Global Clean Air

Making the most of sensor data: How tracking performance of lower-cost sensors allows cities to reveal actionable insights about local air pollution

By Dan Peters / Published: February 23, 2022

Lower-cost air quality sensors can be a game changer for cities looking to understand and improve air quality at the neighborhood level. However, issues with accuracy have been a key barrier to their adoption. Our new paper shows how users can make the most of their data by evaluating sensor performance on a continuous basis.

Collocating sensors to track performance

As part of the Breathe London consortium, we installed 100 sensor devices across the city  to measure key pollutants including nitrogen dioxide (NO2) and particulate matter for more than two years. Lower-cost sensors like the ones we installed are more sensitive than reference-grade instruments to environmental factors like temperature, relative humidity, or even levels of other pollutants. That can make their measurements less reliable in some environments, or even in certain seasons of the year.

To make sure our data was both accurate and useful, the Breathe London consortium developed rigorous quality assurance procedures. For our NO2 dataset, the procedures included multiple methods to calibrate the sensors, as well as applying an algorithm to correct for sensitivity to ozone, which the sensor can mistake for NO2.

While most of our sensors were collecting measurements at new locations across Greater London, we also installed two “test” sensors alongside London reference-grade monitors for most of the project. By tracking when data from these “test” sensors deviated from the more expensive reference instruments, we had an indication of how sensors across our network were performing at different times.

In the left panel, the “test” sensor measurements show a large deviation from the collocated reference monitor (right), indicating a period when the sensor was not performing well.

This approach provided a reality check for our pollution data. If the sensor network reported high NO2 values but the “test” sensors were completely off track from the reference at that time, we could infer that the network result may have been affected by poor sensor performance and adjust accordingly. This kind of ongoing sensor evaluation is important. Without it, users could mistake erroneous sensor data as evidence of major pollution events or local hotspots.

Why performance matters

Our NO2 sensors performed well most of the time, producing data that revealed a variety of actionable insights, including:

  • Times of day and days of week with the highest pollution levels
  • Regional pollution episodes (for example, a multi-day period with high pollution caused by weather conditions)
  • Hotspot detection
  • Impacts of sources on pollution patterns at different locations
  • Long-term trends (for example, seasonal changes or year-over-year improvements)

Improving our understanding of air pollution in cities around the world

While the uncertainties associated with lower-cost sensors may make them unsuitable for some applications, our project demonstrates a way to generate actionable insights from sensors. The Breathe London network’s NO2 data shows that with rigorous quality assurance and ongoing evaluation of sensor performance, cities can utilize lower-cost sensors to better understand local air pollution. That can allow more communities to take advantage of this relatively new technology, even if they do not have the resources to purchase a network of more costly  reference-grade monitors.

Also posted in Academic, London, Monitoring / Comments are closed

Meet Ethan McMahon, Chief of Party, Clean Air Catalyst

By lambrogi / Published: February 1, 2022

Ethan McMahon is the new Chief of Party for the Clean Air Catalyst, a flagship program launched by the U.S. Agency for International Development and a global consortium of organizations led by WRI and EDF. He brings 27 years of experience with the U.S. EPA, where he worked with cities and states to build capacity to address climate change and air pollution, and he advocated to make environmental data more accessible.

What first got you interested in environmental science, and what do you find most interesting about this field?
I started my career as a mechanical engineer, doing things such as evaluating alternative refrigerants. Within a few years I learned about the impacts of climate change and I realized that I wanted to apply my analytical skills to issues that make a difference. This work on the Catalyst interests me because it involves so many dimensions. Technology, human health, collaboration–you need all of these ingredients and more to affect change on many environmental issues.

Why is open, publicly available data so crucial for solving environmental problems?
It’s hard to solve environmental issues because the causes and effects are complicated. In order to present a convincing case to decision-makers you need to speak their language, using numbers and sometimes stories. But you can only crunch the numbers if you can get the data, so it’s critical that data collectors make their data accessible and usable. 

