Looking for info on how the NWS determines the height of the 1000mb surface when it lies below the station elevation.
Comparing the values in TTAA rawinsonde messages to values calculated using the equations in FMH-3 (page D-2) gives differences that suggest some details are missing in my understanding of how the 1000mb heights are being calculated.
Even if you accept that an extrapolation like this may be inaccurate, it should be possible to at least replicate the TTAA values by using the same protocol…which I apparently do not know. The differences in the values means I am not using the same method as the NWS
For example: for station 72357 (Norman, OK or OUN):