, lubricating oil, bushing wear, brake linings, and asphalt paving
12. Nitrogen from animal wastes, on-site sewage systems, vegetative matter, and fertilizers
13. Petroleum compounds from paving, fuels spills, engine blow-by, lubricant leaks, antifreeze, and hydraulic fluids
14. Phosphorus from animal wastes, on-site sewage systems, vegetative matter, and fertilizers
15. Potassium from fertilizers
16. Sulfate from roadbeds, road salts, and fuels
17. Zinc and lead from tire fillers, motor oil additives, automotive and radiator repairs, grease, and paint manufacturing
a
Source: Ministry of Water, Land and Air Protection, Government of British Columbia.

and chlorides, and has reduced the oxygen content.
In addition, introduction of pathogenic bacteria and
viruses into groundwater has caused many outbreaks of
waterborne disease.
Solid materials are frequently stockpiled near indus-
trial plants and construction sites. These may be the raw
materials awaiting use, or they may be solid wastes placed
for temporary or permanent storage. Precipitation falling
on unsheltered stockpiles may transport heavy metals,
salts, and other inorganic and organic constituents as
pollutants to the groundwater.
Urban runoff pollutants are many and varied, depend-
ing on land uses and pollutant sources in an urban area.
Typically, loadings of urban pollutants are greatest from
industrial and commercial areas. Although sources of spe-
cific pollutants may vary widely in urban areas, motor
vehicles are recognized as a major source of pollutants;
they contribute oils, greases, hydrocarbons, and toxic met-
als such as lead, zinc, copper, chromium, and arsenic.
BIBLIOGRAPHY
1. Todd, D.K. (1995). Groundwater Hydrology. Wiley, Toronto,
Canada, p. 332.
2. Gillham, R.W. and Webber, L.R. (1969). Nitrogen contamina-
tion of groundwater by barnyard leachates. J. Water Pollut.
Control Fed. 41(10): 1752–1762.
SALINE SEEP
JOHN E. MOORE
USGS (Refired)
Denver, Colorado
Jackson (1) defines saline seep as an intermittent or
continuous saline water discharge at or near the
soil surface under dryland conditions that reduces or
Excess water
Infiltration
Deep percolation
Water
table
XXX Saline seep
Salt particles Low permeability
SALINE SEEP 453
eliminates crop growth. Human induced salinization of
land and water resources due to water table rise is as old as
the history of human settlement and irrigation (2). Early
settlements in the valleys of the Tigris and Euphrates
flourished about 4000 B.C.: however, the irrigated farming
of wheat resulted in salt accumulation, and in time
the area was abandoned. Historic evidence from Russia,
China, India, Pakistan, South American, United States,
Canada, and Australia show that saline seeps are a
worldwide problem. The number of saline seeps will
probably increase in the future because an increase
in world population will be accompanied by increases
in irrigation.
Saline seeps result from a combination of geologic,
climatic, and management factors. Development of a seep
starts at the recharge area. As the water moves through
the aquifer, the water dissolves and accumulates salts.
Saline seeps are formed by mobilization of salt stored in
the soil profile. The very act of irrigation is the cause
for salinity because every irrigation event adds some
salt to the soil. Land clearing or irrigation can cause
the groundwater to rise. The additional water moves the
water table closer to the land surface. When the water
table is within 2–3 feet of the surface, capillary action can
lift the water to the surface where it evaporates and leaves
salts behind (Figs. 1 and 2).
The salt can also move laterally to affect streams.
The common soil forming minerals calcium, sodium,
magnesium, and potassium form water-soluble salts in
the soil.
Techniques for controlling seeps originated in Montana
and Alberta (3). The two major methods to control seeps
are (1) cut off the recharge water and (2) reclaim the
seep area. Installing subsurface tiles to intercept the
groundwater flow might cut off recharge water. The second
method is to increase water use in the recharge area by
changing the type of crop. Alfalfa is an excellent crop
substitute because it is a deep-rooted perennial capable of
Discharge area
Evapotranspiration
XXX
XXXXXXXXXXXXXXXXXX
Figure 1. Water movement to seep.
454 GROUNDWATER SAMPLING TECHNIQUES FOR ENVIRONMENTAL PROJECTS
Water evaporates
Salts remain behind xxxxxxxxxxxxxxxxxx
Soil surface
Salty water moves
upward from the
water table by
capillary action
Water table
Groundwater
Figure 2. Evaporation and salt concentration.
using large amounts of water. Seeps have been reclaimed
within 5 years by planting alfalfa in the recharge area (4).
In southeast Australia, the replacement of deep-rooted
native vegetation by dryland and irrigated agriculture
has resulted in a major change in surface water and
groundwater systems in the Murray–Darling basin. The
Murray River has a salinity of less than 25 mg/L in the
headwaters and 480 mg/L downstream. Water that was
previously used by vegetation for transpiration is now
leaking past the vadose zone and entering the water table.
The salinization process is closely linked to groundwater
changes. In both irrigation and dryland farming, salts
are remobilized into the production zone. The salts were
originally stored in the aquifer or the vadose zone.
Evaporation concentrates the salts in the near surface
zone to levels where the plants are. For salinization to
occur, it is necessary to have both a hydrologic change and
a source of salt.
Saline soils and sodic soils are quite common in parts
of western North Dakota and Montana. Most saline seeps
have developed recently (postsettlement). The formation
of saline seeps is closely related to the practice of summer
fallow for moisture conservation (5).
Colorado River has 25 mg/L salt in the headwaters
and 825 mg/L downstream at Imperial Dam. In the
San Joaquin River in California, salinity increased from
330 mg/L in the 1930s to 600 mg/L in the 1970s. In the
former Soviet Union, the salinity of the Syr Darya River
which discharges to the Aral Sea increases from 300 mg/L
to 2000 mg/L downstream.
The water table rise in agricultural areas of the world
has contributed to the salinization of large areas. The
development of land and water resources of the earth will
be more difficult and more expensive than in the past. The
control of population and the efficient use of developed
resources will be essential in the future.
BIBLIOGRAPHY
1. Jackson, J. (1997). Glossary of Geology. American Geological
Institute, p. 769.
2. Ghasemi, F. et al. (1994). Water table rise and the expansion
of land and water resources salinsation: a global view:
IAH Congress Proceedings. Water Down Under, Adelaide,
Australia, pp. 607–610.
3. Brown, P.L. et al. (1983). Saline seep diagnosis control and
reclamation. USDA-ARS conservation Research Report No. 30.
4. Lamon, R.E. and Whitney, D.A. (1992). Management of Saline
and Sodic Soils:MF-1022 Cooperative Extension Service.
Kansas State University, Manhattan, KS.
5. Wentz, D. (1997). Dryland Saline Seeps-Types and Causes.
Alberta Agriculture.
GROUNDWATER SAMPLING TECHNIQUES
FOR ENVIRONMENTAL PROJECTS
JAMES A. JACOBS
Environmental Bio-Systems, Inc.
Mill Valley, California
Groundwater sampling techniques are important for
accurate project results and understanding aquifers
whether the project is sampling a groundwater monitoring
well as part of an environmental assessment or regular
sampling of a potable water supply well. Detailed
groundwater sampling information can be obtained
from Testa (1), Jacobs (2), the American Society for
Testing Materials (3), the California Regional Water
Quality Control Board (4), and Driscoll (5). Although
there are numerous differences in groundwater sampling
and record keeping techniques, regional variations, and
specific regulatory requirements that vary from state to
state, a generalized summary of selected environmental
groundwater sampling techniques is given below.
PURPOSE OF GROUNDWATER SAMPLES
Groundwater samples are collected for a variety of
reasons. Groundwater can be collected as a ‘‘grab’’
sample or from a monitoring well. Groundwater samples
are used to define the vertical and aerial extent of
groundwater contamination, to monitor target chemical
concentrations over time, to provide a measurement
for detecting a contaminant plume front, to evalvate
unexpected changes in plume size or direction of flow,
to determine the extent of interaquifer movement of
contaminants, to determine aquifer characteristics such
as permeability, transmissivity, etc.), to estimate the rate
of contaminant plume movement, to develop a data base
for designing remedial measures, to determine the effects
of the remedial measures, to assist in performing the
remedial work (provide hydraulic control and contaminant
removal), to provide data base for groundwater modeling,
and to evaluate the aquifer during regular sampling over
the yearly hydrologic cycle (6).
GRAB GROUNDWATER SAMPLING
A ‘‘grab’’ groundwater sample is collected without the use
of a fully installed groundwater monitoring well. Methods
include using a bailer or jar to collect groundwater from an
excavation or trench made with a backhoe or excavator.
Other ‘‘grab’’ samples can be collected using a drilling
rig wherein a sampler is drilled or pushed to the target
depth and the sampler is pushed a bit further into the
aquifer. Water is allowed to collect in the sampler. The
 GROUNDWATER SAMPLING TECHNIQUES FOR ENVIRONMENTAL PROJECTS
water is removed from the sampler using a bailer or
by using a peristaltic pump and tubing, check-ball valve
with tubing or other similar methods. These methods,
frequently called punch methods, are commonly used with
hollow stem auger rigs or direct push technology rigs. Cone
penetrometer (CPT) rigs can use specialized groundwater
containers that are inserted into a cavity of the water
sampler tip. The groundwater containers are under a
vacuum. At the proper depth, the sampler is opened,
the groundwater container is pierced, and water enters
the container.
Temporary wells are used occasionally in environmen-
tal investigations. A temporary well consists of a well
screen below a casing riser placed in an open borehole.
Several well volumes can be purged prior to sampling.
A bailer or other method is used to bring the water to
the surface. After groundwater sampling is completed, the
borehole is usually tremie grouted to seal the annulus.
WATER WELLS
Water wells are installed using a variety of rigs: hollow
stem auger, mud or air-rotary, air percussion, diamond
coring, cable tool, and others. After well installation,
wells are developed within a day to several days of
installation to remove residual drilling materials from
the well bore and to improve well performance by
removing any fine material in the filter pack that can
pass from the native soil into the well. Well development
techniques include pumping, bailing, surging, jetting,
and airlifting. In most cases, surging and pumping are
satisfactory. Development water is inspected for turbidity,
product sheen, odors, and sediment. A minimum of
several casing volumes is removed during development.
The well is considered fully developed when consistent
pH, temperature, and conductivity readings indicate
characteristic groundwater for the aquifer. If the aquifer
is slow to recharge, development will continue until the
well is pumped dry. Cross contamination of wells from
pumps is avoided by using dedicated equipment and proper
decontamination procedures.
Due to the large volumes of purge water generated
during groundwater sampling and the costs for stor-
age and disposal of the water, some regulatory agencies
are now allowing various low-flow groundwater sampling
techniques for compliance groundwater monitoring. Propo-
nents of micropurging suggest that evacuation of a specific
volume of water may increase turbidity from agitation or
mixing of the well water column. Larger purge volume may
increase the risk of colloid mobilization. Proper well devel-
opment immediately after well construction is essential
when using low-flow sampling (7).
Micropurging is the evacuation of tubing, sampling
devices, and other items from the well prior to sampling
groundwater. There is no purging prior to sample
collection with micropurging, so there is no method for
determining if the well contains stagnant well water or
water representative of the aquifer. For this reason, many
regulatory agencies do not approve micropurging. Instead,
when minimal purging is requested, low-flow sampling is
preferred.
455
The low-flow sampling method, typically used in 2-
inch diameter groundwater monitoring wells, relates to
the amount of drawdown in a well during purging. The
indicator parameters are collected before and during
purging and allowed to stabilize prior to groundwater
sample collection. Purge rates may be higher than
sample rates to maximize purge efficiency. Prior to the
collection of the groundwater sample, the pump rate
may be lowered to lower sample turbidity and entrained
air in the sample, and to mimic conditions that may
exist in the natural aquifer. High turbidity in water
samples being analyzed for metals requiores filtering
to remove colloidaland suspended sediments that may
contain metals.
Discrete sampling systems are designed to collect
groundwater samples form the formation at a specific
depth or multiple target depths without significant
purging prior to sample collection.
GROUNDWATER SAMPLING PROTOCOL
Depth to the top of the water and free-product thickness
within wells are monitored by using a water level meter or
an interface probe. Water level data collected from areas
that have multiple wells are used to develop groundwater
contour maps for the project site. Groundwater flow is
perpendicular to equipotential lines drawn on the map. If
free-product is detected, a product sample is sometimes
collected for source identification. In environmental
projects, where several chemicals are to be analyzed for
a given well, individual samples are collected in order
of decreasing volatility. When the results from previous
sampling are known, it is recommended to start sampling
the cleanest wells first, moving to the most contaminated
wells later in the sampling event.
Each chemical analysis requires specific sample preser-
vation techniques or additives. Sealed chemical ice is
placed in the coolers to maintain samples at a temper-
ature of 4 ◦ C. Where several types of chemicals are to be
analyzed for a given well, individual samples are collected
in the following order: volatile organic compounds, purge-
able organic compounds, purgeable organic halogens, total
organic compounds, total organic halogens, extractable
organic compounds, total metals, dissolved metals, phe-
nols, and cyanide.
QUALITY CONTROL OF GROUNDWATER SAMPLES
A QC program, which is independent of the laboratory’s
program, verifies sample integrity and includes the
submittal of duplicates, field blanks, and travel blanks to
the laboratory. The QC samples are packaged and sealed
in the same manner as the other samples and are assigned
independent sample numbers.
A duplicate sample is collected for about 5% of the
samples or one per sampling round, whichever is greater.
The duplicate sample is acquired by filling separate
containers from the same well bailer as the actual sample.
The contents of the bailer are evenly divided between
the actual and duplicate samples to ensure duplication.
The field blank is submitted to the laboratory for the
456 GROUNDWATER SAMPLING WITH PASSIVE DIFFUSION SAMPLERS
same analyses as the rest of the sampling set. The
field blank is acquired by dispensing deionized water
from a clean sampling bailer into the containers in the
same manner as groundwater samples. When sampling
groundwater for analyzing volatile compounds, travel
blanks are used to detect the introduction of contaminants
during transportation from the field to the laboratory. The
travel blank generally contains deionized water, which
has been poured into the container prior to being taken
into the field.
BIBLIOGRAPHY
1. Testa, S.M. (1994). Geologic Aspects of Hazardous Waste
Management. CRC Press, Boca Raton, FL, pp. 145–187.
2. Jacobs, J. (2000). Monitoring well construction and sampling
techniques. In: Standard Handbook of Environmental Science,
Health, and Technology. J. Lehr (Ed.). McGraw-Hill, New
York, pp. 11.46–11.68.
3. American Society for Testing Materials. (1984). Standard
Practice for Description and Identification of Soils (Visual-
Manual Procedure), Method D 2488-84. Philadelphia, PA.
4. California Regional Water Quality Control Board. (1989).
Leaking Underground Fuel Tank (LUFT) Field Manual:
Guidelines for Site Assessment, Cleanup, and Underground
Storage Tank Closure, Sacramento, CA.
5. Driscoll, F.G. (1986). Groundwater and Wells, 2nd Edn.
Johnson Filtration Systems, St. Paul, MN, pp. 268–339.
6. Sisk, S.J. (1981). NEIC Manual for Groundwater/Subsurface
Investigations at Hazardous Waste Sites. EPA-330/9-81-002.
U.S. Environmental Protection Agency, Denver, CO.
7. Puls, R.W. and Barcelona, M.J. (1996). Low-flow (minimal
drawdown) ground water sampling procedures. EPA Ground-
water Issue, April.
GROUNDWATER SAMPLING WITH PASSIVE
DIFFUSION SAMPLERS
CRAIG E. DIVINE
Colorado School of Mines
Golden, Colorado
INTRODUCTION: DIFFUSION-BASED GROUNDWATER
SAMPLING METHODS
Passive diffusion-based sampling methods have been used
to monitor environmental contaminants, particularly trace
metals and certain lipophilic compounds including poly-
chlorinated biphenyls (PCBs), polynuclear aromatic hydro-
carbons (PAHs), and pesticides, for many years. In particu-
lar, dosimeter-type and semipermeable membrane devices
(SPMDs) are commonly used in limnology, oceanography,
and ecotoxicology for evaluating ultralow concentrations
and potential bioavailability of contaminants (1–4). These
types of samplers are most frequently used for sam-
pling sediment porewater and surface water (i.e., rivers,
estuaries, harbors) and have occasionally been used for
groundwater characterization in contaminant toxicology
studies (5). These types of sampling devices are composed
of a ‘‘trapping’’ phase, such as hexane, activated carbon,
or triolein (a prominent fish lipid), contained within a
semipermeable membrane (such as dialysis-type cellulose
or a synthetic polymer). On deployment, contaminants dif-
fuse through the membrane and are subsequently retained
in the trapping phase until analysis. After the SPMD is
removed, the contaminants must typically be extracted
from the trapping phase before analysis. A significant dis-
advantage with SPMDs and dosimeter-type devices for
groundwater sampling is that back-calculation to actual
in situ aqueous concentrations may not be straightforward
due to nonlinear and/or nonequilibrium sorption behavior
of the trapping phase (5).
Passive diffusion (PD) samplers are similar to
dosimeter-type and SPMD sampling methods in that they
generally consist of a sealed container with a semiper-
meable membrane. The PD sampler is suspended in
a monitoring well, and the solutes of interest diffuse
across the membrane into the sampler. However, unlike
dosimeter-type devices and SPMDs, PD samplers do not
contain a trapping phase but are instead filled with water
or air. Therefore, once the PD sampler is retrieved from
the well (after equilibration), the concentration measured
in the sampler is assumed to relate directly to in situ
analyte concentrations at the deployment location. Pas-
sive diffusion samplers are less expensive to construct
than SPMDs and easier to analyze because no extraction
step is required. Furthermore, the correlation between
the contaminant concentration measured in the PD sam-
pler and the surrounding environment is straightforward:
for water-filled PD samplers deployed in groundwater the
concentrations are equivalent.
Passive diffusion-type sampling methods have been
used in a variety of applications for many years,
including dissolved gas measurement in seawater and
groundwater (3,6,7). Generally, PD samplers offer several
potential advantages for monitoring many common
groundwater contaminants, compared to conventional
sampling techniques. For sites with ongoing monitoring
programs, significant cost savings may be realized due to
reduced sampling-related labor and minimal sampling-
related waste generation. For dissolved constituents
with large Henry’s Law constants, such as volatile
organic compounds (VOCs) and dissolved gases, analyte
loss due to volatilization can be greatly minimized.
In some cases, analytical matrix interferences may be
reduced if the membrane is selective against interfering
constituents (i.e., the matrix in the PD sampler may be
comparatively less complex). Additionally, multiple PD
samplers can be deployed to provide solute concentration
measurements across very discrete intervals. For example,
PD samplers with dialysis-type membranes have been
used to characterize ion and trace metal concentration
profiles on the scale of a few centimeters in sediment (3).
MEMBRANE MATERIALS AND SPECIFIC CONSTRUCTION
EXAMPLES
Because of the potentially significant costs savings and
technical advantages noted above, the use of PD samplers
for groundwater monitoring at contaminated sites is
greatly increasing [particularly water-filled PD samplers
using low-density polyethylene (LDPE) membranes].
 GROUNDWATER SAMPLING WITH PASSIVE DIFFUSION SAMPLERS
(a) Grommet
LDPE membrane
(0.010 cm thick)
~0.6 m
Diameter = ~3.2 cm
Lay-flat LDPE tubing
filled with ~220 mL
deionized H2O
Grommet
(b)
LDPE membrane
Open mouth
Rubber band
Threads (seal gasket)
Glass vial
(gas-or water-filled)
Laboratory and field compatibility tests indicate that
LDPE membranes are suitable for most fuel and solvent
VOCs; however, it has been demonstrated that the
LDPE membrane is generally not suitable for measuring
inorganic ions, metals, pesticides, explosives, semivolatile
organics, alcohols, organic acids, and ethers [including
methyl-t-butyl ether (MtBE)] with PD samplers (8,9).
Dissolved gases, such as He, Ne, H2 , O2 , and N2 ,
can be measured by PD samplers with LDPE (10) or
silicone (7) membranes. Passive diffusion-type samplers
(also known as dialysis samplers or ‘‘peepers’’) that use
other membrane materials, such as cellulose acetate, can
be used to monitor major ions, metals, nutrients, dissolved
gases, and pH (3,11).
Passive diffusion samplers can be constructed easily
from inexpensive materials. As shown in Fig. 1a, the open
mouth of a glass laboratory vial can be covered with a
membrane and sealed with a rubber band gasket and
nylon zip tie. Drop sheeting, which is widely available in
the painting supply sections of hardware stores, can be
used for LDPE membranes and is offered in a range
of thicknesses. PD samplers filled with water can be
constructed while submerged in deionized water to prevent
entrapment of air bubbles. Passive diffusion samplers
457
Tig
hte
n to
sea
l
Nylon cable tie
LDPE membrane
Figure 1. Schematics of passive diffusion
samplers.
made of lay-flat LDPE tubing filled with deionized water
are commercially available from EON Products, Inc.
(Snellville, Georgia) and Columbia Analytical, Inc. (Kelso,
Washington). A schematic of this type of sampler, which
is also known as a PD ‘‘bag’’ sampler is shown in Fig. 1b.
For a description of peepers and other PD samplers with
dialysis-type membranes, see Davison et al. (3).
EQUILIBRATION TIME FOR PASSIVE DIFFUSION
SAMPLERS
To prevent biased-low measurements, PD samplers must
be deployed long enough in a well for equilibration to
occur. Based on empirical laboratory and field studies,
a minimum of 2–3 days is required for equilibration
for most VOCs using commercially available LDPE
PD bag samplers (8). Carnigan (12) found that some
dialysis-type samplers require equilibration times of 3–20
days for measuring inorganics in porewater. Based on
these general results, a minimum equilibration time of
approximately 2 weeks is commonly recommended as a
conservative approach in field studies (8,9). However, note
that this is a generic rule of thumb. As equilibration is a
diffusion process, the specific equilibration time depends
458 GROUNDWATER SAMPLING WITH PASSIVE DIFFUSION SAMPLERS

on sampler volume, membrane surface area, membrane
thickness, and the diffusion coefficient of the solute in the
membrane material.
Sanford et al. (2) present the following solution that
describes the concentration within the PD sampler as it
equilibrates with the surrounding groundwater:
  
−Dm At
Cs (t) = Cr H 1 − e ∀Lm  (1)
where
Cs (t) concentration within the sampler at time t after
PD sampler deployment [M L−3 ],
Cr resident concentration at the measurement location
[M L−3 ],
Dm effective constituent diffusion coefficient for the
membrane material [L2 T−1 ],
H solute dimensionless Henry’s law constant [ ],
A surface area of the PD sampler membrane [L2 ],
t deployment time [T],
∀ internal volume of the PD sample [L3 ], and
Lm membrane thickness [L].
This solution assumes Fickian solute diffusion across
a thin membrane, a constant Cr concentration at the
measurement location, and a well-mixed reservoir in
the PD sampler. As presented, Equation 1 incorporates
air–water partitioning and is applicable to gas-filled PD
samplers deployed in groundwater, such as those described
by Sanford et al. (7) and Divine and McCray (10). For
conditions where no phase partitioning occurs (such as
a water-filled PD sampler deployed in groundwater),
Henry’s law constant (i.e., the gas-water partition
coefficient) is treated as unity.
PD samplers, including both constructed (Fig. 1a) and
commercially available (Fig. 1b) samplers. The specific
parameter values and resulting equilibration times are
summarized in Table 1, and the predicted PD sampler
concentration histories are shown in Fig. 2.
These results show the relative importance of the
various parameters for equilibration times. For example,
increasing the A/∀ ratio reduces the equilibration time,
and this can be accomplished for constructed PD samplers
(Fig. 1a) by filling the sampler with inert sand or glass
beads (i.e., compare results of Scenario 3 with Scenario
4). It is also important to note that Eq. 1 is valid only for
advection-dominated systems. If the groundwater velocity
is very low (diffusion-dominated systems), a concentration
gradient will develop outside of the sampler, and
equilibration times can be much longer (13). Equation 1 is
appropriate, however, for most shallow aquifer systems.
CASE STUDY: APPLICATION OF PASSIVE DIFFUSIONS
SAMPLERS AT A HETEROGENEOUS SITE
Passive diffusion samplers measure water quality directly
in contact with the membrane surface; therefore, they
generally measure concentrations at relatively discrete
intervals within a well. Conversely, samples collected by
conventional purge-and-bail methods tend to integrate
water quality across the entire well screen region and
therefore are most significantly influenced by water
quality conditions in the highest permeable zones across
the well screen interval. In other words, PD samplers
tend to measure porewater concentrations over relatively
small intervals, and purge-and-bail type methods tend
to measure flux-weighted concentrations. For vertically
uniform plume distributions and homogeneous aquifer
1.2
Normalized PD sampler

As clearly shown in Eq. 1, the rate at which a PD 1.0

concentration

sampler equilibrates relates directly to the Dm value

0.8
of the solute, the sampler dimensions (A, ∀), and the Scenario 1
thickness of the membrane (Lm ). Therefore, PD samplers 0.6
Scenario 2
can be constructed to minimize equilibration times by 0.4
using a thin membrane and increasing the A/∀ ratio. Scenario 3
0.2
The effective Dm value is temperature-dependent and Scenario 4
specific to both the membrane and solute. Divine and 0.0
0 1 2 3 4 5 6 7
McCray (10) estimated that a practical range is 1–4 × Deployment time, day
10−7 cm2 s−1 for Dm values of several dissolved gases and Figure 2. Passive diffusion sampler concentrations after deploy-
VOCs for LDPE membranes. Sanford et al. (2) measured ment, predicted by Equation 1. The normalized concentration is
Dm values for helium and neon of 4 × 10−7 cm2 s−1 defined as Cs /Cr (at equilibrium Cs = Cr ; therefore, Cs /Cr = 1.0).
for silicone membranes. Based on these Dm ranges, The parameter values used in the models are summarized in
equilibration times are predicted using Eq. 1 for several Table 1.

Table 1. Passive Diffusion Sampler Dimensions, Membrane Thickness, Membrane Diffusion Coefficients, and Predicted
Equilibration Times for Various Constructed (Fig. 1a) and Commercially Available (Fig. 1b) PD Samplers. The
Equilibration Time is Defined by Equation 1 when Cs ≥ 0.95Cr )
A, ∀, Lm , Dm , Equilibration
cm2 mL cm (cm2 s−1 ) Time, Day

Scenario 1 Commercial ‘‘bag’’ sampler 350 220 0.010 4.0 × 10−7 0.7
Scenario 2 Commercial ‘‘bag’’ sampler 350 220 0.010 2.5 × 10−7 2.7
Scenario 3 Constructed sampler: sand-filled 20-mL glass vial 3.2 12.9 0.005 2.5 × 10−7 2.8
Scenario 4 Constructed sampler: 20-mL glass vial 3.2 23 0.005 2.5 × 10−7 5.1
 GROUNDWATER SAMPLING WITH PASSIVE DIFFUSION SAMPLERS 459

1,1-DCE concentration, µg/L

Lithology
0 200 400 600 800 1000 1200 1400 1600 1800 2000
16 Weathered
110 sandstone
17 110 1300
Sandstone/
18 claystone
19 140 380
Sandstone
Depth, ft bls

20
21 910 1500
Claystone,
22
1100 unfractured
23 1600
Purge-and-bail
Sandstone
24
Low-flow sampling
25
Figure 3. Comparison of sampling results for
Passive diffusion sampler Claystone, few
fractures
purge-and-bail, low-flow, and passive diffusion sam-
26 pling methods for a single monitoring well.

conditions, results from these two sampling methods may 10000

be equivalent. However, these two methods may yield
significantly different concentration values for a specific
well at a heterogeneous site with plume stratification.
To illustrate this potential behavior, the concentrations Purge-and-bail method, µg/L
1000
of 1,1-dichloroethene (1,1-DCE) (a common VOC) mea-
sured at a monitoring well by purge-and-bail, low-flow 1:1 line
[another discrete-type sampling technique; see Puls and
Barcelona (14)]. PD sampling methods are shown in Fig. 3.
100
The aquifer at this study site is highly heterogeneous, and
this is clearly expressed in the results of the various sam-
pling methods. Only one value (1300 µg/L) is obtained by
the purge-and-bail method; however, the low-flow and PD
sampling methods characterize concentrations at various 10
discrete depth intervals and show significant contaminant Not detected
stratification (approximately an order of magnitude over by purge-and-bail
several vertical feet).
Figure 4 compares the results of purge-and-bail with 1
1 10 100 1000 10000
PD sampling methods at several wells at the same
Passive diffusion sampler method, µg/L
heterogeneous site. The purge-and-bail measurements are
generally higher than the average discrete PD sampler Figure 4. Comparison of sampling results (various VOCs) for
purge-and-bail and passive diffusion sampling methods for a
measurements, possibly indicating that the majority of the
series of wells at a heterogeneous site.
contaminant mass exists in highly permeable zones. These
data provide additional characterization information and
have potentially important implications for the site. At measurement within site objectives or regulatory require-
some sites, a similar targeted dual-method sampling ments may be challenging. For example, an analytical
strategy (i.e., use of both purge-and-bail and PD sampling water quality goal or regulatory limit may be met by one
methods) may be useful for evaluating contaminant method and not by another at the same well simply due
distribution, remedial system design and performance, to the inherent averaging characteristics of the sampling
and long-term plume behavior. methods. For additional discussion of the results of this
Clearly, using PD samplers under heterogeneous study, see Divine et al. (15).
aquifer conditions, particularly in an ongoing compli-
ance monitoring program, requires careful consideration BIBLIOGRAPHY
of ways to interpret measurement data and satisfy mon-
1. Sordergren, A. (1987). Solvent-filled dialysis membranes
itoring program objectives. For example, determining the
simulate uptake of pollutants by aquatic organisms. Environ.
appropriate intervals for PD sampler placement may be
Sci. Technol. 21(9): 855–859.
challenging (i.e., Fig. 3). Additionally, as indicated by the
2. DiGiano, F.A., Elliot, D., and Leith, D. (1988). Application of
circled values in Fig. 4, the PD sampling method mea-
passive dosimetery to the detection of trace organic contami-
sured low concentrations of some nontarget VOCs at a few nants in water. Environ. Sci. Technol. 22(11): 1365–1367.
discrete intervals in certain wells, but these compounds 3. Davidson, W., Fones, G., Harper, M., Teasdale, P., and
were not measured above the analytical quantitation limit Zhang, H. (2000). Dialysis, DET, and DGT: In situ diffusional
by the purge-and-bail method. Assuming that these dif- techniques for studying water, sediments, and soils. In: In
ferences are related to the differing averaging natures of Situ Monitoring of Aquatic Systems. J. Buffle and G. Horvai
the sampling methods, interpreting the discrete interval (Eds.). John Wiley & Sons, New York, Chapter 11.
460 SPECIFIC CAPACITY
4. Dunn, R.K., Teasdale, P.R., Warnken, J., and Schleich, R.R.
(2003). Evaluation of diffusive gradient in a thin film
technique for monitoring trace metal concentration in
estuarine waters. Environ. Sci. Technol. 37(12): 2794–2800.
5. Gustavson, K.E. and Harkine, J.M. (2000). Comparison of
sampling techniques and evaluation of semipermeable
membrane devices (SPMDs) for monitoring polynuclear
aromatic hydrocarbons (PAHs) in groundwater. Environ. Sci.
Technol. 34(20): 4445–4451.
6. Dyck, W. and Da Silva, F.G. (1981). The use of ping-pong
balls and latex tubing for sampling the helium content of lake
sediments. J. Geochem. Exploration 14: 41–48.
7. Sanford, W.E., Shropshire, R.G., and Solomon, D.K. (1996).
Dissolved gas tracers in groundwater: Simplified injection,
sampling, and analysis. Water Resour. Res. 32(6): 1635–1642.
8. Vroblesky, D.A. and Campbell, T.R. (2001). Equilibration
times, compound selectivity, and stability of diffusion
samplers for collection of ground-water VOC concentrations.
Adv. Environ. Res. 5: 1–12.
9. Vroblesky, D.A. (2001). User’s Guide for Polyethylene-Based
Passive Diffusion Bag Samplers to Obtain Volatile Organic
Compound Concentrations in Wells, Part 1: Deployment,
Recovery, Data Interpretation, and Quality Control and
Assurance. United States Geological Survey, Columbia, SC.
10. Divine, C.E. and McCray, J.E. (2004). Estimation of diffusion
coefficients and equilibration times for LDPE passive
diffusion samplers. Environ. Sci. Technol. 38(6): 1849–1857.
11. Webster, I.T., Teasdale, P.R., and Grigg, N.J. (1998). The-
oretical and experimental analysis of peeper equilibration
dynamics. Environ. Sci. Technol. 32(11): 1727–1733.
12. Carnigan, R. (1984). Interstitial water sampling by dialysis,
methodological notes. Limnol. Oceanogr. 29: 667–670.
13. Harrington, G.A., Cook, P.G., and Robinson, N.I. (2000).
Equilibration times of gas-field diffusion samplers in slow-
moving ground water systems. Ground Water Monitoring
Remediation 20(2): 60.
14. Puls, R.W. and Barcelona, M.J. (1996). Low-Flow (minimal
Drawdown) Ground-Water Sampling Procedures. EPA/540/S-
95/504, Ada, OK.
15. Divine, C.E., Madsen, L., Andrews, S.D., and Santangelo-
Drieling, T. (in press). Passive diffusion ground water
samplers at a heterogeneous site: Pilot study results. Ground
Water Monitoring and Remediation.
SPECIFIC CAPACITY
ERIC S. WILSON
E. L. Montgomery & Associates,
Inc.
Tucson, Arizona
The specific capacity of a well is the pumping rate of the
well divided by the drawdown at that rate:
Q
Sc =
h
where Sc =specific capacity
Q=constant pumping rate, and
h=drawdown
Water well contractors often report specific capacity to
characterize the performance of a well and use specific
capacity to estimate the optimum depth for installing a
pump to obtain the desired yield. Units typically associated
with specific capacity are liters per second per meter
of drawdown (l/s/m), cubic meters per day per meter
of drawdown (m3 /day/m), cubic feet per day per foot of
drawdown (ft3 /day/ft), or gallons per minute per foot of
drawdown (gpm/ft).
When taking measurements to compute specific capac-
ity, the pumping rate should be constant, and the period
of pumping should be sufficiently long that the rate of
change in drawdown is small. The specific capacity of
a well tends to become smaller as the pumping rate and
duration of pumping increase. At high pumping rates, non-
laminar flow losses are larger and result in an increased
rate of drawdown. Water level drawdown in unconfined
aquifers results in reduced aquifer transmissivity near
pumping wells; larger drawdown at high pumping rates
then results in progressively smaller transmissivity and
smaller specific capacity. A specific capacity value remark-
ably smaller than expected may result when the well is
not screened in the correct water-bearing zone or when
well screen openings are clogged due to incomplete well
development, precipitation of minerals in the screen open-
ings, or biofouling. After many years of pumping, if the
specific capacity declines to about half the original value,
well rehabilitation should be considered.
Specific capacity may be used to estimate the maximum
potential pumping rate of a well by multiplying the
available drawdown by specific capacity. The available
drawdown is the difference between the depth at which
the pump is set and the depth to the nonpumping water
level in the well.
Specific capacity may be used to estimate the
transmissivity of aquifer media in the vicinity of the
pumping well. Theis (1) suggested that specific capacity
could be used to estimate the transmissivity of an aquifer
using the equation,
Q 2.3 2.25T ∗ t
T∗ = log
h 4π r2 S
Q
where =specific capacity
h
t=time
r=radius of pumping well
S=aquifer storativity
T ∗ =transmissivity; an initial guess of the value
of T is made and substituted in the
equation. The equation is then solved for
h. After the initial solution, the value of T
must then be adjusted until the calculated
value of h is reasonably close to the
measured value.
Driscoll (2) indicated that transmissivity could be
estimated as
T = Sc × 2, 000 for confined aquifers
or
T = Sc × 1, 500 for unconfined aquifers
 SOIL WATER 461

where T is in gpd/ft and where the range in water content is from a minimum,
Sc is specific capacity. or residual, saturation, θr , to a maximum equal to the
saturated porosity, θs , so that θr < θ < θs .
Razack and Huntley (3) gave an empirical relationship The water content can also be defined with a
between transmissivity and specific capacity as follows: gravimetric, or mass, measurement, θg , such that

T = 15.3(Sc )0.67 Mw
θg = (2)
2
Ms
where T is in m /day
Sc in m3 /day/m, where Mw is the water mass and Ms is the soil mass. We
can readily convert between these two measures by using:
or
T = 33.6(Sc )0.67
ρw θg
θ = θg = (3)
where T is in ft2 /day ρs ρs
Sc in ft3 /day/ft.
where ρw = Mw /Vw = 1 kg/L is the water density and
Using specific capacity data to estimate transmissivity ρs = Ms /Vs (kg/L) is the bulk density of the soil.
can be useful when trying to estimate potential pumping The soil water content can also be reported as a relative
rates or estimating how much available head is needed to saturation, , which is defined using
produce a specified pumping rate. But, as in any estimation
technique, it is not always applicable or advisable. θ − θr
= (4)
These instances may include situations where substantial θs − θr
drawdown occurs because the pumping duration is too
short for the specific capacity test, too small a pumping where θr and θs are the residual and saturated water
rate is used for an adequate specific capacity test, or well contents, respectively. The relative saturation varies from
construction is inefficient. zero for a dry soil to a maximum of one for a fully saturated
soil, 0 < < 1.

BIBLIOGRAPHY
MEASURING WATER CONTENTS
1. Theis, C.V. (1963). Estimating the Transmissibility of Aquifers
from the Specific Capacity of Wells. In: Methods of Determining Water contents are commonly measured with time-domain
Permeability, Transmissibility, and Drawdown. R. Bentall, reflectometry (TDR) devices, as shown in (Fig. 1). TDRs
compiler, U.S. Geological Survey Water-Supply Paper 1536-I,
determine the water content with the electromagnetic
pp. 331–340.
properties of a wave pulse passing through a conducting
2. Driscoll, F.G. (1986). Groundwater and Wells, 2nd Edn.
set of rods (such as 3-mm stainless-steel welding rods)
Johnson Filtration Systems, Inc., St. Paul, MN.
placed in the soil.
3. Razack, M. and Huntley, D. (1991). Assessing transmissivity
The principle of TDRs is that the velocity of electro-
from specific capacity in a large and heterogeneous alluvial
aquifer. J. Ground Water 29(6): 856–861.
magnetic waves along a conductor is a function of the
dielectric coefficient of the media around the conductor.
Liquid water has a dielectric constant of 80.2 at 20 ◦ C,
SOIL WATER whereas ice is 3.2, petroleum is 1.8 to 2.2, quartz is 4.3,
and air is only 1.00.
TODD RASMUSSEN
The University of Georgia
Athens, Georgia Exterior
shield
Coaxial cable

Soil water is a measure of the quantity of water present

within a unit volume of soil. Soil-water content can vary Block
Conductor
from fully saturated, in which case the water content
equals the soil porosity, to oven-dry, in which case all free
water has been evaporated. Soil
TDR rods
Wet soils are characterized as having high water
contents, which means that more soil pores are filled
with water, whereas dry soils have more pores that are
filled with air.
The volumetric water content, θ , is defined as the Figure 1. Time-domain reflectometer (TDR) probe that mea-
volume of water, Vw , per unit volume of soil, Vs : sures soil-water content. The three-rod assembly is placed within
the soil. Coaxial cable connects probe to the water content tester.
Vw Center wire of coaxial cable connects to the central rod, whereas
θ= (1) the exterior shield is connected to the two outer rods.
Vs
462 SOIL WATER

As the water dielectric constant is so much greater for
water than for soil or air, the wave velocity is substantially
retarded as the water content of a soil increases, because
larger dielectric constants cause slower wave velocities
and, hence, longer travel times.
Before the use of TDRs, neutron probes used the
thermalization of neutrons to determine the water
content (1). This older method has fallen out of favor,
however, because of the rigorous licensing and inspection
requirements associated with radioactive materials.
MATRIC TENSION
Water molecules in the unsaturated zone are bound by
adhesive forces to soil surfaces and by cohesive forces to
other water molecules, which are surface tension forces,
and they result in fluid pressures that are less than
atmospheric.
Rather than use negative pressure heads, p < 0, to
describe the soil potential, we employ soil, or matric,
tensions, ψ > 0. The matric tension is the negative fluid
pressure head, ψ = −p, which increases as the pressure
head becomes more negative (more tightly bound) and
decreases as the pressure head becomes less negative (less
tightly bound).
The total head within the soil pore fluids, h, is the
difference between the elevation, z, and the matric tension:
H =z+p=z−ψ (5)
Capillary theory can relate the matric tension, ψ, to the
fluid potential associated with the radius of curvature, r,
of fluid films within pores:
2σ cos α
ψ= (6)
γr
where σ (Pa·m) is the surface tension and α is the contact
angle at the air–water–solid interface.
The matric tension of a wet soil, 0 < ψ < 0.7 bar, can be
measured with tensiometers, as shown in (Fig. 2), which
determine the negative pressure head of fluids inside a
probe. The fluid within the probe is allowed to reach
Vacuum gage
equilibrium with the soil fluids using a semiporous ceramic
or metal filter that prevents air from entering the probe.
Psychrometry is a useful technique in drier soils,
ψ > 5 bar (2). The dewpoint temperature of the pore atmo-
sphere is measured with a thermocouple psychrometer.
The dewpoint temperature can then, along with the pore
temperature, determine the relative humidity of the pore
atmosphere. The matric tension can be readily determined
from the relative humidity, as shown in Table 1.
MOISTURE CHARACTERISTIC CURVES
The amount of water contained within soil can be related
to the fluid tension, or negative pressure, present within
the water. Dry soils contain water that is held by strong
adhesive forces to soil surfaces along with strong cohesive
forces to other water molecules. These forces are reflected
in the tension that holds the water against the force of
gravity.
Each soil has a unique relationship between soil tension
and water content. This relationship, termed the soil
moisture characteristic curve, is a function that describes
how water content and matric tension are related. The
characteristic curve for each material varies with the pore
size distribution—a soil with larger pores releases water
at a lower tension than does a soil with smaller pores.
A number of formulations have been proposed that
relate the relative saturation, , to matric tension, ψ. The
Brooks and Corey equation has the form:

−λ
ψ
= (7)
ψb
where ψb is the bubbling pressure and λ is a pore-
size distribution index. This equation only holds for
ψ > ψo − = 1 otherwise.
Another common formulation is the van Genuchten
equation:
 (1−n)/n
= 1 + (αψ)n (8)
where α and n are empirical coefficients.
Moisture characteristic curves are commonly estimated
with pressure plates (shown in (Fig. 3)) or Tempe pressure
cells. In both methods, one or more soil samples are placed
on top of a porous plate within a sealed chamber. Nitrogen
gas under high pressure is forced into the chamber, and
water is forced out of the soil samples through the porous

Table 1. Measures of Soil Tension

Water-filled tube Fluid Pressure Soil Tension Relative
(bars) (m) Pore Radius Humidity (%)
Soil
0 0 ∞ 100
0.1 −1 15 µm 99.993
Porous cup 1 −10 1.5 µm 99.93
15 −150 100 nm 98.9
100 −1000 15 nm 93.
Figure 2. Soil-water tensiometer probe that monitors soil- 500 −5,000 3 nm 70.
tension. A porous cup allows water to exchange between the 1,000 −10,000 1.5 nm 48.
soil and a reservoir of water within the tube assembly. A vacuum 5,000 −50,000 3 Å 2.6
pressure gauge monitors the water pressure inside the water 10,000 −100,000 1.5 Å 0.068
reservoir.
 SOIL AND GROUNDWATER GEOCHEMISTRY AND MICROBIOLOGY
Pressure chamber
Gas inflow
Soil samples Porous plate
Water outflow
Figure 3. Pressure plate procedure for determining the soil
moisture characteristic curve. A gas pressure is applied to soil
samples within the chamber, and water is released through the
porous plate. The water content, θ, is found for each pressure, ψ.
plate, which is permeable to water but not to air. The
resulting water content of the samples can be determined
once the outflow has stopped.
SAMPLING SOIL WATER
Soil water samples are more difficult to collect in the
unsaturated zone than in the saturated zone because
water does not freely flow into a well or piezometer.
Commonly, a soil lysimeter collects water samples by
placing a suction within a hollow tube that has a porous
cup placed within the soil. Water flows through the porous
cup into the hollow tube where it can then be collected.
Soil lysimeters are limited to conditions where the soils
are wet, ψ < 0.7 bar. For drier conditions, soil samples
can be collected and returned to the laboratory, where
mechanical or gas pressure can be applied to the samples
to force pore fluids out of the sample.
BIBLIOGRAPHY
1. Elder, A.N. and Rasmussen, T.C. (1994). Neutron probe
calibration in unsaturated tuff. Soil Sci. Soc. Am. J. 58(5):
1301–1307.
2. Rasmussen, T.C. and Rhodes, S.C. (1994). Energy-related
methods: Psychrometers. In: Vadose Zone Characterization
and Monitoring; Principles, Methods and Case Studies.
L.G. Wilson, L. Everett, and S. Cullen (Eds.). Lewis Publish-
ers, Boca Raton, FL.
SOIL AND GROUNDWATER GEOCHEMISTRY
AND MICROBIOLOGY
STUART SMITH
Smith-Comeskey GroundWater
Science LLC
Upper Sandusky, Ohio
DEFINITIONS
Geochemistry can be defined as the study of the chemical
composition of, and of actual or possible chemical changes
463
in, a geological matrix, including soil, that complex
mixture of minerals, life, and organic material on the
surface of the earth. Groundwater, of course, is water
located in the saturated zone of the earth. Microbiology
is the study of small forms of life, usually including the
bacteria and archaea domains and protists and fungi of the
eukaryaceae domain, and naturally also covers numerous
subcategories.
The interdisciplinary studies of the interactions among
microflora, earth materials, and chemical reactions are
often linked under geomicrobiology or biogeochemistry.
ORIGINS
The relationship between life in soil and earth materials
appears to go back to the very origins of life on
earth. Bacterial and archaeans (superficially bacteria-like
microflora that have entirely different genetic operations,
biochemistry, and cell membranes) were interacting with
iron and other minerals as part of energy transformations
since within a few 100 million years of the coalescence of
the planet. Many archaeans specialize in methanogenesis
and adaptation to extreme environments at the present
time, such as geothermal waters, undersea hydrothermal
vents, and hypersaline lakes (e.g., Dead Sea, Great
Salt Lake, and solar salt lagoons). Biochemical evidence
suggests that they may have developed in association
with such conditions, which were common early in
earth history.
Bacteria and cyanobacteria (colloquially known as blue-
green algae) strongly affected the earth environment
from early in the Archean era (±4 billion years before
present, b.y. BP). Cyanobacteria, with a photosynthetic
mechanism similar to algae and plants, used carbon
dioxide and expended oxygen as a waste product in
shallow anoxic seas. Ferrous iron (Fe), common in water
and the earth’s crust, scavenged the oxygen, forming
various FeIII minerals, including magnetite, hematite
(recently confirmed on Mars), goethite, and less-organized
mineral forms. Such iron-containing Archean sediments
were deposited as banded iron formations, with iron
bands in a cherty matrix, which are important iron ores
worldwide where they occur (such as the Mesabi Range in
Minnesota). When much of the available free oceanic FeII
was depleted, oxygen began to build up in the atmosphere,
with a recognizably modern atmosphere dating from 1.7
b.y. BP.
Bacteria have the most diverse biochemical and
respiratory capabilities among forms of life. Although
others require oxygen for respiration, bacteria can respire
using sulfur, forms of nitrogen, carbon dioxide, and
metals (primarily Fe and Mn) as electron acceptors. Such
respiration is used to process a wide variety of organic
compounds, including recalcitrant ring structures. Such
versatility appears to be a very ancient capability.
In ancient shallow marine settings, cyanobacteria
formed stromatolites, laminations of organic and sedi-
mentary structures formed as sand and debris, and copre-
cipitated minerals were trapped in cyanobacterial masses.
Although many stromatolites are found as fossil struc-
tures, a number of locations exist on the modern-day Earth
464 SOIL AND GROUNDWATER GEOCHEMISTRY AND MICROBIOLOGY
where stromatolites are still forming, which are one form
of the common microbial phenomenon, biofilm formation.
From there on, microbial processes and mineral-
biological interactions, and the environment-manipulating
‘‘behavior’’ known as biofilm formation (all developed in the
Archean era), have continued in action until the present.
• Bog ores (important to the development of an iron
industry in the United States during the nation-
forming period) and gold and copper placer ores
are younger manifestations of the ancient microbial
practice of depositing metals.
• Bacteria reduce oxides of Fe, Mn, N, and S to accu-
mulate ‘‘reducing power’’ to oxidize (‘‘digest’’) complex
organic compounds. These capacities are exploited in
groundwater contamination remediation. The pres-
ence of high levels of Fe and Mn in well water samples
is also an indicator that organic carbon compounds
are present in groundwater accessed by a well.
• Microbe-mediated biogeochemical transformations
alter groundwater quality by oxidizing or reducing Fe
and Mn, and mobilizing them into solution. Recently,
the reduction of insoluble arsenic minerals to soluble
forms and As mobilization by microbes has attracted
interest as a factor in As presence in groundwater.
• Biofilm-forming microflora coat, clog, and corrode
engineered structures such as wells and drains in
the saturated zone, resulting in biofouling (further
discussed in the article BIOFOULING IN WATER WELLS).
SOIL MICROBIOLOGY
Soil microbiology is a special category of geomicrobiology
focused on the earth’s regolith, and specifically soil,
which is a complex of mineral, organic, and biological
components. Although beyond the scope of a short article,
soil microbiology has traditionally been distinguished
from groundwater or geomicrobiology. This distinction
has origins in academic discipline; however, studies
indicate that surficial soils have species richness and total
microbial biomass values that are much higher than in
earth materials below the soil zone. Protists and fungi are
common in soil but limited in type and numbers (but not
entirely absent) in aquifers.
Soil microbiology is much more thoroughly known com-
pared with that of less-accessible earth materials. It is
also much more thoroughly influenced by human activi-
ties, which include cultivation, fertilization, construction,
and application of contaminants (both deliberately and
accidentally). Soil microflora and soil structure experience
close interactions with animals and plants. Changes in
such macroflora and fauna communities can trigger sweep-
ing changes in soil microbiology. Clearing and cultivation
of forest and prairie soils result in simplification in micro-
bial community structure. The introduction of Eurasian
earthworms into North American boreal soils (where they
have been absent since the Pleistocene glaciations) has
resulted in measurable shifts in the availability of organic
matter and nutrients to native plants. In turn, changes in
vegetation affect soil and watershed hydrology.
OVERCOMING CHALLENGES IN MICROBIOLOGICAL
METHODS IN GROUNDWATER STUDIES
Background
Environmental microbiology and geomicrobiology are the
oldest disciplines within the broader field of microbiology.
However, industrial, public health, and clinical interests
quickly overshadowed the study of the microbiology
of natural environments. Until the 1970s, scientific
preconceptions in many circles, based on observations
of engineered filters, limited methods (and limited
reading) and placed limits on groundwater microbiology.
Considerations of the chemistry of groundwater assumed
that chemical transformations occurred in an environment
devoid of life. A notable exception was that of the
microbiologists in the Soviet petroleum industry of the
1950s and 1960s, who described the microbiology of
petroleum reservoirs and systems.
Analysis
Methods for the analysis of groundwater environments
for microbes were very limited prior to the 1980s. The
principal analytical method was culturing, which recovers
only a small fraction of bacteria. It was generally assumed
that water passing through the soil was purified by active
microbial processes and by filtration.
Interest in the occurrence of biofouling microflora in
groundwater and concerns about groundwater contam-
ination in the 1970s and 1980s motivated innovation
in methods suitable for groundwater microbiology. As
groundwater contamination became more and more evi-
dent in the United States and other western nations during
the 1980s, the motivation for understanding groundwa-
ter environments increased. In addition, new methods
in microbiology, based on advances in molecular biol-
ogy, provided microbiologists with new tools to explore
this difficult-to-sample microbial habitat. Among these
are phospholipid fatty acid methyl ester (PL-FAME) and
nucleic acid testing, described further in EVALUATION OF
MICROBIAL COMPONENTS OF BIOFOULING. Such methods provide
the means to describe microbial associations in a mean-
ingful way, as cultivation and phenotypic classification
typically fail to do so.
Classification, in particular, is a problem in description.
With micro-organisms, it often depends on making
comparisons with libraries of known genotypes or
phenotypes. Some deep subsurface communities are so
isolated and specifically adapted that comparisons with
described forms are weak. A general recognition exists in
the geobiological community that the task of describing
aquifer and other deep subsurface communities compares
with the task of describing the recently discovered deep-
sea hydrothermal vent ecosystems.
Sampling
Acquiring samples is perhaps the most difficult aspect of
the microbiology of the deep subsurface. Such sampling
is relatively expensive, at times technically challenging,
and prone to sample compromise. Typically, methods for
both solid and liquid sample acquisitions are adapted

from those developed for other purposes (borehole and
core drilling, pumped or zonal sampling). Only coring can
deliver solid samples isolated from drilling contamination.
A variety of water sampling techniques exist to obtain
discrete samples.
However, if properly interpreted, available sampling
and analytical techniques appear to be sufficient to gain
an understanding of deep subsurface microbial ecology.
NOTABLE DISCOVERIES
1. Microbial community diversity: PL-FAME and rRNA
analysis reveal subsurface communities that are
diverse and typically adapted to coping with
starvation and exploiting resource ‘‘windfalls’’ when
they occur. Bacteria can persist in viable but
nonculturable and inactive states when conditions
dictate. Microbial communities can have biomass
that rivals surface soils and may include fungi and
protozoa where pore sizes permit.
2. Microbial communities and numbers vary signifi-
cantly between hydraulically transmissive and con-
fining formations in the same wells and boreholes.
3. Community biochemical adaptability: Bacteria that
form biofilms and readily exchange genetic poten-
tial can be quite plastic in their ability to exploit
resources. Early in the effort to employ bioreme-
diation in treating contaminated groundwater and
soil, it was recognized that natural communities
were capable of degrading a variety of anthropogenic
compounds, especially hydrocarbons.
4. Iron and sulfur biogeochemical cycles are particu-
larly important in deeper systems, whereas nitrogen
is strongly cycled in the soil and shallow saturated
flow zones. In deeper systems, carbon and nitrogen
are cycled locally within biofilms.
5. Groundwater contamination locally alters oxidation-
reduction systems, nutrient cycling, and community
structure. Microbial communities react to influxes
of organic carbon by increasing biomass and tipping
toward anaerobic respiration as oxygen is depleted
by hydrocarbon-using or other aerobic heterotrophs,
which is closely analogous to eutrophication in
surface water systems. For energetic reasons, the
following are favored sequentially, switching to the
next when the electron acceptor is depleted: nitrate,
iron (and other metal) oxides, sulfate, and carbon
dioxide (reduced to methane).
6. Microbial communities may include very long-lived
and ancient lineages. Studies of communities in
deep, isolated systems reveal communities that
may be very ancient and existing in isolation from
the surface.
An article of this length is insufficient to consider
the subject comprehensively (considering that entire
textbooks, conferences, and hundreds of webpages are
devoted to the topic). Related subjects are considered
elsewhere in the Encyclopedia of Water.
CHARACTERIZING SOIL SPATIAL VARIABILITY 465
READING LIST
Ehrlich, H.L. (2002). Geomicrobiology, 4th Edn. Marcel Dekker,
New York.
Krumbein, W.E. (Ed.). (1983). Microbial Geochemistry. Blackwell
Scientific Publications, Oxford, UK.
Bitton, G. and Gerba, C.P. (Eds.). (1984). Groundwater Pollution
Microbiology. Wiley-Interscience, New York.
Sinclair, J.L. and Ghiorse, W.C. (1989). Distribution of aerobic
bacteria, protozoa, algae and fungi in deep subsurface
sediments. Geomicrobiology. 7(1–2): 15–32.
Chapelle, F.H. (1993). Ground-Water Microbiology and Geochem-
istry. John Wiley & Sons, New York.
Chapelle, F.H. (2000). Ground-Water Microbiology and Geochem-
istry, 2nd Edn. John Wiley & Sons, New York.
Madsen, E.L. and Ghiorse, W.C. (1993). Groundwater microbiol-
ogy: subsurface ecosystem processes. In: Aquatic Microbiology.
T.E. Ford (Ed.). Blackwell Scientific Publications, Boston, MA.
LeClerc, H. (2004). Microbiology of Natural Mineral Waters.
Centre Evian. Available: http://www.centre-evian.com.
Madsen, E. (2004). Bibliography of important works in envi-
ronmental microbiology. Available: http://www.micro.cornell.
edu/faculty/Madsen/Bibliography of important works
in Environmental Microbiology.htm.
Sylvia, D.M., Fuhrmann, J.J., Hartel, P.G., and Zuberer, D.A.
(2005). Principles and Applications of Soil Microbiology, 2nd
Edn. Prentice-Hall, New York.
CHARACTERIZING SOIL SPATIAL VARIABILITY
DENNIS L. CORWIN
SCOTT M. LESCH
USDA-ARS George E. Brown,
Jr., Salinity Laboratory
Riverside, California
SPATIAL VARIABILITY: DEFINITION, ORIGIN,
SIGNIFICANCE, AND MEASUREMENT METHODS
Spatial variability refers to the variation in soil physical,
chemical, mineralogical, and biological properties both
laterally across the landscape and vertically through
the soil profile. The spatial heterogeneity of soils is a
consequence of interacting meteorological, pedological,
and biological processes and anthropogenic factors over
time. Within-field spatial variability of soil properties
has been a primary concern within soil science ever
since the classic paper by Nielson et al. (1) concerning
the variability of field-measured soil water properties.
The significance of soil spatial variability stems from its
complex influence on any landscape-scale soil-related issue
including solute transport in the vadose zone, precision
agriculture, and soil quality assessment, to mention a few.
A variety of measurement methods are available for
potentially characterizing soil spatial variability includ-
ing ground penetrating radar (GPR), aerial photography,
multispectral and hyperspectral imagery, time-domain
reflectometry (TDR), and apparent soil electrical conduc-
tivity (ECa ). However, none of these approaches has been
as extensively investigated with respect to characteriza-
tion of within-field soil variability as ECa (2).
466 CHARACTERIZING SOIL SPATIAL VARIABILITY
Since its early agricultural use for measuring soil
salinity, the application of ECa has evolved into a widely
accepted means of establishing the spatial variability
of several soil physicochemical properties that influence
the ECa measurement. Geospatial measurements of ECa
are well suited for characterizing spatial variability for
several reasons. Geospatial measurements of ECa are
reliable, quick, and easy to take. The mobilization of ECa
measurement equipment is easy and can be accomplished
at a reasonable cost. Finally, and most importantly,
ECa is influenced by a variety of soil properties for
which the spatial variability of each could be potentially
established.
NEED FOR CHARACTERIZING SOIL SPATIAL VARIABILITY
The prospect of feeding a projected additional 3 billion
people over the next 30 years poses formidable, but not
insurmountable, challenges. Feeding the ever-increasing
world population will require a sustainable agricultural
system that can keep pace with population growth.
Sustainable agriculture hinges on a delicate balance
of economic sustainability, crop productivity, resource
utilization, and minimized detrimental environmental
impacts. These factors are influenced either directly or
indirectly by soil spatial variability.
Spatial variability within a field amounts to an order
of magnitude or more for soils (3,4) and to a factor of
2–4 or more for crops (5). Spatial variation in crops is the
result of a complex interaction of biological (e.g., pests,
earthworms, microbes), edaphic (e.g., salinity, organic
matter, nutrients, texture), anthropogenic (e.g., irrigation
efficiency, soil compaction due to farm equipment),
topographic (e.g., slope, elevation), and climatic (e.g.,
relative humidity, temperature, rainfall) factors. To a
varying extent from one field to the next, crop patterns
are influenced by soil-related properties. Conventional
farming is not as efficient as it could be because it
currently treats a field uniformly. To maintain uniform
crop production, conventional farming overcomes the
inherent variability of soil and crop conditions between
and within fields with an over application of inputs (i.e.,
irrigation water, pesticides, fertilizers, etc.) intended to
meet the needs of areas of lowest productivity. Bullock
and Bullock (6) pointed out the imminent need for efficient
methods to accurately measure within-field variation in
soil physical and chemical properties as a key component
for precision agriculture.
The effective utilization of soil as a resource depends
on assessing and maintaining its quality. A fundamental
component of assessing field-scale soil quality is establish-
ing the spatial distribution of the soil properties affecting
the soil’s intended management goal (e.g., maximize agri-
cultural productivity, minimize environmental impact,
and/or maximize waste recycling) and its intended func-
tion (e.g., biodiversity, filtering and buffering, nutrient
cycling, physical stability and structural support, resis-
tance and resilience, and water and solute flow). It is not
sufficient to take a single measurement within a field to
characterize its soil quality. Rather, a sufficient number
of measurements must be taken and at specific locations
to representatively characterize the spatial distribution
of the existing soil conditions that influence the soil’s
intended use. Therefore, assessing soil quality requires
quantitative knowledge of each indicator property asso-
ciated with a soil’s quality and the spatial variability of
those indicator properties.
Furthermore, spatial variability has a profound influ-
ence on environmental impacts, particularly of non-point
source pollutants. This result is because of its influence
on the fate and transport of agrochemical inputs (e.g.,
pesticides, fertilizers, etc.) in an agroecosystem. In fact,
it has become clear that the real constraint on modeling
solute transport is not the detail of the model structures,
but defining the characteristics of individual places (7).
Jury (3) provides an excellent fundamental discussion of
the spatial variability of soil properties and its impact
on solute transport in the vadose zone. As Jury points
out, ‘‘any hope of estimating a continuous spatial pattern
of chemical emissions at each point in space within a
field must be abandoned due to field-scale variability of
soils.’’
The characterization of spatial variability is without
a doubt one of the most significant areas of concern
in soil science because of its broad reaching influence
on all field- and landscape-scale issues. Research in
the characterization of spatial variability has advanced
tremendously over the past decade in large part because
of georeferenced ECa measurements.
EDAPHIC FACTORS INFLUENCING ECa MEASUREMENTS
The earliest field applications of geophysical measure-
ments of ECa in soil science, which occurred in the early
1980s, involved the determination of salinity within the
soil profile of arid zone soils. However, it soon became
apparent that the measurement of ECa in the field as a
quantification of soil salinity was more complicated than
initially anticipated because of the complexity of current
flow pathways originating from the spatial heterogeneity
of properties influencing current flow in soil.
Three pathways of current flow contribute to the ECa
of soil: (1) a liquid phase pathway via dissolved solids
contained in the soil water occupying the large pores, (2) a
solid–liquid phase pathway primarily via exchangeable
cations associated with clay minerals, and (3) a solid
pathway via soil particles that are in direct and continuous
contact with one another (8). Of these three pathways,
the solid pathway is usually negligible; hence, most
soil can be approximated with a dual-pathway parallel
conductance equation.
An analysis of this dual-pathway system indicates that
ECa will be influenced by a complex interaction of soil
properties including salinity, soil texture, water content,
ρb , and temperature. The ρb is directly influenced by clay
content (soil texture) and organic matter (OM). Further-
more, the exchange surfaces on clays and OM provide a
solid–liquid phase pathway primarily via exchangeable
cations; consequently, the cation exchange capacity (CEC)
and OM are recognized as additional factors that can influ-
ence ECa measurements. Measurements of ECa must be
interpreted with these influencing factors in mind.

