Context: As weather patterns grow more unpredictable due to the climate crisis, India has launched Mission Mausam to improve weather understanding and forecasting through expanded observation networks, better modelling and advanced tools like AI and machine learning.
Mission Mausam: An Overview
Launch Year: 2024
Budget: ₹2000 crores over two years
Initiative by: Ministry of Earth Sciences
Aim:
To enhance weather and climate services across sectors like agriculture, disaster management, rural development, and more — with the long-term goal of making India weather-ready and climate-smart.
Key Objectives of Mission Mausam
Enhanced Observational Networks:
Expand both in-situ and remote sensing infrastructure using advanced radars, satellites, and automated weather stations.AI & High-Performance Computing (HPC):
Use Numerical Models, Artificial Intelligence, and Machine Learning to improve data assimilation and generate more accurate forecasts.Multi-Scale Forecasting Capability:
Improve prediction abilities across:Short-term
Medium-term
Extended-range
Seasonal timescales
Sector-Specific Advisories:
Deliver actionable weather-based insights for:Agriculture
Water Resources
Energy
Public Health
Disaster Management
Implementation Strategy
1. Infrastructure Development
Installation of Doppler Weather Radars, Automatic Weather Stations, and Rain Gauges nationwide.
2. Supercomputing Power
Use of India’s top-tier HPC systems — Pratyush and Mihir — for climate and weather modeling.
3. Collaborative Research
Partner with international organizations such as the World Meteorological Organization (WMO) to advance forecasting techniques.
4. Public Outreach
Disseminate forecasts through:
Mobile Apps (e.g., Mausam App)
SMS Services
TV and Media Channels
Implementation Phases
Phase 1 (Till March 2026)
Focus: Expansion of observation networks
Add ~70 Doppler radars
Deploy high-performance computers
Set up 10 wind profilers and 10 radiometers
Phase 2
Focus: Enhanced observational capability
Incorporate satellites and airborne platforms (aircraft) for advanced data gathering
Cloud Chamber at IITM, Pune
Timeline: To be established within 1.5 years
Purpose: Study cloud processes in the context of climate change and rising temperatures
Key Features & Functions
Creation of artificial clouds in a lab setting
Research focus:
Identify which clouds can be seeded
Determine ideal seeding materials
Assess optimal seeding quantities for enhancing or suppressing rain/hail
Examine how rising temperatures affect:
Cloud dynamics
Electrical activity
Rainfall patterns
Outcome
Better parameterisation in weather models
Scientific basis for artificial rain enhancement or suppression in the next 5 years
Traditional vs AI-Based Weather Forecasting
1. Traditional Weather Forecasting
Approach:
Physics-based models simulate atmospheric processes using complex mathematical equations.Data Sources:
Relies on observational data from weather stations, satellites, and radars (e.g., temperature, pressure, wind, humidity).Features:
Highly computationally intensive
Can take hours to days for model runs
Limited in resolving localized and chaotic phenomena due to the inherent non-linearity of the atmosphere
Example:
Numerical Weather Prediction (NWP) models like the Global Forecast System (GFS) or European ECMWF
2. AI-Based Weather Forecasting
Approach:
Data-driven models use Machine Learning (ML) or Deep Learning to identify patterns from historical and real-time data.Strengths:
Learns from correlations between variables like wind, ocean temp, humidity, and outcomes like rainfall or cyclone intensity
Captures non-linear relationships often missed by traditional models
Can be faster once trained
Example:
Models like GraphCast (by Google DeepMind), which use AI for medium-range forecasting
3. Challenges in AI-Based Forecasting
| Challenge | Details |
|---|---|
| Data Quality & Availability | Requires clean, large datasets. Sensor errors, inconsistent formats, and remote area gaps limit effectiveness. |
| Human Resource Gap | Few professionals with expertise in both climate science and AI/ML. |
| Interpretability & Trust | AI models often act as black boxes, making it hard to explain and trust predictions. |
| Infrastructure Needs | Requires GPU-based computing and high-end infrastructure for training and real-time deployment. |
4. The Way Forward: Hybrid Models
Hybrid Approach:
Combine the interpretability of traditional physics-based models with the adaptability and speed of AI/ML.Goal:
Retain scientific transparency
Improve accuracy
Address challenges like uncertainty and bias