September 2025

Oceanography | Vol. 38, No. 3

34

et al., 2023), as well as incomplete understanding of the carbonate

pump in Earth system models (Planchat et al., 2023).

Some model approaches go beyond biogeochemistry and

attempt to model impacts on species and ecosystems. However,

the complexity and uncertainty of responses to OA, coupled

with lack of representation of biodiversity in numerical models,

make them generally less certain, but still informative. For

instance, pelagic calcification is included in several models as a

back-​calculation from particulate organic carbon production

(e.g., Aumont et al., 2015; Yool et al., 2015). Few models explicitly

represent calcifying plankton functional types (e.g., Krumhardt

et  al., 2019), and the benthos is usually represented in a very

simple manner focused on the recycling of organic matter into

nutrients and CO2. Feedbacks on pelagic carbonate systems are

represented in some models (e.g., DIC and TA fluxes associated

with remineralization of organic matter), but benthic calcifica-

tion is not usually represented.

The implementation of these coupled physical-​biogeochemical-​

ecosystem models has allowed identification of further areas of

uncertainty that require both deeper understanding of processes

and further model development to improve the representation of

the carbonate chemistry system. For example, improving repre-

sentation of freshwater input of DIC and TA and providing bet-

ter constraint of the continuous representation of the carbonate

system along the salinity gradient are particularly important for

simulating the spatial and temporal variability in coastal envi-

ronments. Including the impact of sediment processes on car-

bonate chemistry dynamics is similarly important for reducing

uncertainty in the coastal environment.

HOW GOAL 3 SUPPORTS mCDR

Data and knowledge synthesis for use in products and models

is essential for mCDR. Before deployment (i.e.,  in the research

stage), they are needed to predict feasibility, scalability, efficiency,

ecosystem response, and impact, and thereby to support decision-​

making to optimize mCDR approaches. If the mCDR activities

are taken forward beyond the research and development stage to

deployment (i.e., field trials and commercialization), these tools

are needed to evaluate MRV, carbon accounting, and environmen-

tal monitoring, therefore supporting the regulation and imple-

mentation of safe and sustainable practices.

The spatial and temporal scales over which mCDR may ulti-

mately be deployed may be very large, making observational

monitoring potentially (even prohibitively) expensive. A joint

model-observation approach to MRV is therefore recommended,

especially to assess alterations to ecosystems that may be too

small to measure observationally. Uncertainties associated with

both models and observations will still limit accurate prediction

(Bach et  al., 2023) and provide challenges for governance and

social license to operate, thus requiring interdisciplinary, collab-

orative approaches.

Models that already incorporate carbonate chemistry and bio-

logical response provide the foundation for the mCDR commu-

nity to build upon. However, some of the mCDR methods could

push the chemical composition of seawater outside of the nor-

mal range and therefore could require a more detailed approach

like the one developed in the SCOR working group Modelling

Chemical Speciation in Seawater to Meet 21st Century Needs

(MarChemSpec; Clegg et al., 2023). In relation to OA, models can

help to predict how different mCDR strategies will interact with

OA processes and ultimately whether they will result in long-term

benefits or result in disruptions to marine ecosystems.

Models provide a reliable basis for assessing the long-term

effectiveness of upscaled CDR as governed by macroscale hydro-

dynamics and the biological pump, which operate on decadal

to millennial timescales. While large-scale approaches provide

the climate context for mCDR impacts and benefits, the efficacy

and impacts of mCDR at local to regional scales can only be con-

veyed through higher-resolution and local-scale modeling. Such

models often have sub-kilometer resolution and can be adapted

to address individual or clusters of CDR deployments, assessing

environmental impact, dispersion of chemical and biological par-

ticles and plumes, local sequestration, and export. These models

will require validation studies to demonstrate that they are fit-for-

purpose, and local model approaches will need to be site-specific

(Khangaonkar et al., 2024).

Modeling is already offering a holistic picture of different

mCDR deployment scenarios; for example, regional simulations

(Wang et al., 2023) and century-long simulations (González and

Ilyina, 2016) show artificial OAE can effectively remove atmo-

spheric CO2 and alleviate OA. However, emissions-driven Earth

system modeling demonstrates that an abrupt ending of OAE

might act to accelerate OA and atmospheric warming, thus

threatening vulnerable ecosystems that are struggling to adapt to

existing environmental pressures (González et al., 2018).

The community has already made critical investments in

ocean biogeochemical and ecosystem models. Such models

are crucial for simulating present-day and future predictions

of mCDR impacts and ecosystem consequences. Better under-

standing of the implications of greenhouse gas emissions and

CDR for the coupled carbon climate system is essential for pro-

viding reliable guidance to policymakers and other stakeholders.

While such global-scale approaches provide the climate context

for CDR impacts, answering questions about the effectiveness

and ecosystem impacts of local to regional-scale CDR approaches

may require both higher-resolution and regional-scale model-

ing as well as incorporation of additional modeling strategies.

These tools will allow the combination of CDR scenario assess-

ment, detection, and attribution; observation system simulation;

and process studies to increase understanding and inform sound

management decision-making.