MOBILE ECa MEASUREMENT EQUIPMENT
An ECa survey to characterize spatial variability uses
mobile geophysical equipment coupled with a global
positioning system (GPS) to georeference each ECa
measurement. Mobile ECa equipment has been developed
by a variety of researchers (9–11). The development of
mobile ECa measurement equipment has made it possible
to produce ECa maps with measurements taken every
few meters.
Mobile ECa measurement equipment has been devel-
oped for both electrical resistivity (ER) and electromag-
netic induction (EMI) geophysical approaches. In the case
of ER, four stainless-steel electrodes are inserted into the
soil generally at equal distances and connected to a resis-
tivity meter. Current is applied to the two outer electrodes
with the two inner electrodes serving as the potential
electrodes. Fixing the electrode array and mounting it
on or pulling it from a vehicle saves considerable time
for measurements. By coupling the fixed electrode array
to a datalogger and GPS georeferences the ECa mea-
surement, geospatial ECa measurements can be taken.
Veris Technologies1 (Salinas, KS; www.veristech.com) has
developed a commercial mobile system for measuring ECa
using the principles of ER. In a similar fashion, an EMI
unit can be mobilized by pulling a sled-mounted EM-38
unit (Geonics Ltd., Mississaugua, Ontario, Canada) behind
a vehicle and logging the EMI data simultaneously along
with real-time GPS position information. Measurements
of ECa with EMI generally involve taking readings in
the horizontal coil configuration (EMh ), which is a shallow
reading penetrating to a depth of about 0.75–1.0 m, and/or
in the vertical coil configuration (EMv ), which is a deep
reading penetrating to a depth of about 1.5 m.
ECa -DIRECTED SAMPLING DESIGNS
Despite the complexities of the ECa measurement, the
fact that ECa is influenced by several soil properties
has broadened its utility and enabled it to play a
major role in addressing the issue of characterizing
spatial variability. Geospatial measurements of ECa
have been successfully used for (1) identifying the soil
physicochemical properties influencing crop yield patterns
and soil condition, (2) establishing the spatial variation
of these soil properties, and (3) characterizing the spatial
distribution of soil properties influencing solute transport
through the vadose zone (4,12–14).
In a typical field survey, the spatial distribution of
various soil properties can often be adequately determined
with conductivity survey information in conjunction with
some type of soil sampling plan. Either probability
(design) or prediction (model) based sampling strategies
can identify the soil properties that correlate best with ECa
in a particular field. Prediction-based sampling strategies
objectively select sampling locations based on the observed
spatial ECa pattern and thus are often referred to
as ‘‘sensor-directed’’ sampling plans (15). Characterizing
1
Product identification is provided solely for the benefit of the
reader and does not imply the endorsement of the USDA.
CHARACTERIZING SOIL SPATIAL VARIABILITY 467
spatial variability with ECa -directed soil sampling is
based on the hypothesis that when ECa correlates with
a soil property or properties, then spatial ECa information
can identify sites that reflect the range and variability
of the property or properties. In instances where ECa
correlates well with a particular soil property, an ECa -
directed soil sampling approach can optimally establish
the spatial distribution of that property while minimizing
the corresponding sampling and labor costs (2,15,16). Also,
if ECa is correlated with crop yield, then an ECa -directed
soil sampling approach can identify those soil properties
that are causing the variability in crop yield (13). Details
for conducting a field-scale ECa survey for the purpose
of characterizing the spatial variability of soil properties
influencing soil quality or crop yield variation can be found
in Corwin and Lesch (16).
Figure 1 shows EMh and EMv maps of an ECa survey
performed on a 6.5-ha field in Maricopa, AZ, along with the
associated soil sample sites selected with a specific type
of ECa -directed sampling methodology known as spatial
response surface sampling design (15,16). Once soil cores
are taken at the selected sites and the corresponding soil
physicochemical properties are analyzed, the properties
that are well correlated with ECa can be identified.
Statistical prediction models can then be derived that
spatially map (predict) these correlated soil properties
with the calibrated ECa survey data. Figure 2 shows
the corresponding ECa -predicted 0–0.9-m percent water
content patterns for the field shown in Fig. 1.
GUIDELINES FOR CONDUCTING AN ECA -DIRECTED SOIL
SAMPLING SURVEY TO CHARACTERIZE SPATIAL
VARIABILITY
The basic elements of a field-scale ECa survey for
characterizing spatial variability include (1) ECa survey
design, (2) georeferenced ECa data collection, (3) soil sam-
ple design based on georeferenced ECa data, (4) soil
sample collection, (5) physicochemical analysis of per-
tinent soil properties, (6) spatial statistical analysis,
(7) determination of the dominant soil properties influenc-
ing the ECa measurements at the study site, and (8) GIS
development. The outlined steps of an ECa -directed soil
sampling survey are provided in Table 1.
LIMITATIONS AND PRECAUTIONS
At present, no other soil measurement provides a greater
level of spatial information than that of geospatial ECa
measurements. Even so, are limitations exist to the use
of ECa to characterize soil spatial variability. Awareness
of these weaknesses is crucial for the proper use and
interpretation of geospatial ECa measurements. The
complex spatial heterogeneity of the soil system has
subtle influences on geospatial ECa measurements that
can have significant interpretive impacts. The ability
to recognize and interpret these influences can be the
difference between the successful or failed application of
ECa measurements for characterizing spatial variability.
First and foremost, in the absence of soil sample
information, geospatial measurements of ECa by them-
selves do not directly quantify soil spatial variability.
 (a) 3661004

EMh
mS/m
< 35 3660904
35 - 40
40 - 45
> 45
3660805

Coord system
UTM (m) 3660705
X: Easting
Y: Northing

3660606
408638 408681 408725 408768 408812

(b) 3661004

EMv
mS/m
< 80 3660904
80 - 85
85 - 90
> 90 3660805

Coord system
UTM (m) 3660705
X: Easting
Figure 1. Maps of ECa survey data and ECa -directed soil
Y: Northing
sampling sites for a field in Maricopa, AZ: (a) EM38 hor-
izontal and (b) EM38 vertical. White square symbol rep-
resents the soil sample site locations. UTM = universal 3660606
transverse mercator grid system in meters. 408638 408681 408725 408768 408812

GWC (0-90 cm) 3661004

%

3660904
< 12
12 - 14
14 - 16
> 16 3660805

Coord system
UTM (m)
Figure 2. Map of spatial variability of percent gravi- 3660705
metric water content (GWC) for the 0–0.9-m sam- X: Easting
pling depth, as determined from the survey data Y: Northing
and ECa -directed soil sampling sites shown in
Fig. 1 (R2 = 0.85 for predicted and measured GWC).
UTM = universal transverse mercator grid system 3660606
in meters. 408638 408681 408725 408768 408812

468
 CHARACTERIZING SOIL SPATIAL VARIABILITY 469

Table 1. Outline of Steps to Conduct an ECa Field Survey

1. Site description and ECa survey design

a. record site metadata
b. define the project’s/survey’s objective
c. establish site boundaries
d. select GPS coordinate system
e. establish ECa measurement intensity
2. ECa data collection with mobile GPS-based equipment
a. georeference site boundaries and significant physical geographic features with GPS
b. measure georeferenced ECa data at the predetermined spatial intensity and record associated metadata
3. Soil sample design based on georeferenced ECa data
a. statistically analyze ECa data using an appropriate statistical sampling design to establish the soil sample site
locations
b. establish site locations, depth of sampling, sample depth increments, and number of cores per site
4. Soil core sampling at specified sites designated by the sample design
a. obtain measurements of soil temperature through the profile at selected sites
b. at randomly selected locations, obtain duplicate soil cores within a 1-m distance of one another to establish
local-scale variation of soil properties
c. record soil core observations (e.g., mottling, horizonation, textural discontinuities, etc.)
5. Laboratory analysis of appropriate soil physicochemical properties defined by project objectives
6. If needed, stochastic and/or deterministic calibration of ECa to ECe or to other soil properties (e.g., water content
and texture)
7. Spatial statistical analysis to determine the soil properties influencing ECa and/or crop yield
a. soil quality assessment:
1. perform a basic statistical analysis of physicochemical data by depth increment and by composite depth over
the depth of measurement of ECa
2. determine the correlation between ECa and physicochemical soil properties by composite depth over the depth
of measurement of ECa
b. precision agriculture applications (if ECa correlates with crop yield, then):
1. perform a basic statistical analysis of physicochemical data by depth increment and by composite depths
2. determine the correlation between ECa and physicochemical soil properties by depth increment and by
composite depths
3. determine the correlation between crop yield and physicochemical soil properties by depth and by composite
depths to determine depth of concern (i.e., depth with consistently highest correlation, whether positive or
negative, of soil properties to yield) and the significant soil properties influencing crop yield (or crop quality)
4. conduct an exploratory graphical analysis to determine the relationship between the significant
physicochemical properties and crop yield (or crop quality)
5. formulate a spatial linear regression (SLR) model that relates soil properties (independent variables) to crop
yield or crop quality (dependent variable)
6. adjust this model for spatial autocorrelation, if necessary, using restricted maximum likelihood or some other
technique
7. Conduct a sensitivity analysis to establish dominant soil property influencing yield or quality
8. GIS database development and graphic display of spatial distribution of soil properties

Source: Ref. 16. With permission.

Rather, ECa -survey measurements provide the spatial
information necessary to direct soil sampling. It is the
combined use of both the ECa -survey and the corre-
sponding sample data that makes this sort of surveying
technology invaluable for characterizing spatial variabil-
ity. In other words, the primary strength of geospatial
ECa measurements lies in their effectiveness as a means
to direct soil sampling (using a minimum number of sam-
ple sites) that best characterizes the spatial variability
of those soil properties influencing ECa in the landscape
of interest.
Second, ECa -directed soil sampling can only spatially
characterize soil properties that correlate with ECa . This
correlation may be caused by a direct or indirect influence
on the ECa measurement, or the correlation may be
a complete artifact. For example, salinity and water
content will directly influence ECa and CEC will indirectly
influence ECa through its influence on current flow at the
470 CHARACTERIZING SOIL SPATIAL VARIABILITY
surface of soil minerals. Likewise, in many instances, B
and salinity distributions are similar. Consequently, a
strong correlation between B and salinity (and thus ECa )
can result, even though no cause-and-effect relationship
exists between B and ECa . Thus, a solid understanding
of the soil properties that directly influence and/or are
indirectly measured by ECa at a specific site is particularly
essential in temporal surveying applications, because over
time these correlations may or may not persist.
Third, as previously mentioned, often a dynamic
and complex relationship exists between ECa and those
properties that influence ECa . Ground-truth soil samples
are obligatory if spatial measurements of ECa are
to be used for quantifying soil variability. Geospatial
measurements of ECa do not supplant the need for soil
sampling, but they do minimize the number of samples
necessary to characterize spatial variability. Additionally,
users of ECa survey data must exercise caution and
be aware of what ECa is actually measuring at the
site of interest. Normally, the only way to establish
those soil properties that influence ECa at a specific
location is by collecting representative soil samples and
then determining the relationship between ECa and the
particular property or properties of interest. This process
requires that every ECa survey has an associated soil
sampling design. The location and number of sites can be
established (after acquiring the ECa survey data) using
model-based sample design software such as ESAP (17).
Finally, the temporal stability of ECa measurements
at a site may be of potential concern because ECa is a
product of both static and dynamic factors, which adds
another dimension to the complexity of understanding
and interpreting geospatial ECa measurements. For this
reason, caution must be taken to characterize spatial
variability with ECa when dynamic conditions influencing
ECa are significant. Apparent soil electrical conductivity
surveys are generally conducted (1) within a set time
frame to minimize the effects of dynamic properties (such
as temperature and water content), (2) when the soil is
at or near field capacity (to avoid current-flow problems),
and (3) with regard for subtle topographic effects (e.g.,
bed-furrow). Protocols for conducting an ECa survey that
consider all previously discussed limitations are presented
by Corwin and Lesch (16).
FUTURE DIRECTION
Because of the heterogeneous nature of soils, characteriza-
tion of soil spatial variability is a fundamental component
of any landscape-scale process that cannot be overlooked or
superficially addressed. Significant technological advances
have occurred over the past two decades, particularly in
the area of sensor technology and in precisely locating
a geographic position with GPS. These advances have
helped the progress of research in characterizing spatial
variability at field and landscape scales.
Geospatial measurement of ECa stands out among
the ground-based sensor technologies, contributing to
an improved ability to characterize spatial variability.
Numerous soil samples are required for representative
estimates of field-scale spatial variability using traditional
grid sampling, which makes grid sampling impractical
because of labor and cost intensiveness. Soil sampling
directed by geospatial ECa measurements provides a
viable alternative for characterizing spatial variability
of a variety of soil-related physicochemical properties.
Geospatial ECa measurements provide a means of
significantly reducing the number of soil samples needed to
characterize spatial variability, provided that the target
soil properties are well correlated with the conductivity
survey data.
Directed soil sampling with geospatial ECa measure-
ments has its limitations in characterizing spatial vari-
ability, particularly when the soil properties of interest
are not measured either directly or indirectly by the ECa
survey data. In these instances, other types of sensors
should be used (if available) or a more intensive soil sam-
pling strategy must be adopted. A robust methodology for
the integrated use of multiple remote and ground-based
sensors has yet to be developed. The successful integra-
tion of multispectral and hyperspectral imagery, TDR,
GPR, aerial photography, and ECa sensors is needed to
provide the redundant and supplemental data necessary
to unravel the spatial complexity of soil. This type of inte-
grated methodology represents the future direction of soil
spatial variability research.
BIBLIOGRAPHY
1. Nielsen, D.R., Biggar, J.W., and Erh, K.T. (1973). Spatial
variability of field-measured soil-water properties. Hilgardia
42(7): 215–259.
2. Corwin, D.L. and Lesch, S.M. (In press). Apparent soil
electrical conductivity measurements in agriculture. Comput.
Electron. Agric.
3. Jury, W.A. (1986). Spatial variability of soil properties. In:
Vadose Zone Modeling of Organic Pollutants. S.C. Hern
and S.M. Melancon (Eds.). Lewis Publishers, Chelsea, MI,
pp. 245–269.
4. Corwin, D.L. et al. (2003). Assessment and field-scale map-
ping of soil quality properties of a saline-sodic soil. Geoderma
114(3–4): 231–259.
5. Verhagen, A., Booltink, H.W.G., and Bouma, J. (1995). Site-
specific management: Balancing production and environmen-
tal requirements at farm level. Agric. Syst. 49: 369–384.
6. Bullock, D.S. and Bullock, D.G. (2000). Economic optimality
of input application rates in precision farming. Prec. Agric. 2:
71–101.
7. Beven, K.J. (2002). Towards an alternative blueprint for
a physically-based digitally simulated hydrologic response
modeling system. Hydrol. Process. 16(2): 189–206.
8. Rhoades, J.D., Chanduvi, F., and Lesch, S. (1999). Soil
Salinity Assessment: Methods and Interpretation of Electrical
Conductivity Measurements, FAO Irrigation and Drainage
Paper #57, Food and Agriculture Organization of the United
Nations, Rome, Italy, pp. 1–150.
9. McNeill, J.D. (1992). Rapid, accurate mapping of soil salinity
by electromagnetic ground conductivity meters. In: Advances
in Measurements of Soil Physical Properties: Bringing Theory
into Practice. G.C. Topp, W.D. Reynolds, and R.E. Green
(Eds.). SSSA Special Publication No. 30, ASA-CSSA-SSSA,
Madison, WI, pp. 201–229.
10. Rhoades, J.D. (1993). Electrical conductivity methods for
measuring and mapping soil salinity. In: Advances in

Agronomy. D.L. Sparks (Ed.). Vol. 49, Academic Press, San
Diego, CA, pp. 201–251.
11. Jaynes, D.B., Colvin, T.S., and Ambuel, J. (1993). Soil type
and crop yield determinations from ground conductivity
surveys. ASAE Paper No. 933552, 1993 ASAE Winter
Meetings, 14–17 Dec. 1993, Chicago, IL, ASAE, St. Joseph,
MI.
12. Corwin, D.L. et al. (1999). Evaluation of GIS-linked model of
salt loading to groundwater. J. Environ. Qual. 28: 471–480.
13. Corwin, D.L. et al. (2003). Identifying soil properties that
influence cotton yield using soil sampling directed by
apparent soil electrical conductivity. Agron. J. 95(2):
352–364.
14. Corwin, D.L., Lesch, S.M., Oster, J.D., and Kaffka, S.R. (In
press). Characterizing spatio-temporal variability with soil
sampling directed by apparent soil electrical conductivity.
Geoderma.
15. Lesch, S.M. (In press). Sensor-directed response surface
sampling designs for characterizing spatial variation in soil
properties. Comput. Electron. Agric.
16. Corwin, D.L. and Lesch, S.M. (In press). Characterizing soil
spatial variability with apparent soil electrical conductivity:
I. Survey protocols. Comput. Electron. Agric.
17. Lesch, S.M., Rhoades, J.D., and Corwin, D.L. (2000). ESAP-
95 Version 2.10R: User Manual and Tutorial Guide,
Research Rpt. 146, USDA-ARS George E. Brown, Jr. Salinity
Laboratory, Riverside, CA.
DEEP SOIL-WATER MOVEMENT
TODD RASMUSSEN
The University of Georgia
Athens, Georgia
Soil water refers to subsurface water between the
Earth’s surface and the regional water table, called the
unsaturated, or vadose zone. Figure 1 illustrates the
components of flow within this zone. Deep soil-water
movement usually implies the migration of water below
the root zone, and is also called deep percolation.
Soil water is held by adhesive forces to soil surfaces and
by cohesive forces to other water molecules. Measurements
Precipitation
Evapotranspiration
Percolation
Overland flow
Infiltration
Clay
Recharge Perched water
Watertable
Saturated zone
Figure 1. Flow components in the unsaturated zone.
DEEP SOIL-WATER MOVEMENT 471
of soil-water pressures are usually negative—water does
not freely drain into a well. Some soil water may exist
with positive pressures, such as water that is perched in a
saturated zone above a layer of lower permeability. Water
under positive pressure flows freely into a well.
FORCES THAT CONTROL WATER MOVEMENT
Gravity is the dominant force that controls soil-water
movement. The downward force of gravity acts in
conjunction with geologic media to control the magnitude
and direction of fluid flow.
A second important force is the matric tension (or soil-
water tension), which causes water to move from zones of
lower tension to zones of higher tension. The soil tension,
ψ, is the negative of the water pressure head, ψ = −P/γ ,
where P is the fluid pressure relative to atmospheric
pressure and γ is the freshwater specific weight.
The total head, H, combines the gravitational and
tension forces using H = z − ψ, where z is the elevation of
the point of measurement.
In some circumstances, osmotic forces can also affect
water movement. Osmotic forces develop when solute
concentration gradients are present, causing water to
move from zones of low salt concentration to zones of
higher concentration. When osmotic forces are present,
the total head must include the osmotic potential, φ, as
noted in Section GW456 (head).
MATERIAL PROPERTIES
In addition to the total head, the characteristics of the
geologic medium are also important for controlling the
magnitude and direction of deep soil-water movement (1).
The important material properties that affect soil-water
movement are the unsaturated hydraulic conductivity
functions and the soil-water characteristic curves.
Unsaturated Hydraulic Conductivity
The unsaturated hydraulic conductivity, K(ψ), is com-
monly written as:
K(ψ) = Kr (ψ) Ks (1)
where Kr is the relative unsaturated hydraulic conductiv-
ity function that varies with matric tension, ψ, and Ks is
the saturated hydraulic conductivity.
As soil dries, the tension increases and the relative
hydraulic conductivity decreases, whereas wetting soils
show a decrease in tension and an increase in relative
conductivity. The relative conductivity ranges from a
theoretical minimum of zero in dry soil, Kr → 0 as ψ → 0,
to a maximum of one in saturated soil, Kr → 1 as ψ = 0,
so that 0 < Kr ≤ 1.
Both the relative and saturated hydraulic conductiv-
ities vary with the pore-size distribution of the geologic
media. Sandy materials may have a large saturated
hydraulic conductivity, but a rapidly declining relative
hydraulic conductivity, whereas finer materials may have
a smaller saturated hydraulic conductivity but a slowly
472 DEEP SOIL-WATER MOVEMENT
Coarse-grained
Hydraulic conductivity, K
Fine-grained
Wet Dry
Tension, y
Figure 2. Unsaturated hydraulic conductivity, K, as a function
of tension, ψ, for two materials. The coarser material is more
conductive under wet conditions (smaller values of tension),
whereas the finer material is more conductive under drier
conditions.
declining relative conductivity. This effect is shown if
Fig. 2.
The crossover of the two curves shown in Fig. 2 results
in the interesting phenomenon that coarser materials have
a larger unsaturated hydraulic conductivity than finer
material in the wet range, but a smaller conductivity in
the drier range.
Thus, the flow behavior for wet conditions may be
very different from the behavior under dry conditions
in heterogeneous materials. Coarser materials (such as
sands, fractures, and other large openings) generally
dominate flow under wet conditions, whereas finer
materials (such as silts, clays, and consolidated materials)
can dominate flow under drier conditions.
Moisture Characteristic Curves
The ability of the unsaturated zone to hold water as
a function of the soil tension is called the moisture
characteristic curve. The characteristic curve for each
material varies with the pore size distribution—a soil
with larger pores releases water at a lower tension than a
soil with smaller pores.
Figure 3 shows two soil characteristic curves, one for a
coarser material and a second for a finer material. Like
the unsaturated hydraulic conductivity function, the finer
soil can hold more water under dry conditions, whereas
the coarser soil holds more water under wetter conditions.
The water retention properties of soils also affect water
movement. Changes in water flux with depth can be
attributed to the uptake and release of water as the soil
tension changes. For example, a wetting front may move
downward with time from the surface after a large rain
event, only to be absorbed by the dry soil underneath.
The ability to absorb the moisture is a function of the
ambient moisture tension—an already wet soil will have
little ability to absorb additional water.
In some cases, water percolating down from above may
not immediately reach the water table because the dry
Coarse
Water content, q
Fine
Wet Dry
Tension, y
Figure 3. Soil water content, θ, as a function of tension, ψ, for
two materials. The coarser material holds more water under wet
conditions (smaller values of tension), whereas the finer material
holds more water under drier conditions.
soils mediate the pulse of water. This water is retained
by the soil and then slowly released over time. The initial
pulse of water at the surface is transformed into a slow
drainage at depth.
Only in materials with little water storage capacity
(i.e., in wet soils or in materials with limited water holding
capacity) will the downward movement of water be rapid.
In these cases, the water that reaches the water table
may or may not be the same water that was added at
the surface. In some situations, the added water merely
displaces the water below it, causing an equivalent volume
of water to move at depth. In other situations, the added
water moves quickly through the unsaturated zone.
Spatial Variability
Local subsurface conditions can also affect deep soil-water
movement. Hydrogeologic units that dip in one direction,
such as bedded sedimentary materials, can cause water
to flow in the down-dip direction. Microscopic layering in
features such as sand dunes can also affect the direction
of movement.
Vertical features, such as fractures and joints, may
focus water movement during wet conditions. Localized
zones of preferential flow can also be found in soils with a
defined structure, such as in soil with aggregates. These
soils form a dual-porosity network consisting of fine pores
within the aggregates and larger macropores between the
aggregates.
Many explanations exist for preferential flow within
soils, including bypass flow, macropore flow, fracture flow,
boundary layer flow, mobile zone flow, finger flow, funnel
flow, media heterogeneities, ion exclusion, and colloid
transport (2).
WATER VAPOR MOVEMENT
Water vapor movement in deep systems can also be
important, especially in very dry media where liquid flow

is small. Water vapor is generally in equilibrium with the
liquid phase and can be related to the matric tension of
the pore fluids (3):


e gψ
RH = = exp (2)
es RT
where RH is the relative humidity, e is the actual
vapor pressure, es is the saturated vapor pressure, R
is the water vapor gas constant, and T is the absolute
temperature. Dissolved solids in soil water can lower the
relative humidity—and increase the tension—because of
osmotic forces.
The diffusion of water vapor through the subsurface
can be predicted using a variety of approaches that
depend on the specific conditions within the subsurface,
which is because additional forces become important
when considering subsurface vapor movement, including
thermal effects and barometric pressure changes.
Temperature variation affects water movement because
the saturated vapor pressure increases with temperature.
The natural temperature increase with depth is called the
geothermal gradient. Deeper water is warmer and has
a higher vapor pressure than water at a shallow depth.
This vapor pressure gradient causes water vapor to move
upward toward the cooler zones, where it may condense.
Also, barometric fluctuations cause oscillations in the
vertical movement of soil gasses, leading to the ver-
tical advective transport of water vapor. The vertical
movement may be amplified when extreme barometric
pressure conditions are present, such as during a hur-
ricane when extremely low surface barometric pressures
cause an upward air pressure gradient within the unsatu-
rated zone.
BIBLIOGRAPHY
1. Evans, D., Rasmussen, T.C., and Nicholson, T.J. (2001). Flow
and transport through unsaturated fractured rock: An
overview. In: Flow and Transport Through Unsaturated
Fractured Rock, 2nd Edn. D.D. Evans, T.J. Nicholson, and
T.C. Rasmussen (Eds.). Geophysical Monograph 42. American
Geophysical Union, Washington, DC.
2. Rasmussen, T.C., Baldwin, R.H., Dowd, J.F., and Williams,
A.G. (2000). Tracer vs. pressure wave velocities through
unsaturated saprolite. Soil Sci. Soc. Amer. J. 64(1): 75–85.
3. Hillel, D. (1971). Soil and Water: Physical Principles and
Processes. Academic Press, New York.
SPECIFIC GRAVITY
TODD RASMUSSEN
The University of Georgia
Athens, Georgia
The density of water and other materials varies with
temperature as well as with the amount of dissolved and
suspended matter within them. The specific gravity is
the observed density of a material divided by a reference
density, which is usually taken to be water at its maximum
SPECIFIC GRAVITY 473
density, i.e., 1 kg/L. Thus, variations in the specific gravity
are a direct measure of the variation in density. The
specific weight of a material equals the product of its
density and the gravitational constant. The specific gravity
is important because it is needed to correct for the
water density in wells to determine the freshwater head.
Well construction also relies on the specific gravity of
drilling muds to prevent the collapse of borehole walls in
unconsolidated media.
The specific gravity is a measure of the density, or
weight, of a fluid or solid. Denser materials have a larger
specific gravity than do less dense materials. The specific
gravity, , is defined using
ρ
= (1)
ρo
where ρ is the density (mass per unit volume) of the
material and ρo is the density of a reference material.
The density of water at 4 ◦ C, ρo = 1 kg/L, is generally
the reference density for liquids and solids. Table 1
provides the specific gravity for a range of elements,
including very light metals, such as aluminum and
titanium, to very dense metals, such as gold and platinum.
Note that for liquids and solids, the specific gravity equals
the density because the standard density of water is 1 kg/L.
The density of hydrogen is common for gasses.
The specific gravity of a pure substance can be affected
by temperature. Table 2 provides the specific gravity and
specific weight for water from 0 ◦ C to 100 ◦ C. Note that the
specific gravity decreases more rapidly as the temperature
increases.
Changes in physical properties of solids, such as their
crystalline structure or porosity, can also result in a change
in their density. The specific gravity can also be affected
by any components present within the material, including
dissolved gasses, liquids, and solids, as well as suspended
gasses, liquids, and solids.
For example, water with high concentrations of
suspended solids (such as drilling fluids) or dissolved solids
(such as brines) has higher specific gravities than does
clear water. Also, a column of air bubbles rising within a
water column can reduce the fluid density.
The specific gravity is also related to the specific weight,
γ = ρg, where g = 9.807 m/s2 is the gravitational constant.
The specific weight commonly converts measured pres-
sures to an equivalent hydraulic head because an increase
Table 1. Specific Gravity
of Selected Elements
Material Specific Gravity
Aluminum 2.70
Titanium 4.51
Iron 7.87
Copper 8.96
Silver 10.50
Lead 11.30
Mercury 13.50
Uranium 19.10
Gold 19.30
Platinum 21.50
474 SPECIFIC GRAVITY

Table 2. Variation of Pure Water with found using

Temperature f − 1
C= (2)
Temperature Density Specific Weight s − 1
( ◦ C) (kg/L) (Pa/m)
where f is the measured specific gravity of the fluid and
0 0.99984 9,805.4 s is the specific gravity of the suspended sediments.
10 0.99970 9,804.1
20 0.99821 9,789.4
IMPORTANCE OF SPECIFIC GRAVITY
30 0.99565 9,764.3
40 0.99222 9,730.7
Head Measurement
50 0.98803 9,689.6
60 0.98320 9,642.2 The use of water levels in wells to determine the hydraulic
70 0.97778 9,589.1 head within an aquifer is complicated when the density
80 0.97182 9,530.6 of the fluid within the borehole is not a constant (2). As
90 0.96535 9,467.2 noted in HEAD, changes in density affect the head according
100 0.95840 9,399.0 to the relationship:

in depth, z, is accompanied by a concomitant increase in
pressure, P, which yields γ = P/z.
MEASURING THE SPECIFIC GRAVITY
A traditional technique for measuring the fluid density
uses a hydrometer, which compares the weight of the
fluid to a standard fluid. The hydrometer is allowed to
float within the fluid, and the fluid density is determined
from how much higher or lower the hydrometer level is
compared with pure water (Fig. 1). Heavier fluids cause
the hydrometer to rise relative to clear water because of
buoyancy.
An alternative method for determining the fluid density
is to employ a differential pressure transducer to measure
the change in fluid pressure over a known depth. The
resulting ratio between the measured pressure and the
depth yields an estimate of the specific weight, γ = P/z.
Pressure transducers have recently measured the
concentration and particle size distribution of suspended
sediments (1). The particle size distribution is determined
using Stokes’s law, whereas the sediment concentration is
Hydrometer
H = zo + z (3)
where H is the freshwater head, zo is the midpoint
elevation of the screened zone within the aquifer,  = γ /γo
is the specific gravity of the fluid within the well,
z = z − zo is the height of the column of water within
the monitoring well above the screened zone, and z is the
water surface elevation in the well.
Drilling Fluids
The specific gravity of the drilling fluid is important during
well construction. In loose materials, the process of drilling
a well is complicated by the tendency of the unsupported
walls to collapse.
Borehole walls can be temporarily supported until
a casing is emplaced by using drilling fluids that are
denser than water. The weight of the drilling fluid places
an outward force on the borehole walls that helps to
support them.
Heavier drilling fluids place a correspondingly heavier
force on the borehole wall. One important drawback is the
tendency of the drilling fluid to accumulate on the walls
and within the formation, which causes a reduction in well
yield if it is not removed during well development.
Lighter drilling fluids, such as foams and air, can also be
used. In this case, the inflow of water from the formation
into the borehole prevents the buildup of materials on
the borehole wall, which eases well development. This
technique is suitable in consolidated media where borehole
collapse is not a problem.

Table 3. Specific Gravities of

Selected Drilling Fluids
Drilling Fluid Specific Gravity

Air 0.001
Mist 0.020
Stiff foam 0.100
Wet foam 0.200
Polymers 1.020
Clear water Water with mud Polymers w/ NaCl 1.200
Bentonite clay 1.300
Figure 1. Hydrometer technique for measuring fluid density. Polymers w/ CaCl 1.400
Clear water causes the hydrometer to float at a constant level. Bentonite w/ barite 2.200
Denser fluids cause the hydrometer to rise, whereas lighter fluids Barite (BaSO4 ) 4.200
cause the hydrometer to sink lower in the water column.

Table 3 presents approximate specific gravities for a
range of drilling fluids. Notice the wide range in specific
gravities, from a low of 0.01 for air to a high of 4.2 for
barite.
The values in Table 3 are approximate because the final
density is a function of the relative concentrations of the
liquids, solids, and gasses within the drilling fluid. Also
note that the specific gravity of the air and mist are related
to the density of water.
BIBLIOGRAPHY
1. Lewis, A.J. and Rasmussen, T.C. (1999). Determination of
suspended sediment concentration and particle size distri-
butions using pressure measurements. J. Environ. Qual. 28:
1490–1496.
2. Driscoll, F.G. (1986). Groundwater and Wells, 2nd Edn.
Johnson Filtration Systems, Inc., St. Paul, MN.
HOT SPRINGS
JOHN E. MOORE
USGS (Retired)
Denver, Colorado
Springs have been important in the social, cultural, and
economic history of humankind. They have been an impor-
tant source of water for humans. Our ancestors did not
have drilling equipment, so in the absence of surface water,
Figure 1. Diagram of a spring collection box (1).
HOT SPRINGS 475
they had to rely on springs. Many early towns in Greece,
Egypt, Mesopotamia, India, and China developed around
springs. Many towns in the United States have taken their
names from springs: Steamboat Springs, Colorado; Glen-
wood Springs, Colorado; Springfield, Missouri; and Rock
Springs, Wyoming, are a few examples.
For centuries, hot springs and mineral springs were
considered of medicinal or therapeutic value. Bottled
‘‘springwater’’ is still considered by many people of
higher quality than ordinary tap water. This belief is
at least partially responsible for the current high-volume
sale of bottled water. Unfortunately, this association of
springwater with purity is not based on fact. Springwater
can contain higher concentrations of dissolved solids than
local public water supplies, and springs become easily
contaminated. Springs become contaminated because they
are open, unprotected, and accessible to human beings and
animals. Contamination can also occur from surface water
flow or flooding into spring collection systems.
More than 3400 public water supply systems (1992)
in the United States obtain part or all of their drinking
water from springs. These systems provide drinking water
for more than 7 million people. However, the average
number of people served by a single public supply system
using springs as a water source is small. Springs are
an important component of the drinking water supply in
many states. Public water supplies that use springs are
more numerous in the western United States. Wyoming
has 80 springs that are used for public water supply. A box
used to collect water from springs is shown in Fig. 1 (1).
476 HOT SPRINGS
A spring is a source of water that flows naturally
from an aquifer or soil onto the land or into a body
of surface water. Its occurrence depends on the nature
and relationship of rocks (especially permeable and
impermeable strata), on the position of the water table,
and on topography.
Springs are generally classified as gravity springs and
artesian springs. Thermal springs are typically considered
a type of artesian spring. The vast majority of springs are
gravity springs. Gravity springs are created by water
that moves downgradient (downhill) and emerges at the
surface. Two gravity springs, contact and fault springs
are shown in Fig. 2. Artesian (nongravitational) springs
occur when the potentiometric level of the groundwater
system is above the land surface and water flows at the
land surface under pressure either at the aquifer outcrop
or from fractures or faults.
Springs are replenished by precipitation that recharges
aquifers. The precipitation seeps into the soil and enters
fractures, joints, bedding planes, or pore spaces in sand
aquifers and sedimentary rocks. Springs occur when water
flows through aquifers and discharges at the land surface
through faults, fractures, or by flow along an impermeable
layer. They can also occur where water flows from large
orifices that result when the water dissolves carbonate
rock (karst) and enlarges fractures or joints to create
a passage (Fig. 2).
Springwater is sometimes forced to the surface along
a fault from deep sources by thermal pressure gradients.
Springs are associated with volcanism and fractures that
extend to great depths in the earth’s crust. Springs
are a common feature in some hydrogeologic settings.
Water table
Sandstone Spring
Seep
Stream
Shale
Contact spring
Sandstone
Water table
Spring
Fault zone Swamp
Stream
Slitstone
Fault spring
Figure 2. Contact and fault springs (2).
For example, springs are common in mountains and in
watersheds that are underlain by fractured rock aquifers.
Springs can be regional (long flow paths) or local (short
flow paths). Local springs are comparatively small, can
have low flow, and are typically from shallow aquifers.
The discharge from these springs often fluctuates either
seasonally or in greater cycles, sometimes in response to
local precipitation. Local aquifers are quickly recharged,
and water movement through them is comparatively rapid,
resulting in low water mineralization. Springs supported
by local aquifers are more likely than regional flow
springs to stop flowing periodically. Regional springs more
typically have large discharges. Regional springs typically
have nearly constant discharge and are more mineralized
than local springs. Regional springs rarely stop flowing
even during long droughts.
The discharge rate of a spring is a function of three
main variables:
1. hydraulic conductivity (permeability) of the aquifer
2. area contributing recharge to the aquifer
3. quantity of recharge.
Springs are also classified based on hydrogeologic
characteristics. The Water Resources Division of the
U.S. Geological Survey recognizes eight principal types
of springs based on hydrogeologic characteristics:
1. An artesian spring is a release of pressurized water
from a confined aquifer at the aquifer outcrop or
through an opening in the confining unit.
2. A depression spring is formed when the water
table intersects a steeply sloping land surface. This
type of spring is sensitive to seasonal fluctuations
in groundwater storage and frequently disappears
during dry periods.
3. A fracture spring (fault spring) is formed in fractured
or jointed rocks. Water movement is through
fractures, and springs form where the fractures
intersect the land surface. This type of spring is
particularly sensitive to seasonal fluctuations in
groundwater storage and frequently disappears due
to reduced flow during dry periods. Daily fluctuations
of discharge of small springs are commonly the result
of use of water by vegetation.
4. A contact spring occurs where a permeable water-
bearing unit overlies a less permeable unit that
intersects the ground surface.
5. A geyser spring is a periodic thermal spring resulting
from the expansive force of superheated steam
within constricted subsurface channels.
6. A perched spring occurs where infiltrating water
discharges above the regional water table.
7. A seep spring discharges from numerous small
openings in permeable material. These springs
typically have very low discharge rates.
8. A tubular spring discharges from rounded channels
(karst solution openings, lava tubes).
 SQUEEZING WATER FROM ROCK 477

Springs occur in many sizes, types of discharge points,
and locations with respect to topography. They occur in
the highest elevations of mountainous areas and on valley
floors. Many of the public water supply springs in the
western United States are small; discharges are less than
5 gallons per minute. Springs vary in their physical and
chemical compositions. They can be cold (50 ◦ F) or hot
(more than 10 ◦ C above the air temperature). Shallow
groundwater has a temperature within a few degrees of the
ambient air temperature. Higher temperatures usually
indicate deeper circulation. Thermal springs gain their
temperature increases when the water comes in contact
with recently emplaced igneous masses. Thermal and hot
springs are due to deep-seated thermal sources and are
classed as volcanic springs or fissure springs. Springs
may occur singly or in groups that can include dozens of
habitats in various sizes.
Springs may be highly mineralized, especially thermal
springs and sometimes region springs that have a very long
flow path. Thermal springs have pH values ranging from
7.2–7.6. Electrical conductance of spring water ranges
from 2–10,000 milligrams per million.
A list of question that could be asked to evaluate spring
conditions is shown below.
• Are there standing water and surface drainage
around the spring?
• Is there deeply rooted vegetation in the spring
collection area?
• Are there roots in the collection pits?
• Is there a lock-tight fence around the collection area?
• Is there a diversion channel capable of diverting
surface water away from the collection area?
• Are there privies, septic systems, underground
storage tanks, or barnyards near the spring?
• Is the collection chamber properly constructed
(lockable)?
Moore, J.E. et al. (in press). A Field Guide to Evaluate Springs.
USEPA, Washington, DC.
Sanders, L.L. (1998). A Manual of Field Hydrogeology. Prentice-
Hall, Upper Saddle River, NJ.
SQUEEZING WATER FROM ROCK
NASA Earth Science Enterprise
Data and Services
On December 16, 1811, an earthquake estimated at
magnitude 7.0 to 8.0 on the Richter scale rocked
the frontier town of New Madrid, Missouri, in the
central United States. It was the first of three powerful
earthquakes that would shake the central Mississippi
River Valley that winter. Damages were reported as far
away as Charleston, South Carolina, and Washington, DC.
According to the U.S. Geological Survey, the New Madrid
events were the most powerful earthquakes in recorded
U.S. history.
Survivors of the New Madrid earthquakes reported
not only intense ground shaking and land movement,
as would be expected during an earthquake, but also
an unfamiliar phenomenon: water and sand spouting up
through fissures, or cracks, in the Earth’s surface. In a
letter published in Lorenzo Dow’s Journal, New Madrid
resident Eliza Bryan (1) wrote in 1816:
. . . the surface of hundreds of acres was, from time to time,
covered over in various depths by the sand which issued from
the fissures, which were made in great numbers all over this
country, some of which closed up immediately after they had
vomited forth their sand and water. . . 25 June 2003

BIBLIOGRAPHY

1. U.S. Environmental Protection Agency. (1999). Guidance

Manual for Conducting Sanitary Surveys of Public Water
Systems.
2. Moore, J.E. et al. (1995). Groundwater-a primer. American
Geological Institute, Arlington, VA.

READING LIST

Davis, S.N. and De Wiest, R.J.M. (1966). Hydrogeology. John

Wiley & Sons, New York.
U.S. Environmental Protection Agency. (1997). Investigation of
Hydrogeologic Mapping to Delineate Protection Zones Around
Springs-Report of Two Case Studies. Structures in the village of Bhuj, India, which was 80
kilometers (50 miles) from the earthquake epicenter, sustained
Fetter, C.W. (2000). Applied Hydrogeology. Prentice-Hall, New
considerable damage. (Image Copyright  James Mori, Research
York.
Center for Earthquake Prediction, Disaster Prevention Institute,
LaMoreaux, P.E. and Tanner, J.T. (2001). Springs and Bottled Kyoto University)
Waters of the World. Springer, New York.
Meinzer, O.E. (1923). Outline of Ground Water Hydrology. U.S.
Geology Survey Water Supply Paper 494, p. 71. This article is a US Government work and, as such, is in the
Moore, J.E. (2002). Field Hydrogeology. Lewis, Boca Raton, FL. public domain in the United States of America.
478 SQUEEZING WATER FROM ROCK
Modern earthquake researchers refer to this expulsion
of water and sand from the ground as earthquake
dewatering, which results from liquefaction. Liquefaction
occurs when an earthquake vigorously shakes and
compacts water-saturated sediments. As the compaction
process displaces the water between the sediment pores,
water and sand shoot upward and out of the ground.
‘‘Imagine a cube full of sand and water. If you press
it in from both sides (compressing and releasing it,
then compressing it again), you build up what’s called
pore-water pressure,’’ said Martitia Tuttle, geologist and
consultant for the U.S. Geological Survey. ‘‘It’s like
shaking a coke can—when the pressure builds up and
you release it, the fluid comes shooting to the surface.’’
These fountains of water can sometimes shoot as high as
30 feet into the air, according to Tuttle (2).
Tuttle is part of a team that has been studying
liquefaction in the New Madrid region in an effort to
identify large prehistoric earthquakes in the geologic
record. ‘‘Interpreting historic and prehistoric earthquake
events involves a lot of uncertainty, due to the lack of
data,’’ Tuttle said. ‘‘We often don’t know precisely where
the earthquakes occurred or how big they were.’’ What
researchers needed was a modern-day event in a similar
environment with which to compare the New Madrid
earthquakes.
On January 26, 2001, a magnitude 7.7 earthquake
near the village of Bhuj, located in the Kachchh
region of northwestern India, provided scientists with
just that analogue. Considered one of the two most
damaging earthquakes in India’s recorded history, the
The Bhuj earthquake struck the Kachchh region of northwestern
India on January 26, 2001. The bottom image is a detail of the
area indicated in red in the top map. (Images adapted from maps
from the University of Texas at Austin General Libraries)
Earthquake Severity Table Adapted from What is Richter
Magnitude? (A New Browser Window Will Open.)
Richter Magnitude Earthquake Effects
Less than 3.5 Recorded but generally not felt
3.5 to 5.4 Often felt but rarely causes damage
Less than 6.0 Slight damage at most to
well-constructed buildings; over
small regions, can cause major
damage to poorly
constructed buildings
6.1 to 6.9 Can be destructive in populated
regions up to 100 kilometers across
7.0 to 7.9 Major earthquake that can cause
serious damage over areas larger
than 100 kilometers across
8 or more Great earthquake that can cause
serious damage over areas several
hundred kilometers across
Bhuj earthquake killed about 20,000 people and caused
an estimated $3.3 billion in damages.
According to Tuttle and co-authors (3) of a paper
recently published in Eos, local residents and survey
teams working in the area after the earthquake reported
fountains of water and sediments surging from the ground
during and immediately following the Bhuj earthquake.
The water flow was significant enough to reactivate
streams in previously dry river channels.
‘‘The Bhuj earthquake provided an opportunity to study
an earthquake that was very similar to the New Madrid
event—it was a very large earthquake, and it occurred in
an intraplate-like setting,’’ said Tuttle. The New Madrid
and Kachchh regions are both located more than 185 miles
(300 kilometers) from active plate boundaries.
‘‘Before the Bhuj earthquake, there was a lot of debate
in the seismological community about the magnitude of the
New Madrid earthquakes, but based on the similarity in
intensity between the New Madrid and Bhuj events, there
is a growing consensus among scientists that the New

Madrid earthquakes were probably larger than magnitude
7.5,’’ said Tuttle. ‘‘The Bhuj earthquake is a modern
event that helps us better understand the historic and
prehistoric earthquake record in the New Madrid region.’’
But despite the unique study opportunity presented by
the Bhuj earthquake, the area presented the researchers
with some grave challenges. First, the Kachchh region of
India is remote and difficult to reach. Second, it shares
a politically sensitive border with Pakistan. These factors
made it difficult for research teams to conduct field studies
to verify liquefaction effects or the presence of water in
some areas.
Bernard Pinty (4), research scientist at the Institute
for Environment and Sustainability in the Joint Research
Centre of the European Commission in Ispra, Italy, and
colleagues found that water expelled during the earth-
quake could be detected in imagery from NASA’s Multi-
angle Imaging SpectroRadiometer (MISR) satellite sensor.
The MISR instrument, which flies aboard NASA’s Terra
satellite, views the sunlit face of the Earth at nine
widely spaced angles simultaneously. As it passes over
the Earth, its nine cameras successively view each piece
of the planet’s surface at a spatial resolution of 300 yards
(275 meters).
The change in reflection at different angles provides a
means for distinguishing different types of land surface
covers. Since bright soils in India’s Rann of Kachchh (also
known as the Rann of Kutch) reflect most of the sun’s near-
infrared radiation, and water bodies absorb near-infrared
radiation, MISR is able to detect the contrast and indicate
where dewatering occurred.
According to Pinty, these features made MISR an ideal
tool for studying liquefaction in the Bhuj earthquake.
‘‘MISR’s multi-angle capability was instrumental in
These cracks in the ground surface north of Chobari, India
were caused by liquefaction induced during the Bhuj earthquake.
(Image Copyright  James Mori, Research Center for Earthquake
Prediction, Disaster Prevention Institute, Kyoto University)
SQUEEZING WATER FROM ROCK 479
exposing the presence of free water close to the Pakistani
border, a region where it is difficult, if not impossible, to
take ground measurements,’’ he said.
Most other satellite instruments only look straight
down or towards the edge of the Earth. ‘‘It would take
several more passes for another sensor to capture the
entire region that the MISR imagery covers,’’ said Tuttle.
‘‘The MISR images enabled us to see a large area both
during and immediately after the earthquake, and we
could see that dewatering was occurring over a huge
region.’’
Scientists now know that the New Madrid earthquakes
were not geologic flukes; similar events have occurred
repeatedly in the geologic past. In fact, the central
Mississippi Valley experiences more earthquakes than
any other region in the United States east of the
Rocky Mountains.
The New Madrid Seismic Zone, which includes parts
of Tennessee, Kentucky, Missouri, Arkansas, Illinois,
Indiana, and Mississippi, harbors a large liquefaction field.
Scientists estimate the probability of a magnitude 6.0 or
greater earthquake occurring in the New Madrid Seismic
Zone within the next 50 years at 25–40 percent, according
to the U.S. Geological Survey. But a similar event would
have much more serious consequences today.
In 1811–1812, the central Mississippi Valley was a
remote frontier. Today, the region is home to millions
of people and includes the metropolitan cities of St.
Louis, Missouri, and Memphis, Tennessee. In addition,
most buildings in the region were not built to withstand
earthquake shaking, as they often are in California
and Japan.
Buildings and engineered structures are at much higher
risk in areas where liquefaction occurs, due to increased
ground motion. ‘‘The liquefaction process is extremely
damaging to all sorts of engineered structures; it has
This false-color MISR image was acquired after the Bhuj
earthquake, on January 31, 2001. The image shows numerous
areas where groundwater flowed up to the surface, including
within the Rann of Kachchh, as well as near the Indo-Pakistani
border. (Image courtesy of NASA/GSFC/LaRC/JPL, MISR Team)
Click here for more information.
480 STORAGE COEFFICIENT
caused major destruction in both Anchorage and San
Francisco,’’ said Pinty.
‘‘The ground failure that resulted from liquefaction
during the New Madrid earthquakes was severe. We’re
talking about vertical displacement of 3 to 6 feet (1
to 2 meters), and lateral displacement up to 33 feet
(10 meters),’’ said Tuttle. ‘‘A recurrence of that type of
event would have severe consequences for engineered
structures.’’
According to Tuttle, in 1988 a magnitude 5.9 earth-
quake in Quebec, Canada, produced liquefaction. ‘‘Base-
ments cracked, septic fields were disrupted, and people
described water and sand shooting into their basements,
out of their toilets, and into their bathtubs,’’ she said. ‘‘And
that was just a moderate-sized earthquake, not a big one.’’
MISR imagery, which is archived at the Langley
Research Center DAAC, promises to be a boon to
earthquake researchers seeking to minimize hazard risks.
The MISR instrument views the sunlit Earth simultaneously at
nine widely spaced angles, providing ongoing global coverage at
high spatial detail. (Image courtesy of NASA)
Although earthquakes in the central and eastern United States
are less frequent than in the western United States, they
affect much larger areas. This is shown by two areas affected
by earthquakes of similar magnitude—the 1895 Charleston,
Missouri, earthquake in the New Madrid Seismic Zone and the
1994 Northridge, California, earthquake. Red indicates regions of
minor to major structural damage; yellow indicates areas where
shaking was felt, but little or no damage to objects occurred.
(Image courtesy of the U.S. Geological Survey)
‘‘The interesting thing is that this project started as an
‘after hours’ sort of activity,’’ said Pinty. ‘‘But it soon
acquired its own momentum, giving us results that are
of significant relevance to earthquake study applications.
Our findings revealed that MISR can detect water and
moist conditions on terrestrial surfaces, even under
unfavorable spectral conditions.’’
‘‘If we study modern earthquakes that produce
liquefaction, we can better interpret the geologic record of
liquefaction during past events. This helps us anticipate
what is likely to happen in the future so that we can
make informed decisions about reducing and mitigating
hazards,’’ said Tuttle.
For more information, visit the Langley Atmospheric
Sciences Data Center DAAC. (A new browser window
will open.)
‘‘It’s one of those things where people tend to think—if
it hasn’t happened during my lifetime, then it can’t
happen here,’’ said Tuttle. ‘‘But the liquefaction field in
the New Madrid region is very large. We’re talking about
a huge earthquake that could have a significant impact on
society.’’
BIBLIOGRAPHY
1. Bryan, E. (1849). New Madrid, Territory of Missouri, March 22,
1816. In Lorenzo Dow’s Journal. Published by Joshua Martin:
344–346. Quoted by The Virtual Times, The New Madrid
Earthquake. Accessed June 5, 2003.
2. Tuttle, M.P. (2001). The use of liquefaction features in
paleoseismology: Lessons learned in the New Madrid seismic
zone, central United States. Journal of Seismology. 5: 361–380.
3. Tuttle, M.P. et al. (2002). Observations and comparisons of
liquefaction features and related effects induced by the Bhuj
earthquake. Earthquake Spectra. 18(Supp. A): 79–100.
4. Pinty, B. et al. (2003). Observing earthquake-related dewa-
tering using MISR/Terra satellite data. EOS, Transactions,
American Geophysical Union. 84(5): 37–48.
This article contributed from Distributed Active Archive
Center (DAAC) Alliance: Supporting Earth Observing
Science 2003
For a copy, e-mail mailto:nasadaacs@nsidc.org?subject
=DA 2003 annual request via EO.
STORAGE COEFFICIENT
MATTHEW M. ULIANA
Texas State University—San
Marcos
San Marcos, Texas
DEFINITIONS
A coefficient is a mathematical quantity that relates two
or more variables in an equation describing some physical
process. In equations describing the flow of fluids in a
porous medium (e.g., groundwater in an aquifer), the
storage coefficient is the quantity that relates the
change in fluid potential (i.e., hydraulic head) in the
 STORAGE COEFFICIENT 481

aquifer to the change in the amount of water stored in the can be substituted for q to get the expanded version of the
aquifer at a given point. In practical terms, the storage flow equation:
coefficient tells us the volume of water that can be removed



from an aquifer for a given drop in hydraulic head. ∂h ∂h ∂h
∂ Kx ∂ Ky ∂ Kz
There are three storage coefficients used in groundwa- ∂x ∂y ∂z ∂h
+ + = Ss (3)
ter flow equations (the units of each are given in [brackets] ∂x ∂y ∂z ∂t
following each definition):
Equations 1 and 3 are specific versions of the continu-
Specific Storage (Ss ) is the volume of water removed ity (or conservation of mass) equation,
from a unit volume of a confined aquifer for a unit
drop in hydraulic head [L−1 ]. I − O = S (4)
Storativity (S) is the volume of water removed from
a unit area of an aquifer for a unit drop in where I = flux into a volume of a system [L3 · t−1 ]
hydraulic head; in confined aquifers, it is equal to the O = flux out of a volume of a system [L3 · t−1 ]
specific storage times the thickness of the aquifer, in S = change in the amount stored in that same
unconfined aquifers, it is equal to the specific storage volume of the system [L3 · t−1 ]
times the thickness of the aquifer plus the specific
The left side of Eq. 3 corresponds to the left side of
yield [−].
the continuity equation and describes the divergence of
Specific Yield (Sy ) is the volume of water removed flux at a point in the aquifer. In other words, it describes
from a unit volume of an unconfined aquifer for a the groundwater flow vector resolved into x, y, and z
unit drop in water level; this quantity is also defined components (see Fig. 1).
as the volume of water that will drain by gravity The right side of Eq. 3 corresponds to the right side
from a saturated unit volume of an aquifer and is of the continuity equation and describes the change in
equivalent to the effective porosity [−]. storage at the given point in the aquifer. The term ∂h/∂t
Specific Storage (Ss ) refers to the change in hydraulic head, with respect to
time, at the same point in the aquifer. The term Ss is the
The main equation that governs the flow of a fluid in a specific storage, and it relates the changes in head in the
porous medium is given by aquifer to the change in the amount of water stored in the
∂qx ∂qy ∂qz ∂h aquifer at a certain point. In other words, it is a quantity
+ + = Ss (1) that tells us the volume of water that can be removed from
∂x ∂y ∂z ∂t
a unit volume of a confined aquifer for a unit drop in head.
where q = specific discharge (i.e., groundwater flow) in the
x, y, or z direction [L · t−1 ]
CONFINED AQUIFERS—A CONCEPTUAL
h = hydraulic head [L]
UNDERSTANDING
t = time [t]
x, y, z = lengths in the principal directions in our Consider a confined aquifer that consists of a layer of
coordinate systems [L] sand overlain by an impermeable layer of clay that acts
Ss = specific storage [L−1 ] as a confining layer. If a well is drilled into the aquifer
Darcy’s law and water is pumped from the well, the heads around
dh the well will lower and create a cone of depression in the
q=K (2)
dl potentiometric surface (Fig. 2). The more water that is
pumped from the well, the larger the cone of depression.
where K = hydraulic conductivity [L · t−1 ] In the example shown in Fig. 2, the pumping well has
dh/dl = hydraulic gradient, or change in head with removed a volume of water (represented by the volume
length [−] in the barrel), and in response, the water levels in

Flow vector in 3-D Flow vector in 3-D

z z z

Resolved
vector in
z-direction
y y y
Resolved
vector in y-direction
x x x
Resolved
vector in
x-direction

Figure 1. Flow vector resolved into component vectors in three dimensions.

482 STORAGE COEFFICIENT

the well and in the surrounding aquifer have dropped.

However, the saturated thickness of the aquifer itself has Unit volume of a
not changed (i.e., the aquifer is still full to the base of confined aquifer
the confining layer). What is the source of this water?
Additional
The conceptual answer deals with elastic deformation and
component
compressibility of the aquifer materials and the water. of pressure
This aquifer is made up of sand grains and pore spaces
that are full of water. If there is no confining layer and
the aquifer is simply recharged by water infiltrating from
above, the pore spaces are full of water, and the pore
fluid pressure is equal to atmospheric pressure plus the
overlying column of water. In this case, the volume of
water in the pore space is equal to the volume of the
pore space.
However, if the aquifer is confined and an additional
component of pressure is added to the water, that pressure
will force a greater volume of water into the pores of the
sand. It does this by compressing the sand grains slightly,
pushing the sand grains around so that pore volume
is increased, and compressing the water itself slightly.
Increasing the pore fluid pressure causes the aquifer to
undergo elastic (i.e., recoverable) deformation, so that 1-meter
1 meter
a subsequent decrease in fluid pressure will cause an
Figure 4. Unit volume of a confined aquifer.
associated decrease in volume (Fig. 3).
This is the source of the additional volume of water.
Pumping water from the confined aquifer lowers the excess
fluid pressure (i.e., the pressure head above the top of the aquifer), causing the aquifer to ‘relax’ and yield water that
is stored by the excess pressure.

Q CONFINED AQUIFERS—A MATHEMATICAL

UNDERSTANDING
Potentiometric
surface
Consider a unit volume of a confined aquifer. This unit
volume is saturated with water, and the water has an
additional component of pressure that would cause the
Confining layer water level in a manometer to rise above the top of the
cube (Fig. 4). If a volume of water is removed from the
cube, it will reduce the additional component of pressure
and lower the head in the cube. Removing more water
Confined aquifer
will cause more head loss; less water will cause less head
loss. Therefore, there is a specific volume of water that,
if removed, will cause a unit drop in head. This volume
of water that results in a unit drop in head in the unit
Figure 2. Pumping well in a confined aquifer. volume of aquifer is the specific storage.
The specific storage is also called the elastic storage
coefficient; it deals only with the amount of water
removed from elastic storage. The specific storage is
a function of the compressibility of the aquifer matrix
∆h (i.e., the material that makes up the aquifer) and the
compressibility of the water and is strictly defined as

Ss = ρw g(α + nβ) (5)

where ρw = density of water [M/L3 ]

Sand grains in an aquifer
Figure 3. Elastic deformation in a sand aquifer.
g = gravitational constant [L/t2 ]
α = compressibility of the aquifer matrix [1/M/Lt2 ]
n = porosity [−]
β = compressibility of water [1/M/Lt2 ]
 QANATS: AN INGENIOUS SUSTAINABLE GROUNDWATER RESOURCE SYSTEM
STORATIVITY IN CONFINED AQUIFERS
The specific storage actually refers to an infinitesimal
point in the aquifer. However, in real aquifers, the
thickness of the aquifer influences how much water the
aquifer will yield; a thicker aquifer will transmit and yield
more water than a thinner aquifer. Multiplying the specific
storage by the saturated thickness of the aquifer results
in storativity (S):
S = Ss b (6)
where b = saturated thickness of the aquifer [L].
In other words, the storativity describes the volume of
water removed from a unit area of aquifer for a unit drop
in head.
UNCONFINED AQUIFERS AND SPECIFIC YIELD
In a confined aquifer, all the water pumped out through
a well is provided by the release of water from elastic
storage. In an unconfined aquifer, the same thing
occurs—the changes in head result in a release of water
from elastic storage in the aquifer. However, in an
unconfined aquifer, the pumping well is dewatering a
section of the aquifer (Fig. 5), and this part of the aquifer’s
response adds an additional volume of water to the volume
withdrawn from elastic storage.
This additional volume of water is released from the
pore spaces and is due to the specific yield of the aquifer.
Specific yield is the volume of water that drains from a
rock under the force of gravity and is directly related to
the effective porosity of the aquifer matrix. The volume
of water released from a unit volume of aquifer due to
dewatering of the aquifer is therefore the unit volume
times the specific yield.
So, in unconfined aquifers, the volume of water released
from a unit volume of aquifer is a function of both
the specific yield (i.e., the water removed from the pore
space) and the specific storage (the water released from
elastic storage in the aquifer). Therefore, in an unconfined
Water is removed
from the pore spaces Q
Potentiometric
surface
or water table
Unconfined aquifer
Figure 5. Pumping well in an unconfined aquifer.
483
aquifer, the storativity of the aquifer is the sum of the
specific yield and the thickness times the specific storage.
S = Sy + bSs (7)
However, the magnitudes of these parameters are
drastically different: specific yield values in an aquifer
generally range from 1% to 50%, or 0.01 to 0.5; specific
storage values generally range from 1 × 10−3 to 1 × 10−6 ,
or 0.001 to 0.000001, which means that, in an unconfined
aquifer, the volume of water provided by release from
elastic storage is orders of magnitude less than that
provided by dewatering the pores; so the specific storage is
generally considered negligible, and unconfined storativity
is a function only of the specific yield.
QANATS: AN INGENIOUS SUSTAINABLE
GROUNDWATER RESOURCE SYSTEM
BAHRAM SAGHAFIAN
Soil Conservation and
Watershed Management
Research Institute
Tehran, Iran
A qanat system consists of an underground tunnel that
conveys water by gravity from highland aquifers to
the ground surface at lower elevations. Blended with
ingenuity, its purpose is simply to collect and bring
underground water to the surface for use. Qanat, invented
and originated in Iran more than 2700 years ago, is
considered a hydraulic engineering system of an Old World
era providing water to settlements since its invention. The
overwhelming climate of Iran is arid and semiarid; two
thirds of the country area in central, southern, and eastern
parts receive an average of about 150 mm of rainfall
annually. Yet, Iran is considered a farming country where
diverse crops are grown. This remarkable achievement in
agricultural development of arid lands has materialized
partly by building qanats to tap underground water in
regions where surface water is lacking or unreliable during
seasons of high water demand. Numerous urban and rural
qanat-watered settlements are vivid examples that not all
civilizations were rooted by the sides of rivers. Qanats
have helped to transform barren lands into productive
assets. Thanks to its simple and efficient technique, the
idea behind this sustainable water resource system was
passed from one generation to the next and traversed
geographic boundaries. Qanat’s use is not confined to Iran
and has been adopted in many regions across the world.
QANAT DESCRIPTION
Britannica characterizes a qanat as, ‘‘A qanat taps
underground mountain water sources trapped in and
beneath the upper reaches of alluvial fans and channels
the water downhill through a series of tunnels, often
several kilometres long, to the places where it is needed
for irrigation and domestic use . . ..’’ The dictionary of
484 QANATS: AN INGENIOUS SUSTAINABLE GROUNDWATER RESOURCE SYSTEM
(a)
Qanat
outlet
Qanat gallery
Water table
(b)
Impermeable layer
Figure 1. Schematic plan (a) and cross section (b) of a qanat system.
Islamic Architecture defines qanat as a ‘‘subterranean
canal system usually used to bring water some distance
from a river or mountains.’’ More technically, Lightfoot (1)
describes a qanat as ‘‘a form of subterranean aqueduct, or
subsurface canal, engineered to collect groundwater and
direct it through a gently sloping underground conduit to
surface canals which provide water to agricultural fields.’’
As shown in Fig. 1, a qanat consists of a gently sloped
underground tunnel, or gallery, penetrating the upland
aquifer at one end and appearing in the ground surface
at the other end. The gallery, is also vertically connected
to the ground by several wells, or shafts, dug along its
length, often at 20–30 m to a maximum of 100 m intervals,
providing ventilation and access during construction and
maintenance. Put differently, a qanat consists of a series
of vertical shafts in sloping ground, inter-connected at the
bottom by a tunnel with a gradient flatter than that of the
ground. The excavated soil is piled up around the entrance
of the wells to give some protection in times of storm
surface runoff. The first shaft, called the mother well or
main shaft, is usually sunk into the alluvial fan to a level
below the groundwater table. The line of qanat wells looks
like anthills in a distinctive plan view. The typical cross
section of a qanat gallery is 1.5 m high and about 1 m
wide, which is just sufficient to allow men and boys to do
the digging and cleaning.
Qanats can be classified into hill and river-based. Hill
qanats have their main shafts based in the aquifers located
in hill slopes. This type of qanat is not very long. If
this type of qanat taps a deep aquifer, it will enjoy
a relatively constant discharge. A river-based qanat is
Shafts
Mother well
Shafts
normally recharged from a seasonal river, stretching over
a long distance where the tunnel meets the ground surface.
Other terms for qanat, pronounced kanat, as it is
called in different parts of the world, are karez (e.g., in
Afghanistan and eastern Iran), foggora (e.g., in Algeria),
hattara (e.g., in Morrocco), and falaj (e.g., in Oman).
HISTORY AND GEOGRAPHIC EXTENT OF QANATS
Written records and recent excavations prove that qanats
first appeared in ancient Iran (then Persia). Wulff (2)
describes the origin of qanat: ‘‘As early as the seventh
century B.C., the Assyrian king Sargon II reported that
during a campaign in Persia he had found an underground
system for tapping water in operation near Lake Urmia.
His son, King Sennacherib, applied the secret of using
underground conduits in building an irrigation system
around Nineveh, and he constructed a qanat on the
Persian model to supply water for the city of Arbela.’’
The earliest reports of qanats indicated that the water
for the city of Ula northwest of Lake Urmia, located
in northwest of Iran, was provided by qanats some 700
B.C. (3). English (4) reports that, based on early mining, the
mountains of Kurdistan in western Iran, eastern Turkey,
and northern Iraq were the birthplace of the qanat more
than 2500 years ago.
The technique was then widely diffused during the
expansion of Achaemenid dynasty, ruling the Persian
Empire, beginning sometime in the fifth century B.C.
Qanats were constructed in Mesopotamia to the shores
of the Mediterranean to the west, in Egypt and the
Arabian Peninsula to the south, and in Afghanistan,
 QANATS: AN INGENIOUS SUSTAINABLE GROUNDWATER RESOURCE SYSTEM 485

Central Asia, and western China to the east. The Arabs

introduced the idea further to the west during the second
diffusion phase in the eighth century A.D. The greatest
developments of qanat technology outside the Persian
core area are in southern Morocco (Marrakesh and the
Tafilalet) and southern Algeria (5). Wulff (2) reported
the excavation of several particularly interesting qanats
constructed by Arab rulers of early medieval times in the
Middle East. These include a small qanat built by the
caliph of Damascus to supply water for his palace in 728
A.D. A century later, presumably with the aid of Persian
engineers, the caliph Mutawakkil in Iraq constructed a
qanat that brought water to his residence at Samarra
from the upper Tigris River 300 miles away.
Qanats are reported to exist in Spain, Sicily, Cyprus,
the Canary Islands, and in South American countries such
as Mexico, Peru, and Chile. The Madrid qanat system in Figure 2. A top view of a qanat during construction.
Spain, bringing water for urban use, still functions. The
qanat systems of Tehuacan and Parras in Mexico are
believed of Spanish origin, but those of Parras (Coahuila) requires great skill and endurance. Figure 2 shows a top
and Pica (Chile) may be pre-Columbian (6). English (5) view of qanat during construction.
and Beekman et al. (7) discuss scenarios of the diffusion Muqannis may encounter a number of dangers in their
of qanat technology throughout the world. work. Soft soil in the tunnel can cause collapse of the roof.
In such cases, the tunnel cross section is braced with baked
CONSTRUCTING A QANAT clay collars to control cave-ins (Fig. 3). Another danger is
air low in oxygen. Any change in the burning of the
The laborious construction of a new qanat is a task to illuminating oil lamps is an indication of such danger. And
be engineered by an expert, who usually heads a team of last, but not least, is the hazard during the breakthrough
diggers. The diggers, called muqannis, are responsible for into the mother well, which must be emptied.
building and repairing qanats. The expert must decide on
a number of key issues prior to qanat construction: the
site of the mother well and the alignment and the slope
of the underground gallery. In his work, the expert uses
simple instruments such as a long rope and a level. As the
mother well of a qanat is usually dug in an alluvial fan
at the foot of a mountain or hillside, the expert watches
for signs of seepage on the ground, slight variations in
the vegetation, and available groundwater. Once the site
is chosen, one or sometimes more trial wells are sunk.
The muqannis then start digging the trial well, normally
1 m in diameter. They use a windlass at the top and haul
up the soil excavated by their teammates. They continue
the digging and spoil the material around the rim of the
well until the water table is struck. The expert then should
determine whether the aquifer water has acceptable water
yield. If so, he measures the distance from the mother well
to the qanat outlet, where water is planned for use. Once
the general alignment of the qanat is determined, the
expert calculates the slope of the qanat gallery. This slope
must allow the water to flow, yet prevent erosion of the
bottom and walls of the gallery. The slope for short qanats
may run from 0.5 to 2 in 1000, whereas longer qanats
usually have gentler slopes. In steep land, the slope of
the gallery is maintained in a looping course. The work
then proceeds from the outlet of qanat by excavating the
tunnel back toward the mother well. Vertical shafts at a
distance of 30 m, or so, are dug to provide ventilation and
access. For shafts deeper than 100 m, the muqannis set up
another windlass halfway within the shaft for a two-stage
haul up of spoil. The muqannis maintain a straight course
by sighting along a pair of burning oil lamps. The task Figure 3. Cross section of a qanat gallery protected by a lining.
486 QANATS: AN INGENIOUS SUSTAINABLE GROUNDWATER RESOURCE SYSTEM
Muqannis are highly respected for their courage, hard
work, and know-how. They are paid well for their work,
and older muqannis are considered blessed or lucky.
QANATS AND WATER REGULATIONS
In Iran, the allotment of water delivered by qanats is
regulated by certain laws, which are rooted in a traditional
system of water rights. Normally a water master or bailiff,
appointed by a qanat owner(s), supervises the water
distribution according to each farmer’s time share. As
Wulff (2) has noted, ‘‘the bailiff is guided by an allocation
system that has been fixed for hundreds of years. For
instance, three hamlets in a region west of Iran, still
receive the shares that were allotted to them in the 17th
century by the civil engineer in the reign of Shah Abbass.
The hamlets of Dastgerd and Parvar in central Iran are
entitled to eight shares apiece. These allocations are built
into outlets from the qanat distribution basin.’’ English (4)
provides a detailed overview of qanat ownership, the
complex water distribution system, and, through some
examples, how fragmentation of qanat water rights has
taken place. He makes a note of strict and unforgiving
adherence to communal methods of water rotation and
upkeep of the water supply system serving to maintain
the social cohesion of qanat-watered villages.
QANATS: COSTLY AND TIME-CONSUMING
Qanats are labor-intensive and time-consuming to build.
Both the cost and time to build a new qanat varies with
topography, depth of the water table, underground soil
type, length of the qanat, type of ownership, availability
and experience of muqannis, and local conditions. Thus,
reports differ in figures provided for the cost and duration
of construction of qanats.
Noting that qanats are long-lasting water resources
systems if properly maintained, Wulff (2) believes that the
agricultural production sustained by qanat water repays
the investment in construction and maintenance by 10%
to 25% depending on the size of the qanat, the water
yield, and the crop type. Wulff’s assessment of the cost
in the 1960s showed that a 6-mile long qanat required
US$13,500 to US$34,000 to build. The cost runs to about
$90,000 for a qanat 10 to 15 miles long.
English (8), on the other hand, reported that it took
27 years to construct a 1-km qanat in Kerman, in central
Iran, with a mother well 45 meter deep, due to three
changes in ownership. Another 3-km qanat with two
mother wells, 50 and 55 m deep, and a bifurcated tunnel
took 17 years to build. The costs of these two qanats,
respectively, were approximately $10,000 to $11,000 per
km in the 1960s. As another example, the cost of
constructing a 40-km long qanat near the city of Kerman
with a mother well 90 m deep was roughly $213,000 when
completed in 1950 (4). Such expenses adjusted for inflation
can be raised only through government subsidies or large
corporations in modern Iran. Due to high construction
investment, qanat owners often try to increase the yield.
They may extend branches from the main tunnel to reach
additional aquifers or excavate the floor of the existing
tunnel to tap water at a deeper level (2).
Note that, although agricultural production cannot
repay the investment and maintenance costs of new qanats
in some regions, existing qanats are still the second
water supply system in volume providing for drinking
and domestic use in Iran. Keeping the qanat cost and
the enormous building effort in mind, individual farmers
are unwilling and unable to fund the construction of new
qanats or even maintain old ones. Thus, farmers have
turned to drilled wells as the more affordable alternative.
QANAT VERSUS DRILLED WELL
People in arid regions rely on underground aquifers for
their water supply. The aquifers can be tapped by a
qanat, or alternatively by a drilled well, and the water
is directed to an area where it is consumed. The volume of
water produced by a qanat depends upon several factors:
extent, type, yield, and recharge rate and area of the
aquifer. Comparing a drilled well with a qanat system,
there are advantages for each. The following compare the
two systems:
• As history proves it, qanats are sustainable renew-
able water resource systems. However, they cannot
meet today’s increasing demand. Wells are the alter-
native at the expense of inducing negative water
balance in the aquifers they exploit.
• Qanats need no extra energy to harvest water,
whereas wells need diesel or electric power to bring
water to the surface.
• Qanats can be built and maintained using low-
technology tools and material locally available.
• Qanats can be used as a passive cooling system in
conjunction with a wind tower (9).
• Qanats reinforce social cooperation, an important
issue in traditional societies.
• Wells are dug in a short time, whereas qanat
construction is slow, investment-intensive, and quite
time-consuming. However, technology can be put to
work to accelerate qanat construction by cutting the
time and the cost.
• Only a portion of a qanat’s water is used during
nonirrigation season, unless conserved by other
means. During periods of low water use in fall
and winter, watertight gates can seal off the qanat
outlet damming up the water. In spring and summer,
night flow may be stored in small ponds for daytime
use (10). A well pump can be switched off when water
is not needed.
• A typical qanat’s discharge is less than that of
pumped wells. The latter are favored to meet growing
demands, though, in a nonsustainable manner.
Experts in Iran now suggest an integrated approach of
conjunctive use of qanats and wells. The location of a
well must be optimized with respect to an existing qanat
recharge area.

