广义模型及分类回归树在物种分布模拟中的应用与比较

比较3个应用较广的模拟物种地理分布模型:广义线性模型(GLM)、广义加法模型(GAM)与分类回归树(CART)对中国树种地理分布模拟的优劣,以提出更为合适的模拟物种地理分布模型,并用于预测气候变化对物种地理分布的影响。3个模型对中国15种树种地理分布的模拟研究表明:除对油松、辽东栎分布的模拟精度稍差外,对其余树种分布的模拟精度均较高,其中以GAM模型最好。结合地理信息系统(GIS),比较分析了这3个模型对青冈、木荷、红松和油松4种树种的地理分布模拟效果,结果亦表明:这3个模型均能很好模拟青冈和木荷的地理分布,而GLM模型对红松分布的模拟结果不太理想,3个模型对油松分布的模拟结果均不甚理想,其中以GLM模型最差。基于3个模型对未来气候变化下青冈与蒙古栎地理分布的预测表明:GLM模型与GAM模型对青冈分布的预测结果较为接近,青冈在未来气候变化情景下向西和向北扩展,而CART模型预测青冈在未来气候变化情景下除有向西、向北扩展趋势外,广东和广西南部的青冈分布区将消失;3个模型均预测蒙古栎在未来气候变化情景下向西扩展,扩展面积的大小为:模型的模拟面积>模型>模型。     

Generalized additive models in plant ecology

Abstract. Generalized additive models (GAMs) are a non-parametric extension of generalized linear models (GLMs). They are introduced here as an exploratory tool in the analysis of species distributions with respect to climate. An important result is that the long-debated question of whether a response curve, in one dimension, is actually symmetric and bell-shaped or not, can be tested using GAMs. GAMs and GLMs are discussed and are illustrated by three examples using binary data. A grey-scale plot of one of the fits is constructed to indicate which areas on a map seem climatically suitable for that species. This is useful for species introductions. Further applications are mentioned.     

Propagule vs. niche limitation: untangling the mechanisms behind plant species' distributions

Distinguishing the roles of propagule limitation and niche requirements in controlling plant species distributions is important for understanding community structure, invasion, and restoration. We used species distribution models based on plant and environmental survey data to assess the strength of species' affinities for particular environmental conditions. We hypothesized that species with statistically detectable environmental requirements were primarily niche-limited, while species with weak habitat affinities were primarily propagule-limited. We tested this hypothesis via a seeding experiment in which we compared species' reproductive fitness in occupied and unoccupied sites. Species that appeared to be niche-limited based on distribution models had lower fitness when planted in unoccupied sites, while species that models suggested were propagule-limited had equivalent fitness when planted in occupied and unoccupied sites. Our results demonstrate that within a single community, both species limited primarily by niche availability or primarily by propagule availability can be identified using observational data.     

Predicting the distribution of shrub species in southern California from climate and terrain-derived variables

Abstract. Generalized additive, generalized linear, and classification tree models were developed to predict the distribution of 20 species of chaparral and coastal sage shrubs within the southwest ecoregion of California. Mapped explanatory variables included bioclimatic attributes related to primary environmental regimes: averages of annual precipitation, minimum temperature of the coldest month, maximum temperature of the warmest month, and topographically-distributed potential solar insolation of the wettest quarter (winter) and of the growing season (spring). Also tested for significance were slope angle (related to soil depth) and the geographic coordinates of each observation. Models were parameterized and evaluated based on species presence/absence data from 906 plots surveyed on National Forest lands. Although all variables were significant in at least one of the species’ models, those models based only on the bioclimatic variables predicted species presence with 3–26% error. While error would undoubtedly be greater if the models were evaluated using independent data, results indicate that these models are useful for predictive mapping – for interpolating species distribution data within the ecoregion. All three methods produced models with similar accuracy for a given species; GAMs were useful for exploring the shape of the response functions, GLMs allowed those response functions to be parameterized and their significance tested, and classification trees, while some-times difficult to interpret, yielded the lowest prediction errors (lower by 3–5%).     

Prediction of plant species distribution in lowland river valleys in Belgium: modelling species response to site conditions

In ecological modelling, limitations in data and their applicability for predictive modelling are more rule than exception. Often modelling has to be performed on sub-optimal data, as explicit and controlled collection of (more) appropriate data would not be feasible. An example of predictive ecological modelling is given with application of generalized additive and generalized linear models fitted to presence–absence records of plant species and site condition data from four nutrient-poor Flemish lowland valleys. Standard regression procedures are used for modelling, although explanatory and response data do not meet all the assumptions implicit in these procedures. Data were non-randomly collected and are spatially autocorrelated; model residuals retain part of that correlation. The scale of most site-condition records does not match the scale of the response variable (species distribution). Hence, interpolated and up-scaled explanatory variables are used. Data are aggregated from distinct phytogeographical regions to allow for generalized models, applicable to a wider population of river valleys in the same region. Nevertheless, ecologically sound models are obtained, which predict well the distribution of most plant species for the Flemish river valleys considered.     

