Friday, January 9, 2015

Finding a Signal in the Noise

by Ben Brown-Steiner

(Note: This post follows up on ideas presented in my previous post, and I highly recommend you read that post before this one).

Take a look at the following two graphs.

Screen Shot 2014-11-12 at 5.38.42 PM.png Screen Shot 2014-11-12 at 5.39.00 PM.png

They both cover the same years (1986 - 2007), and I’ve removed the vertical axis labels because that would (for the moment) ruin all the fun.

Before I give hints to what these two plots represent, can you venture any guesses? Is there a signal in either of these plots or is it just noise?

For a first pass I’d say they both are generally increasing, but not consistently. They both wiggle, although the one on the right wiggles more dramatically (higher variability). The left one seems to plateau and then drop off after 2004, while the right shows a large jump around 1998 and seems to plateau after that.

Now for a hint: both of these plots represent something which we suspect has changed or is changing over time, and we have some expectation that we’d be able to detect these changes by studying these graphs. Can you guess where these changes happened (either one year or a range of years)?

A second hint: in one of these graphs, a distinct change happened in 1998. In the other graph the changes have been gradual over time.

Alright. The left graph is the number of home runs hit by Barry Bonds each year throughout his career. It’s generally accepted that Bonds started taking steroids in 1998. The right graph is the average annual temperature anomaly (meaning the mean temperature from 1951 - 1980 has been removed) over the US, and it’s generally believed that the climate has been warming over these years.

And, almost maliciously, the graph of Bond’s home runs doesn’t show a clear jump after 1998 (when he started taking steroids) while the temperature plot does. While we could speculate that the US temperature spikes as a result of Bond’s steroid use, it’s better to look at the 1998 jump in temperatures as a result of the 1997/1998 El NiƱo event (which I’ll write about in a future post) and the plateau afterwards as some form of variability (see my previous post).

What can we say about the influence of steroids on Barry Bond’s home runs? We can confidently look at year-to-year changes and try to explain what we see because we would expect an athlete to improve every year, reach a peak, and then either decline or retire. We expect any changes to his body (i.e. steroids) to be reflected in the amount of home runs he makes in a year. We see that before he started taking steroids, his home run total was in a slight decline. We also see that after he started taking steroids, his home runs spiked. However, after 2001 his home runs dropped again. Perhaps this is because he stopped taking steroids, or maybe he was just getting old (I’m not really a baseball fan so don’t know much about Bond’s career).

[As a side note: steroids actually make an excellent climate change analogy. See this video from AtmosNews.]

What can we say about the temperature records and their fluctuations? Since this time period is over 20 years, and we aren’t really talking about climate until we’re looking at at least 20 years (see my previous post), we can’t really say much. The year-to-year fluctuations are so large that it’s hard to draw any strong conclusions. To get a better idea of the climate, let’s look at the full US temperature record (1880 - 2011):

Screen Shot 2014-11-12 at 5.58.42 PM.png

We can see more clearly now an increasing trend starting in the 1960s, but there’s still a lot of wiggles (or noise). One common method for reducing the noise level (also called smoothing) is to take a moving average. In the following figure, every yearly datapoint is the moving five-year average (we average the two previous years, the current year, and the two future years together) from the same data as the previous graph:

Screen Shot 2014-11-12 at 5.59.57 PM.png

Without the annual noise it’s easier to see a trend, especially after 1960. This particular dataset stops at 2009, and I want to note that the following three years were all warmer than 2009 [1]. This method has allowed us to reduce the “noise” which enables us to detect the “signal” better. We can also see the “warming hiatus” during the last 10 years, but once again, 10 years isn’t long enough to really be climate yet. It’s still weather. I’ll write a post about the warming hiatus in the near future.

There’s so much more we can explore with climate signals and weather noise (and I will address more of these in future posts). But for now, let’s leave it here.

The data for the plots was obtained from these sites:
[1]: http://www.epa.gov/climatechange/images/indicator_downloads/temperature-download1-2014.png

Tuesday, December 23, 2014

Calling Uncle Fred a Jerk Won't Help: The Having of Hard Holiday Conversations, 2014 Edition

For most of the last several years, we've shared ideas about how to talk about climate change at holiday gatherings. Welcome to the 2014 edition! 

With PRI's educational work in evolution, climate change, and, more recently, hydraulic fracturing, we think about how to address controversial issues all the time. I share things about all of this stuff on my Facebook wall and I'm connected to a number of groups where we hash this stuff out. Today's post comes from a Facebook exchange that grew out of a friend's concerns about Thanksgiving with a brother-in-law who doesn't accept that climate change is real and human-caused. 

