Climate downscaling over southern South America for present-day climate (1970-1989) using the MM5 model. Mean, interannual variability and internal variability

Abstract

This work focuses on evaluating the ability of the MM5 regional model to represent the basic features of present climate over South America. The spatial distribution of seasonal means and the interannual variability, as well as annual cycles for precipitation and near-surface temperature have been evaluated. The internal variability has also been investigated. The analysis has two objectives: one of them is to quantify the dynamic downscaling ability to represent the current climate and the other is to identify critical aspects of the regional climate model in South America in order to interpret the reliability of future projections for the end of the twenty-first century in the A2 scenario of the IPCC Special Report on Emissions Scenarios. In general, the MM5 model is able to reproduce adequately the main general features, seasonal cycle and year-to-year variability of near surface variables over South America. The spatial distribution of temperature is well represented, but some systematic errors were identified, such as an overestimation in central and northern Argentina and an underestimation in the mountainous regions throughout the year. The general structure of precipitation is also well captured by the regional model, although it overestimates the precipitation in the Andean region (specifically in central and southern Chile) in all seasons and underestimates the rainfall over tropical latitudes. The annual cycle of precipitation is adequately represented in the subregions analyzed, but its representation is better over La Plata basin (LPB), Cuyo (CU) and southeastern Pampas (SEP). The annual cycle of mean temperature is well represented, too. The model systematically overestimates the interannual variability of temperature and underestimates the interannual variability of precipitation. From the analyses of interannual and internal variability, as well as the biases, it can be concluded that regardless the season, the simulated precipitation is reliable at subtropical latitudes, Uruguay, southern Brazil and east-central of Argentina, but is less reliable over areas of complex topography. For temperature, the regional model is reliable over subtropical latitudes, Uruguay and the south of Brazil only during winter, but it is less reliable or it is even in the limit of reliability over central and southern Chile all along the year. Therefore, it is concluded that the MM5 model is a useful tool for the generation of regional climate change scenarios and for the evaluation of regional climate change scenarios over southern South America.