STATISTICS OF QANATS
No one knows how many qanats were ever built, how many
are still in operation, and how much water they deliver.
Available reports vary in numbers, albeit they have
collected qanat statistics in different times. It is estimated
that, as of the 1960s, between 40,000 to 50,000 qanats had
been built in Iran during its history, of which some 27,000
running qanats were still used (11). Wulff (2) reported
that some 22000 qanats in Iran, totaling 170,000 miles
of underground conduits built by manual labor, had a
total yield of 19,500 cubic feet of water per second, which
amounts to 17 billion cubic meters of water annually. Two
recent reports prepared in Iran put the total number of
qanats, as of 1995, between 25,000 and 33,000, but it
is unclear whether all are currently used. The reports,
however, agree on the total volume of water provided by
qanats at around 9 billion cubic meters. Having accepted
this figure and the figure given by Wulff in 1968, an
alarming decline of about 50% in water yield of qanats is
inferred, which must be partly due to an excessive number
of wells causing underground water tables to drop and
affect qanat yield. The declining groundwater tables and
increasing number of abandoned qanats have shrunk the
share of water supplied by qanats to less than 10% of total
use.
The discharge of individual qanats varies widely with
the location and, in some cases, time of the year. Qanat
discharges of a few liters to as much as 1 cubic meter per
second have been reported.
There are some unique qanats in Iran described
as follows:
1. Vazvan Qanat in Meimeh, Isfahan province, is
1800 m long. At a distance of 600 meter from its
outlet, a stone dam has been built to store the water.
A total of six gates have been installed at different
elevations to control the flow of water, depending on
the time and amount needed. This qanat system is
essentially an underground dam.
2. The well-known and unique Moun Qanat in
Ardestan, Isfahan province, is a twin-gallery qanat
system, the only one of its kind reported in the world.
The two galleries are independent conduits, about
3 m apart in elevation. A thin impermeable layer
is believed to separate the two qanat conduits. The
water in the layers is not mixed, as the upper conduit
is bent around the shafts. The qanat is about 2 km
long, and its total discharge waters 25 ha.
3. Gonabad region, in Khorasan province northeast of
Iran, is known for some of the oldest qanats in
Iran. Reports indicate that one qanat has a mother
shaft with a depth of 140 meters and a length of
about 70 kilometers. Another qanat is reported with
a 270-m mother well.
BIBLIOGRAPHY
1. Lightfoot, D.R. (1997). QANATS in the Levant: Hydraulic
technology at the periphery of early empires. Technology and
Culture 38(2): 451.
LYSIMETERS 487
2. Wulff, H.E. (1968). The qanats of Iran. Sci. Am. April:
94–105.
3. Laessøe, J. (1951). The irrigation system at Ulhu, 8th century
BC. J. Cuneiform Stud. 5: 21–32.
4. English, P.W. (1998). Qanats and lifeworlds in Iranian
plateau villages. Yale F&ES Bull. 103: 187–205.
5. English, P.W. (1968). The origin and spread of qanats in the
Old World. Proc. Am. Philos. Soc. 112: 170–181.
6. Humlum, J. (1965). Underjordiske vandingskanaler: kareza,
qanat, foggara. Kultergeografi 16: 81–132.
7. Beekman, C.S., Weigand, P.S., and Pint, J.J. (1999). Old
World irrigation technology in a New World context: qanats
in Spanish Colonial Western Mexico. Antiquity 73(279):
440–446.
8. English, P.W. (1966). City and Village in Iran. University of
Wisconsin Press, Madison, WI.
9. Bahadori, M.N. (1978). Passive cooling systems in Iranian
architecture. Sci. Am. February: 144–154.
10. Beaumont, P. (1989). The qanat: a means of water provision
from groundwater sources. In: Qanat, Kariz, and Khettara,
Wishbech. P. Beaumont, M. Bonine, and K. McLachlan (Eds.).
Menas Press, pp. 13–31.
11. Goblot, H. (1979). Les Qanats: Une Technique d ’Acquisition
de l ’Eau. Mouton, Paris.
LYSIMETERS
KRISTOFOR R. BRYE
University of Arkansas
Fayetteville, Arkansas
Lysimetry is a measurement technique that employs
sampling devices called lysimeters to measure the
movement of water and solutes through a porous
medium. Lysimeters have most commonly been used to
quantify water movement (i.e., drainage) through soil and
typically consist of a porous plate surface of varying pore
sizes through which water moves from the soil into a
collection reservoir.
BRIEF HISTORY OF LYSIMETERS
According to Joffe (1), lysimeter studies were initiated in
the late eighteenth century to address the fate of pre-
cipitation in soil. Many lysimeter studies then shifted to
consider leaching of soil and precipitation-deposited con-
stituents, such as ammonium and sulfur (1). Gradually,
lysimeters were adopted to investigate fertilizer losses
from agricultural and nonagricultural soil (1).
Early lysimeter research involved greatly disturbed
soil profiles, unrepresentative of natural soil properties.
Lysimeter tanks were dug and refilled with soil after
construction was complete (1). For this reason, the
validity of drainage and leaching measurements were
severely questioned, as they are today. Nonetheless, many
technological and methodological advancements have been
made since the late 1700s, but the same fundamental
question remains today as it did then: how to measure
natural water movement through undisturbed soil?
488 LYSIMETERS
DRAINAGE LYSIMETERS
Two general types of lysimeters exist for monitoring water
and solute fluxes through soil, weighing and nonweighing
lysimeters. Weighing lysimeters are large-scale sampling
devices typically used for measuring evapotranspiration,
drainage, and solute leaching. Inputs and losses of
moisture are evaluated on the basis of changes in mass in
a containerized column of soil within the lysimeter (2).
Weighing lysimeters are expensive, labor-intensive to
construct, and usually involve filling the lysimeter by
reconstructing a disturbed soil profile (3,4), although large
weighing lysimeters that have undisturbed soil columns
have been constructed, such as the monolith lysimeters
near Coshocton, Ohio. However, nonweighing lysimeters
are better suited for less burdensome, replicable field
measurements of in situ water and solute fluxes.
Nonweighing lysimeters are well adapted to monitoring
drainage and leaching from soils. Nonweighing lysimeters,
also known as drainage or pan lysimeters, are typically
much smaller than weighing lysimeters. An advantage
of a nonweighing pan lysimeter is that replicable field
measurements can be obtained below undisturbed soil
profiles. As a result, site-specific information can be
gathered relative to the mass flux of water and leachate
concentrations of soluble organic and inorganic chemical
species (3). Coupling drainage measurements with solute
concentrations provides a means of estimating solute mass
fluxes through an undisturbed soil profile.
DRAINAGE MEASUREMENTS USING NONWEIGHING
LYSIMETERS
The collection of leachate through the porous surface
of a lysimeter depends on two conditions: (a) minimal
discontinuities in the flow path of moving water and (b) a
driving force for water movement. Maintaining a capillary
connection between the soil and the porous surface of the
lysimeter is imperative for high-quality results. Neglecting
to minimize potential sources of error may cause ponding,
preferential, and bypass flow around the lysimeter (3).
Drainage lysimeters have been designed to operate
where a variety of forces act on soil water. Drainage and
solute fluxes are measured and replicated in the field by
some variation of a nonweighing pan lysimeter operating
under tension (i.e., tension lysimeters) or tensionless
(i.e., zero-tension lysimeters). In soil or any other porous
material, the bulk of the water moves in response to
gradients in gravitational and/or matric potentials. Zero-
tension lysimeters rely on the force of gravity for water
movement through the lysimeter’s porous plate surface.
The effect of a matric potential gradient is created
by applying suction to a tension lysimeter to quantify
drainage and leaching.
Both types of nonweighing lysimeters have several
significant disadvantages. Nonweighing lysimeters are
commonly operated after the lysimeter has been filled
with soil, similar to weighing lysimeters. As a result, soil
horizonation, structure, pore size distributions, and many
other physical properties no longer reflect natural soil
conditions. Therefore, using a disturbed profile technique
for general lysimeter installation may generate biased and
uninterpretable results (4).
Kung (6) described another disadvantage of nonweigh-
ing lysimeters, in which funnel flow may divert percolating
water based on the angle of inclination of coarse textured
soil lenses. Lenses, or soil textural discontinuities, may
cause water to bypass the lysimeter or channel water
directly to it. In either case, results may be erroneous.
Tension or tensionless column lysimeters that have
sidewalls extending above the soil surface pose additional
challenges (7–9). Impeding runoff, as the design of
extended-walled column lysimeters do, unnaturally alters
the water status of the contained soil column by permitting
ponding during winter months and following intense
rainfall. Furthermore, crop management becomes more
difficult as the extended sidewalls act as obstacles to
plowing, seeding, and cultivating (9).
Zero-tension lysimeters rely on the formation of a
saturated soil zone above the porous plate before drainage
occurs by gravitational flow, which is an unlikely natural
condition in most deep, well-drained and moderately
well-drained soils and alters the water content and soil
water potential of the rooting zone (7). The prerequisite
of saturation above these sampling devices may cause
drainage water to bypass the lysimeter (3) and create
unnatural matric potential gradients that cause water
movement from the saturated zone to the surrounding
unsaturated soil matrix (10). Zero-tension lysimeters that
have upwardly extended sidewalls (7,8,11) may promote
preferential flow down the side of the soil column upon
drying (11). In addition, lateral flow that normally occurs
under field conditions would be restricted in containerized,
column lysimeters with fine textured, clayey horizons (12).
Fixed-tension lysimeters improve the flow patterns
around and through the lysimeter’s porous plate surface.
Stored water often moves through soil under some tension.
Applying tension (i.e., suction or vacuum) to a porous
surface produces more natural flow patterns through a
soil profile than relying on gravity to generate drainage.
However, fixed-tension lysimeters typically experience
vacuum depletion shortly after vacuum application unless
constant vacuum is maintained. Fixed-tension lysimeters
do not reproduce natural soil water potential variations
that can occur during a relatively short period of
time. The transient nature of soil water potentials
influences unsaturated hydraulic conductivity and thus
alters drainage fluxes.
RECENT USAGE OF NONWEIGHING DRAINAGE
LYSIMETERS
Drainage lysimeters, both tension and tensionless, have
been frequently employed in numerous studies to accom-
plish a variety of research goals. Zero-tension lysime-
ters have been used to study salt and water movement
and leaching losses under conventional and no-tillage
maize [Zea mays; (13)]; collect mobile colloids transport-
ing organic and heavy metal compounds (11); investigate
nitrate (NO− 3 ) and pesticide leaching from removed, but
intact, soil columns (12,14); monitor the fate and cycling
of 15 N on revegetated mine spoils in containerized, soil
 LYSIMETERS 489

mini-columns (7); determine the effect of irrigation on the

Equilibrium - tension lysimeter
nitrate leaching potential of agroecosystems (4); measure
nitrate leaching from maize agroecosystems under zero,
intermediate, and high fertilizer N rates (3); and elucidate
the effect of various cropping systems on solute leaching 25 cm
and 15 N migration (8,9).
Jemison and Fox (10) evaluated the collection efficiency
of zero-tension lysimeters. A bromide (Br− ) tracer method
was used to determine an overall mean collection effi- 75 cm
ciency, based on lysimeter recovery of applied Br− , of 52%,
ranging from 13–92%, and high plot-to-plot variability of
Vacuum
volumes were collected. Cumulative observed drainage, tube
uncorrected for Br− collection efficiency, measured from
zero-tension lysimeters for 3 years, was consistently lower Spring plate
than LEACHM-predicted drainage (10). LEACHM (Leach- Drain tube
ing Estimation And Chemistry Model) is a complex com-
puter model capable of simulating water infiltration and
redistribution, solute transport and distribution in the Figure 1. Stainless steel equilibrium-tension lysimeter (ETL)
for measuring drainage and solute leaching below an undisturbed
soil, and tracer ion uptake by plants (15).
soil column.
The use of fixed-tension drainage lysimeters has
been somewhat more limited. Karnok and Kucharski (16)
described the use of an underground rhizotron-lysimeter
facility constructed to monitor root growth, evapotran-
spiration, and leaching under turf grass cover. Leaching
losses from observation cells were collected under a con-
tinuous, fixed tension of −33.3 kpa. Dolan et al. (17) used
wedge-shaped, fixed-tension microlysimeters to ascertain
the effect of nitrogen placement on nitrate leaching below
maize monoculture grown in sandy soil. Adams et al. (18)
evaluated nitrate leaching from poultry litter and manure-
amended pasture soil using pan lysimeters at a fixed
tension of ∼ −10 kpa. Despite several disadvantages,
fixed-tension and tensionless nonweighing lysimeters have
many potential uses and applications.

EQUILIBRIUM-TENSION LYSIMETERS

Equilibrium-tension lysimetry is an innovative and im-

proved nonweighing drainage lysimeter technique that
allows applied-lysimeter suction to mimic the surrounding
bulk soil matric potential fluctuations; thus it minimizes
the potential for bypass flow and preserves natural
drainage patterns (19). A heat dissipation sensor (19,20)
is used to monitor the bulk-soil matric potential, and
suction applied to the lysimeters is varied as the
bulk-soil matric potential varies. Lysimeter suction was
originally set manually every several days (19), but a
recently developed automated vacuum system maintains
equilibrium between applied-lysimeter suction and bulk-
soil matric potential fluctuations within seconds (21).
The use of equilibrium-tension lysimeters is relatively
Figure 2. Backhoe digging trench for lysimeter installation.
new and even more limited compared to the other types
of nonweighing drainage lysimeters (Fig. 1). Equilibrium-
tension lysimeters have been used to quantify drainage
losses for a multiyear period in a water balance study of a
restored prairie and maize agroecosystems in southern
Wisconsin (22). At the same study site in southern agroecosystems. Equilibrium-tension lysimeters have also
Wisconsin, equilibrium-tension lysimeters have been used been used to quantify heavy metal and antibiotic leaching
to quantify multiple years of nitrogen and carbon (23,24) from the root zone of tall fescue (Festuca arundinacea)
and soluble phosphorus fluxes (25) in drainage water from amended with poultry litter in northwest Arkansas
below the root zones of a restored prairie and maize (unpublished data).
490 LYSIMETERS

GENERAL TENSION-LYSIMETER INSTALLATION

AND SAMPLING PROCEDURES

For the most representative results in a field setting,

nonweighing drainage lysimeters are typically installed
below an undisturbed soil column. A trench is dug by
hand or with a backhoe, then a small cavity that extends
into the plot or measurement area is excavated in the
long sidewall of the trench (Fig. 2). The lysimeter is forced
upward against the base of the undisturbed soil column.
Some apparatus (e.g., a spring plate, shim, or screw jack)
is installed below the lysimeter to provide the necessary
upward force to maintain the capillary connection between
the lysimeter’s porous plate surface and the soil above
(Fig. 3). Once the cavity is outfitted with the lysimeter, the
trench can either be left open and accessible or completely
backfilled. If the trench will be backfilled, tube extensions
Figure 3. Rectangular cavity excavated in sidewall of the trench
that extend upward out of the soil can be connected to the
with nearly completed lysimeter installation.
drain and vacuum tubes attached to the lysimeter. The
lysimeter cavity itself is typically left open but protected
with plywood from damage during backfilling (Fig. 4). The
trench is backfilled by adding soil in the opposite order in
which it was removed and is revegetated, if necessary, so
that only the tube extensions and heat dissipation sensor
wires if used, are visible above the soil surface (Fig. 5).
Water collected in the lysimeter’s collection reservoir
is typically removed under vacuum. The leachate volume
is recorded, and, if necessary, a portion is saved and
refrigerated for subsequent chemical analysis.

BIBLIOGRAPHY

1. Joffe, J.S. (1932). Lysimeter studies. I. Moisture percolation

through soil profiles. Soil Sci. 34: 123–143.
2. Soil Science Society of America (SSSA). (1996). Glossary of
Soil Science Terms. Soil Sci. Soc. Am., Madison, WI.
3. Jemison, J.M. and Fox, R.H. (1994). Nitrate leaching from
Figure 4. Complete lysimeter installation with cavity protected nitrogen-fertilized and manured corn measured with zero-
from damage upon backfilling with treat plywood. tension pan lysimeters. J. Environ. Qual. 23: 337–343.
4. Martin, E.C., Loudon, T.L., Ritchie, J.T., and Werner, A.
(1994). Use of drainage lysimeters to evaluate nitrogen
and irrigation management strategies to minimize nitrate
leaching in maize production. Trans. ASAE. 37: 79–83.
5. Owens, L.B. (1987). Nitrate leaching losses from monolith
lysimeters as influenced by nitrapyrin. J. Environ. Qual. 16:
34–38.
6. Kung, K.J.S. (1993). Laboratory observation of funnel flow
mechanism and its influence on solute transport. J. Environ.
Qual. 22: 91–102.
7. Reeder, J.D. (1986). A non-weighing lysimeter design for
field studies using nitrogen-15. Soil Sci. Soc. Am. J. 50:
1224–1227.
8. Shipitalo, M.J. and Edwards, W.M. (1993). Seasonal patterns
of water and chemical movement in tilled and no-till column
lysimeters. Soil Sci. Soc. Am. J. 57: 218–223.
9. Moyer, J.W., Saporito, L.S., and Janke, R.R. (1996). Design,
construction, and installation of an intact soil core lysimeter.
Agron. J. 88: 253–256.
10. Jemison, J.M. and Fox, R.H. (1992). Estimation of zero-
Figure 5. Soil surface following lysimeter installation and trench tension pan lysimeter collection efficiency. Soil Sci. 154:
backfilling. 85–94.

11. Thompson, M.L. and Scharf, R.L. (1994). An improved zero-
tension lysimeter to monitor colloid transport in soils. J.
Environ. Qual. 23: 378–383.
12. Bergstrom, L. and Johansson, R. (1991). Leaching of nitrate
from monolith lysimeters of different types of agricultural
soils. J. Environ. Qual. 20: 801–807.
13. Tyler, D.D. and Thomas, G.W. (1977). Lysimeter measure-
ments of nitrate and chloride losses from soil under conven-
tional and no-tillage corn. J. Environ. Qual. 6: 63–66.
14. Bergstrom, L. (1990). Leaching of chlorsulfuron and met-
sulfuron methyl in three Swedish soils measured in field
lysimeters. J. Environ. Qual. 19: 701–706.
15. Jemison, J.M., Jabro, J.D., and Fox, R.H. (1994). Evaluation
of LEACHM: I. Simulation of drainage, bromide leaching, and
corn bromide uptake. Agron. J. 86: 843–851.
16. Karnok, K.J. and Kucharski, R.T. (1982). Design and con-
struction of a rhizotron-lysimeter facility at the Ohio State
University. Agron. J. 74: 152–156.
17. Dolan, P.W., Lowery, B., Fermanich, K.J., Wollenhaupt, N.C.,
and McSweeney, K. (1993). Nitrogen placement and leaching
in a ridge-tillage system. pp. 176–183. In: Conference Pro-
ceedings on Agricultural Research to Protect Water Quality.
Feb. 21–24, 1993. Minneapolis, MN. Soil Water Conserv. Soc.,
Ankeny, IA.
18. Adams, P.L. et al. (1994). Poultry litter and manure con-
tributions to nitrate leaching through the vadose zone. Soil
Sci. Soc. Am. J. 58: 1206–1211.
19. Brye, K.R., Norman, J.M., Bundy, L.G., and Gower, S.T.
(1999). An equilibrium tension lysimeter for measuring
drainage through soil. Soil Sci. Soc. Am. J. 63: 536–542.
20. Reece, C.F. (1996). Evaluation of a line heat dissipation
sensor for measuring soil matric potential. Soil Sci. Soc.
Am. J. 60: 1022–1028.
21. Masarik, K.C., Norman, J.M., Brye, K.R., and Baker, J.M.
(2004). An improved automated approach to equilibrium
tension lysimeters. J. Environ. Qual. In Press.
22. Brye, K.R., Norman, J.M., Bundy, L.G., and Gower, S.T.
(2000). Water budget evaluation of prairie and maize
ecosystems. Soil Sci. Soc. Am. J. 64: 715–724.
23. Brye, K.R., Norman, J.M., Bundy, L.G., and Gower, S.T.
(2001). Nitrogen and carbon leaching in agroecosystems and
their role in denitrification potential. J. Environ. Qual. 30:
58–70.
24. Brye, K.R., Gower, S.T., Norman, J.M., and Bundy, L.G.
(2002). Carbon budgets for a prairie and agroecosystems:
Effects of land use and inter-annual variability. Ecol. Appl.
12: 962–979.
25. Brye, K.R. et al. (2002). Phosphorus leaching under a restored
tallgrass prairie and corn agroecosystems. J. Environ. Qual.
31: 769–781.
STEADY-STATE FLOW AQUIFER TESTS
HARENDRA S. CHAUHAN
G.B. Pant University of
Agriculture and Technology
Uttar Pradesh, India
Groundwater studies are conducted to evaluate the
occurrence, availability, and quality of groundwater.
There are numerous practical applications like optimum
development of groundwater basins, safe withdrawal,
STEADY-STATE FLOW AQUIFER TESTS 491
groundwater recharge, seepage from canals and reser-
voirs, and drainage that require knowledge of the
hydraulic characteristics of aquifers. Two basic aquifer
properties needed to make a quantitative assessment of
such applications are transmissivity and storage coeffi-
cient which represent, in a general sense, its abilities
to transmit and store water, respectively. The other
hydraulic properties of an aquifer are hydraulic resistance,
leakage factor, and drainage factor. Performing an aquifer
test is an effective way of determining these hydraulic
properties of aquifers. Steady-state flow aquifer tests are
conducted to determine transmissivity, hydraulic conduc-
tivity, and hydraulic resistance; more detailed unsteady-
state tests are conducted if other hydraulic properties
are also to be evaluated. The objective of this article is
to discuss steady-state methods of determining aquifer
characteristics.
AQUIFER TESTS
The procedure involves pumping water from a well for
a certain time and at a certain rate and measuring its
effect on the water table in the well and at a number
of piezometers installed in the vicinity. When pumping
is carried out in a well, the water level in its vicinity is
lowered; the extent of lowering at any point is termed
the, drawdown. The drawdown is maximum at the well
point and reduces gradually with distance from the well,
becoming zero at a certain distance termed the radius
of influence. This distance varies with the well and
depends on the well discharge, formation constants, and
the duration of pumping. The lowering of the water level
in the well during pumping creates a head difference or
hydraulic gradient from the surroundings which causes
the flow of water in the well. The hydraulic gradient
increases as flow converges toward the well. The rate of
flow is same through all the concentric cylindrical surfaces
around the well, and the surface area decreases as one
approaches the well. The form of the surface resembles a
cone, and therefore it is known as the cone of depression,
which is the characteristic feature of every pumped well.
Its size and shape depend on the discharge, length of
pumping, aquifer characteristic, and other factors.
Steady-State and Unsteady-State Flows
Both steady-state and unsteady-state groundwater flow
conditions occur in nature and cannot be very distinctly
separated. It is a matter of interpretation and choice as to
which concept is used for practical utility. Although every
phenomenon in nature is transient, it could be considered
a series of successive steady-state events.
Steady-state groundwater movement requires that the
boundary conditions of flow and aquifer characteristics
remain unchanging with time. Even though the aquifer
characteristics do not remain constant, they are assumed
so for practical purposes. Thus, if a hydraulic disturbance
like well pumping is created, even though the process
is initially unsteady, eventually it becomes steady
with time.
492 STEADY-STATE FLOW AQUIFER TESTS
DEFINITION OF AQUIFER PROPERTIES
The transmissivity (T) of an aquifer is the rate of flow
under a unit hydraulic gradient through a vertical strip
of the aquifer of unit width and extending through the
whole saturated thickness of the aquifer. It has the
dimensions of length2 /time and, for example, may be
expressed in m2 /day. T is the product of the average
hydraulic conductivity (k) and the saturated thickness (b)
of the aquifer.
Hydraulic conductivity (k) is the proportionality
constant k in Darcy’s law v = ki, where v is the specific
discharge or Darcy’s velocity (L T−1 ) and i is the hydraulic
gradient (dimensionless), which is the volume of water
that will flow through a porous medium in unit time
under a unit hydraulic gradient through a unit area
measured at right angles to the direction of flow. k
is a parameter that depends on both the properties of
the porous medium and of the fluid and has units of
length/time, for example, m/day.
Saturated thickness (b) is equal to the physical
thickness of the aquifer between two confining layers in a
confined aquifer. The saturated thickness for unconfined
aquifers, equal to the difference between the free water
table and the impervious layer, is not constant but varies
with pumping.
Coefficient of storage (S) of an aquifer is the volume of
water released from or taken into storage per unit surface
area of the aquifer per unit change in head. In unconfined
aquifers, S is known as the specific yield which represents
the volume of water released from storage per unit surface
area of aquifer per unit decline of the water table.
Specific yield is also sometimes called effective porosity
or drainable pore space. S is a dimensionless quantity.
Hydraulic resistance (c) is a property of semiconfined
aquifers that characterizes the resistance of a water-
bearing layer to vertical flow. It is also known as the
reciprocal leakage coefficient and is given by b /k where
b and k are the thickness and the hydraulic conductivity,
respectively, of a semipervious layer. The dimension of the
hydraulic resistance is time and is, for example, expressed
in days.
Leakage factor (L) describes the spatial distribution of
leakage through one or more semipervious layers into √ a
leaky aquifer or vice versa. It is defined as L = kbc,
has the dimension of length, and it may be expressed
in meters.
Of these aquifer parameters, transmissivity (T),
hydraulic conductivity (k), and hydraulic resistance (c) are
the properties that can be determined from steady-state
aquifer tests.
STEADY-STATE METHODS OF DETERMINING
AQUIFER PARAMETERS
Different well flow theories have been derived based on
the consideration of a continuity equation in cylindrical or
spherical coordinates, well depth, the fully or partially
penetrating nature of wells; steady-state or unsteady-
state flow conditions; gravity, artesian, leaky or nonleaky
aquifer; constant or variable discharge conditions; and
sinks or source. The methods of determining aquifer
constants depend on the solution of the selected continuity
equation and the specific boundary conditions. Steady-
state solutions can be used to estimate transmissivity;
the storage coefficient (S) cannot be determined because
it does not appear in the solution. The following methods
based on steady-state approaches can be used to determine
the transmissivity of an aquifer.
Unconfined Aquifer Flows
Radial Flow to a Fully Penetrating Constant Discharge
Well (Fig. 1). The flow is considered assuming that the
aquifer is homogeneous, isotropic, of uniform thickness,
and has an infinite aerial extent. The differential equation
for groundwater flow in an unconfined aquifer may be
written in cylindrical coordinates as


d r dh
=0 (1)
dr dr
In a well that fully penetrates the gravity aquifer,
the flow will be axisymmetric. Neglecting the vertical
component of the velocity as small, the equipotential
surfaces may be considered coaxial cylinders. The flow
across
these cylinders
 at a distance r is given by Q = av =
dh
2π rh −k . Equation (1) may be integrated twice. The
dr
two constants may be obtained from an assumed boundary
condition of a piezometric head at the well as h(r0 ) = h0
and the relationship of the well flow at a given distance r.
The solution of Eq. 1 under the boundary conditions gives
the phreatic surface of the flow as
Q r
h2 = ln + h2 (2)
πk r0
Three methods may be developed from Eq. 2 to
determine transmissivity.
Determination of T Using Two Observation Wells:
Dupuit–Thiem Method. It can be seen from Eq. 2 that there
is no fixed saturated thickness in gravity aquifers and only
k appears in the solution describing phreatic surfaces.
Thus, only k can be obtained from the solution. To obtain
the transmissivity, k has to be multiplied by some average
depth. Measuring heads at two distances r1 , r2 (r2 ≥
Ground surfaceQ
Water table
s
Drawdown
curve Unconfined aquifer
r H
h
h R
2r
Figure 1. Flow to an unconfined aquifer.
 STEADY-STATE FLOW AQUIFER TESTS 493

r1 ), h(r1 ) = h1 , h(r2 ) = h2 in Eq. 2, an expression for k may

Ground surface Q
be written as
Q r2 Piezometric
k= ln (3) surface
π h(h22 − h21 ) r1
s
Drawdown
Contrary to Dupuit’s assumption, the vertical compo- curve
nents of velocity are large in gravity flow and cannot be r
ignored. Equation 3, consequently, does not give correct H
R
results. Setting h = H − s, Eq. 3 can be rewritten as h 2r h

Q r2
kH = 

 ln (4) b Confined aquifer Hydraulic
s2 s2 r1 conductivity k
2π s1 − 1 − s2 − 2
2H 2H
Figure 2. Flow to a confined aquifer.
where H is the saturated thickness of the aquifer, s1 and
s2 s2
s2 are drawdowns and s1 − 1 and s2 − 2 are corrected
2H 2H Writing h = H − s, transmissibility may be written as
drawdowns s1 and s2 ,
Transmissibility may be written in terms of corrected Q R Q R
drawdowns or, neglecting correction terms, it may also be kH =
 ln ≈ ln (9)
s20 r0 2π s0 r0
expressed as 2π s0 −
2H
Q r2 Q r2
T = kH = ln ≈ ln Confined Aquifer Flows (Fig. 2)
2π(s1 − s2 ) r1 2π(s1 − s2 ) r1
Similar to the unconfined aquifer, flow is considered to take
Q r2
≈ ln (5) place assuming that the aquifer is homogeneous, isotropic,
2π(h2 − h1 ) r1 of uniform thickness, and of infinite aerial extent.

Determination of T Using a Single Observation Well. Due

Radial Flow to a Fully Penetrating, Constant Discharge
to constraints on funds for pumping tests, it may be
Well. Differential equations of the groundwater flow to a
possible to construct only one observation well. For
fully penetrating well in a confined aquifer may be written
such a situation, one may have observations only in
in cylindrical coordinates as
a single well. For such conditions, setting h(r0 ) = h0
and h(r1 ) = h1 , (r1 ≥ r0 ), an expression for k may be

d dh
written as r =0 (10)
Q r1 dr dr
k= ln (6)
π(h21 − h20 ) r0
In a fully penetrating aquifer of thickness b, the
flow is axisymmetric, and the equipotential surfaces are
In the same way for such boundary conditions, T
coaxial circular cylinders. Flow across these cylinders
may be written in terms of corrected drawdowns. Con-
in a fully penetrating artesian aquifer of thickness b
sidering that corrections have small values, they may at a given distance r from the well center is given as
be neglected to give approximate values of transmissiv-

dh
ity as Q = av = 2π rb −k . Equation 10 may be integrated
dr
twice under the assumed condition of the piezometric head
Q r1 Q at the well, h(r0 ) = h0 , and the relationship of the well at
kH = 

 ln ≈
s20 s21 r0 2π(s0 − s1 ) a given distance r to give
2π s0 − − s1 −
2H 2H
Q r
r1 Q r1 h= ln + h0 (11)
ln ≈ ln (7) 2π kb r0
r0 2π(h1 − h0 ) r0
Thiem’s (1906) Equilibrium Equation Using Two Observa-
Determination of T Using the Head in the Pumped tion Wells. To reduce errors because of losses in the well
Well. Due to further constraints on funds, one may not and through the well screen, Thiem (1) proposed mea-
be able to make even a single observation well. For such a suring piezometric heads h1 , h2 in two observation wells
condition, measuring the head in the pumped well itself, located at distances r1 , r2 from a constantly discharging
h(r0 ) = h0 , and considering an arbitrary radius of influence well Q. Thus, setting h(r1 ) = h1 , h(r2 ) = h2 in Eq. 11 of an
h(R) = H, the hydraulic conductivity may be obtained as expression of transmissivity may be derived as

Q R Q r2
k= ln (8) T = kb = ln (12)
π(H2 − h20 ) r0 2π(h2 − h1 ) r1
494 STEADY-STATE FLOW AQUIFER TESTS

Determination of T Using a Single Observation Well. Con- Considering heads at two observation wells located
sidering the head as h(r1 ) = h1 at a single piezometer at distances r1 , r2 as h1 , h2 , an expression for k may be
installed at distance r1 from the well, T may be obtained obtained as

from Eq. 11 as 1 1
Q −
r1 r2
k= (17)
Q r1 2π(h2 − h1 )
T = kb = ln (13)
2π(h1 − h0 ) r0
Determination of k Considering the Head in the Well.
Using the measurement of the head in the pumped well
Determination of T Using the Piezometric Head in the itself, h(r0 ) = h0 , and considering an arbitrary radius of
Pumped Well. By measuring the piezometric head in the influence R, an expression for k in terms of the head and
pumped well itself and at an arbitrary radius of influence drawdown s = H − h may be written as
R, h(R) = H, T can be obtained from Eq. 11 as



1 1 1 1
Q R Q R Q − Q −
T = kb = ln = ln r0 R r0 R
2π(H − h0 ) r0 2π s0 r0
(14) k= = (18)
2π(H − h0 ) 2π s0

Spherical Flow to a Nonpenetrating, Constant Discharging Considering that R > r0 , 1/R is negligible and can be
Well in An Extensively Thick Artesian Aquifer (Fig. 3). The neglected; this gives
groundwater flow equation for such a flow system is
described by a continuity equation in spherical coordi- Q Q
k≈ ≈ (19)
nates as 2π r0 (H − h0 ) 2π r0 s0


d 2 dh
r =0 (15) Semiconfined Aquifer Flows
dr dr
Aquifer Constants from the Steady-State Flow of a Well
The isopiestic surfaces to a nonpenetrating well consist in a Semiconfined Aquifer: Hantush and Jacob method (2)
of concentric hemispheres. The flow across a hemispherical (Fig. 4). The solution is based on the assumption that the
surface at distance r is given by phreatic surface remains constant, so that the leakage
through the semiconfining layer takes place in proportion
Q = av = 2π r2 (−kdh/dr) to the drawdown of the piezometric level. Hantush and
Jacob (2) obtained a solution for L > 3D, as
Considering the piezometric head at the well, h(r0 ) = Q r
h0 , and the flow equation above, the piezometric surface sm = K0 (20)
2π kb L
can be obtained as

 where sm = steady-state drawdown in meters at
1 1
Q − distance r;
r0 r Q = discharge of the pumped well in cubic
h= + h0 (16)
4π k meters per day;
√
L = kbc, leakage factor in meters;
Determination of k Using Two Observation Wells. It may b
be observed that from Eq. 16, only hydraulic conductivity c =  , hydraulic resistance of a semipervious
k
can be determined as the thickness of the aquifer is layer in meters; and
not finite. K0 (x) = modified Bessel’s function of the first kind
and zero order.
Hantush observed that if r/L is <0.05, Eq. 20 can be
Ground surfaceQ approximated as
Piezometric

2.3Q L
surface sm = log 1.12 (21)
2π kb r
s
Drawdown
curve A plot of sm against r on semilog paper, with r on the
r H logarithmic scale, will show a straight line relationship in
h the range where r/L is small. The slope of the straight line,
h R
2r the drawdown difference per log cycle of r, is expressed as

r 2.3Q
r Thick confined aquifer
sm =
2π kb
2.3Q
Figure 3. Spherical flow to a thick confined aquifer. kb = (22)
2π sm
 STEADY-STATE FLOW AQUIFER TESTS 495

Approximate Methods
Ground surface Q
Initial water Logan’s (3) Method for confined aquifers. Equation 14
table using the observation in the pumped well itself is also
s known as Logan’s method. According to Logan (3), the
Drawdown accuracy of the calculation depends on the accuracy in
curve Ponded water measuring the maximum drawdown in the well s0 and on
r the accuracy of the ratio R/r0 . The ratio R/r0 may have a
H
R large value, but the ratio of the logarithm may be small. He
h h Piezometric head suggested an approximate value of 3.33 for the log ratio.
2r in main aquifer
Putting this value in Eq. 14 gives the transmissivity as
Semi confining Leak- Hydraulic
b′ 2.3Q R 1.22Q
stratum age conductivity k ′
kb = log ≈ (24)
2π s0 r0 s0

b Main aquifer Hydraulic

conductivity k Logan’s (3) Method for Unconfined Aquifers. Using a
similar approach and considering observation in the
pumped well itself, the transmissivity in an unconfined
Figure 4. Flow to a semiconfined aquifer. aquifer may be obtained from Eq. 9 as

2, 3Q R 1.22Q
For any point sm , r is selected on the straight line. kH = log ≈ (25)
2π s0 r0 s0
Putting r in Eq. 21 gives L.c can be obtained as

b L2 s20
c= = where s0 ≡ s0 − .
k kb 2H

Alternatively the extended straight line portion of the Transmissivity from Other Analytical Solutions
curve intercepts the r axis where the drawdown is zero. At
the interception point, sm = 0, r = ro , and Eq. 21 reduces to In principle, any analytical solution of a boundary value

 problem of groundwater flow can be used to estimate
L aquifer parameters, if other variables can be measured
log 1.12 =0
r or estimated. A few steady-state solutions for estimating
transmissivity are presented.
or √
L kbc
1.12 = 1.12 =1
r0 r0 Two Equal, Constant Discharge, Mutually Interfering
Wells. The relationships between two equal and constant
or discharge interfering wells distance B apart may be
(r0 /1.12)2
c= written as
kb 2π kb(ha − h0 )
Q1 = Q2 = (26)
Transmissivity from Steady Flow in a Semiconfined Aquifer: R2
ln
Ernst’s Modification of the Thiem Method. The relationship r0 B
is based on assuming that the well discharge is corrected
by subtracting, from the total discharge, the amount of where r0 = well radius, ha = average piezometric head at
phreatic water obtained from the semipervious layer. the external boundary, h0 = piezometric head at the well,
The determination requires two piezometers installed and R = radius of influence (R > B).
in semipervious layers at distances r1 , r2 . The coaxial Thus if the values of the constant equal discharges.
phreatic water from a circular area may be written Q, h0 , ha , R, and B are known, the transmissivity may be
approximately as r0 = r1 + 1/3(r2 − r1 ). determined from the relationship
The phreatic water delivered from a circular area is
R2
Q = π r20 s S × 24 Q ln
r0 B
kb = (27)

2π(ha − h0 )
where s = average drop of phreatic level in the
semipervious layer during the last hour (m) and S =
Steady-State Flow to a Partially Penetrating Well. A
specific yield of the semiconfined aquifer (to be estimated
relationship was obtained for such a flow system by De
from well logs). The transmissivity may be found from
Glee (4) as
r2
(Q − Q ) ln

r1 Qp 2 π hs 0.20
kb = (23) h2b − h0 = ln +
2π(s1 − s2 ) 4π k hs 2r0 b
496 STEADY-STATE FLOW AQUIFER TESTS
where hs is the depth of the partially penetrating well, b
is the aquifer thickness, h2b is the initial piezometric head
at distance 2b, and h0 is the piezometric head in the well.
hs
Also 1.3 hs ≤ b; ≥5
2rw
From this relationship and measuring other parame-
ters, the transmissivity may be found as
 
Qp b 2 π hs 0.20
kb = ln + (28)
4π(h2b − h0 ) hs 2r0 b test. Observations should be taken at close intervals
For a partially penetrating well in an unconfined
aquifer where the drawdown is also small in relation to
the saturated thickness, Eq. 28 can be used by replacing b
by H.
Kozeny (5) proposed a relationship, by which transmis-
sivity may be determined from
r at hourly intervals should be sufficient.
2.3Qp log
r0
T= (29)
2π(h − h0 ) G
where
 
r0 πα
G=α 1+7 cos
2hs 2
Qp is the discharge of a partially penetrating well, Qf
is the discharge of a fully penetrating well, G = Qp /Qf is
a correction factor giving the ratio of the two discharges,
α is the fractional penetration of the well screen hs /b, and
hs is the depth of penetration of the aquifer.
PRACTICAL ASPECTS OF PUMPING TESTS (after Ref. 6)
Observation Well Size, Location, and Numbers
An observation well should be just large enough to allow
accurate and rapid measurement of water levels. Small
wells provide a fast response. Five-cm observation wells
are usually effective when hand methods of measuring
water levels are used. Larger pipes may be used with
automatic recorders. Observation wells are commonly
installed with 1–2 m long screens at the bottom embedded
in the aquifer to measure the pressure head. Observation
wells should be installed at about the same depth as the
middle of the well screen in the pumped well.
Observation wells should be located at a reasonably
large distance and not too close to the pumped well.
Setting observation wells at distances of 30–100 m from
the pumped well may work out best in most cases.
When observation wells are too close to the pumped well,
correct drawdown readings may be difficult. The cone of
depression in a water table spreads at a slower rate than
in an artesian formation, so the observation wells for
a water table aquifer should be relatively closer to the
pumped well.
The number of observation wells depend upon the
amount of information required and the funds available
for the test program. Drawdown data from the well
itself or from a single observation well permit estimating
the average hydraulic conductivity, transmissibility, and
coefficient of storage of the aquifer. If two or more
observation wells are placed at different distances, the test
data can be analyzed both in terms of the time–drawdown
and the distance–drawdown relationships. It is always
better to have as many observation wells as conditions may
allow. A minimum of three placed at varying distances and
radially at equal angles around the well is desirable.
Measurement of Water Levels. The depth to water may
be measured many times during the course of a pumping
during the first 2 hours, and the time between readings
should gradually increase as the test continues. The water
level measurement may be recorded to the nearest 0.5 cm
in the observation wells. Measurements may be made
every half-minute during the first 5 minutes after starting
the pump, then every 5 minutes for an hour, then every 20
minutes for about two hours. From this point on, readings
Water-level measurements can be taken in various
ways, the wetted-tape method, mechanical sounder, elec-
tric water-level indicator, pressure gauge, or pressure
logger. For detailed information on these devices, see
Kruisman and De Ridder (7). Fairly accurate measure-
ments of water levels can be made manually.
Measurements of Discharges. The other parameter that
needs to be measured is the discharge from wells during
pumping tests. To avoid complicated calculations later,
the discharge rate should preferably be kept constant
throughout the test. The discharge rate can be measured
using various devices like commercial water meters,
flumes, orifice weirs, containers or the jet-stream method.
The procedures for installation and measurements using
these devices are available in standard texts.
SUMMARY
An overview of steady-state approaches for determining
the transmissivity of confined, unconfined, and semicon-
fined aquifers has been presented. Detailed procedures
for determining transmissivity using drawdown data in
a pumped well and one or two observation wells are
discussed in the context of two most common methods of
Thiem and Dupuit–Thiem. A few approximate methods
and analytical solutions for determining aquifer con-
stants have been briefly introduced. General guidelines for
conducting pumping tests including measurement of dis-
charge and groundwater levels and numbers and locations
of observation wells have been furnished.
BIBLIOGRAPHY
1. Thiem, G. (1906). Hydrologische Methoden. Leipzig, p. 56.
2. Hantush, M.S. and Jacob, C.E. (1955). Non steady radial flow
in an infinite leaky aquifer. Trans. Amer. Geophysical Union.
36: 95–100.
3. Logan, J. (1064). Estimating Transmissibility from Routine
Production of Tests of Water Wells Ground Water 2(1): 35–37.

4. De Glee, G.J. (1930). Over Grundwaterstroomingen bij
Wateronttrekking Door Middle van Patten. J. Waltman Jr.,
Delft, the Netherlands, p. 175.
5. Kozeny, J. (1933). Theorie und Berech nug der Brunneu Wasser
Kraft and Wasser Wartschift. 28: 88–92,113–116.
6. Johnson Div. UDP, Inc. (1972). Ground Water and Wells. 2nd
Edn. Edward E. Johnson, St. Paul, MN, p. 440.
7. Kruisman, G.P. and De Ridder, N.A. (1970). Analysis and
Evaluation of Pumping Test Data, International Institute
for Land Reclamation and Improvement. Wageningen, the
Netherlands.
TIDAL EFFICIENCY
TODD RASMUSSEN
The University of Georgia
Athens, Georgia
Ocean tides often cause water levels in coastal aquifers
to fluctuate. These fluctuations result from two different
causes. In confined aquifers, the additional weight of the
water on the surface increases the pressure on the water
at depth, which causes an increase in water levels in
observation wells located near the coast. In this case, no
movement of water between the ocean and the aquifer
may have occurred. In unconfined aquifers, however, the
loading effect does not occur. Instead, water that has
moved directly through the pores within the aquifer
has caused the water level response. Regardless of the
mechanism, both processes can be described with the
phase lag, or delay, between the tidal peak and the aquifer
response peak, as well as the amplitude ratio between the
magnitudes of the response and the tides
Water levels in wells located near the coast often
fluctuate with a rhythm that mimics the rise and fall
of the ocean. The tidal efficiency, TE, is the ratio of the
change in water levels in wells, W, to the change in tidal
levels, T:
W
TE =
T
where both water levels are measured with the same type
of units (e.g., feet, meters, psi). Tidal efficiencies can range
from a maximum of one in poorly consolidated geologic
media to a minimum of zero in rigid aquifers (1,2).
The tidal response in an aquifer can result from two
types of phenomena, one in confined and a second in
unconfined aquifers. As described in greater detail below,
the confined aquifer response is caused by the increased
load placed on the surface, whereas the unconfined aquifer
response is caused by the movement of water in response
to changing tidal water levels.
CONFINED AQUIFERS
In confined aquifers, coastal tides affect water levels
because the surface load changes as ocean levels fluctuate.
The weight of the ocean compresses the underlying aquifer,
which causes a change in the fluid pressure. In this
situation, no water moves between the ocean and the
TIDAL EFFICIENCY 497
aquifer, which is similar to the effect that trains have as
they pass a well in some confined aquifers—water levels
rise and then fall as a train passes by (3).
The loading effect is greater in poorly consolidated
aquifers where a greater proportion of the overlying load
is borne by the fluid. Pore fluids in consolidated aquifers
are less likely to carry the load, and so these aquifers have
lower tidal efficiencies.
Coastal aquifers are not the only type of aquifer that
shows a response to surface loads. The loading effect
on confined aquifers can be observed in response to
the placement or removal of any type of mass on the
ground surface. Components of the water budget (such as
precipitation, evaporatranspiration, and soil moisture) can
immediately affect water levels in confined aquifers—even
when no direct hydraulic communication is present (4).
An increase in water levels in a confined aquifer
under a waste lagoon may be mistakenly attributed to
contamination from the lagoon. The cause could be from
downward leakage of water from above, or it could also be
from the increased surface load causing a rapid increase
in fluid pressure.
Another example of surface loading develops when
barometric pressure changes cause water levels in wells
to fluctuate. In these cases, the response is somewhat
complicated because the barometric pressure also exerts a
force on water levels in open wells. The tidal efficiency in
confined aquifers is related to the barometric efficiency,
BE, in that the two efficiencies must sum to unity,
TE + BE = 1. Thus, aquifers with low tidal efficiencies
have large barometric efficiencies, and vice versa. The
reader is referred to BAROMETRIC EFFICIENCY for additional
discussion of this topic.
UNCONFINED AQUIFERS
A second tidal influence occurs in unconfined, surficial
coastal aquifers. In these aquifers, water levels change
as the hydraulic perturbation created by changing ocean
levels moves horizontally through the saturated zone
toward the well. Unlike the confined case in which the
load is transmitted without concomitant fluid movement,
water is free to move through the unconfined aquifer.
For a simple, one-dimensional flow geometry (i.e., a
well placed at some distance from a linear shoreline), the
response is a function of the distance of the well from
the shoreline, L, and the aquifer hydraulic diffusivity,
D = T/S, where T and S are the aquifer transmissivity and
storativity, repsectively. Nearby wells in more diffusive
aquifers respond more rapidly than do more distant wells
in aquifers with lower diffusivity. The response is a
function of a dimensionless variable, κ = Dτ/L2 , where
τ is the response time.
This type of response is not limited to coastal areas,
however. A similar type of response is found in surficial
aquifers next to streams. In this case, changing water
levels during floods cause a loss of water from the stream
into the adjacent aquifer, with a reversal of flow later in
the flood hydrograph. This type of response is also referred
to as bank storage (5).
498 COMBINED FREE AND POROUS FLOW IN THE SUBSURFACE
SINUSOIDAL OSCILLATIONS
Many types of loading phenomena display periodic
variation, tides being a common example. In addition to the
gravitational affects of the sun and moon on the oceans,
gravitational forces also cause changes in atmospheric
pressure, called atmospheric tides, and even in the solid
earth, called earth tides.
These periodic variations can be characterized with the
period of the oscillation (e.g., 24 hours for solar tides)
and the amplitude, or magnitude, of the oscillation. When
observing the tidal response in confined aquifers, the tidal
efficiency is equal to the ratio of the magnitude of the
water level fluctuation in the well to the magnitude of the
ocean tides. For example, if peak daily ocean levels vary
by 1 m and the peak daily groundwater levels vary by 10
cm, then the tidal efficiency is 0.10, or 10%.
Another influence is the phase lag, which is the delay
between when the ocean tide and the aquifer water levels
peak. Normally, an insignificant delay occurs between the
ocean tide and a confined aquifer response; yet this delay
can be large in cases in which significant borehole storage
exists.
The tidal efficiency and phase lag in unconfined aquifers
can determine the aquifer diffusivity. The transmissivity
and storativity can be only determined, however, if the flow
of water can be determined, which requires measurements
of the flow rate between the ocean (or stream) and the
aquifer.
BIBLIOGRAPHY
1. Bear, J. (1972). Dynamics of Fluids in Porous Media. Dover
Publications, New York.
2. Walton, W.C. (1970). Groundwater Resource Evaluation.
McGraw-Hill, New York.
3. Jacob, C.E. (1940). On the flow of water in an elastic artesian
aquifer. Am. Geophys. Union Trans. 21: 574–586.
4. vanderKamp, G. and Schmidt, R. (1997). Monitoring of total
soil moisture on a scale of hectares using groundwater
piezometers. Geophys. Res. Lett. 24(6): 719–722.
5. Todd, D.K. (1959). Ground Water Hydrology. John Wiley &
Sons, New York.
COMBINED FREE AND POROUS FLOW IN THE
SUBSURFACE
DIGANTA BHUSAN DAS
Oxford University
Oxford, United Kingdom
VAHID NASSEHI
Loughborough University
Loughborough, United
Kingdom
INTRODUCTION
The presence of a fluid layer adjacent to a porous medium
is common in many environmental, chemical, mechanical
and petroleum engineering problems. The flow behavior in
the fluid layer is determined by the properties of the fluid
(e.g., viscosity and density), and surface (e.g., pressure
and shear) and body (e.g., gravity and electric field) forces.
However, flow in the porous medium depends additionally
on the properties of the medium which include, among
others, porosity and permeability.
In fluid dynamic analysis of groundwater systems, it
is commonly found that water flows through combined
free (nonporous) and porous pathways between surface
and subsurface systems. Such coupled regimes may also
be present due to construction of industrial utilities and
structures in the ground. Important engineering processes
that may involve combined free and porous flow domains
are, for example, fluid losses/leaks from underground pipes
and storage tanks in old gas works sites; dig-and-treat,
pump-and-treat, and permeable reactive barrier technolo-
gies for groundwater treatment; and drilling/extraction
of oil from underground reservoirs. However, in most
cases, the associated transport phenomena are deter-
mined by natural hydroenvironmental conditions, which
include, for example, combined surface and subsurface
flow, lake–groundwater interactions, seepage through
preferential flow channels and macropores, circulatory
flows and rise and fall of groundwater, glaciology, and
multiple karstic regions interfaced with granular porous
sections. The subdomains in coupled flow systems are
generally distinguished by an interfacial surface, which
represents the transitional zone for contaminants/fluid
mobility from free to porous sections or vice versa. To
gain on understanding of the combined flow behavior, it is
necessary to develop realistic methodologies for evaluat-
ing mass and momentum transfer across the free/porous
flow interfaces.
The construction of well-posed mathematical formu-
lations for combining different flow regimes can present
difficulties. Two general approaches have evolved for this
purpose: (1) methodologies based on treating the combined
flow domain as a single continuum and (2) formulations
in which the flow domain is considered as a combi-
nation of discrete sections corresponding to different
flow regimes. In the former, the porous section is effec-
tively regarded as a pseudofluid layer, and the whole
domain is analyzed as a whole. As such, a suitably for-
mulated momentum equation for the entire domain in
conjunction with mass conservation equations is solved.
The mathematical model identifies fluid and porous
domains based on spatially varying permeability. Such
a single-domain approach is usually applied where the
flow transition from free to porous flow zones is not
distinct and properties of the porous domain change
progressively (e.g., metal solidification problems involv-
ing mushy zones). However, if the permeability of the
porous medium is constant and remains above a rel-
atively high level, then at defined interfaces between
adjacent zones, transitions between free/porous regimes
occur. Under these conditions, the domain should be
viewed as a combination of adjacent flow fields rather
than a geometric continuum. In these situations, the sec-
ond approach should be used where appropriate equations
of motion describing the flow in different subdomains are
 COMBINED FREE AND POROUS FLOW IN THE SUBSURFACE
solved. In general, a groundwater system corresponds to
this case.
GOVERNING MODEL EQUATION
The main aim of this article is to present a general
methodology that can be used to simulate hydrodynamic
behavior in combined free (nonporous sections) and porous
flow zones in the subsurface based on the assumption of
discrete flow domains, which involves selecting governing
model equations for each subdomain and the matching
conditions at the interface for linking them. Although
fluid mobility in combined free and porous flow on the
surface has been the focus of intense research for more
than two decades, analyses of subsurface flow have
not treated the possible interactions between free and
porous flow regimes efficiently. Addressing these issues,
the hydrodynamic conditions in underground zones of
coupled fluid and porous domains have been analyzed on
a 3-D spatial scale by Das et al. (1). This methodology
is based on the principles of mass and momentum
conservation and primitive variable formulations (i.e., the
equations are expressed in terms of the pressure and
velocity). Conservative forms of the equations have also
been adopted to enhance the accuracy of the numerical
calculations. The model equations are presented below.
Free Flow Region
The momentum transfer for constant density and viscosity
of a fluid at a reference temperature in a free flow region
can be represented by the well known Navier–Stokes
(N-S) equation,
dvf
∇Pf = −ρ + µ∇ 2 vf (1)
dt
where the subscript f refers to the free flowing fluid. P and
v are the pressure and velocity terms, and ρ and µ are the
constant density and viscosity of the fluid, respectively. In
Eq. 1, the body forces are automatically incorporated in
the pressure term as the resultant flow potential. The
following continuity equation that describes the mass
conservation for an incompressible fluid can be used for a
free flow regime:
∇ · vf = 0 (2)
Porous Flow Region
The corresponding macroscopic equations representing
the flow field in the porous section are more uncertain.
The momentum equations have received considerable
attention in the past. The two most suitable are the
Darcy (2) and Brinkman (3) equations shown below in
the transient form for anisotropic porous media where the
body forces have been incorporated in the pressure terms
to represent the driving flow potential:
∂vp
Darcy equation: ∇Pp = −ρ − µK −1 žvp (3)
∂t
∂vp
Brinkman equation: ∇Pp = −ρ − µK −1 žvp µ ∇ 2 vp (4)
∂t
499
In these equations, the subscript p refers to the porous
section; P is the interstitial averaged pressure, v is the
volume averaged seepage velocity; and ρ and µ are the
constant density and viscosity of the fluid, respectively.
µ/ is the effective viscosity which depends on the fluid
viscosity, µ, and the porosity. For flow of an incompressible
fluid in a continuum porous medium, the equation of
continuity is
∇ · vp = 0 (5)
where v is the volume averaged seepage velocity of fluid.
Matching Boundary Condition for the Darcy Equation
as a Porous Flow Equation. To couple the equations
representing the two domains, it is necessary to specify
the matching hydrodynamic conditions at the interfacial
region separating the two subdomains. Keeping the
three-dimensional (3-D) subsurface problem in mind,
the matching conditions for pressure (continuous or
discontinuous) and velocity components (normal and
tangential) must be specified at the interface. The
Darcy equation is a first-order differential equation with
respect to velocity and pressure, so it is not possible to
describe the velocity components fully at the free/porous
domain interface, which makes it impossible to link
the Darcy and the N-S equations directly because the
latter requires specifying all velocity components to solve
them. Special techniques have been invoked, henceforth,
for this purpose. Addressing this issue, Beavers and
Joseph (4) performed their well-known experiment on the
measurement of mass efflux of a Poiseulle fluid flowing
over a permeable block. They hypothesized the existence
of a slip velocity, commonly known as the Darcy slip
phenomenon, at the interface and a shear velocity in the
porous boundary layer. Consequently, an ad hoc boundary
condition was proposed based on the assumption that the
slip velocity is proportional to the shear effects. It allowed
jumps in the tangential velocities and gave a much-
needed relationship between the flow within and outside
porous media. The Beavers and Joseph (4) formulation
was represented as


∂u γ
= √ (uf − up ) (6)
∂y f K
where the x direction was taken as parallel and the
y direction as perpendicular to the interface. K is
the direction-independent geometric permeability of the
porous medium (isotropic), and γ is a slip coefficient that
depends only on the properties of the porous medium and is
independent of the fluid. The Beavers and Joseph condition
has now been modified for cases where the flow and the
interface are perpendicular rather than parallel, as in
the original formulation. Jones (5) derived the following
expression for the interfacial condition, assuming that
the Darcy slip phenomenon is due to shear stress at the
interface rather than a simple shear velocity:


∂v ∂u γ
+ = √ (uf − up ) (7)
∂x ∂y f K
500 COMBINED FREE AND POROUS FLOW IN THE SUBSURFACE
z This interfacial condition for pressure jump arises from
Interface
w
Free Porous
flow flow v Matching Boundary Condition for the Brinkman Equation
Figure 1. Model of combined free and porous flow domain for
extending the Beavers and Joseph (1967) formulation to the 3-D
subsurface region.
Direct extension of the above formulations to 3-D
cases poses further challenges. However, by resolving
the 3-D framework into two planes (x–y and x–z), it
is possible to use the Beavers and Joseph formulation
for three-dimensional modeling of combined flow in the
subsurface (1,2). The extension can be done based on
Fig. 1 where the x component of the velocity is normal
to the interfacial plane and the y and z components
are tangential (parallel). For the 3-D situation in Fig. 1,
postulating that the Darcy slip phenomenon takes place at
both x–y and x–z planes, the following relationships can be
obtained heuristically based on the Jones (5) modification
of Beavers and Joseph’s original formulation (1,2). The
velocity components that are parallel to the interface can
be calculated by using these expressions:


∂v ∂u γ flow problems, Das et al. (1) clearly demonstrated that
Velocity slip at x–y plane: + = 
∂x ∂y f Kyy


∂w ∂u γ in chemical, petroleum and mechanical engineering appli-
Velocity slip at x–z plane: + = √ (wf − wp )
∂x ∂z f Kzz
The slip coefficient is represented in Eqs. 8 and 9 by γyz
to indicate the plane on which it is active. Equation 8
is used to simulate the jump in the transverse (y)
velocity component across the interface, and Eq. 9 is
used for the vertical (z) velocity component. Kyy and
Kzz are the permeability in the y and z directions,
respectively. The above equations, therefore, represent
the slip-flow phenomena at the permeable interface
with respect to directional geometric permeability. Also,
the velocity component normal to the interface, the
longitudinal component, can be taken as continuous across
the interface:
uf = up (10)
However, the pressure does not necessarily remain
continuous across the interface and only under static
conditions, can it be said to be strictly continuous. If
the pressure is to be modeled as discontinuous at the y–z
interfacial plane during flow, the following relationship
is used:

∂u
−P + 2µ = −Pp (12)
∂x f
the fact that stress in the porous side is carried by the
porous materials, whereas in the free flow domain, the
stress is borne by the fluid medium, which causes a
discontinuity in stress, and hence fluid pressure, across
y the fluid/porous interface.
as a Porous Flow Equation. The use of the Brinkman
u
equation for mathematically describing the flow in
x porous flow zones apparently does not cause any
mathematical difficulty. The second-order derivatives of
the N-S equations and the Brinkman equation with
respect to velocity enable specifying all flow properties
as continuous across the interface, which facilitates
the task of coupling the equations of motion for the
free and permeable domains. The continuity of the
flow variables across the interface yields the following
matching boundary conditions:
uf = up (13a)
vf = vp (13b)
wf = wp (13c)
Pf = Pp (13d)
DISCUSSION AND CONCLUSIONS
Although there are superficial similarities in mathemat-
ical formulations of different coupled free and porous
(vf − vp )
combined flow behavior in the subsurface may differ sig-
(8) nificantly from other combined flow processes encountered
cations. Therefore, models to represent specific scenarios
(9) need to be developed. In the above studies by Das et al.,
natural flow behavior has been defined at the exit of the
domain by imposing stress-free boundary conditions. Das
et al. demonstrated that at the open end (exit) of the
domain, coexisting inflow and outflow might take place. In
other words, fluid from the surrounding media might come
inside the domain through the open end in addition to the
defined inlet. For the particular domain investigated, Das
et al. observed that although flow in the fluid domain is
mostly unidirectional toward the free/porous interface and
appears similar to pipe flow, it might become circulatory
on the porous side and reverse its direction. Evidence of
flow circulation, consistent with the simulated results of
Das et al., can also be found in the literature dealing with
various scenarios of subsurface flow (6). However, fluid
circulation is a complex transient process often involving
a moving center of circulation and shifts of the flow rever-
sal front to or from the interface. After a certain time
interval, the circulation may disappear, and the porous
flow may become unidirectional toward the exit of the
porous section. The main cause of such flow behavior has
been attributed to complex/variable pressure distributions
within the porous domain. The effects of a partially imper-
meable interface between the free and porous domains has

been explored by Das and Nassehi (7). It has been demon-
strated that due to the impermeable interface, there is a
possibility of buildup of fluid in the free flow region. The
influence of having a the multiple number of fluid/porous
interfaces in a flow domain has been demonstrated by
Das and Nassehi (8). This study showed that, in gen-
eral, the presence of a multiple number of interfaces
in the free flow domain has a minimal effect on flow
behavior. However, the flow patterns on the porous side
may vary depending on the direction of the free/porous
interfaces. Das and Nassehi (8) also examined concentra-
tion profiles of a conservative pollutant in associated free
flow and porous domains. The effects of different aspect
ratios of the subdomains on the hydrodynamic behav-
ior of free/porous regimes have been investigated by Das
and Nassehi (9). This study shows the influence of the
extent of free and porous domains, which govern the entire
flow behavior.
The study of combined free and porous flow in
the subsurface using different methodologies is still
relatively new in scientific literature. Although a number
of issues have been studied, future research aimed at
better understanding of combined flow behavior must be
continued. For example, investigations into the effects of
heterogeneity in porous media and multiphase flow (e.g.,
oil and water) in combined domains should be carried out.
A general methodology, which is not restricted to systems
where free/porous interfaces are aligned to coordinate
axes, also needs to be developed.
BIBLIOGRAPHY
1. Das, D.B., Nassehi, V., and Wakeman, R.J. (2002). A finite
volume model for the hydrodynamics of combined free and
porous flow in sub-surface regions. Adv. Environ. Res. 7(1):
35–58.
2. Darcy, H.P.G. (1856). Les Fontaines Publiques de la Ville de
Dijon, Exposition et Application des Principes à Suivre et des
Furmules à Employer dans les Questions de Distribution d’Eau.
Victor Dalmont, Paris.
3. Brinkman, H.C. (1947). A calculation of the viscous force
exerted by a flowing fluid on a dense swarm of particles.
Appl. Sci. Res. A1: 27–34.
4. Beavers, G.S. and Joseph, D.D. (1967). Boundary conditions at
naturally permeable wall. J. Fluid Mech. 30(1): 197–207.
5. Jones, I.P. (1973). Low Reynolds number flow past a porous
spherical shell. Proc. Cambridge Philos. Soc. 73: 231–
238.
6. Devito, K.J., Waddington, J.M., and Branfireun, B.A. (1997).
Flow reversal in peatlands influenced by local groundwater
systems. Hydrological Processes 11: 103–110.
7. Das, D.B. and Nassehi, V. (2001). LANDFLOW: A finite
volume model of combined free and porous flow of water in
contaminated land sites. Water Sci. Technol. 43(7): 55–64.
8. Das, D.B. and Nassehi, V. (2003). Modelling of contaminant
mobility in underground domains with multiple free/porous
interfaces. Water Resour. Res. 39(3), doi: 10. 1029/2002
WR001506.
9. Das, D.B. and Nassehi, V. (2002). Development of a new math-
ematical model for subsurface water quality management.
Water Sci. Technol. 45(9): 301–307.
GROUNDWATER TRACING 501
READING LIST
Das, D.B. (2001). Computer Modelling of Contaminant Mobility
in Subsurface Regions. Ph.D. Thesis, Chemical Engineering
Department, Loughborough University, Loughborough, UK.
GROUNDWATER TRACING
HAMID JAHANI R.
Water Research Institute
Hakimieh, Tehran, Iran
INTRODUCTION
Groundwater tracing is now a major field in groundwater
hydrology aiming mainly at demonstrating the direction
of groundwater flow and secondly, at giving as much
information on the hydrologic setting as possible. It has
been used for several decades to study flow conditions,
for groundwater reconnaissance, and for managing water
resources. During the last decades, water protection has
been one of the main issues due to strong pollution of the
environment. A groundwater tracer is a substance or a sort
of energy conveyed by water to be detected at a sampling
point and to be used in analyzing the temporal and
spatial distribution of water and the pollutant. Artificial
and natural tracers can help in various ways to assess
and to predict contaminant transport in surface and
groundwater.
Groundwater tracing can be a component of a large
hydrogeological program, or it can be a hydrogeological
program by itself. Nowadays, new field methods, new trac-
ers, and new measurement techniques are being developed
to make hydrogeologists more satisfied with results.
The advantage of groundwater tracing is that it
provides the possibility of a direct view of a groundwater
system, so the results can be relied on with assurance. In a
groundwater tracing program, depending on the question
and technical problem, both ‘‘natural’’ (those existing in
the water cycle) and ‘‘artificial’’ (those to be introduced
into groundwater system) tracers can be selected. The
main group of artificial tracers, is soluble tracers, such
as fluorescent dyes that are highly detectable, as well as
different salts. Another group of artificial tracers includes
drift bodies, such as spores, bacteria, or phages. The
applicability of each tracer depends on its properties
and on the condition of the flow environment. The use
of certain tracers, especially for groundwater flow, is
strongly restricted due to hydrogeologic conditions and
aquifer properties.
For artificial tracers, the methodology comprises
introducing the tracer into the groundwater system,
letting it travel together with the water, sampling, and
finally analyzing the water samples. For natural tracers,
the natural physical, chemical, or biological composition
of water is traced. For example, the natural temperature
of cold water from a spring could be a guide to identifying
snowmelt as the main water source and differentiating it
from local groundwater.
502 GROUNDWATER TRACING
THE HISTORICAL BACKGROUND
The groundwater tracing method is as old as the interest of
humans in nature. It was used while people were curious
about finding the source of undiscovered or unknown caves
or springs and used chaff, wood bits, or other drifts to
reveal the route of the water. The way to perform the test
is now almost as simple as it was in the past, at least
in most cases. But the way to analyze samples and the
results evolved into complicated analytical methods. The
development of new tracers and new analytical methods
during the last 10 years has led to more accuracy in the
results and also more ability to cover longer distances
in tracing tests. While using the method during the last
decades, it was aimed to find the direction of groundwater
flow and its velocity; now, its major application is in
studying groundwater pollution problems and protecting
the environment.
GOALS OF GROUNDWATER TRACING
The most important goals to be achieved in a tracing
program can be categorized into two groups:
Quantitative goals: To determine
• groundwater recharge amount
• groundwater discharge amount
• components of groundwater recharge
• age of groundwater
• elevation of recharge
• groundwater velocity
• volume of aquifer dynamic storage
• dispersion coefficient
• travel time
Quantitative goals: To determine
• origin of water
• hydraulic connection between different surface water
and groundwater bodies
• hydraulic connection between different basins
• spring catchment area
• aquifer recharge areas
• aquifer discharge areas
• aquifer boundaries
• groundwater flow direction and path
• flow regime
• role of structural features in groundwater flow
• saline water–freshwater interface
• source of pollution
• pollution distribution pattern
• degree of karstification
• seepage/leakage zones in dams and reservoirs
CHARACTERISTICS OF IDEAL TRACERS
Among the characteristics for selecting a tracer as an ideal
applicable tracer are the following:
The tracer
1. should be detectable in low concentrations. In
most cases, the tracer material is to be diluted
in a big volume of groundwater storage, so it
must be easily detectable at low concentrations
in water samples. The tracer concentration should
be easily differentiated from background values in
sampled water.
2. should be applicable as a nontoxic additive to water.
It is compulsory for the tracer material not to harm
flora and fauna.
3. should not be an absorptive material. The tracer
substance should not be easily adsorbed to aquifer
material; otherwise, it is possible that a considerable
portion of the injected tracer is lost during the test,
and it may be a major source of error in interpreting
the results. The tracer should move in the same
manner as water, and should not interact with
aquifer material.
4. should be economically available. In some tracing
experiments, the cost of tracer is a decisive factor.
5. should not change the hydraulic conductivity of the
medium under investigation.
6. should be stable enough during the test period from
injection until detection.
It is obvious that one could not find an ideal tracer that
meets all of these specifications. Although there are some
tracers developed specifically for groundwater tracing,
selecting a suitable tracer, as well as applying a suitable
amount of the tracer, is a state-of-the-art task and cannot
be done without sufficient experience, favorable expertise,
and knowledge about the medium under study.
GROUNDWATER TRACER TYPES
Groundwater tracers can be categorized into two main
groups of artificial (to be introduced into the system)
and natural (environmental or preexisting in the system)
tracers (Fig. 1). Natural tracers are those present in water
molecules or accompany water molecules naturally along
their path, such as deuterium (2 H), carbon-13, or water
temperature. Discharge pulses are often good tracers
for water especially in the response of springs to flash
rainfalls. Artificial tracers are much more diverse in type,
and there are many more tracers not mentioned.
Natural (Environmental) Tracers
Natural or environmental tracers are those that already
exist in the natural environment, the hydrologic cycle.
They could be substances or some kinds of energy.
Environmental Isotopes. Environmental isotopes used as
groundwater tracers are isotopes of hydrogen, oxygen, and
carbon. They could be part of a water molecule (2 H or
deuterium, 18 O, and 3 H or tritium), or they might enter the
hydrologic cycle and move together with water molecules
(13 C, 14 C, 14 N, 15 N, 32 S, 34 S, 36 S . . .).

