Brian Leung, Ph.D.
Associate Professor


Research Interests  

Human activity is a central cause of environmental change, resulting in large-scale physical effects such as habitat loss, pollution and climate change and biological effects such as biodiversity loss, emerging infectious diseases, and species invasions. While we can make such broad generalizations, details and specific predictions needed for policy formulation are often lacking, despite their relevance in the face of competing interests of diverse stakeholders and resource trade-offs in society. These difficulties occur for virtually all environmental issues, given common limitations in data, resources, time and communication.

My lab seeks to address these limitations. The broad underlying theme of my research is to forecast environmental change quantitatively and to provide input for decision making. Currently, research in my lab is focused on invasive species, which was identified in the Millennium Ecosystem Assessment as one of the five main drivers of ecosystems change and has been roughly estimated at $137 billion/yr (to the US), disease ecology, which plays a role in emerging infectious diseases and is linked to environment change, and ecosystem management, which provides a more holistic framework for decision making, explicitly incorporating both ecological and social elements.

We focus on merging mathematical, computational, and statistical models with empirical information derived from field and laboratory studies or through meta-data from the literature. These quantitative tools provide a highly desirable and transferable skill set for prospective students interested in ecology and the environmental sciences. Given the typically limited data and high levels of uncertainty inherent in most of environmental science, there is much work to do.

Examples of research questions (see publication list for citations)

How do invasive species arrive, and how many are there?
Propagule pressure is argued to be an important determinant of invasion success and must be quantified to construct complete management strategies. My lab has developed approaches to statistically estimate propagule pressure across multiple vectors including ballast (Bailey et al. 2011), hull fouling (Sylvester et al. 2011), the aquarium trade (Cohen, Mirotchnick, Leung 2007, Gertzen, Familiar, Leung, 2008), recreational boaters (Gertzen & Leung 2011, Leung et al. 2006), and natural fluvial sources (Gertzen & Leung 2011). This work revealed many interesting phenomena. First, in the Great Lakes, ballast exchange likely removes or kills over 99.99% of freshwater organisms, but given high initial numbers present, significant numbers likely remain (Bailey et al. 2011). Second, propagule pressure from the aquarium trade is much higher than intuitively expected, with over 10,000 fish and 3000 aquatic plants released annually into the St Lawrence from Montreal alone (Cohen, Mirotchnick, Leung 2007, Gertzen, Familiar, Leung, 2008). Third, for one invader, Bythotrephes longimanus, boater traffic accounted for over 99% of the propagule pressure to inland lakes compared to natural fluvial sources (Gertzen & Leung 2011). Each of these models represented a conceptual advance in invasion biology, allowing researchers to quantify the first step in the invasion process based on accessible data.

Where do invasive species establish and spread?
After propagules reach a new area, they need to be able to survive, establish and spread. My lab has developed novel approaches to link propagule pressure and invasion history to predict probability of establishment (Gertzen & Leung 2011, Leung & Mandrak 2007, Leung & Delaney 2006, Muirhead, Leung et al. 2006, Leung et al. 2004). We have examined niche based models (NBM) (Capinha, Leung, Anastacio, 2011, Kulhanek, Leung, Ricciardi 2011, Leung & Mandrak 2007) and biotic factors (Roche, Leung et al. 2010), as predictors of site invasibility. We have developed models to predict spread across a landscape (Delaney, Edwards, Leung 2011, Hyder, Leung, Miao. 2008, Muirhead, Leung et al. 2006). We have demonstrated the ability to extrapolate to new geographical areas using NBM, to predict both presence/absence (Capinha, Leung, Anastacio, 2011, Kulhanek, Leung, Ricciardi 2011) and abundance of NIS (Kulhanek et al. 2011), and have demonstrated that the type of pseudoabsence data are potentially more important than the type of presence data (Capinha, Leung, Anastacio, 2011). We have extended theories on enemy release into a community context (Roche, Leung et al. 2010), tested for Allee effects in mesocosm (Gertzen, Leung, Yan 2011) and regional studies (Leung et al. 2004), and discovered a novel putative mechanism by which organisms may moderate Allee effects (Gertzen, Leung, Yan 2011). Further, we developed approaches to account for unsampled sources of propagules (Gertzen & Leung 2011, Leung & Delaney 2006), to accommodate information on date of discovery (but unknown true invasion date) (Gertzen & Leung 2011), and to integrate propagule pressure with NBM into a joint probability model (Leung & Mandrak 2007). We demonstrated that Allee effects were present in zebra mussels (Leung et al. 2004) and Bythotrephes (Gertzen, Leung, Yan 2011), that the true invasion rate of Bythotrephes was actually slowing, despite an accelerating rate of discovery (Gertzen & Leung 2011), and that without the use of joint models, researchers may underestimate the independent effects of both propagule pressure and habitat suitability (Leung & Mandrak 2007). Finally, for communicable diseases, which have many of the properties of invasive species, we developed integrated models of social demographic factors and spatial spread models, illustrating how local social factors can interact with spread dynamics and can scale up to affect continental patterns (Hyder, Jeanmougin, Leung. submitted), and we explicitly analyzed the predictive validity associated with models of disease epidemics, considering the effect of policy strategies (Hyder, Leung, Buckeridge, submitted).

