Quantifying the Urban–forest Trade-off in a Rwandan Secondary City: A 35-Year Multi-temporal Remote Sensing Assessment of Huye District (1990–2025)
Dieudonne Nshimiyimana *
University of Lay Adventists of Kigali, P. O. Box 6392, Kigali, Rwanda.
Janvier Niyomugabo
University of Lay Adventists of Kigali, P. O. Box 6392, Kigali, Rwanda.
Pancras Ndokoye
University of Lay Adventists of Kigali, P. O. Box 6392, Kigali, Rwanda.
*Author to whom correspondence should be addressed.
Abstract
Rapid urban expansion is transforming land-cover patterns in Rwanda’s secondary cities, with potential consequences for forest resources and sustainable land-use planning. This study assessed long-term changes in built-up and forest cover associated with urban expansion in Huye District, Rwanda, from 1990 to 2025, and examined the spatial relationship between built-up expansion and forest-cover decline. A multi-temporal remote sensing approach was applied using Landsat imagery for four reference years (1990, 1999, 2018, and 2025). The imagery was pre-processed and classified using supervised maximum-likelihood classification into five land-cover classes, with classification accuracy assessed using error-matrix analysis and ground-reference data, achieving overall accuracies exceeding 85% for all years. Post-classification change detection quantified land-cover transitions and changes in built-up and forest areas, while Pearson correlation analysis assessed the association between built-up and forest areas, and a 5-km proximity buffer was used to examine the spatial concentration of forest-cover loss around expanding built-up areas. Built-up land increased from 2.80 km² (0.52% of the district) in 1990 to 49.30 km² (8.94%) in 2025, representing a 1,660.71% increase, while forest cover declined from 74.50 km² (13.79%) to 35.20 km² (6.51%), corresponding to a 52.75% reduction. Consequently, the built-up-to-forest area ratio shifted from approximately 1:26.6 in 1990 to 1.40:1 in 2025, indicating a substantial reversal in the relative extent of these land-cover types. Built-up and forest areas exhibited a strong inverse association across the four observation years (Pearson r = −0.92), although the relationship was not statistically significant at the 5% level (p = 0.078; n = 4). Spatial analysis further indicated that forest-cover loss was disproportionately concentrated within the 5-km zone surrounding expanding built-up areas. Overall, Huye District experienced substantial built-up expansion accompanied by pronounced forest-cover decline over the 35-year period. The strong inverse temporal association and concentration of forest loss near expanding built-up areas suggest that urban growth is an important spatial correlate of forest-cover change, although the limited number of temporal observations constrains statistical inference. The findings highlight the need to integrate forest conservation into secondary-city expansion planning through enforceable urban growth boundaries, protection of forested buffer zones, and targeted reforestation and restoration measures.
Keywords: Urban expansion, forest-cover change, land-use and land-cover change, Landsat, remote sensing, GIS, spatial proximity analysis, secondary cities