Mean water-vapor-weighted temperature of the atmosphere, Tm, is a key parameter in the retrieval of atmospheric precipitable water (PW) from ground-based GPS measurements of zenith path delay (ZPD). The accuracy of GPS-derived PW is proportional to the accuracy of Tm. An uncertainty of 5 K in Tm corresponds to 1.7-2.0% uncertainty in PW. To derive a global PW dataset from GPE-measured ZPD, it is important to evaluate various estimates of Tm. Two common methods to estimate global Tm are 1) using empirical linear relationship of Tm to surface temperature, Ts, and 2) using numerical weather prediction model output and reanalysis products. We will use the ERA-40, NCEP/NCAR reanalysis, global surface observations of Ts, and global radiosonde data to evaluate different Tm estimates. Tm from radiosonde data will be considered as truth. Comparisons of Tm from different datasets will determine if reanalysis data are good enough for Tm estimates, or local specific Tm-Ts relationship should be used. Preliminary results suggest that ERA-40 seems to overestimate Tm, while Bevis92’ Tm-Ts regression equation tends to underestimate Tm in the Tropics.