Impact of ecological redundancy on the performance of machine learning classifiers in vegetation mapping

Vegetation maps are models of the real vegetation patterns and are considered important tools in conservation and management planning. Maps created through traditional methods can be expensive and time‐consuming, thus, new more efficient approaches are needed. The prediction of vegetation patterns using machine learning shows promise, but many factors may impact on its performance. One important factor is the nature of the vegetation–environment relationship assessed and ecological redundancy. We used two datasets with known ecological redundancy levels (strength of the vegetation–environment relationship) to evaluate the performance of four machine learning (ML) classifiers (classification trees, random forests, support vector machines, and nearest neighbor). These models used climatic and soil variables as environmental predictors with pretreatment of the datasets (principal component analysis and feature selection) and involved three spatial scales. We show that the ML classifiers produced more reliable results in regions where the vegetation–environment relationship is stronger as opposed to regions characterized by redundant vegetation patterns. The pretreatment of datasets and reduction in prediction scale had a substantial influence on the predictive performance of the classifiers. The use of ML classifiers to create potential vegetation maps shows promise as a more efficient way of vegetation modeling. The difference in performance between areas with poorly versus well‐structured vegetation–environment relationships shows that some level of understanding of the ecology of the target region is required prior to their application. Even in areas with poorly structured vegetation–environment relationships, it is possible to improve classifier performance by either pretreating the dataset or reducing the spatial scale of the predictions.     

近20年海南岛植被生态质量变化特征分析

海南是国家生态文明试验区,是我国生态环境最好的省份之一.基于2000—2019年MODIS—NDVI海南岛植被覆盖度、植被净初级生产力、植被生态质量指数及地面气象观测资料,分析了海南岛植被生态质量变化情况以及气象条件对植被生态质量的影响.结果表明:2000年以来,海南岛植被生态质量指数呈增大趋势,植被生态质量整体向好....     

Are potential natural vegetation maps a meaningful alternative to neutral landscape models?

Abstract. In this paper, we present a short overview of neutral landscape models traditionally adopted in the landscape ecological literature to differentiate landscape patterns that are the result of simple random processes from patterns that are generated from more complex ecological processes. Then, we present another family of models based on Tuxen’ s definition of potential natural vegetation that play an important role, especially in Europe, for landscape planning and management. While neutral landscape models by their very nature do not take into account vegetation dynamics, nor abiotic constraints to vegetation distribution, the concept of potential natural vegetation includes the effects of vegetation dynamics in a spatially explicit manner. Therefore, we believe that distribution maps of potential natural vegetation may represent an ecological meaningful alternative to neutral landscape models for evaluating the effects of landscape structure on ecological processes.     

Visual ecology: exploring the relationships between ecological quality and aesthetic preference

High ecological quality and an attractive environment are usually two primary goals of landscape design. Understanding the relationships between ecological quality and visual aesthetic preference is needed to reach the two goals simultaneously. However, previous work has not provided definite answers regards these relationships. The current study took place in the riparian zone of Xuzhou, China, to check ecological quality using the qualitat del bosc de ribera (or riparian forest quality) index, and visual aesthetic preference using pictures as surrogates, which were assessed by undergraduate students. Regression analysis revealed a U-shaped distribution of ecological quality in relation to visual aesthetic preference, which indicated that, with higher ecological quality, the visual aesthetic preference could be higher or lower, and a lower ecological quality implied a medium visual aesthetic preference. In addition, enhancing the cover quality (introducing more native plants and reducing man-made elements) of riparian vegetation was an efficient way to align the ecological goal with a visually pleasing landscape.     

