Earth Matters

Using Satellites to Size Up the Severity of Crop Fires in Northern India

February 8th, 2017 by Adam Voiland with Hiren Jethva

NASA Earth Observatory images by Joshua Stevens, using VIIRS data from the Suomi National Polar-orbiting Partnership and the Fire Information for Resource Management System (FIRMS). The map shows fires detected on November 2, 2016.

When I was writing about the crop fires in northern India last fall, it was obvious that 2016 was a pretty severe burning season. For several weeks, large plumes of smoke from Punjab and Haryana blotted out towns and cities along the Indo-Gangetic plain in satellite images.

But I didn’t realize just how severe the fires were until Hiren Jethva, an atmospheric scientist at NASA Goddard Space Flight Center, crunched the numbers. By analyzing satellite records of fire activity, he found that the 2016 fires were the most severe the region has seen since 2002 in regards to the number of fire hot spots satellites detected. In regards to the amount of smoke detected, the 2016 burning was the most severe observed since 2004. He used data from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor on Aqua and the Ozone Monitoring Instrument (OMI) on Aura to reach his conclusions.

Smoke and fire in northern India have become common in October and November during the last three decades because farmers increasingly use combines to harvest rice and wheat. Since these machines leave stems and other plant residue behind, farmers have started to use fire to clear the leftover debris away in preparation for the next planting.

For more details about how 2016 compared to past years, see the charts below, which Jethva prepared. His explanation for each chart is in italics.

Aqua Detected More Fires in 2016 Than During Any Year Since 2002

Chart by Hiren Jethva based on MODIS data.

The satellite-based sensor MODIS can detect the signal of fire hot spots, also called thermal anomalies, because the signal measured by the sensor in space in the thermal infrared bands appears to be an anomaly compared to the signal emanated from the background land. Since its launch in 2002, the MODIS on NASA’s Aqua satellite has detected thermal anomalies such as wildfires, agricultural fires, and gas flares on a daily basis.

The yearly evolution of total number of fires and Fire Radiative Power (FRP) — the heat energy produced from these fires — detected over Punjab and Haryana showed 2016 to be an anomalous year, with the highest number of crop residue fires (18,707) and the highest FRP in relation to the fires in all other years over the region. In comparison to 2015, the total number of fire hot spots detected over the region in 2016 was 43 percent higher; the difference is 25 percent if the hot spot counts are averaged over the last five years, i.e., 2011-2015. A careful look at the time-evolution of fire counts also reveals an increasing trend in the total number of fires over the region.

Punjab Skies Were Unusually Smoky 

Chart by Hiren Jethva based on OMI data.

These fires produced huge amounts of fine aerosol particles and trace gases, which can potentially impact the climate and degrade air quality drastically at ground level. NASA’s A-train sensors such as the Ozone Monitoring Instrument (OMI) on the Aura satellite and the MODIS on Aqua offer capabilities to measure the total amounts of airborne particles. The UV Aerosol Index (UV-AI), which is an excellent indicator of the column amounts of light-absorbing particles in clear as well as cloudy atmospheres, showed 2016 was the smokiest season on record since 2004.


Greener Fields and Larger Harvests Lead to More Fires

Many studies have shown that satellite measurements of the “greenness” of  crop fields prior to harvest and crop yield after the harvest are strongly correlated. The normalized difference vegetation index (NDVI), which is derived from satellite measurements of radiation at the red and near-infrared light, is one useful measure of greenness. As seen in the charts above, there seems to be a one-to-one relationship in NDVI measured by the MODIS sensor on Aqua prior to harvest (September) and the total number of fire hot spots observed during harvest season (Oct-Nov). This suggest that the increase in the number of fires is likely related to increasing crop yields.

