Projects

Understanding biases and drivers of extinction risk across the world's birds

In this project, I address the relative and combined roles of traits, human-imposed threats and human-related bias in estimating species extinction risk. For birds globally, we first ask how key traits (body mass, range size, and migration) interact with human-posed threats (agriculture, climate change, hunting, invasive species, logging and pollution) to shape extinction risk and assess whether the overall influence of traits varies across different threats.

We then examine how popularity (based on number of Google searches or hits) affects the likelihood of a species being listed, holding all other factors constant. Finally, we use our model to identify regions of the world where extinction risk predictions align or deviate from the IUCN listings, highlighting areas where species and ecosystems may be more vulnerable than currently recognized. I pursued this project as part of my PhD research at McGill University, under the supervision of Laura Pollock, and in collaboration with Katherine Hébert at the Pollock Lab. 

Objectives

The aim of this study is to understand how biological traits, human-induced threats, and species popularity interact to influence the extinction risk of bird species globally. We seek to improve the accuracy of extinction risk assessments and inform conservation strategies.

Results

We find that biological traits, human-induced threats, and species popularity all explain extinction risk. Some interactions are important (e.g. larger species are more likely threatened by hunting and small-ranged and migratory species are more threatened by agriculture). We also find that more popular species are more likely to be listed as at-risk than unpopular species with similar traits and threats.

Our study highlights the need of incorporating biological traits, anthropogenic threats and human bias into extinction risk assessments. These factors interact in complex ways, influencing the vulnerability of bird species. By accounting for these interactions, conservation efforts can be more effectively targeted to protect species with the highest risk.

In this project, we examine how different frameworks used to assess species vulnerability to climate change relate to one another and to observed population declines. Conservation assessments increasingly rely on either trait-based vulnerability assessments (TVAs), which emphasize intrinsic sensitivity and adaptive capacity, or model-based climate change vulnerability assessments (CCVAs), which focus solely on projected exposure derived from species distribution models. Despite their widespread use, these approaches are rarely compared within a common analytical framework or evaluated against independent demographic data. We  explicitly compare trait-based and model-based climate vulnerability metrics for breeding birds in Canada, and to test whether these vulnerability signals align with long-term abundance trends from Rosenberg et al. (2019).

Objectives

We aim to determine whether trait-based and model-based climate vulnerability frameworks identify the same species as vulnerable, or whether they capture distinct dimensions of risk. Specifically, I aimed to (1) reconstruct vulnerability using both a TVA and a model-based CCVA rather than relying on fixed categorical outputs, (2) evaluate the agreement and divergence between these frameworks across species, and (3) test whether vulnerability classifications and underlying traits are associated with observed population change (from . A key motivation was to assess whether current climate vulnerability assessments detect early demographic signals or remain decoupled from contemporary declines.

Results

Results show that trait-based and model-based vulnerability frameworks only partially overlap in the species they identify as climate-vulnerable, indicating that they capture related but non-identical dimensions of risk. Projected loss of suitable habitat was more consistently associated with population declines than categorical climate-change listings, while traits (particularly body mass, habitat breadth, range size, and migratory strategy) explained abundance change. Notably, IUCN threat status and climate-change listings alone were weak predictors of the magnitude of observed declines. Together, these findings highlight the importance of integrating trait-based and exposure-based information when assessing climate vulnerability and show that demographic responses to climate change are structured primarily by species’ ecological characteristics not necessarily being captured by current vulnerability classifications.

BC Parks Biodiversity Monitoring and Conservation Planning

This collaborative project focused on improving biodiversity monitoring and conservation decision-making across British Columbia's protected areas. By integrating community science observations, environmental data, and spatial analyses, we evaluated monitoring coverage, identified biodiversity data gaps, and explored strategies for optimizing future sampling efforts.

Using reproducible workflows in R and GIS, we developed analytical products to support conservation planning, including monitoring priority maps, species richness assessments, and decision-support tools. The project highlights how large-scale biodiversity datasets can be transformed into actionable information for protected area management and conservation practice.

Link: https://doi.org/10.1016/j.biocon.2025.111007 

This project grew out of a collaboration led by Katherine Hébert at McGill University, where I contributed alongside colleagues from Canada and Europe. It started during the “Tracking a Moving Target” workshop at the GEOBON: Monitoring for Biodiversity conference in Montreal. We were motivated by the realization that while many biodiversity indicators exist to track progress toward the Kunming–Montreal Global Biodiversity Framework (GBF), most are designed to capture national or global trends—leaving a critical gap in our ability to detect fine-scale, short-term changes that actually drive conservation success on the ground.

Objectives

Our main goal was to explore whether existing biodiversity indicators could effectively monitor local, near-term biodiversity change, which is essential for adjusting conservation strategies in time to meet 2030 targets. We synthesized expert input from 78 scientists and practitioners to assess indicator coverage across spatial and temporal scales, identified where gaps exist, and proposed concrete recommendations to improve how we track biodiversity in places and timescales that matter most for conservation action.

Results

We found a worrying blank space: current indicators largely fail to track biodiversity changes at local scales or within the tight timeframes required to evaluate short-term progress. This means we may lack the tools to know if on-the-ground conservation actions are actually working before 2030. To close this gap, we outlined five recommendations, from integrating local data and testing indicator performance at multiple scales, to developing new indicators that directly measure fine-scale biodiversity change. This work underscores the need to better connect local monitoring to global targets, ensuring that small-scale conservation decisions truly scale up to achieve international biodiversity goals. 

Strategic efforts needed to address major gaps in documenting world’s threatened and useful plants

As part of my work in the Pollock Lab at McGill University, I contributed to a project led by Nina Obiar, with Isaac Eckert, examining the sampling effort still needed to accurately map global plant diversity. Reliable spatial models are essential for effective conservation planning, yet major gaps persist in our understanding of where plant species occur. In this project, we leveraged both herbarium records and community science data (iNaturalist) to assess how well these sources capture useful and endangred plants across the world’s botanical regions. Using the USAGE index (Obiar et al. 2025) to quantify human value and IUCN assessments to identify conservation status, we assessed whether herbarium records and community science observations provide sufficient coverage to map and model the distributions of these priority plants. This approach allows us to evaluate whether species that are essential for food security, medicine, livelihoods, and cultural practices, or those most at risk of extinction, are adequately represented in global biodiversity databases.

Objectives

Our goal was to determine whether existing herbarium and community science data meet the minimal requirements to build range maps and species distribution models for the world’s useful and threatened plants. Specifically, we aim to quantify  how many useful and threatened species have at least three georeferenced locality records, the minimum needed to draw a spatial range. We explore whether threatened, and overlapping useful–threatened species differ in their documentation gaps compared to other plants and if species with higher USAGE index values are better represented in existing databases, and if do biases persist across categories and regions.

Results

We found striking shortfalls in coverage for useful and threatened plants. Fewer than half of all useful species, and fewer than one-third of threatened species, had enough records to construct even a simple range map. Only about 15% of useful species and 8% of threatened species had sufficient data for species distribution models. Species that were both useful and threatened showed the greatest data deficiencies. Interestingly, while globally important crops and ornamentals were well sampled, wild relatives and culturally important medicinal species were severely underrepresented. Both herbaria and community science contributed complementary strengths: herbaria captured a wider taxonomic scope, while community science provided denser records for common useful species, yet both sources were strongly biased toward the Global North. These results highlight an urgent need to digitize herbarium specimens and expand community science in biodiversity-rich regions of the Global South to safeguard the knowledge and conservation of the world’s most valuable and vulnerable plants.

Link: https://doi.org/10.1139/facets-2024-0357