The mean rock volume percent was determined for each of 11 standard layers for each map unit of each state using data from the STATSGO Comp and Layer tables. The standard layers were introduced because of the wide variation in the number, thickness, and depth to top and bottom of soil layers in the STATSGO data from one soil component to another, even within the same map unit.

Determining the volumetric percent of rocks for the 11 standard layers required three main steps:

1. Computing the percent of rock by volume in each component layer.

2. For each component, determining the contribution of each component layer to the 11 standard layers.

3. For each map unit, combining the contributions of all components to compute the mean rock volume percent for each standard layer.

Since the size variables are given in terms of percent by mass, computation of the volumetric rock fraction requires combining the mean bulk density (BD) of the fine soil component with the mean particle density (PD) of the rocks. In terms of the variables recorded in the STATSGO Layer table, the mean rock volume RVOL for a layer is given by

RVOL = 100 - FINES / DENOM

where the mass percent of fines soil is given by

FINES = NO10 * (100 - INCH3 - INCH10)

and the denominator is

DENOM = FINES + (100 - FINES) * (BD / PD)

The contributions of each component layer to the standard layers for a given map unit were determined using the component layer depths specified by Layer table variables LAYDEPL and LAYDEPH, the mean depth to bedrock for each component calculated by averaging Comp table variables ROCKDEPL and ROCKDEPH, and the percent of the area of the map unit covered by each component as specified by COMPPCT. For each component, the layers defined in the Layer table were compared with each standard layer in turn. If the standard layer was entirely included within one of the component layers, the rock volume value for the layer was multiplied by the COMPPCT value to determine the weighted contribution of the component to the standard layer. If the standard layer overlapped two or more component layers, the rock volume values for each component layer were first weighted in proportion to the amount of overlap before multiplication by the COMPPCT value. The region from the bottom of the last component layer to the bottom of the last standard layer, if any, was assumed to be the same as the lowest component layer down to the mean bedrock depth, below which the rock volume percent was set to 100.

The weighted contributions of all components to each standard layer were then summed to obtain the mean rock volume values for the map unit. However, if a component was identified as all water (COMPNAME = "WATER") or if the Layer table records contained contradictory values for particle size and density (see below), the component was omitted from the computation. If the map unit was entirely water, it was assigned a rock volume of zero. Otherwise, if all non-water components were unusable because of invalid or contradictory information, the map unit was assigned a flag value of 101%.

Two major problems were encountered which affect the validity of the computed values.

1. Many components specify ROCKDEPL = ROCKDEPH = 60 inches (152 cm) to infer that the soil was not examined below this depth. In most cases, bedrock is not actually present. However, there was no way to determine whether this was the case for any given component. Accordingly, the rock volume values for layers extending below this depth (the deepest two standard layers) will frequently be misleading.

2. A number of non-water components, and in some cases entire map units, have BDL = BDH = 0 for all component layers, even though other variables for the layer (e.g., texture) contradict the BD values. In addition, values for NO10, INCH3, and INCH10 were often missing, and in some cases contradictory. In particular, there were about 250 components for which INCH3L + INCH10L exceeded 100%, suggesting that when the values were entered the INCH3 variable was incorrectly interpreted as the total percent by mass of rock fragments larger than 3 inches, rather than being only the fraction between 3 and 10 inches as specified in the STATSGO variable definitions. It was not possible to determine for how many additional components, with INCH3L + INCH10L < 100%, this same incorrect interpretation was made. In all cases, an attempt was made to compute the rock volume using the values as specified, and setting missing NO10, INCH3, and INCH10 values to zero. In a number of cases in which the Layer table specified BDL = BDH = 0, the top of the layer (LAYDEPL) was at or below the mean depth to bedrock; for these cases the layer was assumed to be bedrock, and the rock volume was set to 100%. A component was omitted from the computations for a map unit only if the given data values led to nonsensical results, such as a negative value for FINES or a zero or negative value for DENOM.

4/19/99