In an attempt to improve the quality of Nimbus-6 soundings, the Man-computer Interactive Data Access System (McIDAS) at the University of Wisconsin is used to manually edit individual (scan spot) High-resolution Infrared Sounder (HIRS) soundings. Unlike the Nimbus-6 Data System Test (DST) soundings that are derived from averages of up to 84 spot radiance measurements and the TIROS-N and NOAA-6 operational soundings that can involve averages of up to 63 spot measurements, the HIRS soundings used in this study were derived from single spot radiances or averages of five single spot soundings. Also, unlike the DST soundings, the HIRS McIDAS retrievals used contemporary surface (instrument shelter) temperatures as a pseudoinfrared window channel to aid in cloud filtering. These McIDAS-generated soundings were used to analyze 1) level temperature fields at 850, 500, and 300 mb; 2) the 850–300 mb thickness field; and 3) the 500 mb height field over eastern Europe for a 20 August 1975 case. The excellent radiosonde network in this area is used for verification purposes.
While the individual spot soundings offer little improvement over the DST data, the averages of five of these soundings provide analyses that are distinctly superior to the operational DST soundings. Although improvement is noted in the root mean square and bias scores, the largest improvements are found in the S1 score, which is a measure of gradient comparison. The radiosonde verification data are partitioned into subsets and in some tests are combined with the HIRS soundings. Other tests involve comparisons in which the radiosonde data consist of only mandatory-level data. The various tests indicate that the edited HIRS soundings averaged in small groups are comparable to mandatory-level radiosonde data in constructing 500 mb height analyses.
1 Present affiliation; National Weather Service, Minneapolis, Minn. 55450.