[gdal-dev] Vertical datum conversion for SRTM data

[Greg Troxel]

Nikos Alexandris <nik at nikosalexandris.net> writes:

> * Nikos Alexandris <nik at nikosalexandris.net> [2017-11-21 10:25:42 +0100]:
>
>>Dear list, can GDAL handle the following exercise?
>>
>>Of interest is the Mean Sea Level, based on NASA’s SRTM. I am new in this
>>topic. Is the following equation correct?
>>
>>SRTM(MSL) = SRTM(WGS84) - MSL(WGS84) [*]
>
> Here https://gitlab.com/NikosAlexandris/mean-sea-level the exercise I
> did back then.  I never got to continue on this.
>
> The last three days of the past week, I've been an observer of a
> "geomorphometry" meeting. So, I asked someone knowledgeable to get some
> feedback.  It seems it's on the right track.  Perhaps it'll be of
> interest to some people.

I would caution that the phrase "mean sea level" is potentially
confusing.  It was used to refer to what was thought to be an
equipotential surface that lead to the Sea Level Datum of 1929, and
later called NGVD 1929 when it was realized that it was not actually an
equipotential surface.

I think you are perhaps talking about an actual mean of sea levels over
some area, where the levels are expressed relative to some particular
geoid.  But that's not clearly inferrable from the words, and I think it
would be helpful to be painfully pedenatically precise about terms.

The phrase "SRTM(WGS84)" does not clearly  distinguish between WGS84
ellipsoidal height, apparently the native measurement and height
relative to the WGS84 geoid.    I also have a hard time understanding
what "MSL(WGS84)" is.