The Risk-Impact Paradox of Global Cooling
Part 2 of my Global Cooling series
This is part 2 in my series on global cooling interventions. You can find part 1 and an outline for the multi-part series here:
Break Glass, Cool Planet
Here's the unvarnished truth: We need to buy more time for carbon removal to scale, and that means we need global cooling interventions – soon. Not as a replacement for carbon removal, but as a bridge to give carbon solutions the runway they need.
When I cofounded Nori in 2017, carbon removal was a challenging but reasonably contained problem. We were creating methodologies, building verification systems, and designing market infrastructure—complicated, yes, but with manageable risks. If we got something wrong, we might waste money or delay progress, but we weren't putting billions of lives at immediate risk.
Global cooling and other geoengineering interventions represent an entirely different beast.
The difference is so fundamental that it requires us to rethink everything about how we approach these solutions. Let me explain using a framework that helps classify different types of problems.
A Different Domain: Why Cooling Isn't Carbon Removal 2.0
The Cynefin framework (pronounced kuh-NEV-in) helps organizations identify what kind of situation they're in and how to respond accordingly. It defines four domains:
Simple: Clear cause-and-effect relationships. Follow best practices.
Complicated: Requires expertise, but cause-and-effect can be understood. Follow good practices.
Complex: Cause-and-effect can only be understood in retrospect. Probe, sense, respond.
Chaotic: No clear relationship between cause and effect. Act first to establish order.
Carbon removal in 2015-2017 operated primarily in the "complicated" domain with occasional ventures into "complex" territory for first-of-kind projects. We needed expertise and good practices, and there were risks, but they were generally bounded and local.
Cooling interventions in 2025 exist simultaneously in the "complex" and "chaotic" domains. We're dealing with:
Interventions where cause-and-effect can only be understood after deployment
Systems with so many feedback loops that modeling becomes inherently limited
Situations where small actions can trigger cascading, unpredictable consequences
Stakes so high that failures could affect billions of people
This isn't just a matter of degree—it's a categorically different challenge that demands fundamentally different approaches.
Consider the sort of response that Make Sunsets, the startup that is releasing balloons with small amounts of reflective particles, has received from regulators.
This very small-scale action by a tiny player triggered international backlash and policy responses. The reality is that in the chaotic domain of cooling interventions, individual actors can create ripple effects that shape the entire field—for better or worse.
There's another critical difference: carbon removal has a relatively clear commercial story. Remove carbon, get paid. It has not been easy to build this market, but the unit economics could eventually make sense. Cooling interventions lack that straightforward commercial incentive. Who pays? How? For what specific unit of service? These questions remain largely unanswered.
But the most profound difference is the potential for irreversible harm. With carbon removal, the worst-case scenario for most approaches was wasting resources without removing much carbon. With cooling interventions, we face the possibility of making things exponentially worse in ways we can't take back.
The Manhattan Project Scale of Ice Preservation
An area that I’ve been personally researching for months now (DM me if you want some good book recommendations) is ice sheet preservation. While more geoengineering than global cooling, I think the scope of the challenge and associated risks are quite illustrative.
The scale of what we're contemplating for ice sheet preservation is staggering. A recent Atlantic article put it bluntly:
"Halting the mighty flow of Thwaites would be a daunting challenge... A Thwaites field team could number 5,000 people—that's roughly the peak population of Los Alamos during the Manhattan Project, except in this case, they'd be deployed across one of the world's most remote glaciers."
Let that sink in. We're talking about potentially deploying thousands of people to one of the most inhospitable environments on Earth. The logistical challenges alone are mind-boggling:
Shipping containers by the thousands
Constructing facilities in conditions where "calm breezes can become gale-force winds in just minutes"
Operating in areas where "ice fog can white out the surroundings" without warning
Maintaining safety and operational capability in temperatures that can drop to -40°F
Perhaps the greatest irony? Such an enormous operation would likely require "a great and sustained conflagration of fossil fuels" to power the equipment, housing, and transport.
And here's a paradox that makes my head hurt: Interventions designed to preserve ice sheets might potentially accelerate their collapse if they disrupt existing dynamics. The Thwaites Glacier already has immense forces at work. As one experienced glaciologist told me recently, "Even if you are successful in removing the groundwater and freezing it, locking it down, you're still going to have accumulation of snow happening up glacier, which might increase flow." And that flow could result in even more ice calving off than would have happened if we had done nothing!
