Disentangling the effects of global climate and regional land-use change on the current and future distribution of mangroves in South Africa

The mangrove distribution in South Africa is fragmented and restricted to small forest patches occupying only 16 % of the estuaries within the current range. In this study we used species distribution models to test (1) whether the absence of mangrove forest and its species (Avicennia marina, Bruguiera gymnorrhiza and Rhizophora mucronata) within their current range is driven by climate or by climate combined with human or geomorphic perturbation and (2) how climate change may potentially affect the latitudinal limit of the mangrove forests and its species in South Africa. We used three modelling techniques (generalized linear models, generalized additive models and gradient boosting machines) and a set of three climate-based predictive variables (minimum air temperature of the coldest month, waterbalance and growing-degree days) combined separately with an index of human or geomorphic perturbation. Climate variables for the future projections were derived from two general circulation models driven by two socio-economic scenarios (A2a and B2a). Within the range of the mangrove forest, the fragmented distribution of the mangroves in South Africa was not explained by our set of climate variables alone. The index of human perturbations slightly improved the predictions but the index of geomorphic perturbation did not. Climate change will create climatically suitable sites for the mangrove forest and the two species A. marina and B. gymnorrhiza beyond their current limits, but model outcomes did not agree on the future potential distribution of R. mucronata. We were able to successfully predict range limits and to detect future climatically suitable sites beyond the current limits. Factors controlling mangrove distribution within its range are still to be identified although absences were partly explained by human perturbations.