Individual-Based Ecological Modeling at Humboldt State University

Research on the use of individual-based models (IBMs) for applied and theoretical ecology is affiliated with the Mathematical Modeling Program, HSU Mathematics Department. This research is a collaboration of mathematicians, ecologists and biologists, environmental engineers, and software professionals. See below for our research goals.

What's New

• Short course on teaching individual- and agent-based modeling. Each summer, we teach a one-week short course intended to help instructors, including those new to individual-based modeling, develop and teach courses at the university level. The course is based on Agent-based and Individual-based Modeling: A Practical Introduction by Steve Railsback and Volker Grimm, the first textbook devoted to individual-based (or "agent-based") modeling in science. The book and short course are designed for instructors as well as students who are new to this kind of modeling. They are not specific to ecology, but general to all scientific fields. More information about the book is at its web site. The next course will be July 31 - August 4, 2017. Additional information is at the course web site. If you are interested in future courses, please contact Steve Railsback.
  
  
• Summer short courses in individual- and agent-based modeling. Our friends at the Institute of Forest Growth and Forest Computer Sciences of Technical University Dresden, Germany, conduct intensive summer courses in individual- and agent-based simulation modeling. The courses are intended primarily for graduate students, and taught in English at a lovely country location near Dresden. See the course web site for information.
  
  
• New individual/agent-based modeling and NetLogo interest group at Humboldt State. HSU faculty and staff, grad students, and advanced undergraduates interested in using NetLogo for individual-based modeling are encouraged to participate in our new interest group. We meet approximately biweekly to help beginners get started, solve problems with more experienced users, and pass on new discoveries. Contact Steve Railsback if you are interested.
  
  
• New model of river management effects on breeding success of foothill yellow-legged frogs. This individual-based model simulates how river flow and temperature regimes (at daily or shorter time steps) affect this unique frog species that breeds along river margins. Modeled processes include timing of egg laying, incubation timing and survival of egg masses, and survival and timing of tadpole development. The first publication using the model (Publications and Products page) addresses potential conflicts between frog breeding and river management for salmonids, using sites on the Trinity River.
  
  
• inSTREAM in NetLogo ?? !! with evolution!! Our good friend Daniel Ayllón paid an extended visit to Humboldt State in 2013, and then re-programmed inSTREAM in the popular agent-based modeling platform NetLogo. Daniel carefully tested his new software (and ours) by comparing the two implementations and eliminating the discrepancies. The NetLogo version (pictured here; click on the picture for a larger version) makes it possible for people lacking C programming skills to make changes in inSTREAM; however, Daniel's version currently can only use rectangular habitat cells. Dr. Ayllón then added genetic evolution in several traits of trout to the model and hence calls it inSTREAM-GEN. His first study using the model (Publications and Products page) examines ecological and evolutionary responses of trout to land use and climate change in the Spanish Pyrenees.
  
  
• Publication addressing cumulative effects of river flow fluctuations and exotic species on an endangered warmwater fish, by former MS student Tyler Belarde and Steve Railsback. This paper is our contribution to a special issue of Ecological Modelling that commemorates the 70th birthday of IBM pioneer Donald DeAngelis. Tyler used a model in which cumulative effects emerge from individual variation in size, competition for food, and predation by fish on each other. The piscivory component of the model is based on the very early IBM by DeAngelis et al. (1980) of fish cannibalism. The study is unique in simulating over 1000 sets of input so that the type of cumulative effect (independent, antagonistic, synergistic) could be related to habitat type. Tyler found that habitat disturbance (flow fluctions) could suppress the effects of the exotic species, but only under limited conditions. See the (Publications and Products page).
  
  
• Brooke Penaluna of Oregon State University used inSTREAM to examine cumulative effects of forestry and climate change on trout populations and has several publications resulting from her work. See the (Publications and Products page).
  
  
• Effects of land use on bird populations and pest control services on coffee farms published in Proceedings of the National Academy of Sciences. This new paper analyzes the coffee farm IBM developed to understand Matt Johnson's field sites in Jamaica. In this system, the availability of different habitat types (forest, trees, shade- and sun-grown coffee) determines bird population densities, which determine how well birds control an insect pest that infests coffee. We found that patches of trees, or a mix of shade-grown coffee, support bird populations well enough to provide economically important pest control; separate areas of intact forest and coffee monoculture did not. See the Publications and Products page.
  
