POTABLE GROUNDWATER SUPPLIES AND lOW - COST SANITARY ENGINEERING - HOW COMPATIBLE?

s. S. D. Foster

ABSTRACT

Many developing countries rely on groundwater to supply an increasing proportion of their demand for drinking water. The concomitant drive for major improvements in excreta disposal will include such measures as the intensive use of unsewered sanitation and as sewage treatment in stabilisation lagoons followed. especially in arid regions, by effluent reuse for irrigation. In certain hydrogeological conditions these low=cost technologies may be in conflict. The principal potential hazards are identified and illustrated. A more detailed understanding of these problems, coupled with improved design, careful siting, and integrated planning of the installations involved, are required to reduce the groundwater pollution hazard.

RESUME

Un grand nombre des pays en voie de developpernent dependent de leurs eaux souterraines pour pouvoir subvenir aux exigences augmentantes de I'alimentation en eau potable. Les tendances concomitantes vers une amelioration dans l'elimination des eaux usees veulent inclure de telles mesures comme l'usaqe intensif d'un assainissement sans egouts et d'un traitement des eaux residuaires dans des etangs de stabilisation, suivi par la reutilisation des effluents pour I'irrigation. II est possible dans certaines conditions que ces technologies moins dispensives tombent en conflit. Les principaux dangers potentiels sont identifies et illustres, Pour etre capable de diminuer les risques de la contamination des eaux souterraines, une comprehension plus profonde de ces problemas liee avec une planification perfectionnee , un choix approprie des sites et des projets integres concernant les installations y comprises sont necessaires.

INTRODUCTION

In developing countries, groundwater will be increasingly used, since it is normally the cheapest and safest source of potable water-supply. Most commonly involved will be large numbers of production boreholes of simple design, tapping comparatively shallow water-table aquifers, and providing individually small, untreated and unmonitored supplies, vulnerable to pollution from the land surface.

Improvements in sanitation are also widely and urgently needed in these countries. It is becoming accepted that low-cost technologies can be effective and are more appropriate where purely domestic effluents are involved. These include:

(a) On-site excreta disposal units, such as the ventilated pit and

pour-flush latrines, for the rural population, in villages and

smaller towns and, in some circumstances, in parts of larger towns

and cities.

(b) Wastewater treatment in unlined stabilisation lagoons, where waterborne domestic sewerage systems already exist, prior to discharge

1029

to surface watercourse or, especially in arid regions, to reuse for agricultural irrigation or for basin groundwater recharge.

It has long been recognised that the natural soil profile can be an effective media for the purification of human wastes, including the elimination of fecal microbes and the breakdown and/or adsorption of numerous chemicals. However, not all soil profiles, and their related hydrogeological environments, are equally effective in this respect. Current knowledge of the migration and attenuation of related water pollutants is derived largely from investigations, in North America and Europe, on the effectiveness of septic tank soakaways and on the operation of sewage land disposal schemes.

It is important to recognise that there are significant differences between septic tanks and latrines including:

(a) Much of the upper 2 m of the soil profile, the biotope, which is biologically most active and contains the greatest concentration of pathogen-antagonistic microbes, is normally removed during latrine construction.

(b) The hydraulic loading of septic tank soakaways is designed generally not to exceed about 30 mm/d, whereas substantially higher loading (80-120 mm/d) may often develop in pour-flush latrines.

(c) Fecal bacterial populations, and in particular pathogen counts, are much higher in latrines, especially in developing countries (109 Fe/lOO ml).

(d) The use of septic tanks is usually restricted to low density urban or rural settlements, whereas latrines are being installed at very high density in some villages and towns of developing countries.

There is pressure to reduce the normally-accepted minimum spacing of 15 m between excreta disposal unit and groundwater source (under favourable site conditions) to as Ii ttle as 5 m Ln the most densely populated settlements of some developing countries, such as Bangladesh.

There are also significant differences between sewage land disposal schemes in developed countries and the wastewater reuse schemes being introduced in some arid regions of the developing world:

1030

(a) The normal method of wastewater treatment (or stabilisation) involves retention in shallow unlined oxidation (facultative) lagoons, which may have high rates of seepage loss (especially after initial construction and subsequent cleaning) for example frequently in excess of 40 mm/d at one Peruvian site.

