LandScan USA: a high-resolution geospatial and temporal modeling approach for population distribution and dynamics

Budhendra Bhaduri,Edward Bright,Phillip Coleman,Marie L. Urban
DOI: https://doi.org/10.1007/s10708-007-9105-9
2007-09-29
GeoJournal
Abstract:High-resolution population distribution data are critical for successfully addressing important issues ranging from socio-environmental research to public health to homeland security, since scientific analyses, operational activities, and policy decisions are significantly influenced by the number of impacted people. Dasymetric modeling has been a well-recognized approach for spatial decomposition of census data to increase the spatial resolution of population distribution. However, enhancing the temporal resolution of population distribution poses a greater challenge. In this paper, we discuss the development of LandScan USA, a multi-dimensional dasymetric modeling approach, which has allowed the creation of a very high-resolution population distribution data both over space and time. At a spatial resolution of 3 arc seconds (∼90 m), the initial LandScan USA database contains both a nighttime residential as well as a baseline daytime population distribution that incorporates movement of workers and students. Challenging research issues of disparate and misaligned spatial data and modeling to develop a database at a national scale, as well as model verification and validation approaches are illustrated and discussed. Initial analyses indicate a high degree of locational accuracy for LandScan USA distribution model and data. High-resolution population data such as LandScan USA, which describes both distribution and dynamics of human population, clearly has the potential to profoundly impact multiple domain applications of national and global priority.
English Else
What problem does this paper attempt to address?