About

Michael Bergin is the Steinberg Family Professor of Civil & Environmental Engineering and a Professor in the Department of Civil and Environmental Engineering at Duke Kunshan University. His research focuses on the influence of air pollution on both climate and human health, with a particular emphasis on particulate matter (PM). He has conducted extensive studies on the emission, formation, deposition, and impacts of PM, exploring how it affects climate by modifying the radiation balance of the atmosphere. His research has taken him to pristine regions such as Greenland and the Himalaya, as well as hazy regions including the Southeastern US, China, and India. More recently, Dr. Bergin has been studying the influence of PM on human health, especially in determining the contributions of sources like biomass burning and vehicular emissions to acute health impacts. He is also involved in developing and deploying next-generation air quality sensors to inform citizens about air quality, enabling informed decisions to improve air conditions. His vision involves a multidisciplinary, multicultural approach to research and education, bringing together global researchers to work collectively toward cleaner air.

Research topics

• Artificial Intelligence
• Internal medicine
• Medicine
• Computer Science
• Geography
• Cardiology
• Meteorology
• Pediatrics
• Engineering
• Physical therapy
• Psychiatry
• Remote sensing
• Emergency medicine
• Environmental science

Selected publications

• Comparison of PM2.5 trends and source factors in urban and rural locations in Bangladesh

  Atmospheric Pollution Research · 2025-09-10

  articleOpen access

  Air pollution remains a critical environmental and public health concern, particularly in developing countries like Bangladesh. Although urban air pollution has received significant attention, rural areas also experience high PM 2.5 concentrations. In this study, a low-cost sensor (LCS) network was deployed across five locations: Dhaka, Rajshahi, Panchagarh, Netrokona, and Bhola from April 2022 to September 2023 to assess local and regional sources of PM 2.5 . A Generalized Additive Model (GAM) was applied to analyze the influence of meteorology and source contributions on observed PM 2.5 concentrations. The highest PM 2.5 levels were recorded in Netrokona (212.81 ± 64.5 μgm −3 ), followed by Panchagarh (128.6 ± 71.7 μgm −3 ), Rajshahi (110.4 ± 48.5 μgm −3 ), Dhaka (105.1 ± 55.7 μgm −3 ) and Bhola (82.2 ± 36.2 μgm −3 ). A consistent diurnal pattern was observed across all sites, characterized by two peaks in the morning and evening. GAM analysis revealed that the boundary layer height had the lowest influence in Bhola and Panchagarh, while Dhaka exhibited the highest contribution. The contribution of long-range transport was found uniform at all the sites. The Trajectory Cluster Concentration Impact (TCCI) showed that the Indo-Gangetic Plain (IGP) is responsible for the enhancement of 50 μgm −3 at all the sites. However, wind transported from the Bay of Bengal associates PM 2.5 reduction of 20–40 μgm −3 . Impacts of local winds on the PM 2.5 concentrations in the GAM simulations suggested that winds from the northwest are associated with higher PM 2.5 . These findings emphasize the need for comprehensive air quality management strategies that extend beyond major urban centers to include rural and semi-urban areas. • Low-cost sensors deployed at five sites to evaluate local and regional PM 2.5 trends. • Long range air mass influences PM2.5, but local sources remain primary contributors. • Northwest winds amplify PM2.5 concentrations alongside local emissions. • Indo-Gangetic Plain contributes approximately 50 μgm-3 to PM2.5 across all sites. • Findings highlight the need for air quality monitoring in both urban and rural areas.

