Work

An Exploratory Analysis on the Use of AI Technology in predicting weather, climate modelling and Disaster forecasting

Year: 2026
Type: article
Abstract: Weather and climate extremes are increasing in frequency, severity and duration, placing growing pressure on ecosystems, economies and human lives. Accurate prediction is therefore essential for preparing countries against climate-driven disasters. This paper presents an exploratory analysis of the use of Artificial Intelligence (AI) in weather prediction, climate modelling and disaster forecasting, and examines how nations can act on these forecasts to reduce adverse impacts. Using a mixed-method approach, the study combines quantitative comparison, presented in tables and graphs, with qualitative interpretation drawn from authentic secondary sources. It compares leading AI models, including GraphCast, Pangu-Weather, AIFS, FourCastNet, Aurora, FuXi and FengWu, with conventional numerical weather prediction systems such as the IFS, GFS, WRF and the Unified Model. It also reviews AI-based downscaling of climate projections over India and the role of AI in multi-hazard early warning systems. The findings show that AI models match or outperform traditional systems on many forecast variables while producing forecasts far faster and at lower computational cost. However, they tend to underestimate the frequency and severity of record-breaking extremes, where physics-based models still perform better. Challenges remain in data quality and bias, model explainability, energy consumption and unequal access to technology. The paper further finds that accurate forecasts reduce disaster impact only when integrated into trusted, community-centred early warning systems guided by principles of fairness, accountability, transparency, ethics and sustainability. The paper concludes that a hybrid approach, combining the physical grounding of traditional models with the speed and pattern-recognition ability of AI, offers the most reliable route to anticipating extreme events. It recommends stronger observation networks, explainable and ethical AI, regional capacity-building, sustained funding and peoplecentred system design. Ultimately, AI complements rather than replaces meteorological science, and its value depends on how well it is integrated with governance, local context and community needs.
Language: English
Open Access status: diamond
OpenAlex ID: W7223180285
FWCI:
Cites: 7
Cited by:
Related to: 9