altimeter_to_sea_level_pressure potentially not working as expected?

[stefan-hofer]

Situation: I have a station reading of pressure (hPa), a temperature (C) and a height of the station (m). If I want to calculate the sea level pressure with altimeter_to_sea_level_pressure then I get weird results. In case of mslp_one below, I would assume that a pressure of 850 hPa at 1500 m above sea level with a temperature around freezing would yield a sea level pressure of around the standard value of 1013.25 hPa or similar. Am I simply interpreting the MetPy function incorrectly here? It also seems very insensitive to changing the elevation (mslp two).

from metpy.calc import altimeter_to_sea_level_pressure
from metpy.units import units

mslp_one = altimeter_to_sea_level_pressure(850 * units.hPa, 1500 * units.meter, 1 * units.degC)
mslp_two = altimeter_to_sea_level_pressure(850 * units.hPa, 150 * units.meter, 1 * units.degC)
mslp_three = altimeter_to_sea_level_pressure(800 * units.hPa, 1500 * units.meter, 20 * units.degC)

print(mslp_one, mslp_two, mslp_three)

850.0590185132253 hectopascal 850.5282600553385 hectopascal 788.6961434502869 hectopascal

[dcamron]

The METAR altimeter report of an ASOS does not represent the current pressure observed by the station. If you calibrate an aircraft's onboard altimeter using the value reported in the METAR, it will estimate your craft's elevation at ground level to be the same as the known station elevation. The altimeter reading itself is basically an estimate of sea level pressure given a standard atmosphere.

Internally MetPy works backwards from altimeter readings to calculate the pressure observation at the station via altimeter_to_station_pressure [see docs link], and then includes the observed temperature in its estimation of Sea Level Pressure in altimeter_to_sea_level_pressure. [See docs link] for equations and references.

MetPy currently has no public functions to estimate MSLP from observed pressure accounting for temperature, moisture, etc. We may add this in a future release and we definitely welcome community PRs for this functionality! Functions like scale_height and virtual_temperature might help calculations done by hand in the meantime if needed. I hope this helps.