Guest Post by Willis Eschenbach

There’s a new study out claiming that tropical forests are losing their ability to sequester carbon dioxide. Bad forest, no cookies for you! The study is entitled “Asynchronous carbon sink saturation in African and Amazonian tropical forests”, and is available through Sci-Hub here. In their press release it says:

In the 1990s intact tropical forests removed roughly 46 billion tonnes of carbon dioxide from the atmosphere, declining to an estimated 25 billion tonnes in the 2010s.

The lost sink capacity in the 2010s compared to the 1990s is 21 billion tonnes carbon dioxide, equivalent to a decade of fossil fuel emissions from the UK, Germany, France and Canada combined.

Overall, intact tropical forests removed 17% of human-made carbon dioxide emissions in the 1990s, reduced to just 6% in the 2010s.

Now, the first thing I noticed was that this study falls under what I modestly call “Willis’s Rule Of Authors”, which states that

Q ≈ 1/N2

Where “Q” is the quality of a scientific study, and N is the number of study authors. In English, this means that the quality of a study decreases in proportion to the inverse of the square of the number of authors … which in this case is … wait for it … no less than one hundred and six authors.

So … how did the one hundred and six authors determine that the tropical forest are no longer sequestering as much as they used to sequester?

According to the study, they used both observations and models. How much of each? Well, a clue is that the word “model” appears 158 times in their study, while “observations” appears only 44 times … here’s a typical comment.

A statistical model including carbon dioxide, temperature, drought, and forest dynamics accounts for the observed trends and indicates a long-term future decline in the African sink, whereas the Amazonian sink continues to weaken rapidly.

Now, it turns out that we can actually measure roughly how much human-generated CO2 is sequestered each year. The calculation is not all that complex. We know how many gigatonnes (“GT”, 109 metric tonnes) of carbon are emitted each year from the burning of fossil fuels plus cement manufacture. We know that for every 2.13 GT of carbon emitted, the atmospheric carbon dioxide level increases by one part per million by volume (ppmv). We routinely measure the atmospheric carbon dioxide level at the Mauna Loa Observatory.

So we take a look at how much carbon is emitted in a given year and we calculate how much the CO2 level should have increased. We subtract from that the actual increase in CO2 during that year, and the difference is the amount sequestered in that year by various carbon sinks around the planet.

Using that calculation, Figure 1 shows the amount of CO2 sequestered by year since 1959.

Figure 1. Percentage of anthropogenic CO2 sequestered by year.

There are some interesting things about this graphic. First, the amount sequestered varies widely year by year. This depends on two factors, the amount of natural emissions of CO2, and the variations in the natural carbon sinks. These are weather-driven. Some years there’s a lot of green growth on land and at sea, and other times there’s much less, which varies the amount of the carbon sinks. On the other side of the ledger, El Nino years like 1998 show an increase in natural CO2 emissions, so the percentage sequestered is smaller. With very large El Ninos, the amount of CO2 released may actually overwhelm the sinks, as in 1967.

Next, the smoothed value (blue/black line) varies between about 40% and 50%. It has hit both 40% and 50% twice in the past and most recently is about in the middle. There is no statistically significant long-term trend in the data (a least-squares analysis shows sequestration increased by 9% over the period, but the p-value is 0.20, far from significant).

Of interest for our purposes is the average for the 1990s and the 2010s mentioned in the study. In the 1990s the average sequestered was 49.8%. In the 2010s the average was 46.0%. The difference indicates the reduction in sequestration is 3.8%, but again it is not statistically significant …

But the one hundred and six authors say that forest loss reduced the amount sequestered by 11%. Hmmm …

Ah, well, I guess we shouldn’t expect models to actually agree with reality, that would be far too boring. In any case, we’re left with several possibilities, which include:

  1. The new paper is correct, forest sequestration has decreased by 11%, and coincidentally some other unknown carbon sink has increased by 7.2%, or …
  2. The new paper is wrong, or …
  3. The uncertainty is too great to come to any conclusion.

My conclusion? I’m not taking firm sides in this question, although option 2) seems most probable, followed by option 3). In any case, there’s no indication of a long-term reduction in the sequestration rate. And finally, I find it … well … discouraging that the authors didn’t include the actual sequestration calculations and discuss the implications of reality as opposed to models.

My best to everyone from a very dry hillside near the ocean in Northern California,


As Usual: I politely request that you quote the exact words you are discussing, to avoid at least some of the misunderstandings that are the bane of the internet.

Technical Note: As you might imagine, there is a lag between when CO2 is emitted at the surface and when that surface CO2 affects the background CO2 as measured at Mauna Loa. This lag is on the order of ten months, and it has been adjusted for in Figure 1 above.

via Watts Up With That?

June 7, 2021