Introduction—Analysis disagrees with radiosonde data

The operational 22-km Eta model analysis differed considerably from radiosonde 700 hPa temperature reports over a broad swath of Oklahoma and northeast Texas on 13 August 2001. The radiosonde temperatures are almost the same across this entire area. Although it turns out that the first guess (not shown) agreed well with the 700 hPa temperature from FWD (Fort Worth, TX), the analysis has the largest disagreement in this area. This caselet includes a series of questions to help a forecaster probe what happened and the implications for detecting and analyzing mesoscale atmospheric features.

Learning objectives

In this case study, the student will gain concrete awareness of:

1) Limitations caused by inadequate observation sampling
2) Limitations on the representativeness of observations
3) Effects of horizontal spreading of information in the analysis ("correlation length")
4) The role of variable linkages in an analysis
5) A key motivation and goal of analysis changes currently under development

Credits

By Dr. Stephen Jascourt, UCAR/COMET at NCEP

Thanks to:

• Bill Bua, UCAR/COMET at NCEP for assistance with aircraft plots and gathering data
• Pete Manousos, HPC SOO for reviewing this presentation
• Dave Parrish, NCEP/EMC for providing aircraft data and helpful discussion
• Eric Rogers and Geoff Manikin, NCEP/EMC for Eta guess and analysis soundings
• Steve Silberberg at AWC for calling attention to this case