If governments collect data for one purpose, it makes sense to get more value out of the data by making it available for other purposes. For example, EPA collects data on air quality for regulatory purposes, but community groups may want to use that same data to understand if their air quality has suddenly shifted to be worse. AirNow is a great example of how EPA makes their data available for non-regulatory purposes.

How can more data on air quality improve people’s lives?
Air quality affects some portions of the public more than others. For example, some people can only afford to live or work where pollution levels are high, such as near power plants, roadways or outdoor waste burning. The Clean Air Catalyst is finding ways to help people in the pilot cities (Jakarta, Indonesia; Indore, India; and soon, a third city). We use data from existing air quality monitors and analyze where the pollution is coming from. Then we increase awareness of the pollution – and ask people what they experience in their daily lives. Then we collect more air quality data to complement the existing monitors. After we analyze a few dimensions – health, climate change and gender – we evaluate which actions provide lasting benefits and work with communities to implement them.

Is there something about air quality monitoring that you’re especially excited about right now?
I’m really excited about people using data to affect change. They’re thinking beyond the accuracy of individual sensors and focusing instead on how they can use data to make decisions. That’s where the true value is, the benefit to health and society. Communities can use data from a few nearby sensors to understand if air quality is getting better or worse. That might be enough information for people to change their habits and protect themselves, for example by not exercising during hours when pollution levels are high.

What are some goals you have for the Clean Air Catalyst program?
I want the Clean Air Catalyst program to help cities improve their air quality in ways that are effective and sustainable. We’re using a lot of innovative methods in our pilot cities so we don’t exactly know which activities will be the most successful. However, we’ll learn from the experience and share the lessons with other cities so they can make progress easier. In parallel, we’re fostering two types of coalitions. First, we’re bringing several sectors together at the local level. Second, we’re connecting global and local experts so they can collaborate about feasible interventions. Follow our progress and feel free to suggest ways to make lasting improvements to air quality.

Learn more about the Clean Air Catalyst program here.

Also posted in Health, Homepage, Monitoring, Partners / Comments are closed

Why emission intensity matters

By lpadilla / Published: January 21, 2022

High-intensity emitters disproportionately pollute the air we breathe. Understanding where sources contribute the most potent emissions can help drive smarter clean air solutions.

Cutting the most damaging emissions from the air can be a bit like picking which foods to limit in your diet. You know the concept—fruits, vegetables, whole grains and lean proteins contribute far less to obesity than chocolate cake, cheesy pizza or greasy burgers. Healthy eating means paying attention not only to how much we consume but also the composition of each item.

The same can be true for controlling emissions of harmful pollutants like nitrogen oxide (NO), nitrogen dioxide (NO2) and small particles. Some “high-intensity” sources—like ships, diesel generators and heavy-duty trucks—produce more potent pollution than new, gasoline-fueled passenger vehicles.  In addition, conditions like stop-and-go traffic, larger cargo loads, and driving up hill can increase emission intensity, compared to freely flowing, lighter-duty traffic. Pollution varies from block to block and city to city, so understanding where sources contribute the most potent emissions can help us tailor more effective, local solutions. Our recent paper maps London’s air pollution and hotspots of emission intensity on an unprecedented street-by-street scale.

How to spot high-intensity emissions

In London, our teams used Carbon Dioxide (CO2), a key indicator of combustion, to determine the intensity of NO and NO2 pollution (NOx, in combination). Taking on-road air pollution measurements every second using mobile instruments, we identified local peaks in CO2, signaling recent emissions.  Then we calculated the emission intensity for these events as the ratio of NOx to CO2 concentrations.

Why emission intensity matters

Our measurements coincided with the implementation of Central London’s Ultra-Low Emissions Zone (ULEZ), where highly-polluting vehicles must pay a fee to enter the city center. This policy led to a cleaner vehicle fleet in and around the ULEZ and 35% lower total NOx emissions in the first year, even as overall traffic volume stayed about the same, by effectively reducing the emission intensity of individual vehicles. In fact, the ULEZ has been so successful that the Greater London Authority expanded it to an even larger area.

