TOVS/HIRS Water Vapor Retrieval

Current Research
UTH retrievals

Comparison between TOVS/HIRS and SSM/T-2 derived Upper-Tropospheric Humidity

Richard J. Engelen and Graeme L. Stephens
Department of Atmospheric Science, Colorado State University

Bulletin of the American Meteorological Society, 79, 2748-2751.

Summary

In this paper we used the simple relationship log(p₀ · UTH) = a + bT_b to retrieve upper tropospheric humidity from both TOVS infrared brightness temperatures and SSM/T-2 microwave brightness temperatures. This simple relationship between satellite observed brightness temperatures in the water vapor absorption bands and upper tropospheric humidity have real utility for interpreting existing long-term data sets from TOVS, SSM/T-2, GOES, and Meteosat. However, when interpreting results from a simple retrieval method, the following issues have to be carefully addressed:

Although the method is not heavily dependent on the a priori data, the a priori data has to be representative of the whole geographical area where the retrieval is used. A too limited database overly constrains the retrieval and leads to biases in retrieved UTH.
Clouds have an important effect on the retrieval. In general, TOVS retrievals will be too dry in humid regions since the sample is biased to cloud-free regions or the radiances are cloud-cleared. On the other hand, without cloud screening or proper cloud correction SSM/T-2 retrievals will be too moist in humid regions due to the effect of ice clouds on the brightness temperatures. The launch of the AMSU instruments with both infrared and microwave channels on the same satellite will be a significant step forward in understanding effects of clouds more clearly.
The radiative transfer model implied in the retrieval is also a source of error. Although knowledge on line absorption and continuum absorption is improving, uncertainties in continuum absorption, in particular, are problematic, especially for dry atmospheric conditions. For the microwave calculations the contribution of the far-wings of the absorption lines could also be an source of error, which is underestimated.
A final source of uncertainty in the absolute calibration of the UTH retrievals is the uncertainty in the absolute calibration of the radiance measurements. This error source is hard to address and is still a topic of discussion.

In conclusion, the simple UTH retrieval usefully presents geographical distributions and temporal variations in upper tropospheric humidity, but we have to be careful in interpreting the absolute values. More research is required to assess the effect of clouds on the retrieval and it is important to continue to compare the various radiative transfer models used in developing retrieval schemes.

Main figures from the paper

Figure 1. Upper tropospheric humidity for July 1993 obtained from the TOVS 6.7 micrometer channel brightness temperatures.

Figure 2. Upper tropospheric humidity for July 1993 obtained from SSM/T-2 183.3 ±1 GHz brightness temperatures.

Richard J. Engelen
Richard.Engelen@atmos.colostate.edu