Nearly ten years of air pollution data for dozens of Indian cities – among the most polluted in the world – was deleted from the latest World Health Organization Air Quality database just before WHO published the report on 4 April 2022, Health Policy Watch has learned. 

The Indian air quality data, spanning the years 2010 to 2019, and including 1,139 data points in all, was collected largely from published government sources, including India’s Central Pollution Control Board, and the government’s SAFAR air quality research network, supplemented by the US State Departments AirNow network, which monitors air quality from missions and embassies around the world.  

The suppressed Indian data is reproduced in full here by Health Policy Watch from the embargoed version of the WHO release, alongside the data of other South-East Asian Region cities and communities.

India is among eight other countries in WHO’s South-East Asia region whose data was originally included in the report.  Others included neighbouring Bangladesh, Nepal, and Sri Lanka, as well as Bhutan, Indonesia, Maldives, Myanmar, and Thailand. 

Except for India, all of the other countries’ data from the region remained intact in the public excel version of the database that was put online by WHO at its 4 April launch. 

Why is this important?

The omission is significant as the WHO database was for many years regarded as a gold-standard baseline for assessment and comparison of countries’ air quality challenges. While it has been superseded, somewhat, by a range of other open-access and research-based data collection efforts, WHO’s work remains a critical benchmark for national, regional and global analysis of air pollution challenges in different cities and regions. 

Alongside more cutting edge satellite data and other ‘modelled’ data sources, the ground station monitoring data collected by WHO also is a major input to the global bosy’s assessments of air pollution exposures, illness and premature deaths. 

An estimated 4.2 million people die every year from diseases attributable to ambient (outdoor) air pollution levels – with the largest burden in the South East Asia region, according to WHO.  About 7 million people die from combined ambient and household air pollution exposures, with some overlap between the two risk factors.  WHO estimates that in India alone, nearly 1.8 million people die from air pollution every year. 

WHO – No response to queries  

Health Policy Watch repeatedly sent queries to WHO officials and spokespeople asking why the Indian data had been wiped clean from the 2022 version of the WHO database. WHO initially did not reply. 

However, following publication of this report Thursday, a WHO scientist told Health Policy Watch, “The data for India will be released soon. The data is still going through final checks to ensure accuracy.”  The scientist rejected suggestions that the data had been pulled at the last minute as a result of political pressures – but rather out of an abundance of caution with the sensitive data set.

The WHO trumpeted that its latest WHO Air Quality database was its largest collection of ground station monitoring data ever. And indeed it was. 

The publicly available excel sheet contains new data on average annual particulate concentrations from around 6,000 cities and settlements worldwide – but Indian data is nowhere to be found in the new collection of concentrations of PM10 and PM2.5 – the pollutants most directly linked with a range of cardiovascular, respiratory diseases and cancers, and related premature deaths. 

The 2022 database, the first WHO update since 2018, also included some more limited data on NO2 concentrations for the first time ever – a common pollutant from vehicle emissions that is a precursor to ozone, stunting food production exacerbating asthma and other chronic respiratory diseases.  The embargoed version of the WHO database also reflected new reporting on Indian NO2 data, which also was deleted from the published version.  

India’s data – controversial from the beginning of the WHO effort

India’s data has been a regular feature of the WHO Air Quality database for nearly a decade when the global health agency first started amassing data in order to build a more accurate model of disease and deaths from air pollution – a then poorly-recognized factor in cardiovascular and respiratory diseases as well as cancers. 

The collection of such data was a particularly bold move for the agency at the time.  And it caused a big stir worldwide – particularly in India, which emerged as one of the world’s most polluted countries – overtaking China which had begun making strenuous efforts to clean up its skies. 

Both in private and publicly, Indian air quality officials repudiated WHO’s findings, which placed Indian cities at the top of the global air pollution charts, saying that the WHO assessment provided a distorted view of the state of the country’s air. The Indian officials complained that WHO was overestimating air pollution levels in Delhi, despite the fact that the city faces an air pollution emergency almost every winter as crop waste burning in rural areas to the north drive smoke to the city, combining with Delhi’s already dirty urban sources, transforming the city into a pollution nightmare.   

The government’s objections were often linked to highly technical issues such as: a) the reliance upon one air pollution monitoring station in many Indian cities; b) alleged bias toward data from pollution hot spots where stations are placed; c) bias from stations that don’t operate continuously and; d) a lack of reference to the fact that some of India’s urban pollution is due to dust, rather than human sources. 

