Skip to main content

Oak Ridge National Laboratory (ORNL)

Oak Ridge National Laboratory (ORNL) has expertise in advanced materials, supercomputing, neutrons, and nuclear science to national priorities in energy, security, and scientific discovery.

Data Access

  • LandScan USA provides estimated population counts at 3 arc-second resolution for nighttime and daytime scenarios, for the Continental United States, Hawaii, and Alaska; nighttime population estimates at 3 arc-second resolution are also available for Puerto Rico.  Residents, prisoners, workers, students, and shoppers are modeled as baseline population estimates, capturing the diurnal variations of the U.S. population. 
HIFLD Secure
  • LandScan HD provides population estimates at 3 arc-second resolution for selected countries and regions. LandScan HD modeling is tailored to the unique data conditions of individual countries or regions. Each LandScan HD dataset incorporates current land use and infrastructure data from a variety of sources; applies population density estimates from census sources, from ORNL’s Population Density Tables (PDT) project, and/or from microcensus field surveys; and leverages the high resolution settlement and building layers created at ORNL. 
  • LandCast comprises locally adaptive, spatially explicit population projections for the contiguous United States for 2030 and 2050 represented as gridded datasets. For each target year, the projected distribution models an ambient population at a spatial resolution of 30 arc-seconds (`~1 km) based on a business as usual scenario. LandCast builds on the intelligent dasymetric modeling techniques employed by LandScan Global and LandScan USA to down-scale national level population projections and model spatial population growth at the local level. The underlying model is both locally adaptive and spatially explicit to account for local subtleties unaccounted for in most large-scale projection models.
  • LandScan Settlement and Building Layers are the result of advances in computer vision and access to high performance computing (HPC) at ORNL for identifying settled or built-up areas in support of high resolution population modeling. Both texture-oriented machine learning as well as deep convolutional neural networks have been used in conjunction with ORNL HPC resources and a variety of high resolution image sources to generate settlement and building areas for select areas ranging in resolution from built-up areas at 8 meters to precise building delineations at 0.5 meters. These layers form the spatial foundation for mapping the distribution of population in LandScan HD and LandScan USA.

Data Products

  • LandScan Global is a 30 arc-second (~1 km) resolution gridded population dataset representing an “ambient” (average over 24 hours) population count. The LandScan algorithm, an R&D 100 Award Winner, uses spatial data and imagery analysis technologies and a multi-variable dasymetric modeling approach to disaggregate census counts within an administrative boundary.  Since no single population distribution model can account for the differences in spatial data availability, quality, scale, and accuracy as well as the differences in cultural settlement practices, LandScan population distribution models are tailored to match the data conditions and geographical nature of each individual country and region.

Related Publications

Supporting Activities

  • Population Density Tables (PDT) ( is a data content management and machine learning system that reports average occupancy (people/1000 sqft) with confidence for night, day and episodic events at the national and sub-national level and for over 50 buildings types worldwide. Open source data such as journal articles, country statistical data surveys, humanitarian reports, real estate databases and more is collected for spatial and temporal information on human activity related to building use (number of people and area), and used to develop occupancy reports that dynamically update the reported occupancy.  In addition, sociocultural knowledge is captured and included through the use of a Bayesian modeling process. For complete transparency of the data input and modeling process used in reporting the average occupancy, the open source data and formulas used to develop occupancy are found within each occupancy report in the PDT system. 
Partner Type