You are on page 1of 11

Spatial Data for Soil Mapping

By Tom DAvello, GIS specialist/soil scientist, NRCS, National Soil Survey Center, Geospatial Research
Unit, Morgantown, West Virginia; Katey Yoast, soil scientist, NRCS, Morgantown, West Virginia; and
Dylan Beaudette, soil scientist, NRCS, Davis, California.

Soil scientists have been using Geographic Information Systems and Image Processing software as tools
to assist with soil mapping since the late 1970s. However, widespread use within NRCS did not occur
until the early 2000s with the adoption of ArcGIS as the GIS platform for USDA. Since that time, all soil
survey offices have become equipped with GIS software and most soil scientists are comfortable using
ArcGIS as needed for their job duties. Fortunately, many soil scientists are only aware of pre-GIS
methods through the stories of seasoned soil scientists.

The best available tools for the given circumstances have been brought to the task of each generations
soil mapping challenges. The software and data currently available to soil scientists offer many
opportunities to aid the soil mapping process. Much has been written, discussed, researched, and taught
regarding digital soil mapping. The best way to think of this without getting bogged down in terminology
and methods is soil mapping using geospatial software and data. The end product is a soil map
produced by soil scientists.

The challenge for the new generation of soil scientists is making use of the large volume of available data.
The following table is meant to provide a brief description of data layers (covariates) commonly used for
soil mapping activities grouped by soil-forming factor (Jenny, 1941; McBratney et al., 2003). In addition,
a listing of software unique to the creation of certain covariates is presented along with links to specific
datasets.
Type of Preferred
Covariate Description Resolution Application
Covariate Software

Varies depending All areas where


Derived from a digital
on source; for most elevation is a
elevation model (DEM);
Order 2-3 soil factor in soil
Relief (R) Elevation elevation is represented as a
survey applications, class or
continuous terrain relief
5-10 m resolution is properties being
surface.
recommended modeled
The maximum rate of
change in elevation calculated
from a DEM to represent the
hydraulic gradient acting All areas where
upon overland and subsurface slope gradient is
Same as input ArcSIE
water flow through the a factor in soil
R Slope Gradient digital elevation SAGA
influence of gravity. class or
model GRASS
Neighborhood size and shape properties being
play a large role in resulting modeled
slope gradient output, as
described in specific Job Aid.

Direction of slope gradient


depicting flow direction.
Neighborhood size and shape
ArcMap
play a large role in resulting Potentially all
ArcSIE
slope gradient output, as areas, except
Same as input SAGA
described in specific Job Aid. where slope
R Slope Aspect digital elevation xTerrain
The xTerrain Toolbox gradient is < 3%
model Tools
contains scripts for and relief is
GRASS
calculating northness, < 150 m
eastness, northwestness, and
northeastness.

Potentially all
areas, especially
Estimated incoming solar Same as input ArcMap
where slope
R Solar Radiation radiation for a user-specified digital elevation SAGA
gradient is > 3%
time period, typically 1 year. model GRASS
and relief is
> 150 m
Geomor- Potentially all
Index to correctly represent
phometry areas, especially
southwest aspects being Same as input
and where slope
R Heat Load Index warmer than southeast digital elevation
Gradient gradient is > 3%
aspects (McCune and Keon, model
Metrics and relief is >
2002)
Toolbox 150 m
Curvature of slope gradient
(direction of the steepest
slope) depicting flow
acceleration. Neighborhood Same as input ArcSIE
Potentially all
R Profile Curvature size and shape play a large digital elevation SAGA
areas
role in resulting profile model GRASS
curvature output, as described
in specific Job Aid.

Curvature perpendicular to
slope gradient depicting flow
convergence. Neighborhood
Same as input ArcSIE
Tangential size and shape play a large Potentially all
R digital elevation SAGA
Curvature role in resulting tangential areas
model GRASS
curvature output, as described
in specific Job Aid.

Convergence index is used


to determine whether water
flow from neighboring cells
diverges or converges. Same as input
Convergence Potentially all
R Convergence is calculated digital elevation SAGA
Index areas
using flow direction between model
adjacent cells based on the
aspects of neighboring cells.

Quantifies downslope
controls on local drainage,
Same as input
Downslope assuming flow accumulation Potentially all
R digital elevation SAGA
Distance Gradient in flat areas downslope areas
model
topography.