Over time, those protests have increasingly sounded hollow as more air quality research was published with similar findings. Meanwhile, India’s own air quality monitoring data was constantly improving revealing a similarly dire picture. Any shortcomings that still exist in Indian data, or reporting about it, can be overcome by careful analysis, experts say. 

 

The Indian data included in the 2018 WHO database released four years ago was received in just that light by Indian air quality junkies who combed it for both strengths and gaps. That reflects exactly the kind of transparent debate that is missing from the 2022 release – along with the Indian data itself, of course. 

So if the Indian government had asked WHO to pull the data, it would be particularly hard to understand why now since other reputable global air quality research is now producing similar order-of-magnitude results, year after year. While precise estimates of air pollution levels and peoples’ exposures in different Indian regions and cities might vary across diverse analyses, they tend to reflect the same unhealthy air pollution concentrations across large swathes of northern India, until the present day. 

Other global databases carefully tracking Indian trends 

One such example is the State of Global Air, a report by the Boston-based Health Effects Institute (HEI). HEI also funds air quality research involving Washington University, St. Louis, the University of British Columbia and the University of Washington. Together, those institutions regularly scrape and analyze publicly available Indian monitoring data from official sites. In combination with NASA satellite data and other data, they, too, produce estimates of average annual PM2.5 concentrations and from those inputs, model the levels of “population exposure” to air pollution in India, and worldwide.   

According to some of those recent findings, further distilled by an Indian analysis, and portrayed graphically here, nearly a quarter of India’s population was exposed to annual average PM2.5 concentrations ranging from 80 micrograms/meter3 (µg/m3) or more in 2020 . This is 16 or more times higher than the WHO guideline level of 5 µg/m3 for PM2.5.  That guideline limit was tightened by WHO in 2021 in the face of mounting evidence about the health impacts of particulate air pollution, even at comparatively low concentrations.  

Air pollution worsened over the past two decades 

Looking back further, however, India’s air pollution has worsened significantly over the past two decades.  A comparison of the same mapping from 2000 and 2020 highlights that trend, showing a much wider belt of northern India etched in the dark browns and orange of excessive pollution levels, as compared to the green shades of two decades ago.  

Annual average air pollution exposures appeared to peak in 2014 then had a slight, gradual decline, stagnating at a high level of around 80 µg/m3 for PM2.5, suggests data from the State of Global Air report series. 

WHO data on Delhi – from the embargoed WHO data set never released – reflects similar trends and fluctuations – even if it is urban, not national, data – and framed somewhat differently in technical terms.

According to the WHO data compilation, Delhi’s PM2.5 levels peaked as well around 2014, and then declined somewhat, levelling off recently to concentrations of about 105 µg/m3  – still some 21 times higher than recommended WHO guideline levels of 5 µg/m3. But a deeper analysis by WHO would be needed to be performed to further gauge trends nationally or regionally over time. Needless to say, with the most recent India data missing altogether, that kind of trend analysis hasn’t been done in WHO’s latest Status Report.  

CPCB  officials  say data is freely available on website, no reply from SAFAR  

Health Policy Watch reached out to several Indian government air quality officials for explanations as to why India’s data had disappeared from the final version of the 2020 WHO air quality database. 

Calls to Dr Prashant Gargava, Member Secretary of the Central Pollution Control Board (CPCB), India’s statutory body, went unanswered. 

But two other senior CPCB officials said in separate phone interviews that updated CPCB data is freely available on their website for the general public – and denied there had been any effort to censor WHO’s report. 

“As far as I know, the WHO has not asked us for any data. They are welcome to take it from the website – anyone can do that,” one official told our India correspondent. 

“We have data from our manual as well as our CAQM (continuous air quality monitoring) stations on our website – anyone can take that data – they just have to mention the data source and acknowledge it,” the official who directly supervises this data said.   

An independent Indian air pollution expert interviewed by Health Policy Watch, also confirmed that since the Indian air quality data is publicly available from the official, Central Pollution Control Board, database, the statutory Air Quality monitoring body in India – WHO certainly shouldn’t censor itself over use.  

“Technically CPCB is posting the data online and it is supposed to be public by default for anyone to use, with due credit,” the data scientist  told Health Policy Watch, “Then why not just use it, and not ask for another permission to use it.”   

Other major Indian data sources: SAFAR and the US State Department’s ‘AirNow’

WHO’s suppressed Indian data set also drew upon data from SAFAR-India, a parallel research network operated by the Indian Ministry of Earth Sciences. 

SAFAR, the “System of Air Quality and Weather Forecasting And Research” is an internationally recognized network that collects data from 11 monitoring stations in Delhi, as well as individual stations in 11 other Indian cities. It’s network includes academic, and meteorological research institutions, as well as some municipal authorities. 

