Projection definitions for epsg:2391-2393 (Finnish KKJ)

[Külli Tiits]

Rahkonen Jukka wrote:

> Hello,
> 
> For some reasons there are two versions of Finnish KKJ parameters hanging around in the projection libraries.  I checked that the situation is the same at least in MS4W and FW Tools Windows packages.  In the files "\FWTools1.2.2\proj_lib\epsg" and "\ms4w\proj\nad\epsg" the definition are like this:
> <2393> +proj=tmerc +lat_0=0 +lon_0=27 +k=1.000000 +x_0=3500000 +y_0=0 +ellps=intl +units=m +no_defs  <>
> 
> In the same directories the corresponding "esri" files include also the towgs84 parameters:
> <2393> +proj=tmerc +lat_0=0 +lon_0=27 +k=1.000000 +x_0=3500000 +y_0=0 +ellps=intl +towgs84=-90.7,-106.1,-119.2,4.09,0.218,-1.05,1.37 +units=m +no_defs  no_defs <>
> 
> In practice this means that if you re-project something to/from anything else than between the KKJ zones
> a) Projections given as epsg:2393 and esri:2393 (or taken from esri .prj file) will lead to different result.
> b) If epsg-reference is used the result is always wrong resulting to something like 100-150 metre position error.
> 
> I think that the definitions in epsg file should be corrected. Or is there some reason, historical or other, for not including the towgs84 parameters there?
> 
> Regards,
> 
> -Jukka Rahkonen-

Hi,

I agree, that EPSG file could be corrected to get more correct results.

The question in my opinion is, what towgs84 parameters should be used. I 
do not know the history of the current ESRI file parameters, but in the 
official recomendation for coordinate transformations from KKJ to EUREF 
(JHS 153, 
http://www.jhs-suositukset.fi/intermin/hankkeet/jhs/home.nsf/files/JHS153/$file/JHS153.pdf 
   , page 12) the parameters have somewhat different values.. And it has 
been said, that the same paramaters as in JHS 153 could be used also for 
transformations to WGS84.

In practice it seemed to me, that though the parameter values in JHS 153 
and ESRI file are quite significantly different, the resulting geometry 
differences are rather small (less than 1 m). At least when compared to 
case, when no towgs84 parameters were used.

The second thing is that the epsg codes 3386 and 3387 (KKJ0 and KKJ5) 
should be updated in a similar way.


Regards,
Külli Tiits