While the focus of what follows is on climate change, it highlights general ideas related to talking about many controversial issues. I'll spill the beans early - once you insult someone's motives or intelligence in an argument, you've likely lost your opportunity to help them understand your point of view. Treating those who disagree with you as the enemy may have its place, but its almost certainly not the holiday dinner table. And, we should consider how much of a place it has at all. The post is largely about how to make an honest change in the direction of the discussion that might yield more productive results. 

While I'm updating this text that was mostly written in November, I'm also thinking about the horrible things that have been going on related to police both giving and receiving brutality. So much of that national discussion has simply been ugly and unhelpful. I invite you to think about how the thoughts I share below about climate change connect to discussing other controversial issues. 

Cutting to the chase


Most of the rest of the post is my response to a long back-and-forth exchange about how to talk to a relative with a vastly different worldview from the person who raised the question. Many of the posts by others in the conversation detailed the science of climate change and pointed to resources to back up evidentiary claims, that is, to ground the argument in the science. I was a late comer to the discussion, and read through the exchange before piping in. I've lightly edited my contribution to fix a typo or two, and to make it more clear, but it's essentially what I wrote in response to a long series of posts about what evidence to use. 
I read through this thread this morning, and have been mulling if for the day. Obviously, the science is essential - but it's also pretty obvious that just dishing science isn't sufficient. My educated guess is that your brother-in-law's worldview makes it next to impossible for him to accept the science.
 That's likely grounded in concerns about what the commonly proposed solutions will do to his way of life (and the American way of life).
 It's obviously not an indication that he's stupid - as you note, he's not. Getting heated about it likely will make matters worse - advocates tend to deepen people's convictions more than they deepen people's understandings, and many advocates manage to deepen convictions at both poles of polarizing issues. That's partly because they often insult those at the pole that's opposite from theirs, and that's a terrible strategy. I don't get the vibe from reading Fred's (not his real name) comments that he's likely to do that, but it's a common problem – a really common problem.
 Think about big ideas you've changed your mind about and what led to the change of heart. It likely wasn't sped along by anyone who insulted you, and it likely took a long time. It may have involved someone with kind reassurances about accepting whatever idea wouldn't ruin your life. That could be explicit, or it could be by example. I think gay people coming out of the closet and basically just being as nice as anyone else is fundamental to the change in attitude about gay marriage. For climate change, it makes more sense to talk about other issues that have the same root cause - wasting energy, particularly from fossil fuels. Almost no one will admit to liking waste as part of his or her value system. And, the waste from fossil fuels is often truly filthy and despoils the environment in a wide range of ways. If he's amused by the rolling coal movement (Google "Rolling Coal" if you don't know what it is), you can point out that it's not only filthy, it's also unpatriotic - wasting oil helps keep its price high which helps support groups like ISIS, even if the petroleum wasted comes from the US. With oil as an international commodity, wasting it anywhere helps support those nasty people.
 It's also allowing freeloaders to use our shared atmosphere as dumpsite for their waste. Nobody likes freeloaders either.
 You might well be able to get him to agree to all of that. It's not global warming, but it's got the same set of fixes.
 Another thing about changing your mind on big issues - it usually takes a long time, and a gradual build up of evidence that fits into a bigger picture. You likely won't win him over this Christmas, but maybe if you keep at it in a kindly fashion, it'll happen down the line.

There's a lot more to say on this issue, but keeping it civil and reflecting on important ideas where you've had a change of heart could go a long way toward improving the discussion and maybe making some progress. 

This year's edition has a somewhat different spin than earlier years. If you'd like to look back to previous years, here are the links:
We've been overwhelmed by spam in the comments on the blog, so they've been shut off. Please comment through social media. 

Happy Holidays!
Don Duggan-Haas

Friday, October 24, 2014

What's the climate like outside today?

By Ben Brown-Steiner

Winter is coming. We all know this. But let’s say that I didn’t know this. What could you do to convince me that winter is coming? If we went for a walk outside what things could you use as evidence?

You could use the colors of the leaves and a description of the seasonal cycle of trees. You could point out the frantic behavior of squirrels and an explanation of what they are burying and why. Both of those things are individual pieces of evidence and a logical explanation of how they fit into a larger pattern, and as such they are pretty convincing.

If it was a cold day you could talk about the decreasing temperatures and talk about how you wished it was still August. That would be pretty convincing. But what if it was an unusually warm day? You would be forced to tell me to forget about the day’s weather and focus on the longer trend. You would tell me to ignore what appears to be counter evidence for your claim. If I were skeptical that winter was actually coming, I may not believe you.

