The coupled uncertainty in negative emission technologies and transient climate response to cumulative CO2 emissions.

<div> <p><span data-contrast="auto">To meet the Paris Agreement targets, in addition to rapid emission reductions, carbon dioxide needs to be removed from the atmosphere with Negative Emission Technologies (NETs). On one hand, these solutions seem promising; on the other hand, they have significant and poorly estimated uncertainties and risks related to their potential to remove atmospheric carbon and wider impacts on the Earth system. One previously largely unexplored aspect of NETs is whether the uncertainty in NETs and e.g. transient climate response to cumulative CO₂</span><span data-contrast="auto">&#160;emissions (TCRE) are coupled to any degree, e.g. could some NETs have lower carbon removal potential if TCRE is high.</span><span data-ccp-props="{">&#160;</span></p> </div> <div> <p><span data-contrast="auto">We estimate how TCRE and selected NETs&#8217; carbon removal potential are dependent on climate system parameters using Perturbed Parameter Ensemble (PPE) with the University of Victoria Earth System Climate Model (UVic ESCM) and Gaussian Process (GP) emulator. Our aim is to explore and quantify any potential correlation between the carbon removal potential of single NETs and TCRE. The NETs considered are afforestation, reforestation, ocean alkalinization, ocean iron fertilization and direct air capture, which all except the last one depend on the perturbed parameters.</span><span data-ccp-props="{">&#160;</span></p> </div> <div> <p><span data-contrast="auto">The parameters of interest are chosen according to their expected impact on the climate and carbon uptake, constrained according to observations, and perturbed based on their prior probability distribution functions (PDFs). Then, to explore the parameter&#8217;s space, we use GP emulators to estimate model outputs as surrogate of actual ESM runs, which would be computationally too expensive. The emulators are created for the preindustrial spin-up, historical period, future control scenario and one for each NET scenario. They are trained through 300 simulations, considering 20 perturbed parameters. This analysis yields the correlation between the carbon removal potential of each NET and TCRE, and the contribution of each perturbed parameter to these two metrics.</span><span data-ccp-props="{">&#160;</span></p> </div> <div> <p><span data-contrast="auto">The preliminary results from the 300-member ensemble give a mean TCRE of 1.63 &#176;C/1000 PgC, which is consistent with the best estimate of 1.65 &#176;C/1000 PgC reported by the IPCC AR6 WGI (2021). The simulations with a high TCRE also tended to have a high ocean iron fertilization&#8217;s potential, meaning that some NETs are potentially more effective in removing atmospheric carbon dioxide if the temperature change per cumulative carbon dioxide emissions is high. Identifying such correlations between TCRE and NETs&#8217; potential allows designing more robust mitigation strategies including portfolios of NETs that hedge against high TCRE.</span><span data-ccp-props="{">&#160;</span></p> </div>