HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Figures Only Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Jana, R. B. Articles by Springer, E. P. PubMed Articles by Jana, R. B. Articles by Springer, E. P. GeoRef GeoRef Citation Agricola Articles by Jana, R. B. Articles by Springer, E. P. Related Collections Pedotransfer Functions Water Retention/Capillary Pressure Soil Hydrology

ORIGINAL RESEARCH

Multiscale Pedotransfer Functions for Soil Water Retention

^a Dep. of Biological and Agricultural Engineering, Texas A&M Univ., College Station, TX 77843
^b Earth & Environmental Sciences Division, Los Alamos National Lab., Los Alamos, NM 87845
^* Corresponding author (bmohanty{at}tamu.edu).

Received 23 March 2007.

Parametric soil water retention and hydraulic conductivity functions are often used for predicting soil hydrologic behavior using hydrologic, hydroclimatic, and contaminant transport models. The prediction accuracy of any such model is critically dependent on the quality of the input parameters. Limited availability of (detailed) soil hydraulic data for large-scale hydroclimatic models (with grids ranging from several kilometers to several hundred kilometers) is a major challenge. To address this need, pedotransfer functions (PTFs) have been used to estimate the required soil hydraulic parameters from other available or easily measurable soil properties. While most previous studies derive and adopt these parameters at matching spatial scales (1:1) of input and output data, we have developed a methodology to derive soil water retention functions at the point or local scale using the PTFs trained with coarser scale input data. This study was a novel application of an artificial neural network (ANN)-based PTF scheme across two spatial support scales within the Rio Grande basin in New Mexico. The ANN was trained using soil texture and bulk density data from the SSURGO database (scale 1:24,000) and then used for predicting soil water contents at different pressure heads with point-scale data (1:1) inputs. The resulting outputs were corrected for bias before constructing the soil water characteristic curve using the van Genuchten equation. A hierarchical approach with training data derived from multiple clustered subwatersheds (with varying spatial extent) was used to study the effect of the increase in spatial extent. The results show good agreement between the soil water retention curves constructed from the ANN-based PTFs and field observations at the local scale near Las Cruces, NM. The robustness of the multiscale PTF methodology was further tested with a separate data set from the Little Washita watershed region in Oklahoma. Overall, ANN coupled with bias correction was found to be a suitable approach for deriving soil hydraulic parameters at a finer scale from soil physical properties at coarser scales and across different spatial extents. The approach could potentially be used for downscaling soil hydraulic properties.

Abbreviations: ANN, artificial neural network • LW, Little Washita • PTF, pedotransfer function