Looking for evidence in an individual day for a change in season, especially if you are talking about temperatures, is really tricky because all we can experience when we step outside is the weather. If I wanted to know what the weather was today you could respond like Calvin’s mother does to his question:
Weather is the day-to-day, hour-to-hour, minute-to-minute fluctuation of the atmosphere we all live in, which is chaotic and highly variable. In order to perceive a seasonal change we need to pay attention to the moving average of daily temperatures over the span of weeks or months. This is because ultimately a change in seasons is not weather. It’s a change in the long-term average temperatures that follows an annual cycle dictated by the tilt of our planet. This cycle is so consistent that during certain times of the year we all expect to see evidence of the change in seasons and therefore we perceive changes in temperatures as evidence of a change in season. During the autumn, it is very easy to experience a cold day and feel that winter is near, or experience a warm day and lament the passing of the summer. But our day-to-day experiences are moments in the constantly fluctuating and highly variable weather. On their own they are not evidence of a change in seasons.

To understand the seasonal cycle, or to understand any long-term average, we have to rely on observations, data, statistics, and pattern recognition. This is especially true if we’re trying to understand the climate. This is because, even more so than the seasons, climate is a long-term average (years to decades) of the weather and averaging is a statistical tool that is abstracted from the weather that we experience every day. For instance, if the high today is 77°F and the low tonight is 35°F, the 24-hour average temperature is around 56°F, which is a poor representation of the hourly temperatures that we actually feel on that day. Even an average of temperatures over a single day is abstracted from real-world experience and is about as useful as that broken clock that is right twice a day.

To get an intuitive feeling for how abstract climate really is, let’s look at Decembers in Ithaca over the past ten years. The following figure is the daily maximum temperature (red), minimum temperature (blue) and average temperature (green) for December 2013.

December of 2013 felt like a weird one. It dropped below 0°F on December 17th only to reach above 65°F on the 23rd. What can we predict about this December based on last December? The unfortunate answer is: not much. It would be foolish to predict that this December would match any of the specific highs and lows from last December. Weather is highly variable. What if we look at the average climate for Ithaca in December? The following figure is the same as the one above but with the climate average for each day in a purple line.

The December climatology shows that, on average, the mean temperature drops from 35°F on December 1st to 25°F by December 31st. Interestingly, last December looks nothing like the climatology. Why is that? How can we say that we would expect any given December to be like the December climatology when last December was nowhere near the climatology? Was last December an unusual December? We can’t answer these questions based on any single December, so let’s look at the past ten Decembers plotted in the following figure. 

Trying to pick a ‘normal’ December is difficult. They generally show decreasing temperatures, although not always. Temperatures actually increased throughout December in 2007. None of them actually match the December climatology. They all show temperature fluctuations, but the fluctuations don’t really show much of a predictable pattern. If we were to make a statement about the weather for this coming December, all we could really do would be to state the climate average (“Decreasing daily average temperatures from 35°F to 25°F…”) plus make a statement about the average temperature fluctuations and their frequency (“...with deviations from that trend around 10°F every 5 – 10 days”). That’s a climate forecast. It’s not a weather forecast.

We’re back to that fundamental difference between weather and climate. The climate average is so abstracted from our actual experience that it’s impossible to feel the climate in any meaningful way. All we feel is weather. It’s only with statistics and averaging that we can experience climate. To get a weather forecast for December, we’ll have to wait until late November when actual weather models can start to make meaningful weather forecasts.

The following four figures should help us get a more intuitive understanding of these differences. They are the averages of the last 2, 5, 10, and 20 years of daily December temperatures. At only two years, you can still see the influence of the warm days from 2013. Early December in 2012 was also warm, and a 2-year average still is subject to the random patterns of weather. We’re not yet looking at climate.

At five years, much of the year-to-year variability is smoothed out.

At 10 years, the temperature plots are even smoother. It’s looking more like the climatology now.

At 20 years, it looks even better. The green line (average daily temperature) lines up pretty closely with the long-term climatology.

But by the time we’ve averaged 20 years together, we’ve moved from the concrete world of daily weather that we can experience into the abstracted world of climate averages. As we increase the amount of time we’re averaging together, we see less of the impact of weather variability and more of the average abstracted climate trend. It’s not until we’ve averaged at least 20 years together (and many climate scientists would rather average 30 or more years together) that we can even talk about climate.

