To address the challenge of understanding the pollutant emissions from freely disseminated satellite data, we tried to understand a deeper aspect i.e.whether the satellite data reflect the ground observations or not. The idea behind the work can be best described as follows:

• The challenge of comparing the sensitivity of freely available ESA’s Sentinel 5P products of NO2, SO2 and CO concentrations with more precise ground-based measurements has been addressed in this work.
• Earlier studies have separately used measurements from satellite (Sentinel 5P) and ground-based stations to observe the changing patterns in the concentration of various gases and comment on their effects due to the sudden change in atmospheric conditions due to COVID-19.
• These studies show that there has been a visible reduction of pollutant levels using either of the sources. However, comparing the data from these sources have not been attempted yet.
• In this work, we have tried to compare the two sources of pollutant gases’ concentration: from ground-based pollution monitors and freely available ESA’s Sentinel 5P products. As the study is concentrated on few Indian cities, the recorded ground data available from the Central Pollution Control Board (CPCB), a statutory body of the Ministry of Environment, Forest and Climate Change, Government of India, has been utilized. The CPCB data is freely available (https://cpcb.nic.in/) for several cities throughout the country.
• To take into account the effect of COVID-19 on the pollutant levels, two year (2019 and 2020) data for 4.5 months (January 1st - May 15th) have been used. A total of 135 and 136 days of data have been processed for 2019 and 2020 respectively.
• An additional challenge while comparing the two datasets is that one is point-data (ground station data) and the other is a coarse resolution (~ 7km x 3.5-5 km) continuous data (satellite data). This was addressed by interpolating ground observations to a scale that could be compared to that of remotely sensed data.
• The correlation between gaseous concentrations from ground data and satellite are calculated for the above-mentioned 4.5 months time period over the interpolated areas.
• To rule out any error due to interpolation, point-wise (ground station location) comparison with satellite data (extracted for that corresponding latitude and longitude), has been carried out. For this, a control experiment scenario (Lockdown period) is identified and analysis is performed for pre-lockdown (month of January) and post-lockdown (month of April) for the ground station points.
• Statistical analysis is performed to derive an empirical relation between ground data and satellite observations.