A climatological comparison of column-integrated water vapor for the third-generation reanalysis datasets

The atmospheric reanalysis datasets have been widely used to understand the variability of atmospheric water vapor on various temporal and spatial scales for climate change research. The difference among a variety of reanalysis datasets, however, causes the uncertainty of corresponding results. In this study, the climatology of atmospheric column-integrated water vapor for the period from 2000 to 2012 was compared among three latest third-generation atmospheric reanalyses including European Centre for Medium-range Weather Forecasts Interim Re-Analysis(ERA-Interim), Modern-Era Retrospective Analysis for Research and Applications(MERRA), and Climate Forecast System Reanalysis(CFSR), while possible explanation on the difference between them was given. The results show that there are significant differences among three datasets in the multi-year global distribution, variation of interannual cycle, long-term trend and so on, though high similarity for principal mode describing the variability of water vapor. Over oceans, the characteristics of long-term CWV variability are similar, whereas the main discrepancy among three datasets is located around the equatorial regions of the Intertropical Convergence Zone, the South Pacific Convergence Zone and warm cloud area, which is related with the difference between reanalysis models for the scheme of convective parameterization, the treatment of warm clouds, and the assimilation of satellite-based observations. Moreover, these CWV products are fairly consistent with observations(satellite-based retrievals) for oceans. On the other hand, there are systematic underestimations about 2.5 kg/m2 over lands for all three CWV datasets, compared with radiosonde observations. The difference between models to solve land-atmosphere interaction in complex environment, as well as the paucity in radiosonde observations, leads to significant water vapor gaps in the Amazon Basin of South America, central parts of Africa and some mountainous regions. These results would help better understand the climatology difference among various reanalysis datasets better, and more properly choose water vapor datasets for different research requirements.

The atmospheric reanalysis datasets have been widely used to understand the variability of atmospheric water va- por on various temporal and spatial scales for climate change research. The difference among a variety of reanalysis datasets, however, causes the uncertainty of corresponding results. In this study, the climatology of atmospheric column-integrated wa- ter vapor for the period from 2000 to 2012 was compared among three latest third-generation atmospheric reanalyses including European Centre for Medium-range Weather Forecasts Interim Re-Analysis （ERA-Interim）, Modem-Era Retrospective Analy- sis for Research and Applications （MERRA）, and Climate Forecast System Reanalysis （CFSR）, while possible explanation on the difference between them was given. The results show that there are significant differences among three datasets in the mul- ti-year global distribution, variation of interannual cycle, long-term trend and so on, though high similarity for principal mode describing the variability of water vapor. Over oceans, the characteristics of long-term CWV variability are similar, whereas the main discrepancy among three datasets is located around the equatorial regions of the Intertropical Convergence Zone, the South Pacific Convergence Zone and warm cloud area, which is related with the difference between reanalysis models for the scheme of convective parameterization, the treatment of warm clouds, and the assimilation of satellite-based observations. Moreover, these CWV products are fairly consistent with observations （satellite-based retrievals） for oceans. On the other hand there are systematic underestimations about 2.5 kg/m2 over lands for all three CWV datasets, compared with radiosonde ob- servations. The difference between models to solve land-atmosphere interaction in complex environment, as well as the pauci- ty in radiosonde observations, leads to significant water vapor gaps in the Amazon Basin of South America, central parts of Africa and some mountainous regions. These results would help better understand the climatology difference among various reanalysis datasets better, and more properly choose water vapor datasets for different research requirements.