Selected Publications

Omnivory is extremely common in animals, yet theory predicts that when given a choice of resources specialization should be favored over being generalist. The evolution of a feeding phenotype involves complex interactions with many factors other than resource choice alone, including environmental heterogeneity, resource quality, availability, and interactions with other organisms. We applied an evolutionary simulation model to examine how ecological conditions shape evolution of feeding phenotypes (e.g., omnivory), by varying the quality and availability (absolute and relative) of plant and animal (prey) resources. Resulting feeding phenotypes were defined by the relative contribution of plants and prey to diets of individuals. We characterized organisms using seven traits that were allowed to evolve freely in different simulated environments, and we asked which traits are important for different feeding phenotypes to evolve among interacting organisms. Carnivores, herbivores, and omnivores all coexisted without any requirement in the model for a synergistic effect of eating plant and animal prey. Omnivores were most prevalent when ratio of plants and animal prey was low, and to a lesser degree, when habitat productivity was high. A key result of the model is that omnivores evolved through many different combinations of trait values and environmental contexts. Specific combinations of traits tended to form emergent trait complexes, and under certain environmental conditions, are expressed as omnivorous feeding phenotypes. The results indicate that relative availabilities of plants and prey (over the quality of resources) determine an individual’s feeding class and that feeding phenotypes are often the product of convergent evolution of emergent trait complexes under specific environmental conditions. Foraging outcomes appear to be consequences of degree and type of phenotypic specialization for plant and animal prey, navigation and exploitation of the habitat, reproduction, and interactions with other individuals in a heterogeneous environment. Omnivory should not be treated as a fixed strategy, but instead a pattern of phenotypic expression, emerging from diverse genetic sources and coevolving across a range of ecological contexts.
Ecology and Evolution,2014

Estimating colony areas, locations, population sizes, and trends, are all important aspects of managing animal populations. The ability to assess population trends and delineate important wildlife areas remains a top priority for managers and conservation biologists. Yet, outdated labourious estimation methods remain in high use. By simulating known populations on known island sizes and using established transect and quadrat survey methods we asked whether using inverse distance weighting (IDW) interpolations in ArcGIS improved estimates of colony area and population size for nocturnal burrow-nesting seabirds over conventional global interpolation methods. We performed 100 simulations for each of three population sizes (500, 1000, and 50 000 breeding pairs) on three island sizes (10 ha, 50 ha, and 500 ha), excluding the largest population size on the smallest island size, for a total of 800 simulated islands. We estimated colony area and population size for each simulated island using both IDW interpolations and an established global interpolation method. Accuracy of each estimate was then calculated and using an information theoretic approach we found that IDW interpolation estimates were overall more accurate when estimating population size but we found no difference in colony area accuracy between interpolation methods. We recommend using IDW interpolations to estimate colony area and population size along with consistency in survey structure both among study sites and years. We also recommend maintaining a consistent transect length whenever possible to ensure observer bias does not influence areas surveyed.
Marine Ecology Progress Series,2012

The link between individual habitat selection decisions (i.e., mechanism) and the resulting population distributions of dispersing organisms (i.e., outcome) has been little-studied in behavioural ecology. Here we consider density-dependent habitat (i.e., host) selection for an energy- and time-limited forager: the mountain pine beetle (Dendroctonus ponderosae Hopkins). We present a dynamic state variable model of individual beetle host selection behaviour, based on an individual’s energy state. Field data are incorporated into model parameterization which allows us to determine the effects of host availability (with respect to host size, quality, and vigour) on individuals’ decisions. Beetles choose larger trees with thicker phloem across a larger proportion of the state-space than smaller trees with thinner phloem, but accept lower quality trees more readily at low energy- and time-states. In addition, beetles make habitat selection decisions based on host availability, conspecific attack densities, and beetle distributions within a forest stand. This model provides a framework for the development of a spatial game model to examine the implications of these results for attack dynamics of beetle populations.
Ecological Modelling,2009

Recent Publications

. State-dependent domicile leaving rates in Anopheles gambiae. Malaria Journal, 2018.

PDF Project

. On the evolution of omnivory in a community context. Ecology and Evolution, 2014.

PDF Project

. To tree or not to tree: The role of energy limitation on host tree acceptance in a bark beetle. Evolutionary Ecology Research, 2014.

PDF Project

. Estimating colony and breeding population size for nocturnal burrow-nesting seabirds. Marine Ecology Progress Series, 2012.

PDF Project

. A Theoretical Approach to Study the Evolution of Aggregation Behavior by Larval Codling Moth, Cydia pomonella (Lepidoptera: Tortricidae). Journal of Insect Behavior, 2011.

PDF Project

. A dynamic host selection model for mountain pine beetle, Dendroctonus ponderosae Hopkins. Ecological Modelling, 2009.