Novelty predictors for shrub (and climbers) ecological niche modeling,based on their successional stage

The identification of species' environmental predictors constitute a key challenge for decision making, especially when using ecological niche modeling based on these drivers and when presence points are limited. More specifically, shrub species are affected by ecosystem dynamics, and appear in degraded formations, in dense mid-stage vegetation formations, or under late climax-forest canopy. In this study, we tested novelty predictors to understand the drivers that affect the selected species distribution in the Mediterranean biome, targeting different vegetation successional stages, and further improve ecological models' performance, when presence points are limited. Land surface temperature (LST) in association with temperature related predictors, allowed differentiating between species thriving in the understory of the forest canopy, from those that are co-dominant with dense vegetation cover and a third group/species, thriving in degraded vegetation. In addition, the Normalized Difference Vegetation Level Index (NDVI) played a key role in the models for species growing in highly degraded ecological niches such as Spartium junceum, Calicotome villosa, but also forest-fringe vegetation like the climber Hedera helix. Our study highlights the importance of integrating remote sensed predictors, combined with appropriate climate drivers, when using ecological niche modeling.     

河流生态廊道提取方法研究及其应用思路

随着交叉学科的发展和技术手段的进步，将河流及其周边生态环境视为整体的理念和研究逐渐受到重视，然而国内对河流生态廊道的研究较少。总结了河流生态廊道的基础理论和提取方法。河流生态廊道可以依据区域规划、通过实地观测等手段直接提取，也可以采用“源地识别-阻力面构建-廊道提取”的框架间接提取。对直接和间接提取方法进行了详细的总结并分析了他们的优势和劣势，发现间接提取方法更具实践意义，且应用的流域面积更大。根据间接提取方式的特点，提出白洋淀-大清河流域河流生态廊道提取的思路。结合区域流域协同治理理念，逐步形成以白洋淀为中心、向周边辐射的河流生态廊道，实现流域水质提升、生物多样性稳定等目标，全面提升生态环境质量。在河流生态廊道提取科学性提升和提取过程中廊道宽度优化等薄弱环节提出未来研究方向。     

Assessing the quality of native vegetation: The 'habitat hectares' approach

Summary Assessments of the 'quality', condition or status of stands of native vegetation or habitat are now commonplace and are often an essential component of ecological studies and planning processes. Even when soundly based upon ecological principles, these assessments are usually highly subjective and involve implicit value judgements. The present paper describes a novel approach to vegetation or habitat quality assessment (habitat hectares approach) that can be used in almost all types of terrestrial vegetation. It is based on explicit comparisons between existing vegetation features and those of 'benchmarks' representing the average characteristics of mature stands of native vegetation of the same community type in a 'natural' or 'undisturbed' condition. Components of the index incorporate vegetation physiognomy and critical aspects of viability (e.g. degree of regeneration, impact of weeds) and spatial considerations (e.g. area, distribution and connectivity of remnant vegetation in the broader landscape). The approach has been developed to assist in making more objective and explicit decisions about where scarce conservation resources are allocated. Although the approach does not require an intimate botanical knowledge, it is believed to be ecologically valid and useful in many contexts. Importantly, the index does not provide a definitive statement on conservation status nor habitat suitability for individual species. It purposefully takes a 'broad-brush' approach and is primarily intended for use by people involved with making environmentally sensitive planning and management decisions, but may be useful within environmental research programmes. The 'habitat hectares' approach is subject to further research and ongoing refinement and constructive feedback is sought from practitioners.     

The phylogenetic interpretation of biological surveys

The ecological attributes of two species may be similar through convergent evolution or common ancestry. The extent of similarity by descent can be evaluated by comparing them with their most closely‐related outgroup in a given phylogeny. I describe a method of nested sister‐group analysis for estimating ecological similarity based on landscape features or on co‐distribution. The phylogeny is dissected into triplets, each comprising two sister taxa and their outgroup. For a triplet at any phylogenetic level, the similarity of sister groups with respect to some given character can be compared with their joint similarity to the outgroup to give a single test of similarity by descent. Each comparison is independent, and the full set of triplets provides a complete accounting of phylogenetic variation at all levels. This procedure was applied to 188 moderately abundant species of dicots in two independent surveys from adjoining districts of midland England, supplemented by physical surveys of landscape attributes obtained from digitized maps of the same districts. The co‐distribution of sister species was consistently more positive than the co‐distribution of random species pairs, demonstrating the existence of a phylogenetic signal at some level. When sister species are compared with their most closely‐related outgroup, however, neither landscape attributes nor co‐distribution showed any overall similarity arising from common ancestry, in the sense that ecological attributes are not generally conserved after lineage splitting. Instead, the distribution of similarity is strikingly similar to random data. The lack of ecological similarity between closely‐related groups was attributed to rapid character change at or shortly after the splitting of lineages, coupled with a lack of correlation between successive lineage splits.     

