Possible causes and consequences of philopatry and breeding dispersal in red-backed shrikes <Emphasis Type="BoldItalic">Lanius collurio</Emphasis>

Studies of animal breeding dispersal have often focused on possible causes, whereas its adaptive significance has received less attention. Using an information-theoretic approach, we assessed predictions of four hypotheses relating to causes and consequences of breeding dispersal in a migratory passerine, the red-backed shrike Lanius collurio. As predicted by the reproductive performance hypothesis, probability of breeding dispersal in females (though not in males) decreased with increasing annual average number of fledglings produced in the past year, but there was no association with conspecific reproductive performance in either sex. The site choice hypothesis, stating that individuals disperse to improve breeding site quality, received support in males only, as dispersal probability was positively associated to a measure indicating low territory quality. The social constraints hypothesis, referring to dispersal in relation to intraspecific interactions, received little support in either sex. The predation risk hypothesis was hardly supported either. Consequences of dispersal were marginal in both sexes because neither fledgling production in females, nor territory quality in males improved after dispersal. In addition, males settled on territories closer to the forest edge than those occupied predispersal, which is opposite to the prediction of the predation risk hypothesis. We conclude that own reproductive success was the major factor determining dispersal behavior in females, whereas territory quality and possibly predation risk were most important in males. Overall, breeding dispersal appeared not to be adaptive in this dense population inhabiting an optimal habitat.     

A spatial clustering perspective on autocorrelation and regionalization

We revisit one of the classical problems in geography and cartography where multiple observations on a lattice (N) need to be grouped into many fewer regions (G), especially when this number of desired regions is unknown a priori. Since an optimization through all possible aggregations is not feasible, a hierarchical classification scheme is proposed with an objective function sensitive to spatial pattern. The objective function to be minimized during the assignment of observations to regions (classification) consists of two terms: the first characterizes accuracy and the second, model complexity. For the latter, we introduce a spatial measure that characterizes the number of homogeneous patches rather than the usual number of classes. A simulation study shows that such a classification procedure is less sensitive to random and spatially correlated error (noise) than non-spatial classification. We also show that for conditional autoregressive error (noise) fields the optimal partitioning is the one that has the highest within-units generalized Moran coefficient. The classifier is implemented in ArcView to demonstrate both a socio-economic and an environmental application to illustrate some potential applications.

基于AIC准则优选AR模型研究我国生产事故

目前,由于生产力水平低,安全投入不足,安全监管体制不健全等因素,国内生产安全事故仍然居高不下.将基于AIC准则进行 AR 模型优选的方法应用到我国工矿死亡人数预测中,为事故预测统计提供一种新方法,预测结果将为预防生产安全事故提供依据.     

Landscape-level impact of tropical forest loss and fragmentation on bird occurrence in eastern Guatemala

Tropical forest destruction and fragmentation of habitat patches may reduce population persistence at the landscape level. Given the complex nature of simultaneously evaluating the effects of these factors on biotic populations, statistical presence/absence modelling has become an important tool in conservation biology. This study uses logistic regression to evaluate the independent effects of tropical forest cover and fragmentation on bird occurrence in eastern Guatemala. Logistic regression models were constructed for 10 species with varying response to habitat alteration. Predictive variables quantified forest cover, fragmentation and their interaction at three different radii (200, 500 and 1000 m scales) of 112 points where presence of target species was determined. Most species elicited a response to the 1000 m scale, which was greater than most species’ reported territory size. Thus, their presence at the landscape scale is probably regulated by extra-territorial phenomena, such as dispersal. Although proportion of forest cover was the most important predictor of species’ presence, there was strong evidence of area-independent and -dependent fragmentation effects on species presence, results that contrast with other studies from northernmost latitudes. Species’ habitat breadth was positively correlated with AIC model values, indicating a better fit for species more restricted to tropical forest. Species with a narrower habitat breadth also elicited stronger negative responses to forest loss. Habitat breadth is thus a simple measure that can be directly related to species’ vulnerability to landscape modification. Model predictive accuracy was acceptable for 4 of 10 species, which were in turn those with narrower habitat breadths.     

Modelling heterogeneity in the recoveries of marked animal populations with covariates of individual animals, groups of animals or recovery time

A general framework is developed for modelling rates of survival and recovery of marked animal populations in terms of auxiliary information collected at the time of marking. The framework may be used to estimate differences in survival or recovery among individual animals, groups of animals, and recovery times. Analyses of the recoveries of tagged fish and banded bird populations are used to illustrate the specification and selection of various models.     

Using a model selection criterion to identify appropriate complexity in aquatic biogeochemical models

Aquatic biogeochemical models are widely used as tools for understanding aquatic ecosystems and predicting their response to various stimuli (e.g., nutrient loading, toxic substances, climate change). Due to the complexity of these systems, such models are often elaborate and include a large number of estimated parameters. However, correspondingly large data sets are rarely available for calibration purposes, leading to models that may be overfit and possess reduced predictive capabilities. We apply, for the first time, information-theoretic model-selection techniques to a set of spatially explicit (1D) algal dynamics models of varying parameter dimension. We demonstrate that increases in complexity tend to produce a better model fit to calibration data, but beyond a certain degree of complexity the benefits of adding parameters are diminished (the risk of overfitting becomes greater). The particular approach taken here is computationally expensive, but several suggestions are made as to how multimodel methods may practically be extended to more sophisticated models.