Spatial predictions for the distribution of woody plant species under different land-use scenarios in southwestern Ethiopia

Abstract

Context Deforestation, forest degradation and intensification of farming threaten terrestrial biodiversity. As these land-use changes accelerate in many landscapes, especially in the Global South, it is vital to anticipate how future changes might impact specific aspects of biodiversity. Objectives The objectives of this study were to model woody plant species richness in southwestern Ethiopia, for the present and for four plausible, spatially explicit scenarios of the future (‘Gain over grain’, ‘Mining green gold’, ‘Coffee and conservation’ and ‘Food first’). Methods We used cross-validated generalized linear models for both forest and farmland, to relate empirical data on total and forest-specialist woody plant species richness to indicators of human disturbance and environmental conditions. We projected these across current and future scenario landscapes. Results In both farmland and forest, richness peaked at intermediate elevations (except for total species richness in farmland) and decreased with distance to the forest edge (except for forest specialist richness in forest). Our results indicate that the ‘Mining green gold’ and ‘Food first’ scenarios would result in strong losses of biodiversity, whereas the ‘Gain over grain’ scenario largely maintained biodiversity relative to the baseline. Only the ‘Coffee and conservation’ scenario, which incorporates a new biosphere reserve, showed positive changes for biodiversity that are likely viable in the long term. Conclusions The creation of a biosphere reserve could maintain and improve woody plant richness in the focal region, by forming a cluster with existing reserves, would be a major step forward for sustainability in southwestern Ethiopia.

Land use intensification causes the spatial contraction of woody-plant based ecosystem services in southwestern Ethiopia

“…In the current landscape, woody-plant based ES hotspots spatially coincided most notably in forest, but numerous ES hotspots also occurred within farmland. The extent of forest (53%) and farmland (47%) in the baseline landscape was approximately balanced 33 , Fig. 3B.…”

“…Compared to the past, present access to ES generated from woody plants is more constrained due to a decrease in forest extent 27,33,39…”

“…In panel A, SNNP stands for Southern Nations Nationalities and People's Region 74 . Panel B shows the present day land cover map that illustrates the distribution of forest (in green color) and farmland (in brown) (from Duguma et al 33 ).…”

“…The mean annual temperature is approximately 20 degrees Celsius 60 , with mean daily minimum and maximum temperatures of 12 and 28 degrees Celsius, respectively 61 . We used the classification of farmland and forest 33 to separately model ES in farmland and forest, because different drivers operate in these two very different environments. Modeling results were finally merged for the whole landscape.…”

“…Combining these datasets, two specific objectives of this paper were (i) to predict the current availability of woody-plant based ES in the landscape, and (ii) to quantify potential changes thereof under alternative scenarios of landscape change. Following Duguma et al 33 , we modeled ES in farmland and forest separately and merged the results. From predicted ES maps, we calculated the richness of ES hotspots-that is, overlaps of (combined) priority areas for different ES-because such areas are especially important for the ongoing provision of services (Fig.…”