Optimizing Measurement of 4G LTE Broadband Access in Virginia
Wireless coverage is not confined to jurisdictional boundaries and populations are not distributed evenly within jurisdictions. However both population and coverage statistics are often reported at the jurisdictional scale. Using census-based data to measure a population’s wireless broadband coverage can be a challenge, as census population values are aggregated into jurisdictional boundaries instead of associated with the actual location of individuals. Furthermore, Census data accounts for population by household, essentially capturing, at a jurisdictional level, where people sleep, and not necessarily where they are all day. Alternatively, Oak Ridge National Laboratory developed the LandScanTM dataset, a 1km resolution raster surface that attempts to represent ambient population (or the average over 24 hours) through spatial data and imagery analysis, and by disaggregating census counts using multi-variable dasymmetric modeling. This analysis compared the use of U.S. Census population data at the block level and LandScan’s ambient population model to calculate the percent of population with 4G LTE broadband access in Virginia. 2013 4g-LTE coverage data was provider-reported and accessed through the Virginia Geographic Information Network. The results of this study show that only 4% of census blocks covered by 4G LTE differed by more than 5% when measuring population using LandScan and Census data, for the most part, at the census block level these data sets are comparable for use in analysis of broadband coverage, in Virginia.