Simulations of Tropical-Extratropical Cloud Bands over South America using Convective-Permitting Models

Tropical-extratropical cloud bands are typical of the subtropical South American climate, occurring mainly during the rainy season and producing more than 60% of the seasonal precipitation. Thus, their correct representation in climate models is fundamental for the accuracy of simulated subtropical precipitation. Here, we investigate the representation of the tropical-extratropical cloud bands in two convective-permitting simulations. The first, produced by the UK Met Office, is a 10-year simulation with 4.5 km spatial resolution, with two different runs: a control run forced by a high-resolution global climate model (HadGEM3-GC3.1-n512) and a hindcast run forced by a reanalysis product (ERA-Interim) downscaled by an RCM. The second model is a 20-year retrospective simulation produced by the NCAR Water System Program, using a WRF model with 4-km grid spacing. The cloud bands are identified using an objective detection algorithm applied to OLR, as described by Zilli and Hart (2021). All simulations represent the location and seasonal cycle of observed cloud bands well, despite warm biases in the mean OLR. Additionally, all simulations adequately reproduced the ERA5 precipitation during the cloudband events, despite biases in the intensity of the rain. In the next step, we will investigate the causes of these biases. Zilli et al. (in prep.) identified similar biases in the UK Met Office GCMs, which were caused by biases in the upper-level circulation.