What effect will invasive species have?
Ultimately, society is largely interested in invasive species because of their impact. It has been suggested that the best predictor of impact is the invasion history of a species i.e. its impact at previously invaded sites. However, we expect the magnitude of impact should differ across space and time, because the impact of species depends in part on the environmental conditions, which are also heterogeneous. This variability has not been generally considered, although it is important if we are to make quantitative predictions regarding the impacts. We examined the available published data on all aquatic invaders identified on the IUCN list of the 100 worst alien invasive species. We found that the type and direction of impact could be assessed for 79% of species examined, and enough information existed to fully quantify the variability for 40% of species (Kulhanek, Ricciardi, Leung 2011). For one invasive species where sufficient data existed, carp (Cyprinus carpio), we conducted further analyses to demonstrate how such quantitative information could be used to improve resolution on predicted impacts across space. We examined the relation between abundance and impact, demonstrating that abundance explained substantial variation in carp impact (R2=0.91). We further showed that we could use neural network approaches to predict carp abundance and occurrence, based on environmental characteristics (Kulhanek, Leung, Ricciardi 2011). In combination, the integration of these works provides a viable means of quantitatively predicting impact in new locations. Further, as part of a collaborative NCEAS project between forest ecologists, economists, and modelers (us), we also estimated the cost of all forest insect pests to the USA (Aukema, Leung, et al. 2011). We showed that local government and homeowners are affected most, that borer pest guild is the most damaging, and that the costs are expected to be over a billion dollars annually. Given the potential costliness of management actions (hundreds of millions of dollars), information on magnitude of expected costs and on who will bear the brunt of the costs is useful information for policy decisions.

What should we do?
Finally, management possibilities occur throughout the invasion process. We developed one of the first bioeconomic risk analysis approach as a vehicle for quantitative decision-making for invasive species (Leung et al. 2002). This work highlighted the underfunding of prevention and formed part of the recommendations for a co-authored Ecological Society of America position paper on invasive species policy (Lodge et al. 2006). Further, it served as the starting point to analyze human behaviour, exploring feedbacks between environment, economics and management (Finnoff et al. 2005a, Finnoff et al. 2005b), and explained why prevention may continue to be neglected despite its greater efficacy - for risk-averse managers, prevention of invaders that may never arrive has more uncertainty than controlling existing invaders that are already causing damage (Finnoff et al. 2007). While this work generated interesting theories, it became clear as we tried to apply these models to new situations that we typically faced four major challenges: limited information about novel invaders, limited time for action, limited resources and limited communication between scientists and manager. To overcome these challenges, my lab developed a suite of simpler models that make the most of available information and resources. We created new approaches for rapidly delimiting invasions (Leung et al. 2010) and for evaluating monitoring design (Delaney & Leung 2010), including the evaluation and use of volunteer monitoring networks (Delaney et al. 2008). We derived rules of thumb for optimal allocation for prevention and control (Leung et al. 2005), and novel ways of thinking about eradication of established populations (Edwards & Leung 2009). More broadly, as part of a longer term initiative, we are developing collaborations with governmental agencies in Prince Edward Island for ecosystem management, to integrate stakeholders, identify opportunities for mutually beneficial cooperation between industries, and develop approaches to make decisions while incorporating ecological and social uncertainty. All of these projects were designed to question and change how we think about managing environmental issues and broaden our opportunities for actions.