Model-simulated humidity bias in the upper troposphere and its relation to the large-scale circulation

Abstract

[1] The depiction of water vapor in the upper troposphere in Geophysical Fluid Dynamics Laboratory (GFDL) climate model and ERA-40 reanalysis is evaluated through a model-to-radiance approach. Brightness temperatures of High-Resolution Infrared Radiation Sounder (HIRS) 6.7 μm channel, Special Sensor for Microwave Water Vapor Profiler (SSM/T-2) 183.31 ± 1 GHz channel, and Microwave Sounding Unit (MSU) 60 GHz channel simulated with data from the GFDL climate model and ERA-40 reanalysis show a distinct cold and moist bias in the upper troposphere compared to satellite observations, particularly over the subtropics. Temperature biases are a common feature in many climate models and complicate the interpretation of radiance-based comparisons with satellite data. We introduce a new method for evaluating the water vapor distribution which combines both HIRS 6.7 μm and SSM/T-2 183.31 ± 1 GHz channels and is much less sensitive to tropospheric temperature biases. Using this method, we show that GFDL climate model has a more humid upper troposphere over dry subtropical area than ERA-40 reanalysis. The geographical distribution of the humidity bias is found to exhibit a close association with differences in the 500 hPa vertical pressure velocity, suggesting that much of the bias in tropical upper tropospheric relative humidity can be attributed to errors in simulating the intensity of large-scale tropical circulation. Given the strong dependence of upper tropospheric water vapor on the large-scale circulation, these results suggest that long-term monitoring of upper tropospheric water vapor from satellites may also offer insight into variations in the large-scale atmospheric circulation.