Skip to main content
ARS Home » Research » Publications at this Location » Publication #112863

Title: DATABASE-RELATED ACCURACY AND UNCERTAINTY OF PEDOTRANSFER FUNCTIONS.

Author
item SCHAAP, MARCEL - UC RIVERSIDE, CA
item LEIJ, FEIKE - UC RIVERSIDE CA
item Van Genuchten, Martinus

Submitted to: Soil Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/15/1998
Publication Date: N/A
Citation: N/A

Interpretive Summary: Soil hydraulic properties (water retention, hydraulic conductivity) are necessary to compute the movement of water and dissolved chemicals in the unsaturated zone between the soil surface and the groundwater. Measurement of soil hydraulic properties is often difficult, time consuming, and expensive. Therefore, pedotransfer functions (PTFs) are becoming more popular to predict soil hydraulic properties from surrogate date such as soil texture, bulk density and organic matter content. PTFs were developed on three different soil hydraulic databases. In addition, we developed PTFs on all available data. Results showed that the PTFs often have systematically different predictions of water retention and saturated hydraulic conductivity. Further, the uncertainty associated with the predictions became smaller when the size of the database increased. We conclude that the performance of PTFs may strongly depend on the data that were used for calibration and evaluation. However, PTFs developed on large databases are more likely to be generally applicable than PTFs based on small databases. The uncertainty estimates that were presented in this study allow insight in the reliability of the prediction, which could be used to compute confidence intervals of simulated fluxes of water and dissolved matter.

Technical Abstract: Pedotransfer functions (PTFs) are becoming more popular to predict soil hydraulic properties from soil texture, bulk density and organic matter content. Thus far, the calibration and validation of PTFs has been hampered by a lack of suitable databases. In this paper we employed three databases (RAWLS, AHUJA and UNSODA) to evaluate accuracy and uncertainty of neural network-based PTFs. Sand, silt and clay percentages and bulk density were used as input for the PTFs which subsequently provided retention parameters and saturated hydraulic conductivity, KS as output. Calibration and validation of PTFs were carried out on independent samples from the same database through combination with the bootstrap method. This method also yielded the possibility of calculating uncertainty estimates of predicted hydraulic parameters. Calibration and validation results showed that water retention could be predicted with a root mean square residual (RMSR) between 0.06 and 0.10 cm3/cm3; the RMSR of log(Ks) was between 0.4 and 0.7 log (cm/day). Cross-validation was used to test how well PTFs that were calibrated for one database could predict hydraulic properties of the other two databases. The results showed that systematically different predictions were made while the RMSR values increased to between 0.08 and 0.13 cm3/cm3 or water retention and to between 0.6 and 0.9 log(cm/day) for log(Ks). The uncertainty in predicted KS was half to one order of magnitude while predicted water retention points had an uncertainty of about 0.04 to 0.10 cm3/cm3. Uncertainties became somewhat smaller if the PTFs were calibrated on all available data. We conclude that the performance of PTFs may strongly depend on the data that were used for calibration and evaluation.