Have you noticed that weather forecasting has gotten much better in the last few years? Thanks to weather satellites, weather stations, and better forecasting techniques. How do scientists predict the weather with any kind of accuracy days or even weeks in the future.
We usually record Astronomy Cast every Friday at 12:00 pm PDT / 3:00 pm EDT/ 19:00 PM UTC . You can watch us live on here on AstronomyCast.com, or the AstronomyCast YouTube page.
If you would like to support Astronomy Cast, please visit our page at Patreon here – https://www.patreon.com/astronomycast. We greatly appreciate your support!
If you would like to join the Weekly Space Hangout Crew, visit their site here and sign up. They’re a great team who can help you join our online discussions!
This episode is sponsored by: Casper.
Transcription services provided by: GMR Transcription
Fraser Cain: Astronomy Cast Episode 462: Modeling the Weather. Welcome to Astronomy Cast, our weekly facts-based journey through the cosmos. We’ll help you understand not only what we know, but how we know what we know. My name is Fraser Cain. I’m the publisher of Universe Today, and with me is Dr. Pamela Gay, the Director of Technology and Citizen Science at the Astronomical Society of the Pacific and the director of CosmoQuest. Hey, Pamela, how’s it going?
Dr. Pamela Gay: It’s going well. How’s it going with you, Fraser?
Fraser Cain: Really well. We are about to do the second episode on what’s happening with the weather. Before we do, how is Image Detectives going?
Dr. Pamela Gay: It’s going well. We need more people doing it, though, because we have 1.5 million photos.
Fraser Cain: That’s a lot of photos. Those astronauts were super busy.
Dr. Pamela Gay: Yes, and they’ve been busy for a long time, so if you want to see some truly stunning photos, all of which you can download and turn into your screen backdrops to your heart’s delight, go to CosmoQuest.org/ImageDetective and help us figure out what the heck the astronauts were taking pretty pictures of, and you will help us be able to use the science to figure out how our planet is changing as it gets impacted by the weather.
Fraser Cain: And help science.
Dr. Pamela Gay: And the faster we get more images done. We’re currently working on building a catalog that will allow you to look things up. In much the same way that you’d look up to find, “Is there a hotel near this place?” we’re now gonna have it, “Is there an astronaut photo near this place?” and instead of filtering by pools, you can filter by, “Did anyone identify any lakes? Did anyone identify any volcanoes?” And we just need to have enough things in the software to thoroughly test the software, and we have – I wanna say it’s 330 images that have all been viewed by 15 people and thousands of images that have all been viewed by a person.
So, go in. Don’t skip too many images that you have half a chance of finding, and go through. And our goal is each image has its center identified by 6 people, or we give it to up to 15 people who don’t find the center, and once something has been viewed by 15 people, none of whom have been able to identify the center, or fewer than 6 of whom have been able to identify the center, we put it to the end of the list, and things will come back until we find the centers of everything or determine centers can’t be found.
Fraser Cain: So, I’ve got one quick thing. I’m gonna do a more formal announcement, but I wanna give the Astronomy Cast listeners a super-sneak-preview, advance-knowledge opportunity to make your reservation, which is that, as you may know, Dr. Paul Matt Sutter and I are going to be going to Iceland in February. That filled up in moments, so now we’ve got another trip planned. We’re gonna be going to the Caribbean for a cruise to see the stars and the ruins and also rockets.
So, we’re gonna be going in late September. We’re gonna be starting at Cape Canaveral. Of course, we’re gonna go see Kennedy Space Center, and then we’re gonna do observing every night. We’re gonna do a bunch of live shows. We’re gonna go see the Mayan ruins and some other places. And so, you can check out information on that at AstroTouring.com. And we’ve got a set number of slots, and when they’re gone, they’re gone. So, September 2018, if you’re interested in going on a vacation from the 20th to the 29th, come with Paul and I. We’ll have a lot of fun.
And I will do a more formal advertisement in the show, but this is for the folks who are interested. Go to AstroTouring.com. Just put down a $25.00 deposit, and you’re in. And then figure out the details later.
All right, let’s get on with this show. Have you noticed that weather forecasting has gotten much better in the last few years thanks to weather satellites, weather stations, and better forecasting techniques? How do scientists predict the weather with any kind of accuracy days or even weeks in the future? Pamela, how does this work?
