Predicting Within-City Spatial Variations in Outdoor Ultrafine Particle and Black Carbon Concentrations in Bucaramanga, Colombia: A Hybrid Approach Using Open-Source Geographic Data and Digital Images.

Outdoor ultrafine particles (UFP, <0.1 µm) and black carbon (BC) vary greatly within cities and may have adverse impacts on human health. In this study, we used a hybrid approach to develop new models to estimate within-city spatial variations in outdoor UFP and BC concentrations across Bucaramanga, Colombia. We conducted a mobile monitoring campaign over 20 days in 2019. Regression models were trained on land use data and combined with predictions from convolutional neural networks (CNN) trained to predict UFP and BC concentrations using satellite and street-level images. The combined UFP model (R2 = 0.54) outperformed the CNN (R2 = 0.47) and land use regression (LUR) models (R2 = 0.47) on their own. Similarly, the combined BC model also outperformed the CNN and LUR BC models (R2 = 0.51 vs 0.43 and 0.45, respectively). Spatial variations in model performance were more stable for the CNN and combined models compared to the LUR models, suggesting that the combined approach may be less likely to contribute to differential exposure measurement error in epidemiological studies. In general, our findings demonstrated that satellite and street-level images can be combined with a traditional LUR modeling approach to improve predictions of within-city spatial variations in outdoor UFP and BC concentrations.

An integrated, modular approach to data science education in microbiology.

We live in an increasingly data-driven world, where high-throughput sequencing and mass spectrometry platforms are transforming biology into an information science. This has shifted major challenges in biological research from data generation and processing to interpretation and knowledge translation. However, postsecondary training in bioinformatics, or more generally data science for life scientists, lags behind current demand. In particular, development of accessible, undergraduate data science curricula has the potential to improve research and learning outcomes as well as better prepare students in the life sciences to thrive in public and private sector careers. Here, we describe the Experiential Data science for Undergraduate Cross-Disciplinary Education (EDUCE) initiative, which aims to progressively build data science competency across several years of integrated practice. Through EDUCE, students complete data science modules integrated into required and elective courses augmented with coordinated cocurricular activities. The EDUCE initiative draws on a community of practice consisting of teaching assistants (TAs), postdocs, instructors, and research faculty from multiple disciplines to overcome several reported barriers to data science for life scientists, including instructor capacity, student prior knowledge, and relevance to discipline-specific problems. Preliminary survey results indicate that even a single module improves student self-reported interest and/or experience in bioinformatics and computer science. Thus, EDUCE provides a flexible and extensible active learning framework for integration of data science curriculum into undergraduate courses and programs across the life sciences.

Combining citizen science and deep learning for large-scale estimation of outdoor nitrogen dioxide concentrations.

Reliable estimates of outdoor air pollution concentrations are needed to support global actions to improve public health. We developed a new approach to estimating annual average outdoor nitrogen dioxide (NO2) concentrations using approximately 20,000 ground-level measurements in Flanders, Belgium combined with aerial images and deep neural networks. Our final model explained 79% of the spatial variability in NO2 (root mean square error of 10-fold cross-validation = 3.58 µg/m3) using only images as model inputs. This novel approach offers an alternative means of estimating large-scale spatial variations in ambient air quality and may be particularly useful for regions of the world without detailed emissions data or land use information typically used to estimate outdoor air pollution concentrations.

Predicting outdoor ultrafine particle number concentrations, particle size, and noise using street-level images and audio data.

Outdoor ultrafine particles (UFPs) (<0.1 µm) may have an important impact on public health but exposure assessment remains a challenge in epidemiological studies. We developed a novel method of estimating spatiotemporal variations in outdoor UFP number concentrations and particle diameters using street-level images and audio data in Montreal, Canada. As a secondary aim, we also developed models for noise. Convolutional neural networks were first trained to predict 10-second average UFP/noise parameters using a large database of images and audio spectrogram data paired with measurements collected between April 2019 and February 2020. Final multivariable linear regression and generalized additive models were developed to predict 5-minute average UFP/noise parameters including covariates from deep learning models based on image and audio data along with outdoor temperature and wind speed. The best performing final models had mean cross-validation R2 values of 0.677 and 0.523 for UFP number concentrations and 0.825 and 0.735 for UFP size using two different test sets. Audio predictions from deep learning models were stronger predictors of spatiotemporal variations in UFP parameters than predictions based on street-level images; this was not explained only by noise levels captured in the audio signal. All final noise models had R2 values above 0.90. Collectively, our findings suggest that street-level images and audio data can be used to estimate spatiotemporal variations in outdoor UFPs and noise. This approach may be useful in developing exposure models over broad spatial scales and such models can be regularly updated to expand generalizability as more measurements become available.

Snowfall, Temperature, and the Risk of Death From Myocardial Infarction: A Case-Crossover Study.

