In Unit 16, we’ll talk about Weather Observation and Prediction Technology. Our objectives here are to understand the spectrum of weather technology that is currently available, understand the benefits and limitations of each type of weather instrument, identify the primary instrument(s) used to observe different types of hazardous weather phenomena, and all of this will also lead to our discussion of the benefits and challenges of computer model projections, we’ll place the technological evolution of meteorology in historical context and understand the important of international and global cooperation in setting data collection and dissemination standards. At the end of this unit will be an activity where we will essentially do a “nowcast”, just to see what type of weather is going on around the country right now. We have a lot of different types of Weather Technologies: the Observation Instruments include, they typically fall into two categories: There are In-Situ Instruments. These are the ones that are deployed on location and actually observe weather conditions at the site of the instrument. Then you have remote sensing instruments. These are the ones that essentially can sense what’s going with the weather from a distance. Those would include satellites and radar. In-Situ Instruments include weather balloons, a simple thermometer, barometer, technology that you’ll find at a weather station. As well as aircraft reconnaissance, when they’re dropping instrument packages into a storm, that’s considered In-Situ Instruments. Numerical Weather Prediction (NWP) is essentially a technical term for computer models. We’ll talk about why those are important later on in this module. There’s something called the Automated Surface Observing System or ASOS and these are weather stations deployed across the United States, mostly found at airports, that will document the changes in weather variables such as temperature, pressure, humidity, and when we take this data, meteorologists can interpret a chart where all of this information can be plotted in what’s called a station plot. And these include different ways to decode different symbols that represent wind direction, wind speed, what type of weather, visibility, pressure change, temperature dew point. All of that can be plotted onto one dot when you look at a weather map. The benefit of that is when you plot all the station plots you can start connecting a lot of the lines and you see lines of constant pressure here in brown showing the, what looks like a topography map. Showing the highs and lows of the pressure systems as well as the, the frontal zones. When we put this into motion, it’s easy to see how we’re able to use this as a weather forecasting tool in order to see the recent changes in the weather systems, and extrapolate into the future. This is a very rudimentary way of understanding where weather systems may go without looking at a computer model, but all forecasts have to start here. You have to know as a meteorologist what the weather has been doing, what it’s doing right now, in order to understand future changes. But we all know weather doesn’t just happen at the ground. Weather Balloons are a mid-20th century technology that we’re still using. Filled with helium or hydrogen and then launched with a radiosonde instrument package tied to it. And as the as the balloon goes up, it returns a thermodynamic diagram such as this one. We won’t go into the details of this obviously, because this is, this has given meteorology students nightmares in the past, but this figure essentially shows in the red line the temperature changes as the balloon goes up in the atmosphere and the green line shows the changes in moisture and the top right corner, you see the wind shear pattern so just looking at this one figure, we can derive a whole lot of information about the state of the atmosphere. Now unlike ASOS stations that occur at various airports around the country, we see the Upper-Air Sounding Network where weather balloons are launched, is a lot sparser. You see that it’s a lot more sparse. It’s, it’s, it’s a, there are a lot more gaps in between these stations. A drawback to this of course is we don’t have as much data, right. So if you look here on this map, So, if you look here on this map, Nebraska on this given day had, had one launch on the eastern side of the state, but that’s it. The rest of the state you have to interpolate from the surrounding data collection points for weather balloons. But these are launched twice a day and coordinated around the world, so that’s why we don’t have an unlimited budget to launch these weather balloons and so we have to deal with a sparser network. But it still gives us a decent understanding of what goes on in the upper atmosphere. “Here you’re showing, hey, here this figure shows what the jet stream looks like.” Which is a fast area of quickly moving air from west to east, showing the, at the air, at the level that aircraft are typically flying. So this is something that weather balloons can tell us. Aircraft reconnaissance is very popular, as a mechanism to help us better understand the inner workings of a tropical cyclone. The Air Force Reserves have a fleet of aircraft that will be flown into tropical cyclones as they approach the United States and they’ll launch dropsondes. These are attached to what looks like a parachute and as it comes down, it’ll measure similar variables as what a weather balloon would measure going up the dropsondes, of course, are going down from the plane to the ocean surface. NOAA also has Hurricane Hunters and these are similar, they’re different aircraft, but they have a similar mission, which is to go in and understand the inner structure of a tropical cyclone to get an observation of the maximum sustained wind speed as well as to try to locate to fix the center of the tropical cyclone as well as it’s lowest, it’s area of lowest pressure. This is an example of what the eye of a tropical cyclone looks like from inside. It’s really nature’s football stadium. The thermodynamically the eye wall actually slopes outward as this, as you go up inside a tropical cyclone and this is something that very few people have seen, but those who fly in Hurricane Hunters will see them quite frequently. Other types of aircraft reconnaissance include the NOAA Gulfstream IV aircraft which is a high altitude jet that will fly around a hurricane in order to get the, a better understanding of the surrounding conditions that can govern not just intensity, but also the track changes of a tropical cyclone. And some of that data can be ingested into our computer models to better understand where the storm may go. But there are two; there are two primary types of satellites. The Geostationary Satellites are the ones that are located farther away from Earth, but they are fixed right over the equator. And so as the Earth rotates, the Geostationary satellite will actually look, be placed at the same point over the Earth. And so it’s able to record images of the changes in the clouds over the same location. However, there are drawbacks and that is it’s harder to see the poles because the polar regions are at a higher angle and as you can see here in this, in this cone. So Polar orbiting satellites are necessary to help us see the northern and southern parts of the Earth. However, their drawback of course is you’re not looking at the same