The Modern Era Retrospective-analysis for Research and Applications (MERRA) provides global reanalysis data at ½ degree spatial resolution and 1 hourly frequency. Reanalyses data have provided lateral boundary conditions for regional climate modeling studies, as they assimilate the best observations. However, the background forecast model also provides a significant source of bias and uncertainty for regional climates. In this study, we investigate the regional climate variability of MERRA for the Mississippi River Basin and United States, and its realism of representing major extreme events including links to ENSO. We find that MERRA precipitation is more strongly correlated to ENSO34 than gauge observed precipitation. The seasonal variations of MERRA precipitation anomalies are reproduced well, however, the peak values of the precipitation and notable anomalies are generally underestimated. Examples for the 1993 Mississippi flooding and 2011 Ohio River flooding will be presented. Temperature analysis of the maximum, minimum and diurnal temperature range will also be given.