  PublisherDOI

• How do workers experience online stress management interventions? A review of the findings from qualitative investigations

  Journal of Workplace Behavioral Health · 2025-12-11 · 1 citations

  article
  PublisherDOI

• Scalable and robust Gaussian processes for reanalysis of urban air temperature with crowdsourced meteorological data

  2025-10-24

  articleOpen access

  Crowdsourced air temperature data from networks like Weather Underground offer dense spatial coverage and are increasingly used to study the canopy urban heat island (CUHI) effect. However, these observations are noisy: siting conditions, environmental interference, and sensor failures introduce spatially and temporally varying bias. This complicates interpolation, limiting our ability to estimate neighborhood-level air temperature. While interpolation techniques such as kriging account for uncertainty, they do so under the assumption of homoscedasticity. Moreover, they struggle to scale beyond a few thousand observations, which limits their utility on crowdsourced data. To overcome these limitations, we develop a sparse variational Gaussian process model that accounts for heteroscedasticity, allowing us to efficiently produce interpolated air temperature fields with calibrated uncertainty estimates. To test our approach, we apply our model to six years of hourly data across Durham County, North Carolina. This area includes a medium-sized city, so we expect our model to generalize to similarly sized regions with sufficient sensor coverage. Compared to ERA5-Land, it improves estimates of canopy temperature (MAE=0.57°C versus ERA5-Land MAE=3.20°C on held-out locations) and enables high-resolution analysis of CUHI patterns over space and time. We illustrate this by visualizing (1) how CUHI patterns vary with synoptic conditions, (2) differential impacts on heating and cooling demand, and (3) annual hours exceeding 35°C by neighborhood. Our method provides a scalable and statistically rigorous framework for transforming crowdsourced climate data into a gridded reanalysis product. Using these data, we can better quantify urban heat exposure and its impact on health and energy.

  PublisherOA PDFDOI

• Urinary Pyrene Carboxylic Acid as a Novel Exposure Biomarker of Woodsmoke

  Environmental Science & Technology Letters · 2025-09-08

  articleOpen access

  Quantifying people's exposure to wildfires is essential for assessing related health risks. While hydroxyl metabolites of polycyclic aromatic hydrocarbons (PAHs) are commonly used exposure biomarkers of combustion-originated air pollutants, methylated PAHs are more abundant in woodsmoke than other sources. Thus, urinary PAH carboxylic acids, which are metabolites of methylated PAHs, may serve as more sensitive biomarkers of wildfire exposure. In this exploratory study, we developed an LC-MS/MS method to simultaneously quantify hydroxylated and carboxylic metabolites of PAHs and methyl-PAHs in urine. This method was then applied to 56 urine samples collected from 8 campers before, during, and after a 4-hour exposure to campfire. Campers also wore silicone wristbands to monitor ambient PAHs. We found that 1-pyrenecarboxylic acid (1-PYRCA) levels increased significantly at 4 h (96.9%, 95% CI: 2.60-101%), 6 h (96.8%, 95% CI: 5.85-107%), and 8 h (92.5%, 95% CI: 3.59-99.2%), and returned to baseline levels at 24 h. In contrast, the campfire exposure did not significantly increase other urinary PAH metabolites. Wristband PAHs also significantly increased during the 4-hour exposure. These results suggest the use of urinary 1-PYRCA as a sensitive exposure biomarker for woodsmoke and potentially for assessing exposure to wildfires.

  PublisherOA PDFDOI

• Mass balance model for estimating indoor and outdoor contributions to residential PM2.5 and indoor source composition profiles

  Journal of Hazardous Materials · 2025-09-16 · 3 citations

  article
  PublisherDOI

• Predicting ground-level PM _2.5 using high-resolution satellite imagery and machine learning algorithms with ground-based validation

  International Journal of Remote Sensing · 2025-09-01

  article
  PublisherDOI

• Development and optimization of the low-cost optical monitor for real-time monitoring of atmospheric black carbon

