Extending probability mapping of non-native species to a website toolbox and intersection with fire

Commonly observed non-native species in the United States include 2,600 plants, 600 insects, and 100 diseases. However, 8,650 non-native species are listed in the contiguous United States, 1,250 non-native species are listed in Mexico, and 1,450 non-native species are listed in Canada. Non-native species currently may not be invasive, as defined by damage, but some may become invasive in the future. Using compiled datasets for every non-native species with more than 500 observations from these lists, it is possible to map and model non-native species observed ranges and potential distributions under current and future climate in North America. Including all of North America is important for detecting species at countries bordering the United States to prevent entry and establishment as a management problem. Additionally, over time, species likely will expand poleward from more southern ranges. Maps of invasion and invasion probability at large extents and multiple temporal resolutions can be used by resource managers as an early warning and detection tool of where and when to expect non-native species to prioritize management strategies aimed at invasive species prevention and control. As an extension to a comprehensive database of maps and models of all North American non-native species under current and future climate, we will build on the comprehensive research foundation to develop interactive tools and assess intersection of invasive species with wildland fire. 

We will build a comprehensive toolbox to incorporate into the EDDMapS as an application and science exchange product for relevant and representative end users. Building upon an already well-recognized platform of data archiving and access will make the results of the project available to the 250,000 annual visitors to the EDDMapS website and their network of stakeholders and managers. The website will allow data downloads, interactive mapping, checklists, and watchlists summarized by management units. While checklists summarize detected species, watch lists are intended to guide surveillance and reporting by including species that pose a risk of introduction as well as those thought to already be introduced, but not yet reported. Consequently, watch lists do not necessarily reflect species occurrence but the potential to occur, based on modeled maps. Summary tables will be provided of recorded and predicted species detections during current and future time intervals, within a range of distances to develop early warning watch lists, as well as summaries by various administrative boundaries, by ecoregions (for example, the Great Basin), counties, and protected areas, including National Forests and National Grasslands. 

The other objective is to use our dataset of occurrences from EDDMapS and Global Biodiversity Information Facility to isolate non-native species and thresholds that may contribute to increased fires. We will modify already tested methods applied to reclassify the wildland-urban interface using fire occurrences for the continental United States and classify large wildfires in the United States by land cover. Fire perimeters will be obtained from the Monitoring Trends in Burn Severity database and Geospatial Multi-Agency Coordination database that did not overlap with the Monitoring Trends in Burn Severity fires. For this modeling, we will apply c5.0, random forests, and extreme gradient boosting classifiers to train models of fire occurrences with non-native species variables, to test for potential associations

Objectives

This project will develop a comprehensive website toolbox for comprehensive probability maps of non-native species and to assess intersection of non-native species with wildfire. 

Expected Project Results

A. Outputs:

1. A toolbox of options including data downloads, interactive mapping, checklists, and watchlists for using comprehensive probability maps of all non-native species in North America, built into EDDMapS.
2. A database of non-native species associated with fires.
3. One journal article reporting the findings of this study.
4. Conference presentations and webinar.

B. Expected Outcomes:

Research will support 40804(b)(6): Invasive Species Detection, Prevention, Eradication, and Research 

• Early Detection and Rapid Response to locate and eradicate potential invasive species – assessment of efficacy and cost-effectiveness of existing approaches; identification and assessment of potential new approaches and new technology; particular need in midwest and west coast areas. 
• Probability Mapping for Invasive Plants – build on Rocky Mountain Research Station mapping of invasive grasses to develop more comprehensive species-specific assessments of threat of territorial expansion of invasive species. 
• Intersection of Invasive Species with Wildland Fire – assessment of invasive plants and animals that are affected by or affect fire; spatial probability mapping for these species; assessment of effects of these species on land management and wildfire response options and on overall ecosystem resilience.

C. Metrics of Success:

Project success will be determined by: 

1. Website development to allow interaction with a database consisting of tens of thousands of non-native species maps by a large network of stakeholders and managers
2. Isolation of non-native species associated with fires
3. Sharing through conferences and webinars
4. Publication of journal article

Geographies and High-Risk Landscaped to be Addressed

National landscapes