Michel Goulard - Academia.edu (original) (raw)

Papers by Michel Goulard

Behavioural consistency is a key assumption when evaluating how between-individual differences in... more Behavioural consistency is a key assumption when evaluating how between-individual differences in behaviour influence life history tactics. Hence, understanding how and why variation in behavioural repeatability occurs is crucial. While analyses of behavioural repeatability are common, few studies of wild populations have investigated variation in repeatability in relation to individual status (e.g. sex, age, condition) and over different timescales. Here, we aimed to fill this gap by assessing within-population variation in the repeatability of docility, as assessed by the individual's response to human handling, in a free-ranging population of European roe deer, Capreolus capreolus. Docility was an equally repeatable behaviour at both short-and long-term timescales, suggesting that this behavioural trait is stable across time. Repeatability did not differ markedly between age and sex categories but tended to be higher in juvenile males than in juvenile females. Finally, contrary to expectation, individual variation in the repeatability of docility was not correlated with individual body mass. Further studies are required to assess the life history consequences of the individual variation in docility we report here.

Seven different strategies for sampling a core collection are compared in large collections of na... more Seven different strategies for sampling a core collection are compared in large collections of natural populations of two fodder crop species (Lolium perenne and Medicago truncatula) previously evaluated for agronomic traits. The first five strategies consist of random or stratified sampling methods with one level of classification of the population (classification based on agronomic traits with or without a geographic contiguity constraint, based on the country or region of origin, or on spatial structures of populations). The last one of these strategies is based on geostatistical analysis, which allows studies of the spatial pattern of genetic diversity. The two remaining strategies are based on the principal component score strategy (PCSS), which consists of maximising a distance criterion between populations, this criterion being applied to agronomic traits of the populations or to their spatial components after geostatistical analysis. The different strategies are compared using two criteria: i) their capacity to capture the phenotypic diversity of the whole collection, as measured by the Shannon index; ii) their ability to restore the spatial structuration of the initial collection, measured by distance between maps for agronomic traits. The comparisons for the two species show that strategies that take into account spatial structuration of diversity give the best results for both criteria. © Inra/Elsevier,

Les données de télédétection par satellite donnent accès à la dynamique de la végétation sur tout... more Les données de télédétection par satellite donnent accès à la dynamique de la végétation sur toute une région par l'acquisition régulière de mesures de réflectances. Les méthodes actuelles de traitement de ces données sont basées essentiellement sur une analyse des images pixel par pixel. Notre objectif est de comparer cette approche avec une modélisation par paramètres aléatoires qui traite l'ensemble des pixels simultanément pour ajuster un modèle non linéaire d'évolution temporelle de l'indice de surface foliaire. Les différentes conditions d'observation (nombre de dates, variance du bruit) et de modélisation (nombre de paramètres à estimer) sont analysées par simulation et une mise en oeuvre sur des données réelles est effectuée. La procédure par paramètres aléatoires est plus robuste aux changements de conditions expérimentales.

The aim of this study was to evaluate the ability of lambs to discriminate familiar from unfamili... more The aim of this study was to evaluate the ability of lambs to discriminate familiar from unfamiliar agemates. Four groups of 4 lambs were raised in adjacent outdoor pens without visual contact among groups. Lambs of adjacent groups were considered unfamiliar. Each lamb (9 females and 7 males) was presented to 3 stimulus lambs (1 familiar and 2 unfamiliar) during 3 successive 5-min trials. The recorded variables were: (1) first visit latency, (2) social category of the first lamb visited, (3) number of visits and (4) total time of visits. The responses of the lambs were affected (ANOVA) by the repeated testing on any given day by a main effect of the factor "trial" on variables (2) to (4); lambs presented low motivation to approach stimulus lambs in the second and third trials. During the first trials, lambs at 5-11 days of age were slower to visit and spent less time near stimulus agemates than at older age. Overall, lambs seemed to discriminate familiar from unfamiliar la...