• Environmental isotopes
• Major ions
Natural
• Temperature
(environmental)
• Pulses
• ...
Groundwater tracers
• Dyes
• Fluorescent
• Nonfluorescent
Artificial
• Drifts
(introduced)
• Salts
• Detergents
• ...
Figure 1. Groundwater tracer types.
Stable Isotopes
Hydrogen and Oxygen Isotopes. Analysis of the natural
concentration of these isotopes in water has helped
in understanding groundwater flow. Because of the
large differences in mass between the two hydrogen
isotopes, they tend to fractionate whenever evaporation
or condensation takes place. Other factors being equal,
waters of higher deuterium content are found near
coastlines, at low elevations, in warm rains, and where
partial evaporation occurs, as in saline lakes. The
variations in oxygen isotope almost follow those of
deuterium. The most common use of studies of 2 H and 18 O
has been to trace the regional movement of groundwater
and to locate recharge areas.
Nitrogen. The two abundant isotopes of nitrogen (14 N
and 15 N) can vary significantly in nature. Ammonia
escaping as vapor from decomposing animal wastes, for
example, tends to remove the lighter nitrogen (14 N) and
leaves behind a residue rich in nitrogen. In contrast, many
fertilizers that have an ammonia base are isotopically
light. Natural soil nitrate is somewhat between these two
extremes. As a consequence, nitrogen isotopes have been
useful in helping to determine the origin of unusually high
amounts of nitrate in groundwater.
Most nitrogen in groundwater is in the form of nitrate
anion (NO3 − ) or dissolved nitrogen gas (N2 ) from the
atmosphere. Locally, in zones devoid of dissolved oxygen,
the chemically reduced form (NH4 + ) may predominate. In
general, nitrate moves as a conservative tracer and is an
important indicator of pollution. If nitrate concentrations
exceed about 10 mg/L, the health of infant mammals
including humans may be adversely affected. Also the
presence of more than about 5 mg/L of nitrate commonly
is an indirect indication of other forms of contamination,
including those from fertilizers and sewage.
Sulfur. Most dissolved sulfur within shallow groundwa-
ter is bound within the sulfate ion (SO4 2− ). The stable
GROUNDWATER TRACING 503
sulfur isotopes (32 S, 34 S and 36 S) found in the sulfate ion
vary quite widely and, under certain circumstances, can
be useful indicators of the origin of the sulfate, which is
particularly true if, for example, one wishes to distinguish
sulfate originating from natural dissolution of gypsum (Ca
SO4 : 2H2 O) from sulfate originating from an industrial
spill of sulfuric acid.
Active Isotopes
Tritium. Tritium (3 H) is an unstable isotope of hydrogen
whose half-life is 12.32 years. It enters the hydrologic
cycle in the same manner as stable hydrogen in the H2 O
molecule. This isotope is naturally produced at low levels
in the atmosphere. During and after the nuclear tests
in the Northern Hemisphere in the 1960s, the tritium
concentration (referred to as TU: Tritium Unit) in the
atmosphere increased from a level of 5–15 TU to up to
4000 TU. 3 H can be used qualitatively to date groundwater
groundwater that has less than 2 to 4 TU is dated prior
to 1953; if the amount is significantly greater than 10 to
20 TU, it has been in contact with the atmosphere since
1953. Tritium analysis in groundwater studies is mainly
for absolute groundwater dating for ages up to 40 years.
Carbon-14. 14 C is the radioactive isotope of carbon.
It has a half-life of 5730 years and is mainly used to
date groundwaters up to 30,000 years of age. Most 14 C
in potable groundwater is contained in the HCO3 − ion
in water. As a first approximation, the initial number of
14
C nuclei per total carbon nuclei, or X0 in Eq. 1, in a
water sample is considered to have been constant prior to
1950 due to the almost constant natural production of 14 C
in the atmosphere by cosmic radiation interacting with
the atmosphere.
Xt = X0 e−λt (1)
where λ = decay constant
t = time
X0 = number of nuclei at zero time
Xt = number of nuclei at time t
If the only source of 14 C in water is originally from
the active biosphere, then the 14 C that is measured in
carbon from a water sample can be considered Xt in the
equation because λ is known from experimental work, the
age of the sample (t) can be determined directly. Due to
some complexities, the use of 14 C in age determination
is not so simple. The existence of carbon from carbonate
formations, or formation of CH4 gas, are factors that affect
the results of age determination. But despite that, 14 C
dating is highly useful in determining the approximate
residence time of old groundwater.
Major ions. The chemical composition of groundwater
is in most cases a good indicator of the environment
through which the water has passed. In almost all
groundwater tracing and isotopic hydrologic programs,
water is analyzed for calcium, sodium, magnesium,
potassium (major cations), chloride, sulfur, carbonate,
bicarbonate, and nitrate (major anions).
504 GROUNDWATER TRACING
Figure 2. Isochlor map for Tehran
plain, 1977, Tehran, Iran. The general
groundwater flow direction is southward.
As can be seen, groundwater gets richer
in ions (specifically in chlor) as it flows
downstream.
Typically, if we follow a recharge event and track the
water down-gradient in the flow path, it can be seen
that the ionic concentration in the water increases as it
migrates down-gradient (Fig. 2). Generally, groundwater
is the bicarbonate type in recharge areas the sulfate type
midway, and the chloride type in terminal zones (e.g.,
basin outlet). The chemical composition of groundwater
would finally be demonstrated on specific presentation
charts such as Piper, Durov, Schuler, . . . and different
groundwater types could be easily recognized.
Temperature. Water temperature is almost a good
indicator or tracer where there is an abrupt difference
in the physical properties of the aquifer (e.g., a change
in hydraulic conductivity, fracture properties . . .) or
where there is a significant temperature gradient (Fig. 3)
Thermometry is a study tool for investigating changes in
rock properties in a borehole column. It comprises a heat
sensor that goes down through the borehole and records
the water temperature along the borehole column. The
record is called a temperature log.
Pulses. Pulses are sudden changes in the amount of a
specific parameter that can be easily seen as a peak in a
graph of the parameter versus time. It could be the graph
of a spring discharge, electrical conductivity, turbidity,
etc. A pulse is normally an indicator of a hydrologic event
upstream. For example, a sudden rise in the amount of
turbidity in a spring could be an indication of a storm in a
spring recharge area. It could also be a sign of a hydraulic
connection between the storm area and the spring.
Artificial Tracers
Artificial tracers are those that are introduced intention-
ally into groundwater. So an ‘‘injection point’’ is necessary
in artificial tracing in addition to a sampling network.1
While in environmental tracing, the study is performed
only through a sampling network. Normally, artificial
tracing and natural tracing are complementary, and some-
times both methods are applied to the same area to achieve
more reliable results. Among all artificial tracers, the most
applicable are fluorescent dye tracers which are discussed
at the end for more emphasis.
Drifts. In areas such as karstified limestone terrains
or highly faulted volcanic rocks where groundwater flow
follows a turbulent pattern, drifts usually are good choices
to be selected as tracers. Signal-emitting floats, yeasts,
1
Sometimes the injection point and sampling point are the same.
This is the case in the so-called single-well tracing method.

bacteria, viruses, and spores have reportedly been used as
tracers in many cases.
Salts. Most tracers that have already been used
extensively are common salts. This category of tracers
includes those compounds that ionize in water, resulting
in separation into charged elements possessing a positive
charge (cations) or a negative charge (anions). The most
common ionic tracers are chloride (Cl− ), bromide (Br− ),
lithium (Li+ ), and potassium (K+ ). These tracers have
been successfully used to determine flow paths, residence
time, and aquifer properties. The most important factors
that should be considered before applying such tracers
are aquifer material texture and natural background
concentrations in the aquifer.
The advantages of salt tracers are their availability,
low cost, and stability, but they must be injected in large
amounts, hundreds of kilograms or tons, and this causes
problems in handling and transportation. As they are
injected in huge amounts, they make water denser around
the injection point, and it may affect the hydrodynamic
behavior of the water.
Fluorescent Dyes. For many reasons, fluorescent dyes
have become very important as tracers in hydrology and
hydrogeology. They can be used in hydrology to measure
transit times in streams and rivers, to measure discharge
in open channels by the tracer dilution method, and
to investigate dispersion processes in rivers, lakes, and
seas. Of course their main application is in hydrogeologic
studies for estimating hydrodynamic parameters as well
as studying sources of pollution.
Application of fluorescent tracers has some great
advantages and a few disadvantages:
• They are visible without the use of any instrument at
higher concentrations (say more than 10 ppb). This
property can be useful for qualitative experiments.
• The limit of detection for the dyes is in the sub-ppb
range and therefore much better (up to a factor of
Flow lines
Isotherms
26.1 25 24
26.8 24.8
25.5
26
Figure 3. Sketch showing a section of a temperature gradient
due to freshwater recharged from the river (up) and the plan of
isotherms at both sides of the river (down).
GROUNDWATER TRACING 505
1000 to 10000) than salt tracers. So a small amount
of tracer is enough.
• Instrument detection is quick and easy without time-
consuming sample preparation.
• The chemical properties of the dyes are in some cases
better because salts (particularly the cationic part)
are influenced by ion exchange in clayey sediments.
• Fluorescent dyes do not influence the density of water
near the injection point.
• Depending on the type of analytical instrumentation,
the costs to determine fluorescent dyes are cheaper
(using an ordinary photometer) or comparable to
a flame photometer which can be used to analyze
salt tracers. The detection of salts by electrical
conductivity measurement is obviously cheaper
but cannot distinguish between different types of
salt tracers.
• A disadvantage is the relatively high cost of
fluorescent dyes, but, on the other hand, up to
1000 times less dye than salt is necessary for a
tracing experiment.
These are the most commonly used fluorescent dyes:
Uranine. For hydrologic or hydrogeologic purposes,
uranine (C20 H10 O5 Na2 , Color index name: Acid Yellow
73) is the most important tracer (Fig. 4). It is also the
most sensitive of all. The dye is supplied as a hygroscopic
orange-red powder: it is very soluble in water and has
a yellowish-red color and an intense yellowish green
fluorescence. In solutions at concentrations higher than
1 g/m3 (equal to 1 ppm), the color is orange-red. The
fluorescence is visible without any instrument above
concentrations of 50 mg/m3 (50 ppb) and approximately
above 10 mg/m3 (10 ppb) using a UV lamp, depending
on the quality of the dye. It has no specific odor. At
high concentrations, the fluorescent light is absorbed by
the sample itself. The intensity of the fluorescence can
be measured exactly only if the concentration is much
lower (less than 100 ppb); otherwise, the solution has to
be diluted.
The solubility of Uranine is 500 mg/L, and its detection
limit is 0.001 mg/m3 . Its fluorescence intensity in solution
is very much dependent on the pH. At pH lower than 6,
the intensity tends to zero. Uranine is an anionic tracer,
and it is adsorbed by clay minerals only to a minor extent.
It has also good temperature stability which is important
for analytics (a temperature increase in the instrument)
and for injection, for example, into thermal systems. From
NaO O O
COONa
Figure 4. Molecular structure of uranine.
506 GROUNDWATER TRACING
the view point of toxicity, it is not dangerous to humans or
animals, in particular, at concentrations normally used in
tracing experiments.
Eosin. Eosin (C20 H6 Br4 Na2 O5 , Acid Red 87) (Fig. 5)
is a powder made of red crystals of a bluish tinge. A
concentrated aqueous solution is deep brownish-red, dilute
solutions are yellowish-red with a greenish fluorescence.
The fluorescence is visible in more concentrated solutions
(above 10 ppb).
It is very soluble in water, less in alcohol, and
insoluble in ether. At high concentrations, fluorescent
light is absorbed by the sample itself. For an exact
analysis, the concentration should not be higher than
100 ppb, otherwise, dilution is recommended. Its solubility
is 300 g/L and the detection limit is 0.01 mg/m3 . The pH
effect for eosin is not as significant as that for uranine.
Eosin is an anionic tracer and is generally little adsorbed
by clay minerals. It can be used in karstic systems as well
as in porous groundwater aquifers. It is not dangerous to
humans or animals.
Among other fluorescent dyes, sulforhodamine G,
Sulforhodamine B, rhodamine WT, rhodamine B, and
pyranine are used most groundwater tracing.
Tracing Experiment
While planning a tracing test, there are some factors
that must be considered. First, the objectives of the
study should be defined well because they affect other
study components such as selecting the tracer. Then
all available resources (labor, money, time,) should be
considered to facilitate formatting the program. Third, all
relevant information and data (maps, reports, well logs)
should be available so that a good idea of the geology and
hydrogeology of the site is gained.
The general direction of groundwater flow and also its
velocity should be estimated before starting the test, which
will allow configuring the injection and sampling network
effectively.
Different Phases of the Test
Planning and General Assessment. Before starting a
tracing test, the test site should be visited and examined
carefully, which is to schedule the program activities
and other components of the work such as selecting the
injection and sampling points, selecting the tracer (which
is among the most important and critical components of the
work), designing the sampling program, transportation
(equipment, tracers, personnel, etc.
Br Br
NaO O O
Br Br
COONa
Figure 5. Molecular structure of eosin.
Tracer Injection. By injecting the tracer into the aquifer,
it will enter the hydrologic cycle and (it is assumed that)
will travel in the same manner as water. The injection
method varies with the type of tracer, study purpose, site
conditions, etc.
Sampling. Sampling is done for the purpose of analyzing
water to find if there is tracer in it. Actually the sampling
work starts before injecting the tracer because there might
be some background concentration of the tracer in the
water; this is called background sampling.
Because of the very high dissemination potential
of fluorescent tracers, it is very important to avoid
any contact between injecting personnel/equipment and
sampling. The sampling time interval for each sampling
point should be scheduled according to the hydrogeologic
factors and distance to the injection point.
Analysis. All samples should be carefully packed and
sent to the lab to be analyzed as soon as possible. The
analytical method and the instrumentation for measuring
the fluorescence of the samples are discussed following. As
mentioned before, concentrations higher than 10 ppb of
most fluorescent dyes may be visible and can be detected
without any instrument.
Detection with a Two Monochromator Spectrofluorom-
eter. To analyze for tracers, the best method is to use
a spectrofluorometer with monochromators for excitation
and emission (Fig 6). The light of the excitation source is
transmitted through a filter of maximum transmittance
near or at the absorption maximum of the dye. The dye
molecules in the solution will be excited, and the fluo-
rescence can then be dispersed by a secondary filter or a
monochromator (prism or grating). Table 1 is a summary
of spectral data of some fluorescent dyes, and Table 2 is a
summary of their approximate detection limits in water.
The final result of an analysis is a time–concentration
(or breakthrough) curve, which presents the temporal
evolution of the sampling point tracer concentration
(Fig. 7).
Interpretation of Results. With regard to the break-
through curve plotted for each sampling point, one should
Excitation Sample
Source
monochromator cell
Slits
Emission
monochromator
Recorder Amplifier Detector
Figure 6. Scheme of a spectrofluorometer.
 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY 507

Table 1. Summary of Spectral Data of Some Fluorescent Jahani, H.R. (1998). Groundwater Tracing and Isotope Hydrolog-
Dyes ical Studies in Rudbar-e Lorestan Area, Iran. Water Research
Center, Tehran, Iran, p. 83.
Wavelength
Fluorescent Fluorescent Difference Standards Office. (2003). Standard Guidelines on Groundwater
Dye Excitation Emission (nm) Tracing in Karst and Hardrock Formations (Draft Edition-
Farsi). Iranian Ministry of Energy, Tehran, Iran, p. 60.
Uranine 491.0 512.5 21.5
Eosin 515.5 536.0 20.5
RhodamineB 553.5 576.0 22.5 HYDRAULIC CONDUCTIVITY/
Pyranine 463.5 512.5 49.0
TRANSMISSIBILITY

STEVEN A. DIELMAN
Table 2. Summary of Approximate Detection Limits for ENVIRON International
some Fluorescent Dyes Corporation
Dye Limit of Detection (mg/m3 ) Arlington, Virginia

Uranine 0.001–0.025
Eosin 0.050–0.5 Hydraulic conductivity is a fundamental parameter that
Rhodamine B 0.010–0.5
governs the flow of liquids such as groundwater through
Pyranine 0.008–0.5
aquifers and other porous media. Specifically, hydraulic
conductivity is a quantitative measure of the capacity of
a geologic formation or other porous media to transmit
14 a specific fluid. It is determined by the characteristics of
Concentration (ppb)

12 both the porous medium and the fluid of interest.

10 Transmissibility (or transmissivity) is a property
8 closely related to hydraulic conductivity that describes
6 the capacity of a specific water-bearing unit of a
4 given thickness, such as an aquifer, to transmit water.
2 Transmissibility is most simply defined as the effective
0 hydraulic conductivity of an aquifer or other water-bearing
0 10 20 30 40 50 60 70 80 90 unit multiplied by the thickness of that unit.
Days after injection
Figure 7. The breakthrough curve for a spring in the Tang-e ORIGINS IN DARCY’S LAW
Shemiran area, western Iran.
The concept of hydraulic conductivity dates back to Henry
Darcy’s famous 1856 paper regarding the water supply for
first of all focus on the first arrival time of the tracer. the city of Dijon, France. In that paper, Darcy described the
It reveals first, that there is an underground hydraulic results of laboratory experiments that he had conducted
connection between the injection point and the sampling to investigate the flow of water through sand. From the
point and, second, the maximum groundwater flow veloc- results of those experiments came the relationship now
ity in that zone. The peak concentration also leads to the referred to as Darcy’s law, it is valid for laminar flow
average transit time. The shape of the curve is a guide conditions that occur in most groundwater systems and
to the approximate volume of water in which the tracer can be expressed as
has been diluted. The area under the curve is equivalent
to the total mass of tracer recovered from the sampling ∂h
point (e.g., a spring). The summation of this figure for all Q = −K A (1)
∂l
sampling points should be close to the injected amount.
Otherwise, there is another outlet for the aquifer which where Q = volumetric flow rate (L3 /t)
was unknown during planning. K = hydraulic conductivity (L/t)
∂h
= hydraulic gradient (L/L)
∂l
READING LIST A = Cross-sectional area perpendicular to the flow
direction (L2 )
Benischke, R. (2002). Fluorescent Tracers in Hydrology. Institute
of Hydrogeology and Geothermics, Graz, Austria, p. 60. As evident from Eq. 1, the volumetric flow of a fluid
Davis, S.N. (1985). Groundwater Tracers. National Groundwater through a porous medium is directly proportional to the
Association, Westerville, OH, p. 200. coefficient K; Darcy recognized that it is dependent on the
Ighanian, R. (1977). Hydrochemical Studies of Tehran Plain. type of soil through which water flows. For this reason, the
Ministry of Energy, Tehran, Iran, p. 32. coefficient K has historically been referred to as Darcy’s
Jahani, H.R. (2000). Groundwater Tracing in Tang-e Shemiran, coefficient of proportionality. Today, the coefficient K is
Western Iran. Water Research Center, Tehran, Iran, p. 85. more commonly called hydraulic conductivity.
508 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY
RELATIONSHIP TO PERMEABILITY
It is intuitive that the flow of a liquid through soil also
depends on the characteristics of the liquid. For example,
if all other factors are equal, the volumetric flow rate
(Q in Darcy’s equation above), and hence the hydraulic
conductivity, will be greater for water than for a more
viscous liquid such as a heavy oil.
The relationship of hydraulic conductivity (K) to
permeability (k)1 is useful in distinguishing the effects
of the porous medium from the effects of the fluid:
kρg
K=
µ
where ρ = density of fluid;
µ = dynamic viscosity of fluid; and
g = gravitational acceleration constant.
Permeability k is a function only of the medium and is
commonly used in petroleum studies where various fluids
of differing characteristics such as oil, water, and natural
gas may be present together in multiphase flow systems.
From Eq. 2, it is clear that hydraulic conductivity is
a function of the characteristics of the fluid as well as
the permeability of the medium (k).2 Specifically, a soil
with a given permeability will exhibit relatively greater
hydraulic conductivity to fluids of lower viscosities and/or
greater densities.
Rearranging Eq. 2 also allows determining the hydrau-
lic conductivity to one fluid if the hydraulic conductivity to
another is known:
µ2 ρ1
K1 = K2
µ1 ρ2
where K1 , µ1 , and ρ1 are properties of fluid 1 and K2 , µ2 ,
and ρ2 are properties of fluid 2.
UNITS OF MEASUREMENT
The units of hydraulic conductivity are (L/t), and the units
of permeability are (L2 ). In the context of groundwater
flow, the units of permeability are sometimes expressed
as ‘‘darcys’’; one darcy approximately equals 1 × 108 cm2 .
Conversion factors for commonly used units of hydraulic
conductivity and permeability are provided in Table 1.
For most applications, the viscosity and density of water
vary over a narrow range, and permeability values can be
converted to hydraulic conductivity (of water), assuming
standard temperature and pressure, without introducing
significant error.
1
In older publications, K was sometimes referred to as the
coefficient of permeability and, therefore, k become known as
‘‘specific’’ permeability or ‘‘intrinsic’’ permeability to differentiate
it from K. More recently, it has become conventional to refer to K
as hydraulic conductivity, allowing k to be referred to simply as
permeability.
2
It is common practice in groundwater hydrology publications to
use the term hydraulic conductivity as it relates to water, unless
a different fluid is specified.
Because transmissibility is defined as hydraulic conduc-
tivity multiplied by aquifer thickness, its units are (L2 /t),
often expressed as volume per unit width of the aquifer
per unit time (e.g., gallons per day per foot or gpd/ft).
REPRESENTATIVE HYDRAULIC CONDUCTIVITY VALUES
Local hydraulic conductivity values can vary over an
extremely wide range. Ranges in representative values
of hydraulic conductivity for various geologic media are
summarized below.
Figure 1 illustrates that hydraulic conductivity values
(2) may range over approximately 12 orders of magnitude,
depending on the type of soil or rock. Even for specific
types of geologic deposits, variations over several orders of
magnitude are not uncommon.
In general, geologic deposits that exhibit the lowest
hydraulic conductivity values are fine-grained unconsoli-
dated materials (e.g., clays) and unfractured sedimentary
and crystalline rocks. The presence of fractures in either
consolidated or unconsolidated materials of otherwise low
permeability can significantly increase the hydraulic con-
ductivity of these materials.
Geologic deposits that exhibit the highest hydraulic
conductivity values are typically coarse-grained unconsol-
idated materials (e.g., coarse sands and gravels) and rocks
that are highly fractured or otherwise exhibit significant
secondary permeability (e.g., solution-enhanced void space
in carbonate formations or basalt shrinkage cracks).
GEOLOGIC FACTORS INFLUENCING HYDRAULIC
CONDUCTIVITY
(3)
For unconsolidated materials, the factors that have the
greatest influence on hydraulic conductivity are generally
particle size and particle sorting. It can be theoretically
shown that hydraulic conductivity is proportional to the
square of the particle diameter for a uniform, porous
medium, that is, doubling the particle diameter results in
quadrupling hydraulic conductivity. In practice, however,
geologic formations are not uniform but contain particles
of varying sizes, and the definition of a representative
particle diameter is not clear-cut. In general, poorly sorted
deposits (i.e., those with greater variation in particle
sizes) will exhibit lesser relative hydraulic conductivity
compared to that of well-sorted deposits, due to infilling of
pore space by finer grained particles.
In consolidated crystalline and sedimentary rocks, the
hydraulic conductivity of the rock matrix tends to be low to
extremely low. Transmission of water through these rocks
is generally via interconnected voids such as fractures,
solution channels, shrinkage cracks or metamorphic (e.g.,
schist) foliations. Transmission of water through such
voids is said to occur via ‘‘secondary permeability.’’
HETEROGENEITY AND ANISOTROPY
Geologic depositional environments typically exhibit
significant variability across very short distances. As a
result, hydraulic conductivity values generally exhibit
 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY 509

Table 1. Hydraulic Conductivity and Permeability Conversion Factorsa

Hydraulic Conductivity Permeability

Multiply
cm/s m/s ft/s ft/day US gal/day/ft2 cm2 Darcy
To obtain by by by by by by by

cm/s 1 100 30.48 3.53E−04 4.72E−05 9.71E+04 9.66E−04

Hydraulic Conductivity

m/s 0.01 1 0.305 3.52E−06 4.72E−07 9.71E−02 9.66E−06

ft/s 3.28E−02 3.28 1 1.16E−05 1.55E−06 3.19E+03 3.17E−05

ft/day 2.84E+03 2.84E+05 8.64E+04 1 0.134 2.75E+08 2.74

US gal/day/ft2 2.12E+04 2.12E+06 6.45E+05 7.48 1 2.06E+09 1.48E+01

Permeability

cm2 1.03E−05 1.03E−03 3.14E−04 3.63E−09 2.71E−08 1 9.87E−09

Darcy 1.04E+03 1.04E+05 3.15E+04 3.65E−01 2.72 1.01E+08 1

a
Conversions between units of hydraulic conductivity (K) and permeability (k) assume the properties of water at approximately 15 ◦ C.

spatial variation across several orders of magnitude within
an area of hydrogeologic interest.
The terms heterogeneity and anisotropy are often
employed to describe variability in hydraulic conduc-
tivity. Heterogeneity refers to variability in space, and
anisotropy refers to variability in direction from a given
point. Figure 2 illustrates the relationship between het-
erogeneity and anisotropy.
The distinction between anisotropy and heterogeneity
often depends on scale and context. For example,
on one scale, layered heterogeneity (see Fig. 4) may
require consideration as a series of distinct geologic
formations that have different representative hydraulic
conductivity values. On another scale, it may be more
convenient and appropriate to describe this system as one
anisotropic layer.
Heterogeneity
A porous medium of uniform characteristics is said
to be homogenous. In reality, no geologic system is
truly homogenous with respect to hydraulic conductivity.
Each system exhibits varying types and degrees of
heterogeneity. An illustration of actual heterogeneity is
shown in Fig. 3. This two-dimensional cross-section was
created from more than 700 measurements within an
area approximately 2 meters deep by 20 meters long
within the intensely studied Borden aquifer, which is
generally recognized as ‘‘mildly’’ heterogenous compared
to most groundwater systems. Even in this small area
of a relatively uniform aquifer, measured hydraulic
conductivity values range across approximately three
orders of magnitude.
In some cases, heterogeneity in hydraulic conductivity
values may exhibit one or more measurable trends due
to geologic conditions. Freeze and Cherry (2) identified
three types of common heterogeneity patterns: trending,
discontinuous, and layered.
Trending heterogeneity refers to spatial trends result-
ing from, for example, gradational coarsening or fining
depositional sequences such as those that typically result
from encroaching or receding seas. Trending heterogene-
ity may also result from postdepositional processes such
as vertical trends in rock fracture frequency.
Discontinuous heterogeneity refers to an abrupt change
in local hydraulic conductivity values, such as those
across a geologic discontinuity, fault, or other sharp
stratigraphic contact.
Layered heterogeneity is a commonly encountered
type of heterogeneity, particularly for unconsolidated
or consolidated sedimentary sequences formed in time-
varying depositional environments. An idealized example
of layered heterogeneity is illustrated in Fig. 4. A more
realistic depiction of layered heterogeneity is illustrated in
the Borden aquifer hydraulic conductivity plot depicted in
Fig. 3.
In reality, distinguishing between types of hetero-
geneities is often difficult. For example, layered het-
erogeneity may be considered a series of discontinuous
heterogeneities. In addition, multiple patterns in hetero-
geneity may be present at any given location.
Anisotropy
If hydraulic conductivity is independent of the direction
of measurement, the medium is described as isotropic;
if, however, the hydraulic conductivity value depends on
the direction of measurement, it is considered anisotropic.
Mathematically, hydraulic conductivity is represented as
510 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY

Unconsolidated deposits Sedimentary rocks Crystalline rocks

101

100
Gravel

Karst and reef limestone

10−1

Permeable basalt
Coarse sand

Fractured igneous & metamorphic rock

10−2
Medium sand

Weathered
granite
Fine sand

10−3

Weathered
gabbro
10−4
Hydraulic conductivity (cm/s)

Limestone, Dolomite
Silt, loess

Sandstone
10−5

10−6

Basalt
Till

10−7
Siltstone
Unweathered marine clay

Unfractured igneous & metamorphic rocks

Clay

10−8
Anhydrite

Shale

10−9
Salt

10−10

10−11

10−12

Figure 1. Representative values of hydraulic conductivity for various rock types (adapted from: Ref. 1).

a scalar (direction-independent) variable in the form of hydraulic conductivity (Kv ) between 3:1 and 10:1 for
Darcy’s law presented in Eq. 1 In anisotropic settings, clays and shales due to mineral grain orientation. On
hydraulic conductivity must be considered a tensor a larger scale, however, Kh :Kv anisotropic ratios for
variable, leading to more complex forms of Darcy’s law sedimentary deposits may be extremely high due to
and other equations that describe the flow of groundwater stratified deposits of greatly different characteristics.
and other fluids through porous media. High Kh :Kv anisotropic ratios for stratified deposits
On a small scale, anisotropy can result from mineral are not surprising, considering the large differences in
grain orientation. On a larger scale, anisotropy can result typical hydraulic conductivities of coarse-grained and fine-
from sedimentary bedding planes, metamorphic structure grained deposits, which may differ by several orders of
(e.g., shistocity) or directional fracturing of crystalline magnitude.
rocks resulting from postdepositional stresses.
In sedimentary deposits of horizontal or near-horizontal EFFECTIVE HYDRAULIC CONDUCTIVITY: STATISTICAL
orientation, hydraulic conductivity is frequently greater TREATMENT OF HYDRAULIC CONDUCTIVITY VALUES
in the lateral directions compared to the vertical
direction. On a small scale, such anisotropy may be Groundwater hydrologists have used a variety of tech-
relatively modest. Freeze and Cherry (2) report ratios niques to characterize the spatial variability of hydraulic
of horizontal hydraulic conductivity (Kh ) to vertical conductivity values and to assess groundwater flow in
 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY 511

Homogeneous, isotropic Homogeneous, anisotropic

KZ

KZ
KZ
KX KZ KX

Z KX KX

Heterogeneous, isotropic Heterogeneous, anisotropic

KZ
KZ

KZ

KZ
KX KX

KX KX Figure 2. Four possible combina-

tions of heterogeneity and anisotropy
(adapted from Ref. 2).

219.14
4.5 5.5
4.0 4.0 5.0
4.0
Elevation, masl

218.64 4.0
5.0
5.0 5.0
5.0 5.0
4.5
218.14 5.0 4.5
4.0
4.5 4.5
5.0 5.0 5.0
217.64 5.0 5.0 4.5
4.0 5.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Distance along section
5.0 In(K ) (cm/s)
Location of core
K < 10−3 cm/s

Figure 3. Hydraulic conductivity distribution at Canadian Forces Base Borden (from Ref. 3, v. 22, p. 2069–2082).

heterogenous geologic settings. One such approach is to expressed as

identify an ‘‘effective’’ hydraulic conductivity that, applied   dh
to a heterogenous water-bearing geologic unit, will allow Q= Qi = Ki Di W (4)
dx
an accurate representation of its ability to transmit water
as a whole. where
Consider horizontal groundwater flow through an Ki = hydraulic conductivity of layer i,
idealized layered geologic system such as that shown in dh/dx = horizontal hydraulic gradient (assumed
Fig. 4 where the a number of layers equals n. The total uniform in all layers),
flow through the system (Q) is equal to the sum of the Di = thickness of layer i, and
flow in all layers (Qi ). From Darcy’s law, this can be W = unit width.
512 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY
For combined flow through the layered system, an effective
horizontal hydraulic conductivity (KH-EFF ) can be defined
for groundwater flow such that
dh
Q = KH-EFF DW (5)
dx
From Eqs. 4 and 5, the following relationship for KH-EFF
is obtained: 
Ki Di
KH-EFF = (6)
D conductivity of a heterogenous groundwater system than
This shows that, for flow parallel to the idealized layered
system shown in Fig. 4, KH-EFF is equivalent to the
arithmetic mean of Ki weighted by the thickness of each
layer Di . If all layers are of the same thickness (or all K
values are uniformly distributed), KH-EFF is equal to the
arithmetic mean of the individual Ki values.
Consider, however, vertical groundwater flow through
the same system. Using a similar analysis, it can be shown
that the effective horizontal conductivity in the vertical
direction (KV-EFF) is equal to
D Empirical Relationships
KV-EFF =  (7)
Di /Ki
If all layers are of equal thickness (or are uniformly
distributed), KV-EFF for the idealized layered system is
equivalent to the harmonic mean of Ki values, which is
always lower than or equal to the arithmetic mean.
These two simplified conceptual models represent two
extreme possibilities of flow uniformly parallel to or
perpendicular to geologic layers, respectively. Therefore,
the arithmetic and harmonic means of a population of
hydraulic conductivity values are generally considered to
be upper and lower bounding values, respectively, for the
effective hydraulic conductivity of a heterogenous geologic
Silt
Clay
Sand and gravel
Weathered bedrock
Crystalline bedrock
Figure 4. Example of layered heterogeneity.
system with respect to groundwater flow through that
system.
Frequently, hydraulic conductivity values and well
yields within a specific geologic setting are best character-
ized by a log-normal distribution. As such, the geometric
mean is often preferred as a descriptive statistic for the
central tendency (e.g., median and mode) of a population
of log-normally distributed hydraulic conductivity values.
Further, several studies have concluded that the geometric
mean provides a better estimate of the effective hydraulic
the harmonic mean or arithmetic means, which are less
than and greater than the geometric mean, respectively.
The preceding provides a simplified discussion of
common statistical treatments for hydraulic conductivity
values. In recent decades, much work has been conducted
to apply more complex statistical models and stochastic
techniques to characterize groundwater flow through
heterogenous systems.
METHODS OF ESTIMATION AND MEASUREMENT
Several efforts have been made over the years to correlate
hydraulic conductivity and/or permeability empirically
with characteristics of porous media through which water
flows. Some of the more commonly used relationships are
provided in Table 2.
Although such empirical relationships correlate
hydraulic conductivity with various factors, nearly all can
be simplified to the form
K = Cdx (8)
where C is a constant or shape factor that is a function
of the fluid density and viscosity and may be correlated
with properties of the porous medium such as porosity
and d is some representative grain size (L). It has been
mathematically shown (e.g., by Ref. 5) that for idealized
systems of uniform grain size, x = 2. For all the empirical
relationships shown in Table 2, x ranges between 1.5 and
2, which indicates a disproportionate increase in hydraulic
conductivity as grain size increases, but suggests an
exponent value of something less than 2 for many real-
world situations.
Laboratory Testing
The hydraulic conductivity of geologic deposits can be
measured in the laboratory by using a permeameter, a
device similar to that originally devised by Darcy. Various
configurations of permeameters have been used; the most
common are the constant-head permeameter and the
falling-head permeameter shown in Fig. 5.
Permeameters consist of a chamber to contain a sample
of soil or rock and allow the application of a hydraulic
gradient across the sample. Flow through the sample is
measured, and hydraulic conductivity is calculated based
on principles founded by Darcy. Careful consideration
must be used in evaluating permeameter test data. For
example, when testing unconsolidated material, was the
 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY 513

test conducted on an intact core or on a repacked sample? If
the former, the test likely represents the vertical hydraulic
conductivity, which may be much lower than the horizontal
hydraulic conductivity. If the latter, the disturbance of the
sample during repacking may result in a value that is
neither representative of the horizontal nor the vertical
hydraulic conductivity.
Field Measurements—Aquifer Tests
Hydraulic conductivity and transmissibility on a larger
scale are estimated by evaluating data from aquifer tests.
There are many different types of aquifer tests and
methods for evaluating the test results, but the most
common types of tests employed in groundwater studies
are pumping tests and slug tests.
Pumping tests involve pumping water from a well
at a known rate and measuring the changes in water
level through time at the pumping well and/or nearby
observation points. Slug tests involve rapid introduction
or removal of a volume of water or other material into or
out of a well and measuring the changes in water level
through time at the well being tested.

Table 2. Examples of Empirical and Semiempirical Relationships for Estimating Hydraulic Conductivity and
Permeabilitya
Source Equation Parameters

Hazen (1911) K = Cd10 2

K = hydraulic conductivity (cm/s)
C = constant; 100 to 150 (cm/s)−1 for loose sand
d10 = effective grain size (cm)
(10% particles are finer, 90% coarser)

Harleman et al. (1963) k = (6.54 × 10−4 ) d10 2 k = permeability, cm2

Krumbein and Monk (1943)
Kozeny-Carmen Bear (1972)
d10 = effective grain size, cm
(10% particles are finer, 90% coarser)
k = 760 d2 e−1.31σ k = permeability (Darcys)
d = geometric mean grain diameter, mm
σ = log standard deviation of the size distribution
K = (ρw g/µ)n3 /(1 − n)2 (dm 2 /180) K = hydraulic conductivity
ρw = fluid density
µ = fluid viscosity
g = gravitational constant
dm = representative grain size
n = porosity

Shepard (1989) K = 40,000 d2 glass spheres K = hydraulic conductivity, ft/day

K = 1,600 d1.75 beach deposits d = mean grain size, mm
K = 450 d1.65 channel deposits
texturally immature
K = 100 d1.5 consolidated sediments
a
Adapted from Refs. 1 and 4.

dt
Constant water level

h0
h
h
Overflow
Falling-head tube

Q Q
dc

Porous plate Sample L Porous plate Sample L

Constant-head permeameter apparatus Falling-head permeameter apparatus

Figure 5. Permeameters (adapted from Ref. 4).

514 GROUNDWATER FLOW AND TRANSPORT PROCESS
In general, pumping tests affect a significantly larger
volume of the aquifer than slug tests, and the resulting
hydraulic conductivity and transmissibility estimates are
considered more representative of conditions across a
broad area. Slug tests can be used to target specific
zones and points within the aquifer but are generally
considered ‘‘point estimates’’ of hydraulic conductivity and
transmissibility.
BIBLIOGRAPHY
1. Domenico, P.A. and Schwartz, F.W. (1997). Physical and
Chemical Hydrogeology. John Wiley & Sons, New York.
2. Freeze, R.A. and Cherry, J.A. (1979). Groundwater. Prentice-
Hall, Englewood Cliffs, NJ.
3. Sudicky, E.A. (1986). A natural gradient experiment on solute
transport in a sand aquifer: spatial variability of hydraulic
conductivity and its role in the dispersion process. Water
Resour. Res. 22: 2069–2082.
4. Fetter, C.W. (1994). Applied Hydrogeology. Macmillan College
Publishing Company, Inc., New York.
5. Hubbert, M.K. (1940). The theory of groundwater motion. The
Journal of Geology 48: 785–944.
GROUNDWATER FLOW AND TRANSPORT
PROCESS
MOHAMMAD N. ALMASRI
An-Najah National University
Nablus, Palestine
JAGATH J. KALUARACHCHI
Utah State University
Logan, Utah
INTRODUCTION
Groundwater is an important source of water for
agriculture and domestic drinking water. Two-thirds of
the world’s freshwater is found in underground aquifers.
Groundwater occurs mainly in the tiny spaces between
soil particles that are called voids or pores. In its natural
state, groundwater is perpetually moving. This movement
is governed by conventional hydraulic principles such as
Darcy’s law, the concept of hydraulic gradient, and the
permeability of the medium (1). With the increasing water
demand due to population growth, management of water
resources is becoming a challenging task. Management
of groundwater resources must consider both sustainable
development and prevention and control of groundwater
contamination. Groundwater contamination is a critical
issue because of the importance of groundwater supply
and quality in sustainable development. In general,
contaminants found in groundwater are associated with
adverse health, social, environmental, and economic
impacts (2). Exposure to contaminated groundwater may
occur accidentally because contaminated water might
be colorless, odorless, and tasteless. It is essential to
understand groundwater flow and transport processes for
the best management of groundwater resources, appraisal
of the sustainable yield of aquifers, reliable evaluation
of the extent of groundwater contamination, assessment
of the corresponding hazards and risks, and design of
effective remediation systems and protective alternatives.
This chapter provides an overview of the main princi-
ples of groundwater flow and transport processes, includ-
ing the basic concepts, definitions and terms, physical
governing equations, and mathematical modeling.
VERTICAL DISTRIBUTION OF SUBSURFACE WATER
The subsurface occurrence of water may be divided into
zones of aeration and saturation (see Fig. 1). The zone of
aeration is partially occupied by water and partially by air.
In the zone of saturation, all pores are filled with water.
The zone of aeration may be further subdivided into the
soil–water zone, the intermediate vadose zone, and the
capillary zone (see Fig. 1).
The saturated zone extends from the upper surface of
saturation down to underlying impermeable rock. In the
absence of an overlying impermeable formation, the water
table forms the upper surface of the zone of saturation,
which is defined as the surface of atmospheric pressure.
Water in the soil–water zone exists at less than saturation
except temporarily when excessive water reaches the
ground surface from rainfall or irrigation. The zone
extends from the ground surface down through the major
root zone. The water content and distribution within this
zone depends on soil texture, vegetation, and atmospheric
conditions. The intermediate vadose zone extends from
the lower edge of the soil–water zone to the upper limit
of the capillary zone. The capillary zone extends from the
Ground surface
Soil-water
zone
Vadose water
aeration
Zone of
Intermediate
vadose
zone
Capillary
zone Water table
or phreatic water
Groundwater
saturation
Zone of
Impermeable rock
Figure 1. Illustration of saturated and unsaturated zones (1).

Recharge
area
Water
Water Piezometric surface table
table Ground Flowing well
surface well
Water table
Unconfined
aquifer
Confining stratum
Impermeable
strata
Confined aquifer
water table up to the limit of capillary rise of water. The
thickness of the capillary zone varies inversely with the
pore size of a soil.
AQUIFER TYPES
An aquifer is a formation that contains sufficient saturated
permeable material to yield significant quantities of water
to wells and springs (3). Therefore, aquifers have the
ability to store and transmit water. Aquifers might be
bounded by confining beds that are relatively impermeable
materials located adjacent to the aquifers. Confined
aquifers occur where groundwater is confined under
pressure greater than atmospheric by overlying relatively
impermeable formations. In a well penetrating such an
aquifer, the water level will rise above the bottom of the
confining bed, as shown by the artesian and flowing wells
of Fig. 2. Unconfined aquifers do not have confining beds
above and the aquifer can be directly recharged by rainfall
(see Fig. 2). An unconfined aquifer is one in which the
water table varies in a rising and falling form.
DARCY’S LAW
In 1856, the French engineer, Henry Darcy, developed
the principal equation in groundwater hydrology based on
laboratory experiments he had carried out to analyze the
flow of water through sand. Darcy’s law can be written as
dh
Q = −KA (1)
dl
Q is the quantity of water per unit of time (L3 /T), A
is the area perpendicular to flow (L2 ), K is a constant
of proportionality called the hydraulic conductivity (L/T),
h is the hydraulic head (L), and dh/dl is the hydraulic
gradient. The hydraulic conductivity expresses the ease
by which water can pass through soil pores and openings
and serves as a measure of the permeability of the
porous medium. The hydraulic conductivity of a soil is
a function of the properties of the porous medium and
GROUNDWATER FLOW AND TRANSPORT PROCESS 515
Artesian
well
Figure 2. Schematic cross section of unconfined
and confined aquifers (1).
the fluid passing through it. In addition to hydraulic
conductivity, the term transmissivity T (L2 T−1 ) is widely
employed in groundwater hydraulics. Transmissivity
equals the hydraulic conductivity multiplied by the
saturated thickness of the aquifer.
Darcy flux or specific discharge is given by the following
relationship:
Q dh
q= = −K (2)
A dl
The specific discharge q has the dimensions of velocity
(L/T). Finally, the cross-sectional area of flow for a porous
medium is actually much smaller than the dimensions of
the aquifer. To find the velocity at which water is actually
moving, the specific discharge is divided by the effective
porosity to account for the actual open space available for
flow. The result is the average velocity at which water
moves between two points as in the following equation:
q Q Kdh
v= = =− (3)
ne ne A ne dl
STORAGE IN CONFINED AND UNCONFINED AQUIFERS
Storage in confined aquifers is associated with the
compressibility of water and the porous matrix. The
change in the potentiometric head affects the arrangement
of mineral grains as well as the density of water in the
voids. If the water pressure increases, the mineral skeleton
will expand; if it drops, the mineral skeleton will contract.
Likewise, water will contract with an increase in pressure
and expand if the pressure drops. The specific storage (Ss )
is the amount of water per unit volume of a saturated
formation that is stored or expelled from storage owing to
compressibility of the mineral skeleton and the pore water
per unit change in head. The value of specific storage is
very small, generally 0.0001 ft−1 or less. In unconfined
aquifers, pumpage causes drawdown in the water table
itself, and the induced hydraulic gradient causes both a
horizontal and vertical flow component. Thus, the drop in
the water table creates a vertical hydraulic gradient which
causes drainage of the porous medium (3–7).
516 GROUNDWATER FLOW AND TRANSPORT PROCESS
The storativity (S) of a confined aquifer is the product of
the specific storage Ss and the aquifer thickness. The value
of the storativity of confined aquifers is of the order 0.005
or less. Storativity under unconfined conditions is referred
to as the specific yield, defined as the ratio of the volume
of water an unconfined aquifer drains by gravity due to
the unit decline in the water table to the total volume
of the aquifer. The coarser the material, the more closely
the specific yield approaches the porosity. The volume of
water drained from an aquifer as the head is lowered may
be found from the formula,
Vw = S A h (4)
where Vw is the volume of water drained or released (L3 ),
S is the storativity (L0 ), A is the surface area overlying
the drained aquifer (L2 ), and h is the average decline in
head (L). Note that h for a confined aquifer is the decline
in the potentiometric head and for unconfined aquifers is
the decline in the water table.
MATHEMATICAL MODEL OF GROUNDWATER FLOW
Groundwater flow is described using Darcy’s law and
the mass conservation equation. The governing equation,
describing transient groundwater flow in a three-
dimensional flow domain, is given as




∂ ∂h ∂ ∂h ∂ ∂h ∂h
Kx + Ky + Kz + W  = Ss
∂x ∂x ∂y ∂y ∂z ∂z ∂t
(5)
where Kx , Ky , Kz are the hydraulic conductivity values
in the x, y, and z coordinates, respectively; Ss is the spe-
cific storage; t is time; h is the hydraulic head; and W  is
the source or sink strength in units of volume/time/area.
To solve Eq. 5, initial and boundary conditions should be
specified for the study area. Solving this equation yields
the head distribution in time and space. Computer codes
can be used to solve Eq. 5 numerically. MODFLOW devel-
oped by the U.S. Geological Survey (8), is one of the most
commonly used groundwater softwares for solving one-
, two-, or three-dimensional groundwater flow problems.
MODFLOW is divided into packages that deal with single
simulation aspects. For example, the Well Package sim-
ulates the effects of injection and abstraction wells and
the River Package simulates the effect of rivers, etc. The
WWW address for MODFLOW model and documenta-
tion is: http://water.usgs.gov/sofware/ground water.htm.
A brief description of MODFLOW can also be found in
Charbeneau (7).
SOURCES OF GROUNDWATER CONTAMINATION
There are three key features that differentiate ground-
water contamination sources: (1) degree of localization,
(2) loading history, and (3) types of contaminants. Sources
of groundwater contamination can be classified into point
and nonpoint sources. Point sources include leaking under-
ground pipelines or tanks, wastewater lagoons, septic
systems, and leaking barrels of chemical wastes. These
sources are identifiable, located at single locations, and
have local impacts on groundwater quality. Conversely,
nonpoint sources are diffuse, originating from many
smaller sources, largely distributed, lack a well-defined
single point of origin, and have a regional influence on
groundwater quality. Examples of nonpoint sources are
agricultural activities such as fertilizer and manure appli-
cations; and polluted atmospheric deposition, including
dry fallout, precipitation, and runoff from roads and park-
ing lots. The loading history defines how the concentration
of a contaminant or its rate of application changed with
time at the source location.
TRANSPORT PROCESSES
The physical processes that control contaminant flux
into and out of a groundwater system are advection,
mechanical dispersion, and molecular diffusion. In the
following subsections, the basic concepts related to these
physical processes are illustrated.
Advection
Transport by advection or convection entails movement
and carrying contaminants with flowing groundwater (6).
In advective transport, contaminants travel at a rate equal
to the average linear velocity of the groundwater (5).
Mathematically, advective flux is expressed by the
following equation (6):
J = vx ne C (6)
where Jx is the advective mass flux (ML2 T−1 ) per unit
area per unit time, vx is the average linear velocity of
groundwater(LT−1 ), ne is the effective porosity through
which flow occurs, and C is the contaminant concentration
(ML−3 ). The average linear velocity can be determined
from Eq. 3. The product of the average linear velocity
and time since contaminants were introduced into the
groundwater system defines the position of the advective
front, as depicted in Fig. 3, where the vertical line
represents the advective front without considering the
effects of dispersion and diffusion (6).
Diffusion
The movement of contaminants from an area of high
concentration to an area of lower concentration is called
molecular diffusion or diffusion. This diffusive flux is
Advection only
1
concentration
Relative
(C /C0)
0.5 Advection + diffusion Advection +
dispersion
0
Distance x
Figure 3. Advective transport and the influence of longitudinal
dispersion and diffusion on the transport of a solute in
one-dimensional flow.

directly proportional to the diffusion coefficient and the
concentration gradient (9). This is known as Fick’s first
law and can be expressed in a one-dimensional system as
in the following:
dC
Fx = −Dd (7)
dx
where Fx is the diffusive mass flux (ML2 T−1 ), Dd is
the diffusion coefficient for open water (L2 T−1 ), C is
the contaminant concentration (ML−3 ); and dC/dx is the
concentration gradient (ML−4 ). The negative sign signifies
that the diffusive flux is in the direction of decreasing
contaminant concentration.
Mechanical Dispersion
Dispersion describes the gradual spreading of contam-
inants beyond the region they are expected to occupy
when considering advection alone. The phenomenon of dis-
persion is attributed mainly to the discrepancy between
contaminant velocities and the average linear velocity
caused by subsurface heterogeneity (5,10,11). According
to Fetter (3), this discrepancy is attributed to three key
factors: (1) the velocity of groundwater flow is higher in
the center of pores than at the edges, (2) the contami-
nants can pass through different flow paths to reach the
same location, and (3) groundwater moves faster in large
pores. Therefore, contaminants will dilute due to mixing
with noncontaminated water. Mixing, called mechanical
dispersion, may occur along the flow path, is called longitu-
dinal dispersion, and is attributed to the aforementioned
key factors. The groundwater flow path may diverge at
the pore scale, so the contaminant will spread in direc-
tions normal to the flow path. This mixing is called lateral
or transverse dispersion (3). The difference between the
straight dashed line and the solid line in Fig. 3 depicts
the influence of longitudinal dispersion on the transport of
a contaminant. Molecular diffusion and mechanical dis-
persion contribute jointly to the dispersive process of
contaminant transport. They are combined to define a
coefficient called hydrodynamic dispersion as given in the
following equations (6):
DL = αL vi + D∗ (8)
DT = αT vi + D∗ (9)
where DL and DT are the longitudinal and transverse
hydrodynamic dispersion coefficients, respectively; and αL
and αT are the longitudinal and transverse dispersivities,
respectively.
Figure 4 illustrates the process of hydrodynamic
dispersion in two-dimensional groundwater flow at a
constant uniform velocity, V. A tracer was instantaneously
injected into the aquifer at t = 0. The groundwater flow
carries the tracer with it. Due to the effect of dispersion, the
solute spreads out such that the maximum concentration
decreases with time, as shown for times t1 and t2 when t2 >
t1 . The spread of the tracer results in an elliptically shaped
concentration distribution that is normally distributed in
both the longitudinal and transverse directions.
As mentioned earlier, longitudinal dispersivity or more
specifically, dispersivity, is a scale-dependent property
GROUNDWATER FLOW AND TRANSPORT PROCESS 517
C3 < C2
L2 = Vt2
C3 < C 2 C2 < C1
L1 = Vt1
C2 < C 1 C1
C1
V
Tracer
injected Contours of C = constant
at t = 0 at t = t1 Contours of C = constant
at t = t2
C max C max
Figure 4. Longitudinal and transversal spreading of a tracer
from an instantaneous point injection at t = 0.
that increases with the characteristic transport length
distance of measurement. This phenomenon is attributed
to the macroscale effects resulting from the heterogeneity
of the subsurface on the field scale. Many researchers have
developed relationships between longitudinal dispersivity
and flow length (6,12–14). The ratio of longitudinal to
transverse dispersivity (αL /αT ) controls the shape of the
contaminant plume. As the ratio decreases, the wider the
plume will be (6). Bear (10) provided a range from 5 to 24
for αL /αT .
MATHEMATICAL MODEL OF CONTAMINANT TRANSPORT
The advection–dispersion–reaction partial differential
equation that governs three-dimensional transport of a
single chemical constituent in groundwater, considering
advection, dispersion, fluid sinks/sources, equilibrium-
controlled sorption, and first-order irreversible rate
reactions, is given in the following (11):

 
∂C ∂ ∂ ∂C qs ρb 
R = Dij (vi C) + Cs − λ C + S
−
∂t ∂xi ∂xi ∂xj n n
(10)
where the retardation factor is defined as
ρb ∂S
R=1+ (11)
n ∂C
and C is the dissolved concentration (ML−3 ); S is the sorbed
concentration (MM−1 ) which is a function of the dissolved
concentration, C, as defined by a sorption isotherm; t is
time (T); Dij is the dispersion coefficient tensor (L2 T−1 );
vi is the average linear groundwater velocity (LT−1 ); qs
is the volumetric flow rate per unit volume of aquifer
and represents fluid sources and sinks (T−1 ); Cs is the
concentration of the fluid source or sink flux (ML−3 ); λ is
the reaction rate constant (T−1 ); R is the retardation factor
(L0 ); ρb is the bulk density of the porous medium (ML−3 );
and n is the porosity (L0 ). A complete derivation of this
equation is provided in Freeze and Cherry (4) and Zheng
and Bennett (11).
Solving time-dependent partial differential equations
governing contaminant transport requires identification of
the initial and boundary conditions. The initial condition
describes the distribution of the contaminant in the
groundwater within the domain at the beginning of
simulation. The boundary conditions can be classified
into three categories: (1) specified concentration boundary,
518 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS
called the Dirichlet condition; (2) a specified concentration
gradient boundary, called the Neuman condition; and (3) a
variable flux boundary, called the Cauchy condition (5).
Solving the mathematical model yields the contami-
nant concentration as a function of space and time. The
approach used in formulating and solving the mathe-
matical model can be classified generally into numerical
and analytical approaches. Analytical solutions of the
advection–dispersion equations in one-, two-, and three-
dimensional space can be found in Baetslé (14), Bear (10),
Domenico (15), Domenico and Schwartz (5), and Fetter (6)
among others.
Numerical methods provide an approximate solution
of the mathematical model by solving a set of alge-
braic equations that replace the original partial differ-
ential equations. Numerical solutions are widely used
and preferred because they have the capability to handle
more general conditions. Numerical methods for solving
the advection–dispersion–reaction equation divide the
model domain into small cells, approximate the governing
equation over each cell at time t, and then predict a new
value at time t + t. This process continues forward in
time for small t values. The most common numerical
methods for solving the advection–dispersion–reaction
equation include the finite-difference method, the finite-
element method, and the random walk method. A detailed
illustration of these methods can be found in Zheng
and Bennett (11) and Sun (16). Bedient et al. (17) cited
many references that provide detailed descriptions of
these numerical methods. One of the common com-
puter codes used in simulating the fate and transport
of contaminants in groundwater is MT3D (18). MT3D
is a three-dimensional contaminant transport model for
the simulation of advection, dispersion, and chemical
reactions of dissolved constituents in groundwater sys-
tems. MT3D is based on a modular structure to per-
mit simulation of transport components independently
or jointly with finite-difference groundwater flow mod-
els, such as MODFLOW. MT3D can be downloaded from
http://hydro.geo.ua.edu. Further information on MT3D can
be found in Zheng (18), Zheng and Bennett (11), and Zheng
and Wang (19).
BIBLIOGRAPHY
1. Todd, D.K. (1980). Groundwater Hydrology. John Wiley &
Sons, New York.
2. Office of Technology Assessment. (1984). Protecting the
nation’s groundwater from contamination, Office of Tech-
nology Assessment, OTA-0-233, Washington, DC.
3. Fetter, C.W. (1994). Applied Hydrogeology. Prentice-Hall,
New York.
4. Freeze, R.A. and Cherry, J.A. (1979). Groundwater. Prentice-
Hall, New York.
5. Domenico, P.A. and Schwartz, F.W. (1990). Physical and
Chemical Hydrology. John Wiley & Sons, New York.
6. Fetter, C.W. (1999). Contaminant Hydrogeology. Prentice-
Hall, New York.
7. Charbeneau, R.J. (2000). Groundwater Hydraulics and
Pollutant Transport. Prentice-Hall, Upper Saddle River,
NJ.
8. Harbaugh, A.W. and McDonald, M.G. (1996), User’s doc-
umentation for MODFLOW-96, an update to the U.S.
Geological Survey modular finite-difference ground-water
flow model: U.S. Geological Survey Open-File Report
96–485.
9. Ingebritsen, S.E. and Sanford, W.E. (1998). Groundwater in
Geologic Processes. Cambridge University Press, Cambridge,
UK.
10. Bear, J. (1979). Hydraulics of Groundwater. McGraw-Hill,
New York.
11. Zheng, C. and Bennett, G.D. (1995). Applied Contaminant
Transport Modeling: Theory and Practice. International
Thomson Publishing, Stamford, CT.
12. Neuman, S.P. (1990). Universal scaling of hydraulic conduc-
tivities and dispersivities in geologic media. Water Resour.
Res. 26(8): 1749–1758.
13. Xu, M. and Eckstein, Y. (1995). Use of weighted least-squares
method in evaluation of the relationship between dispersivity
and field scale. Ground Water 33(6): 905–908.
14. Baestlé, L.H. (1999). Migration of radionuclides in porous
media. Progress in Nuclear Energy Series XII, Health
Physics. A.M.F. Duhamel (Ed.). Pergamon Press, Elmsford,
NY, pp. 707–760.
15. Domenico, P.A. (1987). An analytical model for multidi-
mensional transport of a decaying contaminant species. J.
Hydrolo. 91: 49–58.
16. Sun, N. (1996). Mathematical Modeling of Ground Water
Pollution. Springer, New York.
17. Bedient, P.B., Rifai, H.S., and Newell, C.J. (1999). Ground-
water Contamination: Transport and Remediation. Prentice-
Hall, New York.
18. Zheng, C. (1990). MT3D, a modular three-dimensional
transport model for simulation of advection, dispersion and
chemical reactions of contaminants in ground water systems.
Report to the US Environmental Protection Agency.
19. Zheng, C. and Wang, P.P. (1999). MT3DMS, a modular three-
dimensional multi-species transport model for simulation of
advection, dispersion and chemical reactions of contaminants
in groundwater systems; documentation and user’s guide.
U.S. Army Engineer Research and Development Center,
Contract Report SERDP-99-1, Vicksburg, MS.
REACTIVE TRANSPORT IN THE SATURATED
ZONE: CASE HISTORIES FOR PERMEABLE
REACTIVE BARRIERS
LEN LI
CRAIG H. BENSON
University of
Wisconsin-Madison
Madison, Wisconsin
INTRODUCTION
The permeable reactive barrier (PRB) is an in situ
remediation technology where contaminated groundwa-
ter is treated passively as it flows through a reactive
medium. Most PRBs use granular zero valent iron (ZVI)
to create redox conditions that result in degradation
or immobilization of groundwater contaminants, such
as chlorinated solvents and herbicides, heavy metals,
 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS
and radionuclides (1–5). These redox conditions promote
precipitation of secondary minerals from ions typically
in groundwater as well as some groundwater contam-
inants (2,6,7). The most common minerals formed in
PRBs are magnetite, hematite, goethite, lepidocrocite,
calcite, aragonite, siderite, green rust, ferrous hydrox-
ide, ferrous sulfide, and marcasite (8). Calcium carbon-
ates and siderite typically are found near the entrance
face of a PRB, whereas magnetite, ferrous hydrox-
ide, green rust, and iron oxyhydroxides form through-
out a PRB (6,7). Accumulation of secondary minerals
causes loss of pore space and reactive surface area
of the medium in PRBs, which can alter flow paths,
residence times, and the treatment effectiveness of a
PRB (2,7,9).
Developing a general assessment of mineral precip-
itation in PRBs with field and/or laboratory data is
complicated by site-specific geochemical and hydrogeologi-
cal conditions (7), and the relatively long period over which
mineral deposition occurs (9). An alternative approach is
to conduct analyses with a reactive transport simula-
tor that predicts mineral deposition and its effects on
hydraulics (10–12). Li (12) developed a reactive trans-
port simulator specifically for this purpose. This chapter
describes a comparison between ion concentrations pre-
dicted with this model and concentrations measured in
the field at three PRBs.
REACTIVE TRANSPORT SIMULATOR
Equilibrium and Kinetic Reactive Transport Models
Equilibrium and kinetic models have simulated geochemi-
cal processes and mineral formation in PRBs (10,11,13,14).
Equilibrium models can estimate the aqueous species and
solid phases that exist in a state of thermodynamic equi-
librium, but they do not account for the rate at which
reactions occur and generally do not involve the move-
ment of groundwater. In contrast, kinetic models describe
the rate at which geochemical reactions occur and predict
aqueous and solid phase concentrations as a function of
time. In kinetic models, the geochemical reactions usually
are specified by the user along with reaction rate expres-
sions and the rate coefficients. Thus, when a kinetic model
is developed, an equilibrium model often defines the most
important reactions to be included in the kinetic model
(e.g., see Yabusaki et al. (10)).
Equilibrium models are appropriate when the reaction
rate is fast compared with the flow rate, whereas
kinetic models are appropriate when flow rates are
not negligible relative to reaction rates. For most
PRBs, flow rates are fast enough such that reaction
kinetics must be considered when evaluating mineral
precipitation (10). Reactive transport codes incorporating
kinetic geochemical algorithms have been developed by
Mayer et al. (11) and Yabusaki et al. (10) for simulating
contaminant degradation and mineral precipitation in
PRBs. The code developed by Li et al. (8) is similar to those
developed by Mayer et al. (11) and Yabusaki et al. (10),
but it focuses solely on geochemical reactions that result
in mineral precipitation.
519
Conceptual Model
The conceptual model used by Li et al. (8) consists of
a continuous-wall PRB containing ZVI that is placed
orthogonal to the groundwater flow. Groundwater flowing
into the PRB is assumed to contain dissolved oxygen,
Fe2+ , Ca2+ , Mg2+ , SO4 2− , CO3 2− , OH− , HCO3 − , and
NO3 − . These dissolved species are assumed to form
CaCO3 , CaMg(CO3 )2 , FeCO3 , Fe(OH)2 (am), FeS (am),
and Mg(OH)2 in response to the redox conditions
induced by corrosion of the ZVI. In an actual PRB,
additional ions may be present and minerals other than
those included in the model may also form. However,
the reactions in the model include the most common
minerals observed in laboratory and field studies of
PRBs (6,7,10,11). Contaminant transformation reactions
were not included in the model because the volume
of mineral precipitates formed as a consequence of
contaminant treatment generally is negligible, even
though the contaminant transformations are intrinsic to
the remediation process (2,11).
Codes
The reactive transport simulator is created with the
public domain codes MODFLOW and RT3D (15,16). MOD-
FLOW simulated steady-state flow in the aquifer and
PRB. Geochemical processes within the PRB were sim-
ulated with RT3D, a three-dimensional multicomponent
model that describes advective–dispersive-reactive trans-
port (16). Heads from MODLOW are input to RT3D. A
detailed description of the codes that comprise the simu-
lator can be found in Li (12) and Li et al. (8).
Li et al. (8) developed a geochemical module for
RT3D that simulated the mineral-forming reactions.
Geochemical reactions included in the module are
summarized in Table 1. Three redox reactions are related
to iron corrosion, along with six mineral precipitation
reactions and one reaction for microbial sulfate reduction.
A pseudo-first-order reaction rate was assumed for iron
corrosion by oxygen or nitrate with the reaction rate
proportional to the iron surface area and the concentration
of oxygen or nitrate (11). For iron corrosion by water,
the reaction rate was assumed to be proportional
only to the iron surface area (10,11,17). The rate of
degradation of sulfate to HS− by sulfate-reducing bacteria
was assumed to follow the Monod equation (11), and
the kinetics of mineral precipitation were assumed
to follow transition state theory (18). The kinetic rate
expressions and solubility constants for the mineral
reactions are summarized in Table 1. All geochemical
reactions are assumed to occur in parallel and are solved
simultaneously.
FIELD DATA SOURCES AND MODEL SETUP
Field data from three PRBs verified that the MODFLOW-
RT3D simulator produces reasonable predictions of major
ion concentrations. Comparisons were not made between
measured and predicted solid phase concentrations at
any of the sites because of a lack of field data. The
selected PRBs are located at Moffett Federal Airfield in
520 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS

Table 1. Geochemical Reactions and Solubility Constants

Solubility
Mineral Kinetic Rate Constant
Reaction Formed Expressiona Log(Keq )b

Iron Corrosion:
Fe0 + H2 O + 0.5O2 (aq) → Fe2+ + 2OH− — kan,Fe SaDO —
Fe0 + 2H2 O → Fe2+ + H2 (aq) + 2OH− — kan,Fe S —
4Fe0 + 7H2 O + NO− 3 → 4Fe
2+
+ 10OH− + NH+
4 — kNO3 SaNO−3 —
Microbial Sulfate-Reduction:  

− − [SO2−
4 ] [H2 ]
SO2−
4 + 4H2 (aq) → HS + OH + 3H2 O — k —
KSO2− + [SO2−
4 ]
KH2 + [H2 ]
4
Secondary Mineral Precipitation-Dissolution:
CaCO3 (s) ↔ Ca2+ + CO2−
3 Aragonite/Calcite −kCa (1 − aCa2+ aCO2− /Keq ) −8.1
3
CaMg(CO3 )2 (s) ↔ Ca2+ + Mg2+ + 2CO2−
3 Dolomite −kCaMg (1 − aCa2+ aMg2+ a2CO2− /Keq ) −17.7
3
Mg(OH)2 (s) ↔ Mg2+ + 2OH− Brucite −kMg2 (1 − aMg2+ a2OH− /Keq ) −11.2
FeCO3 (s) ↔ Fe2+ + CO2−
3 Siderite −kFe1 (1 − aFe2+ aCO2− /Keq ) −10.5
3
Fe(OH)2 (am) ↔ Fe2+ + 2OH− Ferrous hydroxide −kFe (1 − aFe2+ a2OH− /Keq ) −15.2
FeS(am) + H2 O ↔ Fe2+ + HS− + OH− Ferrous sulfide −kFeS (1 − aFe2+ aHS− aOH− /Keq ) −18.4
a
k is rate coefficient, S is reactive surface area of ZVI, a is activity of aqueous species, and Keq is solubility constant. Activity coefficients were calculated with
the extended Debye—Huckel equation.
b
At 15 ◦ C and 101 kPa.

