There are substantial variability and low frequency trends in precipitable water (column water vapor amounts) and also model-generated precipitation from the ERA-40 reanalyses. Much of the low frequency variability and trends are spurious, and in turn impact the precipitation P minus Evaporation E, which is derived from the vertically-integrated convergence of moisture transports and the atmospheric moisture budget. Several sources exist for precipitable water: radiosondes over land and SSM/I over the ocean, as well as analyses by NVAP into gridded fields. All have large and very different trends post-1988. The problem of knowing the truth is compounded by difficulties in merging multiple SSM/I satellites and changing radiosonde types. Huge spurious trends in P from the model analyses mean that P-E from the moisture budget is much more stable and useful for the surface freshwater flux, but even there, spurious variability appears to be present. We will report on progress in sorting out the real variability and trends.