MaxEnt模型参数设置对其所模拟物种地理分布和生态位的影响——以茶翅蝽为例

【目的】生态位模型被广泛应用于入侵生物学和保护生物学研究,现有建模工具中,MaxEnt是最流行和运用最广泛的生态位模型。然而最近研究表明,基于MaxEnt模型的默认参数构建模型时,模型倾向于过度拟合,并非一定为最佳模型,尤其是在处理一些分布点较少的物种。【方法】以茶翅蝽为例,通过设置不同的特征参数、调控倍频以及背景拟不存在点数分别构建茶翅蝽的本土模型,然后将其转入入侵地来验证和比较模型,通过检测模型预测的物种对环境因子的响应曲线、潜在分布在生态空间中的生态位映射以及潜在分布的空间差异性,探讨3种参数设置对MaxEnt模型模拟物种分布和生态位的影响。【结果】在茶翅蝽的案例分析中,特征参数的设置对MaxEnt模型所模拟的潜在分布和生态位的影响最大,调控倍频的影响次之,背景拟不存在点数的影响最小。与其他特征相比,基于特征H和T的模型其响应曲线较为曲折;随着调控倍频的增加,响应曲线变得圆滑。【结论】在构建MaxEnt模型时,需要从生态空间中考虑物种的生态需求,分析模型参数对预测物种分布和生态位可能造成的影响。     

ROC曲线分析在评价入侵物种分布模型中的应用

生态位模型（ecological niche models,ENMs）已广泛应用于物种潜在分布区预测,ENMs的应用也为外来入侵物种的风险分析提供了重要的定量化分析工具,但如何评价不同模型之间的预测效果成了当今研究的热点问题。本文介绍了受试者工作特征（ROC）曲线分析在评价不同生态位模型预测效果中的应用原理和分析方法,并以一种植物病原线虫-相似穿孔线虫（Radopholus similis）为例,应用ROC曲线分析法对其5种模型（BIOCLIM,CLIMEX,DOMAIN,GARP,MAXENT）的预测结果进行了比较分析。5种模型的ROC曲线下面积AUC（Area Under Curve）值分别为0.810,0.758,0.921,0.903和0.950,以MAXENT模型的AUC值最大,表明其预测效果最好;方差分析结果表明,除GARP与DOMAIN模型之间AUC值差异不显著外,其余各模型之间差异显著。     

生态位模型的基本原理及其在生物多样性保护中的应用

生态位模型是利用物种已知的分布数据和相关环境变量,根据一定的算法来推算物种的生态需求,然后将运算结果投射至不同的空间和时间中来预测物种的实际分布和潜在分布.近年来,该类模型被越来越多地应用在入侵生物学、保护生物学、全球气候变化对物种分布影响以及传染病空间传播的研究中.然而,由于生态位模型的理论基础未被深入理解,导致得出入侵物种生态位迁移等不符合实际的结论.作者从生态位与物种分布的关系、生态位模型构建的基本原理以及生态位模型和生态位的关系等方面探讨了生态位模型的理论基础.非生物的气候因素、物种间的相互作用和物种的迁移能力是影响物种分布的3个主要因素,它们在不同的空间尺度下作用于物种的分布.生态位模型是利用物种分布点所关联的环境变量来模拟物种的分布,这些分布点本身关联着该物种和其他物种间的相互作用,因此生态位模型所模拟的是现实生态位(realized niche)或潜在生态位(potential niche),而不是基础生态位(fundamental niche).Grinnell生态位和Elton生态位均在生态位模型中得到反映,这取决于环境变量类型的选择、所采用环境变量的分辨率以及物种自身的迁移能力.生态位模型在生物多样性保护中的应用主要包括物种的生态需求分析、未知物种或种群的探索和发现、自然保护区的选择和设计、物种入侵风险评价、气候变化对物种分布的影响、近缘物种生态位保守性及基于生态位分化的物种界定等方面.     