  
• Review of foraging theory that can reproduce trait-mediated indirect interactions and non-consumptive effects emerge. TMIIs and NCEs are exciting areas of ecological research because they illustrate the effects of individual adaptive behavior on population and community dynamics. But how can we model the individual foraging decisions from which these complex trophic dynamics emerge? These decisions must trade off food intake vs. predation risk, when the alternatives available and the intake and risk associated with them are subject to feedbacks from how the individuals behavior. Trait-mediated trophic interactions: is foraging theory keeping up? is our new review of this subject in the journal Trends in Ecology and Evolution. It shows how the foraging theory we use in our trout models can reproduce a wide range of observed TMIIs and NCEs. See the Publications and Products page.
  
  
• Facultative anadromy in salmonids: linking habitat, individual life history decisions, and population-level consequences published in Canadian Journal of Fisheries and Aquatic Sciences. How useful are state-based dynamic models of whether an individual juvenile salmonid should migrate to the ocean or remain to mature in freshwater (e.g., Satterthwaite et al., 2010. State-dependent life history models in a changing (and regulated) environment: steelhead in the California Central Valley. Evolutionary Applications 3: 221-243) for making management decisions? How do habitat conditions that affect growth and survival therefore affect the number of smolts produced by a spawning stream? We addressed these and other questions using inSALMO-FA, our new individual-based model for 'facultative anadromous' species such as steelhead/rainbow trout, in which some but not all individuals migrate to the ocean. A pre-publication version is available at the Publications and Products page.
  
  
• Masters theses that use individual-based models. Ryan Baumbusch of HSU's Wildlife Department completed a thesis based on his IBM of the spatial demography of barred owls and how alternative control strategies might limit the spread of barred owls into spotted owl habitat. Elizabeth Arnold built an IBM of how wild dogs (Lycaon pictus) disperse among game reserves in South Africa. The model imported GIS data on landscape features (roads, towns, etc.) and simulated their effect on dog movement and survival. Kyle Falbo built a model of pelagic ocean dispersal of juvenile Hawksbill sea turtle (Eretmochelys imbricata) that hatch on the Big Island of Hawaii. Kyle's model uses satellite data to examine the effects of both ocean currents and turtle behavior on their location in their first 3 years. In collaboration with Argonne National Laboratory, Tyler Belarde used a model of the juvenile Colorado pikeminnow (Pytchocheilus lucius) in backwaters on the Green River, Utah, to simulate the interacting effects of flow alteration (daily power-generation cycles) and exotic species (an introduced fish that competes with and preys on pikeminnow) on survival and growth of the endangered pikeminnow. Math Modeling MS theses are available here.
  
  
• inSTREAM training opportunities. We occasionally provide training in our inSTREAM individual-based trout model. Classes are usually but not always here in Arcata. Contact Steve Railsback if you are interested.
  
  

Research Goals:

• Developing a conceptual and theoretical basis for individual-based ecology. Differential calculus provides the conceptual basis for classical ecological models, but IBMs lack such a basis. We find the new science of Complex Adaptive Systems to be a fruitful source of concepts for thinking about and designing IBMs.
• Applying fish IBMs to river management issues. We developed several generations of stream salmonid IBM to address such management questions as: How do the magnitude and timing of instream flow releases affect populations? What are the cumulative effects of changes in flow and temperature on fish populations? What effects do habitat alterations (e.g., loss of pools, increased turbidity), competition, predation, and habitat connectivity have on population dynamics?
• Using IBMs to test and develop ecological theory. We use IBMs as 'virtual ecosystems' for testing ecological theory and linking behavioral ecology to population ecology.
• Developing software and software engineering approaches for IBMs. Software engineering is much more important for IBMs than for other ecological models. We develop software engineering guidance, flexible and re-usable code for stream fish models, and a library of Swarm-based code for IBMs.
• Integrating individual-based approaches in ecological and modeling education. We integrate IBMs in ecological modeling courses at HSU, and gladly offer software and guidance to instructors at other institutions.

Two-dimensional hydraulic simulation in version 5.0 of inSTREAM. Habitat cells are shaded by water depth (alternatively, by velocity); pink dots indicate juvenile fish, and rectangles indicate adult trout.

WebMASTER:  steven.railsback@humboldt.edu
Last Updated:  October 2016