HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Figures Only Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Eldeiry, A. A. Articles by Garcia, L. A. PubMed Articles by Eldeiry, A. A. Articles by Garcia, L. A. Agricola Articles by Eldeiry, A. A. Articles by Garcia, L. A. Related Collections Soil Salinity Alfalfa Remote Sensing

SOIL & WATER MANAGEMENT & CONSERVATION

Detecting Soil Salinity in Alfalfa Fields using Spatial Modeling and Remote Sensing

Integrated Decision Support Group, Dep. of Civil and Environmental Engineering, Colorado State Univ., Fort Collins, CO 80523-1372

^* Corresponding author (Luis.Garcia{at}colostate.edu).

A new methodology, which integrates field data, geographic information systems, remote sensing, and spatial modeling, was developed to accurately model soil salinity using statistical tools. Ground data from four alfalfa (Medicago sativa L.) fields in the lower Arkansas River basin in Colorado were compared with data derived from Ikonos satellite images with a 4-m resolution and Landsat satellite images with a 30-m resolution. For each image, the combination of satellite image bands that had the best correlation with soil salinity was used. Three statistical models were applied to estimate soil salinity from remote sensing: the ordinary least squares (OLS) model, the spatial autoregressive (spatial AR) model, and a modified residual kriging model. A number of criteria were evaluated to select the best model. The results show that both the Ikonos and Landsat satellite images can be used to estimate soil salinity, and regardless of the source of the satellite image used, the modified kriging model provided the best estimates of soil salinity. Although the OLS model met most of the model selection criteria, in most cases it involved some autocorrelation among the residuals. When the same data were tested using the spatial AR model, most of the autocorrelation among the residuals was removed, but the R² was reduced. In the modified kriging model, where the kriged residuals were combined with the results of the OLS model, there were significant improvements in the R² for all cases tested in this study. Thus, this study shows that combining field data, geographic information systems, and remote sensing with strong statistical measures can significantly enhance the development of high-quality soil salinity maps.

Abbreviations: AICC, Akaike's information corrected criterion • AR, autoregressive • NDVI, normalized difference vegetation index • NIR, near infrared • OLS, ordinary least squares • RSE, residual standard error