We're contemplating interventions in systems we only partially understand, with consequences we can only partially predict.
Flying Without Instruments: The Monitoring Gap
One of the most alarming aspects of our current state is how inadequate our monitoring capabilities are relative to the scale of interventions we're contemplating.
When I speak with researchers working on ice sheets, I'm consistently shocked by how limited our understanding is of what's happening beneath them. We're talking about potentially deploying massive interventions on systems we're still struggling to observe and measure accurately.
Our current situation is akin to flying a plane through a storm with only partial instrumentation. We know generally where we are, but not precisely. We can see some hazards, but not all. And the instruments we do have sometimes give conflicting readings.
Specifically:
We cannot accurately predict tipping points in ice sheets, ocean currents, or atmospheric systems
We have limited observational data about crucial processes happening beneath ice sheets
Our climate models, while impressive, still contain significant uncertainties when applied to regional impacts
We lack real-time monitoring networks that could detect unexpected consequences quickly enough to respond
This isn't a criticism of climate science—it's a recognition of the inherent complexity of the systems we're dealing with. The challenge goes beyond just technological; it's epistemological. Some aspects of these complex systems may be fundamentally unknowable before intervention.
To continue the aviation metaphor: We're contemplating emergency maneuvers in a storm without complete instrumentation. That doesn't mean we shouldn't act—it means we need to acknowledge and account for our limited visibility.
Learning from COVID-19: Our Pattern of Mismanaging Complex Risk
The pandemic provides a sobering case study in how our society handles complex, high-stakes challenges—and the news isn't good.
Despite focusing our full scientific and governmental attention on a single global threat, we still:
Failed to adequately balance competing risks
Made communication mistakes that eroded public trust
Struggled to coordinate international responses
Created unintended consequences that sometimes rivaled the original threat
The collateral damage in the US was immense:
Drug overdose deaths increased by 30% in 2020
Alcohol-involved deaths jumped 25.5%
Children's mental health deteriorated dramatically, with loneliness increasing by 28.3%
Learning loss affected a generation of students
And perhaps most damaging: public trust in scientific institutions eroded when authorities weren't fully transparent with the public about uncertainties, trade-offs, and changing assessments.
I'm not making a political statement here. I'm pointing out that even with unprecedented focus and resources, we struggled to manage a complex crisis effectively. And cooling interventions would present even greater challenges of coordination, risk balancing, and adaptive management.
One PR Disaster Could Derail Everything
The history of deliberate environmental intervention is littered with PR disasters that set fields back decades.
Consider the case of Russ George's 2012 ocean iron fertilization experiment off the coast of British Columbia. Without international authorization, his team dumped approximately 100 tons of iron sulfate into the Pacific Ocean. The backlash was swift and severe, triggering international investigations and tighter restrictions on ocean-based research.
Set aside whether you think that was acceptable or not for him to do—the outcome was such that this single experiment effectively poisoned the well for an entire field of inquiry. That is the reality of the world we live in, and we have to work within that.
The potential for a similar scenario with cooling interventions is significantly higher because:
The stakes are more visible and dramatic
The interventions are more easily politicized
Social media can rapidly amplify misunderstandings or genuine concerns
Trust in institutions is already at historic lows
The public understandably fears things that feel unnatural or hubristic
This isn't just a PR problem—it's an existential risk to responsible development of cooling approaches. A single botched deployment, unauthorized experiment, or even a well-intentioned but poorly communicated research project could trigger restrictive bans that prevent responsible actors from developing these capabilities.
We've already seen 31 U.S. states considering preemptive bans on solar geoengineering research, often fueled by "chemtrail" conspiracy theories rather than evidence-based concerns. One high-profile disaster could cement those restrictions nationwide or globally. Watch this video of a woman on Dr. Phil last month using the phrase “stratospheric aerosol injection” to describe her concern with chemtrails to a totally credulous HHS Secretary RFK Jr.:
The thing that shocked me most about that clip was the way in which she casually tossed the SAI term out there. What was once a very academic term for some niche researchers has hit the mainstream.
The Intervention Paradox: When Solutions Become Problems
The deepest paradox of cooling interventions is that the solutions that might save us also threaten us in ways we cannot fully anticipate.
