Compared the NUCAPS sounding today (4/25) over WFO TAE. See the point evaluated below (labeled point A below):
This is the “unmodified sounding.” Overall the thermodynamic profile looked realistic. Looking in the PBL – noticed it was about 2C too warm and about 2C to dry.
This is the “modified sounding.” As expected, no change to the mid level thermo profile. The PBL did have the “correction” applied – not the “cliff” in the temp profile. Overall, I was expecting a more sophisticated PBL nudging scheme. In reality – it just appeared to force the T/TD from the nearest METAR, with maybe some minor smoothing. This is something a forecaster could do in about 30 sec – so didnt gain much.
BETTER OPTION: employ a more sophisticated nudging scheme to the nearest RAP/HRRR. If you going to modify the sounding – might as well nudge it to a model analysis – which have a long history of being used in severe wx research.
The reason the modification algorithm adjusted only the surface values is because the scene was cloudy for the 7 hours preceding the satellite retrieval. For each of those hours, the routine deepens the boundary layer only if the surface skin temperature exceeds the surface air temperature, a condition that is typically not met in cloudy conditions. In these cases, the surface values of the retrieval are adjusted to the RTMA values, but that is all. As for adjusting to the RAP or the HRRR, we considered that option, but decided on a model-free approach.
Jack Dostalek(CIRA, Colorado State University)