Modification of Temperature Vegetation Dryness Index (TVDI) Method for Detecting Drought with Multi-Scale Image

Abstract

The objective of this research is to assess the accuracy of Temperature Vegetation Dryness Index (TVDI) methods applied to Principal Component Analysis (PCA) and multi-scale images. The TVDI method will revamp with PCA in vegetation and surface temperature variables. Each variable has three algorithms, which are VCI, NDWI, and SAVI, for vegetation, and TCI, CWSI, and LST for surface temperature. The band input used was the PC1 resulted from PCA in each variable. The regression relationship between vegetation and surface temperature with PCA shows an average value of 0.99. The results of the PCA increased drought area throughout the research area and showed a negative relationship on the TVDI concept. Validation uses TRMM data for MODIS images and field surveys for Landsat imagery. Landsat showed an accuracy value of 75% and influenced by climate change. Besides, multi-scale imaging proves very useful in monitoring and mapping droughts.