A soil moisture accounting-procedure with a Richards' equation-based soil texture-dependent parameterization.

摘要

Given a time-series of potential evapotranspiration and rainfall data, there are at least two approaches for estimating vertical percolation rates. One approach involves solving Richards' equation (RE) with a plant uptake model. An alternative approach involves applying a simple soil moisture accounting procedure (SMAP) based on a set of conceptual stores and conditional statements. It is often desirable to parameterize distributed vertical percolation models using regional soil texture maps. This can be achieved using pedotransfer functions when applying RE. However, robust soil texture based parameterizations for more simple SMAPs have not previously been available. This article presents a new SMAP designed to emulate the response of a one-dimensional homogenous RE model. Model parameters for 231 different soil textures are obtained by calibrating the SMAP model to 20 year time-series from equivalent RE model simulations. The results are then validated by comparing to an additional 13 years of simulated RE model data. The resulting work provides a new simple two parameter (% sand and % silt) SMAP, which provides consistent vertical percolation data as compared to RE based models. Results from the 231 numerical simulations are also found to be qualitatively consistent with intuitive ideas concerning soil texture and soil moisture dynamics. Vertical percolation rates are found to be highest in sandy soils. Sandy soils are found to provide less water for evapotranspiration. Surface runoff is found to be more important in soils with high clay content.