Open AccessAgricultural drought monitoring in Tamil Nadu in India using Satellite-based multi vegetation indices
Author(s) -
R. Kumaraperumal,
S. Pazhanivelan,
Ragunath. K.P.,
Balaji Kannan,
Prajesh. P.J.,
Mugilan. G.R.
Publication year - 2021
Publication title -
journal of applied and natural science
Language(s) - English
Resource type - Journals
eISSN - 2231-5209
pISSN - 0974-9411
DOI - 10.31018/jans.v13i2.2585
Subject(s) - normalized difference vegetation index , tamil , environmental science , agriculture , satellite , vegetation (pathology) , remote sensing , proxy (statistics) , raw data , climatology , climate change , geography , mathematics , statistics , geology , medicine , philosophy , linguistics , oceanography , archaeology , pathology , aerospace engineering , engineering
Drought being an insidious hazard, is considered to have one of the most complex phenomenons. The proposed study identifies remote sensing-based indices that could act as a proxy indicator in monitoring agricultural drought over Tamil Nadu's region India. The satellite data products were downloaded from 2000 to 2013 from MODIS, GLDAS – NOAH, and TRMM. The intensity of agricultural drought was studied using indices viz., NDVI, NDWI, NMDI, and NDDI. The satellite-derived spectral indices include raw, scaled, and combined indices. Comparing satellite-derived indices with in-situ rainfall data and 1-month SPI data was performed to identify exceptional drought to no drought conditions for September month. The additive combination of NDDI showed a positive correlation of 0.25 with rainfall and 0.23 with SPI, while the scaled NDDI and raw NDDI were negatively correlated with rainfall and SPI. Similar cases were noticed with raw LST and raw NMDI. Indices viz., LST, NDVI, and NDWI performed well; however, it was clear that NDWI performed better than NDVI while LST was crucial in deciding NDVI coverage over the study area. These results showed that no single index could be put forward to detect agricultural drought accurately; however, an additive combination of indices could be a successful proxy to vegetation stress identification.