Premium
Spatial evaluation of soil salinity using the WET sensor in the irrigated area of the Segura river lowland
Author(s) -
de Paz José M.,
Visconti Fernando,
Rubio José L.
Publication year - 2011
Publication title -
journal of plant nutrition and soil science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.644
H-Index - 87
eISSN - 1522-2624
pISSN - 1436-8730
DOI - 10.1002/jpln.200900221
Subject(s) - kriging , soil science , geostatistics , environmental science , spatial variability , salinity , soil test , hydrology (agriculture) , soil salinity , sampling (signal processing) , interpolation (computer graphics) , multivariate interpolation , variogram , soil water , mathematics , geology , statistics , geotechnical engineering , oceanography , filter (signal processing) , bilinear interpolation , animation , computer graphics (images) , computer science , computer vision
Abstract The electrical conductivity of the water within the soil pores (EC p ) measured with the WET sensor, appears to be a reliable estimate of soil salinity. A methodology combining the use of the WET sensor along with geostatistics was developed to delimit and evaluate soil salinity within an irrigated area under arid to semiarid Mediterranean climate in SE Spain. A systematic random sampling of 104 points was carried out. The association between EC p and the saturation‐extract electrical conductivity (EC se ) was assessed by means of correlation analysis. The semivariograms for EC p were obtained at three different soil depths. Interpolation techniques, such as ordinary kriging and cokriging, were applied to obtain EC p levels in the unknown places. For each one of the soil depths, a model able to predict EC se from EC p was developed by means of ordinary least squares regression analysis. A good correlation ( r = 0.818, p < 0.001) between EC p and EC se was found. Spherical spatial distribution was the best model to fit to experimental semivariograms of EC p at 10, 30, and 50 cm soil depths. Nevertheless, cokriging using the EC p of an adjacent soil depth as an auxiliary variable provided the best results, compared to ordinary kriging. An analytical propagation‐error methodology was found to be useful to ascertain the contribution of the spatial interpolation and ordinary least squares analysis to the uncertainty of the EC se mapping. This methodology allowed us to identify 98% of the study area as affected by salinity problems within a rooting depth of 50 cm, with the threshold of EC se value at 2 dS m –1 . However, considering the crops actually grown and 10% potential reduction yield, the soil‐salinity‐affected area decreased to 83%. The use of sensors to measure soil salinity in combination with geostatistics is a cost‐effective way to draw maps of soil salinity at regional scale. This methodology is applicable to other agricultural irrigated areas under risk of salinization.