What's This Actually About?

Understanding the Complex: Climate Models — How Do We Know What Happens in 2100?

---

June 23, 1988. Washington, D.C. The Senate chamber is sweltering — the air conditioning has broken down, which a few people in the room will later suspect was deliberate. James Hansen, a NASA scientist from Ohio, wipes his forehead and tells the United States Senate something no scientist has said in that room before.

"The greenhouse effect has been detected," he says, "and is changing our climate now."

Not "may change." Not "could change in the future." Is changing. Now.

The committee room buzzes. This isn't the hedged language of a scientist covering his back. Hansen is stating a conclusion. And he's basing it on something most people in that room have never heard of: a climate model.

In the weeks that follow, Hansen is on the front page of every major newspaper in America. His testimony becomes the moment many people point to as the birth of public climate consciousness. Time runs a cover story. Margaret Thatcher delivers a speech to the Royal Society making many of the same points.

But almost no one asks the obvious question: how does he know?

---

Hansen didn't look out the window and decide the climate was changing. He ran simulations.

He built a mathematical representation of the Earth's atmosphere — a model composed of physical equations, fed with data about solar radiation, ocean temperatures, atmospheric composition, and dozens of other variables — and ran it forward in time on a computer. The model told him temperatures would rise. When he compared those predictions to measurements taken over the following years, they held up.

Not perfectly. Models never predict perfectly. But well enough that the basic conclusion — greenhouse gases warm the planet — has been confirmed by decades of subsequent observation.

Hansen's 1988 congressional testimony cited findings from a paper he co-authored that year, which showed a 99% statistical confidence that the warming trend observed in the 1980s could not be explained by natural variability alone.

The model was right. The politics that followed are a different story — but the science has held.

---

Here's what makes climate models strange territory compared to the other subjects in this series.

For LLMs, quantum computers, or nuclear fusion, the central question is "does this work, and what can it do?" The science and the politics are separable. You can explain how a transformer architecture functions without taking a position on AI regulation.

Climate modeling doesn't offer that luxury.

The science and the politics are so entangled in public discourse that "trust the models" has become a political signal rather than an empirical claim. Meanwhile, "the models are wrong" has become a way to avoid policy conversations that some find uncomfortable — even when the specific critique has technical merit.

This course tries to cut through that.

Climate models are physics. They're built from the same laws of thermodynamics and fluid dynamics you'd use to design a jet engine or predict the path of a hurricane. They're not political documents.

But what they tell us has immense political implications. And the space between "here's what the physics says" and "here's what we should do about it" is larger than most public conversations acknowledge.

Understanding that space is what this course is for.

---

The central question:

How does a climate model work — and how much can you actually trust one?

That question has two parts that are usually run together but need to be kept separate.

The first part is mechanical: how do these things actually function? What goes in, what comes out, and why does the output say what it says? We can answer this. Not completely, but well enough that "the models say so" stops being a conversation-stopper and starts being the beginning of an interesting technical discussion.

The second part is epistemological: what does it mean to trust a projection about conditions a century from now? Models have uncertainties. Some are quantified; some aren't. Some projections are more reliable than others. Understanding why — which parts of the output to hold tightly and which to hold loosely — is the most useful thing this course can give you.

The answer isn't "trust everything" and it isn't "trust nothing." It's more interesting than either.

---

One more thing before we start.

Climate change is one of the most politically polarized topics in the developed world. This course takes the scientific consensus as its starting point: the Earth is warming, human activity is the primary driver, and the physical mechanisms are well understood.

That consensus is strong. A 2021 analysis found that 97% of actively publishing climate scientists agree that recent climate change is primarily human-caused — a finding consistent across multiple independent surveys going back to 2009.

But "the Earth is warming" is a scientific claim. "We should implement a carbon tax" is a policy claim. This course has something to say about the first. On the second, it presents the arguments — and steps back. Reasonable people disagree about climate policy for reasons that have nothing to do with whether climate change is real. Those disagreements are yours to navigate.

What the course gives you is a better map of what the science actually says, and doesn't say, so you can navigate it with more accuracy.

---

Next lesson: Why should I care? — Three reasons climate models matter beyond the science.

---

Reading time: approx. 8–9 minutes