Projecting invasive species distributions as defined by the net reproductive rate

Background/Question/Methods Species distribution models are used to project the distribution of invasive species using occurrence records matched to environmental covariates and extrapolating this relationship to uninvaded landscapes. This extrapolation is used to delimit the geographical extent to which a new invasive species may become established. However, the link between occurrence and successful establishment presumes a relationship with environmental covariates and the organism’s life history that allows for survival and population growth. Here we make an explicit connection between the environment and life history by formulating the net reproductive value (R0) of a quagga mussel’s (Dreissena rostriformis bugensis, Andrusov 1897) life history graph as a function of calcium concentration. The net reproductive value is solved using a graph reduction approach and parameter values are taken from the literature and estimated from experiments. We then project the distribution of quagga mussels in the Western US based on Kriging of observed lake calcium data. 

Results/Conclusions

The R0 projections indicate that quagga mussels have a greater range of establishment than previously predicted by occurrence based species distribution models.  This is due to input functions from experiments showing survival and reproduction at lower calcium levels than previously expected.  Of the life history parameters used, the greatest uncertainty is in fecundity The R0 projections are sensitive to fecundity, yet the predicted range of quagga mussel establishment remains largely unchanged. Using the limited number of observed Western US quagga mussel establishments and detections, we evaluate the performance of the R0 projections and show no quagga mussel invasions have occurred where R0<1 and many of the presumably failed establishment events occured in locations where R0 is close to 1 relative to established locations where R0>>1. The R0 approach to species distribution modeling provides another avenue, more directly connected to life history and population biology, to better predict the likely extent of invading species.