Monthly Archives: April 2022

Approaches to risk of reversal for Carbon Dioxide Removal

Different types of carbon dioxide removal differ greatly in the extent to which they are permanent.  The risk of reversal, with release of CO2 into the atmosphere, is always present for land-based sinks and other types of removal based on short duration carbon cycles.  In contrast, removals to geological storage are largely permanent.  These differences can be recognised in different ways.

The removal of carbon dioxide from the atmosphere (Carbon Dioxide Removal – CDR) is widely acknowledged as having an essential role to play in reducing climate change[1].  It is essential for balancing emissions that are hard to abate, such as some from agriculture and long-haul aviation.  Beyond that, it is essential for eventually reducing atmospheric concentrations of CO2 through net-negative global emissions.

However, securing the full climate benefits of removals requires that they are permanent and irreversible over very long timescales.  Climate change depends, broadly, on the cumulative emissions of CO2 to the atmosphere[2].  Delaying emissions through absorption and subsequent release of CO2 does little to change eventual cumulative total.  (Nevertheless, there may be some benefit to temporarily reducing atmospheric concentrations, especially if this reduces or delays the peak in concentration of carbon dioxide in the atmosphere.  This benefit may be relevant to assessing policies in some cases.)

Methods for CDR have very different risks of reversal.  For geological storage, part of very long carbon cycles, the risk of reversal is very low, and permanence can be largely guaranteed.  In contrast, terrestrial sinks such as forests, which are part of a much shorter carbon cycles, risk substantial reversals over years or decades[3].

Removals to geological storage and land-based sinks also have quite different properties in other respects.  Some of the main differences are summarised in the table. 

Table: Characteristics of different types of carbon dioxide removal from the atmosphere

Type of storageCapture from air for geological storageCapture from air inn terrestrial sinks
Length of carbon cycleLong (Many millennia to millions of years)Short (Decades or centuries)
Risk of reversalVery lowModerate to high
CostHigh to very high at present but with substantial scope for reduction.Low to medium in short to medium term (allowing for some benefits being distant in time because, for example time taken for trees to grow).  However long run MRV costs and need for permanence may greatly increase costs.
ScaleCurrently small – about three decades or more likely needed to reach Gt scaleReadily scalable
Requirement for continued management and MRVLowModerate to high

This creates a challenge in comparing the value for reducing climate change of different approaches to CDR.  This post briefly considers three different approaches to addressing this challenge.  The approaches are not mutually exclusive.  A fuller account can be found here.

Approach 1:  Separate treatment of long and short cycle removals

Under this approach differences are explicitly recognised.  Policy and incentives are largely separate for the different broad types of removal.  Among other things, use of land-based removals may be restricted to balancing land-based emissions.  They would not be eligible for balancing emissions of fossil carbon.  This is illustrated in the Figure 1.

Source:  Bellona

Approach 2:  short cycle removals discounted by a probability weighting

Under this approach, the cumulative probability of reversal set in advance by regulation is used to scale number of removal credits surrendered to meeting obligations, creating an “exchange rate” between different types of credit.  For example, a 25% risk of reversal requires 1.33 credits (“risk adjusted tonnes”) to be surrendered to balance a tonne of emissions.  Calculation of the risk of reversal takes into account estimates of the future direct and indirect effects of climate change, and risks arising from policy and management, ownership, and governance of projects.  This is illustrated in Figure 2.

Source:  Bellona

An additional buffer or safety margin may be built to recognise uncertainties in the estimates of probabilities.  For example, assessment may be based on confidence intervals of a distribution rather than the mean.  Removals with a risk of reversal above a certain threshold could be deemed ineligible.

Under this approach the effect of the scaling parameter on credit value is clear.  However it gives limited incentives for subsequent management, as the probability of reversal is set in advance.  It is also potentially administratively quite burdensome if calculations are specific to detailed project characteristics (e.g. tree species), location and jurisdiction.  Furthermore, the concept of probability based on an average outcome may not be robust to risks correlated across very large numbers of projects, for example mass dieback of forests.

Approach 3:  Credits are required to be “permanent equivalent”, with an obligation to replace reversals

Under this approach, credits for removals carry an obligation to make good any reversals at the time the reversal occurs, by creating or buying credits to match any reversals.  This obligation continues in perpetuity.  There is no buy-out available from simply paying a carbon price.  This is illustrated in Figure 3.

Source:  Bellona

Holders of credits must demonstrate they have the means to meet this obligation, for example through an insurance policy or funds held in escrow.  Government may have a role here due to uninsurable risks.

The price of a credit would reflect the cost of the storage project, cost of insurance or funds held, and continuing MRV costs.  It would thus be set by markets (at least in part).  The equivalent of ratings for bonds may emerge over time for different types of credit.

This approach uses market mechanisms to reveal the value of different types of credit, while ensuring permanence.  It creates direct incentives to manage stores of carbon, and potentially covert to permanence, for example via Bioenergy with CCS (BECCS).  However, continuing MRV is potentially costly.  There may also be an unwillingness among private sector parties to take on the required risks, reducing the supply of credits.  This would nevertheless case reveal information about risks.

Ways forward

It is essential for good policy that the different risks of reversals for different types of carbon dioxide removal are recognised, and that policy takes account of them.  This note outlines some possible approaches to this.  In considering each of these, the real value of permanence of removals in limiting climate change needs to be recognised.

Adam Whitmore – 26th April 2022

[1] See for example IPCC, 2018: Global warming of 1.5°C. An IPCC Special Report


[3] Similar reasoning applies to some man-made sinks, for example buildings.