Predicting an Invasive Species’ Potential Range Expansion and Invasion Potential

Research Issue

Photos of four invasive forest species used in the models: Asian gypsy moth, Asian longhorned beetle, Dutch elm disease, and sudden oak death.The world’s forests face unprecedented threats from invasive forest pests that can cause largely irreversible damage to the environment. The number of new introductions and interceptions of forest invasive species is escalating at an alarming rate. Moreover, native forest insects (e.g. bark beetles) are expanding their traditional ranges and are becoming novel pests, causing substantial damage to forest resources. This threatens the forest’s capacity to provide long-term economic, ecological and social benefits that range from fiber supply, jobs, carbon storage, nutrient cycling, water and air purification, soil preservation and maintenance of wildlife habitat.

The key to reducing this risk lies in vigilant efforts at detecting insects (bio-surveillance) to increase preparedness and facilitate early interventions through robust interception and surveillance frameworks. Unfortunately, bio-surveillance programs are costly and require regional prioritization. Developing efficient detection systems for bio-surveillance of invasive species has become an urgent task.

A couple of factors make developing models for early detection of invasive species challenging. For example, one of the critical assumptions in the models is that the species is in equilibrium with its environment in the area that is used for developing the model, but invasive non-native species are typically in disequilibrium with their environment.

Data for invasive species are also prone to two forms of sampling bias, which can produce poorly predictive models. First, invasive species are often unevenly surveyed (intense survey in new areas and poorly surveyed in native or already naturalized regions). Second, short sampling time frames in invaded areas reduce the length of time the model is predictive and require the models to be adaptive and updated with new data from novel environments often to ensure their continuing usefulness.

Our Research

1) We used different bio-surveillance approaches for surveying and monitoring areas under threat of invasion to aid in finding infestations as soon as possible. In this work we employed different types of models that are designed to predict the distribution of a species in a new area: maximum Entropy (MaxEnt), Genetic Algorithm for Rule-set Prediction (GARP), and the Ordination Method. Each of these models uses the location data for the species, in both their native and introduced habitats, and predictive climatic or other variables specific for the species to estimate potential ranges they could invade or disperse into.

2) We studied the distribution of a variety of invasive species to identify potential for range expansion and invasion potential. To do this we compared the niche the species occupied in its native and introduced habitats to see if it had changed or if they had dispersed into the full area they could use.

Expected Outcomes

The goal of this research is to develop models that can predict the potential range of invasive species in invaded areas. The information on climatic niche expansion and other important niche characteristics can prove to be a useful cost-effective tool in the decision-making on managing and monitoring representative invasive species future spread and currently infested areas. Scientists can also use this information along with more spatial and temporal predictive models, to predict the spread dynamics of invasive species in the infested range, and to design short and long-term management strategies of the invasives.

Research Results

Finding the Best Model to Predict Invasive Species Potential Range

We assessed the potential distribution of Asian gypsy moth in Canada using two presence-only species distribution models, Maximum Entropy (MaxEnt) and Genetic Algorithm for Rule-set Prediction (GARP). MaxEnt had higher model fit and sensitivity scores than the GARP model, indicating better discrimination of suitable versus unsuitable areas for Asian Gypsy Moth.

The human influence index was the variable found to contribute the most in predicting the distribution of Asian gypsy moth. The human influence index includes measures that show direct human influence on the ecosystem under consideration such as human population density, major roads, navigable rivers, urban centers, and land cover categories. These model results can be used to identify areas at risk for this pest, to inform strategic and tactical pest management decisions.

Optimizing Model Development to Predict Invasive Species Ranges, Now and in the Future

[Maps:] Geographical distribution of Sirex Woodwasp, Asian longhorned beetle, sudden oak death, and Dutch elm disease. Native ranges for each invasive are circled in green. Since the origin of DED is not known we assume them to be native to Asia for the purpose of comparison..We focused on four forest invasive species (FIS) to evaluate the best way to design the models using MaxEnt to predict the potential forest invasive species. We also created projections of future forest invasive species distributions under selected climate change scenarios that took the natural dispersal abilities into account. The four forest invasive species we chose to serve as case studies were: two insects (Asian longhorned beetle and Asian gypsy moth) and two pathogens (sudden oak death and Dutch elm disease).