提高生态位模型转移能力来模拟入侵物种 的潜在分布

生态位模型利用物种分布点所关联的环境变量去推算物种的生态需求, 模拟物种的分布。在模拟入侵物种分布时, 经典生态位模型包括模型构建于物种本土分布地, 然后将其转移并投射至另一地理区域, 来模拟入侵物种的潜在分布。然而在模型运用时, 出现了模型的转移能力较低、模拟的结果与物种的实际分布不相符的情况, 由此得出了生态位漂移等不恰当的结论。提高生态位模型的转移能力, 可以准确地模拟入侵物种的潜在分布, 为入侵种的风险评估提供参考。作者以入侵种茶翅蝽(Halyomorpha halys)和互花米草(Spartina alterniflora)为例, 从模型的构建材料(即物种分布点和环境变量)入手, 全面阐述提高模型转移能力的策略。在构建模型之前, 需要充分了解入侵物种的生物学特性、种群平衡状态、本土地理分布范围及物种的生物历史地理等方面的知识。在模型构建环节上, 物种分布点不仅要充分覆盖物种的地理分布和生态空间的范围, 同时要降低物种采样点偏差; 环境变量的选择要充分考虑其对物种分布的限制作用、各环境变量之间的空间相关性, 以及不同地理种群间生态空间是否一致, 同时要降低环境变量的空间维度; 模型构建区域要真实地反映物种的地理分布范围, 并考虑种群的平衡状态。作者认为, 在生态位保守的前提下, 如果模型是构建在一个合理方案的基础上, 生态位模型的转移能力是可以保证的, 在以模型转移能力较低的现象来阐述生态位分化时需要引起注意。     

北美刺龙葵在中国的适生区预测

【背景】北美刺龙葵是一种全球广泛分布的恶性杂草,已被列入我国进境检疫性有害生物名单。近年来北美刺龙葵不断随进口货物传入我国,明确其传入途径和适生区对控制其入侵具有重要意义。【方法】采用GIS、空间统计学、Maxent生态位模型等方法分析了北美刺龙葵的传入途径与潜在分布区,并通过ROC分析法对模型进行检验。【结果】跨区域农产品贸易是北美刺龙葵全球扩散的驱动力与传入我国的主要途径。生态模型预测结果表明,北美刺龙葵在我国具有广阔的适生区,除黑龙江、吉林、内蒙古、青海、甘肃、西藏、四川西北部以外的区域都是其在我国的适生区,其中高风险区主要集中在东部和南部沿海、西南边境和新疆的部分地区。AUC值为0.789,表明本研究建立的Maxent模型的预测能力较强,能够很好地拟合物种已知分布的环境生态位。【结论与意义】北美刺龙葵在我国的传入风险极高。基于北美刺龙葵在我国的主要传入途径与潜在扩散媒介的时空分布,划定了重点监测的区域,建议对适生区内极易传入的高风险区如港口、机场、物流中转站、加工厂等开展早期监测预警,以预防其再次入侵与进一步扩散蔓延。     

入侵害虫蔗扁蛾在我国的潜在分布区

【目的】蔗扁蛾是危害巴西木、甘蔗等园林植物和经济作物的重要入侵害虫。该虫于20世纪90年代初在我国被发现,现已分布在海南、广东和上海等19个省市,并有迅速扩散蔓延的趋势。对入侵害虫的潜在分布区进行预测,可为实施害虫监测和管理提供参考。【方法】根据蔗扁蛾已有分布点的记录,分别在4种地理区域构建Maxent生态位模型,并采用加权平均值法对其进行整合,进而分析蔗扁蛾在我国的潜在分布区。【结果】基于4种地理区域构建的Maxent模型对我国南部地区的预测结果基本一致,4种模型的预测差异主要在新疆北部和西南部、黑龙江东部和西部、吉林西部、山西中部等地区。整合模型显示,华东和华南地区以及东部沿海地区具有较大的分布可能性。【结论】蔗扁蛾在我国尤其是南方具有较大的潜在分布空间。这些地区应警惕蔗扁蛾的入侵,同时采取应对措施防止其进一步扩散。     

