I first heard this phrasing at the American Geophysics Union Fall Meeting last December. I was reminded of it recently when reading The Pleasure of Finding Things Out: The Best Short Works of Richard P. Feynman.
Science is not in the business of certainty. Many people consider the purpose of science to be the pursuit of “Truth”. Or an effort to understand how the universe works. Both of these things suggest that there is a deterministic answer to the questions science asks. That’s just not true in most cases.
In reality, certainty is not the goal. Consider the Heisenberg Uncertainty Principle, for example. It’s a fundamental concept of Quantum Mechanics that you’ve probably heard about. Basically, the more we know about the position of a particle, the less we can know about the momentum (or velocity) of the particle. There’s an inherent uncertainty in the system such that even if we go looking for the answer, we’ll never be able to pin down both the position and the momentum of the particle at the same time. This is a prime example of the inherent role of uncertainty in science.
Not all examples of scientific uncertainty are so straightforward as to have uncertainty in the name. But uncertainty is always there. For example, we can’t see the future. We don’t know for sure what will happen in the future and that will always be the case. Uncertainty is the only “truth” possible in this case. This is why scientific uncertainty is usually expressed as an “expected” outcome — the most likely to occur — and a range of values around it that are also possible, but with lower probability.
I’d like to note here that in science, uncertainty does not mean knowing nothing. Scientific uncertainty is the quantification of an outcome that can change given unknown circumstances, like in the case of future carbon dioxide emissions scenarios. We don’t know what global emissions will be in the future, but we do have a sense for how different levels of emissions will lead to different impacts on society. We also know that certain emissions scenarios are more likely than others to become reality in the future. We know far more than nothing.
For applications such as climate science, making informed predictions about what will happen in the future is very important for decision making. In order for governments, companies, or individuals to mitigate or adapt to climate change, there needs to be a measurable estimate of what we expect climate impacts to be in the future given different scenarios. In this sense, communicating uncertainty is the most important thing we can do.
Let’s consider the example of drought. Perhaps one climate scenario projects that the current drought in California will last for 10 years. Whoa! That’s a long time and the drought is bad; Central Valley is already out of water. What will happen if this goes on for 10 years? As a farmer there, you’d probably be thinking of getting out of the farming business.
Now, what if I told you that the drought will last 10 years give or take 9 years? After all, I don’t have a crystal ball to know exactly what’s going to happen. This means that the drought could be over next year…or it could last as long as 19 years. Suddenly, the picture is much different and, as a farmer, your decision-making will be a lot different (and probably a lot harder!).
It’s uncomfortable to make decisions when faced with uncertainty, but that’s the reality of the situation. Even though it’s more difficult to plan for both the case of the 1 year drought and the 19 year drought, it’s important to consider both possibilities since we don’t know which it will be. Ignoring the uncertainty could mean you sell your farm this year — in anticipation of the 10 year drought — and then it turns out that the next 9 years are the most booming, profitable years yet. It’s important to carefully consider the uncertainty.
In fact, what’s even more important for decision-making than uncertainty is assessing risk. It’s harder to do, but risk is really the thing that dominates the bottom line. Risk is basically uncertainty weighted by the cost of your action. Even if we have high uncertainty, we can still have intolerable risks because the cost of inaction is just too high. This is the dilemma with climate change.
We are uncertain about how exactly climate change will impact society in the future. We have estimates of the impacts, such as sea level rise, but the uncertainty on these estimates can be significant. If the best case scenarios end up happening, then perhaps not much adaptation to higher sea level is needed. However, the cost — in dollars and loss of life — of the moderate to severe possibilities is perhaps too great to risk doing nothing.
In the end, uncertainty is the reason we get insurance. We don’t know what will happen, but we have some idea of the probability that something bad may happen. So we hedge our bets by preparing for things to go badly at some point or another. The amount of hedging we do is based on how great a risk there is that things will go badly. Therein lies the importance of scientific uncertainty: we must quantify the uncertainty in order to understand the risks we face.