When reprojecting data to WGS84 which transformation is the best one to use?
When reprojecting data to WGS84 which transformation is the best one to use? For Context, we
have an IMS site of the continental US and have a large variety of datasets in different
Mapping Center Answer:
You will need to test your data to see which transformation works best. To do that you’ll add the data to a new map and then be working with the Data Frame Properties Dialog’s Coordinate system tab, in particular, the Transformations button. Despite what the popup help for the Transformation button says, here’s how it really works.
You can either set an empty data frame’s coordinate system yourself, or the first dataset you add to an empty data frame will set the coordinate system to whatever that data’s coordinate system is. Also, an empty data frame in this case means a new, empty data frame. If you remove all the layers from an existing data frame, we still consider it “dirty” and won’t change it’s coordinate system when you add a layer.
So, for your case if you add a WGS84 dataset to a new empty map’s default data frame, the data frame’s coordinate system will be set to WGS84. You’ll want to use your data in this case as you’ll be assessing the quality of the transformation method based on your data.
So, that gets us to the meat of your question. If you add three more datasets, each using a coordinate system which uses a different geographic coordinate system (GCS) than WGS84, you must set the transformations; ArcMap won’t do it for you (as the help suggests; there is only one case where ArcMap does something automatically, which is when there is a pairing of NAD27 and NAD83 data, which will automatically add the NADCON transformation).
Once the data frame’s coordinate system is set, each time you add data to your data frame that uses a different GCS, that GCS’s name will be added to the list at the top of the dialog box that is shown when you click the Transformations button on the Data Frame Properties dialog’s Coordinate System tab. So, when you look in that dialog box and see two or more GCS names in that top list, you know your data are stored in multiple GCSs.
To change the transformation, click the Transformations button and select one of the GCSs in the top list. The GCS in the “Into” drop-down combo box should be set to what your data frame is using (WGS84). Then you can set the “Using” method. Depending on the combination of GCSs there may be several that make sense, or none at all. Once you’ve chosen a method, explore your map and see how well the transformed data conforms to the base WGS84 data. If its good, then you’ll know what transformation to use when you reproject the data.
In general, this is a tedious process; however it is far less tedious than finding and fixing errors late in your workflow that are essentially errors of assumption about whether to transform or which transformation to use. Ironically, in most cases I’ve found that most ‘errors’ in my data did not have so much to do with the transformation method as they had to do with quality of data capture/production. Getting datasets that were created by two different organizations to match perfectly is pretty hard to do.
If you would like to post a comment, please login.