Friday, October 4, 2013

Why Hurricane Season Predictions Are Bad for the Environment

I had written this a couple days ago and was going to publish it on Sunday. I had to move the date up after I saw this horrendously asinine article about assessing the economic risks of climate change.

This article is asinine because: (1) You cannot assess the risk of climate change. This would require economic models based on climate models. Neither one of the two lends itself to model making; (2) Economics/Finance has talked a lot about risk assessment and risk avoidance. They suck at both. What was the Great Recession if not proof that these eggheads either can't assess risk or they can't do anything about what they know?; (3) We know the climate is changing. We know that is bad. We know why it's changing. We know how to stop it from changing. Why do we need to assess the economic risks of it? Are the human risks not enough to spur action?

The following was not written as a response to this article, but it addresses it well.

Now onto the main event:

Here in Miami there are three seasons, not four. There is summer. There is not-summer. Then there is hurricane season -- the dates of which are ingrained in our minds: June 1 to November 30. Every year before hurricane season starts, someone decides to predict how active it will be. They include the number of named storms and the number of major hurricanes. This year's prediction was horrendous. See here, here, and a ton of other places. But why is predicting how active a hurricane season will be bad for the environment? It all comes down to models.

Models abound in science. They can be simple or complex. They are found in both the hard sciences (physics, chemistry, etc.) and the soft sciences (economics, sociology, psychology, etc.). They are often helpful. They are sometimes damaging. The degree to which they either help or hurt is related to how accurately they can predict an outcome. Usually, the more complex a system, the less likely its model will be accurate in predicting outcomes. Hurricanes are a perfect example. The region of the earth's atmosphere that influences the formation of ONE hurricane is huge (one could argue it's the entire earth). Current models for predicting whether ONE hurricane will develop, its track, its size, and its strength are definitely important for those who may be affected. They definitely save lives. At the same time, they are only accurate to an extent. Hence the "cone of error." I am in no way saying we should stop tracking hurricanes or shut down the National Hurricane Center. Tracking individual storms is necessary and the accuracy of it is high enough to merit its funding.

The problem is trying to model an entire SEASON of storms. It's ridiculous. If someone claims they have an accurate model to predict an entire hurricane season accurately, they deserve nothing more than to be ignored. Yet every year they make headlines. This year they made two: the first was the initial prediction; the second was how wrong they were. Some things do not allow models to be made. At least accurate ones. Included in this group is also global climate. If ONE season of hurricanes in ONE ocean's tropical basin is too complex to model, then an entire planet's climate is even more so. Yet we try to do it.

It is hard to read an article, let alone a book, on climate change that doesn't make predictions. Whether it is average temperature in 2050 or the amount of sea level rise by 2100, experts seem to think that making predictions is their job. It's not. Now, certainly there is a range of average temperatures or sea level rise that one could predict within a certain probability, but this is never how it's presented. It's usually presented as a fact. I recently saw a link to a site where you enter your birthdate and it will tell you what how much higher sea level will be on your xth birthday. As if it were a simple input/output calculation. There are simply too many feedback loops that all affect each other, some of which we surely don't even know about, to pretend we can predict anything about future climate change with any certainly. This is important to admit, but it doesn't mean science is useless when it comes to climate change.

Unfortunately, when some hack tries to predict the hurricane season and fails miserably, another hack climate change denier or skeptic will seize on that and say "look, we can't even predict this year's hurricane season, how on earth can we predict the climate 50 years from now?" And that is the link that makes hurricane predictions so dangerous. One group makes assertions about science that go beyond what it can actually do. Another group takes that to mean that science can't do anything. The loser in this whole thing is the environment, of which humans are a part.

So if science cannot produce models to accurately predict the climate 50 years from now, what good is it? My answer would be that we should consider what science has accomplished concerning climate change not by looking into the future, but by looking into the past. Science has connected the dots that allow us to know that global temperatures are increasing. That CO2 concentrations are rising. That much of this is because of human actions. That seems like plenty to me. Why look to the future when the past is so certain?

On a recent EconTalk podcast, Robert Pindyck boils it down nicely. When it comes to predicting the future, science is able to tell us fairly accurately some broad generalizations.
  • Temperatures will continue to rise. This is mostly due to humans.
  • This is not a good thing. Rising temperatures will change geographic features. While some areas we see desertification, others will see greening. So we may be able to move from one area to another, but the changes will happen too quickly to adapt easily. Communicable diseases will increase. Sea level rise is highly likely. Etc.
  • We should do something about this. (1) Cut carbon emissions so that CO2 concentrations fall below at least 1990 levels (Kyoto Protocol). (2) Devise and implement ways to mitigate climate change. This is not so that we can skip doing (1), but because the rate of change of the climate, even if we cut ALL carbon emissions TOMORROW, will be decades in the making. So either way, we will need to deal with higher temperatures and what go with them.
I think Pindyck nails it. Science may be limited in what in can predict about future climate change, but when even the generalizations are scary, why do we still sit on our hands?

No comments:

Post a Comment