The GEWEX Water Vapor Assessment: Results from Intercomparison, Trend, and Homogeneity Analysis of Total Column Water Vapor

Abstract

The Global Energy and Water Cycle Exchanges project (GEWEX) water vapor assessment’s (G-VAP) main objective is to analyze and explain strengths and weaknesses of satellite-based data records of water vapor through intercomparisons and comparisons with ground-based data. G-VAP results from the intercomparison of six total column water vapor (TCWV) data records are presented. Prior to the intercomparison, the data records were regridded to a common regular grid of 2° × 2° longitude–latitude. All data records cover a common period from 1988 to 2008. The intercomparison is complemented by an analysis of trend estimates, which was applied as a tool to identify issues in the data records. It was observed that the trends over global ice-free oceans are generally different among the different data records. Most of these differences are statistically significant. Distinct spatial features are evident in maps of differences in trend estimates, which largely coincide with maxima in standard deviations from the ensemble mean. The penalized maximal F test has been applied to global ice-free ocean and selected land regional anomaly time series, revealing differences in trends to be largely caused by breakpoints in the different data records. The time, magnitude, and number of breakpoints typically differ from region to region and between data records. These breakpoints often coincide with changes in observing systems used for the different data records. The TCWV data records have also been compared with data from a radiosonde archive. For example, at Lindenberg, Germany, and at Yichang, China, such breakpoints are not observed, providing further evidence for the regional imprint of changes in the observing system.

The GEWEX Water Vapor Assessment archive of water vapour products from satellite observations and reanalyses

“…The intention of this effort is to showcase differences among the data records and not to characterise uncertainties. Intercomparison results for long-term data records and for profile data records are given in Schröder et al (2016) and Schröder et al (2017a,b).…”

“…This approach was also applied to AMSR-E JAXA. Then, all data records are 25 arithmetically averaged onto the common grid by considering all valid observations within a grid cell (see also Schröder et al, 2016). Note that the regridding procedure impacts the data record's internal variance and characteristics of extremes.…”

Atmospheric Moisture Content over Europe and the Northern Atlantic

“…This then leads to a reduction in the magnitude of errors for humidity-sensitive variables [42]. Many comparative studies have confirmed the usefulness of reanalysis data in the examination of water vapor fields [29,43,44].…”

“…Climatologic reanalysis appears to solve this problem to some extent. Despite the fact that the models used in reanalysis studies are characterized by poor parametrization of the hydrologic cycle, although errors are reduced in the course of the development of each reanalysis, the usefulness of reanalysis data (especially ERA-Interim) has been confirmed by numerous comparative studies) [29,44,59,60] featuring analyses with less temporal resolution (averaged data) and covering larger geographic areas (e.g., studies on climate change. In addition, gridded data now yield spatial information with a resolution high enough to make spatial pattern analysis possible.…”

The ESA GOME-Evolution “Climate” water vapor product: a homogenized time series of H<sub>2</sub>O columns from GOME, SCIAMACHY, and GOME-2

“…However, the spatial coverage of ARSA and GNSS stations is rather poor in some regions and the majority of the open oceans are not covered at all. As the stability can be a strong function of region (Schröder et al, 2016), future efforts are needed to assess stability globally.…”