Assessing Effectiveness of India’s Industrial Emission Monitoring Systems

March 18, 2024

Hemant Mallya, Sankalp Kumar, and Sabarish Elango

Suggested citation: Suggested citation: Mallya, Hemant, Sankalp Kumar, and Sabarish Elango. 2024. Assessing Effectiveness of India’s Industrial Emission Monitoring Systems. New Delhi: Council on Energy, Environment and Water.


Indian industries contribute around 30 per cent to India’s particulate matter pollution. Urgent measures are needed to improve monitoring and compliance to keep industrial pollution in check as the economy continues growing. Continuous emission monitoring systems (CEMS) can bridge this compliance gap. CEMS are devices that measure pollutant concentrations and flow in industrial chimney stacks and relay this information at frequent intervals to the state and central pollution control boards (SPCBs and CPCBs). The CPCB’s guidelines (issued in 2014, 2017, and 2018) and the Supreme Court’s ruling (in 2017) have mandated the installation of online CEMS (OCEMS) in various highly polluting industries and the public availability of the CEMS data in online portals maintained by each SPCB. However, there has been no change in the relevant acts and regulations to make CEMS data legally admissible, i.e., there is no mandate to allow the SPCBs or the CPCBs to issue non-compliance notices to industries on the basis of CEMS data.

This study, as part of the USAID-supported Cleaner Air and Better Health (CABH) project, analyses the public accessibility and quality of the CEMS data relayed by various state OCEMS portals. It makes recommendations for consolidating CEMS regulations and providing legal sanctity to CEMS data. The study ranks 20 publicly accessible OCEMS portals (out of 32 states/UTs requiring CEMS to be installed as of 31 December 2021) for ease of public access and transparency. The study also evaluates the quality of CEMS data relayed by six states and six industry types in 2019 and 2020. The data quality was evaluated on the basis of number of hours of data availability over each year, frequency of missing data, and compliance to guidelines on different emission parameters to be monitored. The study also makes recommendations regarding integrating ambient air quality monitors within industrial facilities into the national network for more accurate air quality monitoring.

Key highlights

  • Out of 32 states/UTs with industries mandated to install CEMS, only 20 had publicly accessible OCEMS portals (as of 31 December 2021). Only 9 of these 20 portals provided the option to view historical data older than one month. Only 6 of the 9 allowed the option to download the data to carry out any analyses.
  • These 20 portals were ranked for accessibility and transparency based on scoring in seven parameters: geotagging of facilities, public access to the portals, ease of access, data download facility, stack compliance indication, data reporting interval, and data duration.
  • The Madhya Pradesh and Himachal Pradesh OCEMS portals were found to be the best in terms of accessibility and transparency, obtaining full scores in six parameters.
  • For the six states (Assam, Chhattisgarh, Goa, Himachal Pradesh, Madhya Pradesh, and Meghalaya) that allowed data download, the quality of CEMS data was evaluated for six highly polluting industry sectors (steel, cement, refinery, petrochemicals, aluminium, and pulp and paper).
  • The CPCB guidelines mandate more than 85 per cent data availability, i.e., data should be available for 85 per cent of 730 hours each month. However, even on an annual basis, none of the six states and industry types had an average availability over 75 per cent.
  • The Assam SPCB had the lowest average availability in 2020 (10 per cent), while Meghalaya had the highest (67 per cent). Among industries, refineries had the lowest (36 per cent) and aluminium had the highest (71 per cent).
  • As per the CPCB guidelines, a single missing data even should not last more than 72 hours; if it does, then the facility may be ordered to stop operations. The study found that while less than 7 per cent of the missing data events lasted longer than 72 hours, these events alone contributed to more than 92 per cent of the total missing hours of data (more than 1000 hours in 2019 and 2020 for most CEMS).