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 16 / 23 The loss of about 40% of natural savanna habitats (76% since 1911) significantly threatens Kilimanjaro’s biodiver- sity, as savanna ecosystems are crucial for plant species diversity [17,68]. This decline also affects traditional medic- inal plants, primarily found in Kilimanjaro’s lowlands [69]. This trend aligns with global patterns; over the last century, tropical savanna ecosystems have decreased by over 50% due to agricultural conversion [4]. In Sub-Saharan Africa, around 20% of former savanna land is now cropland, with much of the remaining area used for livestock grazing [70]. The FAO predicts further encroachment of subsistence agriculture into former savanna regions due to population growth [71]. Forests ecosystems have also significantly declined, from over 500 km2 in 1911–60 km2 in 2022 (Table 3), particularly impacting endemic species (Fig 9B). This decrease varied over time and between wards. Between 1911 and 1976, many natural forest ecosystems on the western and northern slopes were converted to exotic tree plantations, while substantial Fig 9. Biodiversity trends 1976-2022. (A) natural species, (B) endemics, (C) IUCN Red List species, (C) neophytes. Blue symbols 1976, red symbols 2002, green symbols 2022. Species numbers per km2 in correlation to population density at ward level with logarithmic regressions. Correlation indices are presented in Table 6. https://doi.org/10.1371/journal.pone.0334184.g009

Table 6. Pearson correlation coefficient for densities of plant groups vs population (Fig 9). Natural Endemic Redlist Neophytes 1976 (n = 82) −0.673*** −0.399*** −0.717*** 0.696*** 2002 (n = 82) −0.527*** −0.315** −0.589*** 0.459*** 2022 (n = 94) −0.488*** −0.341*** −0.419*** 0.137 ns n: number of wards, which changed after the division of several wards after 2002 (see Supplementary Information S1 Text). ***p < 0.001, **p < 0.01, ns: non significant. https://doi.org/10.1371/journal.pone.0334184.t006