Search

Global warming may force montane species to shift upward to keep pace with their shifting climate niche. How species differences in such distribution shifts depend on their elevational positions, elevation-dependent warming rates, and other environmental constraints, or plant functional traits is poorly understood. Here, we analyzed for 137 Himalayan tree species how distribution shifts vary with elevational niche positions, environmental constraints, and their functional traits. We developed ecological niche models using MaxEnt by combining species survey and botanical collections data with 19 environmental predictors. Species distributions were projected to 1985 and 2050 conditions, and elevational range parameters and distribution areas were derived. Under the worst-case RCP 8.5 scenario, species are predicted to shift, on average, 3 m/year in optimum elevation, and have 33% increase in distribution area. Highland species showed faster predicted elevational shifts than lowland species. Lowland and highland species are predicted to expand in distribution area in contrast to mid-elevation species. Tree species for which species distribution models are driven by responses to temperature, aridity, or soil clay content showed the strongest predicted upslope shifts. Tree species with conservative trait values that enable them to survive resource poor conditions (i.e., narrow conduits) showed larger predicted upslope shifts than species with wide conduits. The predicted average upslope shift in maximum elevation (8 m/year) is >2 times faster than the current observations indicating that many species will not be able to track climate change and potentially go extinct, unless they are supported by active conservation measures, such as assisted migration. Abstract in Nepali is available with online material.

The mitigation of global warming and conservation of biodiversity are two significant environmental challenges today. Estimating and comparing forest carbon stock and plant diversity under different management approaches provide insight into the choice of management approaches for carbon and plant diversity management. We investigated the variation in carbon stock and diversity of plant species in two forest managements under different approaches: the Kakrebihar protection forest (PF) and Sano Surkhet community forest (CF) in Karnali Province, Surkhet, Nepal. In total, 63 sample plots (30 plots in PF and 33 plots in CF) were laid out systematically across the forests. Dendrometric measurements were carried out for trees, poles, and saplings, and representative leaf litter and herb samples were collected. Soil samples were taken at 10 cm, 20 cm, and 30 cm depths using a soil auger. The existing tree volume equations of tree species of interest were used to estimate tree volume, and species-specific wood density and conversion factors were used to obtain total biomass and carbon content. Soil samples were analyzed using the Walkley-Black method to determine soil organic carbon. PF had higher carbon stock, plant species richness, and abundance at the landscape level than CF; however, the scenario differed at the plot level. At the plot level, PF had significantly higher total carbon stock and biomass carbon stock than CF. However, PF and CF were statistically indistinguishable in term of soil carbon stock. At the plot level, PF and CF were statistically indistinguishable regarding richness, Simpson diversity, and Shannon diversity, but PF had significantly higher plant abundance than CF. In conclusion, the value of PF for carbon stock and plant diversity surpassed those of CF. This study suggests that PF might be a better strategy to enhance carbon stock in forests and maintain habitat for various plant species.

Tropical montane systems are characterized by a high plant species diversity and complex environmental gradients. Climate warming may force species to track suitable climatic conditions and shift their distribution upward, which may be particularly problematic for species with narrow elevational ranges. To better understand the fate of montane plant species in the face of climate change, we evaluated a) which environmental factors best predict the distribution of 277 plant species along the Himalayan elevational gradient in Nepal, and b) whether species elevational ranges increase with increasing elevation. To this end, we developed ecological niche models using MaxEnt by combining species survey and presence data with 19 environmental predictors. Key environmental factors that best predicted the distribution of Himalayan plant species were mean annual temperature (for 54.5% of the species) followed by soil clay content (10.2%) and slope (9.4%). Although temperature is the best predictor, it is associated with many other covariates that may explain species distribution, such as irradiance and potential evapotranspiration. Species at both ends of the Himalayan elevational gradient had narrower elevational ranges than species in the middle. Our results suggest that with further global warming, most Himalayan plant species have to migrate upward, which is especially critical for upland species with narrow distribution ranges.

