Soil-Landscape Estimation and Evaluation Program (SLEEP) to predict spatial distribution of soil attributes for environmental modeling

The program utilizes digital elevation models and field observations to generate spatially continuous representation of soil attributes. This information is essential for the Soil and Water Assessment Tool (SWAT), a semi-distributed process-based ecohydrological model used to simulate stream flow, crop yield, sediment transport, and nutrient transport. The accuracy of soil information determines the accuracy of environmental modeling applications. However, collecting detailed soil information is complex and costly. Traditional soil maps, which represent soil characteristics in polygon format, aggregate individual soil attributes and present soil information in a generalized form. This approach is not suitable for environmental modeling, which requires proper representation of the spatial distribution of soil attributes. Research has shown that topographic variables play a significant role in soil differentiation. Soil scientists use qualitative relationships between topography and soil variation in soil mapping, while some researchers have used quantitative relationships to estimate the spatial distribution of different soils. The use of GIS and remote sensing provides promising tools to quantify these relationships and aid digital soil mapping efforts. Digital elevation models (DEMs) are used to derive many topographic variables, which are used to predict the distribution of soil characteristics. Many researchers have found satisfactory statistical relationships between different soil attributes and terrain attributes easily derived from DEM. Some of the promising indicators are pH, organic matter, carbonates, particle size distribution, color, bulk density, and depth to specific horizon boundaries. Soil depth was significantly correlated with slope angle and absolute and relative height. Soil depth and A-horizon depth were correlated with plan curvature, compound topographic index (CTI), and upslope mean plan curvature. Models that utilized only CTI explained 84% and 71% of variation in soil depth and A-horizon depth, respectively. Slope and wetness index accounted for half of the variability in A-horizon depth, sand content, and other soil properties. The increasing availability of high-resolution remote sensing data provides a new window for predicting soil characteristics with acceptable accuracy. Researchers have provided evidence regarding the contribution of remote sensing data in providing acceptable prediction of soil characteristics. Soil data represent a basic input of the Soil and Water Assessment Tool (SWAT), which is a semi-distributed process-based ecohydrological model used to simulate stream flow, crop yield, sediment transport, and nutrient transport. The majority of soil data are available as soil maps in polygon format, which tend to aggregate individual soil attributes and present soil information in a generalized form. The extraction of layers of individual soil attributes, which also reflect the spatial variability within these polygons, is not possible in most cases. Environmental modeling and other soil applications require proper representation of the spatial distribution of soil attributes and favor the representation of attributes as individual layers for each soil parameter to facilitate the integration with other layers of information. The approach described in this study is designed to help users generate higher resolution soil information to cover areas where soil data is not available or available at low resolution. Using digital elevation model and soil observations, the model generates spatially continuous representation of soil attributes that are available in format ready for use as an input to SWAT. This will also benefit users who are demanding spatially distributed soil information for various applications. The prediction accuracy of environmental models, such as SWAT, depends on how well the inputs describe the spatial characteristics of the watershed. For example, the use of different resolutions of soil data can significantly affect the accuracy of SWAT predictions. The use of high-resolution soil data can improve the accuracy of SWAT predictions, while the use of low-resolution soil data can lead to inaccurate predictions.