Gavin Schmidt on the evolution, testing and discussion of climate models

gavin schmidt nasa giss
Credit: Bruce Gilbert

Gavin Schmidt runs the NASA Goddard Institute for Space Studies and develops, pokes, prods, tears down, builds up, and talks about talking about climate models. He also considered opening a child care center at GISS, but thought better of it.

I’m a complete neophyte when it comes to doing interviews. Most of the time, I’m trying to think about my next question, making sure that I’m still recording, and listening for distracting background noises. It’s hard to keep track of the overall theme  of the conversation.

This time, it wasn’t until I was in the editing phase that I understand that Gavin was essentially laying out an entire philosophy of how climate models get built. evaluated, and presented.

At least for the NASA GISS ModelE, the model didn’t leap fully-formed, Athena-like, from any sort of centralized plan. Rather, it evolved over decades, with many of the innovations coming from seemingly off-topic work pursued by not-overburdened GISS employees. As Gavin says, “Just get on with it. Whatever you want to do, just do it”. That sort of attitude led to novel components, like direct simulation of isotopes and atmospheric chemistry.

Model testing is key, of course. You make a change, see what happens, test the results against observations, and make a decision about what to keep, what to simplify, and what to place in the circular file. As the model develops, you can then use it to evaluate the physical mechanisms that drove abrupt changes, for example the 8.2 kyr event.

Models develop by fits and starts, experimenting, replacing parameterizations with processes over time. How do you know a model is improving? At least partially through emergent behaviors. As you increase resolution, you’re not specifying that atmospheric transport across the tropopause gets better … but it does, and that improves other processes like the quasi-biennial oscillation.

We talked extensively about science-government interactions, and some of Gavin’s many  Kafkaesque experiences. In the end, attempts at government muzzling and micromanagement of science communication comes across as impractical, appalling, and … a bit comical.

But it’s not all bad. The complete government failure to engage in any sort of response and discussion of climate-related science fiction like The Day After Tomorrow led Gavin to participate in the RealClimate blog.

Gavin, like many of the other climate scientists on Forecast, has some level of optimism for the future. The carbon-GDP relationship isn’t a law of nature and it does seem like governments are starting to come to grips with the size of the problem.

Leave a Reply

Your email address will not be published. Required fields are marked *