Dr. Pamela Gay: They use a lot of math and big computers and do things very, very carefully.
Fraser Cain: Right. So, we talked last week about the weather satellite, and that’s really only a small portion of the weather monitoring system. So, before we get into the way they do their forecast and their model, let’s just talk briefly about what they’ve got that is capturing the data planet-wide.
Dr. Pamela Gay: So, we’re looking at a variety of different things that are gathered from weather stations on the planet, gathered from aircraft that are, in the case of hurricanes and other dramatic atmospheric events, going out and taking measurements, and then, of course, we have our suites of satellites in geostationary orbit, as well as the polar-orbiting satellites. And the kinds of data that are getting gathered include: How is the surface barometric pressure changing? How is the column density of the troposphere changing? This is measuring pressures higher up, effectively. What is the temperature structure of the ground, of the water, of the atmosphere? What is the wind? What is the cloud cover?
And by taking all of these different things into effect, we’ve started building more and more complicated computer models that take into effect more and more data at higher and higher resolutions to build all of our different models.
Fraser Cain: And so, you’ve got all of this data that’s pouring in. In some cases, they’ve got really good regular data that’s been gathered for hundreds of years, in some cases, depending on the kinds of weather stations that were recording some of this information. It’s actually really amazing. I know that ship captains, they’re pouring through old data from ship captains back in the 1700s, 1800s, and they would keep track the direction the wind is blowing, the weather that they’re seeing when they’re crossing the Pacific Ocean in exploration.
So, there’s all this great data that’s being pulled together and fed into computers. And then, what are the meteorologists – what’s the term? Weather forecast meteorologists?
Dr. Pamela Gay: Meteorologists.
Fraser Cain: The weather forecasters, the researchers doing with it?
Dr. Pamela Gay: It actually has been something that people have been really doing as what are called “citizen scientists,” which we’ve talked about just a few times on this show. They started during the American Revolutionary War, when the United States seceded from the British Empire. And it was our founding fathers who realized that it was national security, in some ways, to be able to predict the weather.
And so, they asked farmers to go out and look to the horizon and report what the sunrise looked like, what the sunset looked like, what is the cloud cover, and then to measure the temperature, to measure the barometric pressure. These are all things you can do with essentially no technology tools, just little pressure gauges and oil or mercury or alcohol thermometers. And all of this data has been getting collected continually. There are people still doing this, still part of the original project that it was Thomas Jefferson who formalized, and the data is going to the Smithsonian.
And that is where it started, and people started in trying to figure out, “Okay, what can we do with this temperature and barometric pressure data and looking at the sky?” And it was realized, “If we have a sudden drop in barometric pressure, that actually means that clouds are going to form, that rain is going to fall out of the sky and make you very wet,” whereas if you see the barometric pressure going up, that pressure increase is going to push away the clouds and actually leave you with a bright, sunny day.
So, this inner play started to give us a way to say, “If this, then that,” but it was a very simplistic “If this, then that” because it was impossible to know that there was actually this massive low-pressure front banding across the entire United States that was headed your way with massive cold temperatures on the other side, with tornados that would be generated. To get to that point, we needed the telegraph machine. And so, as technology has advanced, as the speed of our ability to report data has advanced, we’ve been able to get more and more advanced weather forecasting.
Fraser Cain: And so, you’ve got all this data. What do the researchers do with it just to understand the patterns of what they mean? You just mentioned anyone can do this. You get a barometer, which – by the way, get a weather station. Everyone should have a weather station. They’re so much fun. I did it as a kid, too. It was my job every day to track the rainfall, the temperatures, and what the barometer was doing, and we would store it in a little book, exactly what you’re talking about. I don’t know where the book went. It’s possible the garbage, but I like to think that it was sent to some Canadian data center.
But what are the scientists doing with this? How did they learn to be able to model the weather?
Dr. Pamela Gay: It was in the 20th century when people started to realize we can start to use fluid dynamics; we can start to use the detailed science of how gasses interact to start to put together numerical weather predications. This was work that was largely done by a Brit by the name of Lewis Fry Richardson who wrote a book that was very straightforwardly named Weather Prediction by Numerical Process.