Previous research has associated snowfall with risk of myocardial infarction (MI). Most studies have been conducted in regions with harsh winters; it remains unclear whether snowfall is associated with risk of MI in regions with milder or more varied climates. A case-crossover design was used to investigate the association between snowfall and death from MI in British Columbia, Canada. Deaths from MI among British Columbia residents between October 15 and March 31 from 2009 to 2017 were identified. The day of each death from MI was treated as the case day, and each case day was matched to control days drawn from the same day of the week during the same month. Daily snowfall amount was assigned to case and control days at the residential address, using weather stations within 15 km of the residence and 100 m in elevation. In total, 3,300 MI case days were matched to 10,441 control days. Compared with days that had no snowfall, odds of death from MI increased 34% (95% confidence interval: 0%, 80%) on days with heavy snowfall (≥5 cm). In stratified analysis of deaths from MI as a function of both maximum temperature and snowfall, risk was significantly increased on snowfall days when the temperature was warmer.

Extending the spatial scale of land use regression models for ambient ultrafine particles using satellite images and deep convolutional neural networks.

We paired existing land use regression (LUR) models for ambient ultrafine particles in Montreal and Toronto, Canada with satellite images and deep convolutional neural networks as a means of extending the spatial coverage of these models. Our findings demonstrate that this method can be used to expand the spatial scale of LUR models, thus providing exposure estimates for larger populations. The cost of this approach is a small loss in precision as the training data are themselves modelled values.

Seasonal ambient particulate matter and population health outcomes among communities impacted by road dust in British Columbia, Canada.

In recent years, many air quality monitoring programs have favored measurement of particles less than 2.5 µm (PM2.5) over particles less than 10 µm (PM10) in light of evidence that health impacts are mostly from the fine fraction. However, the coarse fraction (PM10-2.5) may have independent health impacts that support continued measurement of PM10 in some areas, such as those affected by road dust. The objective of this study was to evaluate the associations between different measures of daily PM exposure and two daily indicators of population health in seven communities in British Columbia, Canada, where road dust is an ongoing concern. The measures of exposure were PM10, PM2.5, PM10-2.5, PM2.5 adjusted for PM10-2.5, and PM10-2.5 adjusted for PM2.5. The indicators of population health were dispensations of the respiratory reliever medication salbutamol sulfate and nonaccidental mortality. This study followed a time-series design using Poisson regression over a 2003-2015 study period, with analyses stratified by three seasons: residential woodsmoke in winter; road dust in spring; and wildfire smoke in summer. A random-effects meta-analysis was conducted to establish a pooled estimate. Overall, an interquartile range increase in daily PM10-2.5 was associated with a 3.6% [1.6, 5.6] increase in nonaccidental mortality during the road dust season, which was reduced to 3.1% [0.8, 5.4] after adjustment for PM2.5. The adjusted coarse fraction had no effect on salbutamol dispensations in any season. However, an interquartile range increase in PM2.5 was associated with a 2.7% [2.0, 3.4] increase in dispensations during the wildfire season. These analyses suggest different impacts of different PM fractions by season, with a robust association between the coarse fraction and nonaccidental mortality in communities and periods affected by road dust. We recommend that PM10 monitoring networks be maintained in these communities to provide feedback for future dust mitigation programs. IMPLICATIONS: There was a significant association between daily concentrations of the coarse fraction and nonaccidental mortality during the road dust season, even after adjustment for the fine fraction. The acute and chronic health effects associated with exposure to the coarse fraction remain unclear, which supports the maintenance of PM10 monitoring networks to allow for further research in communities affected by sources such as road dust.

Systematic identification and prioritization of communities impacted by residential woodsmoke in British Columbia, Canada.

Residential woodsmoke is an under-regulated source of fine particulate matter (PM2.5), often surpassing mobile and industrial emissions in rural communities in North America and elsewhere. In the province of British Columbia (BC), Canada, many municipalities are hesitant to adopt stricter regulations for residential wood burning without empirical evidence that smoke is affecting local air quality. The objective of this study was to develop a retrospective algorithm that uses 1-h PM2.5 concentrations and daily temperature data to identify smoky days in order to prioritise communities by smoke impacts. Levoglucosan measurements from one of the smokiest communities were used to establish the most informative values for three algorithmic parameters: the daily standard deviation of 1-h PM2.5 measurements; the daily mean temperature; and the daytime-to-nighttime ratio of PM2.5 concentrations. Alternate parameterizations were tested in 45 sensitivity analyses. Using the most informative parameter values on the most recent two years of data for each community, the number of smoky days ranged from 5 to 277. Heat maps visualizing seasonal and diurnal variation in PM2.5 concentrations showed clear differences between the higher- and lower-ranked communities. Some communities were sensitive to one or more of the parameters, but the overall rankings were consistent across the 45 analyses. This information will allow stakeholder agencies to work with local governments on implementing appropriate intervention strategies for the most smoke-impacted communities.