An official website of the United States government
Community science has increased in popularity in communities where residents hope to investigate the relationship between environmental issues and personal health. This study partnered with neighborhoods in the most polluted residential zip code in the US to conduct community science air quality monitoring. We conducted 60 semi-structured interviews after two monitoring deployments to understand participants’ subjective experiences of pollution exposure, their engagement with low-cost air quality monitors, and their data interpretation. We utilize the environmental health concept ‘exposure experience’ to analyze how participants use personal monitors, understand their data, and reinterpret their pollution exposure as a result. We further explore how participants’ understandings are circumscribed by the technological features of low-cost monitors. We find that participants adopt both protective and mitigating behavioral changes based on information gained from personal experiments and hypothesis testing while using the monitors. Of their own accord, 40% of participants in this study adopted mitigation behaviors after identifying sources that impacted their personal air quality. Our analysis reveals that real-time data accessibility through low-cost monitors builds exposure awareness and enables residents of environmental justice communities to test, validate, or invalidate sensory experiences and challenge existing assumptions. These findings point to specific pathways for using low-cost monitors to support individual decision-making and contribute to behavioral change. Findings also identify some limitations of low-cost monitors; designers of low-cost monitors should consider how composite Air Quality Scores may encourage community scientists to equally value scientifically-established pollutants (e.g., PM) with less scientifically-established pollutants (e.g., TVOCs), without additional scientific training and health-related information.
more »
« less
This content will become publicly available on October 19, 2026
Forecasting and Visualizing Air Quality from Sky Images with Vision-Language Models
Air pollution remains a critical threat to public health and environmental sustainability, yet conventional monitoring systems are often constrained by limited spatial coverage and accessibility. This paper proposes an AI-driven agent that predicts ambient air pollution levels from sky images and synthesizes realistic visualizations of pollution scenarios using generative modeling. Our approach combines statistical texture analysis with supervised learning for pollution classification, and leverages vision-language model (VLM)-guided image generation to produce interpretable representations of air quality conditions. The generated visuals simulate varying degrees of pollution, offering a foundation for user-facing interfaces that improve transparency and support informed environmental decision-making. These outputs can be seamlessly integrated into intelligent applications aimed at enhancing situational awareness and encouraging behavioral responses based on real-time forecasts. We validate our method using a dataset of urban sky images and demonstrate its effectiveness in both pollution level estimation and semantically consistent visual synthesis. The system design further incorporates human-centered user experience principles to ensure accessibility, clarity, and public engagement in air quality forecasting. To support scalable and energy efficient deployment, future iterations will incorporate a green CNN architecture enhanced with FPGA-based incremental learning, enabling real-time inference on edge platforms.
more »
« less
- Award ID(s):
- 2234227
- PAR ID:
- 10656617
- Publisher / Repository:
- The 1st Workshop on Multimodal Continual Learning (MCL) in conjunction with the 2025 IEEE/CVF International Conference on Computer Vision (ICCV 2025)
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract In an era where air pollution poses a significant threat to both the environment and public health, we present a network-based approach to unravel the dynamics of extreme pollution events. Leveraging data from 741 monitoring stations in the contiguous United States, we have created dynamic networks using time-lagged correlations of hourly particulate matter (PM2.5) data. The established spatial correlation networks reveal significant PM2.5anomalies during the 2020 and 2021 wildfire seasons, demonstrating the approach’s sensitivity to detecting regional pollution phenomena. The methodology also provides insights into smoke transport and network response, highlighting the persistence of air quality issues beyond visible smoke periods. Additionally, we explored meteorological variables’ impacts on network connectivity. This study enhances understanding of spatiotemporal pollution patterns, positioning spatial correlation networks as valuable tools for environmental monitoring and public health surveillance.more » « less
-
