Dr. Kristan has worked on a variety of topics in animal and plant ecology. His research mixes field ecology, landscape ecology, and theoretical population biology. Previous student projects have included studies of freeloading ravens and crows at the San Diego Safari Park, characteristics of Mexican fan palm invasion of riparian areas of San Diego County, and effects of water supplementation on bird and mammalian predator distributions in local shrublands. His current work uses individual-based simulation modeling to understand how complex environments affect gene flow, habitat selection, and population dynamics. Mathematical and statistical projects include improving the predictive power of statistical habitat models, and modeling the effects of habitat-induced correlations in demographic population models. Most of Dr. Kristan's work is aimed at topics with applications in conservation biology.

• BS: Wildlife 1989: Humboldt State University
• MS: Wildlife 1995: Humboldt State University
• Ph.D. Biology 2001: University of California, Riverside

• Biol 212 - Evolution
• Biol 215 - Experimental Design and Statistical Analysis
• Biol 365 - Computing Skills for Biologists
• Biol 463/663 - Principles of Conservation Biology
• Biol 420/620 - Ecological Monitoring
• Biol 531 - Biological data analysis I: Linear Models
• Biol 532 - Biological data analysis II: Multivariate Analysis
• Biol 533 - Landscape Ecology

Current work with individual-based models

My work has often blended theoretical work and applied, conservation-oriented studies. Recently, in order to try to better bridge the gap between real populations of plants and animals and theoretical models, I have started to use individual-based simulation models that simulate the behavior of animals in realistic environments. These models have the advantage of omniscience (i.e. I know what they are doing and why), time dilation (i.e. they can be run for arbitrarily long periods of time), and omnipotence (i.e. I can add or remove components in ways that are not possible in the real world to see how strongly results depend on particular properties of the organisms). Because individuals are simulated, the results of these models are analyzed in ways that are very similar to studies of real populations, but with fewer unknowns than in a typical field study.

Two examples of these models are:

Effects of barriers and road mortality on gene flow in a territorial vertebrate

Road mortality is a major concern for species such as the threatened desert tortoise (Gopherus agassizii), both as a risk to population persistence and as a potential barrier to gene flow. Building barrier fences along roads is a preferred method of preventing road mortality, but the potential for barriers with insufficient numbers of safe crossings to reduce gene flow has also been a concern. Desert tortoises are territorial, and their poplation densities decline near roads. The potential for preferential dispersal by juveniles into the lightly-populated areas around roads to offset loss of gene flow due to road mortality would be difficult to study in the wild, but is simple to simulate.

Using the NetLogo platform, a novel allele can be introduced into a population at one side of the simulated world, and its spread over time can be observed. In the center can be either nothing (as a control), a road as a source of mortality, or a barrier with some number of gaps through which the simulated tortoises can pass (from none to as many as are needed to allow the barrier to converge on on the control).

Looked at in this way, the diffusion of the gene across the space is affected by the road, but movement by the juveniles to the nearest road side, and away from the opposite side, is facilitated by the lower densities of territorial adults. The novel allele is able to arrive at the far side of the world in the same amount of time as the control, because the effects of reduced population density on movement by juveniles offsets the reduced rate of movement across the road. This effect is dependent on territoriality by adults towards juveniles, and when the juvenile avoidance of adults is removed from the model the movement of the novel allele is unaffected by the presence of the road.

Habitat selection in complex environments

Selection of breeding habitat by species such as birds is a complex forecasting problem, because they must choose places to build their nests and lay their eggs early in the breeding season. By the time the food demands of chicks are at their peak conditions can change to the point that what appeared to be good habitat proves to be unsuitable. Additionally, different environmental characteristics can be important for different parts of an organism's life history, and the ideal conditions for adult survival may not coincide with ideal conditions for reproduction.

We are using simulation models to explore how individual quality and habitat quality combine to produce the patterns of distribution of organisms we find in nature. Models can be constructed in which the best individuals (i.e. those with the greatest fitness potential intrinsically) are able to select the habitat they prefer, based on the information available to them.

If the environment is very simple, the best individuals (red symbols) can occupy the globally best habitat (light colored background pixels) by following gradients in the environment.

In contrast, in more complex environments, clusters of individuals form around local maxima, if individual search radii are too short to detect better habitat that is further away. By using more realistic underlying habitat gradients we can explore what we can know about the habitat selection process based on the patterns of distribution of the organisms.

If you are interested in simulation modeling of animal/habitat associations feel free to contact me. I am interested in both undergraduate and graduate students who would like to participate in, and expand on, these topics.