You are here:
SMALL POPULATIONS REQUIRE SPECIFIC MODELING APPROACHES FOR ASSESSING RISK
Citation:
Grear, J. SMALL POPULATIONS REQUIRE SPECIFIC MODELING APPROACHES FOR ASSESSING RISK. Presented at The Society of Environmental Toxicology and Chemisrty Annual Meeting and 4th Chemistry World Congress, Portland, OR, November 14, 2004.
Description:
All populations face non-zero risks of extinction. However, the risks for small populations, and therefore the modeling approaches necessary to predict them, are different from those of large populations. These differences are currently hindering assessment of risk to small populations from the combined effects of habitat loss and chemical stressors. A contributing factor to this state of affairs is the contrast in the size of populations usually targeted by the traditionally separate disciplines of conservation biology and ecotoxicology. In the case of large populations, stressor effects on individuals can sometimes be successfully extrapolated by applying a single average effect (e.g., 10% mortality) to an entire population or demographic class. This is rarely adequate for small populations, simply because births and deaths are discrete (non-continuous) binary events. That is, when a survival probability is applied to n individuals as n discrete events, the expected number of survivors cannot be easily approximated by continuous probability models (e.g., normal distribution) if n is small. The usual analogy is a coin toss: three flips of a coins can give a success rate of 33%, 67%, or 100%, none of which is close to the expected success rate of 50%; a larger number of tosses would probably come closer to the expectation. The collective effect of these individually varying discrete events on population size is known as demographic stochasticity and can be a factor in both constant and fluctuating environments. An important practical implication of demographic stochasticity is that the application of protection criteria at the aggregate population level (e.g., percent impairment) may be wholly inappropriate for rare or endangered species. This and other implications of demographic stochasticity will be discussed, along with practical examples of how they have been addressed in ecological risk assessment.