Effects of within-patch characteristics on the vulnerability of a plant to herbivory

Consumption of a focal plant by herbivores depends, not only on the physical and chemical characteristics of that plant, but also on the characteristics of the neighbouring vegetation. Consumption of focal plants has been related to their own characteristics and to the quality of the neighbouring vegetation, but the two have not been combined to examine the relative importance of focal plant and neighbouring vegetation characteristics.
We conducted a series of feeding trials to examine the relative importance of focal plant and various characteristics of neighbouring vegetation to browsing of a focal plant within vegetation patches. We planted Eucalyptus nitens seedlings of high and low nutrient status amongst vegetation patches differing in palatability, abundance and height. Generalist mammalian herbivores, red-bellied pademelons ( Thylogale billardierii ), were allowed to feed in each of these patches one at a time, and seedling consumption was recorded. Results were considered in light of the attractant-decoy and apparency hypotheses, which focus on the outcome to plants, and in terms of foraging theory, which is process-focussed.
Seedling and vegetation characteristics were both important. Seedlings of high nutrient status were preferred over those of low nutrient status. The relative quality, abundance and height of neighbouring vegetation all influenced browsing of a focal plant. Seedlings were more vulnerable amongst vegetation that was of low palatability, of low abundance, or was short. Seedling and vegetation effects were additive in two of three trials.
Results were consistent with both the attractant-decoy and apparency hypotheses, and could be explained in terms of maximising foraging efficiency. They demonstrate the need to consider characteristics of both the focal plant and its neighbouring vegetation when predicting the vulnerability of the former to browsing by generalist herbivores.     

基于地面调查的植被生态质量综合评估指标体系构建

植被是陆地生态系统的主体,是保障生态质量的基础,也是基于自然的生态系统增汇、实现“碳中和”的重要利器。植被是生态质量评价的核心要素,但目前的生态质量评价研究中所用到的植被指标多是通过遥感反演或者气象指数模型计算得到的,而基于典型生态系统尺度地面调查的植被观测数据更直接、更准确,数据也很丰富,却很少用于生态质量评价,也缺乏系统的评价指标体系。通过文献研究、专家研讨和问卷调查,并借鉴群落退化演替和生态系统长期监测研究的理论基础,构建基于地面调查的植被生态质量综合评估指标体系。该指标体系整体分为三级,一级综合指数由群落结构指数、物质生产指数、生物多样性指数、群落发展或者演替趋势4个二级分项指标构成,二级分项指标由12—14个三级指标组成,不同植被类型各有其特征指标。该体系将完善我国多尺度陆地生态系统的生态质量评价指标体系,为新时期国家生态质量评价提供科学建议,为我国生态文明建设提供技术支撑。     

亚热带典型红壤侵蚀区人类活动对植被覆盖度及景观格局的影响

福建省长汀县是中国南方最严重的水土流失区之一,在20世纪80年代初和2000年两次集中治理的推动下,当地生态环境已得到显著改善。基于Landsat系列卫星影像提取长汀县1975—2013年共6期植被覆盖度分布图,分析该区在不同时期植被覆盖度及其空间格局的时空动态,并探讨人类干扰与政策治理对植被覆盖度及景观格局的影响。结果表明:(1)近38年来,长汀县平均植被覆盖度由47.02%(1975)提升至71.47%(2013),在覆盖度结构上逐渐形成以中高和高植被覆盖度占主导的格局;县域中部河田盆地的植被覆盖度由30.83%(1975)提升至60.34%(2013)。(2)在景观格局上,研究期间长汀县极低、低和中低覆盖度斑块平均面积呈波动下降趋势、同时斑块密度增加,而中高、高植被覆盖度区域面积扩大,表明封禁、造林等治理措施导致植被覆盖度较高的区域不断汇聚成片。(3)植被覆盖度的提升在空间上主要集中在海拔600 m和坡度25°以下区域,尤其在海拔400—600 m和坡度5°—15°区域最显著,表明植被的破坏和恢复过程与人类活动的联系密切。(4)空间分析表明,在距离农户居民地边缘1.2 km的范围内,越接近居民地中心的区域植被覆盖度越低、破碎度越大且恢复缓慢,但这种空间差异伴随治理进行正在逐步减弱。总体上看,长汀县生态治理和人类干扰的长期驱动影响,其恢复速度在不断提升。     

