ARS | Publication request: Toward improving estimates of field soil water capacity from laboratory-measured soil properties

Submitted to: World Congress of Soil Science
Publication Type: Proceedings
Publication Acceptance Date: February 3, 2010
Publication Date: August 1, 2010
Citation: Pachepsky, Y.A., Timlin, D.J., Nemes, A. 2010. Toward improving estimates of field soil water capacity from laboratory-measured soil properties. World Congress of Soil Science. 182-185 Published on DVD.

Interpretive Summary: Field capacity (FC), defined as the amount of water remaining in soil after the soil has been saturated and then allowed to drain thoroughly, is a very important soil hydraulic parameter that has multiple uses in hydrological, meteorological, agronomical, and environmental predictions and modeling. The customary way to estimate FC is to utilize laboratory equipment that applies a specific pressure to the soil and subsequently measure the water content. However, there is no one standard pressure accepted world wide. As a consequence, it is difficult to compare published FC values generated in different countries for similar soil types. Based on a large USDA-ARS database consisting of FC values determined concurrently with an exhaustive number of other soil properties (determined both in the field and in the lab), we discovered a soil texture-dependent correlation between pressure and measured FC. Utilizing this correction factor, were able to improve the accuracy of FC estimates by about 70%. The results of this work are important for large scale predictions where the ability of soils to retain water has to be quantified and used.

Technical Abstract: Different recommendations exist world-wide on which – if any - pressure head should be used in laboratory measurements to approximate the ‘field capacity’ (FC) of the soil. Literature often deems any such pressure heads to be inadequate to approximate FC for soils of all textures. We used a data collection from the literature to evaluate if corrections can be made to improve the estimation of FC from -33 kPa water retention (W33). Regression tree modeling coupled with jack-knife cross validation was used to identify the best predictors – sand, silt, clay and the measured W33 value – to estimate the difference between W33 and FC. Such predictions were then successfully used to adjust the W33 value as the estimate of FC. An improvement in estimating FC was seen in general statistical terms, and texture specific bias was also greatly reduced. Such solution may allow the reliable use of a single pressure head in the laboratory to approximate FC, which may be the only feasible option for large scale studies.