Plant specialisation may limit climate‐induced vegetation change to within topographic and edaphic niches on a sub‐Antarctic island

Extreme changes in temperature, rainfall and wind regimes have been correlated with plant species range expansion upslope on sub-Antarctic islands. Ongoing climatic changes are expected to continue driving changes in species distributions globally, but niche specialisations may limit the capacity for range shifts. We hypothesised that non-climatic characteristics of ecological niches of vascular plant species could limit climate induced range shifts. We determined the altitudinal ranges of vascular plant species (n=13) on sub-Antarctic Marion Island and measured air temperature, topographic, foliar and soil properties along transects on geologically distinct substrates. Climatic and non-climatic associations were determined using multiple linear regression and boosted regression tree (BRT) analyses. The degree of niche specialisation was determined using outlying mean index analysis (OMI) within the range of species on the island. Several species (7 of 13) exhibited niche-specialisation. Correlation analysis revealed that edaphic properties including soil depth, loss on ignition, the principal component of most soil nutrients (Mg, Cl, K, Ca, Cu, Zn, P, S), Si, Mn and clay dominated the BRT prediction of overall plant cover. Although air temperature was correlated with plant cover in linear models, model simplification dropped temperature in both BRT and linear models. As a consequence, multiple determinants, including temperature, climate, topography and soils control the distribution of vascular plant species on this sub-Antarctic island.极端气温、降水与风况的变化，已被证实与亚南极岛屿上维管植物（vascular plant）物种向高海拔区域扩张分布范围的现象存在显著关联。当前持续的气候变化预计将继续推动全球物种分布格局发生改变，但生态位特化可能会限制物种的分布迁移能力。本研究提出假说：维管植物物种生态位的非气候特征，可能会限制气候变化诱导的分布范围迁移。我们针对亚南极马里恩岛的13种维管植物，测定了其海拔分布范围，并在不同地质基底的样带上，测定了气温、地形、叶片性状及土壤理化属性。通过多元线性回归与提升回归树（boosted regression tree, BRT）分析，明确了气候与非气候因子间的关联。本研究采用偏离均值指数分析（outlying mean index analysis, OMI），在岛屿上的物种分布范围内量化了生态位特化程度。13个物种中有7种表现出生态位特化现象。相关性分析结果显示，包括土壤深度、烧失量、多数土壤养分（镁Mg、氯Cl、钾K、钙Ca、铜Cu、锌Zn、磷P、硫S）的主成分、硅Si、锰Mn以及黏粒在内的土壤理化属性，主导了提升回归树对植物总盖度的预测结果。尽管在线性模型中气温与植物盖度存在相关性，但模型简化步骤表明，提升回归树与线性模型均剔除了气温变量。由此可见，气温、气候、地形与土壤等多重调控因子，共同决定了这座亚南极岛屿上维管植物物种的分布格局。