Emission intensity mapped in Central London. For more information on the image or to read the article, visit the journal Atmospheric Environment.

While the Central London ULEZ and its recent expansion are effective, air quality remains poor throughout London, and hot spots remain. By measuring emission intensity, we understand more about the overall causes of pollution than if we had relied solely on total concentration measurements. By digging deeper, we can show where higher-intensity sources, like heavy-duty diesel, are having a disproportionate impact on air quality. For example, we saw higher-intensity pollution along the Thames river near shipping piers, heavy construction sites and poorly-timed lights that caused traffic jams.

Crafting smart policies to combat air pollution

Equipped with local, street-scale emission intensity data, in addition to more typical total pollution measurements, policymakers in London and beyond can craft tailored solutions to cut air pollution and improve health. Some changes are easy, actionable and don’t require legislation—like fixing poorly-timed traffic lights or enacting anti-idling rules at passenger bus terminals. Other fixes—like limiting the number of warehouses that can be sited in one area to reduce truck traffic, staggering the timing and location of construction projects in order to reduce emissions from heavy equipment, electrifying buses or reducing the number of used, dirty vehicles in operation—would require more political will.

While we need to reduce all combustion-related emissions to achieve air quality and climate goals, using new methods to identify emissions intensity allows leaders to see where the dirtiest sources are, so they can focus initial efforts where tangible impacts are possible.

 

Also posted in London / Comments are closed

Catalogue of Indian Emission Inventory Reports (Jan 2022)

By Kevin Tuerff / Published: January 11, 2022

 

Indian Emission Inventory Report_DIGITAL FILE

(By PAARTHA BOSU, NEW DELHI, INDIA)  A detailed air emission inventory (EI) is a comprehensive list of pollutants within a pre-defined geographical area and is beneficial for developing clean air action plans. It can also test the effectiveness of pilot interventions towards air quality abatement. Emission inventories have been prepared for several Indian cities and states. However, several of these EI reports have not been given due attention. This report presents a database of all publicly available EI reports and several previously un-referred studies for India to help policymakers and scientists prepare reckoner of all the work done in the area.

EI studies have been tabulated as per the source contribution (total emissions, transport, residential, industrial, power plants, agriculture, waste and others) along with details such as geography, grid size, emission factors used, and type of data collected (primary surveys vs secondary literature). Each sector list also consists of the pollutants studied and highlights those reports that have closely followed the existing CPCB guidelines.

As per various operating sections of the Air Act 1981, air pollution monitoring, calculation of pollution load, preparation of emission inventory, preparation of action plan for air pollution control should be done as per the SOPs issued by CPCB from time to time. Therefore, emission inventory prepared by agencies and experts using other methodology may not be tenable per Air Act 1981. In its order for Critically Polluted Areas and Non-Attainment Cities, the National Green Tribunal mentioned that methodologies recommended by CPCB should be followed for such studies.

Robust EI reports form the mainstay of a city’s source apportionment and mitigation strategies. Therefore, scrutiny of the EI reports is required, especially now that all 132 non-attainment cities have been mandated to carry out source apportionment studies. Furthermore, periodically revised emission inventories could help check each sector’s efficacy of control actions. Finally, regional emission inventories now need to be prioritised as the airshed approach has gained prominence in air pollution management in India. About 200 EI reports have been collated and made available with hyperlinks for researchers and policymakers to use. They have also been sectorally classified for ease.

Key Findings

  1. An easy to use ready reckoner of air pollution emission inventory studies for India was created. These reports were catalogued as per sectors; Total emissions, Transport emissions, Industrial and Power Plant emissions, Residential emissions and Emissions from Agriculture, Waste and other miscellaneous sectors.
  2. It was found that only some of the studies followed the CPCB guidelines closely of using indigenous emission factors and primary data for creating emission inventories
  3. Geographically, most of the studies were concentrated in the Indo-Gangetic Plain, focusing on Delhi and the National Capital Region. Multiple emission inventories for the same city and region leads to uncertainties. Instead, a common framework for EI development should be followed. EIs should be periodically updated every few years to test the efficacy of interventions. For instance, in the transport sector, EI for the current year could help gain insights on the effects of introduction on BS VI mass emission standards on road transport emissions. In the residential sector, the introduction of LPG in rural households would have led to a reduction in emissions, and this should reflect in the latest EI report
  4. Emission factors will determine the accuracy of estimations. However, our Indian conditions are distinct from our western counterparts. Therefore, relying on the emission factors developed by USEPA might lead to inaccuracies. Thus, the transport sector emission factors developed by the Automotive Research Association of India (ARAI) were used.
  5. Inventories need to be developed for toxics like VOCs and heavy metals like mercury. Doing so will enable the development of standards for these pollutants

