Wet troposphere corrections to altimeter measurements calculated from the TOPEX/Poseidon (T/P) Microwave Radiometer (TMR) and the ERS-1 and ERS-2 Microwave Radiometers (EMR I and EMR2) are compared to each other and to European Centre for Medium-Range Weather Forecasts (ECMWF) model data. The most recently published correction algorithm for the EMR1 data [Stum et al., 1998] is applied. The suggested drift correction for TMR data [Keihm et al., 1998, 2000] is also evaluated. The corrected EMR1 data (1991-1996) produce a global (to +/- 66 degrees the T/P latitude range) long-term mean wet troposphere correction 6 and 13 mm lower than TMR and ECMWF, respectively. The EMR2 data (1995-1999) yield a mean wet troposphere correction 2 and 9 mm lower than TMR and ECMWF, respectively. After removing mean differences all three microwave radiometers reproduce similar long-term zonal wet troposphere corrections compared to the ECMWF model (10-14 mm rms) and to each other (5-9 nim rms) with some zonally periodic differences, most < 10 mm. The ECMWF model shows variations compared to the radiometers of over 30 mm before 1995, about 20-30 mm from 1995 to 1997, and up to 20 mm from 1998 to 1999. The intersatellite differences include a latitudinally dependent annual signal, reaching 10 nun in amplitude. Before correcting for the TMR drift there exists a global relative TMR-EMR1 drift of -1.6 +/- 0.4 mm y(-1), from 1992 to 1996. After correcting for the TMR drift the TMR-EMR I trend is reduced to -0.4 +/- 0.2 mm. y(-1), supporting the TMR drift correction. The TMR-EMR2 trend changes sign (direction) after an anomaly that occurred in one of the EMR2 brightness temperatures during June 1996. Before the anomaly, with (without) the TMR correction the relative TMR-EMR2 trend is -3.0 +/- 1.9 (4.0 +/- 2.1) nun y(-1), again supporting the TMR drift correction. After the anomaly, through 1997 the TMR-EMR trend is 3.7 +/- 1.2 mm y(-1), and from 1998 to 1999 it is 0.8 +/- 0.6 mm y(-1).