应用生态位模型研究外来入侵物种生态位漂移

由于基础生态位和实际生态位的改变,外来入侵物种在入侵地成功定殖、扩散后常会发生生态位漂移,而物种生态位漂移往往很难直接证明。生态位模型在假设入侵物种的生态位需求保守的前提下,以物种在其原产地的生态位需求为基础,预测其在入侵地的潜在分布,通过比较预测分布与实际分布的差异可以从一定程度上得到外来入侵物种的生态位是否发生漂移的间接证据。以我国入侵杂草胜红蓟在原产地的生态位需求为基础,应用生态位模型预测其在其他地区的潜在分布。研究结果表明,生态位模型可以很好地预测胜红蓟在亚太平洋地区和非洲地区的分布,但在我国,其预测分布与实际分布存在较大差别。胜红蓟在我国预测分布主要为云南、海南、台湾部分地区,而胜红蓟入侵我国后现已广泛分布于长江以南地区,其实际分布比预测分布广泛得多,由此推测胜红蓟在入侵我国后其生态位已经产生了漂移。     

基于Maxent生态位模型预测球果角胫象在云南的潜在地理分布

掌握球果角胫象Shirahoshizo conifera Chao在云南潜在分布区,对了解其危害和防控具有重要的意义。本文利用球果角胫象的分布点数据和环境因子数据,通过Maxent生态位模型预测其在云南的潜在分布区。结果表明：球果角胫象的潜在分布区主要集中滇中地区;在ArcGIS中进行显示与风险等级划分,按栅格数值的大小将球果角胫象的适生范围分为4级,分级标准为：高适生区、中适生区、低适生区、非适生区,经ROC曲线分析法验证,Maxent生态位模型的AUC值为0998,表明预测获得了较好的效果。     

基于生态位模型预测天麻全球潜在适生区

目前对药用植物天麻(Gastrodia elata)的全球潜在适生区研究较少,基于多个生态位模型预测天麻在全球范围内的潜在适生区,对天麻人工引种栽培及其产业发展具有重要意义。该文收集220个天麻全球分布点和19个生态因子数据,最终筛选出8个环境变量参与模型训练,基于3个生态位模型(BIOCLIM、DOMAIN和MAXENT)预测天麻全球潜在适生区,并采用受试者工作特征曲线ROC和Kappa统计量分析比较不同模型的预测效果。结果表明:3个模型的预测结果基本一致,天麻全球潜在适生区主要分布在20°–50°N的亚洲地区,其中中国、日本和韩国是集中分布地,此外,印度、尼泊尔以及欧洲地中海附近有少量适生区。其中最适宜区主要分布在:中国四川盆地附近的省区以及中东部;韩国中东部的忠清北道、庆尚北道和庆尚南道;日本本州岛、九州岛以及四国岛,因此中国、日本和韩国是天麻的主要产区。3个模型的受试者工作特征曲线下面积(AUC值)平均值均达到0.9以上,Kappa平均值均达到0.65以上,能较好地预测天麻的潜在适生区,其中MAXENT模型的精度较高,其次是DOMAIN和BIOCLIM模型。     

