Agricultural Drought Monitoring Using Multispectral Satellite Data: A Case Study of North Central Province in Sri Lanka

Abstract

Agriculture is the main field of the economy that brings profit to Sri Lanka, and Sri Lanka is also considered as the grain storage of South Asia. Sri Lankan agriculture mainly depends on crops like paddy, tea, etc. Therefore, it is vital to identify the problems that can cause damage to the field of agriculture. One such problem is agricultural droughts, which are caused by the shortage of water for agricultural purposes. There have been many indices developed for the monitoring of the phenomena related to the agricultural fields in remote sensing. This research is mainly focused on monitoring agricultural droughts with the use of soil moisture and some selected vegetation indices that were derived from Landsat multispectral data. It has chosen the Soil Moisture Index (SMI) and some vegetation indices that have an influence to the growth of plants. SMI was computed using the relationship between the Normalized Difference Vegetation Index (NDVI) and the Land Surface Temperature (LST) and can measure the soil moisture of the top layer of soil. SMI and the Delta index values were compared with the Standardized Precipitation Index (SPI) values taken from the Department of Meteorology Sri Lanka, which have been currently used for drought monitoring purposes. Both indices showed different correlations with the SPI values, and SMI showed the highest correlation of 0.39 with the SPI – 1-month data of the study area. Then with the vegetation indices, SMI showed high correlation with Land Surface Water Index (LSWI) than the Enhanced Vegetation Index (EVI) and the Soil Adjusted Water Index (SAVI). The regression model between the vegetation indices and the SMI gave the R squared value of 0.63. And the regression model also emphasizes that the SMI shows a high correlation with the LSWI index in the North Central Province, as the regression coefficient for that index was higher than others. Hence, it was evident that soil moisture alone cannot be used for agricultural drought monitoring, and there are other parameters that can affect crop production.