When Georgia Tech climatologist Kim Cobb isn't busy raising her four young children, she spends her days scuba diving for tropical coral samples and developing models of El Niño events. That's because our current climate change models aren't good enough -- and Cobb is working to build better climate forecasts. On her mission to understand climate change, Cobb has been lauded for her work by the National Science Foundation, PopTech and other groups.
In a recent interview, Cobb explained what the recent past tells us about climate change, why precipitation matters and what it means to be "20 percent Indiana Jones." Below are excerpts from our talk.
Your goal in a nutshell is to better predict the progress of climate change. How do we do that currently and why isn't it good enough?
It's probably never going to be good enough. We'd like to always know important things better, so I'd couch what I'm going to say with that caveat.
We'd like to know about expected changes in precipitation, for example. We use large computer simulations of the climate system, including the ocean, the atmosphere, land and ice. We have some fundamental equations that drive the exchanges between these components of the climate model, such as how water moves from the ocean to the atmosphere. Simulating precipitation is very difficult. We rely on some parameters and shortcuts of physics in coding precipitation into models. It's something we need very high resolution models to do. We have a fundamentally incomplete picture of the physics that drives clouds and precipitation.
This ends up being a large uncertainty in the kinds of global climate models that we run in these 100-year projections. You can see this when you look at projections from a multi-model perspective. The suite of models don't necessarily agree in the sign of precipitation change for any given location on the planet. If you want to get down to a region, like the western U.S., you would be even worse off in looking at expected changes. This remains a strong limitation of our current suite of what we call models that are run in the 100-year simulation modes for future climate change. There are a couple dozen of these sophisticated climate models we use for projections. Our general assumption is that when they tell us something collectively, we believe that. Unfortunately, this is not one of the things they do very well.
What are you doing to improve climate model forecasts?
I'm looking to challenge the models to simulate climate changes that occurred in the distant past. Some of these are very large climate changes. But more often my work focuses on climate changes in the very recent past -- the past several thousand years. We resolve climate signals we believe may relate to a forced signal in the climate system caused by, for example, volcanoes. We look for a response to these changes. If we see it, that's an important answer. If we don't, that's another important answer. For the first time in history, the models have been run in paleo-simulation mode. That's a huge step forward for the climate community. Our best models are being explicitly tested with forcings from the more distant past. We have simulations from 6,000 years ago, from 1,000 years ago. They include solar volcanic variability, changes in carbon dioxide. They include this suite of forcings and we can look at the responses and compare them to actual data.
I generate the data that can be compared to these paleo-simulations. Increasingly, I'm becoming involved in doing those comparisons myself. I can compare apples to apples. It's exciting.
You'd think temperature is the most important piece of the climate change puzzle, but you argue that rainfall is critically important. What does it tell us that temperature doesn't? What does something like last summer's drought tell us?
Precipitation is a difficult beast in climate. It is highly variable. This is apparent to all of us. Ocean temperatures have a profound effect on precipitation distribution, statistically speaking. We can't see this on a climate change 100-year timeline, but we can certainly see this when we look at something as basic as hurricane generation. When do we have hurricanes? When the seasonal temperatures are the warmest they're going to get in the Caribbean. A good example comes from the El Niño phenomenon where warmer temperatures drive heavy precipitation along the eastern United States. That's a fairly subtle change in sea surface temperatures, yet it affects precipitation patterns around the globe in fairly predictable ways. This tells us that temperature is a primary control of precipitation patterns. Our challenge is to understand what drives the longer-term variability in precipitation.
Droughts are not unprecedented if you look at tree ring records as proxies for past precipitation. You can see these episodes of mega-droughts in these records. We can surmise that some of these may have been caused by changes in ocean temperatures. This [recent] mega-drought might have been caused by global warming, but all the ones in the past are presumably [related to] natural variability. It's something our planet does. It tells us that today's levels of drought are certainly not without precedent. Our job is to try to understand how these mega-droughts were caused. We don't have an answer for that right now, but it points to the importance of marrying our best paleo-climate data with our best models that can help us understand what causes drought in this region. That's a very promising avenue for improving projections of precipitation.
In the bio for your Twitter feed, @coralsncaves, you describe yourself as "20 percent Indiana Jones." What does your fieldwork entail?
I'm a field junkie. I have two primary sites in the Tropical Pacific with my coral work and in Borneo with my cave work. That puts me in a rare breed of scientist who is willing to fight through Third World bureaucracy and a lot of heat and rain and bugs to pursue climate records of unique value in reconstructing El Niño. These places have very large climate signals associated with El Niño events. I'm one of the only people knocking down these beautiful, difficult places in pursuit of these records. We've gotten quite a few interesting records out of these sites, which challenge dominant paradigms. You get to scuba dive and sweat and curse. You get the reward after three or five years of working on these records of producing something that is profoundly different.
You've said scientists haven't sold to the public the idea that CO2 is warming the planet. Why haven't they?
There are so many different explanations. From my own perspective, we have stressed impacts so much more than we have stressed the fundamental sets of observations that underpin that simple statement: carbon dioxide is warming the planet. We've relied on a communication that's built from impacts because we want to move people to act. The simple statement that the Earth is warming because of carbon dioxide doesn't seem necessarily like it's going to change hearts and minds. Where are the dollar signs? I actually think it might be enough to win hearts and minds. That's a pretty mind-boggling concept: that we're warming the entire planet. If people accept that as a scientific fact, that has immediate impacts that would be clear. For example, sea levels will rise. There's no physical way around that conclusion if you accept the premise that CO2 is warming the planet. We've sold short the average human's reaction to that piece of information.
As a climate scientist, I feel that if I can convince people through the bounty of evidence that CO2 is warming the planet, I stop. It's a huge victory for me. In order to move this debate forward, we're going to have to have a different set of toolkits and a diversity of voices. I try to get a very simple, one-dimensional, strong message across.
Photo: Kim Cobb / By Zafer Kizilkaya