Croplands cover over half of India’s land area, and the agricultural sector employs about 590 million people in the country.
Researchers can’t rely on census data alone to monitor agriculture in India because many farms are small (less than 2 hectares), whereas censuses typically report statistics at the state or national level and mask important heterogeneity. Remote sensing provides an essential vantage point for monitoring agriculture in India and investigating ways to sustainably improve crop yields.
Dr. Meha Jain, assistant professor in the School for Environment and Sustainability at the University of Michigan, uses remote sensing to research agriculture in India. Jain has used a number of remote sensing instruments to better understand agricultural productivity at the farm level. Her more recent work has uncovered a method to use remote sensing to improve crop yields on poor-performing fields.
Mapping India’s Farmlands
In 2017, Jain developed a method to map farms in India at fine spatial scales using remote sensing data from multiple sensors. Jain mapped winter cropped areas for most of India by combining Landsat data (30-meter resolution) with Enhanced Vegetation Index (EVI) data (250-meter resolution) from the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument aboard NASA’s Terra satellite, taken at the time of peak crop productivity (peak phenology). EVI is used to quantify vegetation greenness and is similar to Normalized Difference Vegetation Index (NDVI). EVI is more sensitive in areas with dense vegetation.
The resulting dataset depicts annual percent winter cropped area per 1-kilometer grid cell from 2001 to 2016. This dataset was published at NASA’s Socioeconomic Data and Applications Center (SEDAC) as part of their India Data Collection.
Jain recently used this dataset, along with census irrigation data, to estimate how much winter cropped area would be reduced if farmers lacked access to critically-depleted groundwater. Published in Science Advances, the research found that India could lose 20% of its winter crop production nationally if farmers in areas with over-exploited water basins lost access to groundwater.
Closing Yield Gaps
Nitrogen fertilizer is one of the major limitations for wheat production in some parts of India. Jain uses remote sensing to better understand yield gaps, which are the difference between actual yields a farmer gets and possible yields based on local growing conditions.
Jain and her colleagues recently published research using commercial small satellite imagery to track how mechanical fertilizer spreaders could close yield gaps in low-yielding farms. Mechanical fertilizer spreaders help farmers distribute fertilizers across a field more evenly than hand-broadcasting. These low-cost tools could improve farmer income and nitrogen use efficiency.
The study authors tested how much mechanical spreaders could improve yields on farms in Bihar, in eastern India, where Jain and her colleagues conducted their research. First, they tested how much wheat yields differed on 127 farms when the mechanical spreaders were used on half of the field, and business-as-usual practices were used on the other half of the field, using direct measurements of crop cuts. They found that switching from hand-spreading fertilizer to using a mechanical spreader improved yields by about 4.5%. Improving nitrogen use efficiency on farms is a key technique for sustainably increasing food production, but these findings relied on crop cut data, which isn’t readily available over large areas.
In order to scale up this research, Jain and her colleagues developed an empirical model of yields using crop cut measurements and satellite-derived measurements of green chlorophyll from commercially available smallsat imagery acquired by PlanetScope and SkySat satellites operated by Planet Labs Inc. The high-resolution commercial imagery enabled them to develop a reliable model of crop yields across a wider area. Planet imagery is available to U.S. government-funded researchers through NASA’s Commercial Smallsat Data Acquisition (CSDA) Program.
They leveraged this model to identify lower yielding fields to see if targeted use of mechanical spreaders on these lower-yielding farms would improve yields even more. They found that targeting the use of mechanical spreaders on low-yielding farms resulted in yield gains that were twice as large as using the mechanical spreaders without data on farm productivity. Jain and her colleagues also found that farmers of low-yielding farms were willing to pay more than the cost of the mechanical spreaders, indicating this technology could be a viable and cost-effective way to improve yields. These findings demonstrate that satellite data can provide crucial information to farmers to sustainably improve crop production.