International Journal of Agricultural and Environmental Information Systems, 2013

Journal of Wildlife Management

To determine the spatial resolution of Global Positioning System (GPS) receiver data, rigorous te... more To determine the spatial resolution of Global Positioning System (GPS) receiver data, rigorous testing is essential. We tested performance of the Lotek 3300 GPS collar for medium-sized mammals (Lotek Engineering, Inc., Newmarket, ON, Canada). To mimic real wildlife monitoring situations, we performed both static (stationary receiver) and mobile tests, placing the receiver collar on a dog. We compared fix locations of the mobile receiver with the actual trajectory described by a portable Trimble high-precision GPS. We determined performance in relation to habitat type and leaf cover. Location error was habitat-dependent, with the best results in open habitat and much poorer ones in forest, particularly coniferous-dominated forest. For both static and mobile tests, location accuracy was higher when the number of satellites contacted was high and when the residual positional dilution of precision (PDOP) value was low. However, location error was highly variable, even for a given PDOP value and a given number of satellites contacted. Finally, mobile collars performed less well than their static counterparts, presumably because of frequent changes of GPS position and orientation. (JOURNAL OF WILDLIFE MANAGE-MENT 71(4):

ABSTRACT The concept of developing a core collection involves identifying a subsample of manageab... more ABSTRACT The concept of developing a core collection involves identifying a subsample of manageable size from a larger collection, which is as representative as possible of the total genetic diversity in the whole collection. Different sampling strategies for developing a core collection were compared in large collections of natural populations of perennial ryegrass ( Lolium perenne), an outcrossing species, and medic ( Medicago truncatula), a self-pollinating species. Natural populations of the two species, representing contrasting mating systems, were evaluated for agronomic traits. Some of the sampling strategies were based on geostatistical analysis, which allows studies of the spatial pattern of genetic diversity. The different strategies were compared using 3 criteria: (1) their ability to restore the spatial or geographic structure of the initial collection, measured by the distance between maps for agronomic traits; (2) their capacity to capture the phenotypic diversity of the whole collection, as measured by the Shannon diversity index; and (3) their effect on the conservation of neutral alleles, which was observed by measuring the mean number of alleles at isoenzyme loci. Comparisons for the two species showed that, although there was a minimal effect on the mean number of neutral alleles, the strategies that took into account the spatial structure of diversity gave the best core collections based on the first and second criteria.

Quantitative Geology and Geostatistics, 1993

We address the problem of prediction in a classical spatial simultaneous au- toregressive model. ... more We address the problem of prediction in a classical spatial simultaneous au- toregressive model. The optimality of prediction formulas in non-spatial regression models is not immediately transposable to the framework of all spatial models. In the geostatistical literature, much attention has been devoted to this topic, with the development of the Best Linear Unbiased Prediction formulas. In contrast, in the

Animal Behaviour, 2015

Vigilance in prey species can serve many purposes, including predator detection and monitoring ot... more Vigilance in prey species can serve many purposes, including predator detection and monitoring other group members, and is generally thought to impose a cost due to reduced food intake. However, previous studies have shown that herbivores are able to reduce the foraging cost of vigilance by chewing their food during vigilance bouts ('vigilance with chewing', as compared to 'vigilance without chewing'). How predation risk, food availability and competition affect both the functions and the foraging costs of vigilance remains an open question. We studied female eastern grey kangaroos, Macropus giganteus, during winter and summer, when available food supplies were poor and rich, respectively, to investigate how group size, distance to cover, proximity between foragers and food patch quality affected decisions of foraging female kangaroos to exhibit antipredator or social vigilance, distinguishing vigilance with and without chewing. The use of antipredator vigilance was mainly driven by the perception of predation risk, and antipredator vigilance without chewing decreased with increased group size whereas antipredator vigilance with chewing increased nonlinearly with group size in winter. Distance to cover affected both forms of antipredator vigilance in summer only but there was no effect of nearestneighbour distance. Social vigilance was affected positively by group size, and distance between foragers affected social vigilance without chewing positively, particularly in winter, and social vigilance with chewing negatively. Finally, patch quality increased the use of social vigilance with chewing in both seasons and decreased the use of antipredator vigilance with chewing in winter. This study provides new information on how animals make decisions about the functions and foraging costs of vigilance and allows a better understanding of how social foragers respond to an ever-changing environment.

Microscopy Microanalysis Microstructures, 1996

ABSTRACT In this paper the modeling of soil surface roughness by Boolean random functions is pres... more ABSTRACT In this paper the modeling of soil surface roughness by Boolean random functions is presented. The interest of this class of functions is to be able to model explicitly the cloddy aspect of the soil through the use of primary functions. From a statistical point of view, the height distribution function for such models is easily got, so that classical estimation and modeling methods are available. The case of parametric and non-parametric estimation of the intensity function, and the case of the modeling of the non-stationarity induced by tilling;Ire presented.