多元回归树与双向指示种分析在群落分类中的应用比较

采用两种群落分类方法——以环境梯度为分类依据的多元回归树(MRT)和多年来广泛应用的双向指示种分析(TWINSPAN), 对吕梁山南段森林群落进行了数量分类, 同时依据植物群落分类和命名原则, 对分类后的各群系进行命名, 并用吻合系数比较两种分类结果的吻合程度, 分析MRT的优劣势, 为以后选择合适的植物群落数量分类方法提供参考。结果表明: (1) MRT将41个森林样方分为4个群系, 分别为侧柏(Platycladus orientalis)群系、青榨槭(Acer davidii)群系、辽东栎(Quercus wutaishanica)群系、柿(Diospyros kaki)群系, 群系类型与TWINSPAN的分类结果相同; (2)根据群系的样方组成, 两种分类结果的吻合度较高, 吻合系数达80.5%; (3)与TWINSPAN相比, MRT同时以物种和环境信息为依据, 对有过渡性质的样方划分更为可靠。因此, 单纯从植被分类的角度来看, 尽管TWINSPAN的分类结果更客观, 但当TWINSPAN分类遇到困难时, 如在划分大样地连续样方或具有过渡性质样方时, MRT更有优势。     

The reliability of a composite biodiversity indicator in predicting bird species richness at different spatial scales

Several biodiversity features can be linked to landscape heterogeneity, that, in turn, can be informative for management and conservation purposes. Usually, the more the landscape is complex the more the biodiversity increases. Biodiversity indicators can be a useful tool to assess biodiversity status, in function of landscape heterogeneity. In this study, we developed a biodiversity indicator, based on Shannon diversity index and built from distribution maps of protected species. With such an approach, we seek to evaluate the feasibility of using a combination of target species as a surrogate for assessing the status of the whole bird community. Our approach was spread over multiple spatial scales, to determine which was the most informative. We selected four species protected by European regulation and generated a presence-absence map from species distribution modelling. We, therefore, used the FRAGSTATS biodiversity metric to calculate Shannon index for the overlapped presence-absence maps, at two spatial scales (500 m and 1000 m). Then, the relationships with the whole community was assessed through generalised least square models, at the spatial scale of 4 ha, 9 ha and 25 ha. Results showed that the higher rate of variability of community was explained by the biodiversity indicator at 1000 m scale. Indeed, the more informative spatial scale for the whole bird community was 9 ha. In addition, a pattern emerged about the relationships between biodiversity indicator and community richness, that is worth of further research. Our study demonstrates that the usefulness of surrogate species for biodiversity and community assessment can become clear only at a certain spatial scales. Indeed, they can be highly predictive of the whole community, and highly informative for conservation planning. Moreover, their use can optimize biodiversity monitoring and conservation, focusing on a small number of noteworthy species.     

Identifying biases at different spatial and temporal scales of diversification: a case study in the Neotropical parrotlet genus Forpus

The temporal origins of the extraordinary biodiversity of the Neotropical region are highly debated. Recent empirical work has found support for alternative models on the tempo of speciation in Neotropical species further fuelling the debate. However, relationships within many Neotropical lineages are poorly understood, and it is unclear how this uncertainty impacts inferences on the evolution of taxa in the region. We examined the robustness of diversification patterns in the avian genus Forpus by testing whether the use of different units of biodiversity (i.e. biological species and statistically inferred species) impacted diversification rates and inferences regarding important biogeographic breaks in the genus. We found that the best‐fit model of diversification for the biological species data set was a declining rate of diversification; whereas a model of constant diversification was the best‐fit model for statistically inferred species or subspecies. Moreover, the relative importance of different landscape features in delimiting genetic structure across the landscape varied across data sets with differing units of biodiversity. Patterns based on divergence times among biological species indicated old speciation events across major geographic and river barriers. In contrast, data sets more inclusive of the diversity in Forpus illustrate the role of both old divergence across major landscape features and more recent divergences that are possibly attributed to Pleistocene climatic changes. Overall, these results indicate that conflicting models on the temporal origins of Neotropical birds may be attributable to sampling biases.     

