A lot of the work that climate scientists, especially observational ones like myself, do is that we look for patterns. Within any climate phenomenon there is going to be some portion of it that is just responding to forcings. That'll be the part that we can predict. There'll be some kind of regular steady beat. But on top of that, there can be randomness. It can be really, really difficult to separate the signal from the noise. But pattern recognition is something that humans have been doing for a really long time. And we're so good at it, that sometimes we find patterns even when there are none there. This has been shown in studies where people are given pages of random x's and o's. Most of them will find the pattern, even if there is no pattern. You see this a lot in people's superstitions, especially around sports. Players who don't change their socks during games. What they're doing is they're taking a pattern that they saw and they're attributing to it some kind of cause and effect. Now they may not actually believe that there's cause and effect, but they figure what's the harm? And they're right, there really isn't any harm. And that's because life punishes false negatives, differently then false positives. Let's think about it from an evolutionary viewpoint. Imagine that you're an early human who hears a loud noise. When you hear that noise, you immediately run and hide because it might be a tiger. Well, let's say you're right. It was a tiger. Well, you ran and hid. So that meant you couldn't spend that time doing anything else. So there's a small cost to it, but not much. And you lived. That's great. Now imagine that you hear that noise. You think it's a tiger, but you're wrong. That's a false positive. In that case, again, there's a small cost to you but not much. You wasted a little bit of time, a little bit of energy. Your heart rate went up unnecessarily. Not a big punishment. Not a big cost. Now imagine a mistake in the other direction. You hear a sound and you think, nah, that's not a tiger. If you're right no harm done, but if you're wrong game over. That's why humans are so good at finding patterns and identifying patterns that we sometimes do it when there aren't any patterns. Climate change is all about looking at patterns on multiple scales of time. Think about something as simple as the tide. If you go to the beach, you're probably pretty aware that there's a 12 hour tidal cycle, 12 hours between one high tide and the next. But there's also a longer cycle than that. There's a daily cycle. There's a fortnightly cycle, that's every two weeks, and there are even longer cycles on top of that. So what we see isn't one pattern, it's the combination of many patterns. Finding patterns isn't just something we do in climate. One of our methods of pattern finding called empirical orthogonal functions or principal component analysis, is actually something that happens in a wide variety of fields. Most of us keep our music files as mp3's, you might have heard that that's a lossy format, it's not perfect. We have the same issue with our photos. We mostly save them as JPEGs. Why? File size. You can fit a whole lot more MP3s or JPEGs than you can the full size files. But how do they do it? Well, it's the same technique that we use in climate. What they do is, for example, imagine a picture of a rose. You could record the exact color of every pixel in that picture. That's more or less what a bitmap is and that takes up a lot of space. But, it's a picture of a rose, most of the pixels are closely related to each other. So, the first piece of information you might want is, well, on average the pixels are kind of red. And then you could add on top of it, well, there's kind of a circular pattern. It turns out, you don't need all the information in the picture to create a reproduction of the picture that will be so close that the human eye can't tell the difference. You only need the first couple of modes of variability, the first couple of layers of information. We can do that scientifically, too. What you see here is salinity, how salty the water is, in the area around Indonesia. I wanted to find out what was making the salinity vary. My eye isn't good enough to see some of the really complex patterns because most of the patterns have to do with large scale forcings like the time of year. So the first few modes here, those are regular patterns. They might be interesting if I were studying those things, but for the purpose of my research I ignored them to get to that third mode. That showed real variability, variability that I couldn't immediately attribute to the sun or the season. That's where the signal that I wanted was hiding. So using these techniques we can decompose really complicated pictures, whether they're photos or data from around the world and we can tease out from that what the actual pattern is. It's not always intuitive and it's not always obvious. But finding the patterns that are actually happening, can allow us to make predictions about what's going to happen in the future. When we look at patterns within climate it's easy to focus on the patterns that we can see in our lives. The pattern of day to night. The pattern of years. But it's also worth considering longer scale patterns. As it turns out, the Earth's orbit around the sun isn't always exactly the same. It's similar, but not the same. You know that the Earth orbits the sun in a circle. That's another one of those things that isn't completely the case. It actually orbits in an ellipse, which is a circle that's been squished a little bit out of shape. How much out of shape? Just a tiny bit. But the amount of that tiny bit varies, and it varies on a 100,000 year cycle. The tilt of the Earth also varies in time, on about a 41,000 year cycle. Sometimes we have the tilt we have now. Sometimes we have less tilt which means that the seasons aren't as exaggerated and sometimes we have more tilt which means that the seasons are even more exaggerated. The third thing we need to consider is the wobble of Earth's orbit. Right now the North Pole points to the star that we've named the North Star. But if you could travel back into the past, the North Star would appear to move. Of course, it's not the star that's moving. It's that the North Pole of the Earth points in different directions throughout time. That varies on a 23,000 year cycle. All of these cycles are happening simultaneously. That means that sometimes the Earth gets a little more sunlight, sometimes it gets a little less. Sometimes the winters are extra cold, and the summers extra hot. And sometimes things sort of level out. The math on all of this isn't particularly difficult, but it is tedious. Amazingly it was worked out during World War I by a Serbian mathematician and scientist names Milankovic. He was a prisoner of war and spent his years in prison making these calculations over and over again. It wasn't until after his death that he was shown to be really close to perfectly accurate. What he figured out was that there were times when those forcings of the Milankovic cycles would line up. There are going to be times when just randomly the Earth gets an extra dose of sunlight. Specifically, an extra dose of sunlight during winter. When we think about glacial and inter-glacial cycles, we're particularly concerned with what happens just at the edge of the glaciers. Over the summer will it get hot enough to melt all of the ice? If so, then ice building in the winter is going to be starting from scratch. It's hard to build a glacier if you only have a few months of the year to do it. But imagine that there are times when it's cool enough, even in the summer, that ice could last. Now each winter as the ice sheet is growing, it grows on a base of ice that's already there. Each year you grow your ice sheet bigger and bigger, and that's how you get into an ice age. These forcings are small. The change in the amount of sunlight you get is very slight, but, again, a small forcing can have a big effect on a system if there are feedback loops. In this case, the feedbacks take what would otherwise be a very small signal and turn it into a global ice age or a global torrid age in which there isn't any ice. Why do climate scientists spend so much time looking for patterns and trying to find explanations for them? Well, in our modern time we're really concerned about global warming. But you might think, hey, it was warmer in the past, why should I worry? Actually, it's because it was warmer in the past that people like me are worried. These climate patterns happen because little forcings can have big results on the system. What we're doing now is we're changing those forcings. So, if we just push the climate system a little bit, we might think that not too much is gonna happen, but what we've seen in the past is that little pushes have big results. The climate system doesn't care if it's being pushed by a Milankovic cycle, or by me driving a car. From its point of view, all it sees is a push. Positive feedback can turn that small forcing into a really big effect. We now know that climate is extremely sensitive to these small forcings. So the fact the climate has been unstable in the past, isn't a reason not to worry about it now. It's a reason to worry about it more. We see that climate responds in these really sometimes abrupt ways to forcings and sometimes in unpredictable ways. It's true that in the past it was so warm that there were crocodiles at the North Pole. I think that's really cool but I wouldn't want to live in that world. If there are crocodiles in the North Pole, Manhattan would have to be underwater, and that's my home. So it's not about what's natural or what isn't natural, it's about pushing the system out of its usual set of patterns into a pattern that might work for some animals, but it's going to be really hard on humans.
