A resource selection model for analyzing pseudoreplicated data due to grouping behavior of animals

Quantifying resource selection is of primary interest in animal ecology. Many analyses of resource selection assume spatial and temporal independence of the sampling unit. Autocorrelation between observations, which is a general property of ecological variables, causes difficulties for most standard statistical procedures of resource selection because autocorrelated data violate the assumption of independence. To overcome this problem, we develop a mixed-effects model to estimate resource selection functions from data that are autocorrelated because of unobserved grouping behavior by animals. In the application of the expectation-maximization (EM) algorithm, the computation of the conditional expectation of the complete-data log-likelihood function does not have a closed-form solution requiring numerical integration. A Monte Carlo EM algorithm with Gibbs sampling can be used effectively in such situations to find exact maximum likelihood estimates. We propose a simple automated Monte Carlo EM algorithm with multiple sequences in which the Monte Carlo sample size is increased when the EM step is swamped by Monte Carlo errors.We demonstrate that the model can detect inherent autocorrelation and provide reasonable variance estimates when applied to nocturnal bird migration data. This approach could also be applied to ecological processes with various types of spatially and temporally autocorrelated data, circumventing serious problems caused by dangerous pseudoreplication.     

Bayesian variable selection in clustering high-dimensional data with substructure

In this article we focus on clustering techniques recently proposed for high-dimensional data that incorporate variable selection and extend them to the modeling of data with a known substructure, such as the structure imposed by an experimental design. Our method essentially approximates the within-group covariance by facilitating clustering without disrupting the groups defined by the experimenter. The method we adopt simultaneously determines which expression patterns are important, and which genes contribute to such patterns. We evaluate performance on simulated data and on microarray data from a colon carcinogenesis study. Selected genes are biologically consistent with current research and provide strong biological validation of the cluster configuration identified by the method.     

Bayesian capture-recapture analysis and model selection allowing for heterogeneity and behavioral effects

In this article, we present Bayesian analysis of capture-recapture models for a closed population which allows for heterogeneity of capture probabilities between animals and bait/trap effects. We use a flexible discrete mixture model to account for the heterogeneity and behavioral effects. In addition we present a solid model selection criterion. Through illustrations with a motivating dataset, we demonstrate how Bayesian analysis can be applied in this setting and discuss some major benefits which result, including consideration of informative priors based on historical data.     

A Bayesian hierarchical model for risk assessment of methylmercury

This article uses a Bayesian hierarchical model to quantify the adverse health effects associated with in-utero exposure to methylmercury. By allowing for study-to-studyas well as outcome-to-outcome variability, the approach provides a useful meta-analytic tool for multi-outcome, multi-study environmental risk assessments. The analysis presented here expands on the findings of a National Academy of Sciences (NAS) committee, charged with advising the United States Environmental Protection Agency (EPA) on an appropriate approach to conducting a risk assessment for methylmercury. The NAS committee, for which the senior author (Ryan) was a committee member, reviewed the findings from several conflicting studies and reported the results from a Bayesian hierarchical model that synthesized information across several studies and for several outcomes. Although the NAS committee did not suggest that the hierarchical model be used as the actual basis for a methylmercury risk assessment, the results from the model were used to justify and support the final recommendation that the risk analysis be based on data from a study conducted in the Faroe Islands, which had found a positive association between in-utero exposure to methylmercury and impaired neurological development. We considera variety of statistical issues, but particularly sensitivity to model specification.     

A hierarchical model for compositional data analysis

This article introduces a hierarchical model for compositional analysis. Our approach models both source and mixture data simultaneously, and accounts for several different types of variation: these include measurement error on both the mixture and source data; variability in the sample from the source distributions; and variability in the mixing proportions themselves, generally of main interest. The method is an improvement on some existing methods in that estimates of mixing proportions (including their interval estimates) are sure to lie in the range [0, 1]; in addition, it is shown that our model can help in situations where identification of appropriate source data is difficult, especially when we extend our model to include a covariate. We first study the likelihood surface of a base model for a simple example, and then include prior distributions to create a Bayesian model that allows analysis of more complex situations via Markov chain Monte Carlo sampling from the likelihood. Application of the model is illustrated with two examples using real data: one concerning chemical markers in plants, and another on water chemistry.     