长白山阔叶红松林不同强度择伐后关键树种的竞争关系

天然林择伐改变了林分的树种组成和结构,导致林木竞争关系发生变化,进而影响树木的生长和种群的动态.关键树种在维持群落结构和生态系统功能方面具有至关重要的作用.为了弄清长白山区阔叶红松林典型林型关键树种竞争关系的特点,本研究以红松、紫椴和水曲柳为研究对象,运用Hegyi竞争指数分析了受不同程度采伐干扰后形成的原始林(未受干扰)、轻度择伐林、中度择伐林和重度择伐林的林分结构和竞争关系.结果表明: 与原始林相比,轻度择伐林中关键树种的种群结构没有显著变化;中度择伐林和重度择伐林中3个关键树种大树(胸径DBH≥20 cm)的密度和平均胸径均显著减少,但幼树(2 cm≤DBH<10 cm)的数量增加.所有样地中关键树种的竞争指数均随径级的增加而减小,且二者服从幂函数分布,而林木胸径生长到20 cm后,竞争指数进入稳定状态.在原始林、轻度和中度择伐林中,3个树种的幼树的竞争指数受到非冠层树种的影响最大,而红松小树(10 cm≤DBH<20 cm)和大树主要受红松种内和非冠层树种的影响;水曲柳主要受红松和非冠层树种的影响,紫椴主要受种内和红松的影响.重度择伐林中,白桦是3个树种的主要竞争来源,贡献率均超过50%.根据以上关键树种竞争关系的特点,抚育非冠层树种有利于3个关键种幼树的更新和生长;对于小树来说,还需要根据树种类型来采取相应措施,而大树不需要采用抚育措施.本研究对关键树种培育和天然林择伐后快速恢复具有指导意义.     

The quest for a null model for macroecological patterns: geometry of species distributions at multiple spatial scales

There have been several attempts to build a unified framework for macroecological patterns. However, these have mostly been based either on questionable assumptions or have had to be parameterized to obtain realistic predictions. Here, we propose a new model explicitly considering patterns of aggregated species distributions on multiple spatial scales, the property which lies behind all spatial macroecological patterns, using the idea we term 'generalized fractals'. Species' spatial distributions were modelled by a random hierarchical process in which the original 'habitat' patches were randomly replaced by sets of smaller patches nested within them, and the statistical properties of modelled species assemblages were compared with macroecological patterns in observed bird data. Without parameterization based on observed patterns, this simple model predicts realistic patterns of species abundance, distribution and diversity, including fractal-like spatial distributions, the frequency distribution of species occupancies/abundances and the species–area relationship. Although observed macroecological patterns may differ in some quantitative properties, our concept of random hierarchical aggregation can be considered as an appropriate null model of fundamental macroecological patterns which can potentially be modified to accommodate ecologically important variables.     

Modelling invasive alien species distributions from digital biodiversity atlases. Model upscaling as a means of reconciling data at different scales

Aim

There is a wealth of information on species occurrences in biodiversity data banks, albeit presence‐only, biased and scarce at fine resolutions. Moreover, fine‐resolution species maps are required in biodiversity conservation. New techniques for dealing with this kind of data have been reported to perform well. These fine‐resolution maps would be more robust if they could explain data at coarser resolutions at which species distributions are well represented. We present a new methodology for testing this hypothesis and apply it to invasive alien species (IAS).

Location

Catalonia, Spain.

Methods

We used species presence records from the Biodiversity data bank of Catalonia to model the distribution of ten IAS which, according to some recent studies, achieve their maximum distribution in the study area. To overcome problems inherent with the data, we prepared different correction treatments: three for dealing with bias and five for autocorrelation. We used the MaxEnt algorithm to generate models at 1‐km resolution for each species and treatment. Acceptable models were upscaled to 10 km and validated against independent 10 km occurrence data.

Results

Of a total of 150 models, 20 gave acceptable results at 1‐km resolution and 12 passed the cross‐scale validation test. No apparent pattern emerged, which could serve as a guide on modelling. Only four species gave models that also explained the distribution at the coarser scale.

Main conclusions

Although some techniques may apparently deliver good distribution maps for species with scarce and biased data, they need to be taken with caution. When good independent data at a coarser scale are available, cross‐scale validation can help to produce more reliable and robust maps. When no independent data are available for validation, however, new data gathering field surveys may be the only option if reliable fine‐scale resolution maps are needed.     

