Regional Climate

Regional climate modeling aims to understand and predict climate variability and change at finer spatial scales than global models allow, supporting improved impact assessments and adaptation strategies. Current research heavily utilizes machine learning, particularly deep learning architectures like diffusion models and Generative Adversarial Networks (GANs), to efficiently emulate high-resolution regional climate model outputs, addressing the high computational cost of traditional simulations. This focus includes developing methods to improve the transferability and explainability of these emulators across different global climate models and scenarios, as well as employing machine learning for parameter sensitivity analysis to optimize model performance. The resulting high-resolution climate projections are crucial for applications such as flood risk assessment and heatwave impact studies.