Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Environmental Microbial & Food Safety Laboratory » Research » Publications at this Location » Publication #275220

Title: Estimating spatial variations in water content of clay soils from time-lapse electrical conductivity surveys

Author
item MARTINEZ, GONZALO - Universidad De Cordoba
item Pachepsky, Yakov
item VANDERLINDEN, KARL - Venta Del Llano Ifapa Center
item GIRALDEZ, JUAN VICENTE - Universidad De Cordoba

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 10/31/2011
Publication Date: 12/5/2011
Citation: Martinez, G., Pachepsky, Y.A., Vanderlinden, K., Giraldez, J. 2011. Estimating spatial variations in water content of clay soils from time-lapse electrical conductivity surveys. [abstract]. Abstract book. H42F-05.

Interpretive Summary:

Technical Abstract: Soil water content (theta) is one of the most important drivers for many biogeochemical fluxes at different temporal and spatial scales. Hydrogeophysical non-invasive sensors that measure the soil apparent electrical conductivity (ECa) have been widely used to infer spatial and temporal patterns of theta. Empirical, time and location-dependent, regressions are usual to relate ECa and theta. These relations typically are weak in soils with high content of clay. The objective of this work has been to demonstrate an opportunity to improve the theta estimation in a clayey soil for any given time using spatially collocated time-lapse ECa surveys. The experiment was carried out at a field site with two different types of soil management. Total of 17 gravimetric soil water content sampling campaigns of the top 0.35 m and 13 topsoil ECa surveys have been performed. The studied periods were from January 2008 to May 2009 for theta and from March 2006 to February 2009 for ECa. Three of soil water content and ECa surveys were collocated in both in space and in time. The relationship between ECa and ' was modeled with multiple linear regression, with principal components of ECa used as the independent variables to prevent multicollinearity. The temporal evolution of ' showed a dry-wet-dry pattern. Different soil conditions lead to differences in ECa average values. The correlation coefficient between ECa and theta for collocated surveys was always lower than 0.03 for both managements. The lowest correlations were found when soil moisture was close to field capacity. Using more than one survey substantially improved estimation of theta for both soil management types. Using all possible sets of surveys for a given total number of surveys showed that, on average, the model accuracy increased with the total number of ECa surveys used in predictions. Generally, the accuracy of the regression models was lower and less variable when theta was higher than 0.25 kg kg-1. We hypothesize that the improvement in the regression accuracy was related to the indirect account for soil spatial variability that manifested itself in the ECa surveys in different ways on different survey times. The further testing of the procedure may lead to both improved estimation of theta and decreasing the number of or even forsaking collocated theta and ECa surveys.