Comparison of Observed and Model-Simulated Atmospheric Moisture Vertical Structure over the Amazon Rainforest

The climate of Brazil’s Amazon rainforest is associated with large amounts of atmospheric moisture. Accurate representation of the atmospheric moisture field is essential for simulating the climate of the Amazon in numerical weather prediction and global climate models. The present analysis evaluates the vertical profiles of specific humidity simulated by models of the Climate Model Intercomparison Project Phase 5 (CMIP5), an ensemble of climate models produced by multiple modeling centers around the world. By comparing simulated vertical moisture structure across the models as well as to observed meteorological profiles from the Amazon, this project seeks to identify and attribute errors or biases in the models’ representations of atmospheric moisture. Using the computer language Python, seasonal-mean vertical profiles of specific humidity from the surface to 100 hPa were computed for each climate model. Preliminary results suggest that the models depart from observations most significantly around 700 hPa as well as near the surface, with the models producing drier conditions than observed. This suggests that the climate models do not sufficiently moisten the Amazon, which is consistent with models producing surface precipitation rates below the observations.