Tags: , , ,

7 Responses to “Using Satellites to Size Up the Severity of Crop Fires in Northern India”

  1. James Varghese says:

    Thanks Adam, Hiren and Joshua for this useful article. No wonder air pollution levels were off the charts in New Delhi during November 2016. The article ‘A Stream of Smoke in Northern India’ published on November 5, 2016 is an eye opener and helped to correlate the facts on the ground with satellite information from above. There are multiple (human activities, natural sources, meteorological conditions – inversion etc.) reasons for New Delhi being the most polluted city in the world. However, at this point I would like to ask specifically, ‘Is there any way to differentiate human caused pollution and the ones produced naturally through geo-information science and earth observation?.’ What is the current state of the art – how far have earth observing satellites combined with ground measurements been able to address the first question? I have been scratching my head with these questions for quite some time now. Any help to solve this mystery would be highly appreciated. Even better if you have any existing articles that I can read and follow.

    Once again, thanks a lot for making science work!

    • Fleta says:

      You forgot bacon. Better than Chanel or Gu-ccii-you smear a little bacon on those pulse points and they'll follow you anywhere! (and change faucets and carry laundry and…) Oh, and stick a piece in your pocket for the afterglow snack. 😉

  2. R. Joshi says:

    In the Indian Himalayan uplands of Uttarakhand State located west of Nepal, the previously busy tourist months of May and June have now become a time of a brown haze, when the Himalayas are no longer visible and the normally blue sky becomes a smokey gray . No one is able to say for sure what is causing it. Conjectures range from local (pine) forest fires which are common in summers, to smoke from factories, power stations and vehicles in the north Indian plains, to the pall of smoke from grassland fires in southern continental Africa, the Sahel or in Indonesia. Usually, the haze gets washed away after the first few Monsoon showers, but 2016 was an exception, as the haze persisted even during the Monsoons, ruling out local forest fires as the main cause. It was only after the winter had set in, when the cold air settles down in the plains as a smog, that the mountain air finally became clear again. It should be possible for NASA to investigate this phenomenon thoroughly, so that some corrective action can be taken.

  3. Hiren Jethva says:

    James Varghese,

    Thanks for your comments on the NASA EO blog on Punjab crop fires.

    As you have pointed out, Delhi’s air pollution is a multi-faceted problem where human activities (vehicles, power plant, industries, construction work), natural sources (local and transported dust), transported pollution such as from crop fires in Punjab and Haryana, and meteorological conditions play their respective roles in modulating the concentration of particulate matter in Delhi. In this blog, we highlighted that crop-residue fires in Punjab and Haryana elevate the concentration of PM2.5 dramatically during October and November.

    With regard to your question on differentiating between human-caused versus natural sources of pollution, satellite observations have shown some potential to address this very important problem. However, there are plenty of challenges that satellite-based algorithms face to separate anthropogenic versus natural sources of pollution, particularly over land due to its heterogeneous spectral and spatial signal. Ground-based remote sensing and in situ measurements are more accurate and meaningful, and therefore play an important role in validating the satellite observations. A synergistic use of ground, airborne, and satellite measurements in conjunction with atmospheric model seems to be the best way to fully characterize the sources of pollution.

    Hope this helps.


    Dr. Hiren Jethva
    Research Scientist
    Universities Space Research Association (USRA)/GESTAR
    Code 614, Room#A422, NASA Goddard Space Flight Center
    Greenbelt, MD 20771

    Ph: +1-301-614-5225

  4. James Varghese says:

    Dr. Hiren, thanks a lot for your valuable response. I was also thinking along similar lines. Indeed a synergistic approach seems to be a reasonable tool in tackling the ‘discrimination between anthropogenic vs. naturally caused pollution’ since a single method such as ground based or airborne or satellite based remote sensing or atmospheric models in isolation may not provide a holistic solution to this problem.

  5. Boss says:

    AFJ. B.L.H. I try to do that as I go along, I tend to revise my work daily. Like, I’ll write 2 thousand words, and then spend the rest of the day working on them till I’m happy with them. Then the next day, I look at them again and do some more revising before writing new stuff. Some writers work right to the end of the story before revising, but that doen’t really work for me.

  6. SHIVANI says:

    Dear Sir

    Please tell me how can we access the images for fire detection for different time period.


    Shivani, JRF
    New Delhi, India