SAFAR also provides real-time air quality forecasting and alerts, based on an Air Quality Index, important for air pollution emergencies. Its reporting is particularly important for Delhi, which suffers some of the worst pollution peaks in the country almost every winter, and along with that, the need to make decisions about drastic public health measures, from mask distribution, to work stoppages and school closures.   

A leading air pollution official at SAFAR contacted by Health Policy Watch about the removal of the Indian data from the 2022 WHO database, also did not respond to emailed requests for clarifications, as of the time of this publication. 

Even so, insofar as authoritative Indian data is plentiful, current and public, there should be no problem with WHO using it freely.

The Indian data scientist interviewed by Health Policy Watch stated: “I think WHO should scrape it [the data], or use a group scraping the data, to make their own current averages and publish. This is public information, though it is a little difficult to access data in bulk from the CPCB site, some groups have already put in their time to make it useful.

“My personal opinion (and that of some others) is that when we get stuck in this circle of permissions to even publish what is already public, it makes policy dialogue very difficult, leading to second-guessing to do the right thing, even as simple a task as public awareness of health impacts of air pollution.”

WHO – a scientific or member state body? 

But the pressures WHO faces as a science body, under the constant scrutiny of its member states, tends to complicate its research. In fact, whenever potentially unflattering data lurks, the WHO tends to walk tiptoe around the sensitivities of member states. 

This has been highlighted during the COVID-19 pandemic in WHO’s public statements and relation to member states from China to Italy. In the case of the latter, WHO withdrew from a 2020 report on Italy’s pandemic response, An unprecedented challenge, just a few hours after it was published. Although the official narrative was that the account was deleted was due to “errors”, the report’s chief coordinator, who has since left the WHO, attributed the report’s withdrawal to intense political pressure from the government of Italy and a leading WHO official affiliated with the government, who had found certain findings unflattering.

A pattern of much-delayed reports in WHO’s Disease Outbreak News about new outbreak threats can also likely be traced to excessive concerns with member states’ sensitivities. 

There was a three-month delay in WHO’s publication about the recent wild polio outbreak in Malawi, which occurred in November 2021, but only was reported by WHO in late February 2022, shortly before it launched a five-nation vaccination campaign. A July 2021 Zika outbreak in India’s Kerala and Maharashtra states was only reported by WHO in October 2021, a full three months after the outbreak came to light – with considerable local media coverage.    

In the case of the WHO air quality database, traditionally, data had to be reviewed by countries prior to WHO’s publication. More recently, an online template for government authorities to enter their own data manually is now available on the WHO website. That, WHO says, has become one of the “primary sources of data” potentially also simplifying the official approval process. 

“The primary sources of data were official reports of countries sent to WHO upon request, official national and subnational reports and national and subnational websites that contain measurements of PM10 or PM2.5 and ground measurements compiled in the framework of the Global Burden of Disease project,” state WHO acknowledgements. 

“Measurements reported by the following regional networks were also used: Clean Air for Asia, the Air quality e-reporting database of the European Environment Agency for Europe and the AirNow Programme from the United States embassies and consulates. If such official data were not available, values from peer-reviewed journals were used.”

Missing Indian data went unnoticed in media reports   

Despite the widespread attention that the WHO Air Quality database updates typically receives, major media reporting of the 2022 Air Quality database update, entirely missed the fact that the Indian data had been suppressed from the update, both in India and media reports published abroad. 

Health Policy Watch also reported on the 4 April WHO database launch, with reference to various Indian cities as global pollution hotspots. 

In our report, we explicitly referred to Indian urban data contained in the embargoed database – which we had downloaded from the embargoed report ahead of publication without realizing that the same data had been deleted from the actual, published database that went online that day. 

South-East Asia, Africa and Middle East are World’s Air Pollution Hot Spots in WHO’s Largest-Ever Data Release 

Only a few days later, did it become apparent in an interview with a global air quality expert, that all of the Indian data contained in the embargoed version of the report had been removed from the final report. 

Indian data removed just days before launch

Insofar as the embargoed version of the WHO database is dated 31 March 2022, it’s clear that the Indian data was removed at the very last minute. For those who look diligently, however, historical Indian data from previous years can still, however, be found online in previous versions of the Air Quality database, at least for now.  For instance, at the time of this publication, the following historical data for India could be found in WHO’s:

• 2011 database, the first version of WHO’s public collection efforts, has data for 31 Indian cities, dating from 2008; 
• 2014 database, which had data from 123 Indian cities, including multiple stations in some cities, dating from 2010-2012;
• 2016 database has average annual data from 131 Indian cities, including multiple stations in some cities, all dating from 2012;
• 2018 database update, in contrast, has annual averages from only 31 Indian cities, including multiple stations in many cities, dating from 2016;  

Heavy reliance on US State Department’s Air Now database for new data on major Indian cities

Along with data from CPCB and SAFAR, a third source for Indian data in the embargoed database version was the US State Department’s “AirNow” monitoring system.