Download the report

For further details on the report:

Parthaa Bosu (pbosu@edf.org)

Swagata Dey (sdey@edf.org)

Also posted in Academic, Government Official/Policymaker, India / Comments are closed

Better data is critical to address health disparities in air pollution’s impacts

By EDF Web Team / Published: March 31, 2021

Ananya Roy, Senior Health Scientist, and Maria Harris, Environmental Epidemiologist 

The last several months have seen a wave of momentum in policies seeking toward advance environmental justice and equity through better data collection and mapping. In his first week in office, President Biden signed an executive order to initiate the development of a screening and mapping tool to identify disadvantaged communities with the goal of informing equitable decision making. And legislation introduced in the House of Representatives and Senate would launch a similar effort. This focus on data and mapping is critical.  

Read More »

Also posted in Environmental Justice, Government Official/Policymaker, Health, Homepage, Oakland, Partners, Public Health/Environmental Official / Comments are closed

Pollution data sharing norms are shifting

By Millie Chu Baird / Published: October 22, 2020

Sharing code in the tech community hasn’t always been considered a virtue. But GitHub, with its easy interface and mammoth user base, has shown how allowing developers to build on one another’s software code can accelerate innovation of new projects and solve bugs with existing applications, all in a transparent, open-source code hosting platform. The air quality data space is ripe for this kind of move.

Opening access to air quality data

Today, in an effort to address this critical need, Environmental Defense Fund (EDF) is unlocking our new Air Quality Data Commons (AQDC), an open-access data platform where people can share and use data from low-and medium-cost air quality sensors. With the introduction of the AQDC, researchers now have access to more than the 60 million plus data points from EDF and our partners’ air pollution studies in Oakland, Houston and London.

Until recently, few outside of government could afford the expensive, specialized equipment needed to measure air pollution other than well-funded scientists, whose data was typically private until after the publication of a peer-reviewed paper. Even then, when they wanted to share their data with others in the field, they could do so only on an ad-hoc basis with limited infrastructure in place to support such collaboration.

Now, as scientists, cities and residents are taking advantage of new low-cost, high-quality sensors, and the amount of air quality data is growing rapidly, as is the need to store and share it. To unlock the benefits of the data for both scientists and society, it must be open and easily accessible.

The Fourth Wave of Environmental Innovation

Transparency drives innovation

Many of our academic partners have long expressed the desire to share their data — once they’ve had the opportunity to analyze it. However, they’ve lacked a platform that would allow them to do so. Similarly, donors are increasingly demanding that the data gleaned from the projects they’ve funded be available for others to use and explore. By building this community, we hope people will see a benefit to not only accessing available data but sharing their own — they can ask questions of fellow air quality scientists about trends they are seeing and learn from others who may have new was of analyzing existing data.

Our partner Karin Tuxen-Bettman, Program Manager for Google Earth Outreach sees value for cities and Google as well. “By adding to the Air Quality Data Commons, cities can feel confident their investments in air monitoring — whether through a fixed stationary network or city-owned vehicle fleets equipped with sensors — are creating enormous value,” she says. “Validated data shared on the AQDC will contribute to the larger database that Google’s Environmental Insights Explorer will pull from, enabling us to build hyperlocal air quality maps for more cities. By making this data available through a transparent process, the AQDC can accelerate action required to improve air quality.”

