An autoregressive model of population density change in an experimental population of Daphnia magna.

The usefulness of autoregression equations as empirical models of population change in an experimental population of Daphnia magna was studied. A two-term and a ten-term autoregression equation were fit to the data. Tests of the significance of the terms of the ten-term equation were carried out, and the fit of the second-order autoregression tested. The problem of autocorrelated error terms is discussed. The ability of the autoregressions to generate the observed changes in population density of the Daphnia magna population was studied by running Monte Carlo simulations using both the two-term and the ten-term models. Both models were fairly accurate up to a period of about 60 days. The second order autoregression model appeared to be more appropriate at high densities than the ten-term model, but the ten-term model was more effective at population densities less than 80 individuals.