Estimating Sunflower Growth Using Earth Observation Satellites
Known for their bright yellow faces, sunflowers are also one of the world’s most crucial oilseed crops. This importance makes it vital to better understand how they are faring throughout their growing season to allow markets and policymakers to respond to potential disruptions as quickly as possible. Scientists with NASA Harvest at the University of Maryland (UMD) are developing new indicators to allow for the satellite-based monitoring of sunflower production around the world.
Oilseeds like soybeans, palm oil, canola, and sunflowers are crushed to produce vegetable oil, an important ingredient for cooking. Oilmeals, protein by-products created from the oil extraction process, can be used as feed in animal farming.
According to a 2020 report the United Nations’ Food and Agriculture Organization (FAO), demand for vegetable oil is expected to increase between 2020 and 2029. One of the major producers of vegetable oil, specifically sunflower oil, is Ukraine. Prior to the Russian invasion in early 2022, Ukraine was the number one producer of sunflower oil, meal, and seed and the number one exporter of sunflower oil and meal.
Challenges to agricultural production and export arising from the invasion and consequent war have changed this status however. Prior to the war, Ukraine produced 1/3 of the world’s sunflower oil and nearly half of global sunflower oil exports. The United State’s Department of Agriculture (USDA) estimates that these numbers have fallen to 21% and 35% respectively; with Ukraine’s share of the global sunflower meal exports falling from 66% to 40% as well. These losses in production and the stress they place on the global oilseed market demonstrate the importance of monitoring remaining sunflower production to ensure policymakers can quickly respond to projected shortfalls.
Fully bloom sunflower flower head facing the east direction.
Pic credit: Sergii Skakun
The Harvest team at UMD decided to exploit a peculiar characteristic of sunflowers—their heliotropism or the ability for the flower’s face to track the sun throughout the day. The sun, rising in the east, meets the flower face of young sunflowers, which then turn to track the sun as it travels across the sky. Once the sun goes down in the west, young sunflowers slowly turn their faces back to the east to greet the sun the next morning. Once sunflowers reach maturity they abandon this practice and remain facing eastward.
Eastward-facing sunflower heads influence development and growth, as they attract pollinators with a stronger visual signal. The morning sun exposure heats them faster, making them more appealing to pollinating insects like honeybees. Additionally, the higher temperature increases evapotranspiration and protects their reproductive organs from excessive heat and harmful ultraviolet light.
Previous studies have utilized optical sensors for sunflower monitoring, but they haven't fully explored the heliotropic and directional behavior or identified flowering stages of sunflowers. The Harvest team was curious if they could capture this heliotropic movement using satellite-based radar systems. The European Space Agency’s Sentinel-1 satellite has a sensor on it that continually sends out radar waves (called synthetic aperture radar or SAR). Similar to how a bat uses echolocation to map its surroundings, the radar waves are emitted, hit the surface of the earth before bouncing back to space, and the sensor records how the waves return.
Heliotropism behavior as observed in the sunflower for descending (green) and ascending (pink) passes. (a) Young sunflower demonstrates heliotropism and gives direct backscattering towards the S1 radar sensor (b) The heliotropism behavior stops in the mature sunflower and the flower permanently orient eastward giving direct backscattering in only the descending pass.
The team found that the amount of returning radar waves did vary based on the sunflower’s phenological phase. Specifically, the stage where they begin to flower (and thus start their daily movement) and the stage where they reach maturity (and thus the ending of their daily movement). By comparing these results to reference data collected from fields within Ukraine, this method was found to be able to detect beginning of flowering (BoF) and end of flowering (EoF) stages within around 4 days.
Traditional crop monitoring, particularly in hard to access areas, can be very time and labor intensive. Developing new methods for remotely monitoring crop production can decrease these resource restraints and allow for improved information that markets and policymakers can act upon. Going forward, the team plans to conduct more field studies and repeat this experiment on a wider range of fields using different types of SAR data for comparison.
This research will pave the way for developing algorithms for operationalizing automatic sunflower mapping. As the war in Ukraine continues, and sunflower production remains volatile, these new methods can help the world respond to supply shocks effectively.
Interested readers can access the published study, including a full breakdown of its methodology and results, here.