On the Discrepancy Between Calculated and Observed Nimbus II 6.7-μm Water Vapor Radiation

Abstract

An attempt is made to resolve the difference in Nimbus II satellite-observed and calculated atmospheric radiances in the 6.7-?m water vapor band. Regression equations were calculated to relate the Nimbus II MRIR (medium resolution infrared radiometer) water vapor channel radiance measurements to radiances calculated assuming different values of the effective water vapor absorption coefficient. A value of log10L*=2.4, where L* is the effective water vapor channel absorption coefficient, produced the maximum correlation between computed and observed values. The regression equation for this water vapor absorption co-efficient may be used to recalibrate the Nimbus II 6.7-?m radiance observations. The radiance upwelling from the atmosphere in the 6.4?6.9 ?m spectral region (the 6.7- ?m channel of Nimbus II) arises mainly from the 200?600 mb atmospheric layer. However, assuming that the water vapor profile can be represented by a power function, the entire water vapor distribution can be estimated from these radiance observations. Results from 250 cases showed that a power law exponent of 3.9 yielded the best correspondence between radiance-calculated and radiosonde-observed water vapor profiles. RMS mixing ratio errors varied from 0.03 gm kg?1 at 300 mb to 1.3 gm kg?1 at 850 mb, yielding relative errors of 10?20%. The mixing ratio discrepancies in the sensed layer (200?600 mb) are close to the errors of radiosonde observations. The rms difference between the radiance-calculated and radiosonde-observed total precipitable water was 0.6 gm cm?2. The relatively large total precipitable water discrepancy is caused by the fact that the 6.7-?m radiance observations are relatively insensitive to the lower atmosphere where the majority of water vapor exists and where the variation of water vapor is largest.