Erratum to "Branching out: Feasibility of examining the effects of greenspace on mental health after traumatic brain injury" [Dialogues in Health, Volume 2, 2023, 100129].

[This corrects the article DOI: 10.1016/j.dialog.2023.100129.].

Human disturbance in riparian areas disrupts predator-prey interactions between grizzly bears and salmon.

Wildlife must increasingly balance trade-offs between the need to access important foods and the mortality risks associated with human-dominated landscapes. Human disturbance can profoundly influence wildlife behavior, but managers know little about the relationship between disturbance-behavior dynamics and associated consequences for foraging. We address this gap by empirically investigating the consequences of human activity on a keystone predator-prey interaction in a region with limited but varied industrial disturbance. Using stable isotope data from 226 hair samples of grizzly bears (Ursus arctos horribilis) collected from 1995 to 2014 across 22 salmon-bearing watersheds (88,000 km2) in British Columbia, Canada, we examined how human activity influenced their consumption of spawning salmon (Oncorhynchus spp.), a fitness-related food. Accounting for the abundance of salmon and other foods, salmon consumption strongly decreased (up to 59% for females) with increasing human disturbance (as measured by the human footprint index) in riparian zones of salmon-bearing rivers. Declines in salmon consumption occurred with disturbance even in watersheds with low footprints. In a region currently among the least influenced by industrial activity, intensification of disturbance in river valleys is predicted to increasingly decouple bears from salmon, possibly driving associated reductions in population productivity and provisioning of salmon nutrients to terrestrial ecosystems. Accordingly, we draw on our results to make landscape-scale and access-related management recommendations beyond current streamside protection buffers. This work illustrates the interaction between habitat modification and food security for wildlife, highlighting the potential for unacknowledged interactions and cumulative effects in increasingly modified landscapes.

WalkRollMap.org: Crowdsourcing barriers to mobility.

Walking is a simple way to improve health through physical activity. Yet many people experience barriers to walking from a variety of physical, social, and psychological factors that impact their mobility. A challenge for managing and studying pedestrian environments is that barriers often occur at local scales (e.g., sidewalk features), yet such fine scale data on pedestrian facilities and experiences are often lacking or out of date. In response, our team developed WalkRollMap.org an online mapping tool that empowers communities by providing them with tools for crowdsourcing their own open data source. In this manuscript we highlight key functions of the tool, discuss initial approaches to community outreach, and share trends in reporting from the first nine months of operation. As of July 27, 2022, there have been 897 reports, of which 53% served to identify hazards, 34% missing amenities, and 14% incidents. The most frequently reported issues were related to sidewalks (15%), driver behavior (19%), and marked crosswalks (7%). The most common suggested amenities were sidewalks, marked crosswalks, connections (i.e., pathways between streets), and curb cuts. The most common types of incidents all included conflicts with vehicles. Data compiled through WalkRollMap.org offer unique potential for local and timely information on microscale barriers to mobility and are available for use by anyone as data are open and downloadable.

A framework for modern time geography: emphasizing diverse constraints on accessibility.

Time geography is widely used by geographers as a model for understanding accessibility. Recent changes in how access is created, an increasing awareness of the need to better understand individual variability in access, and growing availability of detailed spatial and mobility data have created an opportunity to build more flexible time geography models. Our goal is to outline a research agenda for a modern time geography that allows new modes of access and a variety of data to flexibly represent the complexity of the relationship between time and access. A modern time geography is more able to nuance individual experience and creates a pathway for monitoring progress toward inclusion. We lean on the original work by Hägerstrand and the field of movement GIScience to develop both a framework and research roadmap that, if addressed, can enhance the flexibility of time geography to help ensure time geography will continue as a cornerstone of accessibility research. The proposed framework emphasizes the individual and differentiates access based on how individuals experience internal, external, and structural factors. To enhance nuanced representation of inclusion and exclusion, we propose research needs, focusing efforts on implementing flexible space-time constraints, inclusion of definitive variables, addressing mechanisms for representing and including relative variables, and addressing the need to link between individual and population scales of analysis. The accelerated digitalization of society, including availability of new forms of digital spatial data, combined with a focus on understanding how access varies across race, income, sexual identity, and physical limitations requires new consideration for how we include constraints in our studies of access. It is an exciting era for time geography and there are massive opportunities for all geographers to consider how to incorporate new realities and research priorities into time geography models, which have had a long tradition of supporting theory and implementation of accessibility research.