spot on the Earth all the time and so you may only pass over the same area twice a day. So that doesn’t really help you very much in terms of creating satellite loops or animations showing the changing of the clouds. So there are there are pros and cons to both systems but we need both to enable better forecasting. Here’s an example of a geostationary satellite animation, showing the development of Hurricane Sandy over the Central Caribbean Sea. And as it moves northward just east of Florida, it gets caught up by a cold front by a cold front and it undergoes extra tropical transition and becomes a post tropical cyclone. Here it is again: a replay of this animation and you see cluster of tropical thunderstorms growing in size and intensity and then as it encounters the cold front, moving in across the eastern United States, it gets absorbed with a mid-latitude cyclone regime. So this is an example of Earth’s way of, of transitioning from the, the heat energy from the tropics over to the mid-latitude and polar regions. Doppler Radar is one of the more important tools that’s used by meteorologists and we’ve seen a lot of these at airports around the country. And essentially, it’s an antenna mounted on a pedestal that emits microwave radiation. And as that radiation hits a particle and part of the energy scatters back, we’re able to estimate the amount and intensity of the particles as they fall from a cloud. In recent years, an advanced type of radar called Dual Polarized Radar, or Dual Pole, is, has been used at National Weather Service sites to get a better understanding of the cross-section of what’s falling from the sky. And in a situation like this, with this technology, we’re able to emit, not just a horizontal beam of radiation, but also a vertical. And so the computer processor can get an understanding of the cross-section of the particle. If it’s spherical, then it’s probably a hail particle. If it is flat like a pancake, it’s probably a raindrop. If it’s irregularly shaped, it could be a snowflake. And so dual polarized radars have helped us with understanding the type of precipitation that’s falling from a cloud, which enables us to which enables us to make better flood estimates, to make better forecasts for or nowcasts for what’s actually falling from the sky. The radar network across the country can be seen here. Immediately, you can see that there are limitations to the radar network, right. Even though, radar is such an important tool. They don’t, it doesn’t exist everywhere. Coverage is not universal. Particularly in areas with rough terrain, such as the Rocky Mountain West. We actually see a lot of gaps. Areas with less population will also have more gaps, but there’s another limitation. And if you notice on the colors here, the cream color areas show the scope of the radar being approximately 4,000 feet above ground level And as you go farther away from the radar site indicated by the black dots, in the orange, you see that’s about 6,000 feet above ground level And then the farther you go, you’re talking 10,000 feet. So we all know that what people care about in terms of weather impact is what So we all know that what people care about in terms of weather impact is what happens at the ground. And so, the farther you are from a radar site, the less likely you are seeing weather that’s near the ground, right, because radar doesn’t tell you what’s going on between the beam and the ground. And because of this, one of the primary limitations of radar is that the farther you are away from the site, the less you understand about what’s actually falling. And this is an important limitation to consider because nowadays we have a lot of weather radar apps that we can all get on our smartphones But many of us probably do not know that just looking at a radar map that what the colors don’t actually represent what’s actually hitting what the colors don’t actually represent what’s actually hitting the ground because it depends on your location away from the site. Here’s an example of a radar image taken from KTLX, which is the radar site in Oklahoma City. And this is the Moore Oklahoma Tornado of 2013. On the left, you see the radar reflectivity diagram showing the intensity of the particles that’s scattering back the radar energy and you see the hook echo of the supercell thunderstorm where the rain and hail is actually wrapping around the backside of the storm as it’s rotating counterclockwise And then right in the middle, you see a debris ball where the radar is actually picking up signs of debris from the tornado itself. On the right, you see a radar velocity image, just like a traffic radar, weather radar can only detect the velocity of objects as they approach and recede away from the radar site. So what that means is, objects that are moving perpendicular to the radar beam can’t be estimated in terms of velocity only the component of the velocity toward and away from the radar is detectable. So here, we see the green color showing air or particles blowing toward the radar and the red particles, and the red area shows air moving away from the radar. And right at that interface, of the bright red and the bright green is where we can infer counterclockwise circulation. And then this case would be a tornado. So, we’ve discussed the different types of weather technologies. So let’s, let’s think about the, the different types of forecast that we can make based on the data that comes in. So when we look at all that we’ve discussed so far, all of these instruments give us an idea of the temperature, pressure, density, moisture, all of this information needs to be processed somehow, we need to make some sense of it in order to predict the future. So, how do we do that? Well, a lot of math. Through Multivariate Calculus, we are able to come up with relationships between all these variables in order to attempt to predict the future of weather and this is why it is so difficult to process sometimes because no human can possibly make all of the calculations by him or herself. There were two pioneers, throughout history, we have a Norwegian physicist and then a British mathematician, Dr. Richardson. These two gentlemen pioneered essentially, what we can discuss as Numerical Weather Prediction nowadays. They came up with the idea that we can use mathematical relationships to try to figure out the future of the weather. Dr. Richardson in particular, he served during World War I, as an ambulance medic and as he was behind the frontlines, he spent about six weeks attempting a manual calculation of pressure to come up with a pressure forecast. It took him several weeks, to get a forecast, but it was of course incorrect by hand. So he came up with a thought problem: He said that in order for humans to be able to keep up with the weather, we have to have 64,000 people sit in a large room, an arena like room with a conductor in the middle and giving, passing instructions along to the 64,000 people So that each of them when they make a calculation, they pass the result to the next person. And then to the next person. And by doing so, you can keep up with the weather as it occurs. Thankfully, this experiment was never done. It would not be possible. But over the years, this experiment eventually became reflected in our modern day computer modeling. So, using a supercomputer, we can create a similar situation, where numbers are being calