Print PageAll invasive species must successfully invade and establish before they can spread. Early detection is key to limiting the spread of invasives, but can be particularly difficult for invasive microbes. The microbiologist's credo 'Everything is everywhere, but, the environment selects,' infers that microbial invasions are relatively continuous with success of population establishment dependent on environmental conditions. Therefore, environmental factors should be instrumental in predicting invasion success and establishment for microbial invaders. Prymnesium parvum, a toxigenic marine unicellular eukaryote, has invaded many freshwater systems throughout the southern and southwestern US. Like many invading species it can potentially wreak havoc in the invaded system, in this case, leading to massive fish kills during blooms. We've been monitoring P. parvum abundances in Lake Texoma (OK-TX) since the 2004 invasion and recently adopted qPCR as the method of choice due to reduced processing time, error, and detection limits. Using a 3-year data set (N=377 observations) we constructed a predictive model relating P. parvum presence or absence to nutrient concentrations and conductivity. We then used this model in conjunction with environmental samples to predict presence and absence of P. parvum throughout the Red River watershed (N=62) and the adjacent Canadian River watershed to the north (N=26).
Results/Conclusions
Our predictive model accurately classified the presence or absence of P. parvum in Lake Texoma for 82% of the samples. Applying this model to the adjacent watersheds also showed good predictive power, correctly classifying 88% of the sites sampled within the Red River watershed and 81% of the sites sampled in the Canadian River watershed. Because of the high accuracy of our model, predictions of P. parvum presence in the watersheds in which none was detected by qPCR indicates that these sites may be conducive to invasion and establishment by P. parvum, and thus should be monitored for future invasion success. Misclassifications by the model of sites in which P. parvum was detected by qPCR suggests that dispersal has occurred but that the environmental conditons were not conducive to population establishment. Indeed, in these sites, P. parvum abundances, when detected, where very low. While we can not rule out dispersal limitation as a major factor involved in the biogeography of P. parvum, our results do indicate that the establishment and spread of this harmful algal species appears to be limited by environmental conditions in the invaded habitat.
