The forests nobody is watching

The global conversation about deforestation has a familiar cast of villains.

Beef. Soy. Palm oil. Commodities that move markets.

And they deserve the attention. Together, they account for an estimated 70% to 80% of total forest loss driven by agriculture.

They are also just part of the story.

A sweeping new analysis from Chalmers University of Technology, published in recently in Nature Food, has produced what its authors call the most comprehensive global survey yet of how food production drives forest loss. It combines satellite land-cover data with agricultural statistics across 184 commodities in 179 countries, tracking annual deforestation across two decades, from 2001 to 2022. What it found in the familiar territory was expected. What it found everywhere else was not.

Maize, rice, and cassava tell a quieter story. They are staple crops, grown largely for domestic consumption, cultivated primarily by small-time farmers. Yet together, they account for roughly 11% of all agriculture-driven deforestation worldwide.

The finding has implications in the regulatory world, particularly in Europe, where policymakers have spent years reframing deforestation as a supply chain problem rather than a distant environmental one. The premise is simple: if demand in wealthy markets helps drive forest loss abroad, then those markets can also help curb it by requiring companies to prove that what they sell is not linked to recently cleared land. That approach has focused attention on commodities like beef, soy, and palm oil, where global trade is deep, traceability is possible, and regulatory leverage is strongest.

This is not a knock on the regulation. Beef and soy and palm oil still do enormous damage, and the study confirms as much. Pasture expansion alone accounts for 42% of global deforestation and more than half of associated carbon emissions. The point is not that regulators got the wrong answer. The point is that they were working from an incomplete question.

The structural reason staple crops escape notice is geographic. Palm oil deforestation is concentrated in Southeast Asia. Soy deforestation is concentrated in South America. When damage is geographically concentrated, it shows up in satellite monitoring, in supply-chain audits, in the risk benchmarking that compliance teams build. Maize and rice and cassava don’t work that way. Their deforestation is distributed across dozens of countries, driven by millions of small land-use decisions that no corporate sourcing policy reaches and no import regulation touches. The damage is real. It just doesn’t aggregate neatly enough to trip any existing alarms.

The debate on deforestation has focused on how consumption in wealthy countries drives forest loss through imported commodities, and that framing is, of course, directionally correct. But a significant share of deforestation is tied to agricultural production for domestic markets, and that means more holistic solutions must extend beyond import controls to action within producer countries themselves.

This is where things can become politically uncomfortable. The EU deforestation regulation was designed, in part, as a way for wealthy nations to exercise leverage over forest loss without needing to engage in the slow, difficult work of supporting agricultural transformation in lower-income countries. It’s a border measure that’s clean, enforceable, domestically popular. But this new data suggest border measures, however well designed, are’t that great at addressing a lot of the problem.

In total, 122 million hectares of forest disappeared between 2001 and 2022 due to agricultural expansion, an area larger than Peru. That amounted to something like 41.2 gigatonnes of CO₂ emissions. More than 80% of that loss occurred in the tropics, in countries that are simultaneously being told to protect their forests and facing real food security pressures that make staple crop expansion difficult to simply regulate away.

The researchers of the study were careful not to cast their findings as an argument against European regulations or supply-chain accountability. Their hope, explicitly, was that their findings might help fill gaps in existing monitoring systems by providing country- and commodity-specific deforestation estimates that risk managers don’t currently have.

What it can’t do is substitute for the harder conversation the data now makes unavoidable: that the forests nobody is watching are being cleared by the crops nobody is regulating, for the people nobody’s supply chain reaches.