Register a parametric CRS for isobaric surface (air pressure)

[jerstlouis]

Several weather / climate datasets use air pressure as 3rd or 4th dimension, yet there is no registered CRS for this on the definition server. Such a registered CRS is necessary where only a URI reference is supported (e.g., CIS "srs" property, and "CRS by reference" mechanisms in OGC API - Common).

As Roger Lott pointed out:

pressure levels from a surface would be referenced to a parametric CRS. Vertical CRS now is explicitly a geometric coordinate system in the earth's gravity vector. Probably the CRS and Metocean DWGs [ should endorse the content ]

An example WKT2-CRS definition already exists and is seen in WKT2-CRS section 12.4 as pointed out by @ilkkarinne :

PARAMETRICCRS["WMO standard atmosphere layer 0",
  PDATUM["Mean Sea Level",ANCHOR["1013.25 hPa at 15°C"]],
  CS[parametric,1],
  AXIS["pressure (hPa)",up],
  PARAMETRICUNIT["HectoPascal",100.0]
]

A question has been asked whether OGC is the proper authority to register such a CRS. My thoughts on this is that if this the fastest way to get a proper atmospheric pressure CRS, then OGC should go ahead and register it, and it could be deprecated in the future or resolve to the same CRS as a more authoritative definition that comes along in the future.

Once a 1D CRS for the atmospheric pressure exists, we could use it in CIS as part of a compound-crs, using either the current mechanism, or a proposed mechanism to allow a list of component CRS that automatically results in a compound CRS.

We could also explicitly define new compound CRS that combines atmospheric pressure with lat/lon and/or time. We could also possibly have registered lat/lon(/h) and time compound CRS. (Though we are still lacking a proper time CRS).

[ilkkarinne]

Before jumping right in and registering an isobaric surface CRS, I would like to clarify what is meant by a particular height coordinate in this CRS. I know of at least two different reasons to using pressure based vertical CRSes, one in meteorology and the other in aviation.

As far as I know the air pressure is used as the vertical coordinate instead of metric CRS in atmospheric models and weather forecasting because the height from the ground (or the mean sea level) has considerably less meaning in meteorology than pressure, temperature and humidity. By defining surfaces of equal pressure, one of these highly meaningful parameters stays the same, which makes it easier to see meaningful variation in the others. The actual calculation of an atmospheric model may also be simpler using these isobaric surfaces as model levels, but this goes beyond my expertise.

The aviation use case is the definition of so called flight levels (FL), which are defined using air pressure in so called International Standard Atmosphere, and used for keeping the necessary vertical distance between aircrafts. The basic reason for this is that aircraft altimeters at least used to be based on measuring air pressure with particular calibration (see Wikipedia article for more info: https://en.m.wikipedia.org/wiki/Flight_level).

The issue that needs attention IMHO is guidance on vertical precision and transformations between pressure based and metric CRSes. At least in meteorology the transformation function between the pressure based CRS and a metric CRS varies in time and location due to the local atmospheric conditions. That would mean that any good precision coordinate transformation between the two would need information on the atmospheric conditions at the place and time of the measurement / estimation.

Perhaps precision of the CRS transformation does not have to be very high in many of the use cases, and some kind of theoretic model can be used for defining the mapping between a pressure based and a metric CRS, as is done for aviation flight levels which use ISA (https://en.m.wikipedia.org/wiki/International_Standard_Atmosphere).

[jerstlouis]

Thanks @ilkkarinne that's very useful to clarify and providing guidance on conversion to and from metric CRSes. @KathiSchleidt mentioned it was not an obvious task currently.

However, while the understanding of what is being defined is essential to the definition, I think the guidance on conversion from/to metric CRS is something that could be addressed separately from the registration.

[desruisseaux]

The transformation from/to metric CRS is indeed a separated topic than the definition of PARAMETRICCRS, but it still something that may need to be addressed by the naming authority. If we want to address the transformations, then in addition to defining the PARAMETRICCRS, we also need to define a separated object: the operation method. Both are at different paths:

• PARAMETRICCRS definition would be somewhere below http://www.opengis.net/def/crs/
• Transformation definition would be somewhere below http://www.opengis.net/def/coordinateOperation/

Some more explanation is given in MetOcean wiki on pole rotation operation method.

[ghobona]

This issue needs to be discussed by the CRS DWG and resubmitted once a definition is available for registration.

Closing the GitHub Issue and moving it to the CRS DWG mailing list.

[chris-little]

@ilkkarinne @jerstlouis @ghobona The example quoted above: `WKT2-CRS section 12.4

PARAMETRICCRS["WMO standard atmosphere layer 0", PDATUM["Mean Sea Level",ANCHOR["1013.25 hPa at 15°C"]], CS[parametric,1], AXIS["pressure (hPa)",up], PARAMETRICUNIT["HectoPascal",100.0] ]`

is the "WMO Standard Atmosphere" which in reality does not exist. It is the basis for the aviation Flight Levels. For practical purposes, a real, observed or forecast, surface pressure and temperature, and the surface, have to be specified and substituted into the PDATUM object. A duration of validity of the CRS is also needed, as the observed or forecast pressure, and temperature, are constantly changing. Aviation typically specifies a validity of 6 or less hours, but typically updates every hour. Aviation uses the coded tags QNH and QFE to denote surface pressure at MSL or aerodrome runway elevation respectively

[pebau]

plus, there is the elephant of MSL which knows many different (often national) definitions. Would we handle them as different datums? A formalization is most welcome as it opens the door for automatic conversion.

[chris-little]

@pebau And to upset the geodesists even more, the ICAO/WMO definition of MSL is where the atmospheric pressure is 1013.25hPa (corrected for temperature if not 15C). There is a standard formula for extrapolating from the actual measured surface pressure, at 2m about the surface, at a known altitude, down to MSL. And that 'known altitude' probably has a few baby elephants too.

[chris-little]

Thoe other piece of information that is needed for the Aviation Use case is that at low level, such as near airports, aircraft fly according to their elevation above the physical surface (not an ellipsoid), such as measured by a radar altimeter.

At higher levels, such as at cruising height, aircraft use the pressure level altitude as determined by the pressure sensing altimeter, calibrated according to the temporally valid observed or forecast surface pressure (QNH) for that location or region.

There is also a "transition layer" where the aircraft transfer between the two vertical coordinate systems. I am not familiar with the rules, which vary regionally and also whether ascending or descending.

HTH.