Abstract

Informing a Population Distribution Model for Settlement Mapping
Track: Analysis
Authors: Eddie Bright, Jeanette Weaver

The Geographic Information Science and Technology (GIST) Group at Oak Ridge National Laboratory has long been involved in population distribution and dynamics research. The objective of this study was to maximize results from training data for large scale settlement mapping for the country of Afghanistan, and then to inform a population distribution model using the extracted settlements. Training data involves combining the spatial location of physical structures with a spectral feature return. The training data was created using an ArcMap add-in that was developed by the GIST Group in order for analysts to combine settlement research and high resolution imagery. Once the settlement extraction was completed the population distribution model was created using a variety of geostatistical methods and tools, including spatial disaggregation, aggregation, and interpolation. The result of combining demographic and socio-cultural settlement research, high resolution imagery, and feature extraction methods is a spatially fine resolution population distribution.