Genetic parameters and selection of maize cultivars using Bayesian inference in a multi-trait linear model

Variance components must be obtained to estimate genetic parameters and predict breeding values. In studies which take many traits into account, it is reasonable to use the Bayesian approach for the estimation of genetic parameters. The main goal of the present research was not only to consider the genetic correlations of the examined traits, but above all to estimate unknown genetic parameters and to gain profits from the selection. Bayesian inference was also useful for the selection of the best maize varieties. It was applied to predict genetic values in the multi-traits linear model. Thirteen maize cultivars representing the traits of our interest were studied by means of Bayesian inference. The traits are the number of plants before harvest, the grain yield, the length of the ears, the mass of leaves and the number of ears. The experiment involved a randomised block design with four replications and ten plants per plot. The highest correlation estimates were found between the number of plants before harvest and the number of ears, jointly with the grain yield and the number of ears. Lower correlation estimates were found between the length of the ears and the number of ears as well as the grain yield and the length of the ears. The research confirms that the best varieties to be grown are: Clarica, NK Cooler, Drim and PR 39K13. The Bayesian approach proved to be useful in selection studies, which can further be used to improve the studied genotypes.     

A test of the hierarchical model of habitat selection using eastern massasauga rattlesnakes (Sistrurus c. catenatus)

Understanding the process animals follow to select habitat, rather than just documenting the habitat they use, will improve our ability to predict how the animals use habitat in other locations and how they will respond to changes in habitat. Animals are usually assumed to select habitats hierarchically, preferentially using specific macrohabitats at a landscape scale and specific microhabitats within the preferred macrohabitats. We used four years of telemetry data from 34 individuals to test this hierarchical model of habitat selection with eastern massasauga rattlesnakes (Sistrurus c. catenatus) in Ontario. Snakes were selective at the microhabitat scale, preferentially using locations with closer retreat sites and shrubs than random. Gravid females were most selective, using sites with more rock cover and less canopy closure than sites used by males and nongravid females. Snakes preferred forested habitats for hibernation and steadily increased their use of open, wetland, and edge habitats to a peak in mid-summer. Landscape-scale habitat preferences were generally mild and could be explained by the relative availability of suitable microhabitat within habitats, suggesting habitat selection was primarily driven by microhabitat preferences. The lack of selectivity at the landscape scale may be a consequence of fine-grained differences between habitats that allow massasauga rattlesnakes to find suitable microhabitats in all available macrohabitats. For species that select habitat primarily at the microhabitat scale (e.g., the rattlesnakes we studied), landscape-scale modeling of habitat use will only be effective to the extent habitats reflect the availability of suitable microhabitat within.     

Hunting habitat selection by hen harriers on moorland: Implications for conservation management

We examine habitat use by hunting hen harriers Circus cyaneus at three study sites in Scotland to evaluate whether foraging patterns differ between sexes, sites, and stages of the breeding period. We modelled time spent hunting in focal plots as a function of habitat and nest proximity. Male hunting intensity (time spent hunting per hour of observation and km²) varied between sites and breeding periods, being lower during the nestling than the incubation period. Habitat use patterns were mostly consistent among study sites, which is important for developing species management recommendations applicable over the species’ range. Males avoided improved grassland, and selected areas of mixed heather and rough grass (with an optimum at ca. 50% heather cover). The effect of nest proximity was small. In contrast, females hunted mainly within 300-500 m of the nest, with a small additive effect of vegetation cover, areas of fragmented heather being preferred. Habitat management to benefit foraging harriers will involve creating (or maintaining) mosaics of heather/grassland around nest areas. Additionally, it might be possible to manipulate habitat to reduce conflict in areas where harrier predation on red grouse is important by segregating areas holding highest grouse densities (with high heather cover) from those favoured for harrier foraging (heather-grass mosaics). However, it would be necessary to test whether these manipulations might also influence harrier nest distribution, an effect which could negate any benefits from this strategy.     

Accounting for habitat loss rates in sequential reserve selection: Simple methods for large problems

We develop reserve selection methods for maximizing either species retention in the landscape or species representation in reserve areas. These methods are developed in the context of sequential reserve selection, where site acquisition is done over a number of years, yearly budgets are limited and habitat loss may cause some sites to become unavailable during the planning period. The main methodological development of this study is what we call a site-ordering algorithm, which maximizes representation within selected sites at the end of the planning period, while accounting for habitat loss rates in optimization. Like stochastic dynamic programming, which is an approach that guarantees a globally optimal solution, the ordering algorithm generates a sequence in which sites are ideally acquired. As a distinction from stochastic dynamic programming, the ordering is generated via a relatively fast approximate process, which involves hierarchic application of the principle of maximization of marginal gain. In our comparisons, the ordering algorithm emerges a clear winner, it does well in terms of retention and is superior to simple heuristics in terms of representation within reserves. Unlike stochastic dynamic programming, the ordering algorithm is applicable to relatively large problem sizes, with reasonable computation times expected for problems involving thousands of sites.     

