Predictions of saturated hydraulic conductivity dynamics in a midwestern agricultural watershed, Iowa

In this study, a physically-based, modeling framework was developed to predict saturated hydraulic conductivity K_sat) dynamics in the Clear Creek Watershed (CCW), IA. The modeling framework integrated selected pedotransfer functions (PTFs) and watershed models with geospatial tools. A number of PTFs and watershed models were examined to select the appropriate models that represent the study site conditions. Models selection was based on statistical measures of the models' errors compared to the K_sat field measurements conducted in CCW under different soil, climatic and land use conditions. The study has shown that combined Rosetta and the Water Erosion Prediction Project (WEPP) predictions provided the best agreement to the measured K_sat values in the CCW compared to the other tested models. Therefore, Rosetta and WEPP were integrated with the Geographic Information System (GIS) tools by developing a program for data registries. The modeling framework allowed for visualization of the data in forms of geospatial maps and prediction of K_sat variability in CCW due to the seasonal changes in climate and land use activities.

Two seasons were selected to demonstrate K _sat dynamics; specifically, the months of October and April, which corresponded to the before harvesting and before planting conditions, respectively. Baseline saturated hydraulic conductivity (K _b) exhibited higher values on the northern part of the CCW compared to the southern part due to differences in soil texture. For bare saturated hydraulic conductivity (K_br), the month of April had overall higher values than the month of October, because K_br is inversely proportional to cumulative rainfall kinetic energy and spring season in Iowa are characterized by lower precipitation compared to the fall season.

Except for the ungrazed grassland areas, effective K_sat that accounts for land cover only (K_e-nr) did not change significantly with season, exhibiting the lowest values at the forest and urbanized areas in the CCW. The corn fields showed lower K_e-nr values than soybean fields due to different characteristics of the crops. The effects of rainfall on effective saturated hydraulic conductivity (K_e) were demonstrated by the single storm events of October 17^th, 2007 and April 18^th, 2008. The former showed higher K_e in CCW, because K_e is linearly proportional to rainfall depth and the October event had higher precipitation than the April event.

Statistical analysis of the K_sat data in CCW has shown that the geometric mean or median was more representative for the distributions of different expressions of saturated hydraulic conductivities due to their wide ranges. The values of K_b were the highest values among the other expressions of K_sat. K_e-nr values were smaller than K_b values, while K_e values were higher than K_e-nr.

The applicability of the pedotransfer functions and watershed models used within the developed modeling framework is limited to the investigated watershed and other watersheds in Iowa having similar soils, management practices, and climatic conditions, mostly in the semihumid region of eastern Iowa. As the proposed modeling framework was able to successfully capture the spatial and temporal variability of K_sat at the watershed scale, it would be advisable to repeat this study in different counties or even in other parts of the country, where arid or semi-arid conditions are ubiquitous, using different pedotransfer functions and watershed models. This can contribute to the development of ratings for many of the soil interpretations incorporated into the National Cooperative Soil Survey (NCSS) and update the K_sat data stored in the National Soil Information System (NASIS) database.