Biological populations rarely exist in perfectly regular distributions.  Because of gradients in habitat quality, competition, predation, and attraction to conspecifics, we expect fish to be unevenly distributed in streams. The way fish distribute throughout available habitat can tell us a lot about the health of the population. When abundance is low fish tend to congregate in the best habitats, but when abundance is high and competition gets more intense fish tend to spread out to more marginal habitats and thus the proportion of sites occupied is higher. Persistent changes in distribution can also alert managers to problems such as habitat deterioration or loss of genetic diversity.

If individuals are clustered in small areas they are at greater risk of localized catastrophic events. Decreased distribution may also reduce diversity and limit opportunities for the recue effect. Because of this, we want to be able to detect instances where most of the individuals of a population are clustered in a small area. Though it’s easy to understand that intuitively, it proves to be somewhat difficult to quantify spatial distribution or to draw the line when we would say that a population isn’t distributed broadly enough.

There are lots of measures of distribution in ecology. However most of them assume we have points in two-dimensional space whereas freshwater fish inhabit networks of streams. The fact that our data are distributed in one dimension rather than two presents some challenges for geostatistical modelling.

The REDD group is currently exploring three aspects of the distribution of freshwater species:

• We are evaluating several different measures of spatial distribution in streams.
• We are exploring the use of spatial simulation models to evaluate the scales at which distribution might be important.

We are examining relationships between distribution and abundance to see if distribution could serve as a proxy for abundance estimates.

Julie Firman