Mountains support roughly one-third of all land-dwelling species and supply water for nearly half of all people1. The ecology of mountain environments is strongly influenced by climate2–4. For example, because temperatures drop as altitude increases, organisms that have greater cold tolerance are favoured at higher elevations. Accordingly, there is a rapid change in the species present as one moves up a mountain from the warm lowlands to the cold highlands. And because few organisms can withstand the most-extreme conditions, the total number of species tends to be low on mountaintops.
Climate change is now rearranging the pieces of this puzzle5,6, and ecologists are struggling to predict the picture that will emerge7,8. One major source of uncertainty is the extent to which the effects of human land use (activities such as farming and logging) might interact with climatic factors to shape the distribution of species and the operation of biogeochemical processes. Writing in Nature, Peters et al.9 report their analysis of an astonishingly comprehensive ecological data set from Mount Kilimanjaro (Fig. 1), which shows that temperature and rainfall modulate the effects of human land use on biodiversity and ecosystems.
A previous study10 from the same research group revealed that the number of plant and animal species declines at an almost linear rate as elevation increases on Mount Kilimanjaro, suggesting that temperature is the main determinant of species richness. Peters et al. have expanded the scope and scale of that earlier assessment. Their new study reports data gathered over 6 years by 50 researchers at 60 sites ranging from 866 to 4,550 metres above sea level. These sites represented both natural habitats, such as lowland savannahs and alpine heaths, and habitats that had been heavily affected by human activity, including cropland and logged forests. The authors noted the number of species of plants, animals and soil-dwelling bacteria at each study site. They also recorded data for 30 different ecosystem functions, which are processes related to the transfer of energy and matter through the system (for example, the rates at which plants grow, organic matter decomposes and greenhouse gases are emitted).
To convert human impacts into a common currency for use in statistical analyses, Peters et al. devised a quantitative metric of land-use intensity, which integrated information about several types of human disturbance. To assess climate, they monitored the average annual temperature and rainfall at each site. Researchers studying large-scale ecological phenomena are often forced to gather previously published data from disparate sources and stitch this information together for analysis, which can introduce biases and artefacts. By instead measuring a wide range of attributes in many places using standardized methods, Peters and colleagues were able to paint one of the most detailed ecological portraits achieved thus far for any mountain.
Peters et al. report that the combined effects of climate and human land use manifested in a consistent way for both plants and animals. Species richness hardly differed between natural and human-altered habitats at high elevations, but species richness was reduced in the low-elevation habitats that had been transformed by human activity. These trends were best described by statistical models that included interactions between climate and land-use intensity — in other words, the effects of land use were dependent on climate, and the interplay of both these factors was necessary to explain the patterns observed in the data.
The trends in ecosystem functions in relation to climate and human activity are harder to decipher. All but five of the functions studied were affected by land-use intensity, and in most cases the effects of land use depended on climate. But these interactions exhibited all manner of forms, defying attempts to identify a general pattern. A clearer picture emerged when Peters and colleagues amalgamated all 30 of the ecosystem functions into a composite statistical index: overall, ecosystem functioning was more heavily affected at sites with greater land-use intensities, and these effects were stronger at low and high elevations than at intermediate elevations.
Ecologists yearn for simple rules to describe how ecosystems respond to environmental gradients and to perturbations arising from human activity. Yet analysing such complex systems, with their multitude of interwoven parts, requires a level of statistical abstraction that makes it difficult to discover the fundamental mechanisms underlying the patterns in the data. Peters and colleagues have unveiled a rich tapestry of ecological patterns on Mount Kilimanjaro, but explaining why those patterns are shaped by climate and land use in the ways that they are stands as a non-trivial challenge for future investigation.
As with any large-scale comparative analysis, one must also consider potential alternative explanations for the results. For example, on Mount Kilimanjaro, areas higher than 1,800 metres above sea level are part of a national park, and this designation places constraints on human land use. Indeed, Peters and colleagues found that land-use intensity was greatest at low elevations (and therefore positively correlated with temperature) across the mountain, which is typical for mountains worldwide11. These correlations make it difficult to fully disentangle the roles of climate and land use. Is it possible that human impacts were greatest at low elevations simply because human activity was much higher outside the national park?
To address this question, the authors carried out more analyses on different subsets of their data, which reinforced their original conclusions. Nevertheless, further work will be needed to establish the degree to which variation in human impacts at different elevations is governed by biophysical mechanisms, as opposed to reflecting trends in human behaviour that stem from both climatic and legal restrictions on land use.
As mountain environments heat up in a warming world, what can be done to safeguard their great biological wealth? Neither climate change nor human pressure on mountains will stop any time soon, but areas can be protected from intensive land use, and that can make a difference. Peters and colleagues’ results indicate that such protection would need to span a range of elevations, from the low-lying sites that are currently most vulnerable to human impacts to the highland areas that will provide future homes for refugee species moving upslope. Nearly 40% of all mountain ranges lack any strictly protected nature reserves, and vanishingly few have conservation areas that span the entirety of their elevation12. There is an urgent need to expand the world’s protected areas to achieve better coverage of elevation gradients13.
Nature 568, 38-39 (2019)