The Himalayas are amongst the global hotspots of biodiversity. Climate warming is forcing species to shift upslope at an alarming rate in the Himalayas, posing a serious threat to its biodiversity. To ensure present-day conservation and to accurately predict how montane plant species are likely to respond to future climate change, it is urgently needed to understand how environmental conditions and species properties (i.e. functional traits) shape species’ distributions along an elevational gradient (i.e. species’ elevational distributions). In this thesis, I analysed the elevational distribution patterns of 277 plant species along an elevational gradient in the Himalayas in Nepal, and showed which environmental factors best predict species’ elevational distributions and whether highland species occupy wider elevational ranges than lowland species; how functional traits shape species’ elevational positions and lead to species partitioning of the elevational gradient; and how future global warming will affect species’ elevational ranges and distribution areas.In Chapter 2, I asked which environmental factors best predict species’ elevational distributions and whether highland species occupy wider elevational ranges than lowland species because species living at harsher, more stressful and variable highland conditions would have wider physiological tolerances to environmental conditions and therefore occupy wider elevational ranges. For the 277 selected Himalayan plant species, I modelled their environmental niches using MaxEnt, by combining species occurrence data with 19 climatic, topographic and edaphic factors. I found that mean annual temperature, followed by soil clay content and slope are the key environmental factors that best predict species’ elevational distributions in the Himalayas. Apart from obvious changes in temperature and its temperature (e.g. temperature extremes) and non-temperature (such as irradiance and PET) covariates, both soil conditions and topography thus also play a major role in shaping species distributions in the Himalayas. Mid-elevation species (2,000-3,000 m a.s.l.) had relatively wider elevational ranges than lowland and highland species. This suggests that mid-elevation species might be less affected by climate change since their wider elevational ranges may partially buffer against climate change induced range shifts.In Chapter 3, I asked how functional traits shape species’ elevational positions and lead to species partitioning of the elevational gradient. For a subset of 31 tree species that occupied different positions along the elevational gradient, I measured 39 traits related to carbon, water and nutrient use. I found that the highland species have a small size, small and thick leaves with low nutrient concentrations, and narrow conduits that allow them to conserve their slowly acquired carbon and nutrients and survive harsh highland conditions. In contrast, the lowland species have a large size, large and thick leaves with high nutrient concentrations, and wide conduits that allow them to be competitive and acquisitive under benign lowland conditions. I also found that stem traits (i.e. basal area) and branch traits (i.e. conduit diameter, leaf area per xylem area and specific branch length) that integrate multiple plant organs and functions are the best predictors of species’ elevational distributions. Stem and branch traits are therefore more important for species elevational distribution than commonly measured leaf traits.In Chapter 4, I asked how future global warming will affect species’ elevational ranges and distribution areas. For 137 tree species, I used the MaxEnt models obtained in Chapter 2 to project their spatial distributions using ‘near current’ (1985) and future (2050) predicted environmental conditions. Highland species show larger predicted elevational shifts than lowland species, probably because of higher rates of warming predicted for high elevations. The upslope shifts that species have to realize to track their climatic niche is predicted to be, on average, more than 2 times faster (8 m yr-1) than currently observed shifts in the Himalayas (3 m yr-1). This implies that species may not be able to track future environmental change. The majority (71%) of the tree species are predicted to enlarge their distribution area: these are particularly species from the lowlands that may expand their upper limit, or species from the highlands that may have access to new potentially suitable highland intermountain valleys and plateaus, whereas species from mid-elevations are predicted to contract their distribution area because less area is available upslope. Hence, the actual change in distribution area depends on the peculiarities of the Himalayan topography and the area available in each elevational belt. I found that the predicted elevational shift in response to climate change varies strongly across species and depends not only on the magnitude of temperature and precipitation change at that specific elevation, but also on local soil conditions. I also found that species with conservative trait values have the largest predicted upslope shifts suggesting that an easily measurable trait, such as conduit diameter, can be used as a proxy for predicting species elevational range shifts in the Himalayas, and possibly in other montane regions.I conclude that the plant species distribution in the Himalayas are best predicted by temperature and its temperature (e.g. temperature extremes) and non-temperature (such as irradiance and PET) covariates followed by soil clay content and slope, but that future distributions will also be affected by changes in precipitation. Multiple trade-offs in plant size, hydraulic efficiency and light competition determine species’ positions along this elevational gradient. The Himalayan plant species face an uncertain future, because they are projected to shift upslope at much faster rates in response to the climate warming than currently observed rates of shift. Based on these findings, here, I call for the establishment of migration corridors, for assisted migration such as direct seeding or planting in new potentially suitable future habitats, and for urgent and informed actions to conserve and improve biodiversity along the slopes of the Himalayas for the benefit of not only the montane communities but also the global community.

Plant functional traits determine plant performance and have therefore the potential to shape and predict species distributions along environmental gradients. This study analyses how traits affect tree species distribution along an elevational gradient in the Himalayas, Nepal. We addressed three questions: (a) what plant strategies can be distinguished among tree species? (b) how are plant traits and strategies associated with elevation? and (c) what plant traits are the best predictors of species positions along elevational gradient? We quantified for 31 tree species a set of 39 plant traits related to resource uptake, use and conservation. We analysed how traits cluster into separate functions using a cluster analysis, and how traits and clusters associate into distinct plant strategies using a principal component analysis. The cluster analysis showed five clusters of traits, reflecting (a) efficiency in vertical expansion, (b) efficiency in horizontal expansion, (c) efficiency in metabolism, (d) physical defence and (e) tree functional groups (conifers vs. broadleaf species). The first PCA axis reflects trade-offs in traits and clusters linked to elevation: highland species had trait values that increase safety against freezing induced cavitation, high solar radiation and strong wind, and that increase resource conservation. In contrast, lowland species had trait values that increase efficiency of resource acquisition, metabolism and expansion. Additionally, a bivariate analysis showed that the first PCA axis is more strongly related to elevation than the individual traits, indicating that the combination of traits is more important than the individual traits. An all subsets regression analysis showed that a small set of traits best explain species distribution: highland species had smaller size (low basal area), safer hydraulics (low conduit diameter) and lower leaf and branch display efficiency (low leaf area per xylem area, specific branch length) that increase persistence under harsh conditions. Remarkably, leaf traits were poor predictors of species' elevational positions. Synthesis. Multiple trade-offs in plant size, hydraulics and light competitiveness shape species distribution along the elevational gradient. Along this extreme environmental gradient, stem and branch traits that integrate multiple plant organs and functions are better predictors of species' elevational distributions than leaf traits.