Convection-Permitting Regional Climate Models (CPRCMs)

A large source of error in climate modeling comes from the parameterization of inherently chaotic systems such as air flow and precipitation. The most clear example of this is cloud formation. Current state of the art models run at resolutions too high to simulate a cloud in detail. The issue is not that scientists don’t understand how or why a cloud forms, but that the scale on which the cloud formation begins is far too small to be well simulated in grids that are 100km by 100km (the low end for most GCMs). Thusly, convection permitting regional climate models (CPRCMs) were created in order to permit scientists to simulate convective processes, such as cloud formation. In order for a model to drop all of its deep convection parameterizations, its spatial resolution must drop below 4km x 4km. Current development of CPRCMs mimics, in some ways, the original development of RCMs in the late 1980s. Dickinson et al., 1989 published their original regional climate model with the goal of resolving mountain ranges in the US that had been obscured by the 500km x 500 km resolution that was standard for state-of-the-art global circulation models at the time. Similarly, RCMs (namely, CPRCMs) are now being developed to resolve physical processes that significantly impact model performance but can’t be done currently because of computational limitations that restrict GCMs to high spatial resolutions.

Meta-analysis of CPRCMs

Because of their low spatial resolution CPRCMs share much in common with weather models. In some ways, weather models are special types of RCMs that are optimized to precisely simulate the weather in a tiny area Although, this classification would be technically incorrect since the models do not simulate climate. Regardless, the two share significant similarities since many of the mechanisms required to simulate weather are required to precisely simulate climate on a small scale. Most CPRCMs are branches of weather models that have had their structure changed to accommodate the goal of modeling climate rather than weather. Some CPRCMs are developed as a fork of an RCM, but it is notable that the vast majority aren’t since it helps illustrate the significant step forward that convection permitting models are. While most of the “innovation” in convection permitting models comes as a side effect of the endless goal of further shrinking the resolution of humanity’s best climate models, creating climate models that are able to simulate, rather than approximate, key convective processes is a significant step forward. As such, convection permitting models are a different breed of models entirely rather than a form of improved RCMs.