AirNow covers cities in the United States, as well as internationally, in sites where the US has a diplomatic presence. Over the past decade, air quality monitoring systems were established in embassy and mission buildings, as a measure to protect the health of diplomatic staff. As real-time continous data is published online, the system serves the secondary function of awareness-raising about air pollution in local communities.

Among the 100 new data points cited from  2017-2019 in the embargoed report, analyzed by Health Policy Watch, values for most of India’s largest cities are attributed to AirNow. For instance, major cities with annual average PM10 data from 2019, attributed to AirNow, include:

• Haryana – 167 µg/m3
• Ahmedabad -119 µg/m3   
• Bangalore – 81 µg/m3  
• Bengaluru – 71 µg/m3  

Those concentrations are 5 to 11 times above the WHO guideline of 15 µg/m3  for PM10.  

For cities with PM2.5 data, AirNow is also cited as the main source. It includes Delhi’s PM2.5 annual average of 105 µg/m3 (2019) which ranked it as the most polluted city in India for that year – 25 times above the WHO guidelines of 5 µg/m3. Other cities with data sourced from AirNow, reported concentrations 8-12 times above the WHO guideline levels for PM2.5, including: 

• Kolkata, 68  µg/m3 
• Mumbai, 43 µg/m3 
• Chennai, 39 µg/m3 

Outside of large cities, however, the 2017-2019 data for other Indian cities was attributed to CPCB or SAFAR.   

Passive data collection – versus ‘data scraping’ 

One noteworthy feature of recent WHO air quality reporting is the fact that the expanding Air Quality database is soliciting and receiving more active contributions by the countries themselves via the online template. 

“The WHO global ambient air quality database is continuously being updated. If you are an official air quality data provider at either the country or city level, please use this template to submit the data to WHO,” states WHO on its air quality theme pages.

While certainly that helps engage countries more actively in the air pollution data collection process, the question is what happens when officials are reluctant to share their data at all?  

It’s possible that in the absence of an active contribution from the Indian government, WHO scraped data available from AirNow and the CPCB for key Indian locations – and then discarded it at the last minute after India objected. But without clear answers from WHO, that remains speculation. 

Other countries’ air quality data is also outdated or missing    

Those are not the only questions about the database, however, that remain unanswered.

While this year’s WHO Global Air Quality database update can indeed boast the largest number data points ever, coverage by region remains highly inconsistent, as WHO’s own maps reveal. 

In some key regions, broad data networks such as the European Environment Agency and Clean Air Asia, also capture lots of data for WHO from countries uninterested or unable to report for themselves.  

But such networks do not extend everywhere. And so, whether as a result of barriers of language, technical incompetence, or pure indifference, cities and even whole countries with data can fall entirely through the cracks. 

Israel, a high-income country considered to be part of WHO’s European Region, has had no new data entries recorded in the WHO database since 2015. That is despite the fact that Israel has a sophisticated air quality network, publishing real time air quality data and annual average PM2.5 values for dozens of cities and locations within its pre-1967 borders. However, that database is in Hebrew and WHO’s data searches are limited to English, French, Spanish and Portuguese, as described in its Status Report on the 2020 update. 

A spot check by Health Policy Watch also found cities in the United States had outdated data. For example, Columbus, Ohio, a major industrial city in the midwest, has not reported an updated value for average annual PM2.5 concentrations since 2011. 

New African data – laudable but still loaded with flaws 

The newest version of the WHO database also is noteworthy for its growing list entries of data from the WHO African Region. 

This year’s update includes nearly 50 data points across more than 10 major African countries, including Kenya, Ghana, Nigeria, Cameroon, Uganda, Senegal – as well as South Africa, which has the largest and most well-established network including some 140 monitoring stations.  

However, even the African data represented here often appears dated and patchy, yet many countries have rapidly growing cities that desperately need more consistent and transparent reporting on air quality. 

Accra in Ghana is one such example.  Home of one of the largest e-waste facilities in the world, it suffers from considerable air pollution from the incineration of plastics and other hazardous materials. In 2015, the city publicly reported its air pollution data for the first time ever – revealing an annual average PM2.5 value of 55 µg/m3 – 11 times more than the WHO Guideline levels. That measurement is now seven years old. Nothing has been reported since. 