一个用于分析土地覆被变化对径流机制影响的分布式月水量平衡模型

The Miyun Reservoir is the most important water source for Beijing Municipality, the capital of China with a population of more than 12 million. In recent decades, the inflow to the reservoir has shown a decreasing trend, which has seriously threatened water use in Beijing. In order to analyze the influents of land use and cover change （LUCC） upon inflow to Miyun Reservoir, terrain and land use information from remote sensing were utilized with a revised evapotranspiration estimation formula; a water loss model under conditions of human impacts was introduced; and a distributed monthly water balance model was established and applied to the Chaobai River Basin controlled by the Miyun Reservoir. The model simulation suggested that not only the impact of land cover change on evapotranspiration, but also the extra water loss caused by human activities, such as the water and soil conservation development projects should be considered. Although these development projects were of great benefit to human and ecological protection, they could reallocate water resources in time and space, and in a sense thereby influence the stream flow.     

A habitat model for brown bear conservation and land use planning in the central Apennines

We investigated the brown bear habitat suitability in an 8000 km² study area encompassing Abruzzo, Latium, and Molise regions in central-southern Italy. Based on long-term field surveys and published records, we classified bear habitat as occupied or unoccupied in 92 out of 320 sample squares (5 × 5 km). For each sample square 36 habitat variables were measured from topographic maps and Corine land-cover III level digital maps. The influence of habitat features on bear presence was investigated by multivariate and one-way analyses of variance and by logistic regression analysis. The logistic model correctly classified 95.5% of sample squares of bear presence and 93.8% of those where bears were absent. Average altitude, deciduous woodlands and ecotone length, showed a positive relationship with bear presence, whereas vineyard-olive groves and shrublands were negatively correlated with bear presence. No specific land management guidelines or strategies exist for bear conservation in central Italy, based on knowledge of habitat-population relationships. The landscape scale habitat model we developed could be useful to predict bear occurrence, to identify critical areas for a brown bear conservation strategy, and to enhance the arrangement of the protected areas network for the conservation of this species.     

Winter habitat selection by two sympatric forest grouse in western Switzerland: implications for conservation

Two endangered tetraonids, the capercaillie (Tetrao urogallus) and the hazel grouse (Bonasa bonasia rupestris), are sympatric throughout part of their distribution range in central Europe. Precise information on their specific habitat requirements is needed if the coexistence of both species in exploited forests is to be maintained. We quantified winter habitat selection for both species in the upper part (1100-1600 m) of the Jura mountains (Switzerland). No preference for altitude or exposure could be detected. Capercaillie preferred open forests (including grazed forests) with a sparse canopy dominated by spruce (Picea abies) and fir (Abies alba), and avoided dense undercanopy and understorey, especially when dominated by spruce and beech (Fagus sylvatica). By contrast, hazel grouse preferred feeding sites with a dense understorey of rowan (Sorbus aucuparia), willow (Salix sp.), beech and spruce. These preferences can be related to the feeding habits and predator avoidance behaviour of both species. Coexistence thus requires a mosaic distribution of habitat types, with a matrix of open forests (30% canopy cover) where fir is favoured, and understorey kept sparse (20%). Group-cuts of mature trees should allow regeneration patches, where a dense understorey (50% cover) should provide suitable habitats for hazel grouse.     

Mark-recapture with occasion and individual effects: Abundance estimation through Bayesian model selection in a fixed dimensional parameter space

We present a Bayesian mark-recapture method for explicitly communicating uncertainty about the size of a closed population where capture probabilities vary across both individuals and sampling occasions. Heterogeneity is modeled hierarchically using a continuous logistic-Normal model to specify the capture probabilities for both individuals that are captured on at least one occasion and individuals that are never captured and so remain undetected. Inference about how many undetected individuals to include in the model is accomplished through a Bayesian model selection procedure using MCMC, applied to a product space of possible models for different numbers of undetected individuals. Setting the estimation problem in a fixed dimensional parameter space enables the model selection procedure to be performed using the freely available WinBUGS software. The outcome of inference is a full “posterior” probability distribution for the population size parameter. We demonstrate this method through an example involving real mark-recapture data.     

烟草田间作业机械评价模型的构建

为评价农业机械的适用性满意度,探索基于实际生产管理需要的农业机械选型方法。以烟草种植和烟草田间作业机械为研究对象,将烟草种植流程划分为3个阶段10个环节,并设定各环节的评价指标体系,每个环节设计了作业成本、作业效率、作业效果、使用效应等4个二级指标,共44个评价指标。结合田间实地测试采集评价指标数据;依据农艺要求量化评价指标的评判基准;采用层次分析法,构建综合评价模型。以起垄环节为例评价模型的应用,共涉及7台不同类型的起垄机组（编号分别为I,II,...VII）,其中,IV、V为"履带式",VII为"复合式",其余均为"手扶式"。测试过程表明,"手扶式"起垄机组工作状态易调整、作业效果好;"履带式"起垄机组整机质量较大,动力不足时容易导致作业、行进速度变缓;"复合式"起垄机组整机结构复杂,操作舒适性和安全性不高。从评价结果看,参与测试的7台起垄机组中,作业成本指标排序靠前是I、III、II号机组,评价值分别为0.91、0.89、0.85;作业效率指标排序靠前的是V、IV、II号机组,评价值分别为1、0.49、0.41;作业效果指标排序靠前的是II、V、I号机组,评价值分别为0.96、0.69、0.66;使用效应指标排序靠前的是I、II、V号机组,评价值分别为0.93、0.75、0.59;综合评价表现较佳的是II、I、V号机组,适用性满意度评分分别为0.75、0.68、0.66,选型评价结果与当地实际情况相吻合。该模型可为农业机械选型定量化的比较提供依据和参考。     