Wave 1 results of the INTerventions, Research, and Action in Cities Team (INTERACT) cohort study: Examining spatio-temporal measures for urban environments and health.

Built environment interventions have the potential to improve population health and reduce health inequities. The objective of this paper is to present the first wave of the INTErventions, Research, and Action in Cities Team (INTERACT) cohort studies in Victoria, Vancouver, Saskatoon, and Montreal, Canada. We examine how our cohorts compared to Canadian census data and present summary data for our outcomes of interest (physical activity, well-being, and social connectedness). We also compare location data and activity spaces from survey data, research-grade GPS and accelerometer devices, and a smartphone app, and compile measures of proximity to select built environment interventions.

Making Space in Geographical Analysis.

In this commentary we reflect on the potential and power of geographical analysis, as a set of methods, theoretical approaches, and perspectives, to increase our understanding of how space and place matter for all. We emphasize key aspects of the field, including accessibility, urban change, and spatial interaction and behavior, providing a high-level research agenda that indicates a variety of gaps and routes for future research that will not only lead to more equitable and aware solutions to local and global challenges, but also innovative and novel research methods, concepts, and data. We close with a set of representation and inclusion challenges to our discipline, researchers, and publication outlets.

Branching out: Feasibility of examining the effects of greenspace on mental health after traumatic brain injury.

Aim: This pilot study's aim was to determine the feasibility of examining the effects of an environmental variable (i.e., tree canopy coverage) on mental health after sustaining a brain injury. Methods: A secondary data analysis was conducted leveraging existing information on mental health after moderate to severe traumatic brain injury (TBI) from the TBI Model System. Mental health was measured using PHQ-9 (depression) and GAD-7 (anxiety) scores. The data were compared with data on tree canopy coverage in the state of Texas that was obtained from the Multi-Resolution Land Characteristics (MRLC) Consortium using GIS analysis. Tree canopy coverage as an indicator of neighborhood socioeconomic status was also examined using the Neighborhood SES Index. Results: Tree canopy coverage had weak and non-significant correlations with anxiety and depression scores, as well as neighborhood socioeconomic status. Data analysis was limited by small sample size. However, there is a higher percentage (18.8%) of participants who reported moderate to severe depression symptoms in areas with less than 30% tree canopy coverage, compared with 6.6% of participants who endorsed moderate to severe depression symptoms and live in areas with more than 30% tree canopy coverage (there was no difference in anxiety scores). Conclusion: Our work confirms the feasibility of measuring the effects of tree canopy coverage on mental health after brain injury and warrants further investigation into examining tree canopy coverage and depression after TBI. Future work will include nationwide analyses to potentially detect significant relationships, as well as examine differences in geographic location.

Sentinel surveillance of SARS-CoV-2 rates and equity impacts using labor and delivery patients in Phoenix, Arizona.

Proactive management of SARS-CoV-2 requires timely and complete population data to track the evolution of the virus and identify at risk populations. However, many cases are asymptomatic and are not easily discovered through traditional testing efforts. Sentinel surveillance can be used to estimate the prevalence of infections for geographical areas but requires identification of sentinels who are representative of the larger population. Our goal is to evaluate applicability of a population of labor and delivery patients for sentinel surveillance system for monitoring the prevalence of SARS-CoV-2 infection. We tested 5307 labor and delivery patients from two hospitals in Phoenix, Arizona, finding 195 SARS-CoV-2 positive. Most positive cases were associated with people who were asymptomatic (79.44%), similar to statewide rates. Our results add to the growing body of evidence that SARS-CoV-2 disproportionately impacts people of color, with Black people having the highest positive rates (5.92%). People with private medical insurance had the lowest positive rates (2.53%), while Medicaid patients had a positive rate of 5.54% and people without insurance had the highest positive rates (6.12%). With diverse people reporting for care and being tested regardless of symptoms, labor and delivery patients may serve as ideal sentinels for asymptomatic detection of SARS-CoV-2 and monitoring impacts across a wide range of social and economic classes. A more robust system for infectious disease management requires the expanded participation of additional hospitals so that the sentinels are more representative of the population at large, reflecting geographic and neighborhood level patterns of infection and risk.

The Canadian Bikeway Comfort and Safety metrics (Can-BICS): National measures of the bicycling environment for use in research and policy.

