Climate warming and range shifts of four medicinal plants in Africa: Insights from predictive bioclimatic models

Medicinal plant species constitute an accessible and affordable health-care alternative, and are an important source of livelihood for indigenous and rural populations. However, over harvesting and habitat losses due to land use and fragmentation have tremendously accelerated species loss, ultimately threatening their sustainability. The accelerated species loss is also compounded by the paucity of knowledge on the ecology and geography of medicinal species in Africa, posing limitations in conservation of their valuable habitats. Further, it is not clear how the niches and habitats are responding to current and future projected global climate change. This study employed ecological niche models (ENM) to understand the effects of climate change on four important and internationally traded medicinal plants in Africa: Pepper fruit tree (Dennettia tripetala), Olive tree (Olea europaea), Red stinkwood (Prunus africana) and Poison devil’s pepper (Rauvolfia vomitoria). We developed ecological niche models using a genetic algorithm for Rule Set Prediction (GARP) and Bioclimatic analysis and prediction system (BIOCLIM). Model inputs included climatic data (temperature and rainfall) and occurrence records of the four medicinal species cross Africa. The latter records were obtained from the Global Biodiversity Information Facility (GBIF) data portal and from several literature records, including several ethno botany collections across Africa. These records were used to produce georeferenced maps indicating areas that are currently ecologically suitable (1950-2000), in the medium-term (2020), and long-term or future periods (2050). Model evaluation was based on the area under curve (AUC) using receiver operating characteristic (ROC) analysis. To determine the effect of climate change on each species, we developed detailed comparisons of current and future (2050) potential distributions. Results showed that the two models (BIOCLIM and GARP) successfully reconstructed key features of the species nichie and geographic distribution. GARP was more consistent in prediction capabilities, but the two models were generally similar in prediction, with test AUC values ranging from 0.89 to 0.99 and 0.92 to 0.97 for BIOCLIM and GARP, respectively. The potential ranges for D. tripetala, P. africana and R. vomitoria are predominant in western, central and eastern Africa, while O. europaea suitable areas were predominant in southern, and parts of northern Africa. Future climate change scenarios indicated net habitat contractions for R. vomitaria (-7.70%), O. europaea (-4.01%) and P. Africana (-2.62%), while significant net gains were predicted for D. tripetala (9.97%). Overall, this translates to 4.36% net climate change induced medicinal species’ habitat loss in Africa. R. vomitaria is projected to lose habitats in central Africa (D.R. Congo, Angola and Central African Republic), but gain suitable ranges in east Africa (Uganda, Somalia and Kenya). On the Contrary, O. europaea is projected to lose ground in east Africa (Tanzania, Ethiopia and Madagascar), but gain in north Africa (Tunisia and Moroco) and south Africa (Botswana). Habitat gains for D. tripetala are projected in Chad, Nigeria, and Sudan, and losses in the Congo basin (D.R. Congo and Congo), and Mali. P. Africana suitable habitats are projected to shift southwards to Mozambique, Botswana and Zimbabwe, with significant habitat loses in Angola, D. R. Congo and Uganda. Our findings have provided evidence to infer areas suitable for conservation and bio- prospecting of the four medicinal species in Africa. Most of the species under consideration have indicated potential contractions in niches due to climate change, hence the need to devise climate adaptive conservation strategies for important medicinal plants in the continent.