ESTIMATING DISPERSAL FROM PATTERNS OF SPREAD: SPATIAL AND LOCAL CONTROL OF LAKE INVASIONS

Abstract

The spread of exotic species can be limited by dispersal or by constraints imposed by the local environment. Using data collected from 152 Missouri (USA) lakes over seven years, we asked whether models based on dispersal or local-scale processes best predicted invasion by the exotic cladoceran Daphnia lumholtzi. We used multiple logistic regression to test the relative importance of 10 local physicochemical features and proximity to all known potential source populations for predicting which lakes were invaded. The decline in invasion likelihood with distance to source populations was used to estimate the shape of the dispersal kernel. Between 1992 and 1998 the cumulative prevalence of D. lumholtzi increased from 6% to 34% of lakes sampled, with frequent appearances of populations in new watersheds. Spatial position and physical factors were both important for predicting the new colonization events. The probability of colonization increased with lake surface area and epilimnetic temperature, declined with increasing conductivity, and was unaffected by variation in lake fertility. Invasion likelihood declined sharply as a nonlinear function of distance to source populations up to around 30 km, and was relatively constant at greater distances. The results suggest that dispersal and local abiotic constraints jointly limit the spread of D. lumholtzi. This approach illustrates how range expansion can be used to estimate dispersal rates at broad spatial scales.