Mountain View, CA; Monticello, UT; and Copenhagen,
Denmark. These PRBs were selected because hydrological
and geochemical data were sufficiently well defined for
each PRB to permit an assessment of the simulator.
For all sites, the simulations were set up assuming
homogeneous hydraulic conductivity in each medium (i.e.,
aquifer, reactive zone, etc.). Constant-head boundaries
were applied along the upstream and downstream faces of
the domain to establish a hydraulic gradient orthogonal
to the PRB. No flow conditions were imposed at the other
boundaries. Each species in the PRBs was assumed to have
an initial concentration of zero. A constant concentration
boundary was applied at the upstream face of the PRBs,
and other boundaries were assigned the Cauchy boundary
condition with no dispersive flux. Influent groundwater
concentrations entering the PRB were set at values
measured in the field.
Moffett Federal Airfield
A pilot-scale funnel-and-gate PRB consisting of ZVI was
installed in 1996 at Moffett Federal Airfield to remove
TCE and DCE from groundwater (10). The gate is 3.0 m
long (orthogonal to flow direction), 3.0 m wide (along the
flow direction), and 5.5 m deep. The gate consists of a
1.8-m-thick barrier of granular ZVI sandwiched between
two 0.6-m-thick layers of pea gravel (up gradient and
down gradient). Groundwater samples collected within the
PRB between April 1996 (installation) and October 1997
were analyzed for target contaminants, pH, major cations
and anions, and indicator parameters (10). Analysis of
core samples showed that magnetite, hematite, aragonite,
and marcasite formed in the iron media. No direct
measurements were made of siderite, FeS, Fe(OH)2 (am),
Fe(OH)3 , brucite, or green rust.
The MODFLOW-RT3D simulator simulated the nearly
one-dimensional conditions in the gate at the Moffett
Federal Airfield site. The input was based on data provided
in Yabusaki et al. (10). MODFLOW simulated steady-state
flow in the pea gravel zones and iron media zone. The
initial hydraulic conductivity in the PRB was 186 m/d,
and the porosity was 0.66. The hydraulic conductivity in
the pea gravel zones was 86.4 m/d. A hydraulic gradient
of 0.0003 was applied across the PRB to yield the Darcy
velocity (0.04 m/d) reported in Yabusaki et al. (10). Five
minerals reported by Yabusaki et al. (10) were included
in the model: CaCO3 , FeCO3 , Fe(OH)2 (am), Mg(OH)2 ,
and FeS (am). Composition of groundwater entering the
PRB was based on information provided in Yabusaki
et al. (10) (Table 2).
Monticello
A full-scale funnel and gate PRB containing granular iron
was installed in 1999 to treat uranium in groundwater at
a former uranium mill site near Monticello, UT (14). The
gate is 30.0 m long, 2.4 m wide (along flow direction), and
4.5∼6.0 m deep. A 1.2-m-thick reactive zone consisting of
100% granular ZVI was sandwiched between two 0.6-m-
thick layers of pea gravel (up gradient and down gradient)
in the gate. Samples of groundwater were collected from
June 1999 (installation) to September 1999 and analyzed
for target contaminants, pH, major cations and anions,
and indicator parameters. Core samples showed that
magnetite and CaCO3 formed within the PRB.
Reactive transport in the gate at the Monticello PRB
was simulated with the MODFLOW-RT3D simulator
using a one-dimensional model similar to that employed
for the PRB at Moffett Federal Airfield (see previous
section). A one-dimensional model was selected because
flow through the pea gravel zones and ZVI was nearly one-
dimensional and steady state. Hydraulic conductivities of
the ZVI and pea gravel zones were set at 30.9 m/d and 8.64
m/d. Based on data in Morrison et al. (14), the porosity
was set at 0.50 in the ZVI and 0.30 in the pea gravel
and the hydraulic gradient was set at 0.142 to yield a
 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS 521

Table 2. Concentrations of Aqueous Species in Influent Ground Water

Aqueous Species Concentrations (M)

Aqueous Species in Moffett Federal Airfield, Former Mill Site, Copenhagen Freight
Influent Ground Mountain View, Monticello, Yard, Copenhagen,
Water Californiaa Utahb Denmarkc

Fe2+ 1.0 × 10−10 1.0 × 10−10 1.0 × 10−10

Ca2+ 3.9 × 10−3 8.5 × 10−3 3.3 × 10−3
Mg2+ 2.7 × 10−3 3.6 × 10−3 1.2 × 10−3
OH− 1.3 × 10−7 5.2 × 10−8 5.0 × 10−7
HCO− 3 4.7 × 10−3 2.5 × 10−3 1.0 × 10−2
CO3 2− 1.4 × 10−6 1.0 × 10−6 1.0 × 10−6
O2 (aq) 1.0 × 10−10 1.0 × 10−10 1.0 × 10−10
NO3 − 3.9 × 10−5 1.0 × 10−10 1.0 × 10−10
SO4 2− 3.6 × 10−3 1.2 × 10−2 1.2 × 10−3
a
From Yabusaki et al. (10).
b
From Morrison et al. (14).
c
From Killerich et al. (19).

groundwater velocity of 3 m/d. Five minerals were included
in the geochemical model: CaCO3 , FeCO3 , Fe(OH)2 (am),
CaMg(CO3 )2 , and FeS. Composition of the groundwater
entering the PRB was defined based on data reported in
Morrison et al. (14), as summarized in Table 2.
Copenhagen
A full-scale PRB containing granular ZVI was installed
in 1998 at the Copenhagen Freight Yard in Copen-
hagen, Denmark to treat chlorinated solvents in ground
water (19). The PRB is a continuous trench 15.0 m long,
1.3 m wide (along flow direction), and 6.0 m deep and
consists of a 0.9-m-thick zone of 100% granular ZVI sand-
wiched between 0.2-m-thick layers of pea gravel. Samples
of groundwater were collected from July 1998 (installation)
to September 1999 and were analyzed for target contam-
inants, pH, and major cations and anions (19). No direct
measurements were made of minerals forming within the
PRB (19).
The MODFLOW-RT3D simulator was set up to
simulate one-dimensional flow and reactive transport
along the centerline of the Copenhagen PRB. The model
predicted steady-state flow in the surrounding aquifer, pea
gravel zones, and ZVI. Hydraulic conductivity of the ZVI,
pea gravel, and aquifer were set at 6.1, 2.1, and 0.6 m/d,
respectively, based on site conditions (19). The porosity
was set at 0.60 for the ZVI, 0.40 for the pea gravel zones,
and 0.3 in the aquifer (19). A hydraulic gradient of 0.011
was applied across the aquifer based on data reported in
Killerich et al. (19). Five minerals were included in the
model: CaCO3 , FeCO3 , Fe(OH)2 (am), CaMg(CO3 )2 , and
FeS. Composition of the groundwater entering the PRB
was defined based on data reported in Killerich et al. (19),
as summarized in Table 2.
COMPARISON OF PREDICTED AND MEASURED
CONCENTRATIONS
Moffett Federal Airfield
Comparisons of the predicted and measured pH and
concentration profiles are shown in Fig. 1 for the PRB
at Moffett Field after 12 months of operation. The reaction
rate coefficients initially were estimated from data in
the literature, but ultimately they were adjusted by
trial-and-error to reproduce the measured pH, alkalinity,
and concentrations of SO4 2− , Ca2+ , and Mg2+ . A similar
calibration procedure was used at the other two sites. The
calibrated reaction rate coefficients are summarized in
Table 3 along with typical ranges based on data reported
in the literature.
The predicted pH is in general agreement with the
observed pH [Fig. 1(a)], including the rapid jump near
the entrance face and the leveling out in the interior.
The drop in alkalinity just up gradient of the PRB is
not captured [Fig. 1(b)], but the alkalinity predicted by
RT3D inside the PRB is similar to that measured in the
field. The sulfate concentration predicted by RT3D also
is in general agreement with the trends observed in the
field [Fig. 1(b)], as are the Ca2+ and Mg2+ concentrations
[Fig. 1(c)]. However, the predicted Ca2+ concentrations
in the PRB are near the upper bound of the measured
concentrations [Fig. 1(c)]. The field data are also too sparse
to evaluate whether the simulator accurately predicts the
rapid changes in geochemical conditions occurring at the
entrance face.
Monticello Mill Site
Comparisons of measured and predicted pH, alkalinity,
and concentrations of Ca2+ , Mg2+ , and SO4 2− for conditions
at 3 months of operation are shown in Fig. 2. Calibrated
rate coefficients for the PRB at Monticello are summarized
in Table 3. As with the PRB at Moffett Field, insufficient
data exist to determine whether the rapid changes
predicted at the entrance face accurately replicate the
conditions occurring in the PRB at Monticello.
The measured and predicted pH, HCO3 − , Mg2+ , and
SO4 2− concentrations in the Monticello PRB are in close
agreement (Fig. 2). However, the predicted Ca2+ concen-
trations in the PRB are higher than are the measured
concentrations [Fig. 2(b)]. One reason for this discrepancy
may be that minerals other than CaCO3 and CaMg(CO3 )2
are precipitating (these minerals only account for 15%
522 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS

Table 3. Calibrated Rate Coefficients in PRB Case Studies

Rate Coefficients
Geochemical Moffett Federal Former Mill Copenhagen Freight Reported in
Reactions Terms Units Airfieldc Sitec Yardc Literatured

Oxygen Corrosiona m3 /m2 -d 2.8 × 10−2 2.8 × 10−2 2.8 × 10−2 < 2.8 × 104
Water Corrosiona mole/m2 -d 7.0 × 10−7 1.5 × 10−5 1.9 × 10−7 3.0 × 10−8 − 5.4 × 10−3
Nitrate Corrosiona m3 /m2 -d 2.6 × 10−5 — 2.8 × 10−5 1.8 × 10−8 − 2.8 × 10−5
Microbial sulfate reductionb M/d 1.0 × 10−3 1.0 × 10−3 4.0 × 10−3 5.0 × 10−6 − 5.0 × 10−3
CaCO3 M/d 1.0 × 10−4 4.0 × 10−4 1.0 × 10−5 2.7 × 10−9 − 1.4 × 10−4
FeCO3 M/d 1.1 × 10−5 2.1 × 10−5 1.0 × 10−5 1.1 × 10−5 − 2.7 × 10−4
Fe(OH)2 (am) M/d 5.1 × 10−5 5.1 × 10−5 5.0 × 10−5 <2.2 × 10−4
FeS (am) M/d 1.0 × 10−6 1.0 × 10−6 1.0 × 10−6 1.1 × 10−8 − 2.2 × 10−5
CaMg(CO3 )2 M/d — 1.0 × 10−6 1.0 × 10−9 6.9 × 10−6
Mg(OH)2 M/d 2.0 × 10−3 NA NA 3.4 × 10−6
a
Reactive surface area of ZVI set at 3.9 × 106 m2 /m3 .
b
Half-saturation constants for SO4 2− and H2 (aq) were KSO4 = 1.0 × 10−4 M, and KH2 = 1.0 × 10−7 M, respectively.
c
After calibration with field measurements.
d
Li et al. (8).

(a) 12 (a) Pea gravel ZVI

ZVI
10
11 0.015
Concentration (M)

pH
SO42−
10 9
0.010
pH

pH

pH
9 PH - Field data
Pea gravel HCO3− - Field data 8
0.005
SO42− - Field data
8 Field pH data
HCO3−

7 0.000 7

(b) ZVI
0.005
Field alkalinity data
(b) 0.010
Field SO42− data
Concentration (M)

0.004 Alkalinity Ca2+ - Field data

Ca2+ Mg2+ - Field data
0.003 0.008
Concentration (M)

0.002 SO42−
0.006
0.001
Pea gravel
0.004 Mg2+
0.000

(c) 0.005
ZVI 0.002
Field Ca2+ data Pea gravel
0.004 Ca 2+ ZVI
Field Mg2+ data
Concentration (M)

0.000
0.003 0.0 0.6 1.2 1.8
Mg2+ Distance (m)
0.002 Figure 2. Measured and predicted concentrations for the PRB at
the former Uranium Mill Site in Monticello, UT, after 3 months
0.001 of operation: (a) pH, HCO3 − , and SO4 2− and (b) Ca2+ and Mg2+ .
Pea gravel

0.000
0.0 0.6 1.2 1.8 2.4
removal in the ZVI. However, neither of these minerals
Distance into PRB (m)
has been observed in the core samples from the PRB at
Figure 1. Measured and predicted concentrations for the PRB Monticello. The surface complexation of calcium on the
at Moffett Federal Airfield, Mountain View, CA, after 12 months ZVI and secondary minerals surfaces, neither of which are
of operation: (a) pH, (b) alkalinity and SO4 2− , and (c) Ca2+ and
considered in the simulator, may also be removing calcium
Mg2+ .
from the solution (10).

Copenhagen Site
of aqueous calcium). For example, Morrison et al. (14)
included iron-rich calcite (Fe0.4 Ca0.6 CO3 ) and gypsum Comparisons of measured and predicted pH and Ca2+ ,
(CaSO4 ) in their equilibrium model to account for calcium Mg2+ , SO4 2− , and HCO3 − concentrations after 15 month
 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS 523

of operation of the Copenhagen PRB are shown in Fig. 3.
The calibrated rate coefficients corresponding to these
predictions are summarized in Table 3.
Close agreement exists between the measured and
predicted pH and the HCO3 − , SO4 2− , Ca2+ , and Mg2+
concentrations in the ZVI of the Copenhagen PRB
[Figs. 3(a) and (b)], but the field data are too sparse to
evaluate whether the trends in geochemistry predicted by
the model are accurate. In contrast to the PRBs at Moffett
Field and Monticello, the predicted pH climbs gradually
at the Copenhagen site. This more graduate change in pH
may be caused by a lower rate of iron corrosion by water
(1.9 × 10−7 mole/m2 -d, Table 3), which results in less OH−
released into solution.
Comparison of Calibrated Rate Coefficients and Literature
Values
The calibrated rates coefficients for the three PRB case
studies are summarized in Table 3. Most of the calibrated
rate coefficients fall within the range of rate coefficients
from the literature (Table 3). The exceptions are the
rate coefficient for Mg(OH)2 at the Moffett PRB and the
rate coefficient for CaCO3 at the Monticello PRB. Both
rate coefficients are higher than are the maximum values
reported in the literature. However, the rate coefficient for
CaCO3 is only slightly above the upper bound from the
literature, and only one rate coefficient was available in
the literature for Mg(OH)2 . Thus, these exceptions are not
considered unusual.
(a) 0.012 Pea gravel
ZVI pH - Field data
HCO3− - Field data
0.010
SO42− - Field data
HCO3−
Concentration (M)
Although good agreement between measured and
predicted concentrations has been shown in Figs. 1–3, the
accuracy of the simulator when applied to other problems
will depend greatly on the rate coefficients used as input.
For example, the effect of varying the rate coefficient for
iron corrosion by water (Table 1) within a reasonable range
(1.0 × 10−7 to 1.0 × 10−5 mole/m2 -d) on pH is illustrated
in Fig. 4 with the model for the PRB at Moffett Field. The
influent concentrations used for the simulation are shown
in Table 2, and the rate coefficients for the reactions other
than iron corrosion by water are shown in Table 3. When
the rate coefficient for iron corrosion by water is 10−7
mole/m2 -d, the predicted pH is much lower than the pH
observed in the field, especially near the entrance. In
contrast, when the rate coefficient for iron corrosion by
water is 10−5 mole/m2 -d, the predicted pH increases too
rapidly near the entrance and levels out approximately
one pH unit higher than observed in the field.
In situ rate coefficients can vary considerably and
depend on a variety of factors, and no method currently
is available to predict rate coefficients accurately a
priori. Thus, at the current time, simulators must be
calibrated with field data if reliable predictions of mineral
precipitation are to be obtained. Alternatively, parametric
analyses can be conducted to evaluate the potential range
of conditions that may be realized in the field.
SUMMARY
Comparisons between predicted and measured concentra-
tions of common groundwater ions in PRBs have been
described in this article. The simulator used for the pre-
11 dictions is based on the public domain codes MODFLOW
and RT3D, the latter being augmented with a geochemical
algorithm specifically for simulating geochemical reac-
10

0.008 tions in PRBs. Field data from three PRBs were used for
the comparison.
0.006 The comparisons showed that pH profiles and anion and
pH

9
pH cation concentrations in PRBs can be predicted accurately.
0.004 The ability to predict solid phase concentrations has not
8 been evaluated because of lack of field data. However,
0.002
SO42−
0.000
7 12
Pea gravel ZVI
(b)
Ca2+ - Field data
Ca2+ Mg2+ - Field data 11
0.003
Concentration (M)

10
pH

0.002

9
Mg2+
0.001 Field pH data
Pea gravel 8 k = 10−7 mole/m2-d
ZVI k = 7 × 10−7 mole/m2-d
0.000 k = 10−5 mole/m2-d
1.0 1.2 1.4 1.6 1.8 2.0 7
Distance (m) 0.0 0.6 1.2 1.8 2.4
Figure 3. Measured and predicted concentrations at the center- Distance into PRB (m)
line of the PRB at the Copenhagen Freight Yard, Copenhagen, Figure 4. Measured and predicted pH for the PRB at Moffett
Denmark, after 15 months of operation: (a) pH, HCO3 − , and Federal Airfield, Mountain View, CA, after 12 months of operation
SO4 2− and (b) Ca2+ and Mg2+ . using various rate coefficients (k) for iron corrosion by water.
524 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER
the accuracy of the predictions depends to a great extent
on the reaction rate coefficients used as input. In this
study, the rate coefficients were obtained by calibration
because the actual in situ rate coefficients were unknown.
Thus, predictions made a priori are prone to considerable
uncertainty until methods are developed to accurately
define reaction rate coefficients.
In many cases, the simulator predicted that the greatest
changes in geochemistry generally occur at the entrance
face of the PRB. However, the current field data sets
contain insufficient detail to confirm that these changes
in geochemistry are being predicted accurately. More
detailed field observations are needed at the entrance
face of the PRBs to more fully evaluate whether the
simulator can accurately capture these rapid changes in
geochemistry.
BIBLIOGRAPHY
1. Blowes, D.W. et al. (2000). Treatment of inorganic contam-
inants using permeable reactive barriers. J. Contaminant
Hydrol. 45: 123–137.
2. Wilkin, R.T., Puls, R., and Sewell, G.W. (2003). Long-term
performance of permeable reactive barriers using zero-valent
iron: geochemical and microbiological effects. Ground Water
41: 493–503.
3. EPA. (1998). Permeable Reactive Barrier Technologies for
Contaminant Remediation. United States Environmental
Protection Agency, Washington, DC.
4. Lee, T. and Benson, C.H. (2004). Sorption and degradation of
alachlor and metolachlor in ground water using green sands.
J. Environ. Quality 33: 1682–1693.
5. Lee, T., Benson, C.H., and Eykholt, G.R. (2004). Waste green
sands as reactive media for groundwater contaminated
with trichloroethylene (TCE). J. Hazardous Materials B109:
25–36.
6. Mackenzie, P.D., Horney, D.P., and Sivavec, T.M. (1999).
Mineral precipitation and porosity losses in granular iron
columns. J. Hazardous Materials 68: 1–17.
7. Phillips, D.H. et al. (2003). Mineralogical characteristics and
transformations during long-term operation of a zerovalent
iron reactive barrier. J. Environ. Quality 32: 2033–2045.
8. Li, L., Benson, C.H., and Mergener, E.A. (2004). Impact of
mineral fouling on hydraulic behavior of permeable reactive
barriers. Ground Water. In press.
9. Vikesland, P.J. et al. (2003). Longevity of granular iron in
groundwater treatment processes: changes in solute transport
properties over time. J. Contaminant Hydrol. 64: 3–33.
10. Yabusaki, S. et al. (2001). Multicomponent reactive transport
in an in situ zero-valent iron cell. Environ. Sci. Technol. 35:
1493–1503.
11. Mayer, K.U., Blowes, D.W., and Frind, E.O. (2001). Reactive
transport modeling of an in situ reactive barrier for the
treatment of hexavalent chromium and trichloroethylene in
groundwater. Water Resour. Res. 37: 3091–3103.
12. Li, L. (2004). Fouling and the Long-Term Performance of
Permeable Reactive Barriers. Ph.D. Dissertation, University
of Wisconsin-Madison, Madison, WI.
13. Liang, L. et al. (2002). Integration of Geochemical Data and
Hydraulic Changes at Permeable Reactive Barriers Sites. Oak
Ridge National Laboratory, Oak Ridge, TN.
14. Morrison, S.J., Metzler, D.R., and Carpenter, C.E. (2001).
Uranium precipitation in a permeable reactive barrier
by progressive irreversible dissolution of zerovalent iron.
Environ. Sci. Technol. 35: 385–390.
15. McDonald, M.G. and Harbaugh, A.W. (1988). MODFLOW: A
Modular Three-Dimensional Finite-Difference Ground-Water
Flow Model. U.S. Geological Survey Techniques of Water-
Resources Investigations, Washington, DC, book 6, chap. A1.
16. Clement, T.P. (1997). RT3D: A Modular Computer Code
for Simulating Reactive Multi-Species Transport in 3-
Dimensional Groundwater Aquifers, PNNL-SA-11720. Pacific
Northwest National Laboratory, Richland, WA.
17. Reardon, E.J. (1995). Anaerobic corrosion of granular iron:
Measurement and interpretation of hydrogen evolution rates.
Environ. Sci. Technol. 29: 2936–2945.
18. Lasaga, A.C. (1998). Kinetic Theory in the Earth Sciences.
Princeton series in geochemistry. Princeton University Press,
Princeton, NJ.
19. Killerich, O., Larsen, J.W., and Nielsen, C. (2000). Field
results from the use of a permeable reactive wall. Proceedings
of the Second International Conference on Remediation of
Chlorinated and Recalcitrant Compounds, Monterey, CA.
TRANSPORT OF REACTIVE SOLUTE IN SOIL
AND GROUNDWATER
NATALIE L. CAPIRO
PHILIP B. BEDIENT
Rice University
Houston, Texas
INTRODUCTION
Remediation of contaminated aquifers, both vadose zone
(unsaturated portion of an aquifer) soils and groundwater,
has proven to be a challenging environmental endeavor.
Groundwater is the source of about 60% of the United
States’ water usage and, with contaminant plumes
impacting many states, associated human health concerns
of great importance exist. Groundwater can become
polluted both directly by leaking underground storage
tanks and indirectly by leakage of contaminants on the
soil surface that leach to the groundwater. Common
pollutants include agricultural chemicals applied to crops,
industrial spills, landfill leachate, disposal sites, and waste
ponds. Subsurface remediation is complicated by not only
the behavior of the contaminants themselves but by the
heterogeneities found in typical aquifer formations. The
processes that govern the fate and transport of solute
in aquifer systems generally fall into the categories of
hydrodynamic, abiotic, and biotic.
The universal process that impacts all forms of
solute include the hydrodynamic processes, which affect
contaminant transport by impacting the flow paths of
groundwater in the subsurface, including both the vadose
zone and the saturated aquifer. Abiotic processes affect
contaminant transport physically by causing interactions
between the contaminant and the stationary subsurface
materials (i.e., sorption) or by affecting the form of the
contaminant (i.e., through hydrolysis and redox reactions).
Biotic processes are driven by microbial activities and
can affect contaminant transport by metabolizing or
 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER
mineralizing the contaminant (i.e., organic contaminants),
or possibly by using the contaminant in the metabolic
process (i.e., nitrate under denitrifying conditions).
PHYSICAL MECHANISMS OF SOLUTE TRANSPORT
The three basic hydrodynamic processes affecting con-
taminant transport in saturated and unsaturated porous
media are advection, diffusion, and hydrodynamic disper-
sion. Advection represents the movement of a contaminant
with the flowing groundwater according to the seepage
velocity in the pore space. Advection is the most important
process driving the down gradient migration of dissolved
contaminants in the saturated subsurface (1). If advection
was the only process acting on a tracer, a sharp inter-
face would exist continuously as the tracer flushes out
the solute-free water. In essence, the tracer would behave
as a plug. A breakthrough curve, illustrating the rela-
tive effluent concentration (C/Co) vs. time in a continuous
tracer input, would appear as a step function (Fig. 1a).
The average velocity of advective transport is described by
the amended Darcy equation:
K dh
vx = − (1)
n dL
where vx is the average linear groundwater velocity
[L/t], K is the hydraulic conductivity (a measure of a
porous media’s ability to transmit water) [L/t], n is the
effective porosity [unit less], and dh/dL is the hydraulic
gradient [L/L].
The average linear velocity or seepage velocity, vx ,
equals the Darcy velocity divided by effective porosity,
n, associated with the pore space through which water
can actually flow. The average linear velocity is less
than the microscopic velocities of water molecules moving
along individual flowpaths, because of tortuosity. The one-
dimensional (1-D) mass flux (Fx ) because of advection
equals the product of water flow and concentration (C) of
solute, or Fx = vx nC.
Darcy’s law is also still valid for the macroscopic view
(through the entire pore body) of unsaturated flow, except
that hydraulic conductivity is now a function of moisture
content. Darcy’s law is then used with the unsaturated
value for K and can be written as
v = −K(θ )dh/dz (2)
where v is Darcy velocity, z is depth below surface [L], K(θ )
is unsaturated hydraulic conductivity [L/T], θ is volumetric
moisture content, h is potential or head (h = z + ψ) [L/T],
and ψ is tension or suction [L]. For saturated soils,
the hydraulic conductivity was assumed to be constant;
however, in unsaturated soil, it decreases sharply with
decreasing water content, because as water drains from
pores, the flow decreases (2).
Although advection is associated with bulk macroscopic
groundwater movement, diffusion is a molecular-based
phenomenon. The process involves spreading because of
concentration gradients and random motion. Diffusion
causes a solute in water to move from an area of higher
525
concentration to an area of lower concentration. Diffusive
transport can occur in the absence of velocity. Mass
transport in the subsurface because of diffusion in one
dimension can be described by Fick’s first law of diffusion,
dC
fx = −Dd (3)
dx
where fx is the mass flux [M/L2 /T], Dd is the diffusion
coefficient [L2 /T], and dC/dx is the concentration gradi-
ent [M/L3 /L].
Typical values of Dd are relatively constant and range
from 10−4 to 10−5 cm2 /sec (3). In cases of very low
velocities, such as in a tight soil or clay liner, or in
the case of mass transport involving very long time
periods, diffusion is the dominant hydrodynamic process.
On the other hand, in areas of high groundwater velocity,
molecular diffusion can actually be ignored compared with
advection and other dispersion mechanisms.
Freeze and Cherry (4) defined hydrodynamic dispersion
as the process in which solutes spread out and are diluted
compared with simple advection alone. It is defined as
the sum of molecular diffusion and mechanical dispersion,
where diffusion is caused by concentration gradients alone
and dispersion is generally caused by local variations in
velocity around some mean velocity of flow. The net effect of
the dispersive phenomena is to decrease the concentration
gradient between two different solutions of the same
solute. In essence, dispersion reduces the sharp interface
between the two solutions. Mass transport because of
dispersion can also occur normal to the direction of flow.
This transverse dispersion, Dy , is caused by diverging
flowpaths in the porous media and concentration gradients
that cause mass to spread laterally from the main direction
of flow. In most cases involving a two-dimensional plume
of contamination, Dy is much less than Dx , and the shape of
the plume tends to be elongated in the direction of flow (5).
Many typical contaminant plumes in groundwater are
represented by 2-D advective-dispersive mechanisms.
One-Dimensional Transport Equations
The governing mass transport equation is difficult to solve
in field cases of practical interest because of boundary
irregularities and variations in aquifer characteristics,
so numerical methods must generally be employed (6).
However, a limited number of relatively simple, 1-D
analytical solutions exist. The simplifying assumptions
include the following: (a) the tracer is ideal, with constant
density and viscosity; (b) the fluid is incompressible;
(c) the medium is homogeneous and isotropic; and (d) only
saturated flow is considered. Initial conditions (t = 0) in a
soil column are usually set to zero (C(x, 0) = 0) or to some
constant background concentration. Boundary conditions
must be specified at the two ends of the 1-D column. For
a continuous source load at x = 0, the concentration is set
to C(0, t) = C0 for t > 0. The concentration at the other
boundary, x = ∞ is set to zero C(x, ∞) = 0 for t > 0.
For an infinite 1-D column with background concentra-
tion of zero and a continuous input tracer concentration C0
at −∞ ≤ x ≤ 0 for t ≥ 0, Bear (7) solves the problem using
526 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER
the Laplace transform at x = L, length of the column

 
 
C(x, t) 1 L − vx t vx L L + vx t
= erfc √ + exp erfc √
C0 2 2 Dx t Dx 2 Dx t
(4)
where erfc is the complementary error function, erfc(x) =
√ x
1 − erf(x) = 1 − (2/ π ) e−u du, Dx is the coefficient
2
0 dation.
of hydrodynamic dispersion, and vx is the average
seepage velocity. The center of mass (C/C0 = 0.5) of
the breakthrough curve travels with the average linear
velocity, vx , and corresponds to the point where x = vx t.
Note that the second term on the right-hand side of Eq. (4)
can generally be neglected for most practical problems.
The error functions erf (β) and erfc(β) are tabulated
in Domenico and Schawartz (3).
The corresponding solution can be derived for the
injection of a tracer pulse (instantaneous input) at x = 0
with background concentration equal to zero in the 1-D
column. As the slug moves downstream with vx in the +x
direction, it spreads out according to
 
M (x − vx t)2
C(x, t) = exp − (5)
(4π Dx t)1/2 4Dx t
where M is the injected mass per unit cross-sectional area.
Figure 1b shows the resulting Gaussian distribution of
concentration for the instantaneous pulse source in one
dimension.
Plots comparing the shapes of Eqs. (4) and (5) are
shown in Figs. 1a and 1b, respectively, at x = L at the end
of a soil column. The differences between instantaneous
(spike) source and continuous source transport problems
are obvious in one dimension. The continuous source
produces a response or breakthrough curve, which starts
at a low concentration and eventually levels off to the
initial input concentration C0 as a function of time. The
spike source produces a normal or Gaussian distribution,
(a)
Co
(b)
Figure 1. Breakthrough curves in one dimen-
sion for: (a) continuous and (b) instantaneous
tracer inputs.
which continues to decrease in maximum concentration
 because of spreading out with time in the direction of flow.
For the spike source, the amount of mass under each curve
is identical if the tracer is conservative.
Additional information can be found on transport
equations in several sources (3–6,8), including three-
dimensional derivations and those incorporating degra-
Techniques for Determining Hydrodynamic Parameters
Tracers and tracer tests are valuable methods for
estimating dispersion and advection below the water
table. In a tracer test, a chemical species is introduced
to the subsurface and monitored as it is carried by
groundwater through the subsurface in the direction of
flow. Average seepage velocity can be determined by
analyzing the resulting curves in time and space, and
hydraulic conductivity (K) values can be computed from
Darcy’s Law (Eq. 1). One of the most successful field sites
for tracer studies has been the Borden landfill site in
Canada, which resulted in a series of classic papers on
dispersion and adsorption processes measured during a
multiyear field experiment (9,10).
Other methods of determining hydraulic conductivity
exist. Permeameters may be used in a laboratory setting
to determine K, whereas slug tests and pump tests are
recommended for field sites. A slug test involves suddenly
adding or removing a volume of water (called a slug) from
a well and monitoring the time it takes for the water
level to return to normal. This test is convenient, but
not as accurate as a tracer test to determine the speed
with which a contaminant moves through an aquifer. The
pump test involves the constant removal of water from a
single well and the subsequent observation of drawdown
in adjacent wells. This test gives the K value for a large
part of an aquifer, but may produce a large volume of water
that requires proper handling and treatment (if performed
in a contaminated area) (3).
Step function, no dispersion
1.0
Effects of dispersion
C/C0
vx 0.5
Center of mass
L 0.0
t
t = L/vx
No dispersion, only advection
With dispersion
C/C0
L
t
Center of mass
t = L/vx
 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER
ABIOTIC SOLUTE TRANSPORT
Abiotic processes affect contaminant transport by causing
interactions between the contaminant and the aquifer
media or by changing the form of the contaminant
that may subsequently interact with the subsurface
material. The interfaces of most interest in subsurface
fate and transport are liquid/solid interfaces and gas/solid
interfaces. The processes that detail these interactions are
discussed in this section.
Sorption
Sorption includes the phenomena of adsorption, absorp-
tion, and ion exchange. Weber (11) defines adsorption as
the accumulation occurring at the interface, absorption
as the partitioning between two phases (accumulation
from groundwater into organic carbon), and sorption as
including both adsorption and absorption. Often, the terms
adsorption and sorption are interchangeable.
Many organic contaminants are removed from solution
by sorption onto the aquifer matrix. Sorption of dissolved
contaminants onto the aquifer matrix results in slowing
(retardation) of the contaminant relative to the average
groundwater seepage velocity and a reduction of dissolved
organic concentrations in groundwater. In some cases,
sorption is a reversible process; at a given solute concen-
tration (steady state), some portion of the contaminant is
partitioning out of solution onto the aquifer matrix and
some portion is desorbing from the matrix back to solution;
this process is called desorption.
The ability of chemicals to completely desorb from a
solid is an area of current study (12–15). At high con-
taminant concentration, desorption patterns are generally
predictable through equilibrium analysis. However, the
phenomena of solid phase contaminant concentrations
exceeding the predicted concentration, based on the sur-
rounding aqueous phase concentrations, is observed in
materials that have been contaminated for long time
periods and contain low solid-phase contaminant concen-
trations (nominally less than 20 mg/kg). The importance
of this observation is potentially enormous, because site
remediation requirements include the removal of sorbed
contamination. If a contaminant does not desorb as read-
ily as it sorbed, the ability to meet cleanup levels may be
seriously threatened (16).
Ion exchange is a specific category of sorption that
occurs when the absorbent charge deficiency of a surface
can be neutralized more efficiently by ions in solution
than those ions currently absorbed (17). An example of ion
exchange that is familiar is a home water softener. Home
water softeners operate by accumulating the divalent
cations from the groundwater onto the absorbent (ion
exchange resin) while releasing monovalent ions, which
reduces water hardness. In the subsurface, the resin would
be natural minerals such as those found in clays. The most
common ions occurring naturally are as follows (listed
in order of occurrence from high to low): cations, Ca2+ ,
Mg2+ , Na+ , and K+ ; anions, SO4 2− , Cl− , PO4 3− , and NO3 −
(18).
527
Cosolvation and Ionization
Cosolvation and ionization are two examples of processes
that act to increase the chemical solubility, by altering
the solvent phase or by altering the chemical form,
respectively. Cosolvation is the process by which the
mobility of one contaminant is enhanced by the presence
of a solvent (19), whereas ionization refers to the gaining
or losing of protons as a function of pH and thus going
from a neutral form to an ionic form. These processes
are solvent motivated; when a contaminant is in the
presence of a solvent, it will have decreased sorption and
an increased solubility leading to elevated concentrations
in the surrounding solute. Based on these assumptions,
it would be expected that as the fraction of cosolvent
increases in solution, the level of sorption will decrease.
These attributes of cosolvents lead to increased mobility
of originally slow dissolving hydrophobic (lacking affinity
for water) contaminants. An example of this is the affect of
ethanol on petroleum hydrocarbons. The issue concerning
an ethanol releases’ potential to cosolubilize benzene and
other aromatic compounds in groundwater (20) has been
addressed in recent groundwater modeling studies (21,22)
and in laboratory studies (23,24). The reason that
tremendous efforts are being employed to understand this
phenomenon is that benzene, which is known to increase
cancer risks, has a maximum contaminant level of only
0.005 mg/L in drinking water as set by the U.S. EPA and
ethanol is now being used as a fuel additive (an oxygenate).
Oxidation and Reduction
Oxidation and reduction (or redox) refer to the transfer
of electrons and contaminant species change of ions or
compounds. Oxidation refers to the loss of electrons and
reduction refers to the gain of electrons in a reaction.
The two processes often occur in tandem; electrons
given by one compound must be gained by another
compound. Inorganic compounds, such as reduced metals
and reduced sulfur, are possible mediators of oxidation-
reduction reactions (25–27). The presence or absence of
inorganic species is often dependent on aquifer conditions.
A common condition is that many anaerobic environments
will contain elevated levels of Fe(II) or Fe2+ , H2 S/HS− ,
and other reduced species that are capable of donating
electrons and reducing certain contaminants. For example,
the two half reaction equations for the oxidation-reduction
of ferrous iron (Fe2+ ) to ferric iron (Fe3+ ) are:
4 Fe2+ ⇔ 4 Fe3+ + 4e− [oxidation]
O2 + 4H+ + 4e− ⇔ 2H2 O[reduction]
Biologically mediated oxidation-reduction reactions are
often more rapid than abiotic reactions for many con-
taminants of interest in groundwater systems, and thus
abiotic oxidation-reduction are often neglected in trans-
port calculations. In cases where they become important,
oxidation-reduction processes are generally introduced as
a first-order reaction (rate = k1 X) with respect to contam-
inant concentration, analogous to hydrolysis.
528 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER
Hydrolysis
Hydrolysis refers to a chemical reaction between a
contaminant molecule and water. Hydrolysis may function
when direct biodegradation (microbially driven) of the
contaminant cannot, and hydrolysis may produce a
product that is susceptible to biodegradation, which can
be significant in affecting fate and transport (17). With
respect to organics, this process frequently results in the
formation of alcohols and alkenes (28).
The rate that a chemical undergoes hydrolysis is
strongly influenced by the temperature and pH of the
system (27). As temperature increases, hydrolysis rates
increase. As pH becomes either acidic or alkaline (e.g.,
greater than or less than 7.0), rates of acid-catalyzed
or base-catalyzed hydrolysis can increase. Typically,
hydrolysis is described with a first-order rate expression
as shown in Eq. 6
C in aquifers, and thus the ability of aquifers to support
= e−khyd t (6)
C0
where C is the concentration of the chemical species and
khyd is a first-order hydrolysis rate coefficient. Knowing
factors, such as pH and temperature, is important
in determining the appropriate rate coefficient for a
particular contaminant.
Volatilization
The transfer of a contaminant from the aqueous phase,
nonaqueous phase liquid (NAPL), or sorbed phase
directly to the gas phase is a process referred to as
volatilization. The rate and extent to which volatilization
occurs is strongly influenced by a number of parameters
including the contaminant phase, the contaminant’s vapor
pressure, environmental factors (e.g., temperature and
others), proximity in respect to the vadose zone, and
other site-specific parameters (29,30). As a result of the
range of factors, that play an important role in vapor
transport, which include the physical transport processes
of advection and dispersion, it is difficult to calculate the
contribution of volatilization to the fate of chemicals in
groundwater.
For the evaluation of volatilization, however, consider a
closed bottle that contains water and air. If a contaminant
is introduced to the bottle and allowed to equilibrate, some
fraction of the contaminant added will reside in the gas
phase. The distribution of a chemical between the water
and the gas at equilibrium is described by the Henry’s
law coefficient (31,32). Mathematically, Henry’s law is
presented in Eq. (7), where Pc is the partial pressure of
the contaminant, [C]aq is the aqueous phase concentration
of the contaminant, and Hc is the Henry’s constant.
Pc
Hc = (7)
[C]aq
In groundwater systems that are isolated from the
vadose zone, Henry’s law can be applied to estimate the
distribution of contaminants between the aqueous phase
and the gas phase. If direct exchange is possible with the
vadose zone, equilibrium may not be achieved and the
calculation of gas phase concentrations is more difficult.
Additionally, volatilization from NAPLs in the vadose zone
or floating on the water table can yield high concentrations
of contaminants in the gas phase (16).
BIOTIC PROCESSES AFFECTING SOLUTE TRANSPORT
Biotic processes include those biological phenomena that
affect the transport and/or fate of compounds (organics,
inorganics, or the microbes themselves) in the subsurface.
Biodegradation of organics is one process that is of partic-
ular interest to minimize human health risks associated
with the widespread contamination of chlorinated ethenes
and hydrocarbons in the environment, because this process
is able to achieve mineralization (endproducts CO2 ,H2 O,
etc.) and effectively render the contaminant innocuous
(harmless). It has only been in the last two decades that
the variety and magnitude of micro-organisms present
biodegradation, has been realized (33). Laboratory col-
umn studies (34) and improved field sampling methods
are chiefly responsible for these advances.
Energy Flow and Metabolism
All living organisms require energy to survive, and
the capture of usable energy through the process of
catabolism is a significant part of an organism’s overall
metabolism. Bacteria that obtain energy from chemical
forms are classified as chemotrophs. Chemotrophic
organisms are further differentiated as either lithotrophs,
which use inorganic sources of energy, or organotrophs,
which use organic compounds. Examples of lithotrophic
substrates include ammonia (NH3 ), hydrogen (H2 ), ferrous
iron (Fe2+ ), and sulfide (HS− ). Examples of common
groundwater contaminants that serve as organotrophic
substrates include fuel components and chlorinated
solvents (35).
Biodegradation involves the process of extracting
energy from organic chemicals via oxidation of the organic
chemicals during catabolism. As a chemical is oxidized, it
loses electrons. For this reason, chemotrophic substrates
are often referred to as primary electron donors. Electrons
lost during oxidation are coupled with the reduction of
electron acceptors. Eventually, the complex sequence of
oxidation-reduction reactions that occur during catabolism
results in the reduction of a terminal electron acceptor.
Common terminal electron acceptors in groundwater
systems include oxygen (O2 ), nitrate (NO3 − ), ferric iron
(Fe3+ ), sulfate (SO4 2− ), and carbon dioxide (CO2 ). Aerobic
catabolism occurs when oxygen is the terminal electron
acceptor, and anaerobic conditions are in place when
all other terminal electron acceptors are used. Typically,
an individual bacterial strain is capable of using only
one terminal electron acceptor. Facultative aerobes are
the exception to this rule; these organisms can use
nitrate as a terminal electron acceptor in the absence
of the preferred oxygen. An important evaluation of
contaminant biodegradation potential involves an analysis
of the availability of primary electron donors and terminal
electron acceptors, as their presence is required for
organisms to obtain energy for survival and selects the
types of organisms that will be present (16).
 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER
The metabolism of organic contaminants can be broadly
differentiated by the ability of organisms to use the
contaminant for catabolic processes. If the compound is
a primary electron donor and provides the bacterium with
energy for cell growth, the contaminant is referred to
as a primary substrate, whereas secondary substrates
are contaminants that provide the cell with energy, but
are present at concentrations that are not sufficient to
support the energy requirements of the organism. If
a compound is metabolized fortuitously as the cell is
obtaining energy from another primary electron donor,
the transformation is referred to as cometabolic (36,37).
However, cometabolism is generally a slow process and
not all chlorinated solvents are susceptible (e.g., PCE and
carbon tetrachloride). In the case of chlorinated ethenes,
the contaminant serves as an anaerobic terminal electron
acceptor in a process referred to as dehalorespiration,
which is faster than cometabolism.
Metabolic Pathways of Common Contaminants
An overview of how bacteria are capable of metabolizing
two of the primary groundwater contaminants (hydrocar-
bons and chlorinated solvents) is presented in this section.
It is important to note that this continues to be an area
of extensive research. Theoretical frameworks continue
to change and it is difficult to assert definitive analysis
regarding the totality of metabolic processes that may
impact an individual contaminant’s fate.
As a result of accidental spills and leaking under-
ground storage tanks, hydrocarbon compounds consisting
of gasoline, diesel, and jet fuels are common groundwater
contaminants. Of particular concern are benzene, toluene,
ethylbenzene, and xylene isomers (BTEX), which, under
aerobic conditions, are rapidly biodegraded as primary
substrates (38). Unfortunately, the oxygen demand result-
ing from the aerobic oxidation of these compounds can
exceed the solubility of oxygen in water. Dissolved oxy-
gen concentrations are rapidly depleted because of the
biodegradation of these and other fuel constituents, yield-
ing anaerobic conditions in contaminated aquifers (39).
Only recently have laboratory studies confirmed that
the biodegradation of all BTEX compounds occurs under
all anaerobic electron acceptor conditions (listed above).
From these studies, it appears that toluene is the most
degradable under all electron acceptor conditions and
benzene is the least degradable. Certain modeling studies
of certain well-characterized contaminated field sites infer
that anaerobic benzene degradation may occur in situ.
Only several recent laboratory studies conducted under
anaerobic conditions have demonstrated that benzene
may degrade in the absence of oxygen (40). Despite the
possibility of anaerobic degradation of BTEX compounds,
it is still generally slower than rates that are observed
through aerobic processes.
Chlorinated methanes, ethanes, and ethenes comprise
a group of compounds commonly referred to as chlori-
nated solvents. Chlorinated ethenes, particularly tetra-
chloroethene (PCE) and 1,1,2-trichloroethene (TCE), are
commonly used as degreasers and cleaning solvents in the
dry cleaning, aviation, and computer industries. Through
529
accidental spills, improper storage, and disposal tech-
niques, these chemicals have become major subsurface
contaminants (41). The extensive impacts observed are
primarily because of their widespread usage by industry,
their general mobility and persistence in the environment,
and the health risks associated with compound exposure.
The metabolism of chlorinated solvents is perhaps
more diverse than any other group of environmental
contaminants. Depending on the compound of inter-
est, the electron acceptor condition, and the presence
of inducing substrates, the metabolism of chlorinated
solvents may occur through primary metabolism, sec-
ondary metabolism, cometabolism, or through termi-
nal electron acceptor processes (16). Laboratory stud-
ies (34,36,37,42–44) have shown that, in the presence
of a suitable substrate, certain microbes are capable of
anaerobically metabolizing PCE by using it as an electron
acceptor. In this process, known as reductive dechlorina-
tion, PCE may be dechlorinated to 1,1,2-trichloroethene
(TCE), to typically cis-1,2 dichloroethene (DCE), to vinyl
chloride, and eventually to ethene and ethane, an innocu-
ous compound (45).
SUMMARY OF SOLUTE TRANSPORT PHENOMENA
Fate and transport of solutes in subsurface soils and
groundwater is impacted by a combination of many
processes that are difficult to quantify. No two field sites or
aquifers have the exact same characteristics. Soil type, pH,
temperature, groundwater flow velocity, the availability of
electron donors for microbes, and the contaminant itself
(among other parameters) all influence the path and
lifetime of the pollution source. However, measurement
techniques are improving, and understanding these
processes is an area of on going research.
In summary, nonreactive (conservative) chemicals tend
to move unimpeded through the subsurface and are sub-
ject to only hydrodynamic processes such as advection,
dispersion, and diffusion. On the other hand, many actual
contaminants (organic solvent and fuels, and inorganic
metals) can be affected during groundwater transport by a
number of subsurface mechanisms and chemical reactions.
The processes that have the largest impact on contaminant
fate and transport include advection, dispersion, sorption,
cosolvation, volatilization, biological oxidation-reduction
reactions, and cometabolism/mineralization. These pro-
cesses are reviewed in Table 1, and Fig. 2 provides an
illustration of these processes acting on a typical contami-
nant spill.
BIBLIOGRAPHY
1. Wiedemeier, T.H., Rifai, H.S., et al. (1999). Natural Attenu-
ation of Fuels and Chlorinated Solvents in the Subsurface.
John Wiley & Sons, New York.
2. Wierenga, P.J. (1995). Water and solute transport and storage,
Handbook of Vadose Zone Characterization and Monitoring.
CRC Press, Boca Raton, FL, pp. 41–60.
3. Domenico, P.A. and Schwartz, F.W. (1998). Physical and
Chemical Hydrogeology. 2nd Edn. John Wiley & Sons, New
York.
530 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER

Table 1. Subsurface Processes Affecting Solute Transport

Category Process Contaminant (Solute) Behavior

Hydrodynamic Solute Transport Advection Plug flow; movement with bulk fluid
Diffusion Movement from high to low concentration
Dispersion Spreads out and dilutes
Abiotic Solute Transport Sorption Decreased solubility
Ion exchange Neutralizes the charge and sorbs
Cosolvation Increased solubility by solvent
Oxidation-Reduction Transfer of electrons
Hydrolysis More susceptible to biodegradation
Volatilization Phase transfer from aqueous or sorbed to gas
Biotic Transport Influences Cometabolism Metabolized as cell obtains energy
Mineralization Broken down to benign products
Oxidation-Reduction Microbial driven transfer of electrons

Surface spill of gasoline 11. Weber, W.J., Jr. (1972). Physiochemical Process for Water
Volatilization Quality Control. John Wiley & Sons, New York.
12. Fu, G., Kan, A.T., and Tomson, M.B. (1994). Environ. Toxicol.
Chem. 13(10): 1559–1567.
Diffusion
13. Kan, A.T., Fu, G., and Tomson, M.B. (1994). Environ. Sci.
Unsaturated Technol. 28: 859–867.
(Vadose) zone Dispersion
14. Kan, A.T., Fu, G., Hunter, M.A., and Tomson, M.B. (1997).
Environ. Sci. Technol. 31: 2176–2185.
15. Pignatello, J.P. and Xing, B. (1995). Environ. Sci. Technol.
Sorbed contaminant on soil 10: 1–11.
Volatilization 16. Hughes, J.B. (1999). Ground Water Contamination, Trans-
port and Remediation. Prentice-Hall, Upper Saddle River,
Capillary Fiinge NJ, pp. 203–240.
Concentrated LNAPL 17. Knox, R.C., Sabatini, D.A., and Canter, L.W. (1993). Subsur-
Dissolved plume face Transport and Fate Processes. Lewis Publishers, Boca
Water table
Raton, FL.
Diffusion Oxidation-reduction reactions
18. Mitchell, J.K. (1976). Fundamentals of Soil Behavior. John
Ground water flow Advection Wiley & Sons, New York.
Dispersion
Desorption 19. Gorelick, S.M., Freeze, R.A., Donohue, D., and Keely, J.F.
Hydrolysis +
biode gradation (1993). Groundwater Contamination: Optimal Capture and
Saturated zone Sorption Containment. Lewis Publishers, Boca Raton, FL.
20. Rice, D., Powers, S.E., and Alvarez, P.J.J. (1999). Potential
scenarios for ethanol-containing gasoline released into
Clay layer Iron exchange surface and subsurface waters. In: Health and Environmental
Assessment of the Use of Ethanol as a Fuel Oxygenate. Report
Figure 2. A typical scenario showing the processes impacting a to the California Environmental Policy Council in Response
surface gasoline spill, and its transport through the vadose zone to Executive Order D-5-99. Vol. 4.
to the groundwater. 21. Molsen, J.W., Barker, J.F., Frind, E.O., and Schirmer, M.
(2002). Water. Res. Research 38(1).
22. Powers, S.E. et al. (2001). Crit. Rev. Environ. Sci. Technol.
31: 79–123.
4. Freeze, R.A. and Cherry, J.A. (1979). Groundwater. Prentice- 23. Heermann, S.E. and Powers, S.E. (1998). J. Contam. Hydrol.
Hall, Englewood Cliffs, NJ. 34: 315–341.
5. Charbeneau, R.J. (2000). Groundwater Hydraulic and Pollu- 24. Rixey, W.G. and He, X. (2001). Dissolution characteristics
tant Transport. Prentice-Hall, Upper Saddle River, NJ. of ethanol from NAPL sources and the impact on BTEX
6. Bedient, P.B., Rifai, H.S., and Newell, C.J. (1999). Ground groundwater concentrations. In: Proceedings of the Conference
Water Contamination, Transport and Remediation. Prentice- on Petroleum Hydrocarbons and Organic Chemicals in
Hall, Upper Saddle River, NJ. Groundwater. November 14–16, Houston, TX, Ground Water
7. Bear, J. (1979). Hydraulics of Ground Water. McGraw-Hill, Publishing Company, Dublin, OH, pp. 41–52.
New York. 25. Kleeka and Gonsior. (1984). S.J. Chemosphere 13: 391–402.
8. Fetter, C.W. (1993). Contaminant Hydrogeology. Macmillian 26. Macalady, D.L. and Wolfe, N.L. (1984). Abiotic hydrolysis
Publishing Company, New York. of Sorbed Pesticides. ASC Symposium Series, No. 259, pp.
9. Mackay, D.M., Freyberg, D.L., Roberts, P.V., and Cherry, J.A. 221–244.
(1986). Water Resources Res. 22(13): 2017–2029. 27. Schwartzenbach, R.P., Gschwend, P.M., and Imboden, D.M.
10. Roberts, P.V., Goltz, M.N., and Mackay, D.M. (1986). Water (1993). Environmental Organic Chemistry. John Wiley &
Resources Res. 22(13): 2047–2058. Sons, New York.

28. Johnson, R.L., Palmer, C.D., and Fish, W. (1989). Transport
and Fate of Contaminants in the Subsurface. U.S. Environ-
mental Protection Agency, Ada, OK, EPA/625/4-89-019.
29. Mackay, D. and Leinonen, P.J. (1975). Environ. Sci. Technol.
9: 1178–1180.
30. Mackay, D. and Wolkoff, A.W. (1973). Environ. Sci. Technol.
7: 611–614.
31. Mackay, D., Shiu, W.Y., and Sutherland, R.P. (1979). Envi-
ron. Sci. Technol. 13: 333–337.
32. Thibodeaux, L.T. (1996). Environmental Chemodynamics:
Movement of Chemicals in Air, Water, and Soil. John Wiley
& Sons, New York.
33. Young, L.Y. and Cerniglia, C.E. (1995). Microbial Transfor-
mation and Degradation of Toxic Organic Chemicals. John
Wiley & Sons, New York.
34. Maymo-Gatell, X., Tandoi, V., Gossett, J.M., and Zinder, S.H.
(1995). Appl. Environ. Microbiol. 61: 3928–3933.
35. Rittman, B.E. and McCarty, P.L. (2001). Environmental
Biotechnology: Principles and Applications. McGraw-Hill,
New York.
36. Hollinger, C., Schraa, G., Stams, A.J.M., and Zehnder, A.B.
(1993). Appl. Environ. Microbiol. 59(9): 2991–2997.
37. Carr, C.S. and Hughes, J.B. (1998). Environ. Sci. Technol. 32:
1817–1824.
38. Alvarez, P.J.J. and Vogel, T. (1991). Appl. Environ. Microbiol.
57: 2981–2985.
39. Lee, M.D. et al. (1988). CRC Cri. Rev. Environ. Control. 18:
29–89.
40. Lovely, D.R., Coates, J.D., Woodward, J.C., and Phillips,
E.J.P. (1995). Appl. Environ. Microbiol. 61: 953–958.
41. Pankow, J.F. and Cherry, J.A. (1996). Dense Chlorinated
Solvents and Other DNAPLs in Groundwater. Waterloo Press,
Portland, OR.
42. DiStefano, T.D., Gossett, J.M., and Zinder, S.H. (1992). Appl.
Environ. Microbiol. 58(11): 3622–3629.
43. Smatlak, C.R., Gossett, J.M., and Zinder, S.H. (1996). Envi-
ron. Sci. Technol. 30(9): 2850–2858.
44. Yang, Y. and McCarty, P.L. (2000). Environ. Sci. Technol. 34:
2979–2984.
45. Freedman, D.L. and Gossett, J.M. (1989). Appl. Environ.
Microbiol. 55: 2144–2151.
WATER IN THE UNSATURATED ZONE
MUKAND SINGH BABEL
Asian Institute of Technology
Parthumthani, Thailand
The soil profile immediately below the land surface and
above the water table is called the unsaturated zone,
where the pores contain both water and air. Figure 1
shows a schematic diagram of the unsaturated zone. This
zone differs from the saturated zone, where the pores are
saturated with water. The unsaturated zone is sometimes
called the vadose zone.
The thickness of an unsaturated zone can range from
zero (e.g., when a lake or marsh is at the surface)
to hundreds of meters, as is common in arid regions.
It is a zone of natural and human-induced activity. It
experiences physical phenomena such as thermodynamic
WATER IN THE UNSATURATED ZONE 531
Land surface Well
Unsaturated
zone
Capilary fringe
Saturated Water
zone table
Figure 1. A schematic diagram of the unsaturated zone.
interactions, transport processes of various kinds, and
chemical reactions involving both natural and artificial
substances. Several biological activities involving plant
roots, rodents, worms, microbiota, and other organisms
also take place in this zone. The zone is used for cultivating
plants, constructing structures, and disposal of waste.
To a large degree, the unsaturated zone controls the
transmission of water to aquifers (water bearing strata),
as well as to the land surface, to water on the surface,
and to the atmosphere. It also acts as a filter as it
removes undesirable substances before they reach and
affect aquifers. Figure 2 gives a pictorial representation of
the unsaturated zone.
Though the unsaturated zone stores water, plant
nutrients, and other substances, it is not always considered
a major storage component of the hydrologic cycle because
it holds only a tiny fraction of the earth’s fresh water and
this water is usually difficult to extract. But it is of great
importance for storing water and nutrients in ways that
are vital to the biosphere.
SOIL WATER
Soil water is normally considered to include both the water
contained in the soil profile and the subsurface water in
the unsaturated subsoil layers. Soil water may be present
as gravity water in transit in larger pores, as capillary
water in smaller pores, as hygroscopic water adhering
Land
Unsaturated zone surface
Water table Surface water
Saturated zone
Groundwater
Figure 2. Unsaturated zone.
532 WATER IN THE UNSATURATED ZONE
in a thin film to soil grains, and as water vapor. After
a rain, gravitational water moves downward in larger
pores, but it is dispersed into capillary pores or passes
through the unsaturated zone to the groundwater or to
a stream channel. Hygroscopic water is held very tightly
to the soil surface by molecular attraction and is not
normally removed from soil. Capillary water, therefore, is
the important variable element of soil water.
SOIL WATER POTENTIAL
The most basic measure of the water in an unsaturated
soil is water content or wetness, defined as the volume
of water per bulk volume of the soil. Water is held in
unsaturated soil by forces whose effect is expressed in
terms of the energy state or pressure of the water. Like
all matter, water can possess both kinetic and potential
energy. However, in general, the flow of water in soils
is too small to consider kinetic energy. Therefore, the
energy state of soil water is represented by its equivalent
potential energy. Potential energy is defined as the ability
to do work. Work is done whenever an object is accelerated
through a distance or heat is produced or transferred. Soil
water potential thus measures the ability of soil water to
move and/or produce heat. Water potential is important to
any process in which soil water moves, such as infiltration
and redistribution within the soil or removal of water from
the soil by evaporation or plant uptake.
To define the soil water potential parameter, a reference
body of water is defined, and the soil water potential of
any other body of water is defined as the difference in
potential energy concentration (potential energy per unit
quantity of water) between the soil water and the reference
body of water. The reference body of water is usually
called the reference state, which is defined as ‘‘pure (no
dissolved solutes), free (free of adsorptive forces) water at
a specified temperature and specified elevation exposed to
atmospheric pressure.’’ The water potential of water in the
reference state is zero.
Total soil water potential may formally be defined
as the amount of work per unit quantity of pure water
that must be done by external forces to transfer an
infinitesimal amount of water reversibly and isothermally
from the reference state to the soil at the point under
consideration (1).
FACTORS AFFECTING WATER POTENTIAL
A change in the potential energy of the soil water brings
about a change in the soil water potential. There are
several factors that affect the potential energy of soil
water, including adsorption of water onto soil particles,
solutes dissolved in the soil water, the elevation of the
soil water in the earth’s gravitational field, and applied
pressure (both positive and negative).
WATER POTENTIAL COMPONENTS
The total soil water potential can be broken into several
components. A complete list of soil water potential
components consists of gravitational potential, osmotic
potential (also sometimes called solute potential), matric
potential, vapor potential, hydrostatic pressure potential,
and overburden pressure potential. The major potential
components of interest in the unsaturated zone are
gravitational potential, osmotic potential, and matric
potential that give rise to the total potential as follows:
t =  g +  o +  p + . . . (1)
where t is the total potential, g is the gravitational
potential, o is the osmotic potential, and p is the
matric potential.
Gravitational potential is defined as the energy of water
per unit volume that is required to move a specific amount
of pure, free water from an arbitrary reference point to the
soil-water elevation. It is independent of soil properties
and solely dependent on the vertical distance between the
arbitrary reference point and the interest point. Often, the
soil surface is chosen as the point of reference, whereby
the distance z (m) to each point of interest is negative by
convention. Gravitational potential is given as
g = ρl gz (2)
where ρl is the density of the soil water, g is the
acceleration due to gravity (m/s2 ) and g is the
gravitational potential (J/m3 ).
The osmotic potential component arises due to the salts
in the soil water. It may be thought of more a suction than
a pressure. In a dynamic soil–water–plant system, in the
presence of a selective permeable membrane as found in
the roots of plants, the pressure or energy potential (a
suction) on the solution side is less than that of pure,
free water. Water will pass through the membrane to the
solution side until equilibrium is reached. The osmotic
potential, o , is given by

o = − (3)
ρw
where  is the osmotic pressure and ρw is the density
of water.
Soil is porous, so it acts as a semipermeable barrier
to salt solutions. An individual particle has an electric
field that causes a decrease in the concentration of the
soil solution near the particle surface, which results
in an increased cross-sectional area for water transport
compared to the area available for transport of dissolved
salts, causing water to move due to the osmotic difference.
The amount of flow depends on the particle size of the
soil and the gradient of the osmotic pressure. In coarse,
saturated sand, soil pores are relatively larger than those
in clays, and consequently, the thickness of fluid layer with
decreased salt concentration is small compared to the total
area available for transport. Thus, the semipermeability
is small, and the transport of soil solution due to osmotic
potential is negligible. In clays, with many more pores,
the solutes in the soil solution can be completely repelled
from the soil and restricted from flow. In such soils, water
movement in response to an osmotic pressure gradient is
significant.

