Quantifying Urban Expansion and Its Association with Agricultural Land Loss in Busasamana Sector, Nyanza District, Rwanda (2010–2024): A GIS and Remote Sensing Assessment

Abstract

Aims: The study aims to quantify the pace of urban expansion in Busasamana Sector, Nyanza District, Rwanda, over the period 2010–2024, using Geographic Information Systems (GIS) and remote sensing techniques, and to establish the strength and spatial pattern of its relationship with agricultural land loss. Study Design: A quantitative, longitudinal, retrospective, GIS-based observational study. Place and Duration of Study: Busasamana Sector, Nyanza District, Southern Province, Rwanda; study period 2010–2024, with fieldwork conducted in 2026. Methodology: Multi-temporal Landsat 5 TM (2010), Landsat 8 OLI/TIRS (2018), and Landsat 9 OLI/TIRS (2024) imagery, all at 30 m resolution, formed the core classification dataset for the three epochs and was supplemented by Sentinel-2 MSI (10 m) imagery, which was used only as an ancillary reference layer for the 2024 epoch. All imagery was acquired via USGS Earth Explorer and Google Earth Engine, pre-processed, resampled to a common 30 m analysis grid, and classified into four Land Use/Land Cover (LULC) classes (agricultural land, built-up land, forest land, grassland) using Random Forest and Maximum Likelihood Classification algorithms. Post-classification change detection, transition matrices, and annual change-rate calculations were applied. A minimum of 600 stratified random reference points, ground-truthed through field observation (handheld GPS) and key informant interviews, supported accuracy assessment. Results: Built-up land expanded from 2.80 km² (2010) to 7.40 km² (2024), an increase of 164.3% (0.33 km²/year), with the annual growth rate accelerating from 0.26 km²/year (2010–2018) to 0.42 km²/year (2018–2024). Agricultural land declined from 28.60 km² to 22.10 km², a loss of 6.50 km² (22.73%), with the annual loss rate rising from 0.28 km²/year to 0.72 km²/year across the same epochs. A strong descriptive inverse correlation (Pearson r ≈ -0.98, based on three temporal observations and interpreted descriptively rather than as inferential evidence) was observed between built-up expansion and agricultural land area. Transition matrix analysis confirmed agricultural-to-urban conversion as the dominant, essentially irreversible land change trajectory, spatially concentrated in the western and central portions of the sector nearest the Nyanza urban core and major roads. Conclusion: Urban expansion is strongly and consistently associated with agricultural land loss in Busasamana Sector. The co-located, temporally consistent pattern of conversion and the near-complete absence of reverse transitions indicate that urbanisation is the leading proximate correlate of that loss, although the observational design cannot formally rule out other contributing factors; losses have accelerated over time. Spatially targeted urban growth boundaries and agricultural land protection zoning are urgently needed to safeguard the sector's remaining farmland and food security.