Author
 |
Rawls, Walter |
|
Pachepsky, Yakov |
|
Submitted to: Soil Science
Publication Type: Peer Reviewed Journal Publication Acceptance Date: 5/15/2002 Publication Date: 7/1/2002 Citation: Rawls, W.J., Pachepsky, Y.A., 2002. Using field topographic descriptors to estimate soil water retention, Soil Science, vol. 167, no 6. pp. 423-435. Interpretive Summary: The soils' ability to transmit and to retain water is of importance in many agricultural and environmental projects. Soil hydraulic properties reflect soil composition. Soil texture as characterized by the proportion of particles of different sizes is known to be the most informative predictor of soil hydraulic properties. Soil maps are used in large-scale projects to infer soil texture and to predict soil hydraulic properties. Soil cartographers show only field-determined, error-prone prevalent textural class. Soil texture is affected by the slope and land surface shape. Our hypothesis was that using land surface characterization along with texture might improve the accuracy of estimating soil water retention. The National Soil Characterization database presented an excellent source of data to test this hypothesis. We used slopes, position on the slope, and land surface shape across the slope as topographic descriptors. Field characterization of land surface shape is done by classes rather than by numbers. For instance, surface shape is classified into concave, convex, planar, and undulated classes. Data of this type could be combined with percentages of textural components for prediction using regression trees that partitioned soils into homogeneous groups tentatively called topotextural groups (TTG). Using topographic variables and soil horizons as additional predictors appeared to be the way to make up for errors made in field determination of texture. In A horizon, the topotextural grouping resulted in estimates that were more accurate than from laboratory texture only. Topographic information appeared to be useful for improving estimates of water retention to be used in modeling for design and decision support purposes in agriculture and environmental protection. Technical Abstract: In field-, watershed-, and regional-scale projects, soil water retention is often estimated from soil textural classes shown in soil maps. Only the dominating textural class is often shown, cartographers routinely use error-prone field judgement of soil texture, and soil texture is known to vary along slopes and to depend on the land surface shape. We hypothesized that including topographic information in water retention estimation may increase the accuracy. To test this hypothesis, we extracted data on 216 soil pedons for soils of moderate and large extent from the NRCS soil characterization database. Textural classes, genetic horizon numbers, slopes, position on the slope classes, and land surface shape classes were the field descriptors that we used to estimate water retention at -33 and -1500 kPa potentials for each horizon in each pedon. Because our input variables were both categorical and continuous, regression trees were used for subdividing the samples into the smallest number of the most homogeneous groups tentatively called topotextural groups (TTG). The jackknife cross-validation was used to prune the regression trees to prevent an over-parameterization. Ten or less TTGs were defined for both the -33 and the -1500 kPa retention. The TTGs were different for the two matric potential levels. In terms of accuracy, using topographic variables and soil horizon appeared to be the way to make up for errors made in field determination of texture. In A horizon, the topotextural grouping resulted in estimates that were more accurate than from laboratory texture only. Even though most of topographic variables are categorical in this work, those variables appeared to be useful for improving estimates of water retention. |
