Abstract

Drought is a critical environmental issue that affects agriculture, water resources, and ecosystems. Traditional drought monitoring methods rely on ground-based meteorological observations, which have limited spatial coverage and do not provide real-time assessments. This project aims to develop a Vegetative Drought Prediction System by integrating Vegetation Condition Index (VCI) data from the ISRO VEDAS VCI Dashboard, remote sensing indices (NDVI), meteorological drought indicators (SPI, PDSI), and machine learning algorithms (Random Forest, SVM, LSTM) to accurately detect and predict drought conditions. The system utilizes VCI data from the ISRO VEDAS portal, which provides high-resolution drought monitoring for India. Additionally, satellite imagery from MODIS, Sentinel-2, and Landsat-8 is used to compute NDVI (Normalized Difference Vegetation Index) to assess vegetation health. Meteorological indices, such as the Standardized Precipitation Index (SPI) and Palmer Drought Severity Index (PDSI), are used to evaluate precipitation deficits and long-term drought trends. Machine learning models further enhance drought prediction. Random Forest (RF) and Support Vector Machine (SVM) classify drought severity based on VCI, NDVI, temperature, and rainfall data. Long Short-Term Memory (LSTM) networks analyze time-series data to forecast future drought conditions. These models are trained on historical drought records and meteorological data to improve accuracy. A web-based application is developed to visualize drought conditions, enabling users to select a region, analyze VCIbased drought indices, and receive real-time predictions. The frontend is built using HTML, CSS, and JavaScript (React.js), while the backend is implemented using Flask/Django, with PostgreSQL/MongoDB for data storage. This project provides an efficient, scalable, and cost-effective solution for drought monitoring using ISRO VEDAS VCI data, benefiting farmers, policymakers, and researchers in making informed decisions and enhancing drought preparedness.