[Qgis-user] Reproject from WGS84 (EPSG: 4326) to ETRS89/UTM32N (EPSG: 25832) considering the epoch

[Greg Troxel]

Jochem Kail via QGIS-User <qgis-user at lists.osgeo.org> writes:

>>    Does your data have an epoch associated with each data point?  If no,
>>    does it have an epoch associated with the whole dataset?
>
> The data were all mapped in one month, i.e. one epoch, but several
> years ago (2011).

ok, makes sense.

> I would now like to compare these river cross-sections with data
> recently measured using a DGPS receiver using ETRS89/UTM32N as CRS in
> the settings.

sure, I can understand that sort of goal.

>>    Why do you believe that your data is in WGS84 (the ensemble)?
> The colleagues used a DGPS receiver and the setting for the CRS was
> "WGS84".

That is almost certainly confused.   When you use differential GPS, then
the coordinates that come out of the GPS receiver (GNSS?) are not in
WGS84, but they are in the CRS of the reference station.

Your statement is vague, in that "DGPS" implies pseudorange corrections
but *does not specify the kind of reference station or what its CRS is*.

>>    Why do you believe that your data is accurate to on the order of a
>>    mater?
> ...because it was taken with a DGPS receiver

To realize that accuracy you have to get the CRS right, and you have to
avoid the ensemble.

>>    Do the people who managed/prepared this data understand that 4326 is
>>    an ensemble and that it contains WGS84(TRANSIT)?
>
> No, nor did I. Up to now I thought that WGS84 is non-plate fixed and
> therefore a single-fixed geodatic datum, with a plate-fixed geodatic
> datum like ETRS89 "moving" relative to WGS84 with the European plate.

WGS84 (specifically EPSG:4326) is an ensemble of a number of datums, all
intended to realize a definition of a system.  In this respect it is
like ETRS89 and the various ETRF realizations.  The various ensemble
members have names like WGS84(TRANSIT) and WGS84(G2296), the most recent
member.   The system definition is more or less congruent with ITRS, and
WGS84 realizations are each very similar to the corresponding (in time)
ITRFxxxx.

You are correct that WGS84 is a global frame with a no-net-rotation
constraint, as opposed to plate-fixed datums.

So, the next step is to say that DGGPS is still vague, what reference
station, and what frame are the reference station coordinates in.  Then,
label that data as being in the right frame, and you will likely get a
good transformation.

I believe that it is extremely unlikely that the reference station
coordinates are in any WGS84 realization.  There is simply no method to
get accurate coordinates in that frame because carrier phase data from
the GPS reference stations are not published.  Every differential system
I have seen or heard about uses either a national/regional datum (e.g. a
NAD83 realization, some ETRF, etc.), or some realization of ITRS,
e.g. ITRF2008.

It is a major problem in GIS that data are mislabeled as being in WGS84
and that it is treated as an interchange format.  Both practices are
errors, but also the standard approach.