A Look At PW During Summer 2010

The above plot shows Tucson - U of A - GPS versus Raob PW for the 0000 UTC soundings for summer 2010. The GPS sensor is on campus about 2 blocks from the NWS sounding release site, so the data are essentially co-located. Each system has its own quirks, so we should not expect exact agreement. However, the hunidity sensor of the NWS RRS sondes is known to be unreliable. The 0000 UTC data from the RRS system were usually dry wrt the GPS data during summer 2010 - some individual soundings had rather extreme differences of a half to nearly a cm - making sounding diagnostics prone to significant errors. Note also that on a small number of days the RRS data at 0000 UTC were wetter than the GPS data.

The same plot for 1200 UTC is just above. The agreement between the two data sets is better for the morning soundings. Interestingly, the morning RRS data tend to be more moist than the GPS data - so the morning soundings often have a different character than the afternoon soundings. Some days do have lower PW in the morning - usually when there is deep, elevated residual layer aloft.
------------------------------------------------------------------
The bottom line here is that each and every sounding needs to be carefully examined and compared (at sites where this possible) with simultaneous GPS data. My feeling, after watching these data for several years, is that the GPS data are far more accurate than the NWS RRS sonde data.
------------------------------------------------------------------
Many thanks to Seth Gutman of NOAA ESRL in Boulder for helping me obtain these data!