Early Online Release

Oceanography in the Age of Intelligent Robots and a Changing Climate By Chris Scholin

Oceanography | Early Online Release

the capabilities for testing this theory globally were in their infancy.

The consequences of fossil fuel consumption were of tremendous

importance societally, but at the time of Revelle and Suess’s procla­

mation, that was not a part of public discussion and politics as it

is today. Decades of research followed as investigators sought the

means to conduct ocean-basin-scale measurements needed to

assess predicted trends. In 1999, Peter Brewer elegantly recounted

that history during the first annual Revelle Lecture (Brewer, 2000).

Capturing time-space variations in the ocean’s interior, at basin

scales, accurately, is no small challenge. For many years, the only

practical way to tackle this problem was by using crewed ships to

conduct hydrographic surveys. Despite the analytical and logisti­

cal challenges, a picture of the exchange of CO2 between the atmo­

sphere and ocean slowly emerged (Brewer, 2013). Decades of

work were required to establish the connection between the burn­

ing of fossil fuels and the reality of human-driven climate change

and ocean acidification. Ironically, nearly 70 years after Revelle

and Suess issued their “geophysical experiment” proposition, we

now find ourselves scrambling to assess the promise and pitfalls

of artificially stimulating the ocean to absorb more CO2 to mit­

igate a climate crisis of our own making (e.g., Coale et al., 1996;

Brewer, 2013; Bach and Boyd, 2021; NASEM, 2022; Levin et al.,

2023; S.M. Smith et al., 2024; Findlay et al., 2025, in this issue).

ENTER THE ROBOTS

As ocean sensor systems matured, so too did the platforms on

which they could be deployed. In addition to measurements

acquired manually, scientists and engineers developed the means to

automate air-sea CO2 flux measurements aboard ships, moorings,

and autonomous surface vehicles (ASVs; Friederich et al., 1995;

Chavez et al., 2017b, 2018). It was apparent that seasonality and

geographical location played an important role in when and where

there was a net flux of CO2 into or out of the sea (e.g., Takahashi

et al., 2009). Other sensor systems for autonomously acquiring bio­

geochemical measurements, such as pH (Johnson et al., 2016) and

nitrate (Sakamoto et al., 2017), also evolved along with improve­

ments for in situ quantification of oxygen and optical parameters,

all of which were deployable on autonomous underwater vehicles

(AUVs) and ROVs. Development and use of these biogeochemi­

cal sensor suites were greatly aided by the availability of mooring

technology and well-established time-series studies that included

routine ship-based hydrographic surveys (e.g., Karl, 2010, 2014;

Chavez et al., 2017b). Now, after years of observations, the unmis­

takable trend of rising CO2 in the atmosphere with concurrent

changes in ocean pH and temperature has emerged (Figure 1;

e.g., Thorne et al., 2024) along with complex biological and eco­

system manifestations (e.g., Doney et al., 2020; Alter et al., 2024).

Thanks to a remarkable confluence of technologies and dogged

determination on the part of scores of visionary scientists and

engineers, it is now possible to observe ocean basin-scale car­

bon cycling using a distributed fleet of profiling floats—robots—

that offer much more information at a far lower cost compared

to ship-based surveys (Figure 2a,b; e.g.,  Johnson and Claustre,

2016; Claustre et al., 2020; Schofield et al., 2022; Sarmiento et al.,

2023). A global fleet of floats now returns sensor measurement

data from remote regions of the globe in real time, and the infor­

mation acquired is freely accessible to anyone nearly instantly via

the Internet (GO-BGC). This remarkable achievement has given

ocean scientists the equivalent of a medical doctor’s tool kit for

rapidly assessing a patient’s vital signs. As a result, we now know

that the Southern Ocean—one of the most inaccessible and diffi­

cult places to work—plays a major role in ocean-atmosphere car­

bon cycling and global climate modulation (Liniger et al., 2025).

FIGURE 1. Plots show time series from 1900 to the present of (a) atmospheric

carbon dioxide (CO2) measured from ice cores (black) and the Mauna Loa

Observatory (red) on the Big Island of Hawai‘i (Keeling et al., 2001; MacFarling

Meure et al., 2006). The trend in the partial pressure of surface ocean pCO2

(a measure of CO2 entering or exiting the sea) in Monterey Bay, California,

from the early 1990s is also shown (blue; updated from Chavez et al., 2017b).

(b) Surface ocean pH data from the early 1990s to the present are shown

here from the Hawai‘i Ocean Time-series (HOT) program (red; Karl and

Lukas, 1996) and Monterey Bay California (blue; updated from Chavez et al.,

2017b). Note that the pH scale is logarithmic. (c) The figure shows sea sur­

face temperature anomalies (seasonal cycle removed) from the California

Current along the US West Coast (Huang et al., 2017). Clear trends are evi­

dent for all of the measurements.