Regional Modeling Frameworks
Predicting the regional hydrologic cycle at time scales from seasons to centuries is one of the most challenging goals of climate modeling. Because hydrologic cycle processes are inherently multi-scale, increasing model resolution to more explicitly represent finer scale processes may be a key to improving simulations of the hydrologic cycle. The overall objective of the proposed research is to develop frameworks for robust modeling of regional climate and hydrologic cycle, and to improve understanding of factors contributing to uncertainties in simulating future changes in the regional hydrologic cycle. We propose a hierarchical approach to test the veracity of global high resolution, global variable resolution, and nested regional climate model for regional climate modeling. We hypothesize that hierarchical evaluation of different modeling approaches will lead to better understanding of their relative merits and improve the frameworks for robust regional climate simulation. Our evaluation hierarchy has four stages: 1) Idealized, no physics test cases, 2) Idealized, full physics test cases, 3) Real world, atmosphere-only and ocean-only simulations, and 4) Real world, coupled atmosphere-ocean simulations for both current and future climate.
At each stage, four types of experiments will be performed:
- Global Simulations at High Resolution (GS-HR)
- Global Simulations at Low Resolution (GS-LR)
- Global Simulations using Variable Resolution (GS-VR) with high resolution in the area of interest and low resolution elsewhere
- Regional Simulations using High Resolution (RS-HR) within a limited-area domain with the lateral boundary forcing provided by GS-LR.
The GS-HR and GS-LR simulations will be performed using CCSM with three different dynamical coresâ€”Spectral, HOMME, and Model for Prediction Across Scales (MPAS). The latter can be configured for GS-HR, GS-VR, and RS-HR so all three modeling approaches can be compared within a single framework. The RS-HR simulations will be performed using MPAS, WRF, and RegCM. For GS-VR and RS-HR, the high-resolution regions will be located in North and South America where the regional hydrologic cycle exemplifies scale interactions and atmosphere-land-ocean feedbacks that challenge regional climate modeling.
The global and regional simulations will be compared and evaluated to assess (a) the impacts of different dynamical cores for global high-resolution simulations, (b) multiple techniques for mesh refinement, (c) the upscaled impacts of the high resolution region, and (d) the overall value of regional climate simulation. We will also apply regional and global diagnostics, evaluation metrics, and process-based analysis to the simulations to determine (1) whether modeling frameworks that allow scale interactions through global high resolution or variable resolution may be more skillful in simulating the regional hydrologic cycle in climate regimes dominated by convection; (2) whether models that couple atmosphere and ocean at the regional scale are more skillful in simulating regional climate variability in the west coast of North and South America; and (3) whether differences in simulating feedbacks by different modeling approaches may be modulated by surface heterogeneities to amplify differences in simulating regional hydrologic cycle changes in the future climate.