Nesting habitat selection by bitterns Botaurus stellaris in Britain and the implications for wetland management

The rarity and decline of the bittern Botaurus stellaris in Britain has prompted large-scale wetland restoration and more recently, wetland creation projects. In order to guide such habitat management, we investigated whether any large or fine-scale features within British wetland sites best described the selection of female bittern nesting positions. Birds nested in continuous vegetation (usually Phragmites dominated) that was on average 100 m at its narrowest width. When compared with random locations, nests had less scrub and more vegetated open water edge in their wider vicinity and were immediately surrounded by smaller percentage coverage of non-Phragmites species and thicker Phragmites stems. Of most importance, female bitterns nested at points where deeper water was maintained into the driest part of the season, perhaps using the presence of water tolerant plant species as an indication of this. The results can be used to aid the design of new wetland sites to take into account the needs of nesting female bittern.     

基于数据仓库的土壤环境监测综合挖掘模型构架

为解决土壤环境监测全流程优化问题,提出了监测流程综合挖掘方法,采用数据仓库和工作流挖掘技术,构建了一个土壤环境监测综合挖掘模型构架.首先,采用雪花结构建立了数据模型构架,为流程挖掘提供数据底层;然后建立了监测流程综合挖掘模型构架,给出了监测单元挖掘、监测点位挖掘和监测数据挖掘,以及综合挖掘的模型函数和参数,以局部优化和全局优化的思路实现了土壤环境监测全流程优化.研究结果对于提高土壤环境监测效率,拓展LIMS实验室信息管理系统功能,构建土壤环境监测数据仓库和专家系统提供指导.     

Managing landscapes for the little bustard Tetrax tetrax: lessons from the study of winter habitat selection

More than half of the world's population of the little bastard Tetrax tetrax lives in the Iberian Peninsula, where it is mostly dependent for survival on extensive agricultural areas. The species has declined dramatically, chiefly due to changes in agricultural practices, and is now globally ‘Near Threatened’. Knowledge of its habitat requirements is crucial to reverse the trend of habitat deterioration. Winter habitat preferences were studied in a region dominated by extensive cereal farming in Southern Portugal, comparing the characteristics of sites used by 54 flocks with those of randomly selected sites within the study area. The birds preferred recent fallows and grassy vegetation of mid-size (about 11-20 cm) and mid-density (about 11-50% cover). They tended to concentrate on the tops of hills and to avoid disturbed areas near roads and houses. Overall, the observed preferences suggest that predator avoidance is a significant factor in habitat selection. To improve habitat suitability for little bustard, managers of key wintering areas should minimize permanent sources of human disturbance, encourage rotations with frequent fallows, and favour moderate levels of grazing to manipulate vegetation height and cover. Particular attention should be given to the upper parts of the hills.     

柑桔叶片黄龙病光谱特征选择及检测模型

为探索高光谱技术诊断黄龙病及分类的可行性,通过变量筛选方法组合为高维数据实用化提供参考。采集柑桔叶片高光谱图像并进行普通(polymerase chain reaction,PCR)鉴别分为轻度、中度、重度、缺锌和正常5类样品。用无信息变量消除算法(uninformative variable elimination,UVE)剔除无关信息,组合遗传算法(genetic algorithm,GA)和连续投影算法(successive projections algorithm,SPA)筛选变量,对数据进行降维。结合极限学习机(extreme learning machine,ELM)和最小二乘支持向量机(least squares support vector machine,LS-SVM)构建柑桔黄龙病判别模型。对预测样品进行诊断分类,来评价模型判别能力。经对比发现,UVE组合SPA筛选变量后的LS-SVM模型效果最好,该模型以Link_kernel函数为核函数,惩罚因子(γ)最小为1.07,误判率最低为0。用全谱作输入变量时LS-SVM模型复杂程度最高且预测能力最差,误判率最高为11.9%,可能是包含无用信息和冗余信息变量造成的。研究显示,UVE组合SPA筛选变量,结合LS-SVM对柑桔黄龙病诊断并分类具有一定可行性,为高维度数据实用化提供一定参考价值。