The wet tropospheric range correction: Product intercomparisons and the simulated effect for tropical Pacific altimeter retrievals

Abstract

The similarities and differences between four water vapor products currently available to correct for the wet tropospheric effect in altimeter data are quantified. Two of the water vapor data sets are derived from satellite measurements (Nimbus 7 scanning multichannel microwave radiometer (SMMR) and Tiros‐N operational vertical sounder (TOVS)), and two are from operational analyses (European Center for Medium Range Weather Forecasts (ECMWF) and Fleet Numerical Oceanography Center (FNOC)). Subsequently, these analyses serve as the basis from which to assess the simulated effect of these products and their uncertainties on our ability to resolve significant sea level signals in the tropical Pacific Ocean via satellite altimetry. The time frame 1979–1983 was selected for study because it was the only period in which four water vapor data sets overlapped. This period also provided an interesting contrast between a relatively regular seasonal cycle for 1979–1981 and the strong El Niño‐Southern Oscillation (ENSO) event in 1982–1983. Water vapor product intercomparisons showed the greatest consistency between the two satellite data sets, SMMR and TOVS. The standard deviation of the mean seasonal cycle for these two range corrections was 2 to 4 cm compared against an rms difference of 1 to 2 cm between the seasonal cycle of the two data sets. An additional 1‐ to 2‐cm difference was also present between the two mean fields of the data. The largest rms differences, 4 to 8 cm, were found to occur between either the satellite and operational data or between the two operational analyses. Detailed spatial structure apparent in both the SMMR and TOVS data sets corresponded well with underlying sea surface temperature fields and features such as the Intertropical Convergence Zone (ITCZ) and the South Pacific Convergence Zone (SPCZ). Although the spatial resolution of the ECMWF data was coarser than either satellite product, many of these same features were also present. The FNOC water vapor correction was the clear outlier of the four products. Tropical Pacific model sea level solutions for 1979–1983 were degraded with the SMMR and ECMWF water vapor products to simulate altimeter measurements unadjusted for a water vapor correction. The low‐frequency, large‐scale influence of a wet troposphere was analyzed on seasonal and interannual time scales. Seasonally, the SMMR and ECMWF wet tropospheric effects induced a slight increase in the amplitude of the simulated sea level retrievals without seriously distorting the spatial structure. Interannual anomalies of the water vapor correction during the height of the 1982–1983 El Niño reached 12 cm in the eastern tropical Pacific, superimposed on simulated sea level anomalies of 20–30 cm. As the equatorial atmospheric convection shifted from the western to the eastern Pacific during this ENSO episode, the associated increase of columnar water vapor had the effect of significantly decreasing the simulated altimeter measurement of the elevated sea level characteristic of El Niño in the eastern equatorial Pacific.