ased on drought records from the Department of Agriculture for the last 5 years (2017 – 2021), the area of agricultural land in Madiun Regency experiencing agricultural drought is 320 ha. Identification of agricultural drought is very important as a disaster mitigation effort, the use of the TVDI method that utilizes Landsat-8 OLI/TIRS imagery will accelerate spatial information of the drought-affected area. The TVDI or Temperature Vegetation Dryness Index method uses the main parameters in the form of a vegetation index or NDVI (Normalized Difference Vegetation Index) and land surface temperature or LST (Land Surface Temperature). The use of Landsat-8 imagery with the TVDI method can calculate the accuracy of the confusion matrix from the NDVI and LST algorithms obtained by the agricultural drought index. The results of this study showed that the spatial distribution of agricultural drought in Madiun Regency was divided into five classes, namely wet, slightly wet, normal, slightly dry, and dry. The most dominant drought class area in 2019 and 2020 is the normal drought class with an area of 15,426.64 ha and 13,960.01 ha. The distribution of this normal drought class is in the west, which indicates that even though it is located in the lowlands, the existing irrigation channels are running quite well, so that it has a normal class. In 2021, the dominant drought class is slightly wet with an area of 13,641.93 ha. The accuracy value of land use and NDVI using the confusion matrix produces an accuracy value of 92.89% for land use. The NDVI accuracy in 2019 was 91.45%, in 2020 it was 87.13%, and in 2021 it was 91.28%. The LST accuracy value in 2019 is 96.38%, in 2020 it is 97.69%, and in 2021 it is 95.71%.

Keywords: TVDI, Agricultural Drought, NDVI, LST