The default model settings produced complex models that provided an incomplete picture of the potential invasion. When models were simplified and included biologically relevant predictors and included dispersal capabilities of the species, more accurate predictions were obtained, particularly when projecting future spread. This is a case where less (using only species-specific predictive parameters) is more in modeling invasive species distributions with MaxEnt.

Assessing Invasive Species Niches

[graph:] Graph of the proportion of the introduced niche that overlaps with the native niche for each species: Sirex woodwasp, Asian longhorned beetle, sudden oak death, and Dutch elm disease.Species distribution models are based on the classical assumption that species will be able to establish populations in areas outside of their native range that closely match the environmental conditions of their native distributions. However, as invasive species are introduced to novel habitats, they are exposed to a variety of abiotic and biotic conditions that may (or may not) resemble their native range. Invasive species may undergo evolutionary shifts in their niche because introductions are often made up of few individuals with limited diversity. Local adaptation or hybridization with existing populations of the invasive species may also occur, modifying their environmental requirements to match the available conditions in invasive ranges.

[graph:] Graph of the proportion of the invaded niche that is not yet filled for each species: Sirex woodwasp, Asian longhorned beetle, sudden oak death, and Dutch elm disease.Understanding how stable a species niche is and how it may evolve can further improve model predictions. We used an ordination method to compare the native and post-invasion niches of four invasive forest pests: Sirex woodwasps; Asian longhorned beetle, sudden oak death and Dutch elm disease. An ordination method orders and characterizes the data so that similar data are mapped near each other and dissimilar data are farther from each other.

We found that the climatic niche that most of the species occupied in their invaded range was different from that in their native range and they had not fully occupied all the suitable habitat available to them in the invaded range. The overlap between the niche in the invaded and native ranges varied between the four species we studied. The observed niche shifts can be explained by combined effects of the tolerance of each species to a range of environmental conditions and the species ability to adapt to occupy newer areas and spread into novel environments. By comparing native and invasive niches, we can help predict a species’ potential range expansion and invasion potential.

Research Products

Srivastava, Vivek; Roe, Amanda D.; Keena, Melody A.; Hamelin, Richard C.; Griess, Verena C. 2020. Oh the places they’ll go: improving species distribution modelling for invasive forest pests in an uncertain world. Biological Invasions. 53 p. https://doi.org/10.1007/s10530-020-02372-9.

Srivastava, Vivek; Griess, Verena C.; Keena, Melody A. 2020. Assessing the Potential Distribution of Asian Gypsy Moth in Canada: A Comparison of Two Methodological Approaches. Scientific Reports. 10: 22. 10 p. https://doi.org/10.1038/s41598-019-57020-7

Srivastava, Vivek; Liang, Wanwan; Keena, Melody A.; Roe, Amanda D.; Hamelin, Richard C.; Griess, Verena C. 2020. Assessing Niche Shifts and Conservatism by Comparing the Native and Post-Invasion Niches of Major Forest Invasive Species. Insects. 11(8): 479. 19 p. https://doi.org/10.3390/insects11080479.

Research Participants

Principal Investigators

  • Melody A. Keena , Research Entomologist, USDA Forest Service Northern Research Station
  • R. Talbot Trotter, Research Ecologist, USDA Forest Service Northern Research Station
  • Vivek Srivastava, University of British Columbia, Department of Forest Resources Management, Vancouver, British Columbia, Canada, Post-Doctoral Researcher
  • Verena Griess, University of British Columbia, Department of Forest Resources Management, Vancouver, British Columbia, Canada, Associate Professor
  • Amanda Roe, Great Lakes Forestry Center, Natural Resources Canada, Sault Ste. Marie, ON, Canada, Research Entomologist
  • Richard Hamelin, Department of Forest and Conservation Sciences, University of British Columbia, Vancouver, BC, Canada, Professor
    Last modified: November 10, 2020