Building-Level Binary Dasymetric Mapping and Spatial-Statistical Analysis of Population Change in Rural Serbia

Abstract

This study primarily implements a building-level Binary Dasymetric Mapping (BDM) framework to analyse population change between 2011 and 2022 in Barje Čiflik, a rural settlement in southeastern Serbia experiencing long-term depopulation. It extends the analysis with spatial and classical statistical methods. High-resolution ancillary data—including manually digitised building footprints, the number of storeys, and building function, all field-verified with abandoned dwellings identified during survey work—were integrated with census counts to allocate population using volume-based weighting.

Population estimates were assigned to each residential building to derive indicators of absolute and relative change, as well as density variation. The analysis combines spatial statistics (Global Moran’s I and Getis–Ord Gi*) with classical statistical techniques (Ordinary Least Squares regression, Spearman’s rank correlation, and LOWESS smoothing) to detect clustering, structural correlates, and spatial patterns of demographic change.

Results show that depopulation is spatially clustered, particularly in peripheral areas of the village, and that larger and multi-storey dwellings are more prone to decline. While density change was modest and statistically dispersed, the study highlights nuanced household-level transformations that remain obscured in aggregated data.

The findings demonstrate that integrating BDM with statistical analysis provides a replicable and cost-effective tool for fine-scale demographic research in rural environments with limited data availability, thereby supporting methodological development and spatial planning.