Background: The lack of consistent measures of the cycling environment across communities hampers cycling research and policy action. Our goal was to develop the first national dataset in Canada for metrics of the cycling environment at the dissemination area (DA) level - the Canadian Bikeway Comfort and Safety (Can-BICS) metrics. Data and methods: The Can-BICS metrics are area-level metrics based on the quantity of cycling infrastructure within a 1 km buffer of the population-weighted centroid of DAs. The base data are a national cycling network dataset derived from OpenStreetMap (OSM) (extracted January 25, 2022) and classified by high-, medium- and low-comfort facilities. A Can-BICS continuous metric (sum of cycling infrastructure per square kilometre weighted by comfort class) and Can-BICS categorical metric were derived and mapped for all 56,589 DAs in Canada. The Can-BICS metrics were correlated with other national datasets (2016 Canadian Active Living Environments [Can-ALE] and 2016 Census journey-to-work data) to test for associations between Can-BICS and related measures. Additionally, city staff were engaged to provide feedback on metrics during the development phase. Results: One-third (34%) of neighbourhoods in Canada have no cycling infrastructure. According to the categorical measure, 5% of all DAs were assigned as the highest category of Can-BICS (corresponding to 6% of the population) and were nearly all within metro areas. The Can-BICS continuous metric had low correlation with bike-to-work rates (R = 0.29) and was more strongly correlated with sustainable-transportation-to-work rates (R = 0.56) and the Can-ALE metrics (R=0.62). These correlations were variable across cities. Interpretation: The Can-BICS metrics provide national research- and practice-ready measures of cycling infrastructure. The metrics complement existing measures of walking and transit environments (Can-ALE), collectively providing a cohesive set of active living measures. The datasets and code are publicly available, facilitating updates as new infrastructure is built.

Riding through the pandemic: Using Strava data to monitor the impacts of COVID-19 on spatial patterns of bicycling.

COVID-19 prompted a bike boom and cities around the world responded to increased demand for space to ride with street reallocations. Evaluating these interventions has been limited by a lack of spatially-temporally continuous ridership data. Our paper aims to address this gap using crowdsourced data on bicycle ridership. We evaluate changes in spatial patterns of bicycling during the first wave of the COVID-19 pandemic (Apr - Oct 2020) in Vancouver, Canada using Strava data and a local indicator of spatial autocorrelation. We map statistically significant change in ridership and reference clusters of change to a high-resolution base map. Amongst streets where bicycling increased, we measured the proportion of increase occurring on pre-existing bicycle facilities or street reallocations compared to streets without. In all our analyses, we evaluate patterns across subsets of Strava data representing recreation, commuting, and ridership generated by women and older adults (55 + ). We found consistent and unique patterns by trip purpose and demographics: samples generated by women and older adults showed increases near green and blue spaces and on street reallocations that increased access to parks, and these patterns were also mirrored in the recreation sample. Commute ridership highlighted distinct patterns of increase around the hospital district. Across all subsets most increases occurred on bicycle facilities (pre-existing or provisional), with a strong preference for high-comfort facilities. We demonstrate that changes in spatial patterns of bicycle ridership can be monitored using Strava data, and that nuanced patterns can be identified using trip and demographic labels in the data.

COVID-19 healthcare demand projections: Arizona.

Beginning in March 2020, the United States emerged as the global epicenter for COVID-19 cases with little to guide policy response in the absence of extensive data available for reliable epidemiological modeling in the early phases of the pandemic. In the ensuing weeks, American jurisdictions attempted to manage disease spread on a regional basis using non-pharmaceutical interventions (i.e., social distancing), as uneven disease burden across the expansive geography of the United States exerted different implications for policy management in different regions. While Arizona policymakers relied initially on state-by-state national modeling projections from different groups outside of the state, we sought to create a state-specific model using a mathematical framework that ties disease surveillance with the future burden on Arizona's healthcare system. Our framework uses a compartmental system dynamics model using a SEIRD framework that accounts for multiple types of disease manifestations for the COVID-19 infection, as well as the observed time delay in epidemiological findings following public policy enactments. We use a compartment initialization logic coupled with a fitting technique to construct projections for key metrics to guide public health policy, including exposures, infections, hospitalizations, and deaths under a variety of social reopening scenarios. Our approach makes use of X-factor fitting and backcasting methods to construct meaningful and reliable models with minimal available data in order to provide timely policy guidance in the early phases of a pandemic.

What does crowdsourced data tell us about bicycling injury? A case study in a mid-sized Canadian city.