(b) Both the raw wastewater and the stabilised effluent may have very much higher fecal bacteria counts; 107 and 105 FC/100 ml and up to 700 and 60 Salmonella/lOa ml respectively at another Peruvian site. (c) In developing countries it is not currently realistic to consider chlorination of the stabilised effluent.

(d) The land commonly chosen for agricultural irrigation is freely draining, so as to reduce initial farming investment and to avoid soil management problems, and thus irrigation return waters cannot readily be contained.

PRINCIPAL GROUNDWATER POLLUTION HAZARDS

Pathogenic Bacteria & Viruses

Early research on the attenuation of fecal bacteria in the subsurface, including some directly relevant work on latrines, showed:

(a) In unsaturated soils, fecal coliforms were rarely found to penetrate more than 1 m in depth.

(b) When discharged directly to the saturated zone, lateral migration did not generally exceed 10 m in the groundwater flow direction.

A serious limitation of these early investigations, however, was that they were essentially limited to unconsolidated strata with a mean grain size of less than 300 ~m at sites where the saturated groundwater flow velocity probably did not exceed 1.0 mid (Lewis et al. 1982).

Numerous incidents of fecal groundwater pollution from excreta and wastewater disposal systems have been recorded, and no doubt countless more have occurred, at sites with:

(a) Only a thin cover (less than 3 m) of unconsolidated strata overlying fissured non-porous bedrock aquifers.

1031

(b) The groundwater table seasonally or perennially shallow (at less than about 3 m depth).

(c) High hydraulic loading of wastewater (greater than about 30 mm/d).

For example, gross fecal pollution of some water-supply boreholes in a Botswana village was traced to pit latrines, generally of the order of 20-40m distant (Lewis et al. 1980). The hydrogeological conditions involved were a low-transmissivity, weathered basement aquifer with low rates of seasonal rainfall recharge and a groundwater table at

4-6 m depth. Pathogenic species were detected, including some strains exhibiting resistance to common antibiotics. The largest documented distance of microbial travel in the saturated zone of an unconsolidated (non-fissured) strata is in excess of 900 m, for both bacterial and viral tracers, in alluvial-fan gravel deposits of a New Zealand

valley.

Less is known about the migration and attenuation of excreted pathogenic viruses, but they are too small (O.Ol-O.lllm) to be retained by porous-media filtration and their infective dose may be orders-ofmagnitude less. Retardation of virus penetration into the subsurface is believed to be largely dependent upon surface adsorption and subsequent degradation, especially by antagonistic aerobic bacteria.

Infiltration into, and through, the unsaturated zone affords the most important line of defence against fecal pollution of underlying aquifers. Unsaturated vertical hydraulic conductivity decreases,

often dramatically, with decreasing moisture content and increasing moisture tension (Fig. 1). In consequence for many soils and some rocks actual groundwater flow rates in the unsaturated zone do not average more than about 0,2 mid for hydraulic loadings of up to about 30 rom/d,

and unsaturated zone residence times are such as to allow gross attenuation of fecal pathogens through a combination of mechanical straining adsorption on mineral surfaces, and adaptation of the biotope promoting degradation by antagonisitc bacteria. However, under conditions of heavy artificial hydraulic loading and/or occurrence of very high intensity infiltrating rainfall, moisture tensions will

fall and actual groundwater flow velocity may be much higher, especi ally

1032

in consolidated fissured formations (Fi~ 1), reducing greatly the opportunity for pathogen elimination. In soils,fecal pathogens normally survive less than 30 days, normally with 90% population reduction in 10-15 days, but under cool, moist, alkaline conditions a hardy residual fraction may survive for many months.

