Guillaume Larocque

Large-scale, spatially-explicit test of the refuge strategy for delaying insecticide resistance

The refuge strategy is used worldwide to delay the evolution of pest resistance to insecticides that are either sprayed or produced by transgenic Bacillus thuringiensis (Bt) crops. This strategy is based on the idea that refuges of host plants where pests are not exposed to an insecticide promote survival of susceptible pests. Despite widespread adoption of this approach, large-scale tests of the refuge strategy have been problematic. Here we tested the refuge strategy with 8 y of data on refuges and resistance to the insecticide pyriproxyfen in 84 populations of the sweetpotato whitefly (Bemisia tabaci) from cotton fields in central Arizona. We found that spatial variation in resistance to pyriproxyfen within each year was not affected by refuges of melons or alfalfa near cotton fields. However, resistance was negatively associated with the area of cotton refuges and positively associated with the area of cotton treated with pyriproxyfen. A statistical model based on the first 4 y of data, incorporating the spatial distribution of cotton treated and not treated with pyriproxyfen, adequately predicted the spatial variation in resistance observed in the last 4 y of the study, confirming that cotton refuges delayed resistance and treated cotton fields accelerated resistance. By providing a systematic assessment of the effectiveness of refuges and the scale of their effects, the spatially explicit approach applied here could be useful for testing and improving the refuge strategy in other crop–pest systems.

Effects of Local and Landscape Factors on Population Dynamics of a Cotton Pest

Background Many polyphagous pests sequentially use crops and uncultivated habitats in landscapes dominated by annual crops. As these habitats may contribute in increasing or decreasing pest density in fields of a specific crop, understanding the scale and temporal variability of source and sink effects is critical for managing landscapes to enhance pest control. Methodology/Principal Findings We evaluated how local and landscape characteristics affect population density of the western tarnished plant bug, Lygus hesperus (Knight), in cotton fields of the San Joaquin Valley in California. During two periods covering the main window of cotton vulnerability to Lygus attack over three years, we examined the associations between abundance of six common Lygus crops, uncultivated habitats and Lygus population density in these cotton fields. We also investigated impacts of insecticide applications in cotton fields and cotton flowering date. Consistent associations observed across periods and years involved abundances of cotton and uncultivated habitats that were negatively associated with Lygus density, and abundance of seed alfalfa and cotton flowering date that were positively associated with Lygus density. Safflower and forage alfalfa had variable effects, possibly reflecting among-year variation in crop management practices, and tomato, sugar beet and insecticide applications were rarely associated with Lygus density. Using data from the first two years, a multiple regression model including the four consistent factors successfully predicted Lygus density across cotton fields in the last year of the study. Conclusions/Significance Our results show that the approach developed here is appropriate to characterize and test the source and sink effects of various habitats on pest dynamics and improve the design of landscape-level pest management strategies.

Coregionalization analysis with a drift for multi-scale assessment of spatial relationships between ecological variables 1. Estimation of drift and random components

In this and a second article, we propose ‘coregionalization analysis with a drift’ (CRAD), as a method to assess the multi-scale variability of and relationships between ecological variables from a multivariate spatial data set. CRAD is carried out in two phases: (I) a deterministic component representing the large-scale pattern (called ‘drift’) and a random component modeled as a second-order stationary process are estimated for each variable separately; (II) a linear model of coregionalization is fitted to the direct and cross experimental variograms of residuals (i.e., after removing the estimated drifts) to assess relationships at smaller scales, while the estimated drifts are used to study relationships at large scale. In this article, we focus on phase I of CRAD, by addressing the questions of the choice of the drift estimation procedure, which is linked to the estimation of random components, and of the presence of a bias in the direct experimental variogram of residuals. In this phase, both the estimation of the drift and the fitting of a model to the direct experimental variogram of residuals are performed iteratively by estimated generalized least squares (EGLS). We use theoretical calculations and a Monte Carlo study to demonstrate that complex large-scale patterns, such as patchy drifts, are better captured with local drift estimation procedures using low-order polynomials within a moving window, than with global procedures. Furthermore, despite the bias in direct experimental variograms of residuals, good estimates of spatial autocovariance parameters are obtained with the double iterative EGLS procedure in the conditions of application of CRAD. An example with forest soil property and tree species diversity data is presented to discuss the choice of drift estimation procedure in practice.

Coregionalization analysis with a drift for multi-scale assessment of spatial relationships between ecological variables 2. Estimation of correlations and coefficients of determination

In two articles, we present ‘coregionalization analysis with a drift’ (CRAD), a method to assess the multi-scale variability of and relationships between ecological variables from a multivariate spatial data set. In phase I of CRAD (the first article), a deterministic drift component representing the large-scale pattern and a random component modeled as a second-order stationary process are estimated for each variable separately. In phase II (this article), a linear model of coregionalization (LMC) is fitted by estimated generalized least squares to the direct and cross experimental variograms of residuals (i.e., after the removal of estimated drifts). Structural correlations and coefficients of determination at smaller scales are then computed from the estimated coregionalization matrices, while the estimated drifts are used to calculate pseudo coefficients at large scale. The performance of five procedures in estimating correlations and coefficients of determination was compared using a Monte Carlo study. In four CRAD procedures, drift estimation was based on local polynomials of order 0, 1, 2 (L0, L1, L2) or a global polynomial with forward selection of the basis functions; the fifth procedure was coregionalization analysis (CRA), in which large-scale patterns were modeled as a supplemental component in the LMC. In bivariate and multivariate analyses, the uncertainty in the estimation of correlations and coefficients of determination could be related to the interference between spatial components within a bounded sampling domain. In the bivariate case, most procedures provided acceptable estimates of correlations. In regionalized redundancy analysis, uncertainty was highest for CRA, while L1 provided the best results overall. In a forest ecology example, the identification of scale-specific correlations between plant species diversity and soil and topographical variables illustrated the potential of CRAD to provide unique insight into the functioning of complex ecosystems.