Theoretical and Applied Genetics, 1994

Methods based on geostatistics were applied to quantitative traits of agricultural interest measu... more Methods based on geostatistics were applied to quantitative traits of agricultural interest measured on a collection of 547 wild populations of perennial ryegrass in France. The mathematical background of these methods, which resembles spatial autocorrelation analysis, is briefly described. When a single variable is studied, the spatial structure analysis is similar to spatial autocorrelation analysis, and a spatial prediction method, called "kriging", gives a filtered map of the spatial pattern over all the sampled area. When complex interactions of agronomic traits with different evaluation sites define a multivariate structure for the spatial analysis, geostatistical methods allow the spatial variations to be broken down into two main spatial structures with ranges of 120 km and 300 km, respectively. The predicted maps that corresponded to each range were interpreted as a result of the isolation-by-distance model and as a consequence of selection by environmental factors. Practical collecting methodology for breeders may be derived from such spatial structures.

Statistics, 1994

A non-parametric estimation method is proposed for the intensity of the Poisson process associate... more A non-parametric estimation method is proposed for the intensity of the Poisson process associated to a Boolean random function. The case studied is that of marks being half-spheres as primary functions, these half-spheres lying on a given horizontal plane. The proposed method is derived from the method presented in Györfi, Härdle, Sarda et Vieu [1989] for the case of time

SIAM Journal on Discrete Mathematics, 2010

ABSTRACT We consider the problem of specifying the joint distribution of a collection of variable... more ABSTRACT We consider the problem of specifying the joint distribution of a collection of variables with maximum entropy when a set of marginals are fixed. One can easily derive that the structure of the solution joint distribution is that of a graphical model. The potential functions are then marginals at some power. We address the following question, Under which conditions on the set of constraints is it possible to fully identify the canonical exponents in the maximum entropy solution as functions of the problem structure? Literature related to this topic is somewhat scattered in disciplines such as statistical mechanics, information theory, graph theory, and inference in graphical models. In this article we gather and link results from these different fields. From this, we show that for a particular class of constraints set on marginals, the chordal maximal coherent sets of constraints, it is possible to derive the canonical exponents of the graphical model solution of the maximum entropy problem as the numbers of occurrences of separators in an associated join tree. Conversely, we present sufficient conditions to ensure that a graphical model is a solution of a maximum entropy problem.

Mathematical Geology, 1992

The geostatistical analysis of multivariate data involves choosing and fitting theoretical models... more The geostatistical analysis of multivariate data involves choosing and fitting theoretical models to the empirical matrix. This paper considers the specific case of the model of linear coregionalization, and describes an automated procedure for ftting models, that are adequate in the mathematical sense, using a least-squares like technique. It also describes how to decide whether the number of parameters of the cross-variogram matrix model should be reduced to improve stability of fir The procedure is illustrated with an analysis of the spatial relations among the physical properties of an alluvial soil. The results show the main influence of the scale and the shape of the basic models on the goodness of fir The choice of the number of basic models appears of secondary importance, though it greatly influences the resulting interpretation of the coregionalization analysis.

Landscape Ecology, 2008

In this study, we sought to understand how landscape structure affects roe deer movements within ... more In this study, we sought to understand how landscape structure affects roe deer movements within their home-range in a heterogeneous and fragmented agricultural system of south-western France. We analysed the movements of 20 roe deer fitted with GPS collars which recorded their locations every 2-6 h over several months (mean = 9 months). Based on empirical observations and previous studies of roe deer habitat use, we hypothesised that roe deer should avoid buildings and roads, move preferentially along valley bottoms and through the more wooded areas of the landscape. To test these hypotheses we paired each observed movement step with 10 random ones. Using conditional logistic regression, we modelled a step selection function, which represents the probability of selecting a given step as a function of these landscape variables. The selected model indicated that movements were influenced by all the tested landscape features, but not always in the predicted direction: our results suggested that roe deer tend to avoid buildings, roads, valley bottoms and possibly the more wooded areas (although the latter result should be interpreted with caution, as it may be influenced by a bias in the rate of GPS fix acquisition in woods). The distances to buildings and to roads were the most influential variables in the model, suggesting that the avoidance of potential sources of disturbance may be a key factor in determining ranging behaviour of roe deer in human dominated landscapes.