A measure of the
Potentially all
openness/protection
areas, except
calculated by analyzing the Same as input
Morphometric where slope
R degree to which the digital elevation SAGA
Protection Index gradient is
surrounding relief protects the model
dominantly
given cell.
0-3%
Measure of flatness and
Multiresolution lowness depicting Same as input
Potentially all
R Index of Valley depositional areas (Gallant digital elevation SAGA
areas
Bottom Flatness and Dowling, 2003). model

Measure of flatness and


Multiresolution upness depicting stable Same as input
Potentially all
R Index of Ridge upland areas (Gallant and digital elevation SAGA
areas
Top Flatness Dowling, 2003). model
The deviation of a point Potentially all
elevation from the specified areas; good for
Same as input
Topographic local mean, calculated by delineating
R digital elevation SAGA
Position Index dividing the elevation valley bottoms
model
difference by its standard from hilltops and
deviation. ridges
Potentially all
A measure of fine (many)
areas; good for
versus coarse (few)
delineating
topographic spacing, Same as input
Terrain Surface crests and
R calculated as the number of digital elevation SAGA
Texture troughs
pits and peaks within a model
regardless of
specified neighborhood
amplitude and
(Iwahashi and Pike, 2007).
lithology breaks
Potentially all
areas, except
Same as input
The vertical height below where slope
R Valley Depth digital elevation SAGA
summit accumulation. gradient is
model
dominantly
0-3%
The relative height above
Same as input
the closest modeled drainage Potentially all
R Slope Height digital elevation SAGA
accumulation. areas
model
The normalized difference
between slope height and
valley depth. Also referred to Potentially all
as relative position. Same as input SAGA areas, especially
Normalized
R Neighborhood size and shape digital elevation xTerrain those with a
Height
play a large role in this model Tools toposequence
output, as described in pattern
specific Job Aid.

The vertical height above


the channel network. Also Same as input
Vertical Distance Potentially all
R known as Altitude Above digital elevation SAGA
to Channel areas
Channel Network. model

A measure of overland flow


distances to a channel
Overland Flow network based on a DEM and Same as input
Potentially all
R Distance to channel network information digital elevation SAGA
areas
Network Channel best depicting potential model
energy of flow.
A measure of water
Topographic accumulation or soil
Wetness Index saturation calculated as:
aka Wetness ln(sca/sg), where sca is the Same as input ArcSIE
Potentially all
R Index aka upslope contributing area per digital elevation SAGA
areas
Compound unit contour length (or model GRASS
Topographic specific catchment area,
Index SCA) and sg is the local slope
gradient (Moore et.al., 1988).
This is similar to the TWI
but is calculated as ln(fa/sg),
where fa is flow
accumulation and sg is the
local slope gradient. In this
way, the SAGA TWI does
not consider flow as very thin
Same as input
SAGA Wetness film and therefore predicts Potentially all
R digital elevation SAGA
Index cells situated in valley floors areas
model
as having a small vertical
distance to a channel and a
more realistic, higher
potential soil moisture
compared to the standard
TWI calculation.

A measurement of the
distance from the origin of
Potentially all
overland flow along its flow Same as input
areas, except
R Slope Length path to the location of either digital elevation SAGA
where slope
concentrated flow or model
gradient is < 3%
deposition.

A measure of soil erosion


calculated as ln(sca*sg),
Potentially all
where sca is specific Same as input
Stream Power areas, except
R catchment area and sg is local digital elevation SAGA
Index where slope
slope gradient (Moore et.al., model
gradient is < 3%
1988).

Landform classification
technique that characterizes
terrain patterns according to
slope, maximum curvature,
Fuzzy Landform minimum curvature, profile Same as input
Potentially all
R Element curvature, and tangential digital elevation SAGA
areas
Classification curvature based on a linear model
semantic import model for
slope and curvature and a
fuzzy classification (Schmidt
and Hewitt, 2004).
Landform classification
technique that characterizes
terrain patterns into 8, 12, or
Same as input
Terrain Surface 16 landforms using slope Potentially all
R digital elevation SAGA
Classification gradient, local convexity, and areas
model
surface texture (Iwahashi and
Pike, 2007).

Scale-invariant landform
classification technique that
characterizes terrain patterns
using a line-of-sight
Same as input
algorithm (Jasiewicz and Potentially all
R Geomorphons digital elevation GRASS
Stepinski, 2013). There are areas
model
10 output landforms: flat,
peak, ridge, shoulder, spur,
slope, pit, valley, footslope,
and hollow.
An index that is calculated
as the distance from
Same as input Terrain with
Depression Cost depression location using xTerrain
R digital elevation depressional
Surface slope gradient as a cost Tools
model positions
surface.