基于MAXENT模型和Arc GIS预测蜡梅适生域在中国的潜在分布

蜡梅(Chimonanthus praecox)是我国二级濒危珍稀植物,是重要的冬季传统观花植物。利用已报道的246个分布点和worldclim中提取的19个气候因子,基于最大熵(Maxent)模型和地理信息系统(Arc Gis)对蜡梅在中国的潜在适生区分布进行预测分析,采用受试者工作特征(ROC)曲线对预测结果进行检验和评价。结果表明蜡梅的潜在适生范围相对集中,主要集中在西南的四川盆地、华中、华东及华北的中南部地区,其他地区则适应性较低。温度是影响蜡梅分布的决定性因子,其中,当最冷季度平均温度接近0℃,等温性范围为0—10℃,降雨量变异系数约为45时,蜡梅的分布概率最大。与原分布区相比较,蜡梅的适生区范围正向中国东部地区和北部地区迁移。ROC曲线检验评价结果表明,Maxent模型的ROC曲线分析法的面积(AUC)值为0.986,预测结果达到了极高精度。     

Ecological niche modeling re‐examined: A case study with the Darwin's fox

Many previous studies have attempted to assess ecological niche modeling performance using receiver operating characteristic (ROC) approaches, even though diverse problems with this metric have been pointed out in the literature. We explored different evaluation metrics based on independent testing data using the Darwin's Fox (Lycalopex fulvipes) as a detailed case in point. Six ecological niche models (ENMs; generalized linear models, boosted regression trees, Maxent, GARP, multivariable kernel density estimation, and NicheA) were explored and tested using six evaluation metrics (partial ROC, Akaike information criterion, omission rate, cumulative binomial probability), including two novel metrics to quantify model extrapolation versus interpolation (E‐space index I) and extent of extrapolation versus Jaccard similarity (E‐space index II). Different ENMs showed diverse and mixed performance, depending on the evaluation metric used. Because ENMs performed differently according to the evaluation metric employed, model selection should be based on the data available, assumptions necessary, and the particular research question. The typical ROC AUC evaluation approach should be discontinued when only presence data are available, and evaluations in environmental dimensions should be adopted as part of the toolkit of ENM researchers. Our results suggest that selecting Maxent ENM based solely on previous reports of its performance is a questionable practice. Instead, model comparisons, including diverse algorithms and parameterizations, should be the sine qua non for every study using ecological niche modeling. ENM evaluations should be developed using metrics that assess desired model characteristics instead of single measurement of fit between model and data. The metrics proposed herein that assess model performance in environmental space (i.e., E‐space indices I and II) may complement current methods for ENM evaluation.     

Insights into the area under the receiver operating characteristic curve (AUC) as a discrimination measure in species distribution modelling

Aim The area under the receiver operating characteristic (ROC) curve (AUC) is a widely used statistic for assessing the discriminatory capacity of species distribution models. Here, I used simulated data to examine the interdependence of the AUC and classical discrimination measures (sensitivity and specificity) derived for the application of a threshold. I shall further exemplify with simulated data the implications of using the AUC to evaluate potential versus realized distribution models. Innovation After applying the threshold that makes sensitivity and specificity equal, a strong relationship between the AUC and these two measures was found. This result is corroborated with real data. On the other hand, the AUC penalizes the models that estimate potential distributions (the regions where the species could survive and reproduce due to the existence of suitable environmental conditions), and favours those that estimate realized distributions (the regions where the species actually lives). Main conclusions Firstly, the independence of the AUC from the threshold selection may be irrelevant in practice. This result also emphasizes the fact that the AUC assumes nothing about the relative costs of errors of omission and commission. However, in most real situations this premise may not be optimal. Measures derived from a contingency table for different cost ratio scenarios, together with the ROC curve, may be more informative than reporting just a single AUC value. Secondly, the AUC is only truly informative when there are true instances of absence available and the objective is the estimation of the realized distribution. When the potential distribution is the goal of the research, the AUC is not an appropriate performance measure because the weight of commission errors is much lower than that of omission errors.     