毛乌素沙地南缘植被景观格局演变与空间分布特征

以1991、1999和2007年TM/ETM+影像为数据源,利用RS和GIS技术提取研究区植被覆盖度信息,将其按地貌分区,计算景观格局指数,分析植被景观格局变化特征.结果表明:(1) 1999年是研究区16a植被景观结构变化的转折点,植被覆盖状况前期恶化后期好转.1991-1999年两区极低和低级植被覆盖度占绝对优势,继续增强,风沙区55％-70％,丘陵区70％-80％.高等级比例均在5％以下,持续下滑.2007年两区的优势级别提升为低级和中级,风沙区比例55％,丘陵区70％.极低级别比例减小,风沙区40％-20％,丘陵区50％-15％.高级比例不同程度增大.(2)1999年前后植被景观格局反向演变,两区变化趋势基本相同程度不同.植被多样性和均匀度先减小后增大,等级问交替分布规律先弱后强.风沙区景观形状由复杂到简单,破碎度先增后减;丘陵区破碎度持续增大,形状变化微小.整体上来看,风沙区植被多样性高整体性好,植被均匀度指数大,交替分布规律明显,丘陵区植被格局较差.但外界对风沙区的干扰比丘陵区强烈.(3)植被景观格局演变过程中始终保持多核心模式,核心面积、形状甚至类型发生不同程度的变化.风沙区核心多为无植被斑块,丘陵区多为高植被覆盖度斑块.核心外围植被分布具有梯度性.     

Plant functional types as predictors of transient responses of arctic vegetation to global change

Abstract. The plant functional types (growth forms) traditionally recognized by arctic ecologists provide a useful framework for predicting vegetation responses to, and effects on, ecosystem processes. These functional types are similar to those objectively defined by cluster analysis based on traits expected to influence ecosystem processes. Principal components analysis showed that two major suites of traits (related to growth rate and woodiness) explain the grouping of species into functional types. These plant functional types are useful because they (1) influence many ecological processes (e.g. productivity, transpiration, and nutrient cycling) in similar ways, (2) predict both responses to and effects on environment, including disturbance regime, and (3) show no strong relationship with traits determining migratory ability (so that no functional type will be eliminated by climatic change simply because it cannot migrate). Circumstantial evidence for the ecological importance of these functional types comes from the distribution of types along environmental gradients and the known ecological effects of traits (e.g., effects of litter quality on decomposition and of plant height on winter albedo) that characterize each functional type. The paleorecord provides independent evidence that some of these functional types have responded predictably to past climatic changes. Field experiments also show that plant functional types respond predictably to changes in soil resources (water and nutrients) but less predictably to temperature. We suggest that evidence for the validity of arctic plant functional types is strong enough to warrant their use in regional models seeking to predict the transient response of arctic ecosystems to global change.     

