Predicting vulnerability of forest patches to invasion by non-native plants for landscape scale management.

As a leading cause of forest health degradation, non-native invasive plant species are a key focus for many forest management and conservation efforts. These efforts come at a high price for resource-limited agencies and organizations making cost-effectiveness an important objective of invasion response plans. In this paper, we present an approach to guide the prioritization of locations for invasion management using species distribution models that account for the non-equilibrium of invasive species distributions and use readily available land use data as the primary explanatory variables. This approach takes advantage of the relatively high spatial resolution, as well as the broad, continuous geographic coverage, of land use data to provide results at a landscape scale relevant to practitioners responsible for invasive species management. In our example from northern Virginia, we simultaneously modeled a suite of invasive plant species to identify common indicators of invasion. We found that the proportions of surrounding non-forested land use types (grasses, crops, and development) were the most common and strongest indicators of invasion risk. These outcomes can guide managers of large protected areas to focus on major divides between forest and non-forest land over linear disturbances. We also found useful species-specific traits that can inform specific management actions. Additionally, we demonstrate through a case study how organizations that manage multiple smaller properties can take advantage of the projected distribution maps when considering acquiring or administering properties.

Scale-dependent impacts of urban and agricultural land use on nutrients, sediment, and runoff.

We coupled a spatially-explicit land use/land cover (LULC) change model, Dinamica EGO, (Environment for Geoprocessing Objects), with the Chesapeake Bay Watershed Model (CBWM) to project the impact of future LULC change on loading of total nitrogen (TN), total phosphorous (TP) and total suspended solids (TSS) as well as runoff volume in the watersheds surrounding Virginia's Shenandoah National Park in the eastern United States. We allowed for the dynamic transition of four LULC classes, Developed, Forest, Grasses (including both pasture and hayfields) and Crops. Using 2011 as a baseline scenario and observed differences in LULC between 2001 and 2011, we estimated the temporal and spatial patterns of LULC change as influenced by physiographic and socio-economic drivers 50 years in the future (2061). Between transitions of the four LULC classes, the greatest absolute change occurred between the gain in total Developed land and loss in total Forest. New Developed land was driven primarily by distance to existing Developed land and population density. Major findings on the effect of LULC change on watershed model outputs were that: the impact of LULC change on pollutant loading and runoff volume is more pronounced at finer spatial scales; increases in the area of Grasses produced the greatest increase in TP loading, while loss of Forest increased TN, TSS, and runoff volume the most; and land-river segments with a greater proportion of Developed or a smaller proportion of Forest in the 2011 scenario experienced a greater change in runoff than other land-river segments. Results of this study illustrate the potential impact of projected LULC change on nutrient and sediment loads which can adversely impact water quality. Studies like this contribute to a broader understanding of how ecosystem services such as fresh water respond to LULC change, information relevant to those in planning and watershed management.