Page Banner

United States Department of Agriculture

Agricultural Research Service

Title: Software to estimate –33 and –1500 kPa soil water retention using the non-parametric k-Nearest Neighbor technique)

item Nemes, Attila
item Roberts, Ralph
item Rawls, Walter
item Pachepsky, Yakov
item Van Genuchten, Martinus

Submitted to: Journal of Environmental Modeling and Software
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/29/2007
Publication Date: 9/27/2007
Citation: Nemes, A., Roberts, R.T., Rawls, W.J., Pachepsky, Y.A., Van Genuchten, M.T. 2007. Software to estimate –33 and –1500 kPa soil water retention using the non-parametric k-Nearest Neighbor technique. Journal of Environmental Modeling and Software. 23:254-255.

Interpretive Summary: The ability of soil to retain and to transmit water has to be known for many engineering, meteorological, agronomic, and hydrological applications. Measurements of soil hydraulic properties are costly and impractical for large scale projects. For such projects, soil hydraulic properties are estimated from publicly available basic soil data using statistical regression. Such regression equations have been developed in various regions of the World. The serious problem with these equations is that their applicability is unpredictable. We have empirically proven that information regarding soil properties can be used to reasonably predict hydraulic properties. The nearest-neighbor algorithm appears to be applicable to evaluate the similarity for this purpose. This algorithm has been coded in simple software to estimate water contents that are commonly associated with the ability of soil to hold water and with the dryness of soil causing wilting of plants. The substantial advantage of our software is that available local information on soil hydraulic properties can be easily incorporated in the similarity evaluation. The more that is known about local soil hydraulic properties, the better the estimation that can be obtained.

Technical Abstract: A computer tool has been developed that uses a k-Nearest Neighbor (k-NN) lazy learning algorithm to estimate soil water retention at –33 and –1500 kPa matric potentials and its uncertainty. The user can customize the provided source data collection to accommodate specific local needs. Ad hoc calculations make this technique a competitive alternative to published pedotransfer equations, as re-development of such equations is not needed when new data become available.

Last Modified: 05/24/2017
Footer Content Back to Top of Page