When anyone talks about climate, or a change in the climate, they are by definition not talking about weather. It’s important to recognize that a single weather event may be extremely warm, cold, wet, or dry, but that event cannot be used on its own as a line of evidence in a conversation about climate. To talk about climate you must talk about abstracted averages of weather that cannot be interpreted or perceived as a thing that we can directly experience.

So next time you find yourself in a conversation that confounds weather and climate, remember that the two things are not interchangeable. Instead of responding with confusion to a question like:

Respond with a brief explanation of the differences, perhaps utilize some statistics, and then go outside and enjoy the weather.

Friday, October 17, 2014

The Climate System as a Foreign Sports Car

by Ben Brown-Steiner

Imagine the earth’s climate system as a foreign sports car owned by a playful friend. Your friend is happy to let you look and listen to the car and will on occasion give you a ride in it. But he’s unwilling to let you look under the hood or study the user’s manual. If you want to try and understand how the car works, and maybe even build your own, you are going to need to be clever and use all the tools at your disposal to figure out how this car works.

This is very much like how earth scientists try to understand the real world. The real world, even more so than the foreign sports car, is very complex and intricate. We are unable to look under reality’s hood or read reality’s users manual. What we are able to do is to make careful observations, come up with theories as to how the real world works, test our theories with experiments and create models to test and explore our understanding.

When you start to build your own car, the first thing you are going to gather are the major parts: a frame, an engine, wheels, rods, doors, spark plugs, fuel and so on. For the Earth’s climate, instead of parts we have variables: ocean temperature, relative humidity, solar radiation, CO2 concentrations, and many others.

However, you need a plan to put your parts together.  You need an understanding of some of the interactions between the parts you have and the function of your car. You know that fuel needs to be injected into the engine in a particular way. You know that the wheels need to be connected both to the engine and to the brakes if you ever want to actually drive your car. Earth scientists know that sunlight heats the surface of the earth, which in turn heats the atmosphere. They know that warm air holds more water than cold air. Any successful climate model needs to have these elementary parts.

To put these parts together you need a design framework. For our car, this is a engineering schematic of the parts and their functionality. We know when brakes are applied, the velocity of the car decreases. For earth scientists, this is typically in the form of code and equations. We know that precipitation falls as rain if the temperature is above freezing and falls as snow if the temperature is below freezing.

During our first attempt, there are many things we don’t know. We know that there is an interaction between the stick shift and the engine. We know that there is some interaction between the oceans and the atmosphere. But we don’t know the exact details. To begin to understand these unknowns we must make careful observations of the thing we’re trying to model and come up with tentative hypothesis and theories. We can listen to the revving of the engine or make observations of ocean and atmospheric temperatures.

You then put together a scheme that you think might work like the real thing. You design some system (or write some code) that combines what you know with the things you have hypothesized. Because you know that you are uncertain about particular interactions, you make those parts easy to observe and easy to modify or swap out with another part. This is what often is described as a parameterization or a scheme in a climate model. It’s a variable or equation that you know you’ll have to tweak or change out later on.

For instance, after you run the car you notice that your car moves backwards when you think it should be moving forwards or your climate snows when it should be raining. You look carefully at your variables and design and equations and parameters and try to find the error. You may have had your wires mismatched our you may have had inaccurately represented the relationship between moisture and temperature.

The next step takes this newfound understanding and incorporates it into your next model. You fix your mistakes and you tune your parameterizations. Then you test it again and repeat the whole process over and over until your understanding grows. Fundamentally, this is is the way that science functions. It is an interactive process. It’s never ending. You test and tune and observe and reformulate and repeat. Eventually, if you are clever and lucky, your model gets better and you gain a deeper understanding.

Unfortunately, since we are not able to look at the actual foreign sports car (because our friend is too secretive) and since we will never see the inner workings of the real world, we are never going to have a perfect model. Models by definition are simplifications of the real thing. You strive to have a really good simplification that provides insight and understanding. And you keep trying. This is science.