The matric potential component results from the soil
matrix. Water molecules can form hydrogen bonds with
the surface of soil minerals that give rise to adsorption
and with other water molecules that result in cohesion.
In unsaturated soil, these forces of attraction between the
soil mineral surfaces and the soil water exert a ‘‘pull’’ on
the soil water reducing the free energy of the water. It is
usual to deal with the combined effect of both capillary
and adsorptive forces in the way in which water is held
in the soil matrix. The matric potential is the potential of
the water in a pore of the medium relative to the potential
of the air that results primarily from the adsorptive and
capillary forces of the soil matrix. When a medium is
unsaturated, the water generally is at lower pressure
than the air, so the matric potential is negative.
Moisture content increases as the matric potential
increases. Zero matric potential is associated with high
(saturated or nearly saturated) water content. As the
matric potential decreases, the water content decreases.
Figure 3 shows the relation between matric potential
and water content, called a retention curve. It is
a characteristic of a porous medium that depends
on the nature of its pores. This relation strongly
influences the movement of water and other substances in
unsaturated media.
MEASUREMENT OF SOIL WATER
Soil water can be measured as either soil water content
or soil water potential; conversion from one to the other
uses the moisture retention curve. Their combined use
provides a powerful approach to the study of soil fluxes in
unsaturated soil.
The most commonly used method for measuring the
water content of soil directly is the gravimetric method,
which involves taking a number of soil samples of known
volume and determining their weight loss when dried
in an oven at 105 ◦ C. The neutron probe is the most
Moisture retention curves
Clayey
soil
Moisture content
Sandy
soil
Matric potential (−cm water)
Figure 3. Moisture retention curves.
GROUNDWATER AND VADOSE ZONE HYDROLOGY 533
widely used indirect way of measuring soil water. In this
method, a source of high-energy neutrons is used that
is lowered into an aluminum access tube in the soil.
Fast neutrons lose energy through collision with atoms
of low atomic weight, such as the hydrogen in water,
and are converted to slow neutrons. These slow neutrons
are counted with a high count indicating high moisture
content. Time-domain reflectometry (TDR) determines soil
water content by measuring the dielectric constant of the
soil. TDR has two significant advantages over the neutron
probe method; it is nonradioactive and can be set up for an
automatic, nonmanual operation. The satellite-mounted
synthetic aperture radar (SAR) method has the potential
for measuring soil moisture content on a drainage basin
scale. It determines soil water content by monitoring
changes in soil dielectric properties.
Tensiometers are the oldest and most widely used
technique for measuring matric potential. They comprise a
liquid-filled porous cup connected to a pressure-measuring
device such as a mercury manometer or a pressure
transducer. The cup is embedded in the soil, and water can
flow between the soil and cup until the pressure potential
inside the cup equalizes with that of the soil water. In fine-
textured soils, the resistance block method may be useful.
In this method, two electrodes are embedded in a porous
block, which is buried in the soil. The matric potential
in the block comes into equilibrium with that of the soil
water, and the resistance across the electrodes varies
with the resulting water content of the block. Gypsum
resistance blocks can be used for measuring potentials as
low as −15,000 cm (1500 kPa).
BIBLIOGRAPHY
1. Tindall, J.A. and Kunkel, J.R. (1999). Unsaturated Zone
Hydrology for Scientists and Engineers. Prentice-Hall, Engle-
wood, Cliffs, NJ.
GROUNDWATER AND VADOSE ZONE
HYDROLOGY
UWE DANNWOLF
URS Australia Pty Ltd.
Turner, Australia
The vadose zone is the geologic media between the land
surface and the regional water table.
The vadose zone
• is the foundation for most man-made structures;
• provides nutrients and water to all plant and
animal life;
• protects drinking water reservoirs (aquifers);
• governs how much groundwater is replenished;
• provides the pathway for contaminants to penetrate
into groundwater, and
• is one of the least understood media.
534 GROUNDWATER AND VADOSE ZONE HYDROLOGY
The latter point becomes evident when a definition for
the vadose zone is needed. All publications include the
unsaturated soil zone, but some publications include the
capillary fringe; others do not. The term ‘‘vadose’’ is
derived from the Latin noun vadosus, which translates
to ‘‘shallow’’.
The vadose zone includes the weathered soil horizon
and the unconsolidated sediments and rocks below. The
capillary fringe separates the water saturated zone from
the unsaturated zone. The water saturated zone is often
referred to as an aquifer if it allows significant amounts
of water to be transmitted. Several processes take place in
the vadose zone.
PROCESSES IN THE VADOSE ZONE
The processes, which influence the vadose zone and the
recharge to groundwater include
• precipitation,
• evapotranspiration,
• surface water runoff, and
• infiltration and groundwater recharge.
Precipitation
Precipitation is the water that falls on the land surface
as rain, snow, and other forms such as dew, frost, and
irrigation water. It is measured as volume per time
period (e.g., mm/a) or as volume per area and time
(e.g., L/km2 s), which is a relatively simple parameter
to measure and is often available from meteorological
stations. However, care should be exercised when using
this parameter for climatic water balancing because there
is great variability even between neighboring stations, due
to the precipitation and wind characteristics as well as the
various measuring devices.
Evapotranspiration
Evapotranspiration, the reverse of precipitation, repre-
sents the transport of water from the earth back to the
atmosphere through evaporation and plant transpiration.
This process depends on the microclimatic conditions at
a given site. This parameter depends on many factors
(air temperature and humidity, solar irradiation, vegeta-
tion, wind speed, etc.). It is more difficult to measure and
therefore not available from all meteorological stations.
Simple (1) and complex calculations (2–5) can estimate its
range. Its units are generally reported as millimeter per
time period. The vegetation type and cover greatly influ-
ence this parameter and are often neglected in standard
formulas, which can lead to significant variations.
Surface Water Runoff
Surface water runoff is the amount of precipitation that
does not immediately enter the subsurface and contributes
to surface water discharge (e.g., into local creeks). A
portion of this surface water runoff appears on the surface
again after a short underground passage (interflow). The
amount of surface water runoff depends on the size of the
catchment, vegetative cover, relief [(slope or gradient) and
depressions (pond formation)], soil cover and soil type,
initial soil moisture content, and rainfall intensity. For
instance, if the supply rate of water to the soil surface is
greater than the soil’s ability to allow the water to enter,
excess water either accumulates on the soil surface or
becomes runoff. Steep slopes cause greater runoff rates,
allowing the water a shorter time to infiltrate. In contrast,
gentle slopes have less impact on infiltration due to
decreased runoff.
Surface water runoff is even more difficult to measure
than any of the above parameters. Hydrological models
aid in estimating it.
Effective Recharge or Infiltration
Both evapotranspiration and surface water runoff limit the
amount of total precipitation that enters the subsurface,
the effective recharge or infiltration into the vadose zone.
The term infiltrability is generally used in soil physics and
hydrology to define the maximum rate at which rain or
irrigation water can be transmitted through a soil under
a given condition.
As a general rule of thumb, a total of four water relevant
zones exist in the vadose zone (after Reference 6):
1. a near-surface zone to shallow (<1 m) depths
affected by episodic rain;
2. a zone below which shows signs of seasonal water
infiltration;
3. a zone of virtually constant moisture content, in
which the hydraulic gradient is unity (i.e., 1), and
there is a predominantly downward flow toward the
water table; and
4. a capillary zone above the water table, in which the
water moves in the groundwater flow direction.
This simplified zoning is complicated by the fact
that irregularities (heterogeneities) in the soil profile
such as root channels, earthworm burrows, soil fissures
or cracks, spatially varying sediment size therefore
exhibiting variable hydraulic and physical characteristics
Coarse sand Silty sand Clay
Silty
sand
Coarse
sand
Clay h
Coarse
sand
Silty Capillary
sand fringe
Water table 0 1.0
Sw
Figure 1. Saturated water content (Sw ) above the water table
for a layered sequence. Dashed lines represent water content for
homogeneous sediments (from Reference 7, altered).

(see Fig. 1), and distinct layering allow preferential (i.e.,
faster) flow of water in selected areas. For instance, clay
layers generally impede flow due to their lower saturated
hydraulic conductivity, whereas sand layers have a lower
tendency to retard the movement of infiltrating water.
This process, which is also referred to as ‘‘fingering,’’ can
occur in less than a few percent and to up to 70% of the
soil profile. Such layers generally retard water movement
during infiltration. For the remainder of the soil profile,
the entering moisture results in a semicontinuous wetting
front, a zone of increased moisture content and a zone
below of significantly lower moisture content. Beyond the
wetting front, there is no visible penetration of water.
Infiltrating precipitation can appear as
1. retarded interflow discharging as surface runoff
after a brief underground passage,
2. increased soil moisture, and
3. effective recharge.
The effective recharge can be measured by using
lysimeters, but this requires a substantial effort. It
is furthermore dynamic, changing constantly over time
and space.
The measuring units are volume or m per time period.
Infiltrating water also changes the moisture content of
the soil. If rain infiltrates a previously dry soil (typical
moisture content of less than 10% saturation), then the
rainwater is first absorbed into the soil pores, increasing
the moisture content. The soil gas in the soil profile
needs to be replaced, and as a result the infiltration
velocity is reduced initially. Later in the process when
the soil gas dissolves gradually in the infiltrating water,
the infiltration rate increases again before a renewed
reduction occurs. This reduction is from increased solid
particles washed into the soil profile from the rain and the
growth of organisms near the soil surface. Additionally,
the increase in soil moisture leads to a reduction in
capillary pressure.
Snow and frost also impact the infiltration rate. In
frost periods, soils of low initial water content may become
porous, and the infiltration rate is increased. Soils of
high initial water content may freeze and become nearly
impermeable.
In detail, infiltration depends on material properties,
such as
• grain size,
• porosity,
• capillary forces and soil suction,
• permeability, and
• fluid densities and viscosities.
MATERIAL DEPENDENCIES
Grain Size and Porosity
Unconsolidated materials are distinguished by their
grain sizes. Various classifications exist, the grain sizes
commonly decrease in the order block, cobble, gravel, sand,
GROUNDWATER AND VADOSE ZONE HYDROLOGY 535
silt, clay, colloid, and ion. Clays are smaller than 2 µm,
whereas colloids consist of particles less than 0.1 µm
in size. On the other hand, gravels start at 2 mm. The
distribution of soil grain size and bulk densities (often
referred to as ρ) of soil and rock are commonly measured
during geotechnical investigations.
With regard to water and contaminant transport, a
critical parameter of vadose zone materials is the type
and distribution of void spaces (pores). Loose materials
contain widely distributed void spaces that typically are
interconnected. These pores are usually individually very
small, but numerous, and the total volume of pores usually
varies from 30–50% of the total volume. However, not all
of these pores are available for water circulation. A clay
may exhibit a total porosity of >50%, however, only less
than a few percent are available for water circulation. In
contrast, in gravel sediments, almost all of the pore space
(roughly 20% of the total volume) allows water transport.
As can be seen from Fig. 2, classifications exist for pore
spaces. The largest pores are obviously present in the
coarsest materials. Pores of the order of 60–1000 µm are
present in sandy sediments. Small and smallest pores are
present in clay grain size and below.
Also depicted in Fig. 2b is the approximate rate of water
movement in meters per year. The coarser the sediments,
the higher the rate of vertical water movement. It is,
however, important to note that even in silts, the vertical
water movement under unit gradient can be as high as
2 m per year.
Capillary Rise and Capillary Pressure
When water evaporates near the surface, it generates
a low hydraulic potential, which is equilibrated by the
capillary rise at the groundwater table. The rate at
which this supply is determined is governed by the
permeability (and size, respectively) of the sediments.
Consequently, the height of the capillary fringe depends
on the sediment’s permeability, as Fig. 2a outlines. The
water in the capillary fringe is held by capillary forces
against gravity induced flow.
The capillary pressure (Cp , also named matrix or soil
suction) for sands and silts can be estimated roughly from
Cp (kPa) = 0.3/grain size (mm) (1)
The higher the soil suction, the higher the residual
saturation of a sediment.
Soil–Water Characteristic Curve
The soil–water characteristic curve (also referred to as
the water retention curve) for a soil is defined as the
relationship between the water content and the capillary
pressure. This curve is characteristic for each soil, it
governs the rate at which water is taken up or released by
the soil.
As can be seen from Fig. 3, when the water content
decreases (i.e., the soil suction increases), the permeability
of a sediment (here, a silty soil) decreases. The order
of decrease depends on pore sizes. Sands have a much
slighter decrease in permeability. Tight soils stay wet
536 GROUNDWATER AND VADOSE ZONE HYDROLOGY

(a) 100000 1000

Thickness of cap. fringe, m

10000
100
Saturated horizontal

1000
permeability, m/d

100
10
10
Thickness of capillary fringe, m
1 1
0.1 Saturated horizontal permeability,
0.1 m/d
0.01
0.001
0.01
0.0001
0.00001 0.001
10000 1000 100 10 1 0.1 0.01 0.001 0.0001 0.00001 0.000001 1E-07
Vertical water movement, m/year

12 10000
(b)
10 1000

Capillary pressure, kPa

100
8
10 Capillary pressure, kPa
6
1 Water movement, m/year
4
0.1

2 0.01

0 0.001
10000 1000 100 10 1 0.1 0.01 0.001 0.0001 0.00001 0.000001 1E-07 Grain size, mm

Block Cobble Gravel Sand Silt Clay Colloid Ion

Largest pores Large pores Medium Small pores Smallest pores

60–1000 µm <60 µm <10 µm

Figure 2. (a) Generalized diagram depicting the thickness of the capillary fringe and saturated
horizontal permeability. (b) Generalized diagram depicting pore sizes, capillary pressure, and
the vertical annual water movement rate of sediments.

for much longer, however, suffer much more from 1. the drainage of an individual pore, which is
evaporative losses. controlled by the smallest radius. When the
Figure 3 also introduces terms often used by soil meniscus retreats under increasing tension, the
scientists to describe the wetting and drying behavior pore’s radius is decreasing. Rewetting (imbibition),
of a soil. The term field capacity is used when the water on the other hand, is controlled by the pore’s
content of a soil has reached a maximum (often referred to largest radius.
as sat). Any further increase in soil moisture results in a 2. contact angle hysteresis. This is a well-recognized
water release (i.e., gravity induced flow). When the water phenomenon, but it is only significant at relatively
content nearly equals its minimum, then the wilting point high flow rates. The slower the process, the less
is reached. By definition, the wilting point equals a soil contact angle hysteresis. Because drainage and
suction between 500 and 6000 kPa. A soil is referred to as rewetting curves are generally measured at slow
‘‘air dry’’ at or above a capillary pressure of 70,000 kPa. (‘‘quasi-equilibrium’’) rates, this is not likely to be an
Furthermore, two terms are defined through this water important contributor.
retention curve: the air entry value and the residual water 3. Air entrapment effects. During imbibition, air can
content. The air entry value is the meeting point of the become trapped inside the larger soil pores. Given
two tangents shown in Fig. 3. The residual water content sufficient time, the entrapped air bubbles slowly
(referred to as res) is defined similarly by using the dissolve in the water and cause hysteresis
lower tangent. Figure 3 further introduces a fundamental
feature soils display when drying out or becoming wet (the Determining the Soil–Water Characteristic Curve. The
hysteresis effect). most commonly used direct way to measure the soil–water
Hysteresis describes the effect of retardation of forces potential is with a tensiometer. A body of pure water is
acting on a body to change (as if from viscosity or hydraulically connected to the soil water via a permeable
internal friction in magnetic induction and other physical ceramic cup. If the soil water is under tension, then a
phenomena). In soil science, when a soil is dried and then small fraction of water in the tensiometer is pulled out,
rewetted, the drying and wetting soil–water characteristic lowering the pressure inside. This negative pressure can
curves are not identical. The wetting curve is drier than the then be measured by a gauge or pressure transducer.
drying curve over a wide range of soil–water potentials. Alternatively, mathematical models can help in deter-
This hysteresis is caused by: mining the soil–water characteristics of a soil.
 GROUNDWATER AND VADOSE ZONE HYDROLOGY 537
Coefficient of permeability,

1.E-04
1.E-05
1.E-06
k, m/s

1.E-07
1.E-08
1.E-09
1.E-10
1 10 100 1000 10000 100000 1000000 10000000
Soil suction, kPa

Air-Entry Value

50
Drying curve
40 Wetting curve
Water content, %

Field capacity
Wa

Wilting point
te

Air dry
30
r re
ten

20
tio
n
cu

Residual water
rve

10
content
0
1 10 100 1000 10000 100000 1000000 10000000
Soil suction, kPa
Figure 3. Visualization of the relationship between the permeability function and the soil water
characteristic curve (from Reference 8, altered).

Hydraulic Conductivity pumping tests from groundwater monitoring wells, slug

tests, and other indirect measurements like grain-size
The basic formula to determine hydraulic conductivity
distribution. In vadose zone studies, tensiometers are used
was first developed by Henry Darcy in 1856 when he
for determining the field capacity of a soil.
determined that the flux of water in sand columns depends
on hydraulic conductivity and the hydraulic gradient. It
can further be shown that hydraulic conductivity depends FLOW PROCESSES IN THE VADOSE ZONE
on both the water and the porous media properties
according to Two processes control water in the vadose zone. The first
kϕg is gravity, which moves water downward. The second is
K= (2)
µ a capillary process that, similar to water dripped slowly
on a sponge, moves water in all directions and stores
where µ = viscosity, and releases it. The capillary process is controlled by
φ = density the nature of the sediment and rock and the pores and
g = acceleration of gravity fractures within it. In most cases, capillary processes
k = intrinsic permeability of the soil (independent are dominant in fine-grained sediments (clay, silt, and
of fluid). the like), whereas gravity is dominant in coarse-grained
sediment and large fractures. Pores and fractures can be
The intrinsic permeability is generally a function of the completely filled with water in places within the vadose
porous media, and it is proportional to the square of the zone. These saturated portions of the vadose zone are
grain diameter, k ∼ d2 . Consequently, capillary pressure typically referred to as ‘‘perched water’’ or ‘‘capillary
depends on the soil’s conductivity: fringe,’’ depending on where they occur. At most sites,
√ the water content is changing over time throughout the
Cp ∼ 1/ K (3) system representing a dynamic system.
Once saturation is continuous, this groundwater moves
This relationship is commonly used in numerical modeling through the system from the highest potential to the
studies to scale the magnitude of the capillary pressure lowest. Numerical modeling studies show that a range of
curves in the absence of direct measurements of infiltration patterns occur, which can be related to the ratio
capillary pressure. of infiltration rate to unsaturated hydraulic conductivity.
A high value of this ratio reflects a prevailing hydraulic
Measurements of Hydraulic Conductivity. The hydraulic conductivity, which cannot readily redistribute the newly
conductivity of an aquifer with respect to water can infiltrated moisture. Moisture accumulates in the near-
be determined by a variety of field tests, including surface region before advancing down through the soil as
538 VADOSE ZONE MONITORING TECHNIQUES
a distinct wetting front. In contrast, low values of the ratio
of rainfall to unsaturated hydraulic conductivity show
minimal moisture accumulation, as the relatively small
volumes of infiltrating moisture are readily redistributed
through the soil profile.
MODELS
Accurate in situ measurement of unsaturated hydraulic
conductivity is cumbersome and time-consuming. One
alternative to direct measurement of unsaturated
hydraulic conductivity is to use theoretical methods, which
predict the conductivity from more easily measured soil
water retention data. Methods of this type are generally
based on statistical pore-size distribution models, a large
number of which have appeared in the soil science and
petroleum engineering literature during the past several
decades (see Reference 9 for a review). Implementation
of these predictive conductivity models still requires inde-
pendently measured soil–water retention data. Measured
input retention data may be given either in tabular form
or by closed-form analytical expressions, which contain
parameters that are fitted to the observed data.
Water flow in variably saturated soils is traditionally
described by the Richards equation:
 
∂θ ∂ ∂h ∂K(θ )
= K(θ ) + (4)
∂t ∂z ∂z ∂z
where h is the soil water pressure head (or fluid pressure
with dimension L), θ is the moisture content [−], t is time
(T), z is soil depth (L), and K is the hydraulic conductivity
[LT−1]. The solution of the Richards equation requires
knowledge of the unsaturated soil hydraulic functions θ (h)
and K(h) or K(θ ).
Several equations have been proposed to describe the
soil–water retention curve near saturation. Most of these
functions are mathematically too complicated to be easily
incorporated into predictive pore-size distribution models
for hydraulic conductivity. A smooth function with attrac-
tive properties is the equation of van Genuchten (10):
Seff (h) = [1 + (ah)n ]−m (5)
where a, n, and m are empirical constants affecting the
shape of the retention curve (i.e., curve fitting parameters).
Another mathematical solution is provided by Brooks and
Corey (11).
Various computer programs exist to model unsaturated
soil properties, including (but not limited to)
• Soil Vision: www.soilvision.com.
• Rosetta: U.S. Salinity Laboratory USDA; http://www.
ussl.ars.usda.gov/models/rosetta/rosetta.
• RETC: Code for Quantifying the Hydraulic Functions
of Unsaturated Soils, U.S. Salinity Laboratory
USDA; http://www.ussl.ars.usda.gov/models/retc.
• SEEP/W: http://www.geo-slope.com/products/seepw.
asp.
• Hydrus: http://www.ussl.ars.usda.gov/models/tu
torials/hydrus2d/tutorial2d.htm.
• UnSat: Waterloo Hydrogeologic Inc;. http://www.
flowpath.com.
BIBLIOGRAPHY
1. Linacre, E. (1993). A three-resistance model of crop and forest
evaporation. J. Theor. Appl. Climatology 48: 41–48.
2. Penman, H.L. (1948). Natural evaporation from open water,
bare soil and grass. Proc. R. Soc.London A193: 120–146.
3. Thorntwaite, C.W. and Mather, J.R. (1955). The water
balance, Publication in Climat. VIII, Nr.1. Drexel Inst.
Laboratory of Climatology, Centerton, NJ.
4. Turc, L. (1954). Le bilan d’eau des sols, relations entre les
precipitations, l’evaporation et l’ecoulement. Trois Journée
d’Hydraulique. Alger.
5. Haude, W. (1955). Zur Bestimmung der Verdunstung auf
möglichst einfache Weise. Mitt. dt. Wetterd. 2: 11, Bad
Kissingen.
6. Faybishenko, B. (2000). Vadose Zone Characterization and
Monitoring—Current Technologies, Applications, and Future
Developments. In: Vadosc Zone Science and Technology
Solutions. Battelle Press, Columbus, OH.
7. Looney, B.B. and Falta, R.W. (Eds.). (2000). Vadose Zone
Science and Technology Solutions. Battelle Press, Columbus,
OH, pp. 1–1465.
8. Fredlund, M.D. (1988). Unsaturated seepage modelling made
easy. Geotechnical News June.
9. Mualem, Y. (1986). Hydraulic conductivity of unsaturated
soils: Prediction and formulas. In: Methods of Soil Analysis.
A. Klute (Ed.), Part 1. Physical and Mineralogical Methods.
Agron. Monogr. 9, 2nd Edn. American Society of Agronomy,
Madison, WI, pp. 799–823.
10. van Genuchten, M.Th. (1980). A closed-form equation for
predicting the hydraulic conductivity of unsaturated soils.
Soil Sci. Soc. Am. J. 44.
11. Brooks, R.H. and Corey, A.T. (1964). Hydraulic Properties
of Porous Media. Hydrology Papers, No. 3, Colorado State
University, Fort Collins, CO.
VADOSE ZONE MONITORING TECHNIQUES
PHILIP R. RYKWALDER
University of Texas at San
Antonio
San Antonio, Texas
WILLIAM J. BLANFORD
Louisiana State University
Baton Rouge, Louisiana
INTRODUCTION
Monitoring fluids of the vadose zone has multiple
motives, including detecting contaminants, measuring
soil properties and water conditions for construction or
monitoring geologic hazards, and determinations of water
saturation that may be used for optimizing irrigation.
Because fluids must move through the vadose zone to
reach the water table, monitoring of the vadose zone allows
predicting possible sources and processes of groundwater
contamination. Geologic hazards such as landslides,

rising water tables, or land subsidence can be partially
predetermined by monitoring the vadose zone. And in
agriculture, monitoring soil water content to irrigate most
efficiently has a long history.
Monitoring the temporal and spatial variance in the
vadose zone is vitally important to engineers, geologists,
meteorologists and other environmental scientists inter-
ested in studying water and contaminants. There are many
techniques that directly or indirectly assess water con-
tent, sorbed mass, gaseous and liquid contaminants, flow
direction, and other parameters that may be monitored.
The methods discussed here are divided into two cate-
gories, direct and indirect monitoring techniques. Direct
monitoring techniques require a degree of sampling or
large-scale soil disturbances. Conversely, indirect tech-
niques are analogous to remote sensing in which no
disturbance to the soil or water is required, although a
well may be required in some techniques. This section dis-
cusses in short detail some of the attributes and shortfalls
of most of the currently used techniques for monitoring
the three phases of the vadose zone: liquid, soil, and gas.
For a more in-depth understanding, readers are directed
to the references listed after each entry.
INDIRECT MONITORING METHODS
Indirect methods for monitoring the vadose zone commonly
have links to geophysical tools, principles, and techniques.
Many vadose zone characteristics such as water content,
the presence of contamination, and flow direction, among
others may be determined without soil sampling. These
techniques may also be used to determine the nature and
extent of soil stratification for site delineation.
In practice, these methods are generally nonintrusive
and do not require disturbing the soil, but some
require well installation. When drilling or augering is
necessitated, proper techniques must be ensured so that
mud linings, drilling fluids, or other drilling by-products
do not affect hydraulic conductivity, which may then affect
measurements. It is also important to clean any grease or
oil from drilling tools to avoid contamination and to clean
tools between use.
Tension Infiltrometers
Tension infiltrometers are used to characterize hydraulic
properties of saturated and partially saturated porous
media, including saturated and unsaturated hydraulic
conductivities. Infiltrometers are commonly used to
measure hydraulic conductivity as a function of matrix
potential in partially saturated porous media, sorptivity,
steady-state volumetric flow rate (Q), and water content
(Fig. 1).
There are variations of the disk infiltrometer and
multiple methods of measuring hydraulic properties using
infiltrometers, but all consist of three basic parts: the
bubble tower, the reservoir, and a porous baseplate. The
porous base plate lies flat on a wetted soil surface,
and both the bubble tower and reservoir are connected
perpendicularly to it. The purpose of the baseplate is
to establish hydraulic contact between the soil and the
VADOSE ZONE MONITORING TECHNIQUES 539
Air inlet
Water reservoir
Bubble tower
Porous baseplate
Figure 1. Tension infiltrometer. Tension measurements may be
made using a handheld tensiometer on nongraduated models or
on the side of graduated models (pictured).
reservoir. A thin layer of moist sand may be placed under
the baseplate to ensure better hydraulic conductivity
between it and the soil. The reservoir supplies water to
the baseplate and the bubble tower; the latter consists
of a series of tubes that relieve the vacuum in the
reservoir when water infiltrates into the soil. Tension
measurements correspond to the level of the water in the
reservoir. An important point to note is that hydraulic
conductivity must exist between the baseplate and the
wet porous media for the device to operate.
The main advantage of disk infiltrometers is that
in situ measurements of vadose zone hydraulic properties
are rapid, reliable, and replicable and the devices are
inexpensive and easy to operate. There exists, however,
a limited range of tensions (generally 0 to 25 cm below
the ground surface (BGS) in which the infiltrometers
can operate, and measurements can be made only at
the soil surface unless trenches are dug to the desired
measurement level and the device used (1–3). Ideal soil
properties are also assumed (soil is uniform, homogenous,
and nonswelling), and pore closure may result from the
weight of the device if used on freshly cultivated soils.
There are other limitations that may be observed in
Hussen and Warrick (4).
Neutron Moisture Probe
Initially designed for the agricultural industry, the neu-
tron moisture probe has been adapted to environmental
monitoring to obtain soil moisture levels. It is a nonde-
structive method of obtaining soil moisture levels when
the probe is inserted in a well. The only restriction to
depth of use is the length of the well used and the probe
cord length.
Once lowered down a well, the probe begins to emit
neutrons. The velocity at which neutrons return to the
540 VADOSE ZONE MONITORING TECHNIQUES
probe is correlative to the substance by which they
were reflected. Neutrons reflected by heavy atoms of
the soil matrix return with higher velocities compared
to those of neutrons reflected by lighter hydrogen atoms.
Hydrogen may be found in organic matter or in mineral
lattices, but water is the greatest source of hydrogen in
the subsurface. Due to this, the counts of reflected low-
velocity neutrons are correlative to soil moisture content.
A complete moisture stratification along the length of a
well may be obtained, given a well-calibrated instrument.
Attributes of the neutron probe include ease of use
and high accuracy, given a well-calibrated probe. It may
be operated by one person, and soil moisture changes
of the order of 1–5 volume % may be measured. Water
flow pathways may also be located with the instrument.
The largest drawback of the neutron moisture probes
lies with calibration difficulties that may be extreme.
Improper neutron probe calibration nearly renders the
probe useless. Ideally, the probe is calibrated using soils
collected in a drum with a known quantity of water
from the field site of interest. Another drawback is that
neutron probes cannot distinguish between water and
other substances with similar hydrogen densities in the
subsurface, such as gasoline. Problems of applying single
calibrations to multiple media types also arise, which is
the case in heterogeneous soils. Inaccuracies may also
be result from improper well packing and the presence
of certain elements, such as migrating boron or chlorine.
Finally, care should be taken to mitigate exposure to the
neutron source (5–8).
Psychrometers
Psychrometers represent another tool adapted from
the agricultural industry for environmental monitoring.
Estimation of the hydraulic gradient using the matric
suction, osmotic potential, and water vapor potential
are possible.
The generic class of instruments used to measure
relative humidity using both wet and dry bulb ther-
mometers are called psychrometers; they may be used for
both surface (agriculture) and subsurface (environmental)
applications. For the latter, they are placed in a small
diameter well and surrounded by silica flour or air-filled
packers. A dry bulb thermometer is employed to measure
the ambient temperature, and a wet bulb thermometer is
then used to measure the temperature decrease resulting
from fluid evaporation in the subsurface (evaporative cool-
ing). Liquid-filled thermometers are not accurate enough
for the precision required, however, and thus more precise
thermocouple thermometers are employed. The junction
of two wires, a thermocouple, generates an electrical cur-
rent that is a function of the junction temperature. The
temperature of the thermocouple determines the activity
of the water surrounding it, which is then correlated with
the evaporation rate.
The direction and magnitude of water and contaminant
migration may be observed easily and inexpensively with
psychrometers, but regular calibration and monitoring
of the thermocouple are necessary as temperature
deviations of ±0.001 ◦ C between the thermocouple and
the liquid phase can introduce significant errors. Regular
calibration is necessary due to corrosion of metals and
the accumulation of salts on the thermocouple junction.
Also, psychrometer equations assume that water is the
first liquid to condense, and, at a site containing volatile
liquids, this may not always be true (9,10).
Tensiometry
Tensiometers measure the force which holds water to the
soil, or the soil–water potential, at shallow depths (0 to 8 m
BGS). Because water moves from areas of high potential
energy to low, evaluation of water’s energy may be used
to estimate the direction of water flow and the direction
of contaminant plume migration. Rapid measurements
at different locations may be obtained with tensiometers,
which may then be plotted to obtain a spatial distribution
of water and an estimation of the water flow direction.
A tensiometer measures matric potential using a porous
ceramic cup placed in the soil, a connecting tube, and a
vacuum gauge. The cup and tube are filled with water
to provide hydraulic contact from the tensiometer to the
water in the soil. Once placed in the ground, water flows
out of the cup and into the soil until equilibrium is reached.
Because the device is sealed from the atmosphere, the drop
of the water level in the tube creates a vacuum which is
measured by a vacuum gauge or pressure transducer from
the surface.
Tensiomenters represent a low-cost, simple, and
accurate measurement of matric potential (11–13). It is
one of the most widely used devices for monitoring soil
moisture levels in the vadose zone, though tensiometry
values are limited to tension values below 0.85 bar at
atmospheric pressures near 1 bar. For field applications,
air coming out of solution from water or air trapped in the
tube are the major shortcomings. During installation, care
must be taken to ensure good hydraulic contact between
the cup and the soil.
Time-Domain Reflectometry
Time-domain reflectometry (TDR) allows shallow, rapid,
in situ measurements of water content and electrical
conductivity in porous media (14–17). It is commonly
used to locate underground cables. It is a relatively new
development in vadose zone monitoring that was first
described by Topp et al. (18). Accuracy within 1–2% of
volumetric water content is feasible with TDR.
TDR uses a probe that is inserted into the soil equipped
with 2 or 3 metallic rods that transmit electromagnetic
pulses. The travel time of these pulses across the rods
through the soil is used to compute the soil’s bulk dielectric
constant from which the water content is inferred, which
provides an average soil water content over the depth of
probe insertion.
The main advantage of TDR is that simple and rapid
accumulation of shallow soil moisture content with high
volumetric accuracy is possible. Calibration requirements
are minimal compared to other indirect monitoring
methods, and measurements for spatial and temporal
analyses are easily obtained. The drawbacks associated
with TDR include temperature-dependent readings and
probe insertion in rocky or hard substrates because

readings are limited to depth of insertion. Additionally,
water molecules bound by interfacial forces in high surface
area substances, for example, clays, result in lowered
bulk dielectric constant compared to other soils at similar
water content.
Ground-Penetrating Radar
Ground-penetrating radar (GPR) is an adaptation of
seismic reflection for shallow use that images subsurface
features by deploying radar waves. It may be used to
image any homogeneity in the vadose zone, including
contaminant plumes and soil stratification.
GPR is used to observe the electromagnetic impedance
of soils using high-frequency radar waves to measure
subsurface properties. It consists of an enclosed wave
emitter and antenna unit that is dragged across the
ground, generally in a grid pattern. Wave pulses are
emitted and when an inhomogeneity is encountered, part
of the incident energy is reflected back to the radar
antenna within the unit. The reflected signal is amplified
and transformed into a viewable image. The data are in
visual form, not a quantified form, so results must be
interpreted.
Any inhomogeneity in the vadose zone can be imaged
with GPR. For vadose zone monitoring, contaminant
plumes of both light nonaqueous phase liquids (LNAPLs)
and dense nonaqueous phase liquids (DNAPLs), as well
as buried utility lines and pipes may be imaged with
GPR. Additionally, both underground storage tanks (UST)
and leaking underground storage tanks (LUST) may be
imaged and discerned. The most obvious drawback to GPR,
however, is the fact that all data must be interpreted from
visual images. Heterogeneities are viewed as anomalies
in imaged data, which must be deciphered according
to available information on subsurface conditions. Due
to this fact, GPR benefits from use in combination
with other vadose zone monitoring techniques to avoid
misinterpretations (19,20).
Electrical Resistance Blocks
Electrical resistance blocks have been used in agriculture
for more than 50 years to measure soil moisture in dry
areas. In the environmental field, they are employed
where other methods, including tensiometers, lysimeters,
and manometers, are not operable due to low soil–water
tension, generally between 0.5 and 15 bars.
The method consists of two or more electrodes that
are placed in a porous block, commonly made of gypsum.
The block is buried in the soil and allowed to remain
until moisture equilibrium with the native soil is
achieved. Changes in the electrical properties of the block,
measured with the electrodes, reflect changes in the water
content, which are measured with a wheatstone bridge
resistance meter.
The benefits of this technique include low-cost,
extensive historical use in agriculture research, and
simplicity of use. It is generally useful for approximate
soil moisture level changes. The disadvantages include
limited historical use in the environmental industry
and measurement and calibration difficulties at wet soil
VADOSE ZONE MONITORING TECHNIQUES 541
moisture potentials. Standard calibration curves exist, but
as in many indirect techniques, soil-specific calibration
of the instrument is recommended. Other limitations
are associated with dissolution of the gypsum block and
salinity changes in vadose zone water.
DIRECT MONITORING METHODS
Direct monitoring methods, in contrast to indirect
monitoring methods, employ sampling of the vadose zone
for liquid, soil, or gas. Obtaining liquid samples is fairly
straightforward, using the most widely practiced method,
suction lysimeters, but techniques for soil and gas sample
acquisition are generally more complicated. Soil sampling
commonly involves trenching or digging of some sort,
which may be subject to Occupational Safety and Health
Administration (OSHA) guidelines (24). Trenches deeper
than 5 feet are subject to OSHA regulations and
diggings deeper than 20 feet require the approval of a
professional engineer. Soils are also subject to maximum
allowable slope protocol, depending on their nature. In
addition, certain cases may require confined space, heavy
machinery, or decontamination permits. Readers are
referred to OSHA literature for proper digging protocol. In
gas sampling, problems stem from the nature of the phase
being sampled, and, to a lesser extent, safety concerns.
Vadose zone gas readily exchanges with any other gas
with which it comes in contact, and thus samplers must
be fully enclosed. Special safety procedures are required
when considering volatile substances, such as methane.
Additionally, direct monitoring methods have a much
higher risk of introducing contamination into the aquifer
or exposing workers to contamination, and thus proper
field procedures are absolutely necessary. In contaminated
areas, all equipment must be constructed of nonreactive
and low sorbing materials and be decontaminated between
each use (25–27).
Direct Liquid Monitoring Techniques
Directly monitoring the liquid phase of the vadose zone
involves obtaining a representative sample, whether in
the form of water or NAPL. These may be important
for soil–water chemistry analyses or contaminant plume
delineation, among others. The primary method of
collection is by using suction lysimeters placed into
unconsolidated soil. As a suction device, lysimeters rely
on applying a vacuum greater than the soil–water tension
to draw a water sample up to the surface. A porous cup is
inserted into the soil or in a shallow well, and a vacuum
tube connects it to the surface. Once the vacuum is applied,
water is drawn to the surface and a sample is obtained.
Outside of lysimeters, other methods involve driving
devices into unconsolidated sediment or lowering devices
down wells above the water table to acquire liquid samples.
Lysimeters have the advantage of being relatively cheap
and simple to install. Virtually no maintenance is required,
although sample volumes may be small.
Described by Gardner (28), wavelengths of light from
gamma to X ray may be used to monitor the water content
of the vadose zone by direct soil sampling. In practice,
542 VADOSE ZONE MONITORING TECHNIQUES
however, only gamma rays are employed. In this method,
a sample is collected and a narrow beam of gamma
radiation is sent through it. A recorder on the opposite
side of the sample records only those rays that passed
freely through the sample with no reflection along the
way. By this method, the volumetric water content can be
measured (29).
Direct Soil Monitoring Techniques
Due to the sorbing nature of many contaminants, soil
samples are used to monitor the spatial and temporal
distributions of contaminants in the soil. For certain appli-
cations, undisturbed samples are desired, in which case
special precautions must be taken, and certain sampling
tools and techniques must be employed. Generally, soil
sampling involves inserting a device into the ground to
recover a soil sample, although more intensive methods
may be necessary in certain situations. Hand-operated
samplers are typically used for shallow sampling and
mechanically powered devices for deeper sampling. There
are five basic hand-operated samplers, as described by Dor-
rance et al. (30), that include screw augers, barrel augers,
tube samplers, hand-powered augers, and bulk samplers
(shovels, etc). Each type has its own pros and cons depend-
ing on cost, ease of use, availability, size of matrix grains,
degree of sediment cohesiveness, and other factors. Due
to these, sampler choice is often site specific. For deeper
monitoring applications, mechanically powered tools such
as a drill rig must be employed.
Direct Vadose Zone Gas Monitoring
Monitoring the gas phase of the vadose zone is impor-
tant because many contaminants volatilize to a significant
degree. Unlike liquid and soil phases, the gaseous compo-
nent of the vadose zone is problematic because atmospheric
air and foreign vadose zone gases readily exchange and
react upon exposure. When sampling, all gaseous vadose
zone samples must avoid exposure to foreign air, and all
sampling methods must be designed to avoid it. In situ
sampling and devices with internal liners designed to seal
samples from the atmosphere are preferred (31).
Monitoring through sampling the gas phase typically
involves hydraulically ramming a specially designed
device into the soil where it collects samples via a vacuum
and pumps them to the surface for collection. Passive gas
sampling by an in situ absorbent, such as an activated
carbon trap placed in the vadose zone, may be used,
but less information is provided by this method. Direct
sampling of gases is more desirable, but it is more difficult
and costly. An alternative is monitoring gas advection
by using groundwater tracers. A tracer that partitions to
the gas phase is injected, samples are taken at other
sites, typically at wells, and analyzed for the tracer
concentration (32–35).
BIBLIOGRAPHY
1. Ankeny, M.M., Ahed, M., Kaspar, T.C., and Horton, R. (1991).
Simple method for determining unsaturated hydraulic
conductivity. Soil Sci. Soc. Am. J. 55: 467–470.
2. Lin, H.S. and McInnes, K.J. (1995). Water-flow in clay soil
beneath a tension infiltrometer. Soil Sci. 159(6): 375–382.
3. Reynolds, W.D. and Elrick, D.E. (1991). Determination of
hydraulic conductivity using a tension infiltrometer. Soil Sci.
Soc. Am. J. 55: 633–639.
4. Hussen, A.A. and Warrick, A.W. (1993). Algebraic models
for disc tension permeameters. Water Resour. Res. 29:
2779–2786.
5. Amoozegar, A., Martin, K.C., and Hoover, M.T. (1989). Effect
of access hole properties on soil water content determination
by neutron thermalization. Soil Sci. Soc. Am. J. 53: 330–335.
6. Arslan, A., Razzouk, A.K., and Alain, F. (1997). The perfor-
mance and radiation exposure of some neutron probes in
measuring the water content of the topsoil layer. Aust. J. Soil
Res. 35(6): 1397–1407.
7. Dillel, D. (1982). Introduction to Soil Physics. Academic Press,
London.
8. Grismer, M.E., Bali, K.M., and Robinson, F.E. (1995). Field
scale neutron probe calibration and variable analysis for clay
soil. J. Irrig. Drainage Eng.—ASCE 121(5): 354–362.
9. Comstock, J.P. (2000). Correction of thermocouple psychrom-
eter readings for the interaction of temperature and actual
water potential. Crop Sci. 40(3): 709–712.
10. Rawlins, S.L. and Campbell, G.S. (1986). Water potential:
Thermocouple psychrometry. In: Methods of Soil Analysis.
A. Klute (Ed.). Part 1, 2nd Edn. pp. 597–618.
11. Nyhan, J.W. and Drennon, B.J. (1990). Tensiometer data
acquisition system for hydrologic studies requiring high
temporal resolution. Soil Sci. Soc. Am. J. 54(1): 293–296.
12. Williams, T.H.L. (1978). Automatic scanning and recording
tensiometer system. J. Hydrol. 39(1–2): 175–183.
13. Whalley, W.R., Leedsharrison, P.B., Joy, P., and Hoefsloot, P.
(1994). Time-domain reflectometry and tensiometry combined
in an integrated soil–water monitoring system. J. Agric. Eng.
Res. 59(2): 141–144.
14. Jones, S.B., Wraith, J.M., and Or., D. (2002). Time domain
reflectometry measurement principles and applications.
Hydrologic Processes 16(1): 141–153.
15. Long, D.S., Wraith, J.M., and Kegel, G. (2002). A heavy-duty
time domain reflectometry soil moisture probe for use in
intensive field sampling. Soil Sci. Soc. Am. J. 66(2): 396–401.
16. Lee, J., Horton, R., and Jaynes, D.B. (2002). The feasibility of
shallow time domain reflectometry probes to describe solute
transport through undisturbed soil cores. Soil Sci. Soc. Am.
J. 66(1): 53–57.
17. Noborio, K. (2001). Measurement of soil water content and
electrical conductivity by time domain reflectometry: A
review. Comput. Electron. Agric. 31: 213–237.
18. Topp, G.C., Davis, J.L., and Annan, A.P. (1980). Electromag-
netic determination of soil–water content: measurement in
coaxial transmission lines. Water Resour. Res. 16: 574.
19. Daniels, J.J., Roberts, R., and Vendl, M. (1995). Ground
penetrating radar for detection of liquid contaminants. Appl.
Geophys. 33: 195–207.
20. Weil, G.J., Graf, R.J., and Leann, M.F. (1994). Investigations
of hazardous waste sites using thermal IR and ground
penetrating radar. Photogrammetric Eng. Remote Sensing
60(8): 999–1005.
21. Everett, L.G., Wilson, G.G., and Hoylman, E.W. (1984).
Vadose Zone Monitoring for Hazardous Waste Sites. Noyes,
Park Ridge, NJ.
22. Fishback, P.E. (1977). Scheduling irrigations by electri-
cal resistance blocks. NebGuide G 77–340, Institute of
Agriculture and Natural Resources, University of Nebraska,
Lincoln, NE.

23. Strangeways, I.C. (1983). Interfacing soil moisture gypsum
blocks with a modern data-logging system using a simple, low
cost, dc method. Soil Sci. 136: 322.
24. http://www.osha.gov/ (Occupational Safety and Health
Administration).
25. Fetter, C.W. (1999). Contaminant Hydrogeology, 2nd Edn.
Prentice-Hall, Upper Saddle River, NJ.
26. Hudak, P.F. and Loaiciga, H.A. (1999). Conjunctive vadose
and saturated zone monitoring for subsurface contamination.
Environ. Monitoring Assessment 59(1): 15–29.
27. Mickam, J.E., Bellandi, R., and Tifft, E.C., Jr. (1989). Equip-
ment decontamination procedures for ground water and
vadose zone monitoring programs: Status and prospects.
Ground Water Monitoring Rev. 9(2): 100–121.
28. Gardner, W.H. (1986). Water content. In: Methods of Soil
Analysis. Part 1. A. Klute (Ed.). Monograph 9, American
Society of Agronomy, Madison, WI.
29. Wilson, L.G. (1990). Methods for sampling fluids in the vadose
zone. In: Ground Water and Vadose Zone Monitoring. ASTM
STP 1053. D.M. Nielsen and Z.I. Johnson (Eds.) American
Society for Testing and Materials, Philadelphia, PA, pp. 7–24.
30. Dorrance, D.W., Wilson, L.G., Everett, L.G., and Cullen, S.J.
(1994). A compendium of soil samplers for the vadose zone. In:
Handbook of Vadose Zone Characterization and Monitoring.
L.G. Wilson, G.E. Lorne, and S.J. Cullen (Eds.). CRC Press,
Boca Raton, FL, pp. 401–428.
31. Lewis, T.E., Crockett, A.B., Siegrist, R.L., and Zarrabi, K.
(1991). Soil sampling and analysis for volatile organic
compounds. E.P.A. Ground Water Forum Issue, EPA/540/4-
91-001.
32. Fetter, C.W. (1999). Contaminant Hydrogeology, 2nd Edn.
Prentice-Hall, Upper Saddle River, NJ.
33. Hubbell, J.M., Wood, T.R., Higgs, B., Wylie, A.H., and McEl-
roy, D.L. (1998). Design, installation, and uses of combination
groundwater and gas sampling wells. Ground Water Monitor-
ing and Remediation 18(3): 151–157.
34. Marrin, D.L. (1988). Soil-gas sampling and misinterpretation.
Ground Water Monitoring Rev. 8(2): 51–54.
35. Thompson, G. and Marrin, D. (1987). Soil gas contaminant
investigations: A dynamic approach. Ground Water Monitor-
ing Rev. 8(4): 88–93.
VAPOR TRANSPORT IN THE UNSATURATED
ZONE
FRED D. TILLMAN
U.S. Environmental Protection
Agency
Athens, Georgia
JAMES A. SMITH
University of Virginia
Charlottesville, Virginia
INTRODUCTION
The term ‘‘vapor’’ describes the gaseous phase of a sub-
stance that is normally a liquid or solid at standard atmo-
spheric pressure and temperature. Examples of vapors
that are important in environmental systems include
humidity (water vapor), gasoline vapor, and solvent vapor.
VAPOR TRANSPORT IN THE UNSATURATED ZONE 543
Soil and groundwater that are contaminated by organic
solvents and fuel hydrocarbons may emit vapors that can
move through the subsurface and reach the land surface.
Human exposure to these harmful vapors has become an
increasingly important topic in the evaluation and remedi-
ation of contaminated soil and groundwater. The potential
for inhalation of vapors transported from a groundwa-
ter plume to the interior of buildings often controls the
outcome of a risk based corrective action (RBCA) anal-
ysis (1). Additionally, vapors emanating from subsurface
contamination contribute to the reduction of contaminant
mass and may provide a beneficial mechanism of natural
attenuation at sites where there are no human receptors.
Subsurface processes affecting the transport of organic
vapors through the soil above the water table (here defined
as the unsaturated zone) may slow, reduce, or enhance the
release of these vapors to potential receptors at the land
surface. A careful investigation of organic vapor trans-
port in the unsaturated zone is therefore important in
assessing the potential for natural attenuation and the
impact of even deep contamination on receptors at the
land surface. This article describes potential sources, fate
and transport, and measurement of organic vapors in
the unsaturated zone. Although this article deals exclu-
sively with organic vapors, many of the fate and transport
properties discussed are also applicable to other vapors,
including metals and moisture.
SOURCES OF ORGANIC VAPORS
Contaminants of concern in vapor transport in the
unsaturated zone are typically volatile organic compounds
(VOCs), although vapors emanating from inorganic
sources such as mercury vapor may be of concern as
well. A chemical is considered volatile if its Henry’s law
constant is 1 × 10−5 atm-m3 /mol or greater (2). Examples
of VOCs that are important in impacted environmental
systems include chlorinated solvents such as carbon
tetrachloride, tetrachloroethylene, and trichloroethylene
(and their degradation compounds), fuel hydrocarbons
such as benzene, toluene, ethylbenzene, and o,m,p-
xylenes as well as volatile pesticides such as chlordane,
aldrin, and lindane. The U.S. Environmental Protection
Agency lists 107 compounds whose toxicity and volatility
produce a potentially unacceptable inhalation risk to
receptors (2). These VOCs can be released into the
subsurface environment from leaking landfill liners,
improper disposal, accidental spillage, or leaking storage
tanks (LSTs). Once in the subsurface, these compounds can
become bound to the soil matrix, dissolved in groundwater
(or soil water), and/or exist as a separate, residual
phase known as a nonaqueous phase liquid (NAPL).
Soil, aqueous, and NAPL-phase organics may all be
sources of organic vapors in the subsurface. Therefore,
organic vapor transport in the unsaturated zone requires
understanding of interphase mass-transfer processes
because the contaminant can be distributed among soil
gas, water, soil, and NAPL phases. In modeling organic
vapor transport, the equilibrium vapor concentration at
the source of contamination is often estimated from the
following relationships (3):
544 VAPOR TRANSPORT IN THE UNSATURATED ZONE
Soil contamination:
HTS CR ρb
Csource = (4)
θw + Kd ρb + HTS θa
Groundwater contamination:
Csource = HTS Cw Henry’s Law
NAPL contamination:
xi pi MWi
Csource = Raoult’s Law
RT
where Csource is the vapor concentration at the source of
contamination (ML−3 ), HTS is the dimensionless Henry’s
law constant at the system temperature (unitless), CR
is the soil concentration (MM−1 ), ρb is the soil dry
bulk density (ML−3 ), θw is the soil water-filled porosity
(L3 L−3 ), θa is the soil air-filled porosity (L3 L−3 ), Kd
is the soil–water partition coefficient (L3 M−1 ), Cw is
concentration of the contaminant in the groundwater
(ML−3 ), xi is the mole fraction of component i in the NAPL
mixture (unitless), pi is the vapor pressure of component i
at the soil temperature (ML−1 T−2 ), MWi is the molecular
weight of component i (M), R is the ideal gas constant
(ML2 T−2 T−1 M−1 , e.g., 8.314570 J◦ K−1 mol−1 ), and T is the
soil temperature (T). Both Henry’s law constant and vapor
pressure rise with temperature for organic compounds of
interest, so a higher source vapor concentration is expected
from higher soil and groundwater temperature.
TRANSPORT AND FATE OF ORGANIC VAPORS IN THE
UNSATURATED ZONE
Organic vapors emanating from contaminated soil or
groundwater or from a residual phase such as gasoline
floating on the water table may move through unsaturated
zone soil gas by diffusion or soil-gas advection due
to pressure or density gradients or a combination of
these processes.
Diffusion
Molecular diffusion is the spreading out of compounds from
random collisions resulting from thermal motion of atoms.
These collisions may be between molecules themselves or
between molecules and their surroundings. Under most
environmental conditions, molecular diffusion in natural
systems proceeds from locations of higher concentration
toward locations of lower concentrations. In a typical
scenario, organic vapors above a contaminated water table
(high concentration) diffuse toward the land surface (lower
concentration). The well-known relation describing the
diffusion of a compound across a unit of cross-sectional
area is Fick’s first law:
dC
Fx = −Deff
dx
where Fx is the mass flux [ML−2 T−1 ], Deff is the effective
diffusion coefficient of the compound in the gas phase
[L2 T−1 ], and dC/dx is the concentration gradient of
the compound in one dimension [ML−3 L−1 ]. From this
equation, it is apparent that the rate of molecular
(1)
diffusion in the gas phase depends on the concentration
gradient and the effective diffusion coefficient of the
compound of interest. Several relations relate the effective
diffusion coefficient to the free-air diffusion coefficient
(2) of the compound and the total and air-filled porosities
of the diffusing media (5–14). An increase in diffusive
flux is seen in soils with greater air passageways (i.e.,
greater porosities and air-filled porosities). Therefore, in
a layered unsaturated zone, vapor diffusion from depth
(3)
to the land surface is limited by the wettest, least
porous soil layer. Because free-air diffusion coefficients are
compound dependent, ranging from 2.5 × 10−3 cm2 s−1 for
hexachloroethane to 2.71 × 10−1 cm2 s−1 for chloroethane
among the 107 volatile compounds of concern to the
U.S.EPA (2), different chemicals diffuse at different
rates under the same concentration gradients. Increased
temperature also produces an increased free-air diffusion
coefficient, leading to a greater rate of diffusion relative to
the same system at lower temperatures. Fick’s First Law
may not sufficiently describe vapor diffusion in systems
where pore sizes are very small (Knudsen diffusion)
and when volatile species constitute a substantial
fraction of the total soil-gas concentration (nonequimolar
diffusion) (15–17). In these situations, it may be necessary
to employ the Stefan–Maxwell equation.
Advection
The flow of soil gas (advection) in the subsurface may
be caused by gas-pressure gradients or, in certain cases,
gas density gradients. Pressure-driven advection is pro-
duced when differences in soil-gas pressure form, causing
soil gas to flow and carrying any vapors present with
it. Air pressure gradients in the subsurface of natural
systems may result from several phenomena. As diurnal
or weather related atmospheric pressure cycles occur at
the land surface, pressure waves are transmitted into the
unsaturated zone and air may flow in response—a pro-
cess known as ‘‘barometric pumping.’’ Barometric pump-
ing may cause soil gas to flow either toward the land
surface carrying soil vapor or away from the land sur-
face bringing in fresh atmospheric air (13,15,18–25). The
underpressurization of an overlying building will pro-
duce gas pressure differences in subsurface soils. This
underpressurization may be caused by thermal differ-
ences between indoor and outdoor air (stack effects), wind
loading on the building superstructure, and unbalanced
building ventilation (26,27). Soil-gas pressure gradients
may also be produced by a rapidly rising or falling water
table, as in coastal zones (28), or through the buildup
of gas pressure from decomposing organic matter inside
a landfill (29). Finally, natural temperature differences
between warmer deep and cooler shallow soil gas will
cause soil gas to rise (30,31). Density-driven flow of organic
vapors may occur in the vicinity of residual-phase organic
(4) compounds whose saturation gas densities are greater
than that of air. As organic liquids with high vapor pres-
sures and molecular weights volatilize, the density of the
soil gas surrounding the liquid changes. In almost all

cases, organic liquids have molecular weights that are
greater than air, so the resulting density-driven flow is
in a downward direction and is proportional to soil per-
meability and density differences between the vapor and
air (32–34). Organic compounds for which density-driven
advection may be significant include methylene chloride,
1,2-dichloroethylene, 1,1,1-trichloroethane, carbon tetra-
chloride, and 1,1-dichloroethane, among others (32–34).
Sorption
As organic vapors move through the unsaturated zone by
diffusion and advection, they come in contact with soil
moisture, infiltrating rainwater, and the soil matrix itself.
Each of these interactions may affect the concentration
of the contaminant in the soil gas. Depending on the
compound, organic vapors may adsorb to soil grain
surfaces or partition into soil organic matter (35–40).
Adsorption of relatively nonpolar organic vapors is
suppressed by high humidity in the subsurface because as
polar water molecules can effectively out-compete organic
vapors for mineral-surface adsorption sites (41,42). For
these high-humidity conditions, sorption may be limited
to organic vapor partition into soil organic matter (41,42).
Soil moisture trapped in unsaturated-zone pore space
or infiltrating rain water may also sorb organic vapors
to varying degrees (43,44). Finally, gas-phase organic
compounds may adsorb to the air–water interface
in unsaturated zones (45,46). Each of these sorption
processes may act as both a source and a sink of organic
vapors in the unsaturated zone.
Biodegradation
Under favorable conditions, organic vapors in the unsatu-
rated zone that partition into soil moisture may be biode-
graded by indigenous bacteria in oxidation/reduction reac-
tions. The aerobic biodegradation of petroleum hydrocar-
bons requires an abundant oxygen supply as well as neces-
sary nutrients of nitrogen and phosphorus (47–51). When
oxygen is depleted, other possible electron acceptors for
biodegradation of petroleum hydrocarbons include nitrate
(NO3 − ), iron oxides (e.g., Fe(OH)3 ), sulfate (SO4 2− ), and
carbon dioxide (CO2 ) (48). Lightly chlorinated compounds
(e.g., chlorobenzene, dichlorobenzene) may be biodegraded
under aerobic conditions. The more highly chlorinated
hydrocarbons are recalcitrant to aerobic biodegradation
but may undergo direct or cometabolic anaerobic reduc-
tive dechlorination. Reductive dechlorination has been
observed to be most effective under sulfate-reducing and
methanogenic conditions (52). In direct reductive dechlori-
nation, the chlorinated hydrocarbon is used as an electron
acceptor and the bacteria gain energy and grow as a result
of the reaction (53). In cometabolic reductive dechlori-
nation, enzymes produced during microbial metabolism
of another hydrocarbon fortuitously reduce and dehalo-
genate the chlorinated contaminant. For either reductive
dechlorination process to be successful, a primary sub-
strate (electron donor) such as soil organic matter, lactate,
acetate, methanol, or glucose is necessary (54).
VAPOR TRANSPORT IN THE UNSATURATED ZONE 545
VAPOR SAMPLING AND FLUX MEASUREMENT
Organic vapors in the unsaturated zone above contami-
nated soil or groundwater are typically sampled to pro-
vide evidence of biodegradation (49,55), calibrate vapor
flow and transport models for estimating natural atten-
uation (23), or to estimate the region of contaminated
groundwater (42,56,57). Soil gas is often sampled through
low volume, inert pipe or tubing penetrating the sub-
surface to the desired sampling depth (42,44,49,56,58,59).
This tubing can be installed with the help of a Geoprobe-
type device or inserted in a previously drilled borehole.
The sampling portion of the tubing must be sealed from
the atmosphere and other sampling depths to prevent
preferential flow down the borehole during sampling. Sam-
ples may be collected from these vapor probes in various
ways, including direct syringe withdrawal (49), pumping
of soil gas into Tedlar bags (57,60,61), or pumping of
soil gas through adsorbent traps (58,62,63). Analyses of
samples collected by syringe withdrawal require on-site
analysis. Samples collected in Tedlar bags or onto adsor-
bent traps may be transported to an off-site laboratory for
analysis. Alternatively, there are field-portable gas detec-
tors and gas chromatographs with flame-ionization and
photoionization detectors capable of measuring real time
hydrocarbon concentrations (47,57,64).
Efforts to measure the flux of VOCs from the unsat-
urated zone to the atmosphere at the land surface
have included using the aerodynamic method, which
necessitates the measurement of trace concentrations
and atmospheric conditions such as wind speed, tem-
perature and pressure (65); using groundsheets sampled
through a nylon sampling tube (66); and several cham-
ber designs (58,59,67–71). Surface measurements of vapor
flux are important because they provide a direct measure-
ment of total natural flux that can be used to estimate
volatilization rates for a field site. This natural attenua-
tion process for contaminants in soil and groundwater may
be significant, and, at times, approach the removal rate of
active remediation systems (58). Complexities involved in
the measurement of vapor transport from the subsurface
to the atmosphere include possible preferential flow paths
(e.g., underground utilities, macropores, etc.) and spatial
and temporal variability in vapor concentrations. Spatial
variability may be produced by subsurface heterogeneities
that are geologic as well as constructed barriers (e.g.,
storm drains, basement foundations). Temporal variability
in vapor concentrations is produced by changing soil mois-
ture content, temperature, and gas-pressure gradients.
SUMMARY
Analysis of the transport of organic vapors in the unsatu-
rated zone is fundamental in evaluating potential human
exposure to harmful compounds from a contaminated sub-
surface as well as determining a site’s potential for natural
attenuation. Analysis of soil gas vapors may also aid in
estimating the areal extent of a contaminated plume. Soil,
aqueous, and NAPL-phase organics may all be sources
of organic vapors in the subsurface, and interphase mass
transfer to and from soil, soil moisture, groundwater, and
546 VAPOR TRANSPORT IN THE UNSATURATED ZONE
soil gas all play a role in vapor transport in the unsaturated
zone. Diffusion of the vapor phase contaminant proceeds
from higher concentrations near the source toward lower
concentrations at the land surface. Advection of soil gas
results from pressure gradients in the subsurface and
from density-driven flow for certain high molecular weight
compounds. Under favorable conditions, organic vapors in
the unsaturated zone may be biodegraded by indigenous
bacteria in oxidation/reduction reactions. All biodegrada-
tion reactions of organic compounds require nutrients and
electron acceptors in sufficient, usable quantities. Soil-gas
concentrations may be quantified through the use of vapor
probes, adsorbent traps, Tedlar bags, direct syringe with-
drawal with field-portable analytical equipment, and the
use of gas chromatography. The total flux of contaminants
from the subsurface to the land surface may be estimated
by the aerodynamic method, through the use of ground
sheets, or with flux chambers. Surface flux measurement
is useful for estimating the total volatilization of contam-
inants from a site, but all vapor-phase measurements are
complicated by spatial and temporal variability.
BIBLIOGRAPHY
1. Laubacher, R.C., Bartholomae, P., Velasco, P., and Reisinger,
H.J. (1997). An evaluation of the vapor profile in the vadose
zone above a gasoline plume. In: Pet. Hydrocarbons Org.
Chem. Ground Water: Prev. Detection Remediation Conf.,
Houston, TX, pp. 396–409.
2. Environmental Quality Management. (2003). User’s Guide for
Evaluating Subsurface Vapor Intrusion into Buildings. U.S.
Environmental Protection Agency, Office of Emergency and
Remedial Response, Durham, NC.
3. U.S. Environmental Protection Agency. (2002). Draft Guid-
ance for Evaluating the Vapor Intrusion to Indoor Air Pathway
from Groundwater And Soils. USEPA Office of Solid Waste
and Emergency Response.
4. Johnson et al. (1990).
5. Buckingham, E. (1904). Contribution to Our Knowledge of the
Aeration of Soils. USDA Bureau of Soils Bulletin No. 25.
6. Penman, H.L. (1940). Gas and vapor movements in soil I. The
diffusion of vapours through porous solids. J. Agric. Sci. 30:
463–462.
7. Van Bavel, C.H.M. (1952). Gaseous diffusion and porosity in
porous media. Soil Sci. 73: 91–105.
8. Rust, R.H., Klute, A., and Geiseking, G.E. (1957). Diffusion-
porosity measurements using a non-steady state system. Soil
Sci. 84: 453–465.
9. Dye, R.F. and Dallavale, J.M. (1958). Diffusion of gases in
porous media. Ind. Eng. Chem. 50: 1195–1200.
10. Millington, R.J. (1959). Gas diffusion in porous media. Science
130: 100–102.
11. Currie, J.A. (1970). Movement of gases in soil respiration.
Sorption and Transport Processes in Soils. SCI Monogr. Ser.
37: 152–171.
12. Nilson, R.H., Peterson, E.W., and Lie, K.H. (1991). Atmo-
spheric pumping: A mechanism causing vertical transport
of contaminated gases through fractured permeable media.
J. Geophys. Res. 96: 21933–21948.
13. Bartelt-Hunt, S.L. and Smith, J.A. (2002). Measurement of
effective air diffusion coefficients for trichloroethylene in
heterogeneous soil. J. Contam. Hydrol. 56: 193–208.
14. Rolston, D.E. and Moldrup, P. (2002). Gas diffusivity. In:
Methods of Soil Analysis: Part 4-Physical Methods. J.H. Dane
and G.C. Topp (Eds.). Soil Science Society of America,
Madison, WI, pp. 1113–1158.
15. Thorstenson, D.C. and Pollock, D.W. (1989). Gas transport
in unsaturated zones: Multicomponent systems and the
adequacy of Fick’s laws. Water Resour. Res. 25: 477–507.
16. Baehr, A.L. and Bruell, C.J. (1990). Application of the Stefan-
Maxwell equations to determine limitations of Fick’s law
when modeling organic vapor transport in sand columns.
Water Resour. Res. 26: 1155–1163.
17. Partridge, G.P., McLeary, K.S., Showers, T.R., Huebner, R.S.,
and Elliott, H.A. (2002). Single and multicomponent gas
phase diffusion in porous media: Modeling and laboratory
measurements. Soil Sediment Contam. 11: 555–581.
18. Sleep, B.E. and Sykes, J.F. (1989). Modeling the transport of
volatile organics in variably saturated media. Water Resour.
Res. 25: 81–92.
19. Massmann, J. and Farrier, D.F. (1992). Effects of atmospheric
pressures on gas transport in the vadose zone. Water Resour.
Res. 28: 777–791.
20. Auer, L.H., Rosenberg, N.D., Birdsell, K.H., and Whiteny,
E.M. (1996). The effects of barometric pumping on contami-
nant transport. J. Contam. Hydrol. 24: 145–166.
21. Elberling, B., Larsen, F., Christensen, S., and Postma, D.
(1998). Gas transport in a confined unsaturated zone
during atmospheric pressure cycles. Water Resour. Res. 34:
2855–2862.
22. Tillman, F., Choi, J.C., Katchmark, W., Smith, J.A., and
Wood, H.G.I. (2001). Unsaturated-zone airflow: Implications
for natural remediation of ground water by contaminant
transport through the subsurface. In: Physicochemical
Groundwater Remediation. J.A. Smith and S.E. Burns (Eds.).
Kluwer Academic/Plenum, New York, pp. 307–339.
23. Choi, J.W., Tillman, F.D., and Smith, J.A. (2002). Relative
importance of gas-phase diffusive and advective trichloroethy-
lene (TCE) fluxes in the unsaturated zone under natural
conditions. Environ. Sci. Technol. 36: 3157–3164.
24. Neeper, D.A. (2002). Investigation of the vadose zone using
barometric pressure cycles. J. Contam. Hydrol. 54: 59–80.
25. Rossabi, J. and Falta, R.W. (2002). Analytical solution for
subsurface gas flow to a well induced by surface pressure
fluctuations. Ground Water 40: 67–75.
26. Nazaroff, W.W., Lewis, S.R., Doyle, S.M., Moed, B.A., and
Nero, A.V. (1987). Experiments on pollutant transport from
soil into residential basements by pressure-driven airflow.
Environ. Sci. Technol. 21: 459–466.
27. Garbesi, K. and Sextro, R.G. (1989). Modeling and field
evidence of pressure-driven entry of soil gas into a house
through permeable below-grade walls. Environ. Sci. Technol.
23: 1481–1487.
28. Li, L., Dong, P., and Barry, D.A. (2002). Tide-induced water
table fluctuations in coastal aquifers bounded by rhythmic
shorelines. J. Hydraul. Eng. 128: 925–933.
29. Little, J.C., Daisey, J.M., and Nazaroff, W.W. (1992). Trans-
port of subsurface contaminants into buildings: An exposure
pathway for volatile organics. Environ. Sci. Technol. 26:
2058–2066.
30. Gustin, M.S., Taylor, G.E., and Maxey, R.A. (1997). Effect
of temperature and air movement on the flux of elemental
mercury from substrate to the atmosphere. J. Geophys. Res.
102: 3891–3898.
31. Krylov, V.V. and Ferguson, C.C. (1998). Contamination of
indoor air by toxic soil vapours: The effects of subfloor