Stratospheric aerosol injection might cool the planet overall but disrupt the Asian monsoons that billions of people depend on for agriculture. Marine cloud brightening might protect coral reefs but alter regional weather patterns that coastal communities rely on. Ice preservation efforts might inadvertently accelerate glacial flow through changed weight distribution.
This creates an impossible ethical calculus. We're forced to weigh:
Known climate catastrophes against uncertain intervention risks
Present generations against future ones
Human systems against natural ones
Different regions and populations against each other
There's no neutral ground here. Even choosing not to develop cooling interventions is itself a moral choice with consequences! If we avoid intervention research out of an abundance of caution, and then cross tipping points that could have been prevented, future generations may judge our inaction harshly.
“Do no harm” is unfortunately not an option. What we face is the agonizing domain of "which harm is least bad when all options include harm?" It’s trolley problems all the way down.
Navigating the Impossible: A Way Forward
Despite these profound challenges, the consequences of inaction are unacceptable. It is imperative that we develop cooling capabilities if only as an insurance policy against worst-case scenarios. But we need to do it differently than any large-scale technological development that has come before.
I want to emphasize that there are many groups who have done a significant amount of work to define ethical approaches to large-scale interventions. The American Geophysical Union last year published a white paper defining “Ethical Framework Principles for Climate Intervention Research.” The recently announced “Exploring Climate Cooling” research projects funded by ARIA in the UK also include extensive risk analysis in their program thesis document.
Here I want to lay out how I would approach a responsible path forward:
1. A risk-based approach focusing first on the most immediate dangers
Not all cooling interventions carry the same risk profile. We should prioritize approaches that:
Can be deployed locally with more contained effects
Have clear termination options if problems arise
Build on existing technologies and natural processes
Have more predictable outcomes based on analogous natural phenomena
Marine cloud brightening to protect specific coral reefs during heat waves, for example, presents a more manageable risk profile than hemisphere-scale stratospheric aerosol deployment.
2. Clear distinction between research and deployment
We need to draw bright lines between:
Computer modeling (lowest risk)
Laboratory studies (controlled risk)
Small-scale field experiments (managed risk)
Regional interventions (significant risk)
Global deployment (highest risk)
Each category requires different governance, oversight, and decision thresholds. Moving between categories should involve explicit, transparent decision points—not gradual mission creep.
3. Carefully controlled pilots with clear stop conditions
Any field testing should include:
Predefined metrics to monitor
Clear thresholds for pausing or terminating the experiment
Independent monitoring and verification
Transparent reporting of all results, including negative outcomes
Mechanisms to rapidly address unexpected developments
4. Transparency as a fundamental requirement
The COVID-19 experience showed that "noble lies" or simplified messaging ultimately backfires by eroding trust. With cooling interventions, we need:
Full disclosure of uncertainties and potential negative outcomes
Acknowledgment of knowledge gaps
Clear explanation of risk trade-offs
Open access to research and monitoring data
Inclusive deliberation processes that respect diverse perspectives
5. Unprecedented monitoring capabilities
Before deploying any significant intervention, we need to develop:
Comprehensive baseline measurements of affected systems
Real-time monitoring networks with redundant verification
Independent assessment capabilities
Rapid response protocols for unexpected outcomes
Data sharing infrastructure that crosses national boundaries
We Cannot Do This the Way We've Done Everything Else
In the complex world of climate interventions, paradox is inevitable. The solutions that might save us also threaten us. The scale needed for success creates unprecedented risks. The certainty we crave is fundamentally unobtainable.
But this paradox isn't a reason to abandon cooling approaches—it's a reason to approach them with unprecedented care, transparency, and humility. We must build new capabilities for monitoring, governance, and decision-making that match the scale of the intervention we're contemplating.
When I was building Nori, I learned that creating entirely new categories requires new frameworks, new markets, and new institutions. The challenges of global cooling interventions are exponentially more complex—and so must be our response.
This isn't carbon removal 2.0. It's an entirely different category of challenge requiring entirely different thinking. If we approach cooling interventions with the governance models, risk frameworks, and institutional structures we've used for previous technologies, we will fail catastrophically.
The risk-impact paradox demands that we create not just new cooling technologies, but new ways of managing risk, making decisions, and holding ourselves accountable. We have to consider not just whether we can develop effective cooling interventions—it is also whether we can develop the wisdom to deploy them responsibly.