— In this paper, we first develop a low-cost surfacebased air pollutants measurement system for the real-time air pollution monitoring and forecasting applications. Then, we compare the performance achieved by the proposed system in real-time urban environment with currently used static monitoring stations by the governmental environmental protection agency (EPA). The proposed design uses particulate matter, humidity, and temperature sensors to measure the values of the air pollutant that determines the value of the Air Quality Index (AQI). The SD storage device is interfaced with the system to store the large amount of data sensed by the system. The Arduino UNO-based processing unit integrates with the sensing units to process and control the sensed air pollutants data. The proposed system is deployed in indoors and outdoor environment in under served minority communities in big cities to illustrate real-time environmental pollution measurement and monitoring applications. The system can measure, monitor and alert the level of PM2.5 and PM10 components of the AQI as they are often the main pollutant that determines the AQI value. The performance of the proposed system compares with the expensive data logger-based EPA-approved LDEQ sensorsbased air quality monitoring system. Our analysis shows that the measurement and monitoring performance of the proposed system is comparable with the EPA-approved LDEQ sensorsbased air quality monitoring system. The analysis also shows that there is a spatial and temporal variation of PM2.5 and PM10 values even for sites that are less than a mile apart. The interaction interphase of the system is simpler and easier to use as compared with bulky display systems in traditional EPA-based monitoring systems. In contrast with the traditional data logger-based system, the proposed system is smaller and quicker to deploy to test specific air pollutants in interested urban and rural locations .more » « less
-
Breathing in fine particulate matter of diameter less than 2.5 µm (PM2.5) greatly increases an individual’s risk of cardiovascular and respiratory diseases. As climate change progresses, extreme weather events, including wildfires, are expected to increase, exacerbating air pollution. However, models often struggle to capture extreme pollution events due to the rarity of high PM2.5 levels in training datasets. To address this, we implemented cluster-based undersampling and trained Transformer models to improve extreme event prediction using various cutoff thresholds (12.1 µg/m3 and 35.5 µg/m3) and partial sampling ratios (10/90, 20/80, 30/70, 40/60, 50/50). Our results demonstrate that the 35.5 µg/m3 threshold, paired with a 20/80 partial sampling ratio, achieved the best performance, with an RMSE of 2.080, MAE of 1.386, and R2 of 0.914, particularly excelling in forecasting high PM2.5 events. Overall, models trained on augmented data significantly outperformed those trained on original data, highlighting the importance of resampling techniques in improving air quality forecasting accuracy, especially for high-pollution scenarios. These findings provide critical insights into optimizing air quality forecasting models, enabling more reliable predictions of extreme pollution events. By advancing the ability to forecast high PM2.5 levels, this study contributes to the development of more informed public health and environmental policies to mitigate the impacts of air pollution, and advanced the technology for building better air quality digital twins.more » « less
-
Abstract Belmont County, Ohio is heavily dominated by unconventional oil and gas development that results in high levels of ambient air pollution. Residents here chose to work with a national volunteer network to develop a method of participatory science to answer questions about the association between impact on the health of their community and pollution exposure from the many industrial point sources in the county and surrounding area and river valley. After first directing their questions to the government agencies responsible for permitting and protecting public health, residents noted the lack of detailed data and understanding of the impact of these industries. These residents and environmental advocates are using the resulting science to open a dialogue with the EPA in hopes to ultimately collaboratively develop air quality standards that better protect public health. Results from comparing measurements from a citizen-led participatory low-cost, high-density air pollution sensor network of 35 particulate matter and 25 volatile organic compound sensors against regulatory monitors show low correlations (consistently R 2 < 0.55). This network analysis combined with complementary models of emission plumes are revealing the inadequacy of the sparse regulatory air pollution monitoring network in the area, and opening many avenues for public health officials to further verify people’s experiences and act in the interest of residents’ health with enforcement and informed permitting practices. Further, the collaborative best practices developed by this study serve as a launchpad for other community science efforts looking to monitor local air quality in response to industrial growth.more » « less
