Population Density

Silver

People per square kilometer (people/km²)

1km

One-off

2000, 2010, 2020, 2030

2030

Provides spatially explicit estimates of population distribution, showing how people are spread across rural and urban landscapes. Using Global Human Settlement Layer dataset, it integrates UN World Population Prospects to estimate population in one square kilometer grid cells and follows the Degree of Urbanisation (DEGURBA) Stage I framework to delineate settlement typologies and establish national benchmarks over time.

Global Human Settlement Population dataset employs a population disaggregation (or downscaling) methodology to estimate the spatial distribution of residents at a 100-meter resolution. Official census or administrative population counts from the Gridded Population of the World (GPW) v4.11 are distributed into grid cells based on the presence and density of human settlements. The use of built-up volume derived from satellite-based building height and surface dat as the primary covariate for population allocation significantly improves the spatial realism of the grids in vertical urban environments. This modeling is executed through Symbolic Machine Learning (SML) workflows that integrate multi-temporal data from Sentinel-2 and Landsat, ensuring that estimates remain consistent with United Nations World Population Prospects across all epochs from 1975 to 2030.

Validation was conducted through a comparative assessment using novel reference data to evaluate model predictions of resident population. Empirical evidence from these assessments indicates that the dataset achieves a total allocation accuracy of 83% for resident population at the 100-meter resolution. Additionally, the validation process leveraged remote sensing-derived estimates to perform semi-automated revisions of census units previously classified as "unpopulated," which helped reconcile discrepancies between official administrative statistics and Earth Observation data.