Problems and pitfalls in relating climate variability to population dynamics

We discuss 3 methodological issues involved in climate-population dynamics research. Precise alternative hypotheses are the first requirement, and correlational studies are the weakest way to test the multiple working hypotheses required. Large-scale effects are difficult to investigate because of heterogeneity in environmental variables, and long-term predictions require more patience to test adequately. Multifactorial hypotheses are required for the investigation of climatic effects but these are too often vague and qualitative. Complex computer models can almost never be empirically validated and their predictions should be carefully examined. We illustrate these problems with 3 case studies: snowshoe hares in Canada, desert rodents in Arizona, and red kangaroos in Australia. Correlations abound but mechanistic understanding is limited because of long causal chains and indirect effects.