This dataset was created by determining the dominant standard soil texture class for each standard layer of each map unit for each of the 48 conterminous states (mapunits for the District of Columbia are included under Maryland), and then joining the results for the 48 states into a single dataset. This data set is in ArcGRID file format at 1 km resolution. Any 1-km grid cell which contains portions of two or more mapunits was assigned the dominant texture class of the mapunit which occupies the largest fraction of the cell. The dominant soil texture class for each layer may be accessed using the mapunit serial numbers associated with each STATSGO mapunit; the texture class data have been incorporated into the Value Attribute Table (VAT) entry for each mapunit.

The information below was compiled from the following web page: http://www.essc.psu.edu/soil_info/index.cgi?soil_data&conus&data_cov Laboratory and field methods for determining the hydraulic properties of soil are costly and time consuming. A more tractable approach for many applications is the determination of soil hydraulic properties from conventionally measured soil property information, including soil texture, particle-size distribution, bulk density, and organic matter content. Many researchers have developed formulations based on the soil water retention curve and the underlying statistical pore size distribution theory in an attempt to relate commonly observed soil information (e.g., soil texture) to desired soil hydraulic properties. Clapp and Hornberger (1978), for example, examined the correlation of soil texture with a power function model of the soil moisture characteristic. Using desorption data from 1446 different soils, they tested the power function and found considerable variation in saturation suction within and among textural classes. Gradual air entry was determined to have a significant effect on the determination of the wetting front suction. They also discovered that the power function coefficient is highly correlated with soil texture, as represented by mean clay fraction within the appropriate USDA soil texture class. Clapp and Hornberger's work has been the basis for most of the soil-property parameterizations used by the SVATS modeling community. The table below is an example of the type of "lookup table" that is currently used in many of these models. Until recently, there has been a dearth of actual soil texture information over regional scales with which to apply any of these approaches. A wide range of schemes, amounting to little more than poorly-directed guesswork, have been used to generate soil information for model input.

For many STATSGO components, a depth-to-bedrock value of 60 inches (152 cm) was used to indicate that the soil was sampled only to this depth, and no bedrock was encountered. As a result, for many mapunits an entry of "bedrock" for the two lowest standard layers (1.5 to 2.5 m) may actually indicate "no data".

Miller, D.A. and R.A. White, 1998: A Conterminous United States Multi-Layer Soil Characteristics Data Set for Regional Climate and Hydrology Modeling. Earth Interactions, 2. [Available on-line at http://EarthInteractions.org] http://www.essc.psu.edu/soil_info/index.cgi?soil_data&conus&data_cov&ph

Process_Description:

The dominant soil texture class for each map unit was determined from the STATSGO Component and Layer tables for each state. For each layer within a component the Texture1 variable contains the rock fragment and texture class information in a combined description (e.g. "extremely cobbly-sandy loam"). 

However, the number, thickness, and depth to top and bottom of soil layers in the STATSGO data varies widely from one soil component to another, even within the same map unit. The wide variation of layer definitions makes it difficult to use the STATSGO data in models, especially if they require rasterized (gridded) data. To facilitate rasterization of the STATSGO data and the physical and hydraulic properties derived from them, we converted the STATSGO layers to a set of 11 standard layers. 

For each component in a given map unit, the layers reported in the Layer table for the component were compared with each standard layer. If the standard layer was entirely included within one of the component layers, the texture class associated with the Texture1 value (which specifies the dominant soil texture and rock fragment class for the component layer) was multiplied by the Comppct value (percentage of map unit covered by the component) to determine the fractional contribution of the texture class to the standard layer. If the standard layer overlapped two or more component layers, the Texture1 values of each component layer were first weighted in proportion to the amount of overlap before multiplication by the Comppct value. After all components for the map unit were processed, the texture class for which the total fractional contributions from all components in the map unit was largest was determined for each standard layer and entered as the dominant texture class for that layer. 

Many Component table entries for depth to bedrock used 60 inches (1.5 m) to indicate that bedrock was not encountered within this distance of the surface. In many cases, the layer table for such a component included a non-rock layer extending below the specified maximum depth to bedrock. When the Layer table did contain layers extending below the depth to bedrock reported for the component, these layers were used, and the bedrock was assumed to actually start immediately below the deepest such layer. When the bottom of the deepest layer was above the reported bedrock, the bottom layer was extended to the bedrock depth or the bottom of the deepest standard layer (2.5 m), whichever was less. The standard layers are as follows: 

The table below lists the texture classes and their assigned numerical codes in the multi-layer soil characteristics data set. Water, organic materials (peat, muck, etc.) and other non-soil surface bedrock cannot be placed in one of the 12 soil texture classes. These classes have been combined into "water", "organic", and "other" and are also given. Including these classes in the final data product provides maximum flexibility to the modeler for making decisions with regard to the use of the dataset. Additional lookup tables may be used to interpret these classes or re-classification functions found within standard GIS software environments can easily aggregate the listed classes. 


The 11 standard layers are :

Layer     Thickness       Depth to Top    Depth to Bottom

1      5 cm (2 in)        0 cm (0 in)     5 cm (2 in)
2      5 cm (2 in)        5 cm (2 in)    10 cm (4 in)
3     10 cm (4 in)       10 cm (4 in)    20 cm (8 in)
4     10 cm (4 in)       20 cm (8 in)    30 cm (12 in)
5     10 cm (4 in)       30 cm (12 in)   40 cm (16 in)
6     20 cm (8 in)       40 cm (16 in)   60 cm (24 in)
7     20 cm (8 in)       60 cm (24 in)   80 cm (31 in)
8     20 cm (8 in)       80 cm (31 in)  100 cm (39 in)
9     50 cm (20 in)     100 cm (39 in)  150 cm (59 in)
10     50 cm (20 in)     150 cm (59 in)  200 cm (79 in)
11     50 cm (20 in)     200 cm (79 in)  250 cm (98 in)

The above selection of the number and depths of these standard layers reflects three main considerations:
The wide variation of numbers, thicknesses, and depths of layers for different components means that there are no "natural" or "obvious" choices for the standard layers.
Many models are particularly sensitive to the properties of the top few centimenters of soil; hence extra priority should be given to preserving all available information for this region.
To minimize data volumes, layer thicknesses should not be much less than the thicknesses of "typical" component layers at similar depths.
To aid in the selection of standard layers, therefore, the frequencies of depths and thicknesses of layers were tabulated for all components. This tabulation indicated that roughly 50% of components have surface layers thicker than 20 cm (8 inches); only about 4% of surface layers have a thickness of 5 cm (2 inches) or less, and about 16%, 10 cm (4 inches) or less. Deeper layers are in general thicker -- roughly 60% of all layers were at least 50 cm (20 inches) thick. The majority of components did not record layers extending below 60 inches (approximately 1.5 m); only about 10% include layers extending beyond 2.0 m (79 inches).

domtext_grid

Metadata: