A framework for minimizing remote effects of regional climate interventions: Cooling the Great Barrier Reef without teleconnections

Climate interventions like Marine Cloud Brightening (MCB) have gained attention for their potential to protect vulnerable marine ecosystems from the worst impacts of climate change. Early modeling studies raised concerns about potential harmful global side effects stemming from regional interventions. Here we propose a modeling framework to evaluate these risks based on using maximal deployment scenarios in a global climate model to identify potential pathways of concern, combined with more realistic large intervention levels. We demonstrate this framework by modeling a cooling intervention over the Great Barrier Reef using the Community Earth System Model (CESM2). We identify potential impacts on tropical convection that could produce remote impacts, and show that limiting intervention duration to deployment in the key season largely eliminates these risks. Overall we illustrate that the local ecological goals can be achieved at a level of cooling well below what poses a risk of significant remote effects.

Methods
The simulations for this study were conducted using the Community Earth System Model (CESM2, Danabasoglu et al. 2020; Computational and Information Systems Laboratory, 2023)) running in a “slab ocean” configuration. In this mode a simplified ocean simulates the mixed layer of the ocean and its interactions with the atmosphere without simulating full ocean circulation. The exchange of heat between the mixed layer and the deeper ocean is approximated with a prescribed monthly Q-flux that is derived from a fully coupled run of the model. All simulations are carried out under preindustrial conditions, with a 50-year spin up period to ensure the model reaches a stable equilibrium before perturbations are applied. 

Cooling interventions are implemented using an additional Q-flux term to specify a forcing at the surface of the ocean. The advantage of this approach is that it allows direct control of the location and amount of forcing applied, and can be used to represent cooling that is achieved through cloud brightening, surface albedo enhancement, deep ocean pumping, or any combination of intervention technologies. The limitations of this approach are that it does not include any dynamic adjustments or feedbacks that are specific to an albedo modification or cloud intervention. The simplified ocean also neglects the advection of cooled surface waters due to ocean circulation. This modeling configuration is not aimed at evaluating the efficacy of particular intervention technology at producing cooling, but whether a given level of regional cooling would produce significant impacts outside the region of intervention, in line with step 1 of our framework. Despite these limitations, the SST variability in the Coral Sea region in the slab-ocean simulation is quite similar to both a fully-coupled pre-industrial run of CESM2 and with ERA5 reanalysis (Figure S1).

For each level of forcing, three separate simulations are branched from the control run at five year intervals and run for 13 years. The first year of each branch is discarded as the system is adjusting, producing 36 years of data for each scenario. Using several shorter runs rather than a single long run allows us additional statistics to examine impacts that emerge in the first decade of deployment, when increased scrutiny will be placed on both the effectiveness and any unintended side effects of an intervention. Spacing the branches by five years allows more sampling across conditions of natural variability.