September 2025

Oceanography | Vol. 38, No. 3

76

TABLE 1. Abridged summary of priorities for addressing needs and challenges related to biogeochemical observations, models, data management, and col-

laboration for each discussion topic.

OBSERVATIONS

MODELS

DATA MANAGEMENT

COLLABORATION

Biological Carbon Pump (BCP)

The BCP is a key process in the ocean carbon cycle involving transfer and remineralization of organic carbon to depth. Biologically mediated biogeochemical (BGC)

transformations occurring against a backdrop of complex physical dynamics make the BCP a uniquely challenging process to measure and model.

• Prioritize deep (mesopelagic and deeper)

measurements and time series

• Observing gaps: physicochemical particle

properties, key nutrients like iron and

ammonium, community structure and

function, trophic interactions, vertical

migration, biological rates (grazing, viral lysis)

• Uncertainty quantification

• Gridded climatology of particulate organic

carbon (POC) flux

• Data rescue and digitization

• Integrate diverse multi-platform

datasets over space and time

• Perform quantitative evaluation of

observation and model mismatch

• Standardized guidelines for data

collection, metadata reporting, and

data processing (data aggregation

and error propagation)

• Centralized repository or aggregator

of metadata for BCP-relevant

measurements

• Conduct moderately sized BCP

process studies that integrate

sampling and modeling activities

from the outset

• Hackathons and community activities

that build capacity and facilitate idea

and knowledge exchange

Episodic and Extreme Events (EEEs)

EEEs such as storms and wildfires may generate large BGC fluxes over short periods of time, thus serving as major players in BGC cycles and marine ecosystem health.

However, EEEs pose safety and logistical challenges, and observations and models require high spatiotemporal resolution to understand their impacts.

• Develop and deploy robust (able to

withstand EEEs) platforms and technologies

to fill spatial and temporal gaps (including

satellite remote sensing, e.g., geostationary

missions like NASA GLIMR)

• Leverage existing observatories (e.g., OOI,

LTERs) to conduct event-based sampling

• Establish sentinel sites where a dynamic

range of EEE impacts occur (storms,

cyclones, wildfires) for sustained data

collection

• Regional models and/or dynamical

or statistical downscaling of global

outputs to constrain event-scale

dynamics

• Organize early collaboration to

ensure sampling resolution is

adequate for models

• Adopt common definition of

“extreme” (% departure from

baseline)

• Create metadata fields and/or flags

for EEEs

• Funding mechanisms and community

activities (model intercomparison,

data synthesis, comparative

analysis) that require integration

of observations and models for

knowledge sharing and capacity

building

• Mechanisms to support collaborative

international EEE research

Machine Learning (ML)

Models used to predict ocean BGC cycling encode a host of relationships between environmental variables. A key question is whether these mathematical relationships

realistically combine to produce the emergent behavior of ocean BGC systems to allow predictions to be made in areas with sparse observations.

• Increase spatiotemporal resolution of BGC

(especially nutrients like iron, ammonium)

and biology (plankton biomass) observations

to improve ML algorithms

• Increase availability of model outputs

for ML reanalysis

• Standardization of model outputs

of phytoplankton community

composition

• Improve data standardization, quality

control, and open access to facilitate

synthesis and ML training on large

datasets

• Make ML tools available to the

community

• Workshops for ML training

Marine Carbon Dioxide Removal (mCDR)

Anthropogenic CO2 emissions have been the leading driver of climate change over the past century. To limit warming and associated climate and ecosystem impacts, multi-

sector efforts are underway to explore human intervention strategies to remove CO2 from the atmosphere and sequester it long-term in the ocean (NASEM, 2022). Well-

integrated modeling and observing efforts are vital to rigorous assessment of these approaches.

• Strategic BGC observing system

deployments (water column and benthos)

for mCDR projects to assess efficacy and

impacts

• Work with industry to produce and refine

BGC sensors and autonomous platforms

(e.g., AUVs, ASVs, moorings) that specifically

focus on relevant carbonate chemistry

parameters

• Optimize sampling strategies using

OSSEs

• Improve representation of particulate

inorganic carbon distributions within

models

• Prioritize development of models that

simulate regional and mesoscale

dynamics

• Transparency and public availability

of data, methods, and software

emerging from mCDR research

• Adopt common vocabularies and

data/metadata reporting standards

• Create mCDR data flags to note

datasets that contain results from

experiments that modify natural

ocean conditions, and/or novel data

assembly centers for mCDR projects

• Integration of observations and

models starting in early stages of

mCDR projects to build common

vocabularies and understanding

• Research funding must keep pace

with venture capital investment to

ensure rigorous scientific evaluation

of emerging technologies

• Efficient data archiving and peer

review to make information available

more quickly

Table continued on next page…

regions due to sea ice coverage and adverse sea conditions (Heimdal

et  al., 2024). Several regions (e.g.,  polar, tropical Pacific, Indian

Ocean) and depths below the surface layer of the ocean are rela-

tively undersampled (Abrahamsen, 2014; Levin et al., 2019; Smith

et al., 2019). Across biogeochemistry disciplines, specific variables

and processes are necessary to measure but remain undersampled.

Measurement, monitoring, reporting, and verification of mCDR

projects require information about baseline ocean biogeochemis-

try (Fay et al., 2024; Ho et al., 2023). The biological pump observing

community requires the constraint of high trophic level processes

in addition to measurements of physicochemical particle proper-

ties, including sinking velocity, porosity, and chemical composition,