2

I am working on a map in ArcGIS ArcMap v10.2 with coordinate system GDA94 (UTM). I have just added a layer with points in latitude and longitude (WGS84). I specified that the layer was WGS84 and the points appear to be in the correct position on the map.

Do I need to use ArcToolbox (Data Management - Projections and Transformations - Feature - Project) to change this coordinate system for the final map, or is it okay as is?

1

If I read your question in the same context as your previous one, the difference between GDA94 and WGS84 seems to be the underlying concern. (Sorry if I am mistaken.)

These two ellipsoids are defined as follows:

WGS84
Ellipsoid: WGS84
Semi-major axis (a):               6378137.0 meters
Inverse of flattening (1/f or rf): 298.257223563

GDA94
Ellipsoid: GRS80
Semi-major axis (a):               6378137.0 meters
Inverse of flattening (1/f or rf): 298.257222101

You will see their semi-major axis are the same, while flattening values are slightly different. To see the impact of flattening, let's calculate semi-minor axis (N-S axis) (b) using equation f=(a-b)/a.

a <- 6378137.0
rf_wgs84 <- 298.257223563         # inverse f of WGS84
rf_grs80 <- 298.257222101         # inverse f of GRS80
# calculate semi-minor axis (b)
options(digits = 15)
(b_wgs84 <- a * (1 - 1/rf_wgs84))  # b_wgs84= 6356752.31424518 meters
(b_grs80 <- a * (1 - 1/rf_grs80))  # b_grs80= 6356752.31414036 meters

(Above example was done by R, but you can use any software to calculate b=a*(1-1/rf))

The difference between WGS84 (b_wgs84 in the above calculation) and GRS80 (b_grs80) is about 0.1 millimeters, for semi-minor axis. Given semi-major axis does not have such digit-accuracy, I tend to ignore their difference.

To me, it seems OK as is, to view the maps. For detailed analysis and such, please follow other answers.

0

I would recommend you to use only one crs. ArcGIS is viewing your points in the correct position, because of its "on the fly" projection. Which basically reprojects your data without changing the data.

You could experience problems performing analysis, when you use two layers with different crs. To avoid this problems, you will have to work only in one crs.

0

ArcGIS Desktop projects data on-the-fly, which is why you are seeing your data in the correct position on your map.

If each of your layer has the correct system defined for the data it contains, ArcMap can project the layers without any/much interaction from you so that they display in the correct place together. Sometimes you may need to select a Transformation to improve accuracy of the projection (you would need to select the same Transformation if you were permanently projecting the data anyway).

I would recommend not projecting your data unless you require it to be in the same CRS. This may be required for geoprocessing tasks, loading into a Feature Dataset, or if the on-the-fly dataset is very large (which would slow down the projection). However, for "mapping" it is usually quite safe to just leave the data in it's original state.

To answer your question

Do I need to use ArcToolbox (Data Management - Projections and Transformations - Feature - Project) to change this coordinate system for the final map, or is it okay as is?

I would say that it's probably OK as-is, and you should only Project it if you need that new dataset to be permanently in your other CRS or are performing tasks that require it to be projected. This could be especially true if the dataset is provided occasionally/frequently by a third-party, otherwise you would need to convert it every time you receive an update.

See also About editing data in a different projection (projecting on the fly) and What are map projections?

ArcGIS automatically integrates datasets whose coordinate systems are known

All geographic datasets used in ArcGIS are assumed to have a well-defined coordinate system that enables them to be located in relation to the earth's surface.

If your datasets have a well-defined coordinate system, then ArcGIS can automatically integrate your datasets with others by projecting your data on the fly into the appropriate framework—for mapping, 3D visualization, analysis, and so forth.

If your datasets do not have a spatial reference, they cannot be easily integrated. You need to define one before you can use your data effectively in ArcGIS. The spatial reference or coordinate system is metadata. It describes the coordinate framework that the data is already using.

-2

WGS84 is the most universal and popular CRS right now because it is great worldwide and great on mobile devices as well. There are other projections that are better suited for certain datasets though.

I live in San Diego, so if I am making a map with datasets specific to this area, then I normally go with the California State Plane VI. If I was up in Northern California then I would use another State Plane. If the datasets are going to be updated and added to locations not restricted to my region, then that's when I use WGS 84.

The 'Define Projection' tool is the best tool for re-setting your projections.This tool overwrites the previous projection, so use this tool to specify the correct one if need be.

There will most likely be problems on the other users end if the coordinate systems aren't identical.

  • Thank you. Just to confirm that I understand: I have to add WGS84 points because they come from an external source, but I need my map to be in GDA94, so I should use 'Define projection' once the WGS84 layer is added to make sure it is also GDA94? – NicoleG Apr 9 '17 at 9:40
  • I have asked this before, but just to confirm in this context, do I need to do this for GCS GDA94 and UTM GDA94 as well or are they okay because they are both GDA94. Thank you - it's confusing when you're starting out. – NicoleG Apr 9 '17 at 9:41
  • One final question: would I get the same result if I used 'convert coordinate notation' in geoprocessing (WGS84 to GDA94) and then add a layer with these GDA94 coordinates? – NicoleG Apr 9 '17 at 10:02
  • 1
    -1 You should use the CRS that is most suitable for each dataset. WGS84 doesn't suit every type of data, so don't just use it because it may be considered "universal" or "popular" (I would argue against both of those, but this isn't the right place). Don't convert your projection if you don't have to. – Midavalo Apr 9 '17 at 14:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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