Ask A Cartographer

Mixing of scales within a same GIS product

February 22 2008 | 2 comments
Categories: ArcGIS Methods, Cartographic Design, Data Modeling

I am contacting you in the context of a large UN project that we are coordinating (SALB, http://www.who.int/whosis/database/gis/salb/salb_home.htm) and for which we are collecting GIS format maps from all around the world to be compiled, edited and then validated by the respective UN Member States.

In order to ensure the seamless aspect of this data set we are basing ourselves on an international inland boundaries and coastline template which unfortunately needs to be improved because of the poor accuracy level observed for some parts of the world.

In this context, and in order to try to accommodate between large and small size territories, we are wondering to what extent it would cartographically correct to mix data presenting a different scale within a same GIS layer ?

For example to mix 1:250'000 scale coastline with 1:1'000'000 scale international inland borders.

Your help on this issue would be very much appreciated.

Many thanks in advance

Mapping Center Answer:

I got to meet some of the folks at the U.N. who are responsible for knowing the state of political boundaries for the world. Talk about a job that's not easy.

That you're asking about the cartographic correctness impresses me; most folks would be consumed by either the potential lack of economic (monetary, resources, and skillsets) parity or the challenge of making sense of a soveriegn nation's depiction of their boundaries not exactly matching their neighbor's perceptions of where a shared boundary is located.

That last point is probably the one that will make your work most interesting; that is, shared boundaries that are depicted at different levels of resolution. Even though the boundaries are as accurate, they are not created equally in terms of resolution or consistency of creation. A set of boundaries for two adjacent countries might be compiled from the same source, but by two different people at slightly different times, and of course they won't line up perfectly. The problem you cited of the small country which can more easily be mapped at larger scales has higher resolution data than their large neighbor, and so the relatively more detailed boundary doesn't match.

As long as there are not any vastly different depictions of these boundaries, i.e., a coarser boundary overlaps it's neigbor and consumes all or a significant portion of one or more subunits, there's some hope in that the data can be edited to create a decent depiction of all the countries involved.

So while it obviously isn't possible to specify a highly precise threshold for positional accuracy for the geometry in these boundary features, can you at least specify a minimally tolerable standard. This would allow the more precisely defined data to be generalized to match the coarser data; making for not only adequate cartography, but potentially excellent cartography.

One tip I would have for folks using such data, which would allow them to get broader use out of some of the more difficult situations that could arise is to use the spatial adjustment tools in ArcGIS to adjust the worst situations. This would allow overlaps to be as diplomacy necessitates, and allow mapmakers to depend entirely on the data on the SALB website.

So, that's just one set of ideas from one person. I'll make sure a few of our folks who know about this area chime in as well.

Mixing of scales within a same GIS product posted by Steeve Ebener on Feb 26 2008 12:22AM
Dear Mapping Center Team,

Thank you very much for your email and the answer to my question.

If I understand you correctly, detailed dataset should anyway be generalized but not to the level where small territories would loose a significant part of their surface, am I right ?

If this is the case, this actually means that mixing scale within a same product can be considered as cartographically correct as long as this is justified by the above.

Many thanks in advance for your justification on this point and one more time for your precious help.

Best regards

Steeve Ebener
Delatyed (sorry) response posted by Charlie Frye on Apr 2 2008 3:54PM
Yes, you've got the right idea. All of this falls under the auspices of best available data. Your intention is that you'd do better cartography if you got better data, and in the meantime you won't be doing a disservice to the contributors.

If you would like to post a comment, please login.

Contact Us | Legal | Privacy |