With only ∼20 % of bicycling crashes captured in official databases, studies on bicycling safety can be limited. New datasets on bicycling incidents are available via crowdsourcing applications, with opportunity for analyses that characterize reporting patterns. Our goal was to characterize patterns of injury in crowdsourced bicycle incident reports from BikeMaps.org. We extracted 281 incidents reported on the BikeMaps.org global mapping platform and analyzed 21 explanatory variables representing personal, trip, route, and crash characteristics. We used a balanced random forest classifier to classify three outcomes: (i) collisions resulting in injury requiring medical treatment; (ii) collisions resulting in injury but the bicyclist did not seek medical treatment; and (iii) collisions that did not result in injury. Results indicate the ranked importance and direction of relationship for explanatory variables. By knowing conditions that are most associated with injury we can target interventions to reduce future risk. The most important reporting pattern overall was the type of object the bicyclist collided with. Increased probability of injury requiring medical treatment was associated with collisions with animals, train tracks, transient hazards, and left-turning motor vehicles. Falls, right hooks, and doorings were associated with incidents where the bicyclist was injured but did not seek medical treatment, and conflicts with pedestrians and passing motor vehicles were associated with minor collisions with no injuries. In Victoria, British Columbia, Canada, bicycling safety would be improved by additional infrastructure to support safe left turns and around train tracks. Our findings support previous research using hospital admissions data that demonstrate how non-motor vehicle crashes can lead to bicyclist injury and that route characteristics and conditions are factors in bicycling collisions. Crowdsourced data have potential to fill gaps in official data such as insurance, police, and hospital reports.

Cyclist crash rates and risk factors in a prospective cohort in seven European cities.

Increased cycling uptake can improve population health, but barriers include real and perceived risks. Crash risk factors are important to understand in order to improve safety and increase cycling uptake. Many studies of cycling crash risk are based on combining diverse sources of crash and exposure data, such as police databases (crashes) and travel surveys (exposure), based on shared geography and time. When conflating crash and exposure data from different sources, the risk factors that can be quantified are only those variables common to both datasets, which tend to be limited to geography (e.g. countries, provinces, municipalities) and a few general road user characteristics (e.g. gender and age strata). The Physical Activity through Sustainable Transport Approaches (PASTA) project was a prospective cohort study that collected both crash and exposure data from seven European cities (Antwerp, Barcelona, London, Örebro, Rome, Vienna and Zürich). The goal of this research was to use data from the PASTA project to quantify exposure-adjusted crash rates and model adjusted crash risk factors, including detailed sociodemographic characteristics, attitudes about transportation, neighbourhood built environment features and location by city. We used negative binomial regression to model the influence of risk factors independent of exposure. Of the 4,180 cyclists, 10.2 % reported 535 crashes. We found that overall crash rates were 6.7 times higher in London, the city with the highest crash rate, relative to Örebro, the city with the lowest rate. Differences in overall crash rates between cities are driven largely by crashes that did not require medical treatment and that involved motor-vehicles. In a parsimonious crash risk model, we found higher crash risks for less frequent cyclists, men, those who perceive cycling to not be well regarded in their neighbourhood, and those who live in areas of very high building density. Longitudinal collection of crash and exposure data can provide important insights into individual differences in crash risk. Substantial differences in crash risks between cities, neighbourhoods and population groups suggest there is great potential for improvement in cycling safety.

INTERACT: A comprehensive approach to assess urban form interventions through natural experiments.

BACKGROUND: Urban form interventions can result in positive and negative impacts on physical activity, social participation, and well-being, and inequities in these outcomes. Natural experiment studies can advance our understanding of causal effects and processes related to urban form interventions. The INTErventions, Research, and Action in Cities Team (INTERACT) is a pan-Canadian collaboration of interdisciplinary scientists, urban planners, and public health decision makers advancing research on the design of healthy and sustainable cities for all. Our objectives are to use natural experiment studies to deliver timely evidence about how urban form interventions influence health, and to develop methods and tools to facilitate such studies going forward. METHODS: INTERACT will evaluate natural experiments in four Canadian cities: the Arbutus Greenway in Vancouver, British Columbia; the All Ages and Abilities Cycling Network in Victoria, BC; a new Bus Rapid Transit system in Saskatoon, Saskatchewan; and components of the Sustainable Development Plan 2016-2020 in Montreal, Quebec, a plan that includes urban form changes initiated by the city and approximately 230 partnering organizations. We will recruit a cohort of between 300 and 3000 adult participants, age 18 or older, in each city and collect data at three time points. Participants will complete health and activity space surveys and provide sensor-based location and physical activity data. We will conduct qualitative interviews with a subsample of participants in each city. Our analysis methods will combine machine learning methods for detecting transportation mode use and physical activity, use temporal Geographic Information Systems to quantify changes to urban intervention exposure, and apply analytic methods for natural experiment studies including interrupted time series analysis. DISCUSSION: INTERACT aims to advance the evidence base on population health intervention research and address challenges related to big data, knowledge mobilization and engagement, ethics, and causality. We will collect ~ 100 TB of sensor data from participants over 5 years. We will address these challenges using interdisciplinary partnerships, training of highly qualified personnel, and modern methodologies for using sensor-based data.