And that is despite intensive efforts to strengthen the city’s air quality monitoring efforts through a WHO-led Urban Health Initiative, co-sponsored by UNEP and a joint US-World Bank initiative, which saw the installation of new air quality monitoring equipment in Accra in October 2021.    

Nairobi’s single data point in the WHO database is a value of 5.91µg/m3 of PM 2.5 for the year 2018. 

That value, provided by the US State Department’s AirNow monitoring, may be unrepresentative of the city as a whole if it is drawn from monitors at the US Embassy itself. The Embassy is located in a posh neighborhood on the northern edge of Nairobi just past the famed Karura Forest, which would act as a natural filter for air pollutants drifting from the center city. 

Other available real-time data, as well as peer-reviewed studies of Nairobi’s air quality also suggest much higher levels for the downtown area. But the WHO data doesn’t yet consider new, and increasingly robust methods of low-cost air quality monitoring being rolled out in low-income cities such as Nairobi.  

Data shortcomings: mix of politics and resource shortages 

Taken together, the gaps and shortcomings described here are a complex mix of objective circumstances (e.g. Africa’s air quality monitoring network was largely non-existent only a few years ago); human resource shortages, and long-ingrained institutional and political constraints. 

WHO insiders point to the fact that WHO’s core air quality database team is led by a single staff scientist. That effort is indeed supported by a wider circle of consultants and advisors, including known global experts such as Michael Brauer, at the University of British Columbia, and Gavin Shaddick, of the University of Exeter.  But collaborations with busy professionals accountable to their own institutions can only go so far.  

So while other global air quality monitoring initiatives now boast flashy websites with interactive maps, charts and graphics, that ease public understanding and highlight trends, WHO’s database remains a single PDF file and flat excel sheet download.  That alone makes the job of sorting and assessing values across cities, countries, regions, and over time onerous, to say the least. 

In its global database effort, WHO largely goes it alone. But acting alone, WHO also may be underpowered for the massive task that it tries to fulfil.  

Even so, until WHO and the Indian government reinstate the India data, as well as explaining more transparently why it went ‘missing-in-action’, dialogue about a way forward that is more robust, accessible, up-to-date, and inclusive, of all countries’ data would be difficult.  

Removal was unplanned and occurred at the last minute

But whatever the drivers, it’s clear is that the removal of the Indian data was unplanned and came only at the very last minute.

For instance, at least some of the new Indian data points suppressed in the technical report can be called up on the interactive search tool of the BreatheLife campaign – a joint awareness-raising effort of WHO, the United Nations Environment Programme’s Climate and Clean Air Coalition (CCAC). A search for Delhi, Chennai, Kolkata and Mumbai showed the gauge arrow responding with the same 2019 value for PM2.5 (105µg/m3) that had been deleted from the technical data base.

In addition, the online WHO Status Report, the technical report that accompanied the 4 April database launch, also displays the Indian data points for PM10 and PM2.5 in its global map of monitoring stations from which data was collected. 

That technical summary, still marked with the big bold words DRAFT on its cover and inside pages, even calls out the Indian data monitoring effort in a caption that states: “More ground measurements are generally found in high- and middle-income countries, in China, Europe, India and North America.”

Ultimately, the removal of the India data was not only hasty but awkward for the entire WHO database effort; the end of the story can only be written by WHO and India’s official authorities.  

_________________________________________________________________________

Elaine Ruth Fletcher is the editor-in-chief of Health Policy Watch. From 2003- to 2018 she was a technical officer and editor at the World Health Organization, engaged mostly in WHO’s air pollution and urban health work – including co-founding the BreatheLife campaign dedicated to awareness raising of the health and climate impacts of air pollution. She is the co-author and/or editor of over two dozen WHO reports and peer reviewed journal articles on air pollution, urban health and other environmental health themes. 

Jyoti Pande Lavakare is a core contributor to Health Policy Watch from Delhi. She is the author of the recently published grief memoir on the human cost of air pollution titled “Breathing Here is Injurious to Your Health”, and a long time contributor to Indian and global media on air pollution themes. She is also co-founder of the Indian NGO, Care for Air.

-Updated at 18:50 CEST 14. April 2022 with initial reply from the World Health Organization.

Image Credits: DYFI Delhi Twitter, Flickr, www.aqicn.org, urbanemissions.info, urbanemissions.info, State of Global Air/Health Effects Institute, Embargoed version of the WHO air quality data base, WHO/A. Kari, WHO Ambient Air Quality Update 2022 Status Report .