Hunting out the big fruit-eaters costs a rainforest about 3% of its carbon — and nothing visible happens

In June we published a piece here saying that when the large fruit-eating animals go, the rainforest canopy itself collapses. We cited nothing. The mechanism is real and it is well studied; the word collapse is not what the measurements say, and the number underneath is more interesting than the word was.

The mechanism, which holds

Big trees tend to make big seeds, and big seeds need a big animal to move them. A hornbill, a gibbon, a tapir, a large monkey — these carry seeds tens or hundreds of metres from the parent and deposit them somewhere with light and without the parent's own pests. Remove the animal and the seed falls under the tree that made it, where its odds are poor.

Hunting removes exactly those animals first, because they are the ones worth carrying home. So a forest can look completely intact from above — closed canopy, no clearing, nothing a satellite would flag — while the mechanism that decides which trees exist in fifty years has already been switched off. Ecologists call it defaunation, and it is invisible in every remote-sensing product we have.

What it costs, measured

A 2019 study in Scientific Reports took a 30-hectare forest dynamics plot in central Thailand — one of the few places with an intact fauna of primates, ungulates, bears and birds of every size — mapped the actual seed-dispersal network between trees and animals, and simulated removing the large frugivores.

Two numbers came out of it. Tree species that depend on large-bodied frugivores hold nearly a third of the above-ground carbon on that plot. And complete defaunation reduces stored carbon by 2.4 to 3.0 percent.

Those two numbers sit oddly together, and the gap between them is the finding. A third of the carbon is in trees whose dispersers are gone, but the forest does not lose a third of its carbon, because other trees grow into the space. They are simply different trees — faster, lighter-wooded, smaller-seeded — and the replacement is close enough in mass that the total barely moves.

The study also found the loss exceeded one percent whenever defaunation reached forty percent or more, and that the effect size is comparable to what has been measured in Neotropical forests, which had been assumed to be more vulnerable than Asian ones.

What we got wrong

Collapse is the wrong word for a three percent carbon deficit and we should not have used it. The honest description is a substitution: an old, slow, dense forest becoming a younger, faster, lighter one with the same outline, over decades, with no single moment at which anything falls down.

That is a harder thing to report and a harder thing to campaign about, which is probably why the stronger word is the one in circulation. It is also more alarming in one specific way: nothing visible happens, so nothing triggers a response.

Why a Sacramento paper covers a Thai forest plot

Because the same substitution is legible closer to home, in a different currency. Oak woodland in this part of California depends on scrub jays and woodpeckers to move and cache acorns; valley oak regeneration has been failing for decades while the standing canopy looks unchanged from the road. A forest that is not reproducing looks exactly like a forest until the day it does not.

What this does not settle

This is one plot, in one forest type, with one modelled scenario, and the carbon figure is a simulation of community turnover rather than an observation of it. It says nothing about what defaunation does to anything other than above-ground carbon — pollination, understory structure, the animals that ate the animals — and a small effect on one measure is not a small effect on a forest.