Broad- and fine-scale vegetation patterns in the Waterberg, South Africa

Abstract

Natural areas are under threat due to land transformation, placing additional pressure on existing protected areas and emphasizing the need for more conserved areas. However, effective conservation planning requires considerable biodiversity data to identify priority areas for conservation, despite many areas being poorly sampled. The Waterberg, a mountainous area in the Savanna Biome in Limpopo, South Africa, is floristically relatively poorly known and currently vulnerable to land conversion. The aim of this study was, therefore, to better understand vegetation patterns in the Waterberg, so to inform conservation efforts in the region.

First, a vascular plant species list was compiled (mainly from online datasets) for the Waterberg Mountain Complex, comprising 2722 taxa from 901 genera, 182 families, and 61 orders. This includes 27 endemic taxa (percentage endemism: 0.99%) and 49 taxa of conservation concern. The Waterberg displays a high diversity of genera and families, comparable to other species-rich areas including the Soutpansberg and the Cape Floristic Region. The current estimates of the vascular plant endemism in the Waterberg are too low for the region to be considered a centre of plant endemism, but with its rich higher-order diversity the Waterberg could potentially be considered a centre of floristic diversity. Species composition in the Waterberg is relatively similar to two other mountainous areas in the Savanna Biome: the Magaliesberg and the Soutpansberg. However, the Waterberg was the most under-sampled of the three areas.

Second, the transferability of plant richness models between two large Waterberg reserves was tested to examine if the richness-environment relationship from one reserve can accurately estimate richness patterns in the other reserve. When richness models demonstrate good transferability, they allow accurate modelling of biodiversity for understudied regions. However, models for six plant richness variables (total species, grass species, herb species, woody species, genus, and family richness), based on a set of 16 predictor variables all consistently showed poor transferability. This result was independent of the three modelling approaches used, which varied the number of predictor variables and/or minimized differences in environmental conditions between the two reserves. The poor transferability of these richness models possibly reflects the degree to which each reserve contains novel environments absent from the other reserve as well as the aggregative nature of richness variables (i.e. they are the sum of the occurrence patterns of different species, with each species potentially responding differently to multiple components in the environment).

The plant richness models from this research are currently not useful for predicting richness to other areas in the Waterberg, due to their poor transferability. However, as the Waterberg region has a high plant diversity and has generally been poorly sampled, it is important to continue conducting more botanical surveys in the area while simultaneously trying to improve plant richness models as a secondary measure to address biodiversity gaps for the region. The combination of additional floristic surveys and spatial models of biodiversity data will help to inform conservation decisions in the Waterberg.