ventilation and other protective measures. Build. Environ.
33: 331–347.
32. Falta, R.W., Javandel, I., Pruess, K., and Witherspoon, P.A.
(1989). Density-driven flow of gas in the unsaturated zone
due to the evaporation of volatile organic compounds. Water
Resour. Res. 25: 2159–2169.
33. Mendoza, C.A. and Frind, E.O. (1990). Advective-dispersive
transport of dense organic vapors in the unsaturated zone 1.
Model development. Water Resour. Res. 26: 379–387.
34. Mendoza, C.A. and Frind, E.O. (1990). Advective-dispersive
transport of dense organic vapors in the unsaturated zone: 2.
Sensitivity analysis. Water Resour. Res. 26: 387–398.
35. Goss, K.U. (1994). Adsorption of organic vapors on polar
mineral surfaces and on a bulk water surface: Development
of an empirical predictive model. Environ. Sci. Technol. 28:
640–645.
36. Goss, K.U. (1994). Effects of temperature and relative
humidity on the sorption of vapors on quartz sand. Environ.
Sci. Technol. 26: 2287–2294.
37. Goss, K.-U. and Eisenreich, S.J. (1996). Adsorption of VOCs
from the gas phase to different minerals and a mineral
mixture. Environ. Sci. Technol. 30: 2135–2142.
38. Popovicova, J. and Brusseau, M.L. (1998). Contaminant mass
transfer during gas-phase transport in unsaturated porous
media. Water Resour. Res. 34: 83–92.
39. Ruiz, J., Bilbao, R., and Murillo, M. (1998). Adsorption of
different VOC onto soil minerals from gas phase: Influence
of mineral, type of VOC, and air humidity. Environ. Sci.
Technol. 32: 1079–1084.
40. Goss, K.-U., Buschmann, J., and Schwarzenbach, R.P. (2004).
Adsorption of organic vapors to air-dry soils: Model
predictions and experimental validation. Environ. Sci.
Technol. 38: 3667–3673.
41. Chiou, C.T. and Shoup, T.D. (1985). Soil sorption of organic
vapors and effects of humidity on sorptive mechanism and
capacity. Environ. Sci. Technol. 19: 1196–1200.
42. Smith, J.A., Chiou, C.T., Kammer, J.A., and Kile, D.E. (1990).
Effect of soil moisture on the sorption of trichloroethene
vapor to vadose-zone soil at Picatinny Arsenal, New Jersey.
Environ. Sci. Technol. 24: 676–683.
43. Cho, H.J. and Jaffé, P.R. (1990). The volatilization of organic
compounds in unsaturated porous media during infiltration.
J. Contam. Hydrol. 6: 387–410.
44. Cho, H.J., Jaffé, P.R., and Smith, J.A. (1993). Simulating
the volatilization of solvents in unsaturated soils during
laboratory and field infiltration experiments. Water Resour.
Res. 29: 3329–3342.
45. Kim, H., Rao, P.S., and Annable, M.D. (1997). Determination
of effective air-water interfacial area in partially saturated
porous media using surfactant adsorption. Water Resour. Res.
33: 2705–2711.
46. Kim, H., Annable, M.D., and Rao, P.S.C. (1998). Influence of
air-water interfacial adsorption and gas-phase partitioning
on the transport of organic chemicals in unsaturated porous
media. Environ. Sci. Technol. 32: 1253–1259.
47. Ostendorf, D.W. and Kampbell, D.H. (1991). Biodegradation
of hydrocarbon vapors in the unsaturated zone. Water Resour.
Res. 27: 453–462.
48. Norris, R.D., Hinchee, R.E., Brown, R., McCarty, P.L., Wil-
son, J.T., Reinhard, M., Bouwer, E.J., Borden, R.C., Vogel,
T.M., Thomas, J.M., and Ward, C.H. (1994). Handbook of
Bioremediation. Lewis, Boca Raton, FL.
49. Lahvis, M.A. and Baehr, A.L. (1996). Estimation of rates of
aerobic hydrocarbon biodegradation by simulation of gas
VAPOR TRANSPORT IN THE UNSATURATED ZONE 547
transport in the unsaturated zone. Water Resour. Res. 32:
2231–2249.
50. Lahvis, M.A., Baehr, A.L., and Baker, R.J. (1999). Quantifi-
cation of aerobic biodegradation and volatilization rates of
gasoline hydrocarbons near the water table under natural
attenuation conditions. Water Resour. Res. 35: 753–765.
51. Hers, I., Atwater, J., Li, L., and Zapf-Gilje, R. (2000). Eval-
uation of vadose zone biodegradation of BTX vapours. J.
Contam. Hydrol. 46: 233–264.
52. U.S. Environmental Protection Agency. (2000). Engineered
approaches to in situ bioremediation of chlorinated solvents:
Fundamentals and field applications. EPA 542-R-00-008,
Washington, DC.
53. McCarty, P.L. (1997). Breathing with chlorinated solvents.
Science 276: 1521–1522.
54. Vogel, T.M., Criddle, C.S., and McCarty, P.L. (1987). Trans-
formations of halogenated aliphatic compounds. Environ. Sci.
Technol. 21: 722–736.
55. Roggemans, S., Bruce, C.L., Johnson, P.C., and Johnson, R.L.
(2001). Vadose zone natural attenuation of hydrocarbon
vapors: An empirical assessment of soil gas vertical profile
data. Am. Pet. Inst. Tech. Bull. 15: 12.
56. Kerfoot, H.B. (1987). Soil-gas measurement for detection of
groundwater contamination by volatile organic compounds.
Environ. Sci. Technol. 21: 1022–1024.
57. Moseley, C.L. and Meyer, M.R. (1992). Petroleum Contami-
nation of an elementary school: A case history involving air,
soil-gas and groundwater monitoring. Environ. Sci. Technol.
26: 185–192.
58. Smith, J.A., Tisdale, A.K., and Cho, H.J. (1996). Quantifi-
cation of natural vapor fluxes of trichloroethene in the
unsaturated zone at Picatinny Arsenal, New Jersey. Environ.
Sci. Technol. 30: 2243–2250.
59. Tillman, F., Choi, J.-W., and Smith, J.A. (2003). A comparison
of direct measurement and model simulation of total flux
of volatile organic compounds from the subsurface to the
atmosphere under natural field conditions. Water Resour.
Res. 39: 1284–1294.
60. Fischer, M.L. et al. (1996). Factors affecting indoor air
concentrations of volatile organic compounds at a site of
subsurface gasoline contamination. Environ. Sci. Technol.
30: 2948–2957.
61. Jeng, C.Y., Kremesec, V.J., and Primack, H.S. (1996). Models
of hydrocarbon vapor diffusion through soil and transport into
buildings. In: Pet. Hydrocarbons Org. Chem. Ground Water:
Prev. Detection Remediation Conf. National Ground Water
Association, Houston, TX, pp. 319–338.
62. Tillman, F. (2003). Design and testing of a chamber device
to measure total flux of volatile organic compounds from
the unsaturated zone under natural conditions. Ph.D.
Dissertation. University of Virginia, Charlottesville, VA.
63. Hodgson, A.T., Garbesi, K., Sextro, R.G., and Daisey, J.M.
(1992). Soil-gas contamination and entry of volatile organic
compounds into a house near a landfill. J. Air Waste Manage.
Assoc. 42: 277–283.
64. Ririe, G.T., Sweeney, R.E., Daugherty, S.J., and Peuron, P.M.
(1998). A vapor transport model that is consistent with field
and laboratory data. Pet. Hydrocarbons Org. Chem. Ground
Water: Prev. Detection Remediation Conf., Ground Water
Association, Houston, TX, pp. 299–309.
65. Chen, C., Green, R.E., Thomas, D.M., and Knuteson, J.A.
(1995). Modeling 1,3-dichloropropene fumigant volatilization
with vapor-phase advection in the soil profile. Environ. Sci.
Technol. 29: 1816–1821.
548 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS
66. Wyatt, D.E., Richers, D.M., and Pirkle, R.J. (1995). Baromet-
ric pumping effects on soil gas studies for geological and
environmental characterization. Environ. Geol. 25: 243–250.
67. Eklund, B., Balfour, W.D., and Schmidt, C.E. (1985). Mea-
surement of fugitive volatile organic emission rates. Environ.
Prog. 4: 199–202.
68. Batterman, S.A., McQuown, B.C., Murthy, P.N., and McFar-
land, A.R. (1992). Design and evaluation of a long-term soil
gas flux sampler. Environ. Sci. Technol. 26: 709–714.
69. Reinhart, D.R., Cooper, D.C., and Walker, B.L. (1992). Flux
chamber design and operation for the measurement of
municipal solid waste landfill gas emission rates. J. Air Waste
Manage. Assoc. 42: 1067–1070.
70. Eklund, B., Anderson, E.P., Walker, B.L., and Burrows, D.B.
(1998). Characterization of landfill gas composition at the
Fresh Kills municipal solid-waste landfill. Environ. Sci.
Technol. 32: 2233–2237.
71. Tillman, F. and Smith, J.A. (2004). Design and laboratory
testing of a chamber device to measure total flux of volatile
organic compounds from the unsaturated zone under natural
conditions. J. Contam. Hydrol. 75: 71–90.
APPLICATIONS OF SOIL VAPOR DATA TO
GROUNDWATER INVESTIGATIONS
BLAYNE HARTMAN
H&P Mobile Geochemistry
Solana Beach, California
JAMES A. JACOBS
Environmental Bio-Systems, Inc.
Mill Valley, California
This article describes the uses and interpretation of soil
vapor data for a variety of common applications related
primarily to groundwater. It begins with an overview of
the applications of soil vapor surveys, a brief discussion
of the various soil vapor methods, and then focuses on a
number of specific issues relating to soil vapor surveys for
groundwater investigations. The intent of this article is to
guide professionals on when, why, and how soil vapor data
can be useful for groundwater-related investigations and
how to interpret the results.
BACKGROUND AND APPLICATIONS
Soil vapor in the pore spaces, also called soil gas, soil
air, pore gas, and interstitial gas refers to the vapor
that lies among the soil grains in the soil column lying
between the ground surface and the water table (vadose
zone). Soil vapor data are collected in a variety of
ways, and the data can be used for various purposes.
1
Some of this material is excerpted from previous work,
appearing in ‘‘Petroleum Hydrocarbon Contamination-Legal and
Technical Issues’’ and ‘‘Chlorinated Solvent Contamination-
Legal and Technical Issues’’ Copyright  1998, 1999, Argent
Communications Group. Reprinted by permission of Argent
Communication Group, P.O. Box 1425, Foresthill, CA 95631.
Telephone 1-800-419-2741; or sales@ argentco.com.
Soil vapor samples are most commonly analyzed for
volatile organic compounds (VOCs). VOCs include simple
hydrocarbons such as methane; lighter hydrocarbons such
as gasoline; gasoline constituents such as benzene and
methyl tert-butyl ether (MtBE); and chlorinated solvents
such as trichloroethylene (TCE), trichloroethane (TCA),
and tetrachlorethylene (PCE).
Soil vapor is surveyed for a variety of reasons.
Historically, soil vapor surveys have been used primarily
for site assessment to identify soil and groundwater
contamination (1–4). More recent applications include
evaluation of upward and downward vapor transport to
monitor site remediation (natural and engineered) and to
monitor landfills and migrating VOCs (5–7).
Soil vapor surveys are quick, relatively nondisruptive,
and generally inexpensive. The surveys are frequently
performed early in the site investigation, when investiga-
tion for numerous VOCs is performed. The data obtained
are often used to focus subsequent soil and groundwater
investigatory sampling programs.
Regulatory agencies across the country are increasingly
concerned with human health risk due to the upward
migration of contaminant vapors through the vadose
zone into buildings. Vertical soil vapor profiles enable
measuring actual soil vapor concentrations that give a
more accurate calculation of the upward vapor migration
than values calculated from soil or groundwater data.
These data are used in risk based corrective action (RBCA)
computer models (8).
Use of soil vapor data for monitoring site remediation
takes many forms. Most commonly, concentrations of
vapor contaminants are monitored to measure the
progress of a vapor extraction system. Remediation of
groundwater or soil by natural attenuation requires
monitoring the concentrations of contaminants, by-
products, and oxygen in the soil vapor to measure the
progress of the degradation of the contaminant plume.
Soil vapor is frequently measured at the fringes
of landfills to ensure that harmful compounds such
as methane, chlorinated compounds (primarily vinyl
chloride), and petroleum related carcinogens such as
benzene are not migrating from landfills to neighboring
developments.
SOIL VAPOR METHODS
Collection of Soil Vapor Samples
Three methods exist for collecting soil vapor data:
(1) active, (2) passive, and (3) flux chambers. A brief
overview of these methods follows. More comprehensive
descriptions of soil vapor collection methods and protocols
are given in documents written by Hartman (9,10),
Hartman and Jacobs (5,11), EPA (1), ASTM (12), and
Devitt et al. (4).
The active approach is quantitative and consists of the
withdrawal of a volume of vapor from the subsurface,
typically using a steel sampling rod driven into the ground
from 0.9 to 6 meters below ground surface, and subsequent
analysis of the withdrawn vapor. Originally, surveys were
limited to shallow depths of 3 meters because the sampling
 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS
rods were driven and extracted using hand techniques.
Currently, active surveys can routinely achieve depths
exceeding 30.5 meters due to the use of direct-push probe
rigs, cone penetrometer (CPT) sampling rigs, and full-
size drill rigs to collect soil vapor using newly developed
sampling tools.
Sample containers for active vapor sampling include
syringes, gas sampling Tedlar bags, preevacuated
stainless steel canisters (e.g., SUMMA canisters), and
gastight vials (e.g., 40 mL preevacuated volatile organic
analysis (VOAs) vials or headspace vials). Analysis
can be performed on-site or off-site, usually within 48
hours of collection. The quantitative results are typically
presented in concentration units (e.g., ppmv, µg/L vapor).
The active method requires actively withdrawing soil
vapor and is best suited to sites that have higher
vadose zone permeabilities and higher contaminant
concentrations (>0.1 µg/L vapor). The active method can
be used for investigating groundwater contamination,
and as discussed later in this chapter, the active
method is also useful for groundwater issues relating
to contaminant fate and transport because it gives
concentration data.
The passive approach consists of emplacing a natural
or synthetic absorbent into the subsurface (typically 0.15
to 0.9 meters) for an extended time, and subsequently
removing and analysing the absorbent. The absorbent is
typically placed in the upper end of an inverted container
or sleeve that has openings for migrating vapors to enter
and collect. Unlike the active method, which requires one
visit to the site, the passive method requires one visit
to the site to deploy the absorbents and a second trip to
retrieve them.
The passive method is qualitative and does not allow
for real-time data. Measured values are reported as total
absorbed mass (e.g., µg) or relative units because the
amount of vapor meeting the absorbent is unknown.
The passive method offers more sensitivity than the
active method; hence it offers advantages in sites of
low vadose zone permeability and lower contaminant
concentrations (>0.01 µg/L vapor). Contaminants must
have an appreciable vapor pressure to be detected by this
method, regardless of the deployment time or contaminant
concentration in the soil or water. As a rule of thumb, one
should not expect compounds whose vapor pressures are
less than 10−6 atmospheres to be detectable by passive soil
vapor methodology in a reasonable time frame (<50 days
of exposure) for reasonable contaminant concentrations in
the water.
Flux chambers consist of an enclosed chamber that
is placed on the surface for a specific period of time.
Vapor concentrations are measured in the chamber
through time or after a period of time. This method
is also quantitative and yields both concentration data
in the chamber and flux data (mass/area-time). Flux
chambers are the least common soil vapor method and
are seldom used for groundwater assessment. Their most
useful application is for sites where volatilization of
contaminants from groundwater is of concern and direct
fluxes are desired Hartman (9).
549
Analysis of Soil Vapor Samples
The type of analytical instrumentation employed depends
upon the goals of the project and the type of soil vapor
method used. For screening-level active soil vapor surveys,
in which the primary goal is simply to find contamination
qualitatively, simple handheld or portable instruments,
such as organic vapor analyzer (OVAs), organic vapor
meter (OVMs), gas-monitoring meters (e.g., IR detectors),
and gas detector tubes, can be used. For applications
where quantitative results are desired, the analytical
methodology employed is typically gas chromatography
(GC) or gas chromatography-mass spectrometry (GC-MS).
These latter techniques differentiate a wide variety of
compounds. These latter methods are also used to analyze
passive soil vapor samples and flux chamber samples.
More comprehensive treatments of the available
analytical options for soil vapor samples are given by
Hartman & Jacobs (5) and the EPA (1).
APPLICATIONS OF SOIL VAPOR DATA TO
GROUNDWATER ISSUES
This section discusses how to use and interpret soil vapor
data for applications specifically related to groundwater.
Assessing and Defining Groundwater Contamination
Soil vapor surveys have frequently been used for initial
site assessment to identify groundwater contamination,
and the data are used to focus subsequent groundwater
investigatory sampling programs (1,3,4). The soil vapor
data are typically plotted as contours or as a raised
surface. More sophisticated approaches, including three-
dimensional visualizations of the soil column, cross
sections, normalization, and statistical analysis, are
sometimes used depending on the project goals.
The applicability of a soil vapor survey to detect
potential groundwater contamination depends upon a
number of factors, including the local geology, contaminant
properties, and the concentration of the contaminant in
the groundwater (11). The maximum concentration of a
contaminant in the soil vapor due to volatilization from
the groundwater can be estimated from Henry’s law:
Csg = HCw (1)
where Csg is the soil gas concentration in equilibrium
with the groundwater,
Cw is the groundwater concentration, and
H is the dimensionless Henry constant.
As an example, consider groundwater that has a benzene
concentration of 10 µg/L. The dimensionless Henry
constant for benzene is 0.22, giving an equilibrium soil
vapor concentration of 2.2 µg/L vapor at the groundwater
interface. Hence, a soil vapor survey conducted near
groundwater could at best be expected to detect only a
maximum value of 2.2 µg/L in the vapor if groundwater
were the only source.
Equilibrium phase partitioning is rarely achieved
in the subsurface, so measured soil vapor values
550 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS

from groundwater volatilization will be significantly
lower (generally one to two orders of magnitude) than
predicted from Henry’s constants. In addition, soil vapor
concentrations decrease as distance increases from the
groundwater due to other processes active in the vadose
zone, such as biodegradation and sorption onto soils (13).
For these reasons, the chances of detecting groundwater
contamination from soil vapor data increases the closer
the samples are collected to groundwater.
Due to its increased sensitivity for some compounds,
the passive soil vapor method can offer advantages
over the active method for this application. Similar to
the active method, the ability of passive methods to
detect groundwater contamination increases the closer the
samples are collected to groundwater. However, regardless
of sampling depth, the capability of passive methods to
detect contaminants is still limited by the vapor pressure
of the compound of concern (5).
Transport of Contaminant Vapors in the Subsurface
To understand the source of soil vapor and groundwater
contamination, it is necessary to understand how contam-
inants move in the vapor phase. A condensed version of
this topic is addressed in this section beginning with a dis-
cussion of the processes by which vapors move through the
vadose zone. A more detailed presentation can be found in
Hartman (6,7,14).
There are two types of physical processes by which
contaminants are transported in the vapor phase. The
first is advection, the bulk movement of the vapor itself.
In advective transport, any contaminants in the vapor are
carried along with the moving vapor. Advective transport
processes can be important in soil vapor, for example, near
the surface due to atmospheric pressure differences (15),
near buildings that can create pressure gradients (16), and
movement of vapor by density-induced flow (17).
The second type of transport process is gaseous
diffusion, the motion of contaminants by molecular
processes through a nonmoving vapor column, which is
the primary mechanism for contaminant transport in the
vapor phase through the vadose zone (18).
Contaminant transport by gaseous diffusion is
described by Fick’s first law as
gradient with respect to the depth below the surface.
Diffusional transport is a three-dimensional process, so
contaminants move away from a source in all directions,
similar to an expanding balloon.
Differentiating the Source of Soil Vapor/Groundwater
Contamination
A common problem confronted by environmental geolo-
gists is determining the source of contamination at a site
that has both soil vapor and groundwater contamination.
Is the soil vapor responsible for the groundwater contami-
nation, or is the groundwater responsible for the soil vapor
concentration? Contaminant transfer by gaseous diffusion
moves from areas of high concentration to areas of low con-
centration, so vertical profiles of the soil vapor through the
vadose zone down to groundwater can provide this answer.
Figure 1 shows the expected shape of soil vapor profiles
in the vadose zone for three common conditions. Processes
active in the vadose zone such as sorption, biodegrada-
tion, and advection may alter these ideal profiles, but the
source of contamination must still be in the area of highest
concentration, as dictated by Ficks law.
Vertical soil vapor is profiled by installing vapor
monitoring wells at multiple depths through the vadose
zone to the groundwater. For soil vapor, vapor wells
are often constructed of small diameter 0.5 to 1.0 cm
(1/8 in. to 1/4 in. OD) flexible tubing and ‘‘nested’’ in the
same borehole (Fig. 2).
Repeated sampling of the vapor wells allows averaging
any temporal effects. Thomson (19) describes two case
histories of vertical profiling using nested vapor wells.
Contaminant Vapor Transport Rates
The distance that contaminant vapors can move by
gaseous diffusion can be approximated from
Distance = (2 × De × t)1/2 (3)
where De is the effective diffusivity and t is time.

De Csg Depth
Surface source
FLUX = (2) Flux
X

where Flux is the rate of movement of a compound

per unit area,
Flux
De is the effective diffusion coefficient in the Depth Deep source
vadose zone,
Csg is the contaminant concentration gradient in
the soil vapor, and
X is depth in the vadose zone
Depth Flux Surface and deep sources
Similar to momentum transfer (e.g., water running
downhill) and heat transfer (movement from hot areas
to cold areas), contaminant transfer by gaseous diffusion
moves from areas of high concentration to areas of low Increasing concentration
concentration. The flux is always down the concentration Figure 1. Common soil vapor profiles in the vadose zone for
gradient, regardless of the orientation of the concentration several one-dimensional situations (8).
 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS 551

laterally underneath a neighboring room or building or

downward to the groundwater surface.

Into or out of Groundwater. In contrast to movement

through the vadose zone where contaminant vapors move
Surface cover
J-plug through the soil vapor, the movement of contaminant
vapors into or out of groundwater is controlled by the
Bentonite Concrete rate at which vapors partition into and move through the
liquid. Because groundwater movement is so slow, the
18" to 24" long water interface remains relatively undisturbed (laminar
2" diameter PVC flow), and vertical mixing of the water is minimal. The
pipe primary exchange process is again molecular diffusion,
Third soil gas but in this case, the exchange rate is controlled by liquid
implant #10/20 silica diffusion, not gaseous diffusion. A general value for the
sand pack liquid diffusion coefficient for compounds is approximately
0.00001 cm2 /s. Using the same factor of 10 reduction due
Bentonite to soil porosity, De for most liquids can be approximated
as 0.000001 cm2 /s.
The distance that contaminants can move into and
through groundwater in a year can be estimated as

Distance = (2 × 0.000001 cm2 /s × 31,536,000 s)1/2

Second soil #10/20 silica
gas implant sand pack ∼ 8 cm

These calculations show that although contaminant

Bentonite
First soil
gas implant #10/20 silica
sand pack
2" diameter
borehole
Figure 2. Schematic diagram of a multidepth, nested vapor
monitoring well (19).
Through the Vadose Zone. For contaminant vapor
transport through the vadose zone, the effective diffusion
coefficient is the gaseous diffusion coefficient corrected
for soil porosity. For many vapors, the gaseous diffusion
coefficient is approximately 0.1 cm2 /s. The effect of soil
porosity varies depending upon the type of soil, and several
equations are available to calculate the effect of porosity
on diffusivity. A general approximation is that the porosity
reduces the gaseous diffusivity by a factor of 10. Thus, for
vapors, De can be approximated as 0.01 cm2 /s.
The distance that contaminant vapors can move
through the vadose zone in a year can be estimated as
vapors can move through the vadose zone relatively
quickly, they partition into and move through groundwater
extremely slowly. The reverse situation is also true;
the movement and partitioning of contaminants through
groundwater and into soil vapor is also extremely slow.
Estimation of the Downward Transport of Contaminant
Vapors into Groundwater
The calculations summarized in the preceding sections
indicate that although contaminant vapors can move
quickly down to groundwater, they do not partition into
groundwater very quickly because phase partitioning
across a liquid laminar interface is kinetically controlled.
For this reason, phase partitioning rarely reaches
equilibrium, and the use of Henry’s constants to estimate
groundwater concentrations from soil vapor data (or the
reverse) often gives estimated values one to two orders of
magnitude from measured values.
Using a modification of Fick’s first law, the transfer
rate of a contaminant from soil vapor into groundwater
can be estimated, and in turn, reasonable estimates of
the expected contaminant concentration in groundwater
due to contamination in the overlying soil vapor can
be estimated.
The flux is calculated as
Flux = KL (Csg /H − Cw ) (4)

Distance = (2 × 0.01 cm2 /s × 31,536,000 s)1/2
∼ 800 cm
This calculation shows that contaminant vapors can move
long distances through the vadoze zone in a short period of
time. Within a few years, vapor contamination can move
where KL is the gas exchange coefficient (length/time),
Csg is the soil gas concentration time
(mass/volume),
H is the dimensionless Henry’s law constant, and
Cw is the groundwater concentration
(mass/volume).
552 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS

The ratio Csg /H represents the groundwater concentration
in equilibrium with the overlying soil vapor at the interface
between the soil vapor and groundwater.
The gas exchange coefficient has units of velocity
and represents the distance that contaminants move
through the groundwater per unit time. This coefficient
is highly dependent on a compound’s effective diffusivity
and advective mixing of the groundwater interface.
For groundwater velocities less than 30.5 meters per
year and soil grain sizes less than 0.5 mm, diffusional
transport dominates over dispersion, and KL can be
approximated as:
KL = 1.1 (De v/d)1/2 (5)
where De is the effective diffusivity,
d is the horizontal distance (downgradient) over
which the soil vapor plume and groundwater
are in contact, and
v is the horizontal groundwater flow velocity.
For the groundwater conditions described above, the
interface remains laminar (i.e., no turbulent mixing), and
the gas transfer coefficient is approximately 8 cm/yr. Using
this value for the gas exchange rate, estimates of the
expected contaminant concentration in groundwater due
to contamination in the overlying soil vapor for various soil
vapor concentrations are summarized in Table 1. These
calculations assume equilibrium partitioning at the soil
vapor/groundwater interface, transfer into groundwater
by molecular diffusion only, and uniform mixing of
the contaminant into the groundwater over the well
screen length.
Table 1 shows that for equilibrium concentrations
(Csg /H) up to 100 µg/L, the resulting ground water
concentration after 5 years will be low (<10 µg/L) if liquid
molecular diffusion is the only exchange process.
Importance of Downward Vapor Transport for Several
Common Contaminants
The potential importance of downward vapor migration for
several of the more common compounds encountered by
environmental consultants are given special consideration
in this section.
Leaking Vapors from Chlorinated Solvents (Dry Clean-
ers and Vapor Degreasers). Many compounds, primarily
chlorinated solvents, have relatively high vapor pressures
and vapor densities three to six times greater than air.
Because of these physiochemical properties, vapors may
emanate from containers or pipes holding gaseous or liq-
uid products. If air flow is restricted, such as in a closed
room containing a washer unit, degreaser, or tank, leak-
ing dense vapors can penetrate the floor and enter the
upper vadose zone. Density-driven advection can continue
to drive vapor downward through the vadose zone until
the vapor is diluted to low enough concentrations (<1%)
that density-driven advection is no longer an important
factor in vapor transport and molecular diffusion takes
over. Vapor clouds reaching tens of feet into the upper-
most vadose zone have been documented in thousands
of cases. Businesses and commercial operations that are
most susceptible to this situation are those that deal with
chlorinated solvents (dry cleaners, vapor degreasers, spray
facilities, etc.).
The soil vapor concentration to yield an equilibrium
water concentration of 100 µg/L is 35 µg/L vapor for
TCA, 60 µg/L vapor for PCE, and 100 µg/L vapor for
vinyl chloride. In areas of groundwater deeper than 9.1
meters, soil vapor concentrations for these compounds are
generally less than these values at groundwater, unless in
the presence of contaminated soil, suggesting that surface
vapor leaks are not a significant contaminant pathway
to groundwater in most cases. In areas of groundwater
shallower than 6 meters, soil vapor concentrations
exceeding these values are found, even in the presence
of clean soil, suggesting that leaking dense vapors from
the surface may be a significant contaminant pathway.

Leaking, Hydrocarbon Fuel Vapors. In contrast to

chlorinated solvents, aboveground vapor releases are
Table 1. Expected Contaminant Concentration in
Groundwater for Various Soil Vapor Concentrations at not common for petroleum hydrocarbons because they
the Groundwater Interfacea,b rarely are stored in confined spaces (due to their
flammability). However, subsurface vapor leaks from
Water conc Water conc petroleum hydrocarbons are possible from USTs and
Csg , Csg /H, Flux, at 1 yr, at 5 yrs,
piping associated with them (vent pipes, pipe joints,
µg/L µg/L µg/cm2 -yr µg/L µg/L
and tank bungs. Gasoline vapor contains approximately
10 40 315 0.63 3.2 8000 µg/L vapor of benzene and lower, although still high
20 80 631 1.3 6.3 concentrations (>1000 µg/L vapor) of toluene and xylene.
30 120 946 1.9 9.5 Reformulated gasolines have very high concentrations of
40 160 1261 2.5 13 MtBE or ethanol, and their vapors contain approximately
50 200 1577 3.2 16 130,000 µg/L vapor of MtBE and 11,500 µg/L vapor of
100 400 3154 6.3 32
ethanol. Due to these high concentrations, leaking vapors
500 2000 15,768 31 158
1000 4000 31,536 63 315
must be considered a potential source of groundwater
contamination, as further described below.
a
Reference 5.
b
Assumptions used to compute the tabulated values: Benzene and Light Aromatics (BTEX). For light aromat-
Henry’s law constant: 0.25 (dimensionless)
Gas transfer coefficient: 8 cm/yr ics, the soil vapor concentration to yield an equilibrium
Well screen interval: 5 m water concentration of 100 µg/L is approximately 25 µg/L
 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS
vapor. In areas of groundwater deeper than 9.1 meters, soil
vapor concentrations of aromatics are generally less than
25 µg/L vapor at the groundwater, unless in the presence
of contaminated soil from a leaking UST, suggesting that
vapor leaks are not a contaminant pathway in most cases.
In areas of groundwater shallower than 3 meters, soil
vapor concentrations exceeding 25 µg/L vapor have been
found even in the presence of clean soil, suggesting that
leaking gasoline vapors from the UST are a possible
contaminant pathway. In areas where the groundwater
is near or above the UST, vapor leaks could be very
significant in groundwater contamination.
MtBE and Ethanol. Due to extremely low Henry
constants, the soil vapor concentration to yield an
equilibrium water concentration of 100 µg/L is only 2 µg/L
vapor for MtBE and approximately 0.02 µg/L vapor for
ethanol. Hence, leaking gasoline vapors from a UST
are a likely contaminant pathway to groundwater for
both compounds in shallow and deep groundwater. The
likelihood is even more acute in locations where the water
table is near or above the UST and/or in areas that have
significant amounts of rainwater infiltration.
Summary. In areas of deeper groundwater (>10 m),
slow groundwater flow rates less than 30.5 meters per
year and low recharge, soil vapor concentrations for most
contaminants at the groundwater surface are typically
well below equilibrium values (Csg /H) of 100 µg/L.
Hence contamination of groundwater by downward vapor
transport is not likely to be significant. For compounds
with very low Henry constants (MtBE, ethanol), downward
vapor transport could be important, even under these
groundwater conditions.
In areas where groundwater is shallower than 10
meters or higher groundwater flow rates and high
recharge, soil vapor concentrations for many contaminants
at the groundwater surface may approach high enough val-
ues that contamination of the groundwater by downward
vapor transport may be significant. This potential will be
enhanced for compounds with low Henry constants and
when vapor leaks occur directly into the groundwater.
Additional discussion of the downward migration of
vapors can be found in Hartman (7,14).
The Upward Migration of Vapors from Groundwater
Devolatilization
Attention has been raised in the past several years to
the risk to human health by the upward migration of
contaminant vapors into buildings and other enclosed
spaces (20–22). In cases of contaminated groundwater, it
is common to assess this risk pathway before site closure
is obtained.
Simplified, the upward contaminant vapor flux into a
building is computed by Fick’s first law, requiring the
soil vapor concentration at some depth underlying the
structure. Soil vapor data may be measured directly,
or alternatively, in the absence of actual soil vapor
data, soil vapor concentrations can be calculated from
groundwater data assuming equilibrium conditions using
Henry’s constants (23). As discussed in the preceding
553
sections, soil vapor values calculated from groundwater
using Henry’s constants are likely to be overestimated
by amounts as large as 10 to 100 times; hence the
resulting risk will also be overestimated. The potential
error in the calculated risk introduced by the incorrect
vapor concentration can be orders of magnitude greater
than errors introduced using default values for other
parameters, such as porosity. Thus, if a site fails the
upward risk calculation using calculated soil vapor values
from groundwater data, direct measurement of actual soil
vapor concentrations will yield more representative soil
vapor values to be used in the calculation (13).
When groundwater is extremely shallow (1.5 meters)
and representative soil vapor data are unobtainable, the
use of surface flux chambers may yield the best estimate
of the vapor flux (24).
A more detailed discussion of the upward vapor migra-
tion risk pathway, including a recommended protocol for
determining the upward vapor flux, can be found in Hart-
man (8).
BIBLIOGRAPHY
1. EPA. (1997). Expedited Site Assessment Tools For Under-
ground Storage Tank Sites. U.S. Government Printing Office,
Washington, DC.
2. Marrin, D.L. and Kerfoot, H.B. (1988). Soil gas surveying
techniques. Environ. Sci. Technol. 22 (7): 740–745.
3. Ullom, W.L. (1995). Soil gas sampling. In: Handbook of
Vadose Zone Characterization and Monitoring. L.G. Wilson,
L.G. Everett, and S.J. Cullen (Eds.). Lewis, Boca Raton, FL,
pp. 555–567.
4. Devitt, D.A., Evans, R.B., Jury, W.A., Starks, T.H., Eklund,
B., and Gholson, A. (1987). Soil gas sensing for detection
and mapping of volatiles organics, EPA/600/8-87/036. (NTIS
PB87-228516).
5. Hartman, B. and Jacobs, J. (2000). Applications and inter-
pretation of soil vapor data to volatile organic compound
contamination. In: Standard Handbook of Environmental
Science, Health, and Technology. J. Lehr and J. Lehr (Eds.).
McGraw-Hill, New York.
6. Hartman, B. (1999). Applications and interpretation of
soil vapor data. In: Chlorinated Solvent Contamination:
Legal and Technical Issues. Morrison et al. (Eds.). Argent
Communications Group, Foresthill, CA.
7. Hartman, B. (1998). Applications and interpretation of soil
vapor data. In: Petroleum Hydrocarbon Contamination:
Legal and Technical Issues. Morrison et al. (Eds.). Argent
Communications Group, Foresthill, CA.
8. Hartman, B. (1997). The upward migration of vapors.
LUSTLine Bull. 27, NEIWPCC, Wilmington, MA, December
1997.
9. Hartman, B. (2003). How to collect reliable soil gas data
for risk based applications. Part 2: Flux chamber method.
LustLine # 44, NEIWPIC, August 2003.
10. Hartman, B. (2002). How to collect reliable soil gas data
for risk based applications. Part 1: Active soil gas method.
LustLine # 42, NEIWPIC, October 2002.
11. Hartman, B. and Jacobs, J. (2000). Soil vapor principles. In:
Standard Handbook of Environmental Science, Health, and
Technology. J. Lehr and J. Lehr (Eds.). McGraw-Hill, New
York.
554 GROUNDWATER VELOCITIES
12. ASTM. (1993). Standard guide for soil gas monitoring in the
vadose zone, D5314-93. Annual Book of ASTM Standards.
Philadelphia, PA.
13. McNeel, P.J. and Dibley, V. (1997). Case study comparisons
of vapor inhalation risk estimates: ASTM RBCA model
predictions vs site specific soil vapor data. Proc. 1997
Superfund XVIII Conf., Washington, DC, December 1997,
pp. 509–516.
14. Hartman, B. (1998). The downward migration of vapors.
LUSTLine Bull. 28, NEIWPCC, December 1998.
15. Massman, J. and Farrier, D.F. (1992). Effects of atmospheric
pressures on gas transport in the vadose zone. Water Resourc.
Res. 28 (3): 777–791.
16. Little, J.C., Daisey, J.M., and Nazaroff, W.W. (1992). Trans-
port of subsurface contaminants into buildings. Environ. Sci.
Technol. 26 (11): 2058–2066.
17. Cherry, J. and Smyth, D. (1996). Non-aqueous phase liquids
in the subsurface environment: Assessment & remediation.
Am. Soc. Civ. Eng., 1996 Annu. Convention & Exposition,
Washington, DC, November 10–14, 1996.
18. Thorstenson, D. and Pollock, D. (1989). Gas transport in
unsaturated porous media: The adequacy of Fick’s law. Rev.
Geophys. 27: 61–78.
19. Thomson, K.A. (1995). Soil gas sampling. In: Handbook of
Vadose Zone Characterization and Monitoring. L.G. Wilson,
L.G. Everett, and S.J. Cullen (Eds.). Lewis, Boca Raton, FL,
pp. 569–588.
20. Jury, W.A., Russo, D., Streile, G., and El Abd, H. (1990).
Evaluation of volatilization by organic chemicals residing
below the surface. Water Resourc. Res. 26: 13–20.
21. Johnson, P.C. and Ettinger, R.A. (1991). Heuristic model for
predicting the intrusion rate of contaminant vapors into
buildings. Environ. Sci. Technol. 25: 1445–1452.
22. Jeng, C.-Y., Kremesec, V.J., and Primack, H.S. (1996). Mod-
els of hydrocarbon vapor diffusion through soil and transport
into buildings. In: Proc. 1996 Pet. Hydrocarbons & Org. Chem.
Ground Water: Prevention, Detection, and Remediation, Hous-
ton, TX, November 1996.
23. ASTM. (1995). Standard guide for risk-based corrective action
(RBCA) applied at petroleum release sites, E-1739-95. Annual
Book of ASTM Standards. Philadelphia, PA.
24. Smith, J.A., Tisdale, A.K., and Cho, H.J. (1996). Quantifica-
tion of natural vapor fluxes of trichoroethen in the unsatu-
rated zone at Picatinny Arsenal, New Jersey. Environ. Sci.
Technol. 30: 2243–2250.
GROUNDWATER VELOCITIES
MUKAND SINGH BABEL
Asian Institute of Technology
Pathumthani, Thailand
Groundwater flows through interconnected pore spaces in
subsurface soil in response to differences in fluid pressure
and elevation. Groundwater flow rates are extremely slow,
compared to surface water, and they are also highly
variable. For example, in clayey soils, groundwater moves
as slowly as a fraction of a millimeter per day; however,
in sandy soils, it moves faster, even up to more than
a meter per day. Hydrology is concerned with both
the speed and the direction of groundwater movement.
Groundwater velocity is very important in dealing with
contaminant migration and arrival time, mass flux, site
characterization, reactive barrier design, and calibration
of transport processes.
Hydraulic head drives flow through porous media. Two
main components contributing to the hydraulic head at
a given point in a groundwater flow system are pressure
head and elevation head, as shown in Fig. 1. The pressure
head is due to the porewater pressure at that point, which
can be measured by a piezometer. The elevation head is
due to gravitational energy. The higher the elevation of
a fluid from sea level or some selected datum, the more
gravitation potential energy it has. The third component,
kinetic energy, can be ignored, as groundwater flow
velocities are usually very small in magnitude. Thus, the
hydraulic head (h) is the sum of the pressure head () and
the elevation head (z), expressed usually in the units of
length (meters).
Henry Darcy, a French engineer who studied the
movement of water through sand in 1856, found that
the rate of water flow through a pipe filled with sand
is proportional to the difference in the height of the
water between the two ends of the tube, and inversely
proportional to the length of the tube (Fig. 2). The constant
of proportionality, K, is called hydraulic conductivity. This
law, universally known as Darcy’s law, is given as
Q = −KA[(hA − hB )/L]
or
Q = −KA(dh/dL) (1)
Groundwater
Ground surface
Depth to water
Pressure head (y)
Total head
Measuring point h=y+z
Elevation head (z)
Selected datum
Figure 1. Elevation head, pressure head, and total head
(adapted from Reference 1).
ha
ha − hb
hb
A B
Pipe filled with sand
L
Figure 2. Darcy’s law.

or
q = Q/A = −K (dh/dL)
where Q is the quantity of flow per unit of time (m3 /day);
K = hydraulic conductivity (m/day); A = cross-sectional
area (m2 ); hA and hB are the hydraulic head at points A
and B, respectively; and dh/dL = hydraulic gradient. The
term Q/A in Eq. 2 is also known as the specific discharge
(q), which is the flow per unit area [LT−1 ]. The specific
discharge is sometimes referred to as Darcy flux, Darcy
velocity, or bulk velocity.
The negative sign in Eq. 2 indicates that the fluid moves
in the direction of decreasing hydraulic head. Darcy’s law
assumes that the flow is laminar, that is, the water follows
distinct flow lines rather than mixing with other flow lines,
which is the case for most groundwater flow in porous
media. Most groundwater movements occur in small
interstices, so that there is considerable resistance to flow
imposed by the aquifer material, making the flow laminar.
When the flow velocity increases, especially in media
having large pores, the occurrence of turbulent eddies
dissipates kinetic energy, rendering the hydraulic gradient
less effective in inducing flow. In very large interstices,
such as those in many limestone and volcanic areas,
groundwater flow is almost identical to the turbulent flow
of surface water. Darcy’s law is not valid for turbulent flow,
for unusually high velocity that may be found in fractures
or in granular aquifers.
Equation 2 yields only an apparent velocity value, the
macroscopic velocity, through the cross-sectional area of a
solid matrix and interstices, which, however, is not the true
velocity of water moving through the pores. Clearly, flow
velocities through interstices alone will be much higher
than the macroscopic value. The interstices themselves
vary in shape, width, and direction, so the actual velocity
in the soil or rock is highly variable. Furthermore, the flow
path of a water particle around and between the grains
in an aquifer is tortuous, so water particles must travel
longer than the apparent distance given by the measured
length of the porous medium in the direction of flow, as
shown schematically in Fig. 3.
To calculate the true linear velocity, one must take
into account the fact that water travels only through the
pores of the sediment. The fraction of the area covered by
open space is equal to the effective porosity (ne ). Thus, the
effective velocity (v), also known as the porewater velocity,
Actual flow
paths in
pore spaces
Average
linear flow
path
Figure 3. Concept of average linear velocity (2).
VISCOUS FLOW 555
is obtained by dividing the macroscopic velocity by the
(2) effective porosity:
v = −(K/ne )dh/dL (3)
or
v = Q/ne /A
or
v = q/ne
From Eq. 3, it can be seen that, for sand with an effective
porosity of 33%, the actual velocity is three times the
Darcy velocity. Therefore, to define the actual velocity, one
must consider the microstructure of the material. Usually,
actual velocities tend to decrease with depth as porosities
and hydraulic conductivities also decrease.
BIBLIOGRAPHY
1. Ward, R.C. and Robinson, M. (2000). Principles of Hydrology,
4th Edn. McGraw-Hill, London.
2. Freeze, R.A. and Cherry, J.A. (1979). Groundwater. Prentice-
Hall, Englewood Cliffs, NJ.
VISCOUS FLOW
NAVRAJ S. HANSPAL
VAHID NASSEHI
Loughborough University
Loughborough, United
Kingdom
DIGANTA BHUSAN DAS
Oxford University
Oxford, United Kingdom
INTRODUCTION
This text describes the elementary concepts of viscous
fluid flow. The principal laws that govern the motion of
fluids are also discussed. From our common experience, we
know that fluids deform under the action of external forces.
The deforming forces and rate of deformation are related
to each other through a proportionality constant known
as the coefficient of viscosity. The value of this constant
depends upon the chemical identity, material properties,
and thermodynamic state of a given fluid. To develop
a theory that shows how viscosity affects fluid dynamics,
one must understand the relationships between stress and
the rate of strain. The outlines given in this article provide
basic concepts of viscous flow. For more detailed analysis,
the reader should consult specialized texts on this topic.
Some of these texts are listed in the references at the end
of this article.
VISCOSITY AND NEWTON’S LAW OF VISCOSITY
The physical property by which the resistance to fluid
flow can be characterized is known as viscosity. A more
556 VISCOUS FLOW
fundamental approach to viscosity shows that it is the
property of a fluid that relates applied stress to a
resulting strain rate. By experience, we come across many
different types of fluids. For example, consider water and
lubricating oil such as that used in automobiles. Water
can flow more freely compared to lubricating oil under
isothermal conditions. Hence, we say that the viscosity of
water is less than that of oil.
When we deform a solid, so that it is strained, we know
that the solid will exert a restoring force that opposes the
strain; for small strains, the restoring force is proportional
to the strain, the familiar Hooke’s law. Real fluids also
oppose strains; however, in a fluid it is not the amount of
strain that is important but the rate at which the strain
is produced.
Consider simple fluid flow between the two plates
shown in Fig. 1. If the force per unit area of the top
plate is F, we find that
F v
=µ (1)
A d
where d is the separation between the plates, which defines
the viscosity µ (sometimes called the absolute or dynamic
viscosity). Equation 1 can be written in a more generalized
form given by Eq. 2. The shear stress exerted in the x
direction on a fluid surface of constant y by the fluid in the
region of lesser y is designated as τyx , and the x component
of the fluid velocity vector is designated as vx :
dvx
τyx = −µ = µγ̇ (2)
dy
This equation is commonly known as Newton’s law of
viscosity. It states that the shear force per unit area is
proportional to the negative of the local velocity gradient.
The left-hand side of Eq. 2 represents the shear stress,
and the right-hand side represents the rate of strain γ̇ · τyx
from Eq. 2 can also be interpreted as the viscous flux
of the x momentum in the y direction. It can be seen
that the viscous momentum flux is in the direction of
the negative velocity gradient, that is, the momentum
tends to go in the direction of decreasing velocity. A
velocity gradient can thus be considered a driving force for
momentum transport.
Fluids in which shear stress is directly proportional to
the rate of deformation are termed Newtonian fluids. The
viscosity of such fluids is independent of the rate of shear
and rate of strain and always remains constant under
given conditions. Examples of Newtonian fluids are air,
v Fig. 2:
y d
x Bingham Plastics. Bingham plastics exhibit yield stress
Figure 1. Viscous drag between two plates, which the bottom
plate is held fixed.
water, and gasoline. In SI units, the units of viscosity are
kg/m · s or Pa · s. In the cgs system, the unit of viscosity
is g/cm · s, also called the poise (after Poiseuille) and
equals 0.1 Pa · s. In fluid mechanics, the ratio of absolute
viscosity to density is also termed kinematic viscosity
which is often denoted by η. Generally, the viscosity of
liquids decreases as temperature increases. Gases, on the
other hand, have viscosities which generally increase with
increasing temperature. More detailed explanations are
based on the kinetic theory of gases and can be found
in (1).
NON-NEWTONIAN FLUIDS
There are fluids that do not obey the simple relationship
between shear stress and shear strain rate given by
Eq. 2 for a Newtonian fluid. These fluids are known as
non-Newtonian fluids. Examples are paints, solutions of
various polymers, food products such as applesauce and
ketchup, emulsions and suspensions of various solids, and
coal slurries. The subject of non-Newtonian flow is actually
a subdivision of the science of rheology, which is also known
as the ‘‘science of deformation and flow.’’
We categorize non-Newtonian fluids and discuss them
under three general classifications:
1. The first is the time independent non-Newtonian
fluid in which the shear rate is a unique but
nonlinear function of shear stress.
2. In the second category are the time dependent non-
Newtonian fluids that have more complicated shear
stress and shear strain relationships. In these fluids,
the shearing rate is not a single valued function of
the shear stress. The shear rate depends on shearing
time or on the previous shear stress rate history of
that fluid.
3. In the third category are viscoelastic fluids. Unlike
a truly viscous fluid in which all its energy of
deformation is dissipated, some of the energy of
deformation of a viscoelastic fluid may be recoverable
as in the deformation of an elastic fluid.
Time-Independent Fluids
For time-independent non-Newtonian fluids,
γ̇ = f (τ ) (3)
A Newtonian fluid is a special case of this category where
the function f (τ ) is linear.
Time-independent non-Newtonian fluids can be com-
monly represented by three distinct rheograms shown in
1. Bingham plastics
2. pseudoplastics
3. dilatant fluids
at zero shear rate, followed by a straight-line relationship
between shear stress and shear rate. The characteristics

A B
Bingham plastic
Slope mp
t0
Pseudoplastic
Slope ma
Shear stress
Dilatant fluid
Slope m
Shear strain rate
Figure 2. Shear stress–strain rate relationships for non-New-
tonian fluids.
of such fluids are defined by two constants. The first is the
yield stress denoted by τ0 ; this is the stress that must be
exceeded for the flow to begin. The second is the plastic
viscosity µp which is the slope of the straight-line portion
on curve A, Fig. 2. The model for Bingham plastics can be
represented by Eq. 4:
τyx = µp γ̇ + τ0 if τyx > τ0
This model closely approximates the behavior of many
real fluids such as slurries, plastics, emulsions such as
paints, and suspensions of finely divided solids in a liquid.
However, different constitutive models representing this
type of fluids have been proposed by Herschel and
Bulkley (2), Oldyroyd (3), and Casson (4).
Pseudoplastic fluids. Pseudoplastics (curve B, Fig. 2)
do not have any yield stress. A pseudoplastic fluid is
characterized by a progressively decreasing slope of shear
stress versus strain rate. The slope has been defined as
apparent viscosity,
µa = τyx /γ̇ (5)
At very high shear rates, the apparent viscosity becomes
constant, and the shear stress versus shear strain becomes
linear. A number of empirical relations have been used to
describe the behavior of pseudoplastic fluids. The simplest
of these is the power law model proposed by Ostwald in
1923 [refer to (5)] and expressed in the form of Equation 6.
τyx = kγ̇
where n < 1 and k and n are constants for a particular
fluid. k is known as the consistency coefficient for a given
fluid, and n is the power law index. n also represents the
deviation of a particular fluid from a Newtonian fluid. For
a Newtonian fluid, n = 1, and k = µ, the viscosity of the
fluid. Combining Eqs. 5 and 6,
µa = kγ̇ (n−1)
VISCOUS FLOW 557
The value of n is greater than unity for pseudoplastic
D
Newtonian fluid fluids.
C
Dilatant Fluids. Dilatant fluids (curve C, Fig. 2) are
similar to pseudoplastics in having no yield stress. They
differ from pseudoplastics in a fashion such that the
apparent viscosity increases as shear rate increases. The
power law model given by Eq. 7 is also applicable to
dilatant fluids, and so are Eqs. 6 and 7. The only difference
here is that the power law index n is greater than unity.
Pseudoplastic and dilatant fluids are also known as shear
thinning and shear thickening fluids.
Time-Dependent fluids. Some fluids are more complex
than those just described, and the apparent viscosity
depends not only on the shear rate but also on the time
for which the shear has been applied. Such fluids can be
classified in two categories;
1. thixotropic fluids
2. rheopectic fluids
The shear stress decreases with time for a thixotropic
fluid, whereas the shear stress increases with time for a
rheopectic fluid. Rheopexy and thixotropy are complex
phenomena resulting from transient changes of the
molecular structure of time-dependent fluids under an
applied shear stress. The proposed models for such fluids
are hence based on many simplifying assumptions (6).
(4)
Thixotropic Fluids. The apparent viscosity of thixotropic
fluids depends on the length of time of shearing as well
as on the shear rate. As the fluid is sheared from the
state of rest, it breaks down (on a molecular scale), but
then the structural reformation increases with time. An
equilibrium state is achieved when the breakdown rate is
equal to the build up rate. If allowed to rest, the fluid builds
up slowly and eventually regains its original consistency.
As the shear is decreased, the apparent viscosity is
less than that for increasing shear. For more detailed
discussions, refer to (7).
Rheopectic Fluids. Molecular structure in rheopectic
fluids, is formed by shear, and the behavior is opposite
to that of thixotropy. A simple example of the formation
of shear structure is the beating and thickening of
egg whites. Many substances lose their rheopectic
properties at extremely high shear rates and behave as
thixotropic fluids.
Viscoelastic Fluids. A viscoelastic fluid exhibits both
viscous and elastic properties. The simplest type of such
a material is one which is Newtonian and obeys Hooke’s
law for the elastic part. The shear rate for such fluids can
hence be written as
γ̇ = τyx /µ0 + τ̇yx /λ (8)
where λ is the rigidity modulus. Under steady-state flow,
(7) γ̇ = τyx /µ0 and the fluid behaves like a simple Newtonian
558 VISCOUS FLOW
fluid. However, if the shear is changed, an elastic effect is
added. Maxwell first proposed Eq. 8 in the form
τyx + τ̇yx (µ0 /λ) = µ0 γ̇ (9)
Liquids that obey Eq. 9 are known as Maxwell liquids. The
constant (µ0 /λ)−1 is known as the relaxation time, which
is the time constant for the exponential decay of stress at
constant strain. If the motion is stopped, the stress relaxes
at e−tλ/µ0 . Rather complex models of viscoelastic materials
have been developed in which higher time derivatives of
τyx and γ̇ appear. A very readable text for these models
is (8).
MICROSCOPIC BALANCES OF MASS, MOMENTUM,
AND ENERGY
We define the control volume as a region in space
within which our fluid flow system is present. Laws
of conservation (or change) of mass, momentum, and
energy desribe the behavior of fluid flow within this
system. To yield the detailed descriptions of variables
of interest, the control volume must be of infinitesimal
dimensions that can shrink to zero, yielding a point
volume. This approach reduces the quantities to point
variables. The application of conservation principles to this
infinitesimal system produces microscopic or differential
conservation equations.
Shell Momentum Balances
Shell momentum balances are written to represent a fluid
flow problem mathematically. The momentum balance for
any fluid flow system can be written as
(Rate of momentum in) − (rate of momentum out)
+ (sum of forces acting on the system)
= (rate of momentum accumulation) (10)
Momentum can enter the system by momentum transfer
according to Newtonian (or non-Newtonian) expressions
for momentum flux. Momentum can also enter from overall
fluid motion, bulk transfer. The forces acting on a system in
fluid flow problems are pressure forces (acting on surfaces)
and gravity forces (acting on the entire volume as a
whole). The general procedure for solving any viscous
flow problem is
1. Write a momentum balance in form of Equation 10
for a shell of finite thickness, a microscopic shell
momentum balance.
2. Let this thickness approach zero, and use the
mathematical definition of the first derivative
to obtain the corresponding differential equation
describing the momentum flux distribution.
3. Insert the appropriate Newtonian or non-Newtonian
expression for momentum flux to obtain a differen-
tial equation for the velocity distribution.
4. Integrate these differential equations which yields
the momentum flux and velocity distributions for the
system under consideration.
Refer to (1) to get more acquainted with shell momentum
balance concepts, where this approach has been applied to
simple flow problems.
EULERIAN AND LAGRANGIAN COORDINATE SYSTEMS
Conservation equations and the constitutive equation
are expressed in terms of relative kinematics, velocities,
gradients of velocities, strains, and rates of strain, so it is
very important to quantify these variables.
Flow kinematics, the relative motion of fluid particles,
can be described by using either a Lagrangian or a
Eulerian description. In the Lagrangian system, the
motion of an individual particle is tracked where the
coordinates locate the position of a particle and are
functions of time. In the Eulerian system, dependent
variables such as the velocity vector and pressure are
considered functions of fixed spatial coordinates and
of time.
The Lagrangian framework is considered a more
natural choice to represent the actual kinematics and
stresses experienced by fluid particles. However, to solve
complex flow problems, this approach is limited (e.g.,
due to the necessity of tracking particles along unknown
streamlines). This system is particularly convenient in
flows of viscoelastic fluids with memory that require
particle tracking and calculation of deformation and
stresses along streamlines. The Eulerian formulation
is, in general, more convenient to use because it deals
only with local or present kinematics. In most cases,
all variables of interest, such as strain, rate of strain,
stress, and streamlines, can be obtained from the velocity
field. An additional advantage of the Eulerian system is
that it involves only time as a variable in unsteady-state
flow, whereas the Lagrangian system uses traveling time
as a variable, even in steady-state flow. All equations
described in subsequent sections have been represented in
the Eulerian coordinate system.
EQUATIONS OF CHANGE
The shell balance approach was used to acquaint
the beginner with the application of the principle of
conservation of momentum to viscous flow problems. It is,
however, not always necessary to formulate a momentum
balance. It is quicker and easier to start with the equations
of continuity and motion in general form and then to
simplify them to fit the problem in hand. These two
equations describe all problems of the viscous flow of a
pure isothermal fluid.
EQUATION OF CONTINUITY
The equation of continuity is developed by applying the
law of conservation of mass to a small 3-D volume element
within a flowing fluid, as shown in Fig. 3.
(Rate of mass accumulation)
= (rate of mass in) − (Rate of mass out) (11)
 VISCOUS FLOW 559

We consider each of the six faces of this volume element analogously. Momentum flows in and out of the control
and evaluate the mass entering and exiting the system: volume by two mechanisms; convection, which is a result
of the bulk fluid flow and molecular transfer (i.e., viscous
Rate of mass in = yzρvx + xzρvy transfer), that is, by virtue of velocity gradients.
x y
The important forces acting on the cubic element in
+ xyρvz (12) Fig. 4 are the fluid pressure p and the gravitational force
z gx . Hence, to write Eq. 10 in mathematical form,
Rate of mass out = yzρvx + xzρvy
x+x y+y Rate of momentum in = Rate of momentum in by
+ xyρvz (13) (convection + viscous transfer)
z+z

∂ρ = yzρvx vx + xzρvy vx
Rate of mass accumulation = xyz (14) x y
∂t
+ xyρvz vx + yzτxx
After inserting, Eqs. 12–14 in Eq. 11, dividing the entire z x
equation by xyz, and taking the limit of dimensions
+ xzτyx + xyτzx (17)
approaching zero, we get the equation of continuity. y z
For generality, we have expressed the equation of
Rate of momentum out = Rate of momentum out by
continuity in vector form:
(convection + viscous transfer)
∂ρ
= −(∇ · ρv) (15) = yzρvx vx + xzρvy vx
∂t x+x y+y

Here (∇ · ρv) is called the ‘‘divergence’’ of ρv. Vector ρv + xyρvz vx + yzτxx

is the mass flux and is the net rate of mass efflux per z+z x+x
unit volume. Equation 15 states that the rate of increase
+ xzτyx + xyτzx (18)
of density within a small volume element fixed in space is y+y z+z
equal to the net rate of mass influx to the element divided Sum of forces acting = body forces + pressure
by its volume. Note that the derivation can be performed  
for a volume element of any shape. = yz p − p
x x+x
A very special form of the equation of continuity is
Eq. 16, which is applicable to fluids of constant density, + ρgx xyz (19)
incompressible fluids.
Rate of accumulation of x
(∇ · v) = 0 ∂ρvx
(16) component of momentum = xyz (20)
∂t

EQUATION OF MOTION These stresses developing in viscous transfer can be

Using the shell momentum balance approach for the 3-D
volume element shown in Fig. 5, an equation of motion
is developed. It should be emphasized that momentum is
a vector quantity, and hence the equation of motion has
components in each of the three coordinate directions x, y,
and z, respectively. For simplicity, we consider only the x
component here; y and z components can then be handled
divided into normal and shear stresses. For example, τxx is
the normal stress, whereas τyx is the shear stress resulting
from viscous forces. At each point in a continuous medium,
whether it is solid or fluid, we need six components; each
represents a component of force per unit area to define
(tzx)z + ∆z

z (tyx)y + ∆y
z

∆z (txx)x + ∆x
(rnx)x (txx)x
∆z (rnx)x + ∆x

y y
∆y
∆y
∆x ( x,y,z ) ∆x
(x,y,z)
(tzx)z
(tyx)y
x x
Figure 3. Region of volume in xyz through which fluid Figure 4. Direction of transfer of x component of momentum
is flowing. through the surfaces.
560 VISCOUS FLOW


tyy ∂vz ∂vx
τzx = τxz = −µ + (27)
tnx ∂x ∂z

txy Equations 22–27 have been presented with proof in (9)

and can be applied to various complex situations.
Equations 22–24 are normal stresses and should contain
txx an additional term. For example, Eq. 22 should be
txx
written as


∂vx 2
txy τxx = −2µ + µ − κ (∇ · v) (28)
∂x 3
tnx
Similarly, Eqs. 23 and 24 are also written as


∂vx 2
tyy τxx = −2µ + µ − κ (∇ · v) (29)
∂x 3
Figure 5. Local stresses acting on a cubic element in an

∂vx 2
x–y plane. τxx = −2µ + µ − κ (∇ · v) (30)
∂x 3

the local stress completely. The local stresses acting in the κ is known as ‘‘bulk viscosity.’’ The bulk viscosity is zero for
x–y plane are shown in Fig. 5. low-density monoatomic gases and not too important for
Using Eqs. 17–20, dividing the entire equation by dense gases and liquids which are almost incompressible.
xyz, and taking the limit of the dimensions Experimental measurements to estimate the value of
approaching zero, we get the x component of the equation ‘‘bulk viscosity’’ have been discussed by Karim and
of motion. Analogously, we can find the components for Rosenhead (10). Formulas to estimate the value of bulk
the y and z components. Adding the three components density can be found in (11). When the fluid flows in the
of equation of motion, the equation of motion can be x direction between two parallel plates perpendicular to
expressed in vector notation as the y direction and vx is only a function of y, then this
set of six equations reduces to Eq. 2 described previously.
∂ Equations 22–27 inserted in Eq. 21 yield the equations
ρv = [∇ · ρvv] −[∇p]
∂t of motion for a Newtonian fluid with varying density
(21) and viscosity.
Rate of increase Rate of momentum Pressure on
of momentum gain by convection element per 1. For constant ρ and µ, the equation of motion can be
per unit volume per unit volume unit volume written as


−[∇ · τ ] +ρg ∂v
ρ + v · ∇v = −∇p + µ∇ 2 v + ρg (28)
∂t
Rate of momentum Gravitational
gain by viscous force on Equation 28 is the well-known Navier–Stokes equations,
transfer per unit element per first developed by Navier in France in 1822.
volume unit volume 2. When viscous effects are ignored, Eq. 28 reduces to


τ is known as the ‘‘stress tensor’’, [∇ · ρvv] represents the ∂v
ρ + v · ∇v = −∇p + ρg (29)
rate of loss of momentum per unit volume by fluid flow, and ∂t
(∇ · ρv) represents the rate of loss of mass per unit volume.
To use Eq. 21 to determine the velocity distribution, we Equation 29 model laminar flow when viscous effects
must insert expressions for various stresses in terms of are negligible. These flows are termed Euler, potential,
velocity gradients and fluid properties. For Newtonian or inviscid flows. Equation 29 is the well-known Euler
fluids, these expressions can be written as equation.
3. When the motion of a fluid is ‘‘very slow,’’ the flow
∂vx 2 is said to be creeping or Stokes flow. Viscous forces
τxx = −2µ + µ(∇ · v) (22)
∂x 3 dominate creeping flows and hence, the nonlinear
∂vy 2 inertia term in Eq. 28 is neglected. Then,
τyy = −2µ + µ(∇ · v) (23)
∂y 3
∂v
∂vz 2 ρ = −∇p + µ∇ 2 v + ρg (30)
τzz = −2µ + µ(∇ · v) (24) ∂t
∂z 3

 Equation 30 is the well-known Stokes equations.
∂vx ∂vy
τxy = τyx = −µ + (25)
∂y ∂x The equations of motion together with the equation of


∂vy ∂vz continuity, the equation of state p = p(ρ), viscosity mod-
τyz = τzy = −µ + (26)
∂z ∂y els illustrated in previous sections, and the intial and
 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES 561

the boundary conditions, complete the problem specifica- 5. Bird, R.B., Armstrong, R.C., and Hassager, O. (1977). Dynam-
tion for any viscous flow problem. These equations, when ics of Polymeric Liquids: Fluid Mechanics. John Wiley & Sons,
solved mathematically, determine completely the density, New York.
pressure, and velocity components in a flowing isother- 6. Slibar, A. and Paslay, P.R. (1959). Retarded flow of Bingham
mal fluid. materials. Trans. ASME 26: 107–113.
7. Tanner, R.I. (1985). Engineering Rheology. Clarendon Press,
Boundary Conditions Oxford, UK.
8. Wilikinson, W.L. (1960). Non-Newtonian Fluids. Pergamon
Boundary conditions are constraints imposed on conserva- Press, New York.
tion equations to describe how flow behavior conforms to
9. Lamb, H. (1945). Hydrodynamics. Dover, New York.
the surroundings. Boundary conditions come from nature
10. Karim, S.M. and Rosenhead, L. (1952). Rev. Mod. Phys. 24:
and are mathematical descriptions of the physics at the
108–116.
boundary. These may describe conditions along the bound-
11. Papanastasiou, T.C., Georgiou, G.C., and Alexandrou, A.N.
ary dealing with motion, external stresses, rate of mass
(2000). Viscous Fluid Flow. CRC Press, New York.
and momentum flux, boundary values of field variables
12. Hirschfelder, J.O., Curtiss, C.F., and Bird, R.B. (1954). The
as well as the relations among them. When the solution
Molecular Theory of Gases and Liquids. John Wiley & Sons,
involves the time evolution of flow fields, initial condi- New York.
tions are also needed. The number of boundary conditions
required to solve a particular problem is determined by the
Nomenclature
nature of the governing partial differential equation. In
general, elliptical equations require boundary conditions A Area
on each portion of the boundary; hyperbolic equations F Force
require boundary conditions at upstream but not at down- g Gravity force vector
stream boundaries; parabolic equations require initial n Power law index
conditions and boundary conditions everywhere except p Pressure
downstream. The Navier–Stokes equations are hyperbolic v Velocity vector
at high Reynolds numbers and elliptical at low Reynolds t Time variable
numbers. The Euler equation is the upper limit of hyper-
Greek Symbols
bolicity, whereas Stokes equations are the lower limit of
γ̇ Rate of strain
ellipticity. Different types of boundary conditions in math-
κ Bulk viscosity
ematical form have been described by Papanastasiou (12)
µ Viscosity of a fluid
and are applicable to a wide range of viscous flow problems.
τ Stress
In general, there are three types of boundary conditions:
τ̇ Rate of shear
λ Elasticity modulus
1. First kind or Dirichlet boundary conditions: The
ρ Density of a fluid
value of the dependent variable along the bound-
ary is imposed as the condition. Typical Dirich- Subscripts
let boundary conditions are no-slip boundary a Stands for apparent
conditions for velocity and the specification of p Stands for plastic
inlet/outlet velocities. o Stands for yield stress
2. Second kind or Neumann boundary conditions: x Stands for x component
The normal derivative of the dependent variable y Stands for y component
is specified along the boundary. Examples are z Stands for z component
symmetry conditions and free surface and interface
stress conditions.
3. Third kind or Robin’s boundary condition: The VULNERABILITY MAPPING OF
dependent variable and its normal derivative GROUNDWATER RESOURCES
are related. Slip wall conditions are typical
Robin conditions. THOMAS HARTER
University of California
Davis, California
BIBLIOGRAPHY

