Astronomy & Space Sciences
Title : | Precipitation type classification with INSAT 3D/3DR Satellite Observations using Explainable AI/ML (XAI/ML) |
Area of research : | Astronomy & Space Sciences |
Principal Investigator : | Dr. Shruti Ashok Upadhyaya, Indian Institute Of Technology (IIT) Hyderabad, Telangana |
Timeline Start Year : | 2024 |
Timeline End Year : | 2026 |
Contact info : | shrutiau@ce.iith.ac.in |
Equipments : | Real-time data storage
UPS devices, Institute HPC services and any other field data
Workstation
Realtime Data Monitoring device |
Details
Executive Summary : | Accurate and timely precipitation information is crucial for understanding Earth's water and energy cycles, identifying socio-economic effects of fresh water needs, forecasting extreme climate impacts, and adapting to long-term climate change. The Earth Science Decadal Survey highlights the importance of quantifying precipitation occurrence and rates in the coming decade. The Indian Space Research Organization (ISRO) is a leader in providing long-term access to geostationary satellite observations (INSAT series), which provide unprecedented data at high temporal scale and larger spatial scale. This work aims to develop a precipitation type classification model using state-of-the-art machine learning (ML) techniques using the latest Indian geostationary satellite observations INSAT 3D/3DR. The objective is to understand and study the impact of higher multi-spectral channel observations in INSAT 3D/3DR and its higher temporal resolution (~15min) compared to its previous generation satellite Kalpana-1. The research product will be a research product with ML-based precipitation type product to complement existing INSAT meteorological products. This convective/stratiform classification scheme will aid in better quantification of precipitation from Indian geostationary satellites and can be easily integrated with current operational precipitation products such as Hydro-Estimator and INSAT Multispectral Rainfall Algorithm (IMSRA). The study will also be valuable for other applications such as classifying storm types, identifying extreme events, calibrating satellite biases, and evaluating numerical models. |
Total Budget (INR): | 24,13,300 |
Organizations involved