And at this point, unfortunately, in the 1920s, we didn’t really have computers. And so, Richardson kind of imagined following the model that we saw used in astronomy where they had a whole bunch of women who were performing the work of calculators, and he imagined an entire auditorium filled with human beings who were sitting there, each doing their calculations, passing their paper to the next person who did their calculations. So, you essentially had human beings doing parallel processing, but there was no way to do it.
And so, it wasn’t until computers finally started to get built that we saw a whole group of people using the early ENIAC computers. And what I love is here you have Klára Dán von Neumann, who is a fairly small human, who did amazing software development, and the contradiction between the human and the data capabilities. I love it when powerful brains is contained in little tiny people.
So, you have the mathematician John von Neumann, you had meteorologist Ragnar Fjørtoft, you had Klára all working together in the ‘50s on early computers to start developing software. In that case, it was actually mostly hardware that could do a single calculation. They were programming electronics to take all of this data that humans can’t, by themselves, deal with, and build models that were coarse, that treated smaller countries as one grid on the calculation but that were better than anything we’d previously had.
Fraser Cain: And one of the great things about having all this data is that you can essentially run a prediction historically to see if what you thought was going to happen did actually happen, this sort of regression testing, where you can just go again and again to see how well your model predicts the future. You start by making your model predict the past, and from there, you can then try and have your model predict the future.
Dr. Pamela Gay: And this is actually part of how the different nations decide whether or not to upgrade their data models. Scientists are always trying to find new ways to remove the assumptions, to remove the “this is based on brute force rather than underlying equations” out of our software.
So, as they’re trying to figure out hurricane models, as they’re trying to figure out storm forecasts, they look at all the historic weather models, and they say, “If I feed my software everything up to the Sunday before, is it able to figure out what happens one day later, three days later, ten days later?” And we can’t really go beyond ten days. We’re still not there. And if the new computer models can consistently perform better than the prior computer models, it’s like, “Okay, everyone update your software. Go from Windows 95 to Windows” – I guess it was 3.1 to 95. Don’t do the reverse.
And this, sometimes, leads to controversy. This happened earlier this year when there was the question with the US National Weather Service on whether or not they should upgrade their models to new software or not. And what I find interesting as a complete outsider to this field – and I love doing this episode because it gave me a chance to look up and read about all of this different way of looking at our planet instead of other planets.
There was controversy because the new models don’t always perform better, so the question is, “Do you want to maximize your model to be better at predicting tornadoes but not as good at hurricanes? Do you want to maximize it at being able to get rainfall amounts correct, or do you want to maximize it at getting wind speeds correct?” And so, there’s so many different ways of saying, “This is the best model,” what do you prioritize? And so, they did make the update, but there’s a lot of people who are resolutely like, “That wasn’t a good choice.”
And so, you quite often see, “We’re just gonna show you all the models, all of them.” So, lately, with the weather tracks for hurricanes, we’ve simply been seeing, as consumers of weather content, the meteorologists putting out all of the results from all of the models and then explaining that this cone we have of prediction for weather tracks is actually based on integrating across all these different models and trying to figure out what is the most likely scenario.
Fraser Cain: Right. And this is where we put meteorologists into a cage match and make them fight to see whose weather model is the most accurate. I believe that’s how it works.
Dr. Pamela Gay: Pretty much. It’s more their software’s doing the cage match, and it’s a question of whose assumptions were the most accurate assumptions because, unfortunately, we don’t have complete data, and so you have to make certain amounts of assumptions.
Fraser Cain: Hey, podcast listeners, Fraser here. Once again, it’s time to thank Casper Mattresses for their generous and ongoing support of Astronomy Cast. For those of you who don’t know, I’ve got two Caspers here in the house, one that’s just like the main bed, and then one we’ve got in the spare bedroom, and, man, I look forward to sleeping in my Casper mattress every night. And I’ve had it now for years, so I’ve gone through summer and winter, and I know it works through all of those seasons.
Probably one of the coolest parts about the Casper mattress is how it must bend and break the laws of physics, as they send you this box that’s way too small to contain an entire mattress. And then, you open up the box, and this mattress unfolds, and it’s pretty amazing.
