A 90--150 days signal is identified in the global precipitable water field generated by the Global Data Assimilation Systems (GDAS) of the Goddard Laboratory for Atmospheres (GLA), National Meteorological Center, and European Centre for Medium-Range Weather Forecasts. The finding of this intraseasonal signal in global precipitable water is significant for two reasons: (1) it suggests that there is 90--150 day intraseasonal variability in the atmospheric branch of the global hydrological cycle and (2) it provides a useful parameter to test the sensitivity of the GDAS-generated hydrological data. This newly identified intraseasonal signal in the global precipitable water was verified with Special Sensor Microwave/Imager precipitable water data over oceans and station-mixing ratio data over the continental United States. Based upon some simple statistical analyses and global and regional composite charts, it was found that the 90--150 day low-frequency oscillations contained in different GDAS data sets are more coherent with each other in regions with good data coverage but are poorly correlated over the data-sparse areas. Furthermore, the GLA GDAS provides the most realistic representation of this intraseasonal global precipitable water signal. ¿ American Geophysical Union 1996