We look forward to growing this group of data scientists, companies and cities sharing and analyzing data into a robust community who will contribute to the scientific knowledge base, so we can better understand air pollution problems around the world.

The revolution of smaller, cheaper air pollution sensors has brought us here, but the full potential of this revolution will only be realized when a larger community of scientists, cities, residents and activists use the data we collect to take action and improve local air quality. Join us by downloading our data from the AQDC, or upload your own. We look forward to sharing and learning with you.

We are entering a new era of environmental innovation that is driving better alignment between technology and environmental goals — and results. #FourthWave

This was originally posted on Medium.

Also posted in Academic, Monitoring, Partners / Authors: / Comments are closed

How we used machine learning to get better estimate of London’s NO2 pollution reduction

By Greg Slater and Dan Peters / Published: October 5, 2020

A new analysis for UK Clean Air Day from Environmental Defense Fund Europe (EDF Europe) finds nitrogen dioxide (NO2) pollution was 40% lower than expected across London during the initial COVID-19 lockdown.

But how do we know about pollution that didn’t happen? We used a machine learning model to predict what the concentration of NO2 would have been if lockdown restrictions had not come into effect. Here’s how it works.

Removing the weather impact

Meteorology and seasonal patterns have a big impact on air quality, which needs to be taken into account when measuring changes in pollution. For example, a windy day could improve air quality by dispersing pollutants that might have otherwise accumulated locally. Meteorological and seasonal variations like this make it difficult to directly compare one period to another – are changes in pollution due to a policy intervention or behaviour change, or is it just the weather?

We wanted to isolate the impact of lockdown measures on London’s NO2 pollution, which is produced from fossil fuels and is associated with heart and lung-related health impacts.

Using open-source tools developed by researchers at the University of York (Grange, 2020), and data from over 100 regulatory air quality monitors, we built a machine learning model to help us do this. London’s long-running monitoring network provides years’ worth of historic pollution data, which is used to train and test the model, alongside a series of meteorological and temporal variables.

We can then use this model – with time and weather information from lockdown dates – to predict the pollution levels we would have expected to see had lockdown measures not occurred. These predictions mirror seasonal and meteorological changes in observed pollution levels much more closely than an historical average, for example, which may vary due to different weather during that period.

As a result, with this method the difference between expected and observed levels can be more directly attributed to the impact of lockdown restrictions rather than random weather variations.

London lockdown expected vs observed chart

40% less pollution

The figure above shows a comparison between average expected and observed NO2 concentrations. The gap between what we expected to see and what we actually saw increases dramatically after 16th March, when social distancing was strongly advised. The figure shows the close alignment of trends between expected and observed levels, illustrating how both are similarly influenced by meteorological effects during the period.

Overall, we found a 40% difference from mid-March to mid-June 2020 – i.e. NO2 pollution levels were 40% less than what the model predicted during lockdown. This is the average change across London’s different monitoring site types, including those close to roads (kerbside and roadside) and farther away from busy streets (urban background and suburban).

Changes in meteorology over time typically complicate air quality intervention analysis, but a machine learning method like this allows us to better isolate changes associated with interventions, like lockdown measures. This method has been used successfully in other recent air quality research – for example, Grange and Carslaw (2019) – and we will continue to use cutting-edge methods like this to better understand how London’s pollution levels are changing.

This analysis complements our previous lockdown assessment using data from the Breathe London monitoring network. We used data from the regulatory monitors here rather than Breathe London because training the model requires a longer historical record.

References:

This was originally posted to EDF Europe.

Also posted in Academic, Health, London, UK / Authors: , / Comments are closed

Digging deeper into London air pollution reduction with Waze

By Anne Marie Borrego / Published: June 23, 2020

Editor’s Note: Today’s post is by Meg Dupuy-Todd, a Manager at Environmental Defense Fund. Meg has led projects as a part of EDF’s Global Clean Air initiative, including collaborations with Google’s Project Air View in London and Salt Lake City.

Covid-19 has disrupted everyone’s lives, including providing an unfortunate and unwelcome experiment to examine London’s changing commuting and pollution patterns.