10+2
10+1
'0
] 10°
~
.;;
. ."
c
:J
"0
c 10-1
0
c
.>1
:;
~
"0
>-
s: 10-2
n;
c
'E
'"
>
10-3
10-4 'la sand

1b silt loam

'lc clay

2a sandstone

2b limestone (microporous)

-....,

mllm

probable effect of fissuring

preferred tension range beneath latrines

,,.-------I!I-I1IIIIII11t!WJJ!LIIIIIII-,--~ ----r-, --,-----,"'" ----" ;oat~~~ial

o 0.5 1.0 1.5 2_0 2.5 (mH20) <<Ili!'!III!f------ increasing moisture content

full

saturation

'field capacity'

Fig. 1 In situ hydraulic conductivity-moisture potential (tension) relationship for representative unconsolidated and consol idated strata

The processes responsible for the elimination of excreted pathogens, both bacteria and viruses, will generally be much less active in the saturated zone of groundwater systems. Thus longer survivals than in soils can be generally expected. Since many aquifers are heterogeneous

and contain some highly permeable flow paths of relatively small crosssectional area, actual groundwater flow velocities often may exceed

1033

10 mid and reach 100 mid or more near production boreholes in fissured

formations. Thus in many hydrogeological environments increasing lateral separation between the potential source of fecal pollution and a potable groundwater supply source is not an effective way of increas-

ing protection against pollution, since large increments will be required.

and often impractical,

Nitrogen Species & Salinity

The build-up of nitrogen compounds in groundwater, while not representing as immediate a hazard as contamination by fecal microbes, is likely to be far more widespread and persistent.

An indication of the potential groundwater nitrate pollution from onsite excreta disposal units can be derived from the following considerations; a population density of 100 personslha represents a discharge of some 500 kg Nlhala to the ground which, if oxidised and leached by 100 rom/a of infiltrating rainfall, would result in the local groundwater recharge containing 500 mg N03-N/I. The concentration will be lower only if:

(a) Nitrogen remained immobile in organic forms until latrine contents are finally removed for land application.

(b) Ammonification or denitrification of latrine contents led to

gaseous discharge to the atmosphere.

However, even if only 10% of the nitrogen in excreta is nitrified and leached, local groundwater recharge may still reach very high nitrate concentrations in relation to WHO limits for potable water (Fig. 2); 50 ml N03-N/I in the case cited above. Further decrease could result if any of the following processes are significant:

(a) Dilution from regional groundwater flow.

(b) Denitrification within the groundwater system itself.

(c) Significant dilution from effluent liquid, as may be the case for pour-flush latrines.

Nevertheless, it must be expected that unsewered sanitation schemes

will have a major impact on groundwater nitrate concentrations (Lewis et al. 1982) especially in those arid regions without significant

1034

400

300

~ }

" o

~

"" 200 ,S

regional groundwater flow.

50

150

population density b (no. per hal

250

a. general indication only, depends on soil-vegetation system and proportion of surface run-off

b. assumes initial generation of 5 kgN/Capita/a but 90% return, directly or indirectly, to atmosphere

c. in some cases dilution from regional aquifer throughflow, streambed infiltration or effluent liquid itself will also be significant

Fig. 2 Grossly-simplified estimation of impact of unsewered sanitation schemes on nitrate concentration in local groundwater

Not much systematic field data has, as yet,

been collected but this has been demonstrated to be the case in Botswana (Lewis et al. 1980), where domestic water-supply boreholes and pit latrines have co-existed in certain major villages for some

years now.

In the two examples illustrated (Fig. 3), the groundwater

supplies are derived from relatively shallow boreholes in weathered basement rocks (late pre-Cambrian Waterberg Sandstone above minor escarpments and older metamorphic complex below) with seasonal rain-

fall recharge probably averaging above 50 mm/a.

In a 1976 survey,

N03-N concentrations in groundwater supplies from within the urban

limi ts were shown generally to exceed 20 mgy l , and in some cases 50 mgy I,

1035

1036

o

o

~
1: ~
E ~
e-~
~ ~:E
~o
c o :~
~ .s E
E:=
f c
0
.;
~
Z
I
OM
Z <;
~ ~ <
.~ E E
" g ~ ~ ""
E E I I E
s g 2 in
.8 v
~ A
E ®@ !) 0 0
5. whereas in the surrounding, largely uncultivated, area they were less

than 10 mg/l. For Mochudi (Fig. 3b), this result has been confirmed by more detailed sampling in 1982 and seasonal fluctuations in concentration have been demonstrated. Chloride concentration will generally increase concomitantly with nitrate. In the case of these Botswana villages, values always exceed 200 mg/l, and in most cases 300 mg/l. Potable water-supplies are now reticulated to village standpipes from boreholes outside the urban limits, and efforts are being made to protect these boreholes from new settlements and from cattle enclosures.

