TWP-ICE global model intercomparison: resolution impact and diurnal cycle

Description

Global models, including global forecast and climate models, from seven modeling centers were used to compare tropical cloud and precipitation simulations during the Tropical Warm Pool-International Cloud Experiment (TWP-ICE) period. As part of the ARM GCSS TWP-ICE intercomparison, this intercomparison aims to assess current global model performance of tropical cloud and convection and the resolution impact. Models were initialized from ERA-40 each day with day 2 forecasts compared with available observations. All models capture the transition of three distinct regimes (wet, break, and dry monsoon period) and their associated cloud, radiation, and precipitation variation during the TWP-ICE. However, large discrepancies are found in cloud, radiation, and precipitation among models. Stratiform precipitation fraction is sensitive to cumulus parameterizations used and generally increases with resolution. But cumulus parameterization impact on model cloud fraction and microphysics (LWC and IWC) is not straightforward. Model cloud fraction and hydrometeor fields differ significantly (up to one order of magnitude) among models and are greatly modulated by cloud schemes used. Though most models are able to produce the realistic Q1 and Q2 profiles during different regimes, their components, such as convective and stratiform heating, differ significantly, even in sign, among models. This reflects the model's ability to achieve realistic temperature profile via compensation among various parameterized physical processes. Most models are unable to capture the cloud and precipitation diurnal cycle triggered by localized forcing during the dry period, though resolution indicates slight improvements in this regard. The study illustrates the complexity involved in tropical convection and suggests the importance of using ARM observations to better constrain model physical parameterizations and their interactions.