At the same time, if you live in Canada and the US, you can get a Casper mattress delivered to you for free shipping, and then you can try it out for 100 nights. And then, if it doesn’t work, you can send it back to them, and they will take it away, and you don’t have to worry about it, but I’m sure you’re gonna wanna hang onto it. Now, they call it their “You can be sure of your purchase with Casper’s 100 night risk-free sleep-on-it trial.” And like I said, I’m now closing in on two years, so I definitely like the Casper mattress.
And Casper has a special deal for you as a fan of Astronomy Cast. If you go to Casper.com/Astro and use the promo code ASTRO, you can get $50.00 off the purchase of your own mattress. Terms and conditions do apply. So, give it a shot. Go to Casper.com/Astro, use the promo code ASTRO, and you can try your own Casper mattress.
Thanks again to Casper for supporting Astronomy Cast.
I feel like with the modern age, with all of the sensors that are coming online, with all of the satellites that are coming online, these models are becoming really fine, nuanced versions of one another at this point. Like the broad strokes is understood fairly well, and now we’re able to do things like, “Is it gonna be rainy here on Vancouver Island in two weeks, or is it gonna be partly sunny?” I do find myself starting to make plans around the weather forecast in ways that I never used to.
Dr. Pamela Gay: And what’s amazing about this is it’s allowing us to understand more and more. As our models get better, it requires us to be better about our data. So, we have to take into consideration that cities are heat traps, that as a storm sweeps across the landscape, how it behaves changes whether or not it’s going over that nice cool cornfield or it’s going over a big asphalt, hot, trap-the-heat-and-reradiate-it pavement.
So, our models are now telling us we, as humans, are impacting the models through urban sprawl. And so, our models have to take into consideration the fact that that new suburb got built.
Fraser Cain: Yes, and this is where I think the technique is really getting finely honed is you’ve got these situations, as you said. You’ve got these human-built structures that are actually changing the weather, the croplands, the heat islands of cities. If it feels like it’s hot in the city, it literally is hot in the city, and as soon as you head out into the forest, it’s fine, I promise you. And so, it’s that mixed with – and then we’ve got this other situation, which are these long-term climate changes that are happening, both natural and human-caused.
Dr. Pamela Gay: And are actually mucking up our simulations to a certain degree because the ocean temperatures are dramatically changing, and with these changing ocean currents, as well, as we add more and more fresh water into the oceans, as we drop in new icebergs on a regular basis, this is changing the interplay between the ocean and the atmosphere. As we remove forests, it’s changing the interplay between the landscape and the atmosphere.
And so, when we have a model that is focused on, “I’m going to predict Europe, taking data from these places, and I’m going to ignore what’s going on in Australia,” that might be a mistake because it is a global atmosphere, and what is happening in Australia is related to the Pacific Ocean, is related to what’s happening in South America, is related to China, to India. It’s all one story, and if your model says, “I’m going to take in these inputs, which I consider good enough, but I’m not actually bothering to calculate with everything because processing time, and I want the model to finish someday,” we get inaccuracies because we can no longer assume the Pacific Ocean works this way, the Atlantic Ocean works this way.
There is a hurricane currently barreling towards Ireland as we record this. That is not normal.
Fraser Cain: Right, but what is in this day of crazy climate weirdness? And so, just when the scientists thought they had these data models figured out and could tell me whether or not it was gonna be rainy or partly cloudy in two weeks, now the long-term changes are causing more instabilities. Changes in rainfall patterns, changes in temperature, both up and down, increased hurricanes, increased tornadoes. And it’s those things, those hurricanes, for example, where people are keenly interested to know if there’s gonna be a hurricane, where that hurricane is going to hit.
Dr. Pamela Gay: And we used to be pretty accurate at knowing how strong the storms would be. I remember watching Katrina and Rita and all of those hurricanes back in 2005, and our understanding of just how badly the Louisiana coastline, the Texas coastline were gonna get creamed, it was pretty good, whereas looking at things this summer, I need to go back and do a one-for-one comparison, but it feels like the storms are suddenly getting much stronger than we predicted. I’m seeing this with Ophelia. Ophelia was never predicted to get to Category 2, and then it was a Category 2. It wasn’t predicted to stay a Category 2. It stayed a Category 2. And so, we’re seeing intensification that we haven’t seen before.