Stabilised wastewater effluent will generally contain in excess of

20 mg N/l, mainly in organic and ammoniacal forms, and in the case of coastal Peru, for example, 30 mg N/I is normal. Concentrations of Cl

and S will vary quite widely with the corresponding levels in the associated water-supplY,but values of 10 and 80 mg/l respectively, characteristic of Lima (Peru), have been taken for discussion (Fig. 4). When this effluent is used for agricultural irrigation, all the nonplant nutrients are likely to be leached from the soil by irrigation return water and to recharge local groundwater in permeable terrains. The concentrations will depend essentially on the rate of irrigation and of any infiltrating rainfall. Very large concentrations of Cl and S04-S may be reached in the absence of the latter, especially at irrigation rates only slightly in excess of crop water requirements (Fi~ 4). The situation with respect to nitrate is more complex because

of such factors as variable rate of nitrogen mineralisation in soils,

of crop uptake and of gaseous loss associated with soil denitrification

(Lance et al., 1976; Foster & Young, 1980). However, the lat~ermost is least likely for cultivated soils, for which very high N03-N concentrations must be anticipated in irrigation return waters (Fig.4).

Trace Organics

Insufficient is known about the migration, transformation and attenua-

tion of organic compounds in groundwater systems. However, the corresponding pollution risk should not be dismissed, even where sewage is entirely of domestic origin, since it will contain increasing numbers and concentrations of synthetic organic compounds from numerous house-

hold products.

1037

1000

500

20

CASE I triple cultivation of cash crop with nitrogen ~ uptake of 240 kg N/ha/a C

CI

--

CASE n continuous cover of forage crop with nitrogen ........ <= uptake of 450 kg N/ha/a C

stabilised wastewater a

_ irrigation water requirement

-_

S(10'ol)

N(totol)

I 11 II!

o

4.0

8.0

2.0

6.0

irrigation rate (rn/a)

a composition may vary significantly with water-supply quality, unit water consumption, proportion of industrial effluent, groundwater infiltration to sewers, stabilisation lagoon characteristics

b assumes wastewater is only source of irrigation water end local Tainfaliinsignificant

c additional to that 9lInerated by soil mineralisation (no losses by soil denitrification or otherprocBss)

Fig. 4 Grosslv-simplified estimation of potential nitrate and salinity problems in groundwater recharge from agricu Itural land irrigated with wastewater in arid climates

The migration of most soluble organics will be retarded by absorption and the more readily biodegradable of the absorbed compounds will be

eliminated.

However, increases in anionic ABS detergents and phenol

concentrations have been reported in groundwater from some areas of land disposal of sewage effluents and an increase in total (dissolved;

1973; Baxter, 1982).

organic carbon is also quite commonly observed, (e.g. Hughes & Robson,

In the absence of measurable BOD, this must

1038

represent an increase in more persistent (essentially refractory) organic compounds. Little is known about their character, but a potential for generating widespread trace organic contamination of groundwater must be recognised. The risk will be greater where artificial hydraulic loading is highest, and where waterborne sewage systems are involved.

CONCLUDING REMARKS

In general terms it can be concluded that for a given hydraulic and pollutant loading:

(a) The depth to groundwater table and the character of the strata in the unsaturated zone will be the principal factors controlling the degree of penetration of pathogens and biodegradable organics into groundwater.

(b) The amount of dilution by regional flow, local recharge and effluent fluid, will be the dominant factor controlling the final concentration of nitrates and the more persistent trace organics in groundwater.

However, it remains difficult to quantify the groundwater pollution

risk for a sufficiently

wide variety of hydrogeological conditions.