Estimating plant abundance using inflated beta distributions: Applied learnings from a lichen–caribou ecosystem

Quantifying abundance and distribution of plant species can be difficult because data are often inflated with zero values due to rarity or absence from many ecosystems. Terrestrial fruticose lichens (Cladonia and Cetraria spp.) occupy a narrow ecological niche and have been linked to the diets of declining caribou and reindeer populations (Rangifer tarandus) across their global distribution, and conditions related to their abundance and distribution are not well understood. We attempted to measure effects related to the occupancy and abundance of terrestrial fruticose lichens by sampling and simultaneously modeling two discrete conditions: absence and abundance. We sampled the proportion cover of terrestrial lichens at 438 vegetation plots, including 98 plots having zero lichens. A zero‐inflated beta regression model was employed to simultaneously estimate both the absence and the proportion cover of terrestrial fruticose lichens using fine resolution satellite imagery and light detection and ranging (LiDAR) derived covariates. The probability of lichen absence significantly increased with shallower groundwater, taller vegetation, and increased Sphagnum moss cover. Vegetation productivity, Sphagnum moss cover, and seasonal changes in photosynthetic capacity were negatively related to the abundances of terrestrial lichens. Inflated beta regression reliably estimated the abundance of terrestrial lichens (R² = .74) which was interpolated on a map at fine resolution across a caribou range to support ecological conservation and reclamation. Results demonstrate that sampling for and simultaneously estimating both occupancy and abundance offer a powerful approach to improve statistical estimation and expand ecological inference in an applied setting. Learnings are broadly applicable to studying species that are rare, occupy narrow niches, or where the response variable is a proportion value containing zero or one, which is typical of vegetation cover data.     

Housing arrangement and location determine the likelihood of housing loss due to wildfire

Surging wildfires across the globe are contributing to escalating residential losses and have major social, economic, and ecological consequences. The highest losses in the U.S. occur in southern California, where nearly 1000 homes per year have been destroyed by wildfires since 2000. Wildfire risk reduction efforts focus primarily on fuel reduction and, to a lesser degree, on house characteristics and homeowner responsibility. However, the extent to which land use planning could alleviate wildfire risk has been largely missing from the debate despite large numbers of homes being placed in the most hazardous parts of the landscape. Our goal was to examine how housing location and arrangement affects the likelihood that a home will be lost when a wildfire occurs. We developed an extensive geographic dataset of structure locations, including more than 5500 structures that were destroyed or damaged by wildfire since 2001, and identified the main contributors to property loss in two extensive, fire-prone regions in southern California. The arrangement and location of structures strongly affected their susceptibility to wildfire, with property loss most likely at low to intermediate structure densities and in areas with a history of frequent fire. Rates of structure loss were higher when structures were surrounded by wildland vegetation, but were generally higher in herbaceous fuel types than in higher fuel-volume woody types. Empirically based maps developed using housing pattern and location performed better in distinguishing hazardous from non-hazardous areas than maps based on fuel distribution. The strong importance of housing arrangement and location indicate that land use planning may be a critical tool for reducing fire risk, but it will require reliable delineations of the most hazardous locations.     

黄土高原植被覆盖与水热时空通径分析

探究黄土高原地区气象因子对植被覆盖的影响作用以丰富生态修复理论。基于黄土高原2001—2017年归一化植被指数(Normalized Difference Vegetation Index, NDVI)与气象数据,采用通径分析方法分别从时间和空间尺度上,分析黄土高原气温和降水对植被覆盖变化的直接及间接影响作用,为该地区生态建设提供科学依据。结果如下：黄土高原地区年际间植被明显波动增长,降水变化大体上与植被变化相似;降水整体较气温对植被覆盖变化的作用大。黄土高原植被与水热空间关系的最优分析尺度为80km,在80km空间尺度上,植被与气温有最大相关性,植被、降水由东南到西北递减,而气温分布规律不显著;降水整体呈现促进作用,气温的抑制作用较强,且空间差异明显。在时间与空间尺度上,植被主要受水热促进尤其是降水促进影响,且降水对植被生长的直接作用远大于通过气温的间接作用;不论生态区还是植被类型,气候因子作用均以促进类型为主,但存在明显差异。水热作用在时空尺度上具有明显空间差异性,不同地区影响植被变化的主控因子不同。     

Assessing temporal scales and patterns in time series: Comparing methods based on redundancy analysis

