Measurement and zonation of soil surface moisture in arid and semi-arid regions using Landsat 8 images

نویسندگانرضا دهقانی بیدگلی,حمیدرضا کوهبنانی,علی کشاورزی,Vinod Kumar
نشریهARAB J GEOSCI
شماره صفحات1
شماره مجلد17
ضریب تاثیر (IF)1.327
نوع مقالهFull Paper
تاریخ انتشار2020-08-20
رتبه نشریهعلمی - پژوهشی
نوع نشریهالکترونیکی
کشور محل چاپایران
نمایه نشریهSCOPUS ,JCR

چکیده مقاله

Monitoring of soil surface moisture is an imperative factor in water and energy cycle. Due to the variability of soil characteristics such as topography, vegetation, and climate dynamics, this important factor varies with respect to time and place. Measuring methods can provide soil moisture information in a wide range of short intervals with reasonable accuracy. In present research, Landsat 8 satellite data with various soil moisture content estimation methods were tested. In order to evaluate the accuracy of each method, the real-field data used 80 samples of volumetric soil moisture content in suburban areas of Semnan city that were collected at the time of satellite passage of the area. Some of the indicators used in this study are normalized vegetation index, NDTI index, NDMI index, PSMI index (use full form of these indices), surface temperature, and SMSWIR index. The SMSWIR index with correlation coefficient was 0.78, and the correlation coefficient of regression model was 0.61, and RMSE was 3.69. The results of the regression model and real data were estimated to be 3.69, which are recommended for assessing surface soil moisture in arid and desert regions. Three indicators of SMSWIR index, NDTI index, and NDMI index with a small difference are not suitable indices for measuring soil moisture content in desert areas with vegetation cover. By employing multivariable regression models, soil moisture model was also prepared by using the studied indices. The findings of this research indicate that the simultaneous correlation model is superior to the surface soil moisture mapping

tags: Remote sensing . SMSWIR . LST . Landsat . Multivariate regression models