Choosing a coordinate system when data spans coordinate system zones

4206
7
Jump to solution
07-17-2015 01:02 PM
JasonFultz
New Contributor III

Hi all,

I'm curious to hear how others choose an appropriate coordinate system when creating a new feature class that has a large spatial extent and spans several zones.  Two scenarios come to mind:

1.  When digitizing polygons that fall in various regions across North America.  Is the correct choice to always choose a geographic coordinate system (such as WGS84) for this scenario?

2.  On a smaller scale, A Feature Class that will span a 10 mile region but spans two different state plane zones or UTM zones.  Is there a best approach for choosing which zone to use?

Any insight would be much appreciated.

Thanks!

Jason

0 Kudos
1 Solution

Accepted Solutions
MichaelLohr
Occasional Contributor II

Jason, in making a response to your specific question about choosing a system for a new dataset spanning several eastern US states, I would say this. If my new data points were determined by using Lat/Lon pairs like in Google Earth for a group of specific locations, then that dataset could be appropriately stored as WGS 84 or another similar spherical system. If your source data was based on a planar system, then  you would want to use that. You have to create your data from something. You are correct, the data frame can be set to a different planar system if need be, to match with other local data.

WGS 84 shows up in ESRI as a projected system. I have always used a Geographic Coordinate System for Lat/Lon, Like NAD83 or NAD83 HARN for local Lat/Lon features. WGS is a world based system (that's why Google uses it) and NAD is a North American system. Based on that you would assume the NAD was better suited to continental US. But the differences sometimes are small enough to be ignored. Even so, knowing the use of your data will help keep you out of trouble.

I would suggest contact with a good county or state surveyor with experience in regional coordinate systems. It doesn't take much researching to discover how little you know about this kind of topic. Thanks to Dan and Darren for their insights.

View solution in original post

0 Kudos
7 Replies
DarrenWiens2
MVP Honored Contributor

The main goal in choosing a projection is to minimize a certain type of distortion or accept the fact that all or some will be distorted more than others, between area, distance, or angle. Each projection has a sweet spot, at or between certain lines, where these distortions are minimized. You should choose a projection that fits your needs as far as type of distortion goes, as well as fits your geographical area.

With that said, your local geographers probably have things mostly figured out. See what they use. For example, I live in British Columbia, which spans several UTM zones. If a project fits neatly inside UTM zone 9, we will likely use that coordinate system. If it spans from zone 9 to 10, chances are we will default to BC Environment Albers. For the US, here is an extensive list of what the USGS uses.

DanPatterson_Retired
MVP Esteemed Contributor

Or to add to what Michael noted, there are more recent references in the ArcMap 10.3 which is still applicable

http://desktop.arcgis.com/en/desktop/latest/guide-books/map-projections/what-are-map-projections.htm

On the main help link Documentation | ArcGIS for Desktop  you can examine more topics for current and previous versions as well as ArcGIS Pro     

JasonFultz
New Contributor III

Thanks for the information, Dan.  I found this article very helpful.

0 Kudos
JasonFultz
New Contributor III

Thank You, Darren. You helped clear some of this up for me.  Sometimes I start thinking about coordinate system scenarios in my head then I quickly start over thinking them!  Thanks again

0 Kudos
MichaelLohr
Occasional Contributor II

Sometimes your choice of coordinate systems when data spans across more than one system is based on where the majority of your data will come from. I don't have experience in national mapping, but in Florida, there are three State Plane zones, North, East and West. Different agencies/municipalities will base their data on what the survey coordinate system is for their area. State wide agencies can span more than one of these zones, but they will choose to keep all their data in either one or the other zone, to minimize confusion.

If you will have a deliverable GIS product, a shapefile or geodatabase, you will want to have that delivered in the system that matches your client's base system. Yes, you can reproject the data prior to delivery also, if needed.

One reason you have multiple coordinate systems is to minimize the positional error generated by transforming coordinates from one system to another. Calculating planar positions (2D) from spherical coordinates (3D) (Latitude/Longitude), as one example, is a complex operation. Multiple systems have developed over the years to achieve the best accuracy possible. Most of these systems attempt to project the 3D coordinates onto a flat surface for depiction. Different systems will best preserve different parameters.

So, for example, if your project was 90% in one zone and 10% in a second zone, you would be better off to choose the zone where the majority of your data was located, to minimize distortion at the outside edges.

Since ArcGIS handles the data transformations on the fly when you have properly set projection files, it sometimes doesn't really matter which system you pick to use if you don't have a high accuracy criteria to meet or some other need that drives you to use a specific system.  Don't forget to think about those that might come behind you and use your maps. What will make the most sense for them.

A good reference to overview this mathematically complex topic:

ArcGIS Desktop Help 9.2 - About coordinate systems and map projections

Hope some of that is useful.

JasonFultz
New Contributor III

Michael, Thank you for the response and for shedding some light on this topic.  I originally started thinking of this question in my head from a standpoint of starting a project from scratch without any corresponding data.  If this were a project with various sets of data, I agree with your approach of matching to the existing system being used, or one that would match a clients existing data.

So as far as choosing a coordinate system for new data that would be scattered across the eastern United States (for instance from Illinois to Maryland), would an appropriate workflow be to create the data in WGS84 for the feature class and then put my Data Frame into a projected Coordinate System?

Again, thank you very much for your explanation and shedding some light on this topic, I found it very helpful.

Thanks,

Jason

0 Kudos
MichaelLohr
Occasional Contributor II

Jason, in making a response to your specific question about choosing a system for a new dataset spanning several eastern US states, I would say this. If my new data points were determined by using Lat/Lon pairs like in Google Earth for a group of specific locations, then that dataset could be appropriately stored as WGS 84 or another similar spherical system. If your source data was based on a planar system, then  you would want to use that. You have to create your data from something. You are correct, the data frame can be set to a different planar system if need be, to match with other local data.

WGS 84 shows up in ESRI as a projected system. I have always used a Geographic Coordinate System for Lat/Lon, Like NAD83 or NAD83 HARN for local Lat/Lon features. WGS is a world based system (that's why Google uses it) and NAD is a North American system. Based on that you would assume the NAD was better suited to continental US. But the differences sometimes are small enough to be ignored. Even so, knowing the use of your data will help keep you out of trouble.

I would suggest contact with a good county or state surveyor with experience in regional coordinate systems. It doesn't take much researching to discover how little you know about this kind of topic. Thanks to Dan and Darren for their insights.

View solution in original post

0 Kudos