Development of a Remote Sensing-Based Agriculture Monitoring Drought Index and Its Application Over Semi-Arid Region
Abstract
Agricultural drought is a natural disaster that usually occurs when the available water content goes below the optimal needs of the proper growth of plants during its growing season. It has enormous impacts on economic, environmental, and social sectors. In this study, our overall objective was to develop a fully remote sensing-based method for monitoring agricultural drought conditions and evaluate its performance over a semi-arid heterogeneous rainfed agricultural dominant landscape in Jordan. In general, remote sensing data having both high spatial and temporal resolutions would be required for evaluating agricultural drought conditions, as usually agriculture land cover would be relatively heterogeneous and small in size, while drought could occur during critical short time periods i.e., few days or weeks during the growing season. However, due to different technical and cost issues such high spatio-temporal remote sensing data are still unavailable. Thus, we opted to develop a spatio-temporal image-fusion model (STI-FM) to generate synthetic Landsat-8 like data with 30 m spatial and 8 day temporal resolutions upon combining regular Landsat-8 (having 30 m spatial with 16 day temporal resolutions) with moderate-resolution imaging spectroradiometer (MODIS)-based 8-day composite data having 250-1000 m spatial resolutions. Then, we used these fused data in developing the agricultural drought monitoring index (ADI) as a combination of three uncorrelated remote sensing-based agricultural drought related variables [i.e., normalized difference water index (NDWI), visible and shortwave drought index (VSDI), and land surface temperature (LST)]. Results showed that the proposed STI-FM was able to produce synthetic Landsat-8 data with strong accuracy (i.e., r2 were in the range 0.71 to 0.90). The evaluation of agricultural drought conditions over the study area using the proposed remote sensing-based agricultural drought index showed high agreements such as 85% overall accuracy and 78% Kappa-values, when compared to ground based 8-day standardised precipitation index (SPI) values. These strong results demonstrated that the proposed methods would be great in monitoring agricultural drought conditions at agricultural field scale (i.e., high spatial resolution) and short time periods (i.e., high temporal resolution).
Description
Keywords
Physical Geography, Remote Sensing, Chemistry--Agricultural
Citation
Hazaymeh, K. (2016). Development of a Remote Sensing-Based Agriculture Monitoring Drought Index and Its Application Over Semi-Arid Region (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/25672