Gain and loss: Human and environmental wellbeing – drivers of Kilimanjaro’s decreasing biodiversity

PLOS One | https://doi.org/10.1371/journal.pone.0334184 October 29, 2025 8 / 23 with area. This species-area relationship (SAR) is one of the few very well documented patterns of species richness [45,46]. This model is still the most commonly used and is considered the best for describing species-area calculations [47]. We used the calculated SAR for the 13 major land use and vegetation classes of Kilimanjaro of Hemp et al. [48] to scale up the species counts of the nine classes to 1 km2. Second, these classes show a certain overlap of species. For example, many species of open agricultural land also occur in agroforestry systems. Based on the analysis of the releves, we calculated this overlap which varied between the different classes and depended on the number of classes within a given area, in our case a ward. In one ward with all nine classes, the reduction factor increased to 0.51, meaning that the species numbers of each class had to be halved (see S1 Table with the species numbers and reduction factors for all combinations). Using the same approach, we calculated the number of endemic species, IUCN Red List species and neophytes, from the 772 releves (Table 2). In a final step, we used these reduced species numbers to calculate the final species numbers per km2 by relating them to the proportion of each LULC in the different wards. We used the detailed LULC classification by Hurskainen et al. [75] from the southern slope of Kilimanjaro to demon- strate the correlation between different LULC and species richness on a finer scale. Data analysis

We related changes in human population density at the ward level, to changes in land use and vegetation cover in the respective wards. We then analysed the effects of land cover changes on biodiversity using the number of species of vascular plants as an indicator. The total number of species and the number of different plant groups (natural vegetation plants, endemic and Red List species, neophytes) were used as indicators. To analyse trends in species numbers, we fitted linear functions using R2 as the measure of fit. The bivariate correlation between biodiversity data as response vari- able and population and land use data as explanatory variable was performed with Pearson’s correlation coefficient using SPSS 29.0.2.0. For critical remarks on the methodology see Supplementary Information S1 Text. Permits, inclusivity in global research We obtained the necessary permits to access the sites and conduct fieldwork from the Tanzania Commission for Science and Technology (COSTECH), Tanzania National Parks (TANAPA), and the Tanzania Wildlife Research Institute (TAWIRI). Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the Supporting Information (S1 Checklist). Table 2. Species numbers in the study area based on the 772 releves. LULC/ ecosystem type Number of releves Species number Number endemics Number IUCN Red List species Number neophytes Agriculture* 136 700 2 1 100 Agroforestry 76 581 4 0 138 Forest 101 846 25 16 62 Grassland 163 509 6 0 46

Regeneration 24 300 0 0 15 Savanna 272 1059 9 10 43 Water body 0 0 0 0 0 All classes 772 2057 35 26 180 *Including classes forest plantation and built-up (see Supplementary Information S1 Text) https://doi.org/10.1371/journal.pone.0334184.t002