By David Fickling / Bloomberg Opinion

Did we just … win?

It’s certainly sounding that way. Emissions from fossil fuels — the key driver of global warming since the dawn of the industrial era — are set to peak within two years, according to Rystad Energy, an oil and gas consultancy. Carbon pollution from electricity, the sector with the biggest footprint, will never again hit the levels it reached last year, the International Energy Agency said Feb. 8.

That’s a remarkable and somewhat unexpected achievement. Predictions of large-scale decarbonization are “the academic equivalent of science fiction,” Vaclav Smil, Bill Gates’ favorite energy thinker, argued last year. Exxon Mobil still expects petroleum demand to be growing in 2050. One of the main models used until recently to map out the future of climate assumes emissions won’t fall until the 2090s.

Everything still depends on not just the moment when carbon pollution starts to decline, but the pace at which it shrinks. Still, a near-term peak sharply improves the prospects of keeping warming below catastrophic levels.

The problem that’s looming just out of sight may be the tougher one, though.

While it’s common to think of climate change as almost synonymous with fossil-fuel consumption, burning coal, oil and gas isn’t the only way human activity is heating the atmosphere. A far older practice — changing the landscape by cutting down forests, pasturing livestock, and planting crops — accounts for between 6 percent and 21 percent of emissions. The burning of fossil carbon has risen (and will hopefully fall) at the breakneck pace of the industrial development it fuels. Agriculture and land use is measured by a slower clock, and may be harder to turn back.

Working out our fossil-carbon emissions is a relatively straightforward business. A high school student could calculate the number of carbon atoms in a barrel of oil or a metric ton of coal, and then estimate the tonnage of carbon dioxide emitted when it’s burned. Given publicly available trade and production figures, they could produce a decent estimate of the world’s industrial carbon footprint in an afternoon.

Biological emissions are different. Does planting a biodiverse forest on a bare Scottish moorland lock away emissions, or release them? It’s a surprisingly complicated question that even experts don’t find easy to solve. (If the trees break up masses of carbon locked away in inert peat and release it back to the atmosphere, they may be releasing extra CO2; but that answer may change if you measure it five years after planting, and 20 years, and 50 years.)

Just appraising the volume of carbon locked up in the world’s ecosystems presents formidable challenges. Working out how much is stored in the boughs and leaves of forests comes down to a series of extrapolations taken from estimates of canopy height and species mix, backed up with sample tape measurements of trunk girth taken at the chest height of whichever forester is doing the work that day.

We have even more approximate ideas of how much biomass is stored in plant roots. One 2017 study of cornfields found that traditional techniques overestimated the volume of living roots by about 67 percent; a significant gap when you consider that they’re reckoned to amount to about 130 billion tons of plant matter globally.

The non-plant carbon in soils is usually estimated by taking core samples and burning them to measure how much carbon dioxide is emitted. That technique, again, is subject to vast sampling errors, making it hard to be precise about a category which may lock up 2.5 billion metric tons of carbon, equivalent to 60 years of the world’s emissions.

The good news is that we may already be making progress on the natural world’s carbon ledger. About half of the world’s emissions from deforestation are being counterbalanced by the planting of new trees, according to the Global Carbon Project, a coalition of climate scientists. Land-use emissions appear to have declined from nearly 7 billion tons in 1960 to less than 4 billion now, even as we’ve added about 5 billion people to the global population.

It’s not impossible that further decreases in logging and increases in planting could cause that number to turn negative, meaning our ecosystems would be sequestering more carbon than they’re emitting. Indeed, we’re counting on just that change taking place: Almost all those pledges to hit net zero depend on the environment absorbing the gross bit of emissions we don’t know how to prevent.

That’s just where all the uncertainty is so troubling, however. To date, about a third of our industrial emissions have ended up locked up in land ecosystems, with another third going into the oceans. Climate change results entirely from the remaining third, 14 billion tons a year or so, that goes into the atmosphere.

Should the land and ocean carbon sinks reach saturation and stop sucking carbon out of the skies — something that may occur on land in a matter of decades — then even a reduction of our emissions may be insufficient to stop the atmosphere’s carbon concentration from going up and cooking the planet.

Changing the natural environment was the first way human technology started to bend this world to suit our needs, dating back to the dawn of fire agriculture 80,000 years or so ago. In comparison to two centuries or so of industrialization, that ancient practice may be far harder to reverse.

David Fickling is a Bloomberg Opinion columnist covering energy and commodities. Previously, he worked for Bloomberg News, the Wall Street Journal and the Financial Times.