Socioeconomics drive invasive woody plants in New England through forest fragmentation

Background/Question/Methods

The distribution of invasive plants in New England may be mediated by climate and landscape characteristics.  Our goals were to assess: 1) change in forest fragmentation in New England, 2) the association of fragmented forest types with invasive woody plant richness, and 3) the extent to which socioeconomic and other factors can explain land use/land cover (LULC) and forest fragmentation patterns.  We created a thematically consistent LULC time series (1992, 1996, 2001, 2005/6) for New England that was used to classify forests into four classes: perforated, patch, edge, and core.  Invasive plant distribution points from the Invasive Plant Atlas of New England, physical factors, such as roads and elevation, and socioeconomic factors from the U.S. Census provided the necessary data.  We used spatially-explicit Bayesian Poisson models to estimate the effect of fragmented forests in the landscape around known invasive species locations.  We also developed novel, spatially-explicit, Hierarchical Bayesian compositional data models to demonstrate simultaneously the relationships between all LULC or forest types and socioeconomic and physical factors. 

Results/Conclusions

We observed a New England-wide decline in total (-5.3%) and core (-8%) forest and increases in edge forest (1.6%), developed land (0.7%), and scrub/shrubland (3.6%) from 1992 to 2006, with the majority of change occurring in northern New England.  Spatially explicit, Hierarchical Bayesian Poisson models demonstrated that woody invasive plant richness was higher in edge forest relative to patch, perforated, and especially core forest types. Using novel, spatially explicit, Hierarchical Bayesian compositional data models we showed that physical factors, including road density and elevation range, and time-lagged socioeconomic factors, primarily population density, help drive development and forest fragmentation patterns. Our socioecological approach can help guide early detection and management efforts for invasive plants and highlights the critical role humans play in the regional invasion process.