Some Discussion Re RRS Upper-Air Data

The serious problems with data from the new NWS RRS sondes have not gone away - I've just been blogging about other things. I thought I'd return to the RRS subject, since the NWS program is nearing completion over the CONUS. I hope to finish a formal paper on the various problems I've noted with these data during the next couple of months. For now it suffices to say that upper-air data with various degrees of problems are going into the upper-air archives every day, and I repeat my warning to research users that extreme caution is required if you use these data.

------------------------------------------------------------------

Yesterday I pulled some plots from the GPS upper-air site ( NOAA site formerly known as FSL). The top data plot compares precipitable water (PW) from GPS data to PW from the RRS soundings at TWC (Tucson). The dry bias continues in the RRS data at TWC, with six soundings having PW more than a standard deviation below the GPS running mean value. Interestingly, the plots from Albuquerque (ABQ - bottom two graphs) indicate that the RRS sounding data have been even drier, relative to the GPS running mean, than have the soundings at Tucson. Note that Albuquerque RRS soundings had PW more than a standard deviation lower than the GPS values on 15 different flights.
------------------------------------------------------------------

The reasons why two close-by upper-air sites should display such differences wrt the GPS observations remains a quirky, technical mystery. One that is probably related somewhat to the variance of individual moisture sensors that occurs somehow in the manufacturing and calibration processes at Sippican. Regardless, one can easily imagine how difficult it would be to try to "correct" the sounding data from TWC and ABQ for the past month.