Assessing of Two Planetary Boundary Layer Schemes in RegCM4 Model Over the Tropical Region of Brazil

This work presents a contribution to regional climate modeling studies with emphasis on seasonal variability of precipitation in the Amazon Basin (AMZ) and Northeast region of Brazil (NEB). In this sense, the present study aimed to analyze the sensitivity of the simulations carried out for the period 2001 to 2005 (5 years), using two different Planetary Boundary Layer parameterization schemes (Holtslag and UW-PBL) available in the regional climate model version 4 (RegCM4). As initial and large-scale boundary conditions, ERA15 reanalyses were used. Data from the Climate Prediction Center morphing technique (CMORPH) were used for the precipitation assessment. Statistical metrics were used to evaluate the simulations. The dominant feature in RegCM4 simulations is a dry bias with the largest average value (~ 6.0 mm d−1) in part of AMZ Basin, while the lowest mean bias precipitation (~ 1.0 mm d−1), mainly located in NEB region. It was found that both experiments are able to adequately represent the pattern of annual cycle of precipitation when compared to the CMORPH. The UW-PBL scheme experiment had the best performance in the northern sector of AMZ Basin in the austral summer season, with the highest correlation value (~ 0.6). Both experiments showed persistent dry bias in the southern sector of the AMZ Basin during the austral summer season. The Holtslag scheme experiment showed greater ability to reproduce climate variability in the NEB region, especially during austral winter season, with the highest correlation value (~ 0.8). This study can help RegCM users in choosing the suitable configuration of their experiments. In general, the results show that precipitation is better simulated by the Reg_UW-PBL experiment in part of AMZ Basin, while the Reg_Holtslag experiment has better performance in representing the precipitation in the NEB region.