Impacts of Bicycle Infrastructure in Mid-Sized Cities (IBIMS): protocol for a natural experiment study in three Canadian cities.

INTRODUCTION: Bicycling is promoted as a transportation and population health strategy globally. Yet bicycling has low uptake in North America (1%-2% of trips) compared with European bicycling cities (15%-40% of trips) and shows marked sex and age trends. Safety concerns due to collisions with motor vehicles are primary barriers.To attract the broader population to bicycling, many cities are making investments in bicycle infrastructure. These interventions hold promise for improving population health given the potential for increased physical activity and improved safety, but such outcomes have been largely unstudied. In 2016, the City of Victoria, Canada, committed to build a connected network of infrastructure that separates bicycles from motor vehicles, designed to attract people of 'all ages and abilities' to bicycling.This natural experiment study examines the impacts of the City of Victoria's investment in a bicycle network on active travel and safety outcomes. The specific objectives are to (1) estimate changes in active travel, perceived safety and bicycle safety incidents; (2) analyse spatial inequities in access to bicycle infrastructure and safety incidents; and (3) assess health-related economic benefits. METHODS AND ANALYSIS: The study is in three Canadian cities (intervention: Victoria; comparison: Kelowna, Halifax). We will administer population-based surveys in 2016, 2018 and 2021 (1000 people/city). The primary outcome is the proportion of people reporting bicycling. Secondary outcomes are perceived safety and bicycle safety incidents. Spatial analyses will compare the distribution of bicycle infrastructure and bicycle safety incidents across neighbourhoods and across time. We will also calculate the economic benefits of bicycling using WHO's Health Economic Assessment Tool. ETHICS AND DISSEMINATION: This study received approval from the Simon Fraser University Office of Research Ethics (study no. 2016s0401). Findings will be disseminated via a website, presentations to stakeholders, at academic conferences and through peer-reviewed journal articles.

Monitoring city wide patterns of cycling safety.

Many cities are making significant financial investments in cycling infrastructure with the aim of making cycling safer for riders of all ages and abilities. Methods for evaluating cycling safety tend to summarize average change for a city or emphasize change on a single road segment. Few spatially explicit approaches are available to evaluate how patterns of safety change throughout a city due to cycling infrastructure investments or other changes. Our goal is to demonstrate a method for monitoring changes in the spatial-temporal distribution of cycling incidents across a city. Using cycling incident data provided by the Insurance Corporation of British Columbia, we first compare planar versus network constrained kernel density estimation for visualizing incident intensity across the street network of Vancouver, Canada. Second, we apply a change detection algorithm explicitly designed for detecting statistically significant change in kernel density estimates. The utility of network kernel density change detection is demonstrated through the comparison of cycling incident densities following the construction of two cycle tracks in the downtown core of Vancouver. The methods developed and demonstrated for this study provide city planners, transportation engineers and researchers a means of monitoring city-wide change in the intensity of cycling incidents following enhancements to cycling infrastructure or other significant changes to the transportation network.

Multiuse trail intersection safety analysis: A crowdsourced data perspective.

