Display Full Text | Download Full Text
2.37 MB PDF file

Degree

Master of Arts, Ohio State University, Geography, 2005.

Abstract

This research implements and compares four different methods of areal interpolation for estimating the population within a set of target zones based on the known population within source zones. The methods include the areal weighting method (Lam, 1983), the pycnophylactic method (Tobler, 1979), an implementation of the dasymetric method (Wright, 1936, Eicher and Brewer, 2001) using remotely sensed data as ancillary information, and an implementation of the network method (Xie, 1995) using the road network as ancillary information. Both the dasymetric and network methods make use of ancillary data, whereas the areal weighting and pycnophylactic methods do not use ancillary data. The interpolation is conducted from source zones (U.S. Census tracts) of lower spatial frequency to target zones (U.S. Census block groups) of higher spatial frequency. The source zones are census tracts and the target zones are census block groups. The methods are implemented in three different counties: Franklin County, Ohio, Hamilton County, Ohio and Jefferson County, KY. Measures of the accuracy of the interpolation methods include the root mean square error (RMSE), an adjusted RMSE measure that normalizes the RMSE based on the population within each target zone and percent error maps for spatial errors. Other visualization methods include scatterplot diagrams and histograms of the errors associated with each of the areal interpolation methods.

This research finds that in terms of the RMSE, the network method produces the most accurate results, followed by the dasymetric method. The areal weighting method produces the least accurate results and the pycnophylactic method is better than the areal weighting method. The adj-RMSE shows more complicated results but mirrors the RMSE results in some cases. Patterns of spatial error are also examined by discussion of percent error maps.

Keywords

target zones; AREAL INTERPOLATION; areal weighting; pycnophylactic; dasymetric; source zones

Advisor

Harold Moellering

Pages

xi, 88 leaves : ill. (some col.)

Document number: osu1139949635
Permalink: http://rave.ohiolink.edu/etdc/view?acc_num=osu1139949635

This ETD has been downloaded 841 times (through March 2013)