Changes in the realized niche of the invasive succulent CAM plant Furcraea foetida

Furcraea foetida (Asparagaceae) is a native plant of Central America and northern South America but there is no information about its country of origin. The species was introduced into Brazil and is now considered invasive, particularly in coastal ecosystems. To date, nothing is known about the environmental factors that constrain its distribution and there is only inconclusive information about its location of origin. We used reciprocal distribution models (RDM) to assess invasion risk of F. foetida across Brazil and to identify source regions in its native range. We also tested the niche conservatism hypothesis using Principal Components Analyses and statistical tests of niche equivalency and similarity between its native and invaded ranges. For RDM analysis, we built two models using maximum entropy, one using records in the native range to predict the invaded distribution (forward‐Ecological Niche Model or forward‐ENM) and one using records in the invaded range to predict the native distribution (reverse‐ENM). Forward‐ENM indicated invasion risk in the Cerrado region and the innermost region of the Atlantic Forest, however, failed to predict the current occurrence in southern Brazil. Reverse‐ENM supported an existing hypothesis that F. foetida originated in the Orinoco river basin, Amazon basin and Caribbean islands. Prediction errors in the RDM and multivariate analysis indicated that the species expanded its realized niche in Brazil. The niche similarity test further suggested that the niche differences are because of differences in habitat availability between the two ranges, not because of evolutionary changes. We hypothesize that physiological pre‐adaptation (especially, the crassulacean acid metabolism), human‐driven propagule pressure and high competitive ability are the main factors determining the current spatial distribution of the species in Brazil. Our study highlights the need to include F. foetida in plant invasion monitoring programs, especially in priority conservation areas where the species has still not been introduced.     

Pooling biospecimens and limits of detection: effects on ROC curve analysis

Frequently, epidemiological studies deal with two restrictions in the evaluation of biomarkers: cost and instrument sensitivity. Costs can hamper the evaluation of the effectiveness of new biomarkers. In addition, many assays are affected by a limit of detection (LOD), depending on the instrument sensitivity. Two common strategies used to cut costs include taking a random sample of the available samples and pooling biospecimens. We compare the two sampling strategies when an LOD effect exists. These strategies are compared by examining the efficiency of receiver operating characteristic (ROC) curve analysis, specifically the estimation of the area under the ROC curve (AUC) for normally distributed markers. We propose and examine a method to estimate AUC when dealing with data from pooled and unpooled samples where an LOD is in effect. In conclusion, pooling is the most efficient cost-cutting strategy when the LOD affects less than 50% of the data. However, when much more than 50% of the data are affected, utilization of the pooling design is not recommended.     

Global distribution patterns of highly pathogenic H5N1 avian influenza: Environmental vs. socioeconomic factors

In this report, we quantitatively analyzed the essential ecological factors that were strongly correlated with the global outbreak of highly pathogenic H5N1 avian influenza. The ecological niche modeling (ENM) was used to reveal the potential outbreak hotspots of H5N1. A two-step modeling procedure has been proposed: we first used BioClim model to obtain the coarse suitable areas of H5N1, and then those suitable areas with very high probabilities were retained as the inputs of multiple-variable autologistic regression analysis (MAR) for model refinement. MAR was implemented taking spatial autocorrelation into account. The final performance of ENM was evaluated using the areas under the curve (AUC) of receiver-operating characteristic. In addition, principal component analysis (PCA) was employed to reveal the most important variables and relevant ecological gradients of H5N1 outbreak. Niche visualization was used to identify potential spreading trend of H5N1 along important ecological gradients. For the first time, we combined socioeconomic and environmental variables as joint predictors in developing ecological niche modeling. Environmental variables represented the natural element related to H5N1 outbreak, whereas socioeconomic ones represented the anthropogenic element. Our results indicated that: (1) the high-risk hotspots are mainly located in temperate zones (indicated by ENM)—correspondingly, we argued that the “ecoregions hypothesis” was reasonable to some extent; (2) evaporation, humidity, human population density, livestock population density were the first four important factors (in descending order) that were associated with the H5N1 global outbreak (indicated by PCA); (3) influenza had a tendency to expand into areas with low evaporation (indicated by niche visualization). In conclusion, our study substantiates that both the environmental and socioeconomic variables jointly determined the global spreading trend of H5N1, but environmental variables played a more important role. Consequently, our study is consistent with the assumption that the natural element is more important than the anthropogenic element as the underlying ecological mechanisms explaining global H5N1 transmission.     

