Coupling a micro-genetic algorithm with RegCM5 for improving extreme precipitation simulations over Southeast Asia

Regional climate models (RCMs) are essential for producing fine-scale climate information, but their effectiveness is highly sensitive to the combination of physical parameterizations and optimal settings of key parameters. To tackle this problem, this study develops a coupled modeling system that integrates a micro-genetic algorithm (μGA) with the Regional Climate Model version 5 (RegCM5), focusing on optimizing parameters in the Tiedtke convection scheme, crucial for precipitation simulations. Using the benchmarking version of RegCM5 for Southeast Asia, we aim to identify the optimal parameter set that enhances performance for three extreme precipitation events. The evaluation of this parameter set is then conducted by simulating six additional extreme events. Results show that simulations with optimized parameters improve both precipitation and temperature compared to the default model, significantly reducing biases, particularly over ocean regions. Our coupled RegCM5-μGA system will aid the broader RegCM5 community in enhancing model performance in their target regions.