New Delhi: Delhi is likely to continue grappling with high levels of particulate pollution through the second half of Nov, with the probability of daily average PM2.5 concentrations exceeding 250 micrograms per cubic metre, the threshold for the severe category, on Nov 21, 26 and 28, according to Cambridge Air Pollution Tracker for Delhi.The daily national ambient air quality standard for PM2.5 is 60 micrograms per cubic metre, while World Health Organization’s daily safe limit is 15.This tracker has been developed by Paul Kattuman, professor of economics at Cambridge Judge Business School, and forecasts particulate pollution and estimates the probability of specified concentration thresholds being exceeded. It is designed to give public agencies, health services, schools, employers and residents advance warning of high pollution episodes.“We flag a day when there is at least a one-in-four chance (25%) that the daily average across the network will exceed 250µg/m3, the boundary for Central Pollution Control Board’s severe band for 24-hour PM2.5. The estimated chance was just under 30% for Nov 21 and 27% for Nov 26 and 28,” Kattuman told TOI.“What matters most for mitigation is how reliably we can anticipate the start of a surge and judge how long it is likely to last on average,” he said. “Pollution levels can swing a lot from one day to the next. It is impossible to switch mitigation measures, which affect people’s livelihoods, on and off at that pace.”The tracker’s forecast also flags elevated PM10 levels. The probability of daily average PM10 exceeding 350µg/m3 ranges from 27.8% to 67.9% from Nov 16 to 29.Before this pollution tracker, Kattuman and his colleague Andrew Harvey developed a statistical model to help health authorities plan ahead during the Covid-19 pandemic. “Our forecasts were used by National Health Service in East England and health departments in Kerala, Punjab and Tamil Nadu to plan their response,” Kattuman said.“Having lived in Delhi and with many friends still there, I felt a personal connection to the city’s winter pollution crisis. We hope to extend the work to other Indian cities facing severe pollution,” he added.Drawing hourly data from 32 continuous ambient air quality monitoring stations, shared by CPCB, and past data from Jan 2020, the tracker uses a statistical model. “We forecast PM2.5 and PM10 separately, adjusting the forecasts for the timing of festivals and for the residue-burning season…We also forecast how high pollution may rise during the day,” he said.Asked how this tracker differs from the air quality forecasts issued by IITM and IMD, Kattuman said, “IITM in Pune and IMD use chemical transport models… with estimates of what traffic, industry and fires release into the air and use weather forecasts to work out where winds will carry those emissions, how widely they will disperse and how much rain may wash out… Our tracker complements that work with a statistical outlook over a fortnight.”He added that sudden unanticipated changes in weather and effective mitigation measures could reduce the accuracy of the forecasts issued by the tracker.The previous forecast by the tracker significantly overestimated pollution after Diwali, as actual concentrations fell faster than predicted. “We expected pollution to ease after Diwali, but we got both the timing of the peak and the speed of the fall wrong for the daily PM2.5 average,” said Kattuman.Asked what changes he would recommend to Delhi govt, Kattuman said, “Act on year-round sources of pollution well before the winter peak. Focus inspections on high-emitting vehicles, industrial premises and dust control at construction sites. Publish what inspectors find, and whether follow-up checks showed the improvements lasted.”Suggesting coordination with neighbouring states to provide farmers with practical alternatives to stubble burning, he said, “Agree in advance how forecast warnings should inform action, taking account of both the pollution risk and the disruption restrictions can cause. Keep a record of when each warning was available, what action followed and what the monitors subsequently showed”
