Authors: Hannah Rosenblum*, U.S. Bureau Of the Census
Topics: Remote Sensing, Geographic Information Science and Systems, Spatial Analysis & Modeling
Keywords: Population distribution, remote sensing, built-up area
Session Type: Poster
Start / End Time: 8:00 AM / 9:40 AM
Room: Lincoln 2, Marriott, Exhibition Level
Presentation File: No File Uploaded
Current and accurate population estimates can facilitate disease burden estimation, displacement and resettlement monitoring, poverty mapping, and establishment of baseline data for a wide array of demographic analyses. However, limited access and a lack of detailed geographic and demographic data have compounded the challenges faced by those attempting to implement aid programs in South Sudan. Satellite imagery and ancillary datasets, used with population distribution algorithms, can provide updated population distribution data where up-to-date census or survey data are not available. The combination of an ongoing crisis and a substantial nomadic population that migrates seasonally requires a novel approach that takes into account the local context. This poster demonstrates the methods we used to produce detailed population maps of South Sudan. We employed multi-seasonal imagery and a novel semi-automatic method of unsupervised classification and reclassification to generate a built-up area layer. This layer is used with several other biophysical and anthropogenic datasets in a random forests model and dasymetric redistribution process to generate a gridded population distribution map. Because the last census conducted in South Sudan was completed in 2008, we produced a corresponding 2008 population map using Landsat imagery, as well as a 2017 version using Sentinel-2 imagery.