Author
Behrman, Kathrine | |
WILLIAMS, J - Texas Agrilife | |
Kiniry, James | |
NORFLEET, M - Natural Resources Conservation Service (NRCS, USDA) | |
TAYLOR, R.A.J. - Texas A&M Agrilife |
Submitted to: Annual International SWAT Conference
Publication Type: Abstract Only Publication Acceptance Date: 5/5/2014 Publication Date: N/A Citation: N/A Interpretive Summary: Technical Abstract: Process-based simulation models ALMANAC/APEX/EPIC/SWAT contain generalized plant growth subroutines to predict biomass and crop yield. Environmental constraints typically restrict plant growth and yield. Water stress is often an important limiting factor; it is calculated as the sum of water use from each soil layer divided by the potential plant evapotranspiration. The plant available water in each soil layer is estimated from the difference between water volume at field capacity and wilting point. Reliable estimates of the plant available water are essential for accurate estimates of plant growth and water use. Several pedotransfer methods have been developed to estimate field capacity and wilting. We tested the ability of three methods (Rawls, Baumer, Norfleet, Nearest Neighbor, and KD Tree) to estimate field capacity and wilting point based on commonly measured soil properties (%sand, %silt, %clay, % organic carbon, bulk density, and cation exchange capacity). The Rawls, Baumer, and Norfleet methods compute field capacity and wilting directly from properties of the selected soil, whereas the Nearest Neighbor and K-D Tree methods lookup these values from a database using %sand, %silt, and % organic carbon. Each method was tested for 2,039 cropland soil profiles from the NRCS National Soil Characterization Database. The ability of the five methods to estimate wilting point and field capacity were assessed for accuracy and their relative processing speeds were compared. The relative value of these computational methods for process-based biophysical models will be discussed. |