  Environmental Advances · 2025-06-21

  articleOpen access

  • Machine learning models trained by using spectrum values for Black retrieval. • The GBR model performed better than the RF model in estimating BC, yielding more accurate results. • Machine learning models demonstrate high accuracy in predicting BC. • Carbon Scan is a cost-effective method to monitor BC on large temporal and spatial regions. • Carbon Scan overcomes low-cost methods by enabling continuous and real-time monitoring. Black Carbon (BC) is a major component of atmospheric aerosol produced from incomplete combustion of fossil fuels, biomass, and solid fuel. About 42 % of BC emissions are from open biomass burning with major contributions from Africa and Asia. A low-cost (<1500 US$) monitoring system is designed for large scale monitoring atmospheric BC near the surface for effective mitigation actions. Carbon Scan includes an air sampler with filtration of particulate matter of size 2.5 micron or less, a color sensor that captures image, and a machine learning (ML) model that retrieves BC concentration. Gradient Boosting Regressor (GBR) and Neural Network (NN) were trained and evaluated for the retrieval of atmospheric BC and its fraction from biomass burning. Carbon scan is a significant advancement in atmospheric BC monitoring. It achieves great accuracy, with a low RMSE of 1 µg/m³, Mean Absolute Error (MAE) of 0.40 µg/m³, Mean Absolute Percentage Error (MAPE) of 6.76 %, (SMAPE) of 5.84 % and a high R² of 0.97. The sensor provides an opportunity to monitor real time concentrations of atmospheric BC. Carbon scan is a low power, low cost, that ensures continuous air monitoring in remote areas, while capturing large temporal and spatial variations of BC.

  PublisherDOI

• The Impact of Decarbonization on Particulate Soiling of Solar Panels

  ACS ES&T Air · 2024-10-30 · 2 citations

  articleSenior author

  Climate researchers have examined many impacts of climate change on energy supply and demand under various scenarios. However, the effect of changing particulate deposition onto solar panel surfaces on solar power production efficiency (i.e., soiling) has not been studied. We therefore characterize probabilistic outcomes across multiple climate models and scenarios. We find large current regional losses (up to 40% without manual cleaning, up to 20% with monthly cleaning and rain removal) in generation that grow slightly under a high-emission scenario, largely due to regional increases in windblown dust. In contrast, under a low-emissions scenario, potential production increases significantly (2–8% interquartile range with only rain removal) due to reduced soiling, especially in regions of Asia and Africa where anthropogenic aerosols are major contributors to soiling. Projected changes vary widely across models in many dusty areas outside of the Sahara and Arabia. Differences can also be large in regions dominated by anthropogenic aerosols, such as Nigeria, eastern China, and northern India, where the full range across modeled potential power production changes extends from −1 to +11% for the end of the century (without manual cleaning), underscoring the need to consider multiple climate models. With large increases in projected solar power deployment, the relatively small potential production increases reported here could nevertheless represent a large dividend in additional energy production. Hence, reductions in air pollution attributable to decarbonization could provide positive feedback under which a greater deployment of solar power (or other renewables) increases the production of solar power, facilitating the transition to renewable energy.

  PublisherDOI

• Nasal oxidative stress mediating the effects of colder temperature exposure on pediatric asthma symptoms

  Pediatric Research · 2024-04-11 · 6 citations

  article
  PublisherDOI

• Differentiating local and regional drivers of exceedances of WHO PM2.5 standards with a low-cost sensor network in the greater male’ region (GMR)

  Atmospheric Pollution Research · 2024-10-30 · 4 citations

  article
  PublisherDOI

Recent grants

• Collaborative Research: Particulate Organic Carbon in the Air and Snow at Summit, Greenland

  NSF · $287k · 2004–2008

• Characterization of the Physical and Chemical Properties of Water Insoluble Atmospheric Aerosol

  NSF · $274k · 2001–2006

• Collaborative research: Spatial and temporal variability of surface albedo and light absorbing chemical species in Greenland.

  NSF · $46k · 2015–2015

Frequent coauthors

• Junfeng Zhang

  45 shared

• C. S. Kiang

  Peking University

  37 shared

• James J. Schauer

  37 shared

• J. A. Ogren

  31 shared

• Lauren Schmeisser

  NOAA Earth System Research Laboratory

  31 shared

• Lynn G. Salmon

  Georgia Institute of Technology

  31 shared

• Limin Zeng

  Peking University

  27 shared

• W. L. Chameides

  26 shared