Global warming modelling simulation based on the numerical weather prediction system

Accurately and efficiently simulating the climate and predicting the weather are universal goals in the realm of human advancement. Despite its status as the gold standard, numerical weather prediction (NWP) faces challenges due to inherent atmospheric uncertainty and high processing costs, particularly in the post-Moore's Law era. This article summarises the most significant models and noteworthy advances in climate modelling and data-driven weather forecasting. These models reduce prediction times from hours to seconds, outperforming state-of-the-art NWP techniques in over 90% of the variables. Data-driven climate models can accurately reproduce climate patterns across periods ranging from decades to centuries, significantly reducing computational effort and increasing efficiency. However, despite their numerous advantages, data-driven techniques also have notable limitations. These include difficulty in interpreting forecasts, challenges in evaluating model uncertainty, and overly cautious predictions under extreme conditions. The proposed system achieves an accuracy of 96.7%.