1. Bird, R.B., Stewart, W.E., and Lightfoot, E.N. (1994). Trans-

port Phenomena. John Wiley & Sons, New York.
2. Herschel, W.H., and Bulkley, R. (1926). Flow of Non-
Newtonian fluids. In: Encyclopaedia of Fluid Mechanics. Gulf
Publishers, Houston, TX, Vol. 9, Chap. 4.
3. Oldyroyd, J.G. (1947). A rational formulation of the equations
of plastic flow for a Bingham solid. Proc. Camb. Philos. Soc.
43: 100–105.
4. Casson, N. (1959). Rheology of Disperse Systems. Pergamon
Press, London.
WHAT IS GROUNDWATER VULNERABILITY?
Groundwater vulnerability is a measure of the likelihood
that pollution or contamination at the land surface reaches
a production aquifer. It is a measure of the ‘‘degree of
insulation’’ that natural and man-made factors provide
to keep pollution away from groundwater. The term was
first used in Europe in the 1960s. In defining groundwater
vulnerability, a number of issues must be examined and
562 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES
clarified; all refer to these two questions: vulnerability
of what? and vulnerability to what? Exact definitions of
groundwater vulnerability have therefore varied (1).
In principle, all groundwater is vulnerable to human
activity; no groundwater is completely isolated from the
aboveground environment. The degree of vulnerability
depends on environmental conditions, on how we define
groundwater and the part of groundwater in which we are
interested in (hence the question, ‘‘vulnerability of what?’’),
and on the timescale of interest. In addition, it depends
on whether or not a vulnerability measure is intended to
account also for the presence and type of pollutants (hence,
‘‘vulnerability to what?’’).
Vulnerability of What?
Groundwater is produced from a wide variety of hydroge-
ologic environments. When defining vulnerability, there-
fore, we need to define the portion of groundwater to
be assessed. Because the shallowest groundwater zone is
typically the most vulnerable, vulnerability assessments
are mostly concerned with the vulnerability of the upper-
most aquifer (in a multiaquifer system) or with the water
table (in an unconfined aquifer system). Deep groundwa-
ter is considered less vulnerable than shallow groundwater
because of the longer travel times necessary for a pollutant
to reach these deeper aquifers.
Some methods of assessing vulnerability therefore
account, directly or indirectly, for the travel time between
the water table and a well by considering aquifer
permeability or travel-time zones (e.g., 2). Timescales
are important in defining vulnerability. The practical
question is, should we consider groundwater vulnerable
to today’s activities only if the recharge time is a few
years or also if recharge timescales are tens, hundreds,
thousands, or tens of thousands of years? The potential
travel time between contaminant source and groundwater
or groundwater wells is an important consideration for
protection, monitoring, and assessment. Timescales are
also important when mapping vulnerability to specific
pollutants that become less harmful over time and with
travel distance due to radioactive decay, degradation,
inactivation (in the case of pathogens), filtration, or strong
sorption to soils and aquifer materials.
Vulnerability to What?
An assessment of groundwater vulnerability may or may
not include consideration of the presence of pollutants.
Vulnerability that is independent of whether or not con-
taminants (pollutants) are present and focuses primarily
on a description of natural environmental conditions is
often (though not always) referred to as ‘‘susceptibility,’’
‘‘natural vulnerability,’’ ‘‘aquifer sensitivity,’’ or ‘‘intrinsic
vulnerability.’’ Pollution-type-dependent vulnerability, or
vulnerability to specific land uses, is sometimes referred to
as ‘‘specific vulnerability’’ or ‘‘integrated vulnerability’’ (1).
WHY ASSESS GROUNDWATER VULNERABILITY?
The purpose of vulnerability assessments is to provide
a decision-making tool based on the best available data
and good scientific judgment. It is generally not intended
to be a scientifically rigorous analysis of groundwater
resources. The National Research Council (3) identified
four general objectives typically achieved by groundwater
vulnerability assessments: (1) to facilitate policy analysis
and development at the local and regional level, (2) to
provide program management, (3) to inform land use
decisions, and (4) to provide general education and
awareness of a region’s hydrogeologic resources.
The usefulness of vulnerability maps is limited (some
would say, arguable). The scientific literature lists pros
and cons to vulnerability assessments (1). The facts
that speak for implementing assessments are that land
use decisions have to be made, land management
practices need to be sensitive to the risk for groundwater
contamination, and not all anthropogenic activities can
be carried out isolated from groundwater. Timescales
and distinctions based on travel time are important,
at least with respect to some pollutants. Timescales
are also important in placing groundwater monitoring
wells and in scheduling and providing for planning and
action–response time if pollution does occur. So a need
exists to provide at least some general guidance to land
use planners, decision makers, and water users. Carefully
prepared vulnerability assessments allow them to make
decisions that are economically sensible and at the same
time hydrogeologically reasonable.
On the other hand, some may argue against the use
of vulnerability maps for land use planning because
groundwater flow conditions and the transport properties
of the subsurface are too complex to be appropriately
captured by any vulnerability tool. A second group of
arguments against vulnerability zoning arises from the
fact that all groundwater is vulnerable to land pollution.
The only geographically differing factor is the timescale
for the pollution to reach groundwater.
WHAT FACTORS DETERMINE VULNERABILITY?
We distinguish between intrinsic factors (natural con-
ditions) and extrinsic factors (anthropogenic activities).
Among intrinsic factors, the thickness and hydraulic prop-
erties of the geologic formations above the aquifer—the
unsaturated zone and confining layers above the
aquifer—are key factors in determining the vulnerabil-
ity of an aquifer system. They are the principal natural
controls that determine the recharge rate and recharge
time to the aquifer. The unsaturated zone also provides
key groundwater protection by
• intercepting, sorbing, and eliminating pathogenic
viruses and bacteria;
• sorbing and degrading many synthetic organic
chemicals; and
• attenuating heavy metals and other inorganic
chemicals through sorption and complexation with
mineral surfaces within the unsaturated zone and
through uptake into plants and crops (e.g., fertilizer).
The potential of the soil and the unsaturated zone to sorb,
degrade, or eliminate substances depends on the type of
 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES
pollutant and therefore is considered only in vulnerability
studies that are specific to certain land uses.
The amount of recharge at the land surface is
another important factor that determines vulnerability.
The higher the recharge, the higher the vulnerability
of the aquifer. If climatic conditions are such that little
or no recharge occurs at the land surface, downward
movement of moisture through the unsaturated zone is
very limited, regardless of the hydraulic properties of the
unsaturated zone.
In vulnerability assessments that encompass more than
just aquifer susceptibility (see above), the presence of
contamination sources or possible contaminating activities
is considered an important factor. The amount, spatial
distribution, and chemical properties of the contaminants
(possibly) released into the subsurface are highly specific
to individual sources and regions.
CHARACTERIZATION OF GROUNDWATER
VULNERABILITY
Numerous schemes have been developed for assessing and
mapping vulnerability. These methods can be grouped into
three major categories:
• index-and-overlay methods
• process-based computer simulations
• statistical analyses
Index-and-Overlay Methods
Index-and-overlay methods are based on assembling infor-
mation on the most relevant factors affecting aquifer
vulnerability (soil type, geologic formation type, recharge,
etc.), which then is interpreted by scoring, integrating, or
classifying the information to produce an index, rank, or
class of ‘‘vulnerability.’’ The scoring, ranking, and inte-
gration methods are based on expert opinion rather than
processes and are inherently subjective to some degree.
In the United States, the most prominent vulnerability
assessment method in this category is ‘‘DRASTIC’’ (2).
Others include the vulnerability analysis specified by Cal-
ifornia’s Drinking Water Source Assessment Program (4)
and the British ‘‘GOD’’ method (1). A brief overview of
DRASTIC and GOD is given below.
The advantage of index-and-overlay methods is that
they provide relatively simple algorithms or decision trees
to integrate a large amount of spatial information into
maps of simple vulnerability classes or indexes. The
methods are designed to rely on data that are readily
available from local, state, or federal agencies, such as
information on soils, water level depth, precipitation,
and geology. These methods are particularly suitable for
use with computerized geographic information systems
(GIS), which is a digital form of mapmaking, because they
usually involve overlaying and aggregating multiple maps
showing soil properties, depth to water table, recharge,
and so on.
DRASTIC Vulnerability Mapping (2). DRASTIC is the
most popular vulnerability mapping tool in the United
563
States. Put together by a group of experts and the EPA
in the mid-1980s, it has been applied to a number of
groundwater basins, regions, even states.
The name stands for depth to groundwater, recharge
rate, aquifer media, soil media, topography, impact of
vadose zone media, and hydraulic conductivity of the
aquifer. In DRASTIC, scores of 0 to 10 are assigned
to each parameter, 0 means low risk of groundwater
contamination; 10 means high risk of contamination. The
scores are multiplied by a parameter-specific weight. The
weighted scores of all seven parameters are then added
for the final DRASTIC score. Two sets of weights can be
used: one for general (intrinsic) vulnerability analysis, one
for specific vulnerability to pesticides.
GOD (1). GOD is a vulnerability assessment method
developed in Great Britain, where most groundwater
resources are in hardrock aquifers, primarily sandstone
and limestone aquifers. Unconsolidated overburden or
soil layers cover the fractured hardrock in many places.
Like DRASTIC, GOD is an index-and-overlay method
designed to map groundwater vulnerability across large
regions based on a few important parameters; GOD
stands for groundwater occurrence, overall lithology of
the unsaturated zone or overlying aquitard, and depth
to the groundwater table. Scores are assigned to each of
the three categories and then multiplied to yield a final
score. In developing GOD, the method’s authors have given
particular consideration to the likelihood of fractures or
fracture systems to develop in the soils, overburden, or
overlying geologic units of the aquifer.
Process-Based Computer Simulations
Process-based computer simulations afford a great amount
of realistic complexity and detail to be built into the
vulnerability assessment. Computer models can account
for complex physical and chemical processes on a very
detailed scale. Unlike the two-dimensional maps and
map layers used with other methods, computer modeling
allows for complete three-dimensional resolution. Geologic
and hydrogeologic variations with depth can therefore
be reproduced to evaluate their effect on vulnerability.
Process-based computer models focus on re-creating the
flow and transport patterns within the unsaturated zone
or in an actual aquifer and can be used to compute
travel times or concentrations of a contaminant in the
unsaturated zone or in the aquifer. Computer models do
not compute vulnerability directly. Rather, vulnerability
is defined as a function of what the computer models
simulate. For example, high vulnerability may be defined
as any region in the aquifer for which the computer model
shows a travel time of less than 5 years.
The most advanced computer models also allow the
analyst to compute the uncertainty that is inevitably
associated with the computer model predictions due to
shortcomings in the database fed into the computer
and due to our limited knowledge of the ‘‘underground
world’’. These methods use statistical methods to quantify
uncertainty. However, even simple computer models can
include a sensitivity analysis, that is, an investigation
564 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES
of the sensitivity of the model predictions to the
input provided.
Computer models are not commonly used for vulnera-
bility assessment due to their considerable data require-
ments and the expertise required to implement models.
In other words, computer simulation models are rarely an
economical alternative for vulnerability mapping. How-
ever, computer modeling is an excellent and economical
tool for vulnerability mapping if
• a more localized analysis of specific vulnerability
to particular land uses (particular contaminants) is
required and sufficient data are available or can be
collected to prepare the computer model, and
• a number of ‘‘what-if’’ scenarios involving complex
processes need to be evaluated for making important
land use planning decisions.
Four computer model codes are particularly popular
for computing the fate and transport of contaminants
as those contaminants travel downward through the
unsaturated zone to the water table: VLEACH (EPA),
PRZM (EPA), LEACH (Cornell University), and HYDRUS
(U.S. Department of Agriculture/Agricultural Research
Service Salinity Laboratory, Riverside, California) (for up-
to-date information on these codes, see the web pages of
these agencies). All four codes (listed here roughly in order
of complexity and user expertise required; VLEACH is
least complex and HYDRUS the most complex) compute
the downward movement of water (flow) and contaminants
(transport). These codes are designed specifically to
determine the amounts of sorption and degradation that
occur in soils and in the unsaturated zone and are typically
used to compute the residual, nondegraded mass of a
contaminant that arrives, over time, at the water table. For
vulnerability assessments, subregions or smaller sections
of land with similar unsaturated zone properties are
selected and grouped. One simulation is performed for
each group or section, and the results are mapped in
various ways; a map of the travel time to the water table
(perhaps specific to a contaminant), a map of the percent
removal of a contaminant within the vadose zone, and
so on.
The most common computer code used for groundwater
modeling (in all three dimensions) is MODFLOW (U.S.
Geological Survey). This model and other 3-D computer
models can be used for highly specific site studies of
groundwater contamination that are far beyond the
purpose of a typical vulnerability assessment. The data
requirements for running these models are tremendous
and require careful data preparation and data processing,
particularly in a fully three-dimensional simulation.
Sometimes, groundwater models are used to compute
explicitly the complex source area of the water reaching
a well. In that case, the same computer model can
also be used to compute specific vulnerability for
particular contaminants of concern within the source
area of the well. If a well-calibrated and well-documented
groundwater model exists, it may be preferrable to use
the groundwater model to address specific vulnerability
issues rather than using a basic index or a scoring-based
vulnerability assessment.
Statistical Methods
Statistical methods are used to quantify the risk of ground-
water pollution by determining the statistical dependence
or relationship between observed contamination, observed
environmental conditions that may or may not charac-
terize vulnerability (e.g., unsaturated zone properties,
recharge), and observed land uses that are potential
sources of contamination (e.g., fertilizer applications, sep-
tic tank occurrence). Once a model of this dependence
or relationship has been developed from the statistical
analysis, it can be used to predict—in a similar area
elsewhere—the chance or risk of contamination. Such an
application requires, of course, knowledge of significant
environmental conditions for that area. When statistical
methods are used, the risk of contamination is essentially
a quantitative measure of ‘‘vulnerability.’’ The higher the
contamination risk, the higher the vulnerability.
In principle, statistical methods are not much dif-
ferent from index-and-overlay methods. Both establish
a relationship between inherent natural conditions and
groundwater vulnerability (which the statisticians refer
to as groundwater contamination risk). In overlay meth-
ods, the relationship is established by a team of experts. In
statistical methods, the relationship is established by sta-
tistical analysis. The advantage of the statistical method is
that the statistical significance can be explicitly calculated,
which provides a measure of uncertainty or certainty of the
model. The disadvantage is that statistical methods are
difficult to develop and, once established, can be applied
only to regions that have environmental conditions simi-
lar to those of the region for which the statistical model
was developed.
Few statistical methods have been developed for
vulnerability assessment, primarily because they rely on
large data sets from regions already contaminated. The
key to the development of a statistical method is a large,
high quality data set of a specific contaminant or several
specific contaminants commonly found in a region. That
data set is then correlated in one fashion or another to
specific properties of the subsurface (depth to water, soil
permeability, slope, hydraulic conductivity, etc.) to create
a statistical predictive model that can be used to make a
quantitative statement about the contamination risk, such
as, ‘‘if this kind of unsaturated zone, aquifer, and land use
properties exist in a location, then the risk for groundwater
contamination is X%, with a confidence interval of plus or
minus Y%.’’ Some examples follow.
CALVUL. In California, a statistical method has been
developed by the California Department of Pesticide Reg-
ulations (DPR) to determine the specific vulnerability of
groundwater to pesticide residues. The method is nick-
named CALVUL (California Vulnerability approach; (5)).
DPR uses the vulnerability analysis to determine, for each
section of agricultural land in California, whether ground-
water in that section of land is vulnerable or not. If it is
vulnerable, the land section (square mile) becomes part
of a groundwater protection area with restricted pesticide
use. The DPR method distinguishes two types of vulnera-
bility: (1) vulnerability to pesticide leaching in coarse soils
 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES 565

with a shallow water table and (2) vulnerability to pesti-

cide runoff over hardpan soils (and subsequent leaching
in dry wells). Pesticide management practices depend on
whether pesticide leaching or pesticide runoff is of concern.
A pesticide database for the San Joaquin Valley was
used to determine the statistical relationship between pes-
ticide residue in groundwater and a large number of soil
(unsaturated zone) properties, including water holding
capacity, texture, organic matter content, permeability,
shrink-swell potential, slope, and infiltration. By deter-
mining which soil and geographic properties are good
predictors of pesticide residue in groundwater, the statis-
Probability that nitrate
tical model can be used to identify sections of land that exceeds 4 mg/L
have similarly ‘‘vulnerable’’ soil and geographic conditions 0–.17
but no past or current pesticide residue detections (5). >.17–.33
>.33–.50
Presumably, the most vulnerable areas are identified and >.50–.67
can be targeted for implementation of mitigation manage- >.67–.83
>.83–1
ment practices.
Figure 1. Groundwater vulnerability map for the United States,
where vulnerability is here defined as the probability that nitrate
Texas Case Study. In Texas, a statistical analysis
concentration in groundwater exceeds 4 mg/l. (reprint from (7),
similar to DPR’s San Joaquin study, albeit based on
also published at http://water.usgs.gov/nawqa/nutrients/pubs/est
a different set of statistical tools, was developed by v36 no10/fig3.html).
Evans and Maidment (6). Instead of pesticide residues
in groundwater, this method uses an analysis of nitrate
in groundwater as the basis for delineating vulnerable
The regression model was validated (tested) on a large set
groundwater areas. As in CALVUL, the assumption
of well records not used in developing the regression model
made by the developers of the method was that
and then applied to predict the probability of elevated
where high contamination exists, the aquifer is more
nitrate (a measure of nitrate vulnerability) in groundwa-
vulnerable than elsewhere. Nitrate was used as the target
ter throughout the United States at a resolution of 1 km2
variable because a large amount of data was available
(Fig. 1, from Reference 7). The study demonstrates the
throughout Texas and because nitrate is neither sorbed
advantage of the statistical method over index-and-overlay
nor significantly degraded in groundwater (except under
methods. The statistical assessment model uses objective,
anaerobic conditions). The analysis was done by using data
quantifiable measures to assess vulnerability. Its predic-
from the entire state. From a 30-year statewide database,
tive capability has been quantified as well (through the
the probability (risk) that a well would have nitrate levels
validation test).
in excess of some threshold value (1, 2, 5, or 10 mg/L NO3 -
N) was computed separately for each analysis unit. The
exceedance probability is used as a quantifiable measure of WHAT ARE THE LIMITATIONS OF VULNERABILITY
vulnerability; the higher the chances that nitrate is found ASSESSMENTS?
in wells within an analysis unit (area), the higher the
vulnerability. The study illustrates that overall results of Because of the implied imprecise nature of vulnerabil-
the statistical analysis follow the general patterns laid out ity assessments and the inevitable subjectiveness of the
by the index-and-overlay methods, albeit at considerably underlying interpretive scheme, the National Research
higher expense. Council (3) issued ‘‘three rules of groundwater vulnera-
bility.’’ These rules, or limitations, should be spelled out
National Groundwater Nitrate Vulnerability Assess- explicitly with every vulnerability assessment:
ment. The U.S. Geological Survey recently published a
national assessment of the probability of nitrate contami- • All groundwater is vulnerable to some degree.
nation in recently recharged groundwater (7). The assess-
ment relied on a specialized statistical regression analysis • Uncertainty is inherent in all vulnerability assess-
among major factors affecting the specific vulnerability ments.
of groundwater to nitrate and the measured occurrence • There is a risk that the obvious may be obscured and
of elevated nitrate levels in wells. The data set used to the subtle may become indistinguishable.
develop the regression model included wells from 20 major
groundwater basins across the United States; each con- The last refers to the danger, especially when using
tained 20–30 test wells. Nitrogen fertilizer loading and complex vulnerability assessment tools, that in light of
percent cropland were significant extrinsic (nitrogen use the final vulnerability index or ranking, one may lose sight
related) factors. The presence of well-drained soils, depth of the data used for the analysis and of the assumptions
to a seasonally high water table, and the presence of uncon- underlying vulnerability assessment schemes. The results
solidated sand and gravel aquifers were the key intrinsic of vulnerability assessments may mask real differences
factors that determine groundwater nitrate vulnerability. between specific sites or aquifer conditions.
566 WATER/ROCKS INTERACTION
AFTER THE ASSESSMENT: VERIFICATION AND POSTAUDIT
Verification refers to some independent procedure that can
verify the results of the vulnerability analysis. Postaudit
is essentially the same as verification but typically occurs
years later, when additional data have been sampled
that can be compared with the predictions made by the
vulnerability analysis.
Verification and postaudits of vulnerability assess-
ments can be done in many different ways. The most
common approach, particularly for verifying assessments
done with index-and-overlay methods, is to compare the
vulnerability map with the actual occurrence of some com-
mon pollutant in groundwater. Typical pollutants used are
nitrate and pesticides. However, such verification works
well only where the appropriate pollution source is actually
present and has been present for some time.
BIBLIOGRAPHY
1. Robins, N.S. (Ed.). (1998). Groundwater Pollution, Aquifer
Recharge and Vulnerability. Geological Society Special Pub-
lication No. 130, Geological Society, London.
2. Aller, L., Bennett, T., Lehr, J.H., and Petty, R.J. (1985).
DRASTIC—A standardized System for Evaluating Ground
Water Pollution Potential Using Hydrogeologic Settings. U.S.
EPA, Robert S. Kerr Environmental Research laboratory,
Office of Research and Development, EPA/600/2-85-018.
3. National Research Council. (1993). Ground Water Vulnerabil-
ity Assessment, Contamination Potential under Conditions of
Uncertainty. National Academy Press, Washington, DC.
4. California Department of Health Services. (2000). Drinking
Water Source Assessment and Protection Program. Available:
http://www.dhs.ca.gov/ps/ddwem/dwsap/
DWSAP document.pdf.
5. Troiano, J., Spurlock, F., and Marade, J. (1999). Update of the
California Vulnerability Soil Analysis for Movement of Pesti-
cides to Ground Water. Department of Pesticide Regulation,
Sacramento, CA 95814-3510, Document EH 00-05. Available:
http://www.cdpr.ca.gov/docs/empm/pubs/ehapreps/eh0005.pdf.
6. Evans, T.A. and Maidment, D.R. (1995). A Spatial and Statis-
tical Assessment of the Vulnerability of Texas Groundwater
to Nitrate Contamination. Technical Report CRWR 260, Cen-
ter for Research in Water Resources, Bureau of Engineering
Research, The University of Texas at Austin, Austin, TX.
7. Nolan, B.T., Hitt, K.J., and Ruddy, B.C. (2002). Probability of
nitrate contamination of recently recharged groundwaters in
the conterminous United States. Environ. Sci. Technol. 36(10):
2138–2145.
READING LIST
Blodgett, R. (1993). Drinking Water Vulnerability Assessments
in Texas. Proc., 1993 Ann. Conf., American Water Works
Association, Denver, CO.
Holtschlag, D.J. and Luukkonen, C.L. (1998). Assessment of
Ground-Water Vulnerability to Atrazine Leaching in Kent
County, Michigan: Review, Comparison of Results of Other
Studies, and Verification, Water-Resources Investigations
Report 98-4006. U.S. Geological Survey, Lansing, MI.
Rupert, M.G. (2001). Calibration of the DRASTIC ground water
vulnerability mapping method. Ground Water 39(4):625–630.
Snyder, D.T., Wilkinson, J.M., and Orzol, L.L. (1998). Use of a
Ground-Water Flow Model with Particle Tracking to Evaluate
Ground-Water Vulnerability, Clark County, Washington,
Water-Supply Paper 2488. U.S. Geological Survey, Denver, CO.
Vrba, J. and Zoporozec, A. (Eds.). (1994). Guidebook on Map-
ping Groundwater Vulnerability. International Association of
Hydrogeologists: International Contributions to Hydrogeology
Vol. 16, Verlag Heinz Heise, Hannover, Germany.
WATER/ROCKS INTERACTION
HENRY TENG
The George Washington
University
Washington, DC
Water–rock interaction is a process whereby solid earth
materials and aqueous phases in the Earth system
exchange matter and, ultimately, energy through chemical
and biochemical reactions. The extent of reactions is
determined by how far the system is away from
equilibrium, and the rate of reactions is controlled largely
by the kinetics of mineral-water interfacial processes.
In a broad sense, these reactions can be classified into
two categories based on the reactants involved: (1) those
between H2 O and the mineralogical components of rocks
and (2) those between solutes dissolved in natural waters
and the minerals. Water–rock interactions participate
in many geological processes such as ore deposit
formation, weathering, and pedogenesis. The outcome of
these interactions plays an important role in regulating
geochemical cycling of elements, the atmospheric CO2
level, and the chemistry and nutrient availability of
surface/near-surface waters at both local and global
scales.
Water–rock interactions are heterogeneous reactions
because of the involvement of multiple phases (that is,
solid and liquid phases in most cases). Despite the com-
plexity associated with the heterogeneity and variations in
mineral assemblages, water chemistry, and environmen-
tal conditions, reactions between aqueous phase and rocks
at the fundamental level can be regarded as consisting
of only two types of mechanistic movement (1), chemical
species attachment to and detachment from mineral sur-
faces. Individual or combined development of the two basic
actions over time yields three primary types of water–rock
interactions: adsorption (including ion exchange), dissolu-
tion, and precipitation. Following is a detailed discussion
of the three types of reactions.
ADSORPTION
Chemical sorption results directly from the occurrence
of undercoordinated sites and associated charges on
mineral surfaces. Development of surface charge in
aqueous environments can be understood as a consequence
of surface complexation, by which water molecules
form chemical bonds with the undercoordinated surface
atoms (2). These surface-bound water molecules are then
subject to a proton transfer process that shifts hydrogen

ions onto neighboring surface anions. This dissociative
sorption of water molecules results in one or more
types of surface species, depending on the chemical
composition of the minerals in question. For (hydr)oxide
materials, hydroxylated surface cation centers >X(OH)0 ,
where > symbolizes lattice bonds and X represents
the cationic component, seem to be the only concern.
For salt-type minerals Xm An , an additional, second-type
nonhydroxylated surface anion site >AH0 has to be taken
into consideration.
When present in solution environments, the surface
hydroxyl groups can undergo protonation or deprotonation
to form charged surface species. The relative abundance
of protonated and deprotonated species thus governs the
surface electric properties. Consider an oxide mineral XO.
The sign and quantity of surface charges incurred in pure
water will be controlled by the following reactions
≡ XOH2 + ⇔≡ X(OH)0 + H+
app
Ka1 (1a)
and
≡ X(OH)0 ⇔≡ XO− + H+
app
Ka2 (1b)
app app
where Ka1 and Ka2 are the apparent equilibrium
constants of reactions 1a and 1b. The participation of
hydrogen ions in the above reactions indicates that the
relative concentrations of the protonated and deprotonated
surface species, and hence the amount of surface charges,
are controlled by the pH of the aqueous environment.
The pH at which there are equal amounts of >XOH2 +
and >XO− (i.e., an uncharged surface) is defined as the
pH of zero net proton charge, below which the surface is
positively charged for >XOH2 + is more abundant than
>XO− and above which the surface is negatively charged
because of the excess of >XO− .
In the presence of other sorbing ions, say M (n+1)+ and
L(n+1)− , in solutions, competitive complexation becomes
operative as M can exchange for H and L for OH
at the >X(OH)0 sites to result in the formation
of surface complexes such as >XOM n+ and >XLn− .
Surface complexation under this condition becomes
adsorption where both chemical bonding and electrostatic
interaction take place between solute particles and
surface atoms (3). Empirical approaches to adsorption
employ simple equations to describe the adsorption
isotherm (relationship between the solution concentration
or activity and the surface concentration of an adsorbate),
without concerning the specific interfacial interactions
between adsorbate and adsorbent. For example, the
Langmuir isotherm, one of the most common empirical
models, simply treats adsorption as a free surface site >S
occupied by an adsorbate A to form a sorbed site >SA:
>S + A ⇔ >SA (1c)
The adsorption constant Kad given by the mass law is
[>SA]
Kad = (1d)
[>S][A]
WATER/ROCKS INTERACTION 567
Because the total amount of surface sites is [>S]tot =
[>S] + [>SA], Eq. 1d can be rewritten as
Kad [A]
[>SA] = [>S]tot (1e)
Kad [A] + 1
High concentrations of [A] lead to [>SA] ∼ = [>S]tot ,
whereas low values of [A] reduces Eq. 1e to a linear form
of [>SA] ∼
= Kd [A]. Another common empirical adsorption
model, the Freundlich isotherm, uses a logarithmic
relationship between [>SA] and [A] so that
[>SA] = kf [A]n (1f)
where both kf and n are constants. The Freundlich model
can be thought of as the Langmuir isotherm integrated
over many types of surface sites (4).
In the surface complexation model, adsorption is
treated as a chemical reaction
>X(OH)0 + M (n+1)+ ⇔ >XOM n+ + H+ (1g)
The equilibrium constant K app
[>XOM n+ ][H+ ]
K app = = K int K coul (1h)
[>XOH0 ][M (n+1)+ ]
now has contributions from chemical (intrinsic) as well
as electrostatic (coulombic) interactions between the
reactants. Quantitatively, K coul = exp(FZ), where Z
is the change of charges in surface species because of
adsorption. The surface potential  is a function of surface
charge σ , and the charge-potential relationship can be
derived from electrochemistry theories. A common one
σ = (8RTεε0 I)1/2 sinh(zF/2RT) (1i)
is given by the Gouy–Chapmen theory (diffuse double-
layer model). Here ε is the relative dielectric constant
of water, ε0 is the permittivity of vacuum, I is the ionic
strength of the bulk solution, and z is the valence of
the symmetrical background electrolyte. The quantity of
surface charge can be computed from the concentration of
charged surface species. For Eq. 1g,
σ = F[≡ XOM (n−1)+ ]/AS (1j)
where A and S are the specific area (m2 /g) and solid
concentration (g/l) of the adsorbent material, respectively.
Overall, assuming the total amount of surface sites
in Eq. 1g is approximated by [>XO]tot = [>X(OH)0 ] +
[>XOM n+ ], one can rearrange Eq. 1h into
 
exp(nF)K int [M (n+1)+ ]
[>XOM n+ ] = [>XO]tot
exp(nF)K int [M (n+1)+ ] + [H+ ]
(1k)
It becomes apparent that, after inspecting Eqs. 1e, 1f,
and 1h, the surface complexation approach is advanta-
geous over the empirical isotherms in that the formal pro-
vides additional insight into the dependence of adsorption
upon pH and ionic strength (through Eq. 1i). Moreover, the
568 WATER/ROCKS INTERACTION

use of K int in surface complexation models makes it possi- a system containing the crystal in contact with an
ble to compare results from different experiments because, undersaturated solution can be described by
unlike empirical parameters such as Kad (Eq. 1d) and Kf  
(Eq. 1f), the intrinsic surface complex formation constant dG = − νi µi dn + γj dAj (3a)
is not experimental condition dependent. i j
It is noteworthy to point out that although surface
complexation reactions can occur on any mineral crystals, where µi is the chemical potential of the ith species in
effective adsorption requires small grain sizes. Because the liquid phase; νi is the stoichiometric coefficient of the
of the inverse relationship between particle size and ith species in the solid phase; n is the number of moles of
specific surface area (defined as the ratio of surface area solute that are involved in the dissolution reaction; and γj
to volume of a particles), small grains (<1 µm) tend to and Aj are the surface free energy and the surface area of
have a significant percentage of surface-bound atoms. the jth type of faces, respectively. Using the definition of
Consequently, small particles have large numbers of active chemical potential,
surface sites available to undergo complexation reactions.
Micron-scale suspended particles and sub-micron-scale µi = µoi + RT ln αi (3b)
colloids of minerals are a common occurrence in surface
waters and groundwaters (Ref. 5). It is estimated that one can rewrite Eq. 3a as
80% of the materials transported by rivers and discharged
dG  
into the oceans is in the form of suspended particles. =− νi µoi − RT ln αiνi + γj (dAj /dn) (3c)
Although the exogenic cycle of the major cations and dn
i i j
anions is controlled by solution process, in most cases
the geochemistry of trace elements in natural waters is where µoi and αi are the standard chemical potential
governed by adsorption on separate mineral grains (6). and the activity of the ith species, ! respectively. At
equilibrium conditions, dG/dn = 0, i νi µoi = −RT
" ln Ksp ,
it then follows that the crystal solubility Kb = i αiνi can
DISSOLUTION be expressed as

Minerals dissolve because the chemical potential of solute 1 

ln Kb = ln Ksp + γj (dAj /dn) (3d)
in solid and liquid phases has not reached equality, RT
j
which indicates that the driving force of any dissolution
reaction may be estimated by the free energy difference, Equation 3d states that the solubility of a crystal is
G, between reactants and products. As the reaction always greater than the standard solubility product of the
approaches equilibrium conditions where G becomes solute. Rewrite Eq. 3d using the volume v and density ρ
zero, dissolution should cease to proceed. of the crystal and the molecular weight W of the solute,
Consider the dissolution of a solid compound ML, one gets

ML ⇔ M n+ + Ln− (2a) W/ρ 

The free energy change in this system is described by
G = Go + RT ln(aM aL ) (2b)
where Go is the standard free energy change, R is the
gas constant, and T is the temperature; aM and aL are the
activity of aqueous species M n+ and Ln− , respectively.
At equilibrium conditions, G = 0 and Eq. 2b can be
rewritten as
aM aL = exp(−Go /RT) (2c)
Because Go can be computed as the difference between
the standard free energy of formation for the dissolved
species and the solid phase, Eq. 2c yields a numerical
value defined as the standard equilibrium constant or
solubility constant, Ksp , of the compound ML.
When the solid–fluid interface is concerned, surface
energies and grain sizes have to be taken into account
to define the equilibrium conditions. Consider a mineral
crystal of any geometry with any number of faces that
are not related by any symmetry operations. At constant
temperature and pressure, the free energy change in
ln Kb = ln Ksp + γj (dAj /dv) (3e)
RT
j
For isotropic crystals with any characteristic dimen-
sion l (i.e., length, width, or height), γj (dAj /dv) can
be approximated by (2γ )/(3l) and Eq. 3e can be simpli-
fied to
2W/ρ
ln Kb = ln Ksp + (γ /l) (3f)
3RT
Equation 3f is the statement of the classic Gibbs–Thomson
relationship, which predicts that the difference between
crystal solubility and Ksp is size-dependent, and the
solubility of any crystals can only be approximated by
the standard solubility product of the solute when the
crystal is infinitely large. In a practical sense, the ‘‘infinite’’
size is usually no larger than a few micrometers. For
anisotropic minerals, each crystal face has to be considered
individually because the surface energy is face-specific
and Eq. 3e cannot be easily simplified. Nonetheless,
the conclusions drawn from Eq. 3f are just as well
applicable.
Dissolution not only takes place in the form of
dissociation reactions (Eq. 2a), but it can also result
from reduction–oxidation reactions where the number

of valence electrons changes on one or more reactant
elements. Redox reactions particularly affect water–rock
interactions involving minerals that contain multivalence
state elements such as S and Fe. However, because of
the relatively few minerals concerned, the contribution
of redox chemistry to the exogenic cycle of elements is
considered auxiliary (7). Most natural redox reactions
are mediated by microbial communities or induced by
solar radiation. In the formal scenario, oxidized mineral
species are often used as electron acceptors for metabolism,
whereas in the latter situation, electrons in the reactants’
molecular orbital are excited photochemically to pair up
with more electrons or to be released to lower the energy
state of the system.
Traditionally, geochemists treat redox reactions as
reactants’ gaining or losing oxygens. The more modern
approaches, on the other hand, choose to use electron as
a basic component. In this convention, reductants are
electron donors that release e− and oxidants electron
acceptors that receive e− . The activity of electron can
be expressed as a potential (Eh, in volts) or as an activity
unit (pe = − log ae− ). Quantitatively, Eh and pe can be
converted to each other through
F nonstoichiometric surface leaching) so SiO2 accumulates
pe = Eh (4a)
2.303RT
where F is the Faraday constant. In addition, Gibbs free
energy and the activity of electrons are related by
G = −nFEh (4b)
where n is the number of electrons in the reaction. If a
redox dissolution reaction can be written as
Ox (solid) + ne− ⇔ Rd (aqueous) (4c)
the standard free energy expression (Eq. 2b) becomes
RT
Eh = Eo − ln aRd (4d)
nF
or
1 1 1
pe = pe − log aRd = log K o − log aRd
o
(4e)
n n n Much of the discussion on reaction driving forces in
o o
where E and pe are the activity of electrons under
standard conditions. Notice that the equilibrium constant
of redox reaction K o can still be calculated from the
difference between the standard free energy of formation
for the dissolved species and the solid phase (i.e., Eq. 2c).
In general, the rate R of dissolution reactions can be
related to G by (8)
R = k(1 − eG/RT ) (5a)
where k is the rate constant. When the reaction is
sufficiently close to equilibrium (i.e., |G|  RT), Eq. 5a
reduces to
R = −(k/RT)G (5b)
WATER/ROCKS INTERACTION 569
This expression depicts a linear relationship between the
driving force and the kinetics of dissolution reactions
at near-equilibrium conditions. When applying Eq. 5b
to water–rock interactions, however, caution has to
be used because the dissolution of many minerals in
aqueous environments is initiated by the formation of
etch pits on crystal surfaces. Etch pit nucleation usually
requires one to overcome an energy barrier that decreases
with descending saturation state (9,10). At very close to
equilibrium conditions, the saturation state may be too
high to bring down the barrier to the extent that allows pit
formation, which indicates that dissolution kinetics may
change discretely, rather than in a linear fashion, with
decreasing saturation state (11,12).
A unique characteristic of mineral dissolution is
that some (alimino)silicate minerals may dissolve in an
incongruent or nonstoichiometric manner. This type of
reaction is best exemplified by the dissolution process
of Na- and K-feldspars in acidic environments. It is
generally agreed that the dissolution starts with an
initial fast ion exchange of hydrogen ions for alkali
and alkaline earth metals, followed by a slow process
that disrupt the Si-O-Al-O framework. The disruption
seems to proceed by preferential release of aluminum (i.e.,
in situ. Observations by numerous workers indeed reveal
the presence of a Si-rich surface coating on feldspar crystal
faces when the minerals are exposed to solutions with pH
<3–5 (Ref. 13). What researchers are not in complete
agreement on is to what extent the Si-rich layer is
originated from surface alteration. Although results from
investigations up to the end of the last century imply
that most coating (up to several thousand angstroms
in depth) forms through nonstoichiometric leaching of
the underlying minerals (Ref. 13), recent studies using
more modern experimental techniques suggest otherwise.
Synchrotron radiation and high-resolution transmission
electron microscopy show that the incongruent dissolution
may be operative only for, at most, the first few unit
cells at the mineral-water interface, and the thick
surface alteration thus forms largely by interfacial
dissolution–reprecipitation (14,15).
PRECIPITATION
dissolution section is directly transferable to precipitation
reactions. However, precipitation is not a pure mirror
image of dissolution in that the former is a combination
of two separate but sequentially occurring processes,
nucleation and growth. Nucleation refers to the atomic
process in which solute components congregate to
form a thermodynamically stable cluster, whereas the
following accretion of the cluster is termed growth.
Both nucleation and growth require the occurrence of
supersaturation, but each of the two processes has its
unique physiochemical controls.
Homogeneous nucleation occurs when a thermodynam-
ically stable 3-D cluster is born directly from solution. On
the other hand, solute particles can assemble onto an exist-
ing surface, and hence through heterogeneous nucleation,
570 WATER/ROCKS INTERACTION
to form a stable 2-D cluster. No matter how nucleation
takes place, the classic nucleation theory requires the
sizes of the clusters to satisfy the following energetics for
the atomic assemblage to become stable nuclei: A surface
energy increase associated with the emergence of the solid
phase precisely offsets the chemical potential decrease in
the solution because of losing solute. Mathematically, free
energy change upon formation of a spherical cluster of
radius r can be expressed as
G = 4π r2 γ − (4/3π r3 /v)µ (6a)
where γ is the average surface free energy and v is the
specific molecular volume of the solid. The minimal size of
a stable cluster
rc = 2vγ /µ (6b)
can be found by solving the equation d(G)/dr = 0. Any
cluster with r < rc is unstable and will be dissolved,
whereas the ones with r > rc will move on to the growth
stage. Inserting Eq. 6b into 6a, one has an expression that
estimates the energy barrier, G∗ , for nucleation:
16π γ 3 v2
G∗ = (6c)
3µ
The nucleation rate Rn , according to the conventional rate
theory, is then available through
−G∗
Rn = A exp( ) (6d)
kB T
where kB is the Boltzmann constant and A is a con-
stant related to the efficiency of particle collision (16).
All implications from Eqs. 6a–6d apply to heterogeneous
nucleation, except that, quantitatively, G∗ is usually
smaller when nucleation occurs on a substrate surface.
This result leads to a general conclusion that heteroge-
neous nucleation often occurs at a lower supersaturation
than does homogeneous nucleation.
Precipitation enters the growth stage with the onset
of enlargement of the minimal sized clusters. Nuclei
increase their size through attachment of individual
solute particles onto the cluster surfaces. During growth,
crystallographic control becomes strongly operative. That
is, faces that are more closely packed will prevail, whereas
the ones with smaller packing density will be phased
out. The more closely packed faces usually are the flat
face (F-face) on the ideal Kossel crystals (17); the growth
of flat surfaces is commonly recognized as following a
layer-by-layer mechanism where solute particles attach
to kink sites along steps (i.e., the edges of individual
layers) generated either by crystal imperfection or surface
nucleation. The net rate of step advance, vs , is determined
by the difference between solute fluxes attaching to and
detaching from the step kinks. This relationship can be
described mathematically by (18,19)
vs = (d2 /λo )υ{exp[(µS − E)/kB T] − exp[(µL − E)/kB T]}
(6e)
Here, d is the size of the attaching species in the lattice,
λo is the interlink distance, and υ is the fundamental
frequency (20) that can be approximated by kB T/hp where
hp is the Planck constants; µL and µS are the chemical
potentials of solute in liquid (solution) and solid (crystal),
respectively, and E is the free energy of an intermediate
activated complex.
Equation 6e can be further rearranged into
vs = (d2 /λo )υ exp[(µL − E)/kB T]{exp[(µS − µL )/kB T] − 1}
(6f)
Considering that the term exp(µL − E) has an entropy
factor, exp(S/kB ), and an energy factor, exp(−E/kB T),
and that (µS − µL ) is essentially the driving force for
growth and is quantitatively related to supersaturation,
σ , by
µ = kB T ln(a/ae ) = kB Tσ (6g)
one can rewrite Eq. 6f as
vs = {(d2 /λo )υ exp(−S/kB ) exp(−E/kB T)}/
kB T]}{(a − ae )/ae } (6h)
Notice that step velocity has a linear dependence on
the difference between solute activity and its equilibrium
value (a − ae ). By convention, the quantity between the
first two curly parentheses in Eq. 6h is termed the kinetic
coefficient, β, of step growth. This parameter, as shown
here, is characterized by the activation energy barriers,
E, and the corresponding entropic changes for the
formation of activated intermediate complexes.
Complementary to the classic approach of atom-
by-atom growth, recent findings suggest that crystals
can also grow by oriented aggregation of nano-sized
clusters (21,22). In this growth mode, adjacent nano-
crystallites coagulate after spontaneous self-organization
that situates the two parties to a common crystallographic
orientation. The new, larger particles gain thermodynamic
stability by forming chemical bonding between the original
clusters to eliminate the interface between them. It
is entirely possible that different growth mechanisms
operate at different conditions. When nucleation rates
are low and nuclei are far apart in the media, growth
may only be accomplished by atom-by-atom attachment
on growing surfaces. When sufficient nuclei are present,
they can coalesce, through orientated coagulation, to form
larger crystals, as long as the system is shifting toward a
lower energy state.
It is important to be informed that the kinetics of
interfacial reactions such as dissolution and growth can
be controlled either by the pace of the actual chemical
reactions or the speed at which the reactants and products
are delivered to and removed from the reaction sites,
or both. To date, it is generally agreed that surface
reactions, rather than transport, play a predominant
role in controlling the rate of mineral-water interactions.
This point of view is supported by the nearly ubiquitous
occurrence of etching features on both laboratory dissolved
and naturally weathered grains of many minerals. The
significance of this understanding is directly reflected by
the intensive interests in mineral surface geochemistry
that has been growing for more than two decades.
 GROUND WATER: WELLS 571

BIBLIOGRAPHY

1. Hochella, M.F., Jr. and White, A.F. (1990). Rew. Mineral. 23:
1–16.
2. Schindler, P.W. and Stumm, W.W. (1987). The surface
chemistry of oxides, hydroxides, and oxide minerals. In:
Aquatic Surface Chemistry. W. Stumm (Ed.). John Wiley &
Sons, New York, pp. 83–110.
3. Stumm, W. and Morgan, J.J. (1981). Aquatic Chemistry. 2nd
Edn. John Wiley & Sons, New York, pp. 600–601.
4. Stumm, W. (1992). Chemistry of the Solid-Water Interface.
John Wiley & Sons, New York, p. 95.
5. Langmuir, D. (1997). Aqueous Environmental Geochemistry.
Prentice-Hall, Englewood, Cliffs, NJ, p. 344.
6. Morel, F.M.M. and Hering, J.G. (1993). Principles and
Applications of Aquatic Chemistry. John Wiley & Sons, New
York, p. 509.
7. Morel, F.M.M. and Hering, J.G. (1993). Principles and
Applications of Aquatic Chemistry. John Wiley & Sons, New
York, p. 421.
8. Lasaga, A.C. (1998). Kinetic Theory in the Earth Sciences.
Princeton Univ. Press, Princeton, NJ, p. 188.
9. Frank, F.C. (1951). Acta Cryst. 4: 497–501.
10. Cabrera, N. and Levine, M.M. (1956). Phil. Mag. 1: 450–458. the bottom of the hole. Some wells are still dug by hand
11. Nagy, K.L. and Lasaga, A.C. (1992). Geochim. Cosmochim. today, but more modern methods are available. It’s still a
Acta 56: 3093–3111. dirty job, though!
12. Lasaga, A.C. and Luttge, A. (2001). Science 291: 2400–2404. Wells are extremely important to all societies. In many
13. Blum, A.E. and Stillings, L. (1995). Rev. Mineral. 31: places wells provide a reliable and ample supply of water
291–351. for home uses, irrigation, and industries. Where surface
14. Teng, H.H., Chen, L., Fenter, P.P., and Sturchio, N.C. (2001). water is scarce, such as in deserts, people couldn’t survive
Geochimica et Cosmochimica Acta. 65: 3459–3474. and thrive without ground water.
15. Hellmann, R. et al. (2003). Phys. Chem. Minerals 30:
192–197. TYPES OF WELLS
16. Stumm, W. (1992). Chemistry of the Solid-Water Interface.
John Wiley & Sons, New York, p. 216. Digging a well by hand is becoming outdated today (would
17. Kossel, W. (1927). Nach. Ges. Wiss., Gottingen, p. 135. YOU want to do it?). Modern wells are more often drilled
18. Chernov, A.A. (1961). Soviet Phys. 4: 116–148. by a truck-mounted drill rig. Still, there are many ways to
19. Chernov, A.A. and Komatsu, H. (1995). Topics in crystal put in a well—here are some of the common methods.
growth kinetics. In: Science and Technology of Crystal
Growth. J.P. van Eerden and O.S.L. Bruinsma (Eds.). Kluwer Dug Wells
Academic, Amsterdam, pp. 67–80.
20. Lasaga, A.C. (1998). Kinetic Theory in the Earth Sciences.
Hacking at the ground with a pick and shovel is one way
Princeton Univ. Press, Princeton, NJ, p. 178. to dig a well. If the ground is soft and the water table is
21. Penn, R.L. and Banfield, J.F. (1998). Science 281: 967–971.
shallow, then dug wells can work. They are often lined
with stones to prevent them from collapsing. They cannot
22. Penn, R.L. and Banfield, J.F. (1999). Geochimica et Cos-
mochimica Acta. 63: 1549–1557.
be dug much deeper than the water table—just as you

GROUND WATER: WELLS

ROGER M. WALLER
U.S. Geological Survey

There’s a good chance that the average Joe who had to

dig a well in ancient Egypt probably did the work with
his hands, a shovel, and a bucket. He would have kept
digging until he reached the water table and water filled

This article is a US Government work and, as such, is in the

public domain in the United States of America.
572 WELL SCREENS

cannot dig a hole very deep when you are at the beach. . .
it keeps filling up with water!
Driven Wells
Driven wells are still common today. They are built by
driving a small-diameter pipe into soft earth, such as sand
or gravel. A screen is usually attached to the bottom of
the pipe to filter out sand and other particles. Problems?
They can only tap shallow water, and because the source
of the water is so close to the surface, contamination from
surface pollutants can occur.
Drilled Wells
Most modern wells are drilled, which requires a fairly
complicated and expensive drill rig. Drill rigs are often
mounted on big trucks. They use rotary drill bits that chew
away at the rock, percussion bits that smash the rock, or,
if the ground is soft, large auger bits. Drilled wells can be
drilled more than 1,000 feet deep. Often a pump is placed
at the bottom to push water up to the surface.
WATER LEVELS IN WELLS
Ground-water users would find life easier if the water level
in the aquifer that supplied their well always stayed the
same. Seasonal variations in rainfall and the occasional
drought affect the ‘‘height’’ of the underground water level.
If a well is pumped at a faster rate than the aquifer around
it is recharged by precipitation or other underground flow,
then water levels around the well can be lowered. The
water level in a well can also be lowered if other wells near
it are withdrawing too much water. When water levels
drop below the levels of the pump intakes, then wells will
begin to pump air—they will ‘‘go dry.’’
1. Adequate open area (typically greater than 3–5%)
to maximize well efficiency1
2. Strong pipe-based well screen (e.g., horizontal
louver shutter screen) to provide maximum col-
lapse strength.
3. Constant diameter throughout the entire length of
the well screen.
4. Single material type (e.g., all copper-bearing or all
stainless steel) to avoid galvanic corrosion from two
different metals coming in contact with one another.
Well screens may be constructed with a variety of
different materials and designs. The screen material
can be different steels, such as carbon, copper-bearing,
high strength low-alloy, and stainless, or PVC materials.
The type of material used is dependent on the purpose
of the well, strength requirements, groundwater quality
(corrosion potential), the expected well life, and cost factor.
There are many different types of well screens; however,
the most commonly used are discussed below.
HORIZONTAL LOUVER SCREENS
Horizontal louver screens are a pipe-based well screen
manufactured by punch-formed downward facing horizon-
tal louver-shaped openings (Fig. 1). This type of screen is
particularly suited for use in large diameter, deep, gravel
envelope wells. The horizontal louver screen can be manu-
factured in a variety of patterns, with various percentages
of open areas. Percentage open areas range from approxi-
mately 3.4% to 7.4% for Ful Flo to Super Flo patterns (2).
The smooth, unobstructed interior of the horizontal louver
screen allows for efficient use of highly effective devel-
opment techniques such as swabbing and also facilitates
employment of efficient maintenance and repair meth-
ods. Disadvantages of horizontal louver screen include

WELL SCREENS
DENNIS E. WILLIAMS
Geoscience Support Services
Claremont, California

A well screen is a fairly generic term that denotes

the component of a well that provides (1) an opening
for water to enter the well casing from the aquifer;
and (2) stabilizes material in the ‘‘near-well zone’’ (i.e.,
materials immediately adjacent to the well) and prevents
it from entering the well.
Proper design of well screens therefore is paramount in
successful water wells. Most modern wells are constructed
using a filter pack placed between the well screen and
the aquifer. In a proper well screen/filter pack design (i.e.,
gravel envelope well), the fundamental principle is: The
Figure 1. Horizontal louver screen.
purpose of the filter pack is to stabilize the aquifer; the
purpose of the well screen is to stabilize the filter pack. 1
[It has been demonstrated (1,2) that if the percentage open area
A properly designed well screen therefore allows sand- is above 3–5%, well efficiencies do not significantly increase. It
free water to flow into the well with minimum drawdown. has also been demonstrated that entrance velocity is not a critical
Features of a properly designed well screen include: design factor and only needs to be below 0.5 m/s (1.5 ft/s).]
 WELL SCREENS 573

unavailability of slot size smaller than 1 mm (0.040 in.),
and difficulty in perforating though thick walled pipe [i.e.,
greater than 9.5 mm (0.375 in.)]. However, the superior
strength capabilities and long-life characteristics have
made horizontal louver screens an industry standard in
deep high capacity water supply wells.
CONTINUOUS WIRE WRAP SCREEN
slot screen is usually installed in gravel envelope wells.
Its chief advantages are reasonably high area of opening
(approximately 4.8%) and minimum frictional head losses
at relatively low cost. One important disadvantage is its
low collapse strength due to the large number of vertically
oriented slots. The manufacturing process is restricted
with respect to wall thickness, which is limited to a
maximum of 6.35 mm (0.250 in.).

Wire wrap screen is manufactured by wrapping a shaped MACHINE-SLOTTED (MILLED) SCREENS

wire (steel or plastic) around an internal array of
longitudinal rods (Fig. 2). The wire and rod are joined Horizontal and vertical machine-slotted screens are
by resistance welding, which produces a cage-shaped manufactured from casing in which openings are milled
cylindrical configuration. Continuous slot screens are very with axially oriented precision cutters (Fig. 4). Vertical
effective in naturally developed wells in uniform fine- slotted screens are also manufactured by prepunching flat
grained sands in relatively shallow, thin aquifers since sheets of steel prior to fabrication into cylinders. Machine-
they are able to be manufactured with very small slot slotted screens have disadvantages that include clogging
sizes and yet still maintain the necessary open area due to the parallel surfaces within the opening. As slot
to minimize frictional head loss. Continuous wire wrap
screens typically have open areas greater than 25%.
Additionally, the design allows small particles to pass
completely through the screen without becoming wedged
in and causing clogging of the screen. Disadvantages of
the continuous slot screen are associated with its relatively
low strength, which necessitates special handling during
installation and development, and also restricts the
extent of future well maintenance and repair efforts.
As the interior is obstructed by rods, the use of highly
effective well development techniques such as swabbing
is precluded. Should the screen be damaged, repair by
swedging is impractical.

BRIDGE SLOT SCREENS

Bridge slot screens are manufactured on a press from

flat sheets or plates. The slot opening is usually vertical
and provides two orifices, longitudinally aligned to the
axis. The perforated steel sheets or plates are then rolled Figure 3. Bridge slot screen.
into cylinders and the seam is welded (Fig. 3). A bridge

Figure 2. Continuous wire wrap screen. Figure 4. Horizontal machine-slotted screen.

574 WELL TEST
clogging is directly related to wall thickness, thicker
material encourages greater plugging. Some vertical
machine-slotted casings are machined with an undercut
to reduce this tendency. A second drawback is the low
area of openings (less than 1%), although this can be
overcome at higher cost by increasing the number of slots.
Collapse strength of vertically machine-slotted casings,
however, is substantially reduced when the number of
openings is increased. The main positive characteristics of
machine-slotted screens are low cost, uniform openings,
and availability in a fairly wide range of apertures
and patterns.
WELL CASING PERFORATED IN PLACE
Well casing perforated in place has historically been
used with cable tool wells drilled in unconsolidated
alluvium. Once the casing has been driven or jacked to
its total depth, a mills knife (vertical slots) or hydraulic
louver-type perforator (horizontal slots) is used down the
hole to perforate the casing at specific intervals. The
hydraulic louver-type perforator is superior to the mills
knife in producing slots due to better physical casing
characteristics and more control over the aperture size;
in addition, the louver-shaped openings provide greater
stabilization of particles in relative fine-grained sand
formations. The hydraulic louver-type perforator is limited
to a slot size of 3.2–4.0 mm (0.125–0.156 in.). Typically,
a well casing perforated in place would have a percentage
open area of less than 1%.
The design of the screen slot or aperture size for gravel
envelope wells requires the filter pack to be designed first.
The screen slot size must result in 90% retention of filter
pack material outside the screen.
BIBLIOGRAPHY
1. Williams, D.E. (1985). Modern techniques in well design. J.
AWWA 77: 9.
2. Moss, R. (1990). Handbook of Ground Water Development. John
Wiley & Sons, Hoboken, NJ.
WELL TEST
DENNIS E. WILLIAMS
Geoscience Support Services
Claremont, California
A well test is a test conducted in a completed and developed
well from which data can be collected to calculate certain
hydraulic parameters of the well and surrounding aquifer.
Hydraulic parameters may then be used, for example,
to design the maximum safe pumping rate of the well,
determine well field spacing, calculate aquifer storage,
or determine the rate of movement of contaminants
in the groundwater. Well tests include pumping tests,
recovery tests, and slug tests. Essential measurements
and data collected during well tests include time, depth
to water, and discharge rate. Obtaining accurate data
from a pumping test involves a carefully planned test
program, including consideration of continuous operation
and execution of the necessary steps.
Step-drawdown tests are generally the first type of
test performed on a newly completed production well
following final well development. A step-drawdown test
involves pumping the well at several discharge rates and
measuring the change in depth to pumping level in the
well as time progresses (Fig. 1). At least three ‘‘discharge
steps’’ are required for the test with subsequent discharge
rates increasing over the previous steps. Data obtained
from step-drawdown tests are used to determine well
production and drawdown characteristics, from which
the permanent production pump can be designed. The
discharge step rates are determined based on results from
the final development pumping. For example, if during
final development, the maximum discharge rate reached
2500 gpm with good groundwater level stabilization, then
the three rates for the step-drawdown test might be 750,
1500, and 2200 gpm. The discharge rates for a step-
drawdown test should be reasonably spread out over the
maximum range available for the well (as determined
during development).
Following the step-drawdown test, a constant-rate
pumping test is usually performed. The purpose of
the constant-rate test is to verify the design discharge
rate estimated from the step-drawdown test and also
to measure longer-term (usually greater than 24 h)
drawdown effects on the pumping well and any nearby
wells. Data obtained from both the pumping well and
nearby wells can be used to calculate aquifer parameters,
which can then be used to design spacing between wells,
calculate groundwater storage volumes, or determine the
rate of movement of groundwater. The discharge rate
during the constant-rate test should be kept constant for
the entire duration of the test and should be at the design
discharge rate for the well (Fig. 2). The design discharge
rate may be based on both development pumping and
results from the step-drawdown test. The step-drawdown
test is seldom performed for a duration sufficient to
design the permanent pump and should always be verified
with a longer-term constant-rate test. However, the final
design pumping rate is still subject to ‘‘sound engineering
Time since pumping began, minutes
1 10 100 1,000
0
Drawdown, ft
50
Figure 1. Step-drawdown test data.

Time since pumping began, minutes
1 10 100
0 ERIC S. WILSON
Drawdown, ft
Constant drawdown Recovery
50
Figure 2. Constant-rate test with recovery data.
judgment’’ and should consider both demand requirements
and pumping test results.
Recovery tests are carried out after pumping tests. The
test involves measuring the rate at which the groundwater
level recovers to its initial level after pumping has stopped.
In general, water level data from recovery tests are more
reliable due to the lack of water level fluctuations caused
by discharge variations.
A slug test is a popular and inexpensive method of
estimating the hydraulic properties of aquifers in small
diameter wells when the hydraulic conductivity of the
aquifer materials is too low to conduct a pumping test.
Slug testing involves either rapidly adding or removing
a ‘‘slug of water’’ from a well and measuring the rate
at which the water level recovers (either falling-head or
rising-head). Disadvantages with this method are that
it can only be used to determine the aquifer’s hydraulic
conductivity in the immediate vicinity of the well, and
results can be influenced by filter pack material adjacent
to the well screen.
The type of test to be performed depends on the test
purpose, available resources, and site-specific limitations.
For example, in an area containing only a single well, it
is not possible to calculate aquifer storage parameters
from a pumping test without an observation well
(i.e., nonpumping well). Similarly, observation wells not
screened in the same aquifer as the pumping well cannot
provide reliable interference data.
Once groundwater level data from the well tests have
been collected, compiled, and graphically plotted, the
data are evaluated using various methods to determine
hydraulic parameters of the aquifer and well. In many
cases, the calculation of parameters requires only solving
simple equations or a set of equations. For each method
there are basic underlying assumptions from which the
method equation was derived. These assumptions need
to be carefully considered with respect to the test data
and specific aquifer conditions when a method is used. In
addition to the basic assumptions, each of the analysis
methods has other assumptions specific to the method.
Many of the analytical solutions are based on the basic
differential equation of groundwater flow, which can be
derived from fundamental laws of physics.
SAFE YIELD OF AN AQUIFER 575
SAFE YIELD OF AN AQUIFER
1,000
E. L. Montgomery & Associates, Inc.
Tucson, Arizona
Safe yield is a controversial term and concept that
may be defined for common usage as the amount of
water an aquifer system can yield over a long period
without producing unacceptable effects. Unacceptable
effects may include an excessive decline in groundwater
level, sometimes described as groundwater mining;
reduction or elimination of riparian vegetation; reduction
or elimination of water in streams, lakes, springs,
and wetlands; aquifer compaction and land subsidence;
reduction in groundwater underflow to downgradient
aquifer systems; a change in groundwater temperature;
an undesirable change in the chemical quality of
groundwater; and salt water intrusion into aquifer
systems adjacent to oceans or saline lakes.
The definition of safe yield has evolved for nearly a
century; initially, it was defined by Lee (1) as ‘‘. . .the
limit and quantity of water which can be withdrawn
regularly and permanently without dangerous depletion
of the storage reserve.’’ The definition was modified and
expanded by Meinzer (2), Conkling (3), and Todd (4). Due
to ambiguity and lack of agreement on the definition, many
authors have suggested abandoning the term (5–7) for an
alternate term such as ‘‘sustainable yield.’’
Incorrect concepts, regulations, or laws may indicate
that safe yield is equal to recharge. This incorrect concept
has been described by Bredehoeft et al. (8) as the ‘‘water-
budget myth’’ because the concept oversimplifies the
necessary information needed to understand the effects
of groundwater development on an aquifer system.
It is a great deal more difficult to define safe yield of
an aquifer based on analyzing the unacceptable effects
of overpumping than to assume simply that safe yield is
equal to recharge. When water is pumped from an aquifer,
even in amounts equal to recharge, natural groundwater
discharge continues. The initial result of pumping is the
decline of groundwater levels near the pumping well,
and with time, across larger areas of the aquifer. This
water level decline increases the hydraulic gradient and
assists in drawing groundwater from marginal parts of the
aquifer system into the central part of the aquifer system,
but eventually water levels drop sufficiently that natural
groundwater discharge becomes smaller. As a result, even
by pumping groundwater at rates equal to recharge, a
decline in groundwater discharge may occur. This decline
in discharge could possibly be defined as unacceptable,
and the aquifer may not be in a state of safe yield.
When considering the pragmatic aspects of safe yield,
it is important to address the principal elements of the
water budget expressed by the law of conservation of mass
and the water budget equation:
Inflow = outflow ± change in storage
Pumping from an aquifer system over time changes
the amount of water in storage and subsequently changes

576 SPECIFIC YIELD STORAGE EQUATION
values of inflow and outflow, thus changing values in the
water budget equation. Theis (9) stated that groundwater
withdrawals are balanced, over time, by a loss of water
somewhere. Groundwater withdrawals by pumping result
in one or more of the following: (1) a decrease in
the amount of groundwater in storage; (2) capture of
previously rejected recharge, if available; (3) an increase
in water entering the aquifer system through increased
underflow, if available; (4) decreased outflows by way
of underflow to downgradient aquifers; or (5) decreased
outflow by reduction of discharge to streamflow or
to evapotranspiration. The hydraulics of an aquifer
system evolve toward, and eventually may approach,
a new dynamic equilibrium under increased pumping.
An exception may occur in hydraulically isolated, arid
environments where little recharge is available. Under
these conditions, if pumping exceeds the meager natural
recharge, no subsequent equilibrium will be reached and
groundwater mining is expected.
Groundwater withdrawal or development of water
resources of an aquifer system occurs in stages. These
stages include the predevelopment stage, partially devel-
oped stage, and fully developed stage. Under natural
or predevelopment conditions, an aquifer system is in
dynamic equilibrium among change in aquifer storage,
groundwater underflow, and interaction between ground-
water and surface water. As additional development
occurs, water budget values change as a result of an
extended period of groundwater pumping. A partially
developed aquifer system will see moderate declines in
groundwater level and decreases in groundwater in stor-
age; an increase in recharge may occur, and a decrease
in natural discharge is expected. Full development of an
aquifer system results in a large magnitude of ground-
water level decline, additional depletion of groundwater
in storage, maximization of recharge, elimination of dis-
charge to surface waters, and underflow to downgradient
aquifers may cease.
Safe yield cannot be defined solely on the basis of hydro-
logic conditions. Hydrologists may describe and quantify
the effect of groundwater withdrawals on an aquifer sys-
tem, but safe yield is properly a determination by a team
of natural resource managers, and the results must be
acceptable to concerned citizens. It is often difficult to
reach consensus on the true safe yield of an aquifer sys-
tem. Although difficult, the term and concept are the basis
of many water management programs and groundwater
laws and therefore must be given careful attention.
Wood (10) states, ‘‘. . .scientists and engineers must
avoid the trap of blurring the boundary between scientific
investigations and management decisions. Neither I nor
any other scientist or engineer can say whether any given
volume of water or delivery rate meets the needs of anyone
beyond the minimum personal requirement of two or three
liters per day necessary for long-term survival. ‘Meet-
ing the needs’ requires many value judgments, and their
assessment lies outside the realm of science or engineer-
ing. In many ways, sustainability appears to be analogous
to the management concept of aquifer ‘safe yield’ that has
been extensively used and misused in the last 85 years.’’
Next Page
BIBLIOGRAPHY
1. Lee, C.H. (1915). The determination of safe yield of
underground reservoirs of the closed basin type. Trans. Am.
Soc. Civ. Eng. 78: 148–251.
2. Meinzer, O.E. (1923). Outline of Ground-Water Hydrology
with Definitions. U.S. Geological Survey Water-Supply Paper
494.
3. Conkling, H. (1946). Utilization of ground-water storage in
stream system development. Trans. Am. Soc. Civ. Eng. 3:
275–305.
4. Todd, D.K. (1959). Ground Water Hydrology. John Wiley &
Sons, New York.
5. Thomas, H.E. (1951). The Conservation of Ground Water.
McGraw-Hill, New York.
6. Kazmann, R.G. (1956). Safe yield in ground-water develop-
ment, reality or illusion? J. Irrig. Drainage Div. Am. Soc. Civ.
Eng. 82(IR3): 12.
7. DeWiest, R.J.M. (1965). Geohydrology. John Wiley & Sons,
New York.
8. Bredehoeft, J.D., Papadopulos, S.S., and Cooper, H.H., Jr.
(1982). Groundwater: The water-budget myth. In: Scientific
Basis of Water-Resource Management. National Academy
Press, Washington, DC, pp. 51–57.
9. Theis, C.V. (1940). The source of water derived from
wells—essential factors controlling the response of an aquifer
to development. Civ. Eng. 10: 277–280.
10. Wood, W. (2001). Water sustainability—science or manage-
ment. Ground Water 39(5): 641.
SPECIFIC YIELD STORAGE EQUATION
THOMAS HARTER
University of California
Davis, California
The specific yield, Sy , is the ratio of the volume of water
that can be drained by gravity from a mass of porous
rock (unconsolidated or consolidated) and the volume of
that mass:
VW
Sy = ,
VT
where VW is the volume of drainable water and VT is the
total volume of porous rock or sediments (Fig. 1). Specific
yield, Sy , is typically expressed as a percentage.
Specific yield is part of the total porosity of a porous
rock or sediment. The total porosity (total pore space) of a
rock or sediment formation is much larger than the specific
yield. Total porosity includes the fraction of pore space that
is interconnected (called ‘‘effective porosity’’) and porosity
due to isolated pores. The effective porosity itself includes
specific retention and specific yield. Specific retention is
that volume fraction of water that is held back by adhesion
and capillary forces, when an aquifer is drained. Specific
yield is the amount of water that is actually available
for groundwater pumping, when sediments or rocks are
drained due to lowering of the water table.
Specific yield depends on the distribution of pores, their
shape, and grain size (Table 1). The larger the surface
area of the rocks or sediment particles per unit volume