Areas with broad


An index that is calculated
valleys and
as the number of flowpaths Same as input
Potential xTerrain braided or
R per unit area as defined by a digital elevation
Drainage Density Tools undefined
user-specified neighborhood model
drainage
(Dobos et. al, 2005).
channels
Potentially
where relief is a
The absolute value of factor; relief and
Same as input
elevation difference for a xTerrain slope gradient
R Relief digital elevation
user-specified neighborhood Tools usually have
model
(Riley et al., 1999). > 90%
correlation; pick
one or the other
The value of a cell when
compared to the average Same as input Potentially areas
xTerrain
R Relief Ratio relief for a user-specified digital elevation where relief is a
Tools
neighborhood (Hammond, model factor
1954).
Areas where
relief is a factor
The product of the standard
Same as input and degree of
Roughness by deviation of elevation and xTerrain
R digital elevation dissection or
Relief and Aspect aspect variety for a user- Tools
model surface
specified neighborhood.
smoothness is
variable
Calculates the standard
Roughness by deviation of elevation for a
Same as input Potentially areas
Standard user-specified neighborhood. xTerrain
R digital elevation where relief is a
Deviation of High standard deviation Tools
model factor
Relief corresponds to rougher
terrain.
Calculates the standard
Same as input Potentially areas
Slope deviation of slope for a user- xTerrain
R digital elevation where slope
Heterogeneity specified neighborhood Tools
model gradient is > 8%
(Evans, 1998).
Potentially all
areas if project
Typically mean annual, but Varies, but 4km domain is of a
Climate (C) Precipitation seasonal time slices are also and 800m are PRISM large spatial
used. common extent and if
orographic
effects warrant
Potentially all
areas if project
Typically mean annual, but Varies, but 4km domain is of a
C Air Temperature seasonal time slices are also and 800m are PRISM large spatial
used. common extent and if
orographic
effects warrant
Potentially all
Newhall areas if project
model or domain is of a
C Soil Temperature Typically mean annual. Varies modeled large spatial
from sensor extent and if
data orographic
effects warrant
Often Potentially all
derived areas if project
from domain is of a
Evapo- Typically potential Varies, but 800m is
C PRISM large spatial
transpiration evapotranspiration. common
data or extent and if
Newhall orographic
model effects warrant
Often Potentially all
derived areas if project
The difference between
Water Balance from domain is of a
precipitation and Varies, but 800m is
C aka Effective PRISM large spatial
evapotranspiration for a given common
Precipitation data or extent and if
time period.
Newhall orographic
model effects warrant
Potentially all
Orthorectified Landsat band
Organisms Landsat areas where land
combinations of 7-4-2 and 15m
(O) GeoCover 2000 cover data is
4-3-2.
needed
Bathymetric
mapping,
distinguishing
O Band 1 blue 30m soil from
vegetation, and
detecting
cultural features
Peak vegetation
(plant vigor),
separating
vegetation from
O Band 2 green 30m
soil, and
separating urban
from non-urban
areas
Vegetation
O Band 3 red 30m
slopes
Biomass content,
shorelines,
Band 4 near
O 30m distinguishing
infrared
dry and moist
soils
Soil and
Band 5 short
vegetation
O wave infrared 30m
moisture; land
(MIR)
use
Band 6 thermal
O 60m Soil moisture
infrared
Normalized
ERDAS
Difference Calculation: Resolution of
O ArcGIS Vegetative cover
Vegetation Index (NIR RED)/(NIR+RED) imagery
SAGA
(NDVI)
Comprehensive geospatial
data and databases put
together by multiple Federal
Landfire (multiple agencies that describe ESD and soil
O Varies
datasets) vegetation, wildland fuel, and survey
fire regimes across the U.S.
and insular areas.

