Integrating low-cost sensor monitoring, satellite mapping, and geospatial artificial intelligence for intra-urban air pollution predictions

Lu, Liang; Daniels, Jacob; Bailey, Colleen; Hu, Leiqiu; Phillips, Ronney; South, John

doi:10.1016/j.envpol.2023.121832

Cited by 22 publications

(16 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Wood heater pollution can vary substantially over periods of less than an hour and distances of a few tens of metres, whereas smoke pollution from bushfires can travel long distances. Even a widely spaced network of PA sensors in local towns could therefore provide accurate measurements ('ground truth') to complement satellite data on bushfire smoke, noting that PM 2.5 estimated from satellite measurements of aerosol optical depth (AOD) has a spatial resolution of 500 m to 1 km [52] and temporal resolution determined by the frequency of satellite passes over the location.…”

Section: Application To Other Situations and Other Types Of Low-cost ...mentioning

confidence: 99%

Five Years of Accurate PM2.5 Measurements Demonstrate the Value of Low-Cost PurpleAir Monitors in Areas Affected by Woodsmoke

Robinson,

Goodman,

Vardoulakis

2023

IJERPH

View full text Add to dashboard Cite

Low-cost optical sensors are used in many countries to monitor fine particulate (PM2.5) air pollution, especially in cities and towns with large spatial and temporal variation due to woodsmoke pollution. Previous peer-reviewed research derived calibration equations for PurpleAir (PA) sensors by co-locating PA units at a government regulatory air pollution monitoring site in Armidale, NSW, Australia, a town where woodsmoke is the main source of PM2.5 pollution. The calibrations enabled the PA sensors to provide accurate estimates of PM2.5 that were almost identical to those from the NSW Government reference equipment and allowed the high levels of wintertime PM2.5 pollution and the substantial spatial and temporal variation from wood heaters to be quantified, as well as the estimated costs of premature mortality exceeding $10,000 per wood heater per year. This follow-up study evaluates eight PA sensors co-located at the same government site to check their accuracy over the following four years, using either the original calibrations, the default woodsmoke equation on the PA website for uncalibrated sensors, or the ALT-34 conversion equation (see text). Minimal calibration drift was observed, with year-round correlations, r = 0.98 ± 0.01, and root mean square error (RMSE) = 2.0 μg/m3 for daily average PA PM2.5 vs. reference equipment. The utitilty of the PA sensors without prior calibration at locations affected by woodsmoke was also demonstrated by the year-round correlations of 0.94 and low RMSE between PA (woodsmoke and ALT-34 conversions) and reference PM2.5 at the NSW Government monitoring sites in Orange and Gunnedah. To ensure the reliability of the PA data, basic quality checks are recommended, including the agreement of the two laser sensors in each PA unit and removing any transient spikes affecting only one sensor. In Armidale, from 2019 to 2022, the continuing high spatial variation in the PM2.5 levels observed during the colder months was many times higher than any discrepancies between the PA and reference measurements. Particularly unhealthy PM2.5 levels were noted in southern and eastern central Armidale. The measurements inside two older weatherboard houses in Armidale showed that high outdoor pollution resulted in high pollution inside the houses within 1–2 h. Daily average PM2.5 concentrations available on the PA website allow air pollution at different sites across regions (and countries) to be compared. Such comparisons revealed major elevations in PA PM2.5 at Gunnedah, Orange, Monash (Australian Capital Territory), and Christchurch (New Zealand) during the wood heating season. The data for Gunnedah and Muswellbrook suggest a slight underestimation of PM2.5 at other times of the year when there are proportionately more dust and other larger particles. A network of appropriately calibrated PA sensors can provide valuable information on the spatial and temporal variation in the air pollution that can be used to identify pollution hotspots, improve estimates of population exposure and health costs, and inform public policy.

show abstract

Section: Application To Other Situations and Other Types Of Low-cost ...mentioning

confidence: 99%

Five Years of Accurate PM2.5 Measurements Demonstrate the Value of Low-Cost PurpleAir Monitors in Areas Affected by Woodsmoke

