Yes, roughly 36% of annual carbon uptake at eastern US forest sites occurs after tree growth stops — here's why that's real

The claim is that at eastern US forest sites, roughly 36% of annual carbon uptake occurs after tree radial growth has already stopped — meaning the forest keeps pulling carbon out of the atmosphere well into autumn even after the wood-forming cells have shut down for the year. The verdict is true, and it is supported by multiple independent lines of primary evidence.

The most direct source is Dragoni et al. (2011, Global Change Biology), which explicitly measured and quantified that at Morgan Monroe State Forest in Indiana, approximately 36% of annual carbon uptake occurred after radial tree growth had ceased. This was not an estimate or a model output — it was derived from eddy-covariance flux tower measurements that continuously track the net exchange of carbon dioxide between the forest and the atmosphere, cross-referenced against dendrometer records marking when xylem growth stopped. That combination of direct flux measurement and growth monitoring is exactly the right methodology to answer this question.

The strongest version of a skeptical objection would be that this is a single-site result from one Indiana forest and may not generalize. That objection does not hold. Urbanski et al. (2007, Global Change Biology) found a substantial fraction of annual net ecosystem production occurring in the post-peak-growth autumn period at Harvard Forest in Massachusetts. Gough et al. (2013, Journal of Geophysical Research: Biogeosciences) found that late-season, post-radial-growth-cessation uptake contributed approximately one-third of annual gross ecosystem production at the University of Michigan Biological Station. Curtis et al. (2002, Ecological Applications) synthesized AmeriFlux data from multiple eastern US sites and documented that net carbon uptake persists well into autumn after cambial growth stops, placing the fraction in the 30–40% range. Baldocchi (2008, Australian Journal of Botany) reviewed eddy-covariance records across eastern North American deciduous forests and reached the same 30–40% autumn-season estimate. Five independent studies, spanning Indiana, Massachusetts, Michigan, and multi-site syntheses, all converge on the same range.

The mechanism is well understood and makes the number less surprising once you know it. Leaves remain photosynthetically active and continue fixing carbon for weeks after the cambium stops producing new wood cells in late summer. Crucially, autumn temperatures are cooler, which suppresses the metabolic respiration costs that would otherwise offset photosynthetic gains. The result is that the forest runs a net carbon surplus — taking in more than it releases — even with no new wood being built. Growth cessation and carbon uptake cessation are simply not the same event.

It is worth being precise about what the 36% figure does and does not mean. It refers to net carbon uptake — the balance between photosynthesis and ecosystem respiration — not to gross photosynthesis alone. It also refers to the period after radial growth stops, not after leaves visibly change color or drop. The autumn carbon-gain window is real but finite; once leaves senesce and fall, uptake collapses. The figure also reflects current climate conditions; a warmer autumn that extends the green-leaf period would likely push this fraction higher, which is why the finding matters for projections of forest carbon sequestration under climate change.

The pattern to watch for in misuse of this finding is the reverse error: assuming that because post-growth carbon uptake is large, tree-ring records (which track only the growth period) are adequate proxies for total annual forest carbon balance. They are not. Tree rings capture roughly 64% of the annual carbon story at these sites, by the same math. Anyone using ring width alone to reconstruct historical carbon flux should be treating that as a known underestimate, not a complete account.