AI-Driven Atmospheric Modeling and Environmental Prediction for Sustainable Climate Management

The increase in extreme weather events and environmental degradation demands new computational approaches for accurate forecasting and green management. Traditional numerical weather prediction (NWP) models are computationally expensive and struggle with resolution and data assimilation, despite being effective and high capacity. The proposal in this work is an Artificial Intelligence (AI) hybrid framework featuring machine learning (ML) and deep learning (DL) models with atmospheric and climate datasets to improve forecast accuracy, reduce computing costs and improve temporal-spatial resolution of environmental simulation. The framework consists of transformer models, spatiotemporal neural networks, and physics-constrained AI models to forecast weather events, air quality indices, and greenhouse gas (GHG) levels. Our project develops an extensible and sustainable AI pipeline that continuously learns from satellite images, sensor networks, and community open climate data sources (e.g., NASA EarthData, Copernicus). Expected outcomes are improved accuracy of short-term forecasts, reduced computational energy costs, and actionable insights for strategic decision making by policy-makers and intelligent city infrastructure planners.

Saoud Sarwar1, Aizad Khursheed2 and Abul Saeed Azad*3 Click here to download file

Research Review | Published online : 21-May-2026