Spatial identification of priority urban tree planting sites in Addis Ababa via Google Earth Engine–based multi-criteria decision analysis

Abstract Urban tree planting is a cost-effective, nature-based solution for climate resilience, yet its implementation in rapidly urbanizing metropolises is frequently obstructed by land scarcity and competing developmental priorities. This study develops a reproducible, cloud-based multicriteria decision analysis (MCDA) framework within the Google Earth Engine (GEE) to prioritize urban tree planting sites in Addis Ababa, Ethiopia. The spatial model integrates key biophysical and anthropogenic variables: land use/land cover (LULC), elevation, slope, soil texture, rainfall, and land surface temperature (LST). We applied the analytical hierarchy process (AHP) to derive criteria weights with a validated consistency ratio. Adopting the FAO land evaluation framework, the study classifies site suitability into five distinct tiers: S1 (highly suitable), S2 (moderately suitable), S3 (marginally suitable), N1 (not currently suitable), and N2 (perhaps not suitable). Spatial analysis revealed that 15.44% of the study area was highly suitable and that 20.43% was moderately suitable. Conversely, a substantial majority of the landscape (62.2%) was classified as permanent and not suitable due to hard infrastructural constraints. Validation via a stratified random sample (n = 120) yielded an overall accuracy of 93.33% (Cohen’s kappa = 0.896). Sensitivity testing, performed via a ± 10% perturbation of the LULC weight, shifted the highly suitable area by approximately ± 3–5 percentage points, confirming the absolute dominant influence of urban land use on final suitability. The results demonstrate that the physical availability and zoning compatibility of land, rather than biophysical limitations, are the principal barriers to urban tree planting in Addis Ababa.