Real and perceived concerns about cycling safety are a barrier to increased ridership in many cities. Many people prefer to bike on facilities separated from motor vehicles, such as multiuse trails. However, due to underreporting, cities lack data on bike collisions, especially along greenways and multiuse paths. We used a crowdsourced cycling incident dataset (2005-2016) from BikeMaps.org for the Capital Regional District (CRD), BC, Canada. Our goal was to identify design characteristics associated with unsafe intersections between multiuse trails and roads. 92.8% of mapped incidents occurred between 2014 and 2016. We extracted both collision and near miss incidents at intersections from BikeMaps.org. We conducted site observations at 32 intersections where a major multiuse trail intersected with roads. We compared attributes of reported incidents at multiuse trail-road intersections to those at road-road intersections. We then used negative binomial regression to model the relationship between the number of incidents and the infrastructure characteristics at multiuse trail-road intersections. We found a higher proportion of collisions (38%, or 17/45 total reports) at multiuse trail-road intersections compared to road-road intersections (23%, or 62/268 total reports). A higher proportion of incidents resulted in an injury at multiuse trail-road intersections compared to road-road intersections (33% versus 15%). Cycling volumes, vehicle volumes, and trail sight distance were all associated with incident frequency at multiuse trail-road intersections. Supplementing traditional crash records with crowdsourced cycling incident data provides valuable evidence on cycling safety at intersections between multiuse trails and roads, and more generally, when conflicts occur between diverse transportation modes.

Spatiotemporal patterns of paralytic shellfish toxins and their relationships with environmental variables in British Columbia, Canada from 2002 to 2012.

BACKGROUND: Harmful algal blooms produce paralytic shellfish toxins that accumulate in the tissues of filter feeding shellfish. Ingestion of these toxic shellfish can cause a serious and potentially fatal condition known as paralytic shellfish poisoning (PSP). The coast of British Columbia is routinely monitored for shellfish toxicity, and this study uses data from the monitoring program to identify spatiotemporal patterns in shellfish toxicity events and their relationships with environmental variables. METHODS: The dinoflagellate genus Alexandrium produces the most potent paralytic shellfish toxin, saxitoxin (STX). Data on all STX measurements were obtained from 49 different shellfish monitoring sites along the coast of British Columbia for 2002-2012, and monthly toxicity events were identified. We performed hierarchical cluster analysis to group sites that had events in similar areas with similar timing. Machine learning techniques were used to model the complex relationships between toxicity events and environmental variables in each group. RESULTS: The Strait of Georgia and the west coast of Vancouver Island had unique toxicity regimes. Out of the seven environmental variables used, toxicity in each cluster could be described by multivariable models including monthly sea surface temperature, air temperature, sea surface salinity, freshwater discharge, upwelling, and photosynthetically active radiation. The sea surface salinity and freshwater discharge variables produced the strongest univariate models for both geographic areas. CONCLUSIONS: Applying these methods in coastal regions could allow for the prediction of shellfish toxicity events by environmental conditions. This has the potential to optimize biotoxin monitoring, improve public health surveillance, and engage the shellfish industry in helping to reduce the risk of PSP.

Intertidal resource use over millennia enhances forest productivity.

Human occupation is usually associated with degraded landscapes but 13,000 years of repeated occupation by British Columbia's coastal First Nations has had the opposite effect, enhancing temperate rainforest productivity. This is particularly the case over the last 6,000 years when intensified intertidal shellfish usage resulted in the accumulation of substantial shell middens. We show that soils at habitation sites are higher in calcium and phosphorous. Both of these are limiting factors in coastal temperate rainforests. Western redcedar (Thuja plicata) trees growing on the middens were found to be taller, have higher wood calcium, greater radial growth and exhibit less top die-back. Coastal British Columbia is the first known example of long-term intertidal resource use enhancing forest productivity and we expect this pattern to occur at archaeological sites along coastlines globally.

A geospatial approach to the prediction of indoor radon vulnerability in British Columbia, Canada.

Radon is a carcinogenic radioactive gas produced by the decay of uranium. Accumulation of radon in residential structures contributes to lung cancer mortality. The goal of this research is to predict residential radon vulnerability classes for the province of British Columbia (BC) at aggregated spatial units. Spatially referenced indoor radon concentration data were partitioned into low, medium, and high classes of radon vulnerability. Radon vulnerability classes were then linked to environmental and housing data derived from existing geospatial datasets. A Balanced Random Forests algorithm was used to model environmental predictors of indoor radon vulnerability and values at un-sampled locations across BC. A model was generated and evaluated using accuracy, precision, and kappa statistics. The influence of predictor variables was investigated through variable importance and partial dependence plots. The model performed 34% better than a random classifier. Increased probabilities of high vulnerability were associated with cold and dry winters, close proximity to major river systems, and fluvioglacial and colluvial soil parent materials. The Kootenays and Columbia-Shuswap regions were most at risk. Here, we present a novel method for predictive radon mapping that is broadly applicable to regions throughout the world.