Checked: “the world has only 60 harvests left”

A claim audit, done in the open. What holds, what is sloppy, and what I could not verify. Written 26 September 2026.

What holds

Soil erosion is a real and serious problem. It thins the topsoil that stores carbon, water and nutrients, and it is not uniform: measured erosion rates across the world span five orders of magnitude.

Some soils are eroding fast. In the largest global inventory, 16 percent of conventionally managed soils have a lifespan under 100 years. That is a near-term problem, and it is not confined to poor countries: short-lived soils appear on every continent, including the United States, Australia, Spain, Italy, Brazil and China.

Conservation works. No-till, cover cropping, contour cultivation and land-use change to forest or grassland extend soil lifespans by orders of magnitude.

Where the number comes from

The line traces to a spoken remark, not a study. In December 2014, at a forum marking World Soil Day, Maria-Helena Semedo, the FAO’s deputy director-general for natural resources, said that generating three centimetres of topsoil takes 1,000 years, and that “if current rates of degradation continue all of the world’s top soil could be gone within 60 years.” The Thomson Reuters Foundation (Chris Arsenault) reported it. Scientific American ran it under the headline “Only 60 Years of Farming Left If Soil Degradation Continues.”

Note the shape of that. A caution spoken at a conference became a deadline, and the deadline became a headline about farming ending.

The trail goes cold

Two careful tracers have tried to find a published basis for the figure and failed. The botanist James Wong tried in New Scientist. Hannah Ritchie, for Our World in Data (January 2021), did the same and concluded: “Although they continue to be repeated, there is no scientific basis to them.”

The sibling figures are worse. “Britain has only 100 harvests left” (the Independent) appears to link back to a University of Sheffield study of city allotments in Leicester, a paper that does not mention the number. Michael Gove’s warning that the UK was “30 to 40 years away from the ‘eradication of soil fertility’” has no published basis either.

And the number mutates. Sixty harvests, 100 harvests, 30 to 40 years. Different figures, the same rhetorical move: a countdown to global collapse. A 2019 Guardian piece still carried it as “an estimated 60 harvests left before they are too degraded to feed the planet.”

What the best global data says

The largest assessment is Evans and colleagues, “Soil lifespans and how they can be extended by land use and management change” (Environmental Research Letters, 2020; DOI 10.1088/1748-9326/aba2fd). It gathered 10,030 plot-years of erosion data from 255 sites across 38 countries. “Lifespan” here is the time for a 30 cm topsoil to erode at the measured rate.

Its findings: just under a third of conventionally managed soils have lifespans under 200 years, and 16 percent under 100 years. Under conservation management, 39 percent exceed 10,000 years. Our World in Data’s reading of the same dataset: the median lifespan for thinning soils is around 491 years, half exceed 1,000 years, about a third exceed 5,000 years, and some soils are not eroding at all. Some are even thickening, forming faster than they erode.

So the honest picture is a serious, uneven problem with a long median. The “60 harvests” line takes the worst tail, the one soil in six under a century, and states it as the whole world.

The tell

The claim converts a lifespan, how long a particular soil layer lasts at its current erosion rate, into a global calendar: the world has X years of farming left. Those are not the same claim, and they do not have the same evidence.

It also fixes a deadline that no measurement supports. For the same idea, the figures in circulation are 60, 100 and 30 to 40. A real measurement does not come in three different sizes.

Verdict

Say it this way and you will not be wrong:

Do not quote the countdown as a finding. It is a conference remark that outlived its evidence.

Edges: what I could not verify

• The original basis of the FAO official’s remark. Neither I nor two earlier tracers found a published calculation behind it.

• Whether “about a third of the world’s soil has already been degraded,” which Semedo also said, traces to a specific assessment. I did not verify that.

• The Sheffield allotment link is reported by Our World in Data with a hedge (“seems to”). I did not read the paper myself to confirm the absence of the figure.

Method

I started where the number circulates (Scientific American, the Guardian, the Independent, a former UK environment secretary), traced it to its origin, then checked it against the largest peer-reviewed global inventory (Evans and others 2020) and Our World in Data’s review. Every number above is attributed to a named source with a date. Where sources disagreed, I said so instead of picking a winner.