The multivariate index of ecosystem multifunctionality.
To quantify the overall differences in ecosystem functions, we calculated the multivariate index of multifunctionality of the average change in ecosystem functions in the PTE-polluted habitats compared to the natural habitats (Fig. 6). The multivariate index of multifunctionality was best explained by the climate-PTEs-topography interaction model (Supplementary Materials Table 9, R2 = 0.86). However, the patterns of dissimilarity in the multivariate index of multifunctionality could not fully explain the dissimilarity structure of the species communities (Fig. 1d and Fig. 6), indicating that the relationship between species composition and ecosystem functioning in actual ecosystems was more complex than expected. Across the elevation zones, the models showed that the changes in ecosystem multifunctionality (i.e., the average changes in the ecosystem functions of the PTE-polluted habitats compared with those of the natural habitats) decreased with PTE intensity (Fig. 5b (1)). However, as we took both climate and topographic factors into account, the opposite trend of ecosystem multifunctionality with m-PTE intensity (m-PTE is a complex variable taking all PTEs, topography, and climate variables into account) was observed (Fig. 5b (2)). This finding indicated that the changes in ecosystem multifunctionality would deviate from the actual situations if climate factors and topographic variables were not included as explanatory variables. In detail, in the four studied climate zones of the Qilian Mountains, the impact of climate and other factors on ecosystem multifunctionality was higher in PTE-polluted habitats of the relatively humid mountain grasslands and alpine deserts than in the arid low-elevation desert steppes (Fig. 5c). Moreover, stronger changes were observed for the soil-mediated ESFs than for the plant-mediated ESFs (Supplementary Materials Fig. 5 a-d), and a higher change in the soil-mediated ESFs occurred in the lowland mine areas (both desert and grassland) (Supplementary Materials Fig. 5 e, f).
———————-
Place Figs 5-6 here
———————-
The results of our study demonstrate the importance of fully considering climatic and topographic impacts while assessing the changes in plant diversity caused by mining activities in different elevation zones of mountain ranges. We highlighted that the presence of PTEs can unevenly affect plant diversity trends along elevation gradients. First, compared with natural habitats, the plant turnover in PTE-polluted habitats at mid-elevation sites may be greater than that at sites in any other climate zone of the Qilian Mountains. Indeed, at mid and low elevations, species habitats have a higher spatial coincidence, with many species cooccurring in narrow geographical ranges38,32,33. Moreover, extreme climatic conditions (e.g., hurricanes and heavy precipitation) have a greater impact on the habitats of mid- and low-elevation species, where plant richness is higher and more sensitive to PTE sediments. Second, speciation increases in the habitat mosaic of topographically complex areas34,35,. Interestingly, we found that in the mountains with very high peaks and more rugged terrain, changes in the PTE deposition in the topsoil were more likely to promote the upslope/downslope movement of rare species, leading to their disruptive selection and displacement39,45(Supplementary Material Fig. 9). Similar phenomena have occurred historically in other contexts32,36, for instance, when vegetation belts moved upslope during warm and wet interglacial periods, leading to the fragmentation of populations and genetic divergence. As temperatures dropped again in glacial episodes, vegetation belts moved downslope, forcing secondary contact of populations and leading to founder effects32,36. Third, the succession mechanisms of plants showed significant differences in different PTE-polluted habitats. In the mine areas of the lowland grasslands, several toxic and harmful herbaceous plants (i.e.,Stellera chamaejasme , Potentilla multicaulis, andPedicularis Linn) occupied the niche of the PTE-polluted areas in the humid mountain grasslands (Supplementary Materials Fig. 7 and Table 8). Species evenness showed greater changes, especially in the sites containing more ultramafic rocks (such as AR, an asbestos plant, XTS, and multimetal mines), where ultrahigh magnesium content and low phosphorus availability act as powerful and selective plant filters11. In contrast, in the mine areas located in the arid mountain desert, the large and medium shrubs restricted the dispersal of PTEs and protected the nearby small shrubs and grasses with high fertility, leading to the establishment of new plant communities. In general, compared with the vulnerable ecosystems in low-elevation dry deserts and warm grasslands, the mine areas located in alpine deserts and cold humid grasslands may have a stronger resistance to disturbance and may possibly support more rapid restoration processes.
Despite the lack of time-series data for the Qilian Mountains, the ”space-for-time-approach” could explain the changes in biodiversity and ecosystem functions caused by mining activities37. Based on the long-term effect of mining activities over the broad elevation gradients of the mountain range, the models we proposed demonstrated, from a holistic perspective, the importance of climate change on the impact of metal pollution on different spatial and temporal scales. Although previous studies have considered the critical importance of local adaptation along natural elevation gradients, they only explored the potential effects of zinc (Zn) and climate change (20 and 24 ℃) on certain species (Ischnura elegans )23. Our models, instead, took into account the majority of the potential drivers (12 variables) that shape the contemporary patterns of biodiversity and generated predictions, which we tested with independent data. More particularly, the models showed that the impact on biodiversity and ecosystem functions depends largely on the original climatic environment and geological conditions. In vulnerable mining ecosystems, the processes driven by climatic or other factors may present contrary effects on species richness and vegetation cover, which may conceal the actual succession effects and habitat expansion or contraction at landscape level due to the presence of PTEs. From a long-term perspective, at larger spatial scales (hundreds of kilometers) encompassing a range of ecosystems and climate types, the importance of topography declines, while broad-scale climatic variations become substantially more important38. Given that most mining areas in the Qilian Mountains, as well as in other mountain ranges, have been closed for only some decades and that the community succession in PTE-polluted habitats cannot usually reach a stable level after decades or even hundreds of years39, our findings have worrying implications for the impacts of global climate change. As a consequence, we strongly suggest considering the role of global climate change when developing management strategies for plant diversity conservation in the mine areas of high-altitude mountain ranges.
Acknowledgements We thank the Land and Resources Bureau, Environmental Protection Agency (EPA) (Qinghai and Gansu Province), and the Qilian Mountain National Park authority for their support, and for granting us access to the Qilian Mountain National Park; as well as all of the companies who allowed us to work and helped to collect data at mine areas of the Qilian Mountains. This work was supported by the Second Tibetan Plateau Scientific Expedition and Research Program (STEP) (Grant No. 2019QZKK1003), the Strategic Priority Research Program of Chinese Academy of Sciences (Grant No. XDA20040301), the National Natural Science Foundation of China (Grant No. 41890824), and CAS Interdisciplinary Innovation Team-Global Change Hydrology.
Author contributions J.Z.L and H.W.L designed the concept for ecological research of mine areas in Qilian Mountains. M.L.B. and L. H. developed the mathematical theory. J.Z.L and A.G. processed the data of plant diversity. J.Z.L, H.W.L., T.C.Y., P.P.T., S.S. F., Q.Y., Q.W.N., Y.Y.Y., C.Y., M.T., W.F., Y.X.X and F.P.Y. established study sites and collected data. J.Z.L., S.S.F, M.T., Y.Y.Y., Q.W.N. and C.Y. conducted the chemical analysis of the plant and soil samples. T.C.Y., P.P. T and Q.Y. processed the topography and climate data of the Qilian Mountains. All authors contributed to the subsequent drafts.
Competing interests The authors declare no competing interests.