Monday, September 15, 2014

The Grid Box

by Ben Brown-Steiner

The above image is taken from a popular indie game called Minecraft that gives players free access to an environment with in which they can interact with water, land, lava, and other natural resources. It’s blocky textures are a unique style for modern games, but serves as a wonderful example of how climate modelers look at our world.
The Earth is huge and complex. It’s easy to lose yourself observing the infinite details in an ant colony, a thunderstorm, or a sunset. And while that can be fun, if we’re to understand what’s happening over the entire planet, we need to divide it up into parcels and we need to try and understand how each chunk interacts to its surroundings.
Climate scientists who use models are become comfortable with viewing the Earth as a set of boxes. In order to get our increasingly powerful computers to create realistic simulations of our world, we have to divide up the atmosphere, the oceans and the land into grid boxes. For each grid box the computer tracks a single value for each variable. A grid box has one temperature, one value for cloudiness, one relative humidity, and so on. And a typical climate model has a time step of roughly 20 to 30 minutes.
Currently, we consider a high-resolution global climate model to have boxes that are around 60 kilometers or 35 miles per side. That’s roughly the length of Cayuga Lake. So a global climate model isn’t capable of seeing Cayuga Lake! How is it possible that a grid box that encompasses nearly all of Cayuga Lake and that jumps forward in time every 30 minutes has any chance of being realistic?
It’s because, somehow, astonishingly, the movement of mass and energy in the real world is amazingly organized. To some degree, it’s intelligible, which means that we can study it, take some notes, and understand the basic principles of how the it operates.
            Even though you could be standing in a parking lot in direct sunlight blinded and sweating and someone 100 feet away could be lounging in the breeze under a shady tree, if you average all of the temperatures in a region and track those values over time, you find smooth and regular patterns. If you average every temperature in a day (from the high temperature in the afternoon to the low temperature at night) and draw a graph, you can see the regular and repeating cycle of our seasons.
If you are examining a region’s climate you look at these average cycles. You ask questions like: Are the seasonal highs and lows changing over time? Is it drier, on average, this decade than it was last decade? How does the average state of our region influence the surrounding regions and the global climate? These questions are the questions of climate science and the purpose of these questions is to examine averages. In this sense, a large, abstracted grid box, is a really great way of looking at the big picture.
What, however, if you care about the high temperature tomorrow? If you want to know whether you should bring an umbrella on your walk right now, you do not ask a climate scientist. If you did, you would get a funny look and perhaps a response like this: the average daily high for September in Ithaca is 70 degrees and on average it rains 3.5 inches.
If you ask the question “What’s the weather going to be tomorrow?” you would ask a meteorologist, and the meteorologist, because they don’t have to worry about the entire globe all at once, can zoom in on smaller atmospheric patterns. They can run simulations with time steps as short as 10 seconds and grid boxes as small as one mile (~1.5 km) per side.
To look at a concrete example, let’s say you’re a climate modeler trying to capture interactions between clouds, the Earth’s surface, and solar radiation. You recognize that the real world is complicated and chaotic, but you know that there is some underlying structure that you can model. You take a look at a satellite photo over the ocean and it looks like this (from NASA):


How are you going to recreate this reality in your model? Clearly, you’ll have to simplify. If this particular picture is 100 miles per side, you know that your computers can’t capture that level of detail and have any chance of running on your computer. You need grid boxes. What size grid box is appropriate? If you had the computational power, you could build something like this:


Even though you know you’ll have to make simplifications, and parameterize (we’ll talk about these in another post) some of the small cloud features, you can still represent the overall cloud system you see in the satellite photo. But then your IT staff tells you that there’s no way you can run the model at that resolution. You need to try a somewhat coarser resolution:


You aren’t all that happy with this one since you lose a lot of detail and you’ll have to make different and broader assumptions. For instance, you’ll have to completely forget about resolving individual clouds. You’ll have to start representing cloudiness in a grid box as a percentage (real climate models do this). After some time, you hear from your IT staff that you can run this resolution, but it will take three months to run 1 year of your simulated Earth at this resolution. For your purposes, that’s not practical. Since you don’t want to give up you decide to go to an even coarser resolution with grid boxes of roughly 35 miles per side:


This one leaves out even more detail. You can hardly recognize this as a system of clouds anymore. But in exchange you can run your model at this resolution much more quickly and you’ll be able to examine the details of what you think is going on with much more confidence and data points. Right now, this is the grid box size of many climate models. And even though they use this resolution, they can be used to understand our Earth. The following image is an example of North America viewed through a variety of resolutions used today:


            The top left resolution is a course resolution used in the past. The top right is roughly the resolution of the average climate model today. The bottom left image is roughly the resolution that is considered high resolution today (the resolution that doesn’t quite see Cayuga Lake) and the bottom right resolution is one used more by meteorological models than climate models.
            As we’ve mentioned before, a model by definition is a simplification. It’s not going to simulate reality, and you are going to have to make sacrifices. But if you are careful and you understand what parts of reality you’re ignoring and what parts of reality you’re including in your model, you are able to interpret your results and hopefully discover something new that wasn’t understood before. The history of weather and climate modeling is a wonderful history of practical limitations, amazing ingenuity and cleverness, and glorious tales of scientific advancement of our understanding or our Earth.