Empiric recommendations for population disaggregation under different data scenarios

Area of Science:

• Geographic Information Science
• Remote Sensing
• Urban Planning

Background:

• High-resolution population maps are crucial for crisis management and urban planning.
• Earth Observation aids in disaggregating population data into fine-grained maps.
• Optimal methods, spatial units, and accuracy metrics for population mapping remain unclear.

Purpose of the Study:

• To provide recommendations for producing high-resolution population maps using remote sensing and geospatial data in urban areas.
• To evaluate the suitability of different population disaggregation methods based on data availability and resolution.
• To clarify the impact of spatial units and accuracy metrics on validation processes.

Main Methods:

• Conducted experimental research on 36 population disaggregation scenarios.
• Combined five top-down methods (dasymetric, statistical, hybrid) with varying data resolutions and availability (poor, average, rich).
• Systematically validated resulting maps using a two-fold approach and six accuracy metrics.

Main Results:

• Combining statistical and dasymetric methods yields better results with only remotely sensed data.
• Simpler methods are more suitable for highly-resolved input data.
• Using at least three relative accuracy metrics is highly recommended for robust validation.

Conclusions:

• The choice of population disaggregation method and validation strategy depends on data characteristics and landscape heterogeneity.
• Recommendations are provided to optimize efforts and time in future high-resolution population mapping.
• Understanding the influence of spatial units and accuracy metrics is key for reliable population mapping.