An evaluation of Soil Moisture and Evapotranspiration as derived from CLARIS-LPB regional models

PABLO C. SPENNEMANN; A. CELESTE SAULO; PATRIC SAMUELSSON

Punta del Este

Encuentro; CLARIS LPB M48 Final Meeting; 2012

Surface states evolve slower than atmospheric states, and RCM are expected to represent these states since they include surface models that are coupled to the atmospheric model. A miss-representation of surface states can affect simulated climate and may be responsible of ?climate drifts?. The objectiva of this study is to analyze how well soil moisture, evapotranspiration and precipitation monthly means are represented by CLARIS-LPB models as compared with Global Land Data Assimilation System (GLDAS 2) data set and ERA Interim. Also, an emphasis is set on two particular regions of interest: the continental portion of the South Atlantic Convergence Zone (SACZ) and southeastern South America (SESA). The evapotranspiration, compared with GLDAS-2, the RCA model show larger mean values in the southern portion of LPB and smaller in the northern part. RegCM3 and PROMES show lower values over the same area. Over the Amazon region, CLARIS models tend to overestimate ET. The ERA Interim and CLARIS models tend to overestimate the variability over SACZ and slightly underestimate it over southern LPB. The soil moisture representation is hard to determine, because the different values are related with different total depths. Despite this, the mean fields and standard deviations have similar patterns. Over LPB region, it seems that errors in precipitation are correlated with errors in evapotranspiration, while this is not the case over the Amazon. More analysis is needed to determine the role of soil moisture variability to explain reduced precipitation variability over SESA.