Addressing risks to biodiversity arising from a changing climate: The need for ecosystem restoration in the Tana River Basin, Kenya

Species distribution modelling and the Wallace Initiative This study uses models developed by the Wallace Initiative, a global project designed to assess climate change impacts on the distribution of species [13, 47]. It links outputs from ClimGen [48] with the species distribution model, MaxEnt [49], and data from the Global Biodiversity Information Facility [50]. The current version of the Wallace Initiative has a spatial resolution of 20km x 20km and uses a range of 21 alternative climate models (GCMs) in order to account for uncertainties in the projections. For this research, five terrestrial taxa (plants, amphibians, birds, mammals and reptiles) were analysed. The Wallace Initiative projects changes to species distribution with specific levels of global warming above pre-industrial levels: those corresponding to the Paris Agreement global tem- perature targets of 1.5 and 2˚C; the nationally-determined contributions (NDCs) which have a lower estimate of 2.7˚C and an upper estimate of 3.2˚C; and a no-mitigation ‘business as usual’ scenario of 4.5˚C [51]. This research involved three steps. First, the extent to which the terrestrial biodiversity of the Tana River Basin is affected by climate changes was determined. Specific changes consid- ered are (i) proportion of species richness lost, and (ii) presence of climate refugia. Within the Wallace Initiative, refugia are defined as cells where, in at least 11 of the 21 GCMs used, at least 75% of the current species richness can remain in a changing climate. Current land use, and

potential land use change are excluded from this mapping of refugia. Secondly, the extent to which the projected climate refugia are threatened by present and potential future land use was assessed. The locations of projected refugia were compared to (i) locations of protected areas, (ii) current agriculture within the basin, to determine which potential refugia have already been converted to alternative land use, and (iii) spatial informa- tion from management plans produced by the GoK to determine the areas most at risk from future development. Data on the location of agricultural land was obtained from the Land Cover (LC) project of the European Space Agency (ESA) Climate Change Initiative (CCI), which is a global land cover map at 300m spatial resolution [52]. The agricultural land classes were extracted and reclassified. This is necessary in order to determine whether historical land use change has altered the landscape in a way potentially preventing the movement of species and restricting their ability to respond to climate change. These areas could also be important areas for the focus of restoration activities. Spatial information from the National Spatial Plan 2015–2045 (NSP; [53]) which was produced by the Government of Kenya, was digitised using GIS software as original digitised files were not readily available. Third, the importance of facilitating movement was considered. The Wallace Initiative employs different dispersal scenarios to examine whether species can move to track their pre-

ferred climate. Within the Wallace Initiative, dispersal refers to the average long-term shift of an entire species’ range based on a review of rates in the published literature [13]. It is the geo- graphic speed of the whole population rather than the velocity of movement of individuals across a species’ climatically suitable geographic range (climate envelope) which is being mea- sured [10]. The previous steps of this research only considered a scenario in which species are not able to shift their geographic ranges. This is in common with many previous studies [54] and reflects the fact that, in practise, many species’ ability to shift their ranges will be severely limited by the present land use. However, it is likely that some species will be able to move with sufficient speed to shift their ranges in response to warming, so a sensitivity analysis was also carried out to show the potential importance of facilitating species’ movement. Warren et al. [13] reviewed average dispersal speeds in order to create a scenario where range shifts may occur at a realistic rate. This review concluded that mammals and birds are able to disperse at a realistic rate of, on average, around 1.5 km/yr, while plants, amphibians PLOS ONE Risks to Kenya’s biodiversity from climate change PLOS ONE | https://doi.org/10.1371/journal.pone.0254879 July 21, 2021 5 / 25