Maxent模型复杂度对物种潜在分布区预测的影响

生态位模型在入侵生物学和保护生物学中具有广泛的应用, 其中Maxent模型最为流行, 被越来越多地应用在预测物种的现实分布和潜在分布的研究中。在Maxent模型中, 多数研究者采用默认参数来构建模型, 这些默认参数源自早期对266个物种的测试, 以预测物种的现实分布为目的。近期研究发现, Maxent模型采用复杂机械学习算法, 对采样偏差敏感, 易产生过度拟合, 模型转移能力仅在低阈值情况下较好。基于默认参数的Maxent模型不仅预测结果不可靠, 而且有时很难解释。在本研究中, 作者以入侵害虫茶翅蝽(Halyomorpha halys)为例, 采用经典模型构建方案(即构建本土模型然后将其转移至入侵地来评估), 利用ENMeval数据包来调整本土Maxent模型调控倍频和特征组合参数, 分析各种参数条件下模型的复杂度, 然后选取最低复杂度的模型参数(即为最优模型), 综合比较默认参数和调整参数后Maxent模型的响应曲线和预测结果, 探讨Maxent模型复杂度对预测结果的影响及Maxent模型构建时所需注意事项, 以期对物种潜在分布进行合理的预测, 促进Maxent模型在我国的合理运用和发展。作者认为, 环境变量的选择至关重要, 需要综合分析其对所模拟物种分布的限制作用和环境变量之间的空间相关性。构建Maxent模型前需对物种分布采样偏差及模型的构建区域进行合理地判断, 模型构建时需要比较不同参数下模型的预测结果和响应曲线, 选取复杂度较低的模型参数来最终建模。在茶翅蝽的分析中, Maxent模型的默认参数和最优模型参数不同, 与Maxent模型默认参数相比, 采用调整参数后所构建的模型预测效果较好, 响应曲线较为平滑, 模型转移能力较高, 能够较为合理反映物种对环境因子的响应和准确地模拟该物种的潜在分布。     

An evaluation of transferability of ecological niche models

Ecological niche modeling (ENM) is used widely to study species’ geographic distributions. ENM applications frequently involve transferring models calibrated with environmental data from one region to other regions or times that may include novel environmental conditions. When novel conditions are present, transferability implies extrapolation, whereas, in absence of such conditions, transferability is an interpolation step only. We evaluated transferability of models produced using 11 ENM algorithms from the perspective of interpolation and extrapolation in a virtual species framework. We defined fundamental niches and potential distributions of 16 virtual species distributed across Eurasia. To simulate real situations of incomplete understanding of species’ distribution or existing fundamental niche (environmental conditions suitable for the species contained in the study area; N* _F ), we divided Eurasia into six regions and used 1–5 regions for model calibration and the rest for model evaluation. The models produced with the 11 ENM algorithms were evaluated in environmental space, to complement the traditional geographic evaluation of models. None of the algorithms accurately estimated the existing fundamental niche (N* _F ) given one region in calibration, and model evaluation scores decreased as the novelty of the environments in the evaluation regions increased. Thus, we recommend quantifying environmental similarity between calibration and transfer regions prior to model transfer, providing an avenue for assessing uncertainty of model transferability. Different algorithms had different sensitivity to completeness of knowledge of N* _F , with implications for algorithm selection. If the goal is to reconstruct fundamental niches, users should choose algorithms with limited extrapolation when N* _F is well known, or choose algorithms with increased extrapolation when N* _F is poorly known. Our assessment can inform applications of ecological niche modeling transference to anticipate species invasions into novel areas, disease emergence in new regions, and forecasts of species distributions under future climate conditions.