ATMOSPHERIC MOISTURE MEASUREMENTS - A MICROWAVE RADIOMETER - RADIOSONDE COMPARISON

The NASA/Goddard Space Flight Center Crustal Dynamics Project microwav e Water Vapor Radiometer (WVR-J03) is used to measure the thermal emis sion of the sky at three frequencies (20.7, 22.2, and 31.4 GHz). Measu rements were taken during the Atmospheric Moisture Intercomparison Stu dy (ATMIS) held at Wallops Island, VA during April 1989. These measure ments were compared with brightness temperatures inferred from measure ments from VAISALA radiosonde packages launched every 3 hours during t he experiment period. An error analysis for the radiosonde-inferred br ightness temperatures was performed assuming reasonable random uncerta inties for the pressure, temperature, and humidity measurements and pr opagating these uncertainties through the analysis algorithm. For the assumed uncertainties (sigma(P) = sigma(T) = 0.84 K, and sigma(RH) = 5 % RH) the dominant contribution to the total uncertainty comes from th e temperature measurement (66% - 88%) whereas the relative humidity me asurement contributes only 2% - 8%, except in the vicinity of the wate r vapor line, where the contribution is 10% - 20%. Sky brightness temp erature random errors range from 0.03 K to 0.6 K, with systematic erro rs between 0.4 K and 1.8 K. Two different water vapor emission models were used for the derivation of the brightness temperatures from the r adiosonde measurements. The Liebe model and VAISALA radiosonde data gi ve better agreement with the WVR for the 20.7 and 22.2 GHz frequencies (mean differences [WVR - VAI] of -0.32 +/- 0.56 K and -0.22 +/- 0.77 K, respectively) than does the Waters model and VAISALA radiosonde dat a. Agreement is best at 31.4 GHz using the Waters model (mean differen ce [WVR - VAI] -0.28 +/- 0.5 K), although in all cases the differences are less than the estimated uncertainties on the radiosonde inferred brightness temperatures. Differences between the two models increase a s the moisture content increases and vary as a function of frequency.