基于最大熵模型的不同尺度物种分布概率优化热点分析: 以红色木莲为例

单一空间尺度构建的最大熵(maximum entropy, MaxEnt)模型是否具有代表性, 是MaxEnt模型应用与发展中面临的重要问题。本研究基于有效的地理分布位点数据, 利用最小凸多边形法(the minimum convex polygon method)在三江并流、云南省及全国3个空间尺度下分别识别了红色木莲(Manglietia insignis)的建模区域, 并进一步建立MaxEnt模型: 使用ROC曲线分析法与遗漏率(omission rate, OR)检验评估MaxEnt模型预测精度; 基于ArcGIS分析分布概率及其热点区域的分布趋势, 并通过分区统计工具Zonal识别潜在适宜分布区域的质心位置; 采用刀切法检验环境因子贡献率。结果表明: (1)不同尺度下红色木莲的MaxEnt模型都有良好的预测效果, 三江并流、云南省及全国尺度下的AUC值分别为0.936、0.887和0.930, OR值分别为0.18、0.15和0.20; (2)各尺度红色木莲的适生区格局呈现一致性分布趋势, 集中在独龙江、怒江和澜沧江3个流域; (3) 3个空间尺度下红色木莲的地理分布受不同环境因子影响, 存在着尺度依赖效应。由此可见, 红色木莲在不同空间尺度下的预测模型有着稳定的性能表现与良好的预测效果。此外, 我们建议在野外实地调查与野生生物资源保护中加强对普通物种的关注, 在预测物种地理分布的研究中将MaxEnt模型与热点分析结合使用。     

Sensitivity of plant functional types to climate change: classification tree analysis of a simulation model

Question: The majority of studies investigating the impact of climate change on local plant communities ignores changes in regional processes, such as immigration from the regional seed pool. Here we explore: (i) the potential impact of climate change on composition of the regional seed pool, (ii) the influence of changes in climate and in the regional seed pool on local community structure, and (iii) the combinations of life history traits, i.e. plant functional types (PFTs), that are most affected by environmental changes. Location: Fire‐prone, Mediterranean‐type shrublands in southwestern Australia. Methods: Spatially explicit simulation experiments were conducted at the population level under different rainfall and fire regime scenarios to determine the effect of environmental change on the regional seed pool for 38 PFTs. The effects of environmental and seed immigration changes on local community dynamics were then derived from community‐level experiments. Classification tree analyses were used to investigate PFT‐specific vulnerabilities to climate change. Results: The classification tree analyses revealed that responses of PFTs to climate change are determined by specific trait characteristics. PFT‐specific seed production and community patterns responded in a complex manner to climate change. For example, an increase in annual rainfall caused an increase in numbers of dispersed seeds for some PFTs, but decreased PFT diversity in the community. Conversely, a simulated decrease in rainfall reduced the number of dispersed seeds and diversity of PFTs. Conclusions: PFT interactions and regional processes must be considered when assessing how local community structure will be affected by environmental change.     

Shallow environmental gradients put inland species at risk: Insights and implications from predicting future distributions of Eucalyptus species in South Western Australia

As climate changes, tree decline in Mediterranean‐type ecosystems is increasing worldwide, often due to decreased effective precipitation and increased drought and heat stress, and has recently been observed in coastal species of the iconic Eucalyptus (Myrtaceae) genus in the biodiversity hotspot of south‐west Western Australia. To investigate how this drought‐related decline is likely to continue in the future, we used species distribution modelling techniques to generate broad‐scale predictions of future distribution patterns under three distinct projected climate change scenarios. In a moderate climate change scenario, suitable habitat for all species was predicted to decrease by, on average, 73% by the year 2100, with most receding into southern areas of their current distribution. Although the most severe Eucalyptus declines in south‐west Western Australia have been observed in near‐coastal regions, our predictions suggest that inland species are at greater risk from climate change, with six inland species predicted to lose 95% of their suitable habitat in a moderate change scenario. This is due to the shallow environmental gradients of inland regions causing larger spatial shifts of environmental envelopes, which is likely to be relevant in many regions of the world. The knowledge gained suggests that future research and conservation efforts in south‐west Western Australia and elsewhere should avoid focussing disproportionately on coastal regions for reasons of convenience and proximity to population centres, and properly address the inland region where the biggest future impacts may occur.     