Land Cover National Land Cover


O 30m
(NLCD) Database
Derived
from Areas with
LiDAR relatively
point cloud; undisturbed
O Canopy Density Density of vegetative canopy. Varies dependent vegetation where
on canopy density
sufficient differences infer
point soil differences
density
Derived
from Areas with
LiDAR relatively
point cloud; undisturbed
O Canopy Height Height of vegetative canopy. Varies dependent vegetation where
on canopy height
sufficient differences infer
point soil differences
density
Hydrothermally
altered rocks
Parent Band 7 short
associated with
Material wave infrared 30m
mineral deposits;
(P) (MIR)
separating land
from water
ERDAS
Differentiate
Carbonate Resolution of Model
P (Band 3 - 2)/(Band 3 + 2) carbonate-rich
Difference Ratio imagery ArcGIS
areas
SAGA
ERDAS Differentiate
Clay Difference Resolution of Model areas of high
P (Band 5 - 7)/(Band 5 + 7)
Ratio imagery ArcGIS clay hydroxyl
SAGA influence
ERDAS Differentiate
Ferrous Minerals Resolution of Model areas of higher
P (Band 5 - 4)/(Band 5 + 4)
Difference Ratio imagery ArcGIS ferrous mineral
SAGA influence
ERDAS Differentiate
Iron Difference Resolution of Model areas of higher
P (Band 3 - 7)/(Band 3 + 7)
Ratio imagery ArcGIS iron mineral
SAGA influence
Differentiate
ERDAS
sedimentary rock
Rock Outcrop Resolution of Model
P (Band 5 - 2)/(Band 5 + 2) (lime/dolostone)
Difference Ratio imagery ArcGIS
from igneous
SAGA
rock
Potentially all
areas where
P Geology 1:250,000
parent material
is pertinent
Data generated by aerial
sensing of radiation
Radiometric data
emanating from the earth's
may be helpful
surface provides general
Aeroradiometric in differentiating
P estimates of the geographic
Data various parent
distribution of uranium,
materials (Cook
thorium, and potassium in
et al., 1996)
surficial and bedrock units.

Soil polygons or regions


Potentially all
from SSURGO or SSURGO,
Soils (S) Soil Map 1:12,000-1:24,000 areas where data
gSSURGO. gSSURGO
are available
Soil description with field
observations and/or physical, Potentially all
KSSL and
S Soil Pedon chemical, and mineralogical areas where data
NASIS
characterization. are available

Data in x,y,z format, such


Potentially all
as EM, GPR, Gamma, or User
S Proximal Sensing Various spacing areas where data
Bathymetric surveys. developed
are available
Literature

Cook, S.E., R.J. Corner, P.R. Groves, and G.J. Grealish. 1996. Use of airborne gamma radiometric data
for soil mapping. Soil Research 34(1):183-194.

Dobos, E., J. Daroussin, and L. Montanarella. 2005. An SRTM-based procedure to delineate SOTER
Terrain Units on 1:1 and 1:5 million scales. EUR21571 EN, Office for Official Publications of the
European Communities, Luxembourg.

Evans, I.S. 1998. What do terrain statistics really mean? In S.K. Land et al. (eds.) Landform Monitoring,
Modeling and Analysis, John Wiley and Sons, Ltd., Chichester, UK, pp. 119-138.

Gallant, J.C., and T.I. Dowling. 2003. A multiresolution index of valley bottom flatness for mapping
depositional areas. Water Resources Research 39(12).

Hammond, E.H. 1954. Small scale continental landform maps. Annals of the Association of American
Geographers 44:32-42.

Iwahashi, J., and R.J. Pike. 2007. Automated classifications of topography from DEMs by an
unsupervised nested-means algorithm and a three-part geometric signature. Geomorphology 86:409440.

Jasiewicz, J., and T.F. Stepinski. 2013. GeomorphonsA pattern recognition approach to classification
and mapping of landforms. Geomorphology 182:147-156.

Jenny, H. 1941. Factors of soil formation: A system of quantitative pedology.

McBratney, A.B., M.M. Santos, and B. Minasny. 2003. On digital soil mapping. Geoderma 117(1):3-52.

McCune, B., and D. Keon. 2002. Equations for potential annual direct incident radiation and heat load.
Journal of Vegetation Science 13(4):603-606.

Moore, I.D., G.J. Burch, and D.H. Mackenzie. 1988. Topographic effects on the distribution of surface
soil water and the location of ephemeral gullies. Transactions of the ASAE 31(4):1098-1107.

Riley, S.J., S.D. DeGloria, and R. Elliot. 1999. A terrain ruggedness index that quantifies topographic
heterogeneity. Intermountain Journal of Sciences 5:23-27.

Schmidt, J., and A. Hewitt. 2004. Fuzzy land element classification from DTMs based on geometry and
terrain position. Geoderma 121(3):243-256.