Several years ago, we performed a preliminary study, which compared passive microwave total precipitable estimates with those obtained from the first NCEP/NCAR reanalysis data set. Our findings then showed large differences in the two fields, particularly in the tropical oceanic regions. In the original study, retrievals from two different SSM/I precipitable water algorithms were compared with the Re-analysis output. Differences between the SSM/I retrievals of precipitable water were small, while the largest differences were between the satellite and modeled fields.

During the past five years, there have been several upgrades to the NCEP/NCAR reanalysis data sets and there now exists several passive microwave precipitable water data sets (e.g., SSM/I, 14-years in length; TRMM/TMI, 4 years in length; NOAA/AMSU, 3-years in length) as well as the GEWEX/GVAP data set (a blended satellite and in-situ product). A unified comparison of all of the data types (satellite and model generated) is warranted.

The primary purpose of this study will be to identify regional and seasonal differences amongst the various time series data sets. These differences will be quantified in terms of zonal and spatial biases, and spatial and temporal correlation fields. Another purpose is to document the differences noted in the interannual anomaly fields of TPW derived from all the sources, particularly, the anomalies attributed to the warm and cold SST phases in the tropical Pacific. A variety of analysis techniques will be utilized.