Abstract
This “Shor Simulations” task explores the feasibility and usefulness of short (2-10 days) simulations for the purpose of efficient and effective testing, tuning and evaluation of high-resolution models. Based on the established framework for short simulations tests, in the past two quarters, we started to estimate the maximum likelihood of model parameter space using efficient-sampling-based short ensemble simulations. More particularly, we calculated the posterior joint 2D marginal distribution for liquid water content (LWC), cloud fraction, and shortwave (SW) forcing using surrogate model based on short ensemble simulation results and corresponded observations. We also identified optimal parameter sets with various objectives, e.g. LWC, Cloud Fraction, and SW Cloud Forcing, along GPCI transect. Finally we applied those optimal parameter sets in short simulations and compared the simulation results with default and optimal parameter sets. The results obtained from those analysis are not only useful in identifying the optimal parameter sets for different objectives but also helpful in improving our understanding of the model behavior at the process level.