The role of digital data entry in participatory environmental monitoring

Many argue that monitoring conducted exclusively by scientists is insufficient to address ongoing environmental challenges. One solution entails the use of mobile digital devices in participatory monitoring (PM) programs. But how digital data entry affects programs with varying levels of stakeholder participation, from nonscientists collecting field data to nonscientists administering every step of a monitoring program, remains unclear. We reviewed the successes, in terms of management interventions and sustainability, of 107 monitoring programs described in the literature (hereafter programs) and compared these with case studies from our PM experiences in Australia, Canada, Ethiopia, Ghana, Greenland, and Vietnam (hereafter cases). Our literature review showed that participatory programs were less likely to use digital devices, and 2 of our 3 more participatory cases were also slow to adopt digital data entry. Programs that were participatory and used digital devices were more likely to report management actions, which was consistent with cases in Ethiopia, Greenland, and Australia. Programs engaging volunteers were more frequently reported as ongoing, but those involving digital data entry were less often sustained when data collectors were volunteers. For the Vietnamese and Canadian cases, sustainability was undermined by a mismatch in stakeholder objectives. In the Ghanaian case, complex field protocols diminished monitoring sustainability. Innovative technologies attract interest, but the foundation of effective participatory adaptive monitoring depends more on collaboratively defined questions, objectives, conceptual models, and monitoring approaches. When this foundation is built through effective partnerships, digital data entry can enable the collection of more data of higher quality. Without this foundation, or when implemented ineffectively or unnecessarily, digital data entry can be an additional expense that distracts from core monitoring objectives and undermines project sustainability. The appropriate role of digital data entry in PM likely depends more on the context in which it is used and less on the technology itself.

Statistical analysis of animal observations and associated marks distributed in time using Ripley's functions

Biologists regularly collect behavioural observations of animals distributed in time from a fixed location. Examples include counts of migrating whales from an observation point (Marcoux et al. 2009), birds captured at banding stations during migration (Marra et al. 2005), measurements of fish passing through a fish ladder (Quinn et al. 1997) and characteristics of individuals captured by a camera trap (Karanth & Nichols 1998). The observer notes the time of passage of each individual. Additional characteristics might also be noted such as qualitative (sex, age group) or quantitative descriptors (size, weight) for each individual. This type of data set is called a marked linear point pattern, where each observation is traditionally referred to as an event or point and the characteristics of the observations are termed marks (Gatrell et al. 1996; Stoyan & Penttinen 2000). For the remainder of this paper, we will use the terms ‘observation’ and ‘mark’. Observations of travelling narwhals, Monodon monoceros, swimming past a fixed land location will be used through the remainder of the paper as an illustrative example of a marked linear point pattern. In this example, the observation is the passage of a narwhal and the mark is an index of age or age class of each individual. Like most whales, narwhals are a social species, such that individuals travel in groups. As a result, narwhal observations are often clustered in time (Marcoux et al. 2009). Furthermore, groups are likely to be composed of nonrandom aggregations of individuals of different ages and genders. By investigating the temporal distribution of observations and their associated marks, we gain insight into the size, composition and social organization of narwhal groups. Unfortunately, point patterns such as these are difficult to analyse statistically because observations are sporadic and marks are temporally autocorrelated. The use of classical statistical methods in this context is problematic since the presence of autocorrelation in the marks violates the assumption of independence among observations (Diggle 1990). For example, a comparison of the average age of the narwhals among different seasons requires careful attention because the age of narwhals may be correlated at short time intervals. A number of methods available for the analysis of time-series data are based on the premise that the observed pattern can be perceived as a surface pattern (e.g. * Correspondence: M. Marcoux, Department of Natural Resource Sciences, McGill University, 21 111 Lakeshore, Ste-Anne-de-Bellevue, QC H9X 3V9, Canada. E-mail address: marianne.marcoux@mail.mcgill.ca (M. Marcoux).

Embracing heterothermic diversity: non-stationary waveform analysis of temperature variation in endotherms

Recent research is revealing incredible diversity in the thermoregulatory patterns of wild and captive endotherms. As a result of these findings, classic thermoregulatory categories of ‘homeothermy’, ‘daily heterothermy’, and ‘hibernation’ are becoming harder to delineate, impeding our understanding of the physiological and evolutionary significance of variation within and around these categories. However, we lack a generalized analytical approach for evaluating and comparing the complex and diversified nature of the full breadth of heterothermy expressed by individuals, populations, and species. Here we propose a new approach that decomposes body temperature time series into three inherent properties—waveform, amplitude, and period—using a non-stationary technique that accommodates the temporal variability of body temperature patterns. This approach quantifies circadian and seasonal variation in thermoregulatory patterns, and uses the distribution of observed thermoregulatory patterns as a basis for intra- and inter-specific comparisons. We analyse body temperature time series from multiple species, including classical hibernators, tropical heterotherms, and homeotherms, to highlight the approach’s general usefulness and the major axes of thermoregulatory variation that it reveals.