Harvesting Insights From Crop Data

In “When Farmland Becomes the Front Line, Satellite Data and Analysis Can Fight Hunger” (Issues, Winter 2024), Inbal Becker-Reshef and Mary Mitkish outline how a standing facility using the latest satellite and machine learning technology could help to monitor the impacts of unexpected events on food supply around the world. They do an excellent job describing the current dearth of public real-time information and, through the example of Ukraine, demonstrating the potential power of such a monitoring system. I want to highlight three points the authors did not emphasize.

First, a standing facility of the type they describe would be incredibly low-cost relative to the benefit. A robust facility could likely be established for $10–20 million per year. This assumes that it would be based on a combination of public satellite data and commercial data accessed through larger government contracts that are now common. Given the potential national security benefits of having accurate information on production shortfalls around the world, the cost of the facility is extremely small, well below 0.1% of the national security spending of most developed countries.

Second, the benefits of the facility will likely grow quickly, because the number of unexpected events each year is very likely to increase. One well-understood reason is that climate changes are making severe events such as droughts, heat waves, and flooding more common. Less appreciated is the continued drag that climate trends are having on global productivity, which puts upward pressure on prices of food staples. The impact of geopolitical events such as the Ukraine invasion then occur on top of an already stressed food system, magnifying the impact of the event on global food markets and social stability. The ability to quickly assess and respond to shocks around the world should be viewed as an essential part of climate adaptation, even if every individual shock is not traceable to climate change. Again, even the facility’s upper-end price tag is small relative to the overall adaptation needs, which are estimated at over $200 billion for developing countries alone.

Third, a common refrain is that the private sector (e.g., food companies, commodity traders) and national security outfits are already monitoring the global food supply in real time. My experience is that they are not doing it with the sophistication and scope that a public facility would have. But even if they could, having estimates in the public domain is critical to achieving the public benefit. This is why the US Department of Agriculture regularly releases both its domestic and foreign production assessments.

The era of Earth observations arguably began roughly 50 years ago with the launching of the original Landsat satellite in 1972. That same year, the United States was caught by surprise by a large shortfall in Russian wheat production, a surprise that reoccurred five years later. By the end of the decade the quest to monitor food supply was a key motivation for further investment in Earth observations. We are now awash in satellite observations of Earth’s surface, yet we have still not realized the vision of real-time, public insight on food supply around the world. The facility that Becker-Reshef and Mitkish propose would help to finally realize that vision, and it has never been more needed than now.

David Lobell

Given the current global food situation, the importance of the work that Inbal Becker-Reshef and Mary Mitkish describe cannot be emphasized enough. In 2024, some 309 million people are estimated to be acutely food insecure in the 72 countries with World Food Program operations and where data are available. Though lower than the 2023 estimate of 333 million, this marks a massive increase from pre-pandemic levels. The number of acutely hungry people in the world has more than doubled in the last five years.

Conflict is one of the key drivers of food insecurity. State-based armed conflicts have increased sharply over the past decade, from 33 conflicts in 2012 to 55 conflicts in 2022. Seven out of 10 people who are acutely food insecure currently live in fragile or conflict-affected settings. Food production in these settings is usually disrupted, making it difficult to understand how much food they are likely to produce. While Becker-Reshef and Mitkish focus on “crop production data aggregated from local to global levels,” having local-level data is critical for any groups trying to provide humanitarian aid. It is this close link between conflict and food insecurity that makes satellite-based techniques for estimating the extent of croplands and their production so vital.

This underpins the important potential of the facility the authors propose for monitoring the impacts of unexpected events on food supply around the world. Data collected by the facility could lead to a faster and more comprehensive assessment of crop production shortfalls in complex emergencies. Importantly, the facility should take a consensual, collaborative approach involving a variety of stakeholder institutions, such as the World Food Program, that not only have direct operational interest in the facility’s results, but also frequently possess critical ancillary datasets that can help analysts better understand the situation.

While satellite data is an indispensable component of modern agricultural assessments, estimation of cropland area (particularly by type) still faces considerable challenges, especially regarding smallholder farming systems that underpin the livelihoods of the most vulnerable rural populations. The preponderance of small fields with poorly defined boundaries, wide use of mixed cropping with local varieties, and shifting agricultural patterns make analyzing food production in these areas notoriously difficult. Research into approaches that can overcome these limitations will take on ever greater importance in helping the proposed facility’s output have the widest possible application.

In order to maximize the impact of the proposed facility and turn the evidence from rapid satellite-based assessments into actionable recommendations for humanitarians, close integration of its results with other streams of evidence and analysis is vital. Crop production alone does not determine whether people go hungry. Other important factors that can influence local food availability include a country’s stocks of basic foodstuffs or the availability of foreign exchange reserves to allow importation of food from international markets. And even when food is available, lack of access to food, for either economic or physical reasons, or inability to properly utilize it can push people into food insecurity. By combining evidence on a country’s capacity to handle production shortfalls with data on various other factors that influence food security, rapid assessment of crop production will be able to fully unfold its power.

Friederike Greb

Rogerio Bonifacio

Inbal Becker-Reshef and Mary Mitkish use Ukraine to reveal an often-overlooked impact of warfare on the environment. But it is important to remember that soil, particularly the topsoil of productive farmlands, can be lost or diminished in other equally devastating ways.

Globally, there are about 18,000 distinct types of soil. Soils have their own taxonomy, and the different soil types are sorted into one of 12 orders, with no two being the same. In the case of Ukraine, it has an agricultural belt that serves as a “breadbasket” for wheat and other crops. This region sustains its productivity in large part because of its particular soil base, called chernozem, which is rich in humus, contains high percentages of phosphorus and ammonia, and has a high moisture storage capacity—all factors that promote crop productivity.

Even as the world has so many types of soil, the pressures on soil are remarkably consistent across the globe. Among the major source of pressures, urbanization is devouring farmland, as such areas are typically flat and easy to design upon, making them widely marketable. Soil is lost from erosion, which can be gradual and almost unrecognized, or sudden, as following a natural disaster. And soil is lost or degraded from salinization and desertification.

So rather than waiting for a war to inflict damage to soils and flash warning signs about soil health, are there not things that can be done now? As Becker-Reshef and Mitkish mention, “severe climate-related events and armed conflicts are expected to increase.” And while managing such food disruptions is key to ensuring food security, forward-looking polices and enforcements to protect the planet’s base foundation for agriculture would seem to be an important part of food security planning.

In the United States, farmland is being lost at an alarming rate; one reported study found that 11 million acres were lost or paved over between 2001 and 2016. Based on those calculations, it is estimated that another 18.4 million acres could be lost between 2016 and 2040. As for topsoil, researchers agree that it can take from 200 to 1,000 years to form and add an additional inch in depth, which means that topsoil is disappearing faster than it can be replenished.

While the authors clearly show the loss of cultivated acreage from warfare, to fully capture the story would require equivalent projections for agricultural land lost to urbanization and to erosion or runoff. This would then paint a fuller picture as to how one vital resource, that of topsoil, is faring during this time of farmland reduction, coupled with greater expectations for what each acre can produce.

Joel I. Cohen