I have a shapefile that contains data of a country from the following Geographic Coordinate System ( GCS_North_American_1927 aka NAD27 )

contents from the shape file .prj: GEOGCS["GCS_North_American_1927",DATUM["D_North_American_1927",SPHEROID["Clarke_1866",6378206.4,294.9786982]],PRIMEM["Greenwich",0.0],UNIT["Degree",0.0174532925199433]]

I have the following questions regarding this data.

1.) When i load this shapefile into SQL server in NAD27 Geographic coordinate system and also convert it to WGS 84 and load it into sql server. When i render both on google maps they look as if they overlap perfectly. Is this because google doesn't have the precision to go out to the proper decimal places to render the spatial data?

NOTE: Google maps assumes you will give it WGS 84 spatial data to work with.

2.) Are Transformations lossy? for example if i perform the following transformation

NAD27 -> WGS 84 and then go back WGS -> NAD27

do i get the same exact spatial information back or is some precision lost?

3.) If i transform my information from NAD27 to WGS84 to be displayed on google maps will my coordinate alway show the right position of the spatial data relative to WGS84 or will it be slightly off?

  • 2
    First off, GSE's question/answer format works best with one question per Question. Please edit the for your most pressing question. But you should also consider the possibility that the original data was incorrectly marked NAD27 (this was the default projection in ArcInfo back before folks paid attention) -- if it lines up with WGS84, it's not likely to have been NAD27.
    – Vince
    Mar 7, 2014 at 15:04

2 Answers 2


projections of vector datasets are not lossy in theory (if you don't have rounding errors, of course, which will depend on your software and your data storage precision).

You should however be careful with projection of raster datasets, which are lossy because of the resampling.

Now, if you change the datum (from one geographic coordinate system to another one), you need to use some approximation and different approximation do exist. So your transformation will be inaccurate and you could degrade your information when you go back to the original CRS with another approximation than the one used at first.


Some kinds of transformations are lossy. Transverse mercator for example can only be used for the visible half of the globe. The back side gets lost when you try to re-transform.

Mercator only looses the poles, because it is accepted that the cylinder is unrolled in a world map.

This is important if you have huge shapefiles, e.g. complete Russia or the Natural Earth shapefiles.

  • This is more a projection change issue than NADCON lossiness, but NADCON on non-reversible projections is likely to be lossy.
    – Vince
    Mar 7, 2014 at 16:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.