A Comparison of Lightweight Deep Learning Models for Particulate-Matter Nowcasting in the Indian Subcontinent & Surrounding Regions

TL;DR

This study develops and compares lightweight deep learning models for 6-hour particulate matter nowcasting in the Indian subcontinent, demonstrating improved accuracy and efficiency over existing models using regional atmospheric data.

Contribution

It introduces three novel lightweight deep learning architectures tailored for particulate matter nowcasting in a specific regional domain, with comprehensive evaluation against baseline models.

Findings

Models outperform Aurora baseline in RMSE, MAE, Bias, and SSIM metrics.

Proposed models achieve rapid inference suitable for real-time applications.

Region-specific models improve forecast accuracy over generic models.

Abstract

This paper is a submission for the Weather4Cast~2025 complementary Pollution Task and presents an efficient framework for 6-hour lead-time nowcasting of PM$_1$, PM$_{2.5}$, and PM$_{10}$ across the Indian subcontinent and surrounding regions. The proposed approach leverages analysis fields from the Copernicus Atmosphere Monitoring Service (CAMS) Global Atmospheric Composition Forecasts at 0.4 degree resolution. A 256x256 spatial region, covering 28.4S-73.6N and 32E-134.0E, is used as the model input, while predictions are generated for the central 128x128 area spanning 2.8S-48N and 57.6E-108.4E, ensuring an India-centric forecast domain with sufficient synoptic-scale context. Models are trained on CAMS analyses from 2021-2023 using a shuffled 90/10 split and independently evaluated on 2024 data. Three lightweight parameter-specific architectures are developed to improve accuracy, minimize systematic bias, and enable rapid inference. Evaluation using RMSE, MAE, Bias, and SSIM demonstrates substantial performance gains over the Aurora foundation model, underscoring the effectiveness of compact & specialized deep learning models for short-range forecasts on limited spatial domains.