The urgent need for improvement in sanitation in most developing countries is such that the groundwater pollution risk associated with the sanitary engineering measures being introduced is generally not such as to warrant the abandonment of such activities. It should, however, be more widely recognised that a hazard exists, espec~ally in certain hydrogeological conditions, and significant degradation of groundwater quality is likely to result. Some of the groundwater quality problems are likely to be surprisingly slow to appear in watersupply sources, but to prove exceptionally persistent once having appeared. Under the most unfavourable hydrogeological conditions lowcost, safe, potable groundwater supplies and unsewered sanitation may be incompatible without substantial design modifications and great care in construction of latrines and/or boreholes.

A more detailed understanding of the character and scale of these

1039

problems is required to allow more effective, and not unduly conservative, planning of low-cost groundwater supplies and sanitary engineering measures on an integrated basis.

Where technically feasible, scientifically justified and economically viable, numerous measures can be considered to reduce the risk of pollution of groundwater supplies. These are listed below.

(a) Unsewered Sanitation

(i) Incorporate artificial filter media in base of latrines to improve retention and elimination of pathogens.

(ii) Artificially suppress mineralisation of organic nitrogen in latrines and thus reduce the seepage of ammonium and/or nitrate.

(iii) Ensure effective emptying of latrine pits when contents are stabilised and thus application of nitrogen to land, rather than accumulation in ground.

(iv) Reduce fluid hydraulic loading by increasing latrine areadepth ratio, thus increasing pollutant retention time in unsaturated zone.

(v) Impose minimum separations between groundwater supply boreholes and pit latrines considerably greater than the normally accepted minimum of 15 m.

(vi) Modify water-supply boreholes by extending the depth of grouted surface lining tubes to significant depth below pumping groundwater level, ensuring longer groundwater flow paths and increased pollutant residence in the groundwater system.

(b) Wastewater Stabilisation Lagoons & Irrigation Reuse Areas

(i) Control location with respect to hydrogeological conditions and to groundwater supply installations, and prohibit use of groundwater for potable supplies within defined zone.

(ii) Consider lining, or otherwise sealing, the base of lagoons to reduce seepage.

(iii) Increase lagoon area-depth ratio in order to reduce artificial hydraulic,pathogen and nutrient loading and increase pollutant residence time in unsaturated zone.

1040

(iv) Select types of crop grown to minimise mineralisation of nitrogen, and thus reduce nitrate leaching losses.

(v) Control irrigation schedules to maximise soil denitrification and to dilute concentrations of other mobile non-nutrient

salts being leached.

ACKNOWLEDGEMENTS

This paper is published by permission of the Director, NERC-Institute of Geological Sciences, London, and with financial support from the British Overseas Development Administration.

REFERENCES

Baxter K. M. (1982):

The effects of the disposal of sewage effluents on groundwater quality in the United Kingdom. Water Services 86: 219-220

Foster S.S.D.& Young C.P. (1980):

Groundwater contamination due to agricultural land-use practices in the United Kingdom. UNESCO Studies & Reports Hydrology 30: 268-282

Hughes J.L.& Robson S.G. (1973):

Effects of waste percolation on groundwater in alluvium near Barstow, California. Proc 2nd AAPG-USGS-IAHS Symp "Underground Waste Management & Artificial Recharge" I: 91-129

Lance J.e, Whisler F.& Rice R.C. 1976):

Maximising denitrification during soil filtration of sewage water. J Environ Qual 5: 102-107

Lewis W.~, Farr J.L.& Foster S.S.D. (1980):

The pollution hazard to village water supplies in eastern Botswana. Proc lnst Civil Engrs II: 69: 281-293

Lewis W.J., Foster S.S.D.& Drasar B.S. (1982):

The risk of groundwater pollution by on-site sanitation in developing countries: a literature review. IRCWD (Dubendorf) Report 01-82: 1-79

AUTHOR'S ADDRESS

S.S.D.Foster

(Groundwater Flow & Pollution Section Leader) Institute of Geological Sciences (Hydrogeology Unit) Wallingford (Oxford) OXI0 8BB

United Kingdom

1041