At Environmental Defense Fund, we have been collecting air quality data as part of Breathe London, a collaborative project to map and measure pollution in London. We could see that the social distancing measures had changed nitrogen dioxide (NO2) pollution levels, but we wanted to provide a fuller picture and determine what factors might be at play, including traffic.

Here’s how we used Waze for Cities data to dig deeper into the cause of observed NO2 pollution reductions.

NO2 pollution changes

The Breathe London consortium, with partners including Google Earth Outreach, has collected hyperlocal data to help Londoners better understand pollution in the city. Displayed in real-time on an interactive map, the Breathe London project measures and displays data from air quality sensors across the city.

After social distancing was strongly encouraged in the city in mid-March, the Breathe London monitors saw substantial NO2 pollution reductions — across the full network and especially in the city’s Ultra Low Emission Zone (ULEZ) in Central London. The greatest reductions occurred during daytime hours, between 6:00–22:00, when we also expected traffic to be the highest.

FIGURE 1: Breathe London network hourly mean NO2 measurements during 17 March to 13 April 2020 compared to pre-confinement levels.

FIGURE 1: Breathe London network hourly mean NO2 measurements during 17 March to 13 April 2020 compared to pre-confinement levels.

Exploring Waze data

It was a given that restrictions to help reduce the spread of Covid-19 would change traffic patterns, as workplaces and businesses closed and many people began to work from home. As road transport is a major source of NO2 pollution, naturally, our assumption was there was a link with fewer vehicles on the road.

Through partnership with Waze for Cities program, we had access to Waze-generated anonymous incident and slow-down information in Greater London. Could the data help us understand where and how traffic had reduced?

From previous reviews of Waze data, we knew it would not provide us direct information on traffic volume — the number of vehicles on the road. Instead, we relied on the reported congestion information comparing speed on roads as compared to free-flow traffic. Using Google BigQuery, we were able to analyze the large volume of spatio-temporally resolved data to begin to get a picture of daily traffic congestion in London. Using two distinct geographical boundaries (Greater London and the ULEZ), we decided to calculate, for each hour of the day, the total length of roads where traffic was less than 60% of free flow speed as our proxy for traffic.

With our analysis, we saw that traffic congestion reduced to such an extent that it was approaching free-flow in the vast majority of Greater London roads after the stay-at-home order, even during what used to be peak commuting hours. This is most pronounced in the mostly commercial ULEZ.

FIGURE 2: Waze data — mean total length of congested roads by hour during 17 March to 13 April 2020 compared to pre-confinement levels. (Note difference in scales for Greater London and the ULEZ).

FIGURE 2: Waze data — mean total length of congested roads by hour during 17 March to 13 April 2020 compared to pre-confinement levels. (Note difference in scales for Greater London and the ULEZ).

Putting it all together

With both datasets in hand, our team was able to look at the differences over time of NO2 pollution and traffic congestion. We knew that traffic varied by day of week, so we compared the data to a pre-confinement median by day of week and hour of the day.

We found an apparent association between the reduced pollution levels and lower traffic congestion in London. Examining the daily pattern of traffic congestion also suggests a tie between the biggest drops in pollution and the biggest drops in congestion — which both occur in the late afternoon from around 3 to 7 pm. You can see this depicted in the chart below. Note the difference in “variability in road congestion due to traffic during confinement” (shown in red) and the “weekly congestion average” (shown in green).

FIGURE 3: Breathe London network NO2 measurements during 13 March to 13 April 2020 in comparison to the typical hourly pre-confinement levels.

FIGURE 3: Breathe London network NO2 measurements during 13 March to 13 April 2020 in comparison to the typical hourly pre-confinement levels.

There is still a lot more to learn about the changes in road transport emissions and the relationship to measured pollution, but by incorporating the Waze data into our Covid-19 analysis we were able to shed new light on why NO2 pollution has recently gone down in London.

To learn more about our assessment please visit BreatheLondon.org/covid19.

 

This was originally posted on Medium.

Also posted in London, Monitoring, Partners, UK / Authors: / Comments are closed