I am reconstructing historical time series of land use and other related data from a variety of sources with different spatial and temporal resolutions and would be interested in what possible (hidden) problems could arise when aggregating and combining the data.

I am especially curious about the ecological fallacy and whether choices in the aggregation of temporal data affect change of support in the spatial dimension and vice versa?

References to literature are fine, but I am more interested in examples of how those interactions would arise if at all.

  • I've read this on stats.se and I am familiar with the basics of spatial statistics, but haven't yet read too many actual research papers that deal with this issue. – Torsti Feb 28 '13 at 10:35
  • I found a blog entry and an abstract, but they don't quite answer the question on their own. – Torsti Feb 28 '13 at 12:47

Have you read Openshaw's (1984) and Cressie's (1996) papers? This is a good place to start.

Openshaw, S. (1984). The Modifiable Areal Unit Problem. CATMOG 38. Norwich: Geo Books. ISBN 0-86094-134-5. http://qmrg.org.uk/files/2008/11/38-maup-openshaw.pdf

Cressie, N. (1996) Change of Support and the Modifiable Areal Unit Problem. Geographical Systems, 3:159-180.

If you examine the primary literature you will realize that the Modifiable Areal Unit Problem is inherently temporal in nature. Ecological Fallacy arises when you attempt to draw inference across levels of organization.

|improve this answer|||||
  • I haven't got my hands on Cressie (1996) yet, but Openshaw does not directly discuss the statistical consequences of temporal aggregation and choice of time periods and whether they modify the effects of spatial grouping and zoning. de Jong and de Bruin (2012) doi:10.5194/bg-9-71-2012 discuss temporal aggregation, but not in the context of MAUP. – Torsti Mar 1 '13 at 10:13
  • 2
    This short paper by Cressie (1998) seems to be directly relevant: Aggregation and interaction issues in statistical modeling of spatiotemporal processes. Geoderma, 85: 133-140. doi:10.1016/S0016-7061(98)00017-2. – Torsti Mar 24 '13 at 20:24

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.