Time-series modelling techniques are powerful tools for studying temporal scaling structures and dynamics present in ecological and other complex systems and are gaining popularity for assessing resilience quantitatively. Among other methods, canonical ordinations based on redundancy analysis are increasingly used for determining temporal scaling patterns that are inherent in ecological data. However, modelling outcomes and thus inference about ecological dynamics and resilience may vary depending on the approaches used. In this study, we compare the statistical performance, logical consistency and information content of two approaches: (i) asymmetric eigenvector maps (AEM) that account for linear trends and (ii) symmetric distance-based Moran's eigenvector maps (MEM), which requires detrending of raw data to remove linear trends prior to analysis. Our comparison is done using long-term water quality data (25 years) from three Swedish lakes. This data set therefore provides the opportunity for assessing how the modelling approach used affects performance and inference in time series modelling. We found that AEM models had consistently more explanatory power than MEM, and in two out of three lakes AEM extracted one more temporal scale than MEM. The scale-specific patterns detected by AEM and MEM were uncorrelated. Also individual water quality variables explaining these patterns differed between methods, suggesting that inferences about systems dynamics are dependent on modelling approach. These findings suggest that AEM might be more suitable for assessing dynamics in time series analysis compared to MEM when temporal trends are relevant. The AEM approach is logically consistent with temporal autocorrelation where earlier conditions can influence later conditions but not vice versa. The symmetric MEM approach, which ignores the asymmetric nature of time, might be suitable for addressing specific questions about the importance of correlations in fluctuation patterns where there are no confounding elements of linear trends or a need to assess causality.     

Bayes' theorem and its applications in animal behaviour

Bayesian decision theory can be used to model animal behaviour. In this paper we give an overview of the theoretical concepts in such models. We also review the biological contexts in which Bayesian models have been applied, and outline some directions where future studies would be useful. Bayesian decision theory, when applied to animal behaviour, is based on the assumption that the individual has some sort of "prior opinion" of the possible states of the world. This may, for example, be a previously experienced distribution of qualities of food patches, or qualities of potential mates. The animal is then assumed to be able use sampling information to arrive at a "posterior opinion", concerning e.g. the quality of a given food patch, or the average qualities of mates in a year. A correctly formulated Bayesian model predicts how animals may combine previous experience with sampling information to make optimal decisions. We argue that the assumption that animals may have "prior opinions" is reasonable. Their priors may come from one or both of two sources: either from their own individual experience, gained while sampling the environment, or from an adaptation to the environment experienced by previous generations. This means that we should often expect to see "Bayesian-like" decision-making in nature.     

Land use changes affect distributional patterns of desert birds in the Baja California peninsula, Mexico

Reliable plans for desert bird conservation will depend on accurate prediction of habitat change effects on their distribution and abundance patterns. Predictive models can help highlight relationships between human‐related and other environmental variables and the presence of desert bird species. Presence/absence of 30 desert bird species of Baja California peninsula was modelled on the basis of explanatory variables taken from the field, maps, and digital imagery. Generalized linear models were fit to each bird species using both variables representing human activity and other environmental factors as predictors that might influence distribution. Probability of species presence was used as a habitat suitability index to evaluate the effect of human activity when the model contained a significant human activity variable. No differences were found in bird species richness between natural sites and those transformed by agriculture or urbanization. Of 59 bird species recorded in surveys, 34% were positively or negatively associated with human‐transformed habitats. Fourteen species seem to benefit from transformation of natural vegetation by agriculture or urbanization, while six were negatively affected. Sensitivity analyses of final models indicated all were robust. Results suggest that the occurrence of a large percentage of bird species inhabiting scrub habitats is sensitive to human habitat transformation. This finding has important conservation implications at regional scale as fragmentation and conversion of desert ecosystems into agricultural and urban areas affect the distribution of species that are highly selective for scrub habitat. Land use and anthropogenic activities seem to change ecological patterns at large spatial scales, but other factors could drive species richness distribution too (i.e. individual species response, species–energy relationships). The spatial modelling approach at regional scale used in this study can be useful for designing natural resource management plans in the Sonoran desert scrub.