Jj. Simpson et al., The NVAP global water vapor data set: independent cross-comparison and multiyear variability, REMOT SEN E, 76(1), 2001, pp. 112-129
Space-time variability in the global distribution of atmospheric total colu
mn water vapor (tcwp) greatly impacts the hydrologic cycle. NASA's Water Va
por Project (NVAP) produced a global 1 degrees x1 degrees tcwp data set for
use as a tool to investigate, among other things, atmospheric variability.
An independent cross-comparison of the NVAP tcwp product was performed usi
ng the TOPEX/POSEIDON (TIP) TOPEX microwave radiometer (TMR) data and the E
uropean Centre for Medium-Range Weather Forecasts (ECMWF)-based range delay
data set produced by Meteo-France (MF) and distributed with T/P data. When
these T/P range delay data are converted to tcwp, they show that NVAP is b
iased dry and ECMWF/MF is biased wet relative to the independent TMR measur
ement. Although the absolute accuracy of the NVAP tcwp product is uncertain
, results indicate its relative accuracy is sufficient for variability stud
ies. Empirical orthogonal function (EOF) analysis and spectral analysis app
lied to this data set show that seasonal variability over the annual cycle
accounts for about 20% of the variance (EOF1). An El Nino-southern oscillat
ion (ENSO) signal is found in the annually demeaned data; the magnitude of
the crosscorrelation between the temporal amplitude (TA) of EOF1 and the Ni
no 3.3 (SST) time series is .9. Comparisons also were made between the NVAP
patterns of variability in tcwp and independent reanalysis and interpretat
ion of numerical model generated atmospheric fields. In general, there is g
ood agreement between the NVAP data and the reanalysis fields. Finally, spe
cific recommendations are made for: (1) improvement of the NVAP data set up
on reanalysis and (2) use of the NVAP data, in place of ECMWF/MF-based rang
e delay data, for T/P retrievals when TMR data are not available if and whe
n T/P data are reanalyzed. This latter recommendation is especially importa
nt for regions of the tropical Indo-Pacific (e.g., Indonesia) where islands
can interfere with valid TMR retrievals. (C) 2001 Elsevier Science Inc. Al
l rights reserved.