Repository logo
 

Use of remote sensing to estimate soil salinity and evapotranspiration in agricultural fields

Date

2007

Authors

Elhaddad, Aymn, author
Garcia, Luis A., advisor
Loftis, Jim C., committee member
Albertson, Maurice L., committee member

Journal Title

Journal ISSN

Volume Title

Abstract

In recent years, methods for detecting soil salinity have improved greatly. This research describes methods to detect soil salinity levels in agricultural lands based on crop conditions and evapotranspiration (ET) using satellite imagery. Elevated levels of soil salinity affect the growth of most crops as well as their appearance. For this research, satellite images of the study area, the Arkansas River Basin in Colorado, are used to classify the condition of the crops being grown in fields according to their different reflectance values. Using spatially referenced ground data collected in the study area, each class in the satellite image is related to a level of soil salinity. These classes are then used to create a signature file to classify other areas within the same image having the same crop. For the purpose of detecting soil salinity in this study, two satellite scenes were used: a multi-spectral Ikonos image from July 27, 2001 and a Landsat 7 image from July 8, 2001. While the Ikonos image provides more details, the results of this study indicate that the Landsat imagery also performed remarkably well. Evapotranspiration (ET) is one of the processes that are affected by soil salinity. Reliable estimates of evapotranspiration from vegetation are needed for investigations of the relationship between soil salinity and ET. Satellite-derived information has been found useful for estimation of aerial ET. For this purpose, a surface energy balance-based model (RESET) was developed using remotely sensed data from satellite imagery. The RESET model takes into consideration the spatial variability in weather. Moreover, the model implements a spatiotemporal interpolation methodology in order to obtain ET information between satellite scenes. The RESET model was applied to estimate ET values in the study area. A geographic information system (GIS) was used to spatially relate the ET values to soil salinity data. The ET values were regressed against the spatially corresponding soil salinity values to develop a relationship between ET and soil salinity. The ET values were found to correlate well with the soil salinity levels in the study area, with correlation coefficients of up to 0.92.

Description

Pagination errors include two pages numbered ix and an omitted page number 2. No text is repeated or missing.

Rights Access

Subject

Citation

Associated Publications