Modelling potential habitat suitability of three invasive alien plant species under projected climate scenarios and land-use/land-cover change in the Lake Zone of Tanzania

Abstract

Invasive alien plant species pose significant ecological and socio-economic risks across East Africa, yet invasion risk assessments in Tanzania have largely relied on single-species and climate-only projections. Such approaches may overestimate invasion potential by neglecting land-use constraints and inter-model uncertainty. This study evaluates the current and future habitat suitability of three invasive species: Chromolaena odorata and Lantana camara (terrestrial) and Pontederia crassipes in the Mara–Simiyu Region of northern Tanzania. Habitat suitability was modelled using MaxEnt and spatial block cross-validation. Projections were generated under three CMIP6 climate models (ACCESS-CM2, MIROC6 and MRI-ESM2-0) and four emission pathways (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5). We quantified suitable area using the 10th percentile training presence (P10) threshold and derived ensemble predictions by averaging across climate models. Climate-only projections indicate that terrestrial species retain moderate suitability under low-emission scenarios, but experience increasing fragmentation and contraction of highly suitable areas under higher-emission pathways. Niche overlap between L. camara and C. odorata was moderate to high across scenarios (Schoener’s D = 0.45–0.81), suggesting substantial climatic similarity and potential spatial convergence under future warming. In contrast, P. crassipes maintained broad climatic suitability across scenarios. When land-use projections were incorporated, suitable habitat estimates were substantially reduced and exhibited greater spatial uncertainty. For C. odorata , suitable habitat covered approximately 55–65% of the region under SSP1–2.6, but this uncertainty increased under high-emission scenarios (coefficient of variation ≈ 65–75% by 2090). L. camara showed even stronger contractions, with uncertainty exceeding 85–90% under SSP5–8.5. These findings suggest that climate-only models may overestimate potential niche space by ignoring land-use constraints. Therefore, integrating multiple environmental drivers provides a more realistic assessment of invasion risk and supports climate-adaptive management strategies.