With technological advances in Earth observation, we may be only a few years away from screening commodity index funds for environmental risks the same way we account for other financially material risks in public equities today.
The expansion of agricultural commodities production — especially in the tropical context — now contributes to the loss of an area of forest the size of 48 football fields every minute. In addition to creating the second-largest source of anthropogenic greenhouse gas emissions, this is utterly unsustainable for producers and buyers that want to ensure long-term price and supply stability in commodities markets. Understanding and managing land use and land conversion in global supply chains, in a spatially explicit way, is therefore increasingly critical for sustainable agribusiness sourcing and operations.
Fortunately, global agribusinesses and investors alike are beginning to understand and manage unsustainable land use as the financially material risk it is. Leading agribusinesses now have actionable ambitions to monitor and improve land use and land conversion in their global supply chains — but struggle with the cost and feasibility of monitoring solutions.
You see, historically, monitoring of forests and land use change through remote sensing has involved a compromise between resolution and frequency of images available, as well as limited computing capacity for analysis. In general, the optimization of these choices has been visually interpreting annual, low-resolution data at national scales — severely limiting our capacity for actionable insights, particularly in the private sector, to quantify and correlate deforestation risks.
However, recent developments in the spatial and temporal resolution of satellite imagery (Fig. 1), as well as in machine learning and image analysis at global scales, offer new opportunities to disrupt these historical limitations. Planet — a mission-driven aerospace and data analytics company — has leveraged the consumer electronics revolution to launch the largest constellation of Earth-observing satellites in human history: over 100 small satellites are now imaging the full Earth, in four spectral bands at high resolution, every single day. Paired with the power of cloud computing and machine learning, Planet’s dataset makes real-time insights on global change — at the pace and scale at which they occur — both practical and cost-effective.
The 2020s: The decade of regenerative agriculture?
Join us as PepsiCo, Timberland and more discuss their efforts to optimize and future-proof their agricultural supply chains through regenerative practices — October 19 at SB'21 San Diego.
Fig. 1. Traditional full-Earth observation data has been limited in spatial and temporal resolution; whereas Planet’s new technologies image the full Earth in relative high-resolution daily.
Take land use monitoring in Brazil, for example. Brazil is one of the most biodiverse countries in the world, home to the largest rainforest, but it is also one of the world’s largest producers of grains and beef. Despite zero-deforestation pledges from agribusinesses, rapid deforestation for agricultural expansion has persisted as a difficult problem to manage. This could soon change. Leveraging Planet’s daily satellite imagery and Google Earth Engine’s cloud-computing capabilities, Mapbiomas has developing automated land use change detection and classification models at scale in the country. By using algorithms that can detect objects in the daily high-resolution satellite imagery — objects such as pivot irrigation systems or cattle troughs, for example — these models can send an automated alert when forested lands are beginning to be converted for agricultural production.
Fig 2. Combining Planet imagery & machine learning to detect land conversion to agricultural production in Brazil
In an increasingly climate-constrained world, agribusinesses and investors alike increasingly recognize the financial materiality of managing land use in global commodities production and supply. Finally, their ambitions can now be met by practical and cost-effective management tools.
With technological advances in Earth observation, cloud computing and machine learning, it is not crazy to think that we are only a few years away from automatically screening commodity index funds for environmental risks the same way we account for other financially material risks in public equities today.