A First Quantification of Plant Endemism in the Manica Highlands (Zimbabwe–Mozambique) and the Significance of Open Habitats

Abstract

The Manica Highlands is a tropical transboundary mountain system shared by Zimbabwe and Mozambique, covering 7400 km². Although extensively botanized for over 100 years, with many known endemic plant species, these have never been comprehensively enumerated or their patterns of distribution determined. Based primarily on herbarium specimens, we present here the first quantitative assessment of plant endemics, totaling 216 taxa (including subspecies and described varieties) and representing over 9% of the estimated total flora. The endemics are not evenly distributed across the Manica Highlands-77 taxa (36%) are confined to the Chimanimani Mountains (only 9% of the total area) 22 (10%) are confined to the Nyanga-Rukotso-Serra Choa area, while 67 (31%) are distributed across the whole Manica Highlands, including 27 on Mount Gorongosa. These data necessitate spatial changes to the original description of the Chimanimani-Nyanga Centre of Floristic Endemism and its renaming to the Manica Highlands Centre. Regarding ecological preferences, 173 (80%) are confined to open habitats (notably montane grassland) despite these habitats comprising only 22% of the area. By comparison, only 36 (17%) are found in moist forest or on forest margins. Some 24 (11%) are predominantly found above 2000 m elevation. Although it is evident that the unique quartzite outcrops and soils associated with the Chimanimani Mountains are a strong local driver of endemism, there is insufficient evidence to determine other edaphic drivers at present, but we suggest that the high levels of endemism seen may be driven by the low-nutrient and stressed nature of many of the open habitats. These findings show the significance and conservation importance of open montane habitats, which should help focus future conservation efforts across the region. Future structured fieldwork to explore the potential drivers of endemism would be very valuable, given that ecological data are very limited.