Sensitivity of seasonal climate and diurnal precipitation over Central America to land and sea surface schemes in RegCM4

Climate Research

A. M. Tompkins¹  G. T. Diro¹  F. Giorgi¹  S. A. Rauscher¹ 

关键词: RegCM4;    CORDEX;    Central America;    Diurnal precipitation;    Climate;    Diurnal SST scheme;    CLM;   

DOI  :  10.3354/cr01049

来源: Inter-Research Science Publishing

【 摘 要 】

ABSTRACT: Multi-annual simulations over the Central America CORDEX domain are conducted with the latest version of regional climate model RegCM4 driven by ERA-Interim reanalysis fields. The RegCM4 system can reproduce both the annual cycle and the spatial patterns of mean summer precipitation over Central America and Mexico. Regional circulation features are also reproduced, although the intensity of the Caribbean Low-Level Jet is underestimated and it is located too far south. Over most land areas, RegCM4 surface air temperatures are lower than observations by 1 to 3°C, which however may also be related to biases in the reanalysis forcing data. The model can realistically simulate the amplitude of the convective diurnal cycle in areas where the convective triggering is dominated by non-local gravity wave effects. However, the simulation of the phase of the diurnal cycle of convection is less satisfactory, with the peak precipitation occurring earlier than observed, a common fault in atmospheric models. Sensitivity experiments are carried out to investigate the model sensitivity to land surface and a prognostic diurnal sea surface temperature scheme. Use of the Community Land Model (CLM) instead of the Biosphere-Atmosphere Transfer Scheme (BATS) results in a warmer and drier land surface and a better simulation of the seasonal average spatial pattern of precipitation. However, with BATS, RegCM4 has a more realistic simulation of the mid-summer drought over the region. The impact of the prognostic sea surface temperature (SST) scheme is generally small. In general, neither of these surface physics upgrades results in a clearly superior model performance.

【 授权许可】

Unknown