Robinson,

Goodman,

Vardoulakis

2023

IJERPH

View full text Add to dashboard Cite

show abstract

“…Additionally, effective preprocessing techniques, such as data cleaning, normalization, and feature extraction, play a vital role in enhancing the accuracy and reliability of AI models. Choosing the appropriate algorithm for a specific sensor application is another key factor in maximizing the potential of AI . Different algorithms possess varying strengths and weaknesses, and researchers must carefully evaluate their suitability based on the desired objectives and characteristics of the sensor data.…”

Section: Maximizing the Potential Of Ai: A Practical Approach To Sens...mentioning

confidence: 99%

“…Choosing the appropriate algorithm for a specific sensor application is another key factor in maximizing the potential of AI. 11 Different algorithms possess varying strengths and weaknesses, and researchers must carefully evaluate their suitability based on the desired objectives and characteristics of the sensor data. This informed selection process helps ensure that the chosen algorithm aligns with the research goals and maximizes the model's performance.…”

Section: ■ Maximizing the Potential Of Ai: A Practical Approach To Se...mentioning

confidence: 99%

AI in Sensor Research: A Reality Check and the Underestimated Potential of ChatGPT

Ray

2023

ACS Sens.

View full text Add to dashboard Cite

“…However, numerous challenges are associated with this endeavour, including the limited size and representativeness of ground reference stations for model development, the integration of data from multiple sources, and the interpretability of machine learning models. This research endeavours to address these challenges by utilising a strategically deployed and extensive network of low-cost sensors (LCS), which have undergone rigorous calibration through an optimised neural network approach [12][13][14].…”

Section: Introductionmentioning

confidence: 99%

Air Pollution PM10 Forecasting Maps in the Maritime Area of the Bay of Algeciras (Spain)

Rodríguez-García,

Carrasco-García,

Ribeiro

et al. 2024

JMSE

View full text Add to dashboard Cite

Predicting the levels of a pollutant in a given area is an open problem, mainly because historical data are typically available at certain locations, where monitoring stations are located, but not at all locations in the area. This work presents an approach based on developing predictions at each of the points where an immission station is available; in this case, based on shallow Artificial Neural Networks, ANNs, and then using a simple geostatistical interpolation algorithm (Inverse Distance Weighted, IDW), a pollutant map is constructed over the entire study area, thus providing predictions at each point in the plane. The ANN models are designed to make 1 h ahead and 4 h ahead predictions, using an autoregressive scheme as inputs (in the case of 4 h ahead as a jumping strategy). The results are then compared using the Friedman and Bonferroni tests to select the best model at each location, and predictions are made with all the best models. In general, to the 1 h ahead prediction models, the optimal models typically have fewer neurons and require minimal historical data. For instance, the best model in Algeciras has an R of almost 0.89 and consists of 1 hidden neuron and 3 to 5 lags, similar to Colegio Los Barrios. In the case of 4h ahead prediction, Colegio Carteya station shows the best model, with an R of almost 0.89 and a MSE of less than 240, including 5 hidden neurons and different lags from the past. The results are sufficiently adequate, especially in the case of predictions 4 h into the future. The aim is to integrate the models into a tool for citizens and administrations to make decisions.

show abstract

Integrating low-cost sensor monitoring, satellite mapping, and geospatial artificial intelligence for intra-urban air pollution predictions

Cited by 22 publications

References 58 publications

Five Years of Accurate PM2.5 Measurements Demonstrate the Value of Low-Cost PurpleAir Monitors in Areas Affected by Woodsmoke

Five Years of Accurate PM2.5 Measurements Demonstrate the Value of Low-Cost PurpleAir Monitors in Areas Affected by Woodsmoke

AI in Sensor Research: A Reality Check and the Underestimated Potential of ChatGPT

Air Pollution PM10 Forecasting Maps in the Maritime Area of the Bay of Algeciras (Spain)

Contact Info

Product

Resources

About