An Assessment of Upper Tropospheric Water Vapor in the MERRA-2 Reanalysis: Comparisons with MLS and In Situ Water Vapor Measurements

Abstract:

The recently released MERRA-2 reanalysis represents a significant evolution of the GEOS-5 atmospheric general circulation model and data assimilation system since the original MERRA project, and it is expected that MERRA-2 will be widely used in climate change studies as has its predecessor. A number of studies have demonstrated critical sensitivities of the climate system to the water vapor content of the upper troposphere and lower stratosphere (UT/LS) and it is therefore important to assess how well the MERRA-2 reanalysis represents the mean structure and variability of water vapor in this part of the atmosphere. Recent comparisons with MLS water vapor indicate that the ECMWF and original MERRA reanalyses overestimate water vapor throughout the global upper troposphere by 50-80%. These overestimates are particularly acute at 147 hPa and 215 hPa and occur in all seasons. In this presentation, we analyze differences between the MLS v.4.2 water vapor data and the new MERRA-2 reanalysis to assess improvements in the treatment of water vapor in the GEOS-5 system since MERRA. We also include in our analysis a comparison of MERRA-2 profiles with water vapor and relative humidity profiles from frostpoint hygrometers at five sites with long-term records and a sixth with an intensive campaign of one month. Three of the long-term sites, Boulder, Colorado, Lindenburg, Germany and Lauder, New Zealand, lie in middle latitudes, and two sites, San José, Costa Rica and Hilo, Hawaii, are in the tropics and subtropics, respectively. The campaign-only database is from the NASA SEAC4RS mission at Ellington Field, Houston, TX in 2013.