Forest carbon's back-end durability problem

Most reforestation projects today ignore natural regrowth in making carbon claims

The natural world is a key ally in combating climate change. Studies estimate that reforestation – restoring areas where forests have been removed or fragmented – could remove upwards of 300 gigatons of CO2 (GtCO2) from the atmosphere. Even more could be removed by protecting existing forests and allowing to recover and maturity.

The world needs to invest more in protecting the forests we have and restoring those that have been degraded. However, the way we have chosen to finance it – by selling offsets into voluntary or compliance markets – introduces unintended consequences that could undermine our ability to stabilize global temperatures and achieve our climate goals.

Climate scientists have a saying that “carbon is forever”¹. While natural sinks will absorb about half of what we emit today over the next century, it takes on the order of 400,000 years for the carbon cycle to fully remove current emissions. The extremely long atmospheric lifetime of CO2 means that even if we get emissions down to zero, warming will stop but the world will not cool down for centuries to millennia to come.

In short, warming can be approximated as a somewhat time-invariant function of cumulative emissions. This insight has driven the development of net-zero frameworks that seek to achieve zero emissions by balancing anthropogenic sources of CO2 with anthropogenic sinks. Forest carbon projects are used at fairly large scales in both voluntary markets (corporations offsetting their footprints) and compliance market (state and national-level cap and trade systems) to “neutralize” fossil CO2 emissions today.

The challenge is that once fossil CO2 enters the atmosphere, its going to stay there for an extremely long time. Its fiendishly difficult to make comparable claims for carbon removed from the atmosphere and stored in the biosphere.

There are a host of durability challenges with biosphere carbon storage such as reforestation or afforestation. A warming world will bring multiple challenges from more frequent fires, pests, and droughts. These concerns are far from academic; for example, California set aside a “buffer pool” of forest carbon projects in addition to those used as offsets in its cap-and-trade system to cover risks of disruption. The buffer pool was intended to last for a century, but forest carbon projects representing nearly twice the entire pool have already burnt up in just the past decade.

But even if we could solve these “front-end” durability issues – ensure forests are restored after disruptions and design legal structures to ensure preservation over millennia – there is an even thornier problem that nearly all forest carbon projects are ignoring today: back-end durability.

The counterfactual growth question

Over the last century the Eastern US coast forest regrowth has been nothing short or miraculous. The region was largely deforested in the 1800s and 1900s to make way for agricultural expansion. However, significant gains in agricultural productivity reduced the need for marginal farmland, leading to large scale reforestation. It is a pretty common occurrence to come across old stone fences and foundations of farm houses when walking in the woods there. This large-scale regrowth has even counteracted the impact of climate change in the region due to the localized cooling effect of trees.

However, this did not occur as a result of early 20th century carbon markets paying landowners to transition from agricultural land to forests. Rather, it came about as a broader shift toward higher yields in farms in the midwest and western states and a decline in east coast agriculture.

In some ways this represents a preview of larger trends to come. The world has already reached peak pasture, and global population is expected to peak and decline over the next 60 years. This will reduce the demand for agricultural land, and other regions may follow the lead of the US Southeast with natural afforestation.

For reforestation to permanently remove carbon requires proving that forests never regrow in the absence of human intervention. Otherwise projects are just capturing the time delta between when the forest regrows due to human intervention and when vegetation would have naturally recover.

However, the very areas that are usually targeted for reforestation are the most likely to naturally recover given the absence of more economically viable utilization options (which would kill the economics of reforestation in many cases).

There may be afforestation opportunities that could be more permanent in regions where trees have not historically grown, but that involves transforming whatever ecosystem (e.g. grasslands) was previously there, and has its own challenges for both biodiversity and carbon impacts.

The diagram below shows an example of the expected carbon sequestration for a reforestation project compared to a counterfactual natural regrowth 30 years later. The red curve shows the net carbon sequestration over time associated with this hypothetical example:

Illustrative example of forest carbon sequestration over time in human reforestation and the natural regrowth counterfactual, with the net removals over time shown below.

This issue also applies to avoided deforestation credits (which represent avoided emissions rather than net removals) as well as reforestation credits. To credibly neutralize fossil CO2 with avoided deforestation requires demonstrating that the forest would remain deforested indefinitely.

This is clearly a problem in carbon markets today, as almost no projects account for counterfactual natural regrowth. So how do we solve this?

First of all, we can sidestep the whole issue by valuing temporary removals like afforestation outside of a framework that requires making neutralization claims against fossil fuels. In this world, any natural reforestation that occurs is a feature rather than a bug, and does not result in a warmer world in the way that using reforestation in lieu of fossil fuel mitigation would.

If we insist on making neutralization claims with forest carbon, we need to either explicitly plan for their temporary nature through mechanisms that transition from temporary to permanent removals over time (called hybrid or “blended” tons), or figure out ways to credibly “horizontally stack” temporary removals by replacing one with another over time.² However, developing systems to indefinitely replace temporary removals over time is quite challenging, particularly in a world where the lifespan of emitting companies is a flash in the pan compared to that of atmospheric CO2 (and the private sector has historically proven quite adept at figuring out ways to discharge environmental liabilities).

Finally, we could try estimate what would have happened in the absence of reforestation through the use of dynamic baselines, something that the community has started to use over the past year in a limited number of projects. However, dynamic baselines require some inherently subjective decisions in the selection of control regions to compare. There is some irreducible selection bias in the project design as true randomized controls are not possible.

But the current prevailing practice of making neutralization claims with forest carbon credits from reforestation in the absence of considering natural regrowth potential fundamentally overestimates the magnitude and effective durability of reforestation projects, and results in a warmer future than if we eschewed offsets and mitigated fossil fuels instead.

1. The is particularly true for fossil carbon emissions. Arguably land use emissions are not as long-lived if the sink is restored through regrowth, but processes to form new fossil carbon like coal, oil, and gas take place over extremely long timeframes.

2. Some folks propose “vertical stacking” as an alternative through ton-year accounting, assuming a certain amount of tons removed for one year (say, 100 tons) is equivalent to a ton permanently removed from the atmosphere. However, this approach is fundamentally inconsistent with net-zero frameworks.