PDF Project



  • fpCompare: Reliable comparison of floating point numbers. CRAN; GitHub; Website
  • grainscape: Efficient modelling of landscape connectivity, habitat, and protected area networks. CRAN; GitHub; Website
  • NetLogoR: A Port of ‘NetLogo’ Functions and Language to R. CRAN; GitHub; Website
  • quickPlot: Develop and run spatially explicit discrete event simulation models. CRAN; GitHub; Website
  • reproducible: Develop and run spatially explicit discrete event simulation models. CRAN; GitHub; Website
  • SpaDES: Develop and run spatially explicit discrete event simulation models. CRAN; GitHub; Website
  • SpaDES.addins: Tools and RStudio addins for ‘SpaDES’ and ‘SpaDES’ module development. CRAN; GitHub; Website
  • SpaDES.core: Core functionality for Spatial Discrete Event Simulation (SpaDES). CRAN; GitHub; Website
  • SpaDES.shiny: Utilities for building shiny-based apps for SpaDES simulations. GitHub; Website
  • Additional modelling tools for Spatial Discrete Event Simulation (SpaDES) module development. CRAN; GitHub; Website


  • miniCRAN: Create a mini version of CRAN containing only selected packages. CRAN; GitHub

Recent Posts

More Posts

Just noticed the Queen is listed as an author in this (and probably other) R packages #rstats — David Smith (@revodavid) September 7, 2017 A number of people have been surprised to learn that Her Majesty the Queen is listed as an author on a number of our packages. For example, in the SpaDES package: Author: Alex M Chubaty [aut, cre], Eliot J B McIntire [aut], Yong Luo [ctb], Steve Cumming [ctb], Her Majesty the Queen in Right of Canada, as represented by the Minister of Natural Resources Canada [cph] Note, however, that Her Majesty’s role not as a package author (aut) but as the copyright holder (cph).


I wanted to play around with the new sf package, which requires the latest GDAL (>= 2.0.0), GEOS (>= 3.3.0), and PROJ.4 (>= 4.8.0). However, the version of GDAL installed via brew is 1.11.4, so I needed to update to the latest version and reinstall a few R packages in order to get sf to work on macOS. Update GDAL ## unlink the previous GDAL version brew unlink gdal ## update GDAL to the latest version (2.


In preparing a presentation on developing R packages using RStudio for the Victoria R Users Meetup Group this month, organizer Kiri Whan and I put together a very simple demo R package. UPDATE 2015/02/09: I didn’t notice the conflict with the package name; thanks Scott Chamberlain for pointing it out. I’ve renamed the package to meow and updated the links below. From the README at meow: Get random pictures of cats.


This is actually really easy to do, but most of the google hits I came across were old (from 2010) or horribly complex (building gdal and proj4 from source then building rgdal itself). First, this assumes you already have homebrew installed 1. If not, see for the one-liner terminal install. Next, install gdal: brew install gdal Then open RStudio (for some reason it didn’t work using R in the terminal…) and install the package from source:


Whenever I used to sit down in front of the computer to write anything other than an email, I would immediately open Microsoft Word and start clacking away on the keyboard. When I switched operating systems (away from Windows) I began using Open Office Writer and even now I still use Libre Office Writer for some of my writing needs. Each of these word processors offer similar sets of tools and can effectively be used for a wide range of writing tasks, from letters to essays, yet I feel most people use Word not because it’s always the best tool for the job, but because it’s ubiquitous and familiar.



Habitat Connectivity

We have recently reimplemented and updated the grainscape package for R for use in landscape connectivity, habitat, and protected area network analyses.

Boreal forest ecosystem forecasts using integrated dynamic simulation models

Working with others in the McIntire lab at the Pacific Forestry Centre, we are working on integrating various simulation models dealing with forest vegetaiton dynamics, fire, insect disturbance, and wildlife populations, to inform decision making in these management areas.

Simulating mountain pine beetle eastward spread

The continued eastward spread of mountain pine beetle (Dendroctonus ponderosae Hopk.; MPB) now threatens the boreal forests of eastern Alberta, Saskatchewan, and beyond. Predicting the outbreak and spread dynamics of this insect in jack pine, and to evaluate control measures to mitigate a potentially devastating loss of forest habitat and timber supply requires not only a complex understanding of the various inputs to this system and their interactions (e.g., MPB population dynamics, climate impacts, landscape features) but also the technical capacity to run large-scale spatial simulation models, and to update them quickly as new data are acquired and new models are developed. …

SpaDES: Spatial Discrete Event Simulation

A collection of R packages for implementing a variety of event-based models, with a focus on spatially explicit models. These include raster-based, event-based, and agent-based models.

Evolution of omnivory in a community context

We developed an evolutionary simulation model to consider the feeding strategies (broadly classified as three feeding types: herbivores, omnivores, and carnivores) of a community of individuals and explore how the intrinsic properties of these foragers and extrinsic characteristics of their environment determine the prevalence of omnivores and other feeding types.

PhD thesis: Individual host selection decisions and population-level responses in a time- and energy-limited forager, Dendroctonus ponderosae Hopkins

My doctoral work involved the development of analytical and simulation models of mountain pine beetle (MPB) habitat selection and dispersal behaviour, and the resulting population distributions and dynamics. Specifically, I considered the role of energy- and time-limited host search in MPB, and I developed individual-based state-dependent dynamic game models of beetle host selection.