A Method for Estimation of Atmospheric Water Vapor Profiles by Microwave Radiometry

P. W. Rosenkranz Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge 02139

Search for other papers by P. W. Rosenkranz in
Current site
Google Scholar
PubMed
Close
,
M. J. Komichak Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge 02139

Search for other papers by M. J. Komichak in
Current site
Google Scholar
PubMed
Close
, and
D. H. Staelin Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge 02139

Search for other papers by D. H. Staelin in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Simultaneous measurements of microwave emission from the earth, in the oxygen band near 60 GHz and the water vapor line near 183 GHz, permit inference of atmospheric temperature as a function of two variables: pressure and water vapor burden (integrated water vapor content above any level). Combination of these two profiles yields a profile of water vapor burden versus pressure. Occasional singularities in these retrievals can readily be identified and excluded. Numerical simulations have been made, using temperature and water vapor profiles from subtropical and midlatitude radiosonde stations, and assuming surface reflectivities typical of either land or ocean. Over a land surface, the residual rms errors in estimated water vapor burden profile between 300 and 1000 mb, range from 23 to 43% of the a priori standard deviation of water vapor burden for the corresponding climate. The relative humidity profile was also estimated with rms errors ranging from 4 to 17% of saturation. Over a seawater surface, using three additional channels at 18.5, 22.2 and 31.7 GHz, the corresponding results are 3–46% of a priori standard deviation for water vapor burden and 4–15% rms error for relative humidity.

Abstract

Simultaneous measurements of microwave emission from the earth, in the oxygen band near 60 GHz and the water vapor line near 183 GHz, permit inference of atmospheric temperature as a function of two variables: pressure and water vapor burden (integrated water vapor content above any level). Combination of these two profiles yields a profile of water vapor burden versus pressure. Occasional singularities in these retrievals can readily be identified and excluded. Numerical simulations have been made, using temperature and water vapor profiles from subtropical and midlatitude radiosonde stations, and assuming surface reflectivities typical of either land or ocean. Over a land surface, the residual rms errors in estimated water vapor burden profile between 300 and 1000 mb, range from 23 to 43% of the a priori standard deviation of water vapor burden for the corresponding climate. The relative humidity profile was also estimated with rms errors ranging from 4 to 17% of saturation. Over a seawater surface, using three additional channels at 18.5, 22.2 and 31.7 GHz, the corresponding results are 3–46% of a priori standard deviation for water vapor burden and 4–15% rms error for relative humidity.

Save