Now, what this could be is we have models that work perfectly well based on a set of assumptions. So, this is the problem of you’re going up the stairs and, all of a sudden, one of the stairs is two inches taller than your leg anticipated, and you fall on your face. We know how stairs work, but because we had an assumption, the model falls on its face. And so, we have to figure out which assumptions do we change so that we can still calculate the models and they’re accurate for our new normal.
Fraser Cain: Yeah, and this is why the scientists are saying, say, in terms of global warming, “The temperatures are gonna rise this amount by this period,” while others, based on different models, different assumptions are saying they’re gonna rise by a different amount. And really, each year, everyone is looking at how well their model predicted what actually happened. Did the sea ice drop again? Did glacier melts cause? Did the methane vents at the northern latitudes cause a problem? Each of these issues is making the job of climate prediction so much tougher.
Dr. Pamela Gay: And volcanoes, don’t forget the volcanoes.
Fraser Cain: Sure, absolutely, just the normal variations, what’s happening with the sun, what’s happening with the volcanoes, what’s happening with, as you say, the heat islands, things like that. It’s a very complicated job. Does just really powerful computers make the problem go away?
Dr. Pamela Gay: It doesn’t make it go away, but it starts to make it so that fewer assumptions are required, so that we can do a finer-grid mesh of calculations. Where it starts to get messy is just what all do we have to include. How salty the North Atlantic is actually makes a difference in understanding the currents and understanding the heat capacity. And as these things change, which we’re not used to these things changing, it changes our models.
And it’s important to remember that, with climate change, it’s not just everyone’s going to get warmer. It’s actually the weather is going to get more extreme. So, everything that we’re experiencing this year is completely consistent with the models that scared the bejesus out of me in the early 2000s. And I wish that was a lie. There was literally a Friday night where I’m like, “Have you seen these journal article papers?!”
And the issue is that, as we end up with more energy stored in the ocean at tropic latitudes, you have these storms that spin up much stronger. As we end up with difference in how the jet stream flows and how the tides dissipate that energy in the equatorial regions, we see that the northern- and southern-most latitudes actually get colder. So, we’re going to have more extreme winters in the Canadas of the world. And I live right at the crux of all of it, so I’m gonna get more tornadoes.
So, this is what climate change is. It’s a global on-average increase, which we’ll see as a significant warming in equatorial regions and more extreme weather, which includes more extreme cold during the winters, in the northern and southern latitudes.
Fraser Cain: So, once again, I give you a blank check here to fund your weather forecasting. What would you like to see that money spent on?
Dr. Pamela Gay: A think tank with a set of supercomputers dedicated to running models. Imagine if you could get the best meteorologists from around the world working in their own Max Planck Institute kind of place with unlimited computing technology, with that mesh of satellites that we discussed in the last episode, as well as full access to all of the planet’s weather stations. I’d want to see those people with the best computer scientists in the world working to advance the models, advance the software, and find ways to get fewer assumptions and more calculations going into how we see into the future.
Fraser Cain: I’ll see what I can do.
Dr. Pamela Gay: Thank you.
Fraser Cain: Thanks, Pamela.
Dr. Pamela Gay: My pleasure.
Announcer: Thanks for listening to Astronomy Cast, a nonprofit resource provided by Astrosphere New Media Association, Fraser Cain, and Dr. Pamela Gay. You can find show notes and transcripts for every episode at AstronomyCast.com. You can email us at Info@AstronomyCast.com, tweet us @AstronomyCast, like us on Facebook, or circle us on Google+.
We record our show live on YouTube every Friday at 1:30 p.m. Pacific, 4:30 p.m. Eastern, or 2030 GMT. If you miss the live event, you can always catch up over at CosmoQuest.org or on our YouTube page.
To subscribe to the show, point your podcatching software at AstronomyCast.com/podcast.xml, or subscribe directly from iTunes. If you would like to listen to the full unedited episode, including the live viewers’ questions and answers, you can subscribe to AstronomyCast.com/Feed/FullRaw. Our music is provided by Travis Earl, and the show is edited by Chad Weber.
This episode of Astronomy Cast was made possible thanks to donations by people like you. Please give by going to Patreon.com/AstronomyCast.
[End of Audio]
Duration: 29 minutes