Hierarchical models facilitate spatial analysis of large data sets: a case study on invasive plant species in the northeastern United States

Many critical ecological issues require the analysis of large spatial point data sets – for example, modelling species distributions, abundance and spread from survey data. But modelling spatial relationships, especially in large point data sets, presents major computational challenges. We use a novel Bayesian hierarchical statistical approach, 'spatial predictive process' modelling, to predict the distribution of a major invasive plant species, Celastrus orbiculatus , in the northeastern USA. The model runs orders of magnitude faster than traditional geostatistical models on a large data set of c . 4000 points, and performs better than generalized linear models, generalized additive models and geographically weighted regression in cross-validation. We also use this approach to model simultaneously the distributions of a set of four major invasive species in a spatially explicit multivariate model. This multispecies analysis demonstrates that some pairs of species exhibit negative residual spatial covariation, suggesting potential competitive interaction or divergent responses to unmeasured factors.     

Patterns of woody plant species richness in the Iberian Peninsula: environmental range and spatial scale

Aim Climate‐based models often explain most of the variation in species richness along broad‐scale geographical gradients. We aim to: (1) test predictions of woody plant species richness on a regional spatial extent deduced from macro‐scale models based on water–energy dynamics; (2) test if the length of the climate gradients will determine whether the relationship with woody species richness is monotonic or unimodal; and (3) evaluate the explanatory power of a previously proposed ‘water–energy’ model and regional models at two grain sizes. Location The Iberian Peninsula. Methods We estimated woody plant species richness on grid maps with c. 2500 and 22,500 km² cell size, using geocoded data for the individual species. Generalized additive models were used to explore the relationships between richness and climatic, topographical and substrate variables. Ordinary least squares regression was used to compare regional and more general water–energy models in relation to grain size. Variation partitioning by partial regression was applied to find how much of the variation in richness was related to spatial variables, explanatory variables and the overlap between these two. Results Water–energy dynamics generate important underlying gradients that determine the woody species richness even over a short spatial extent. The relationships between richness and the energy variables were linear to curvilinear, whereas those with precipitation were nonlinear and non‐monotonic. Only a small fraction of the spatially structured variation in woody species richness cannot be accounted for by the fitted variables related to climate, substrate and topography. The regional models accounted for higher variation in species richness than the water–energy models, although the water–energy model including topography performed well at the larger grain size. Elevation range was the most important predictor at all scales, probably because it corrects for ‘climatic error’ due to the unrealistic assumption that mean climate values are evenly distributed in the large grid cells. Minimum monthly potential evapotranspiration was the best climatic predictor at the larger grain size, but actual evapotranspiration was best at the smaller grain size. Energy variables were more important than precipitation individually. Precipitation was not a significant variable at the larger grain size when examined on its own, but was highly significant when an interaction term between itself and substrate was included in the model. Main conclusions The significance of range in elevation is probably because it corresponds to several aspects that may influence species diversity, such as climatic variability within grid cells, enhanced surface area, and location for refugia. The relative explanatory power of energy and water variables was high, and was influenced by the length of the climate gradient, substrate and grain size of the analysis. Energy appeared to have more influence than precipitation, but water availability is also determined by energy, substrate and topographic relief.     

Disregarding the edaphic dimension in species distribution models leads to the omission of crucial spatial information under climate change: the case of Quercus pubescens in France

Species distribution modelling is an easy, persuasive and useful tool for anticipating species distribution shifts under global change. Numerous studies have used only climate variables to predict future potential species range shifts and have omitted environmental factors important for determining species distribution. Here, we assessed the importance of the edaphic dimension in the niche‐space definition of Quercus pubescens and in future spatial projections under global change over the metropolitan French forest territory. We fitted two species distribution models (SDM) based on presence/absence data (111 013 plots), one calibrated from climate variables only (mean temperature of January and climatic water balance of July) and the other one from both climate and edaphic (soil pH inferred from plants) variables. Future predictions were conducted under two climate scenarios (PCM B2 and HadCM3 A2) and based on 100 simulations using a cellular automaton that accounted for seed dispersal distance, landscape barriers preventing migration and unsuitable land cover. Adding the edaphic dimension to the climate‐only SDM substantially improved the niche‐space definition of Q. pubescens, highlighting an increase in species tolerance in confronting climate constraints as the soil pH increased. Future predictions over the 21st century showed that disregarding the edaphic dimension in SDM led to an overestimation of the potential distribution area, an underestimation of the spatial fragmentation of this area, and prevented the identification of local refugia, leading to an underestimation of the northward shift capacity of Q. pubescens and its persistence in its current distribution area. Spatial discrepancies between climate‐only and climate‐plus‐edaphic models are strengthened when seed dispersal and forest fragmentation are accounted for in predicting a future species distribution area. These discrepancies highlight some imprecision in spatial predictions of potential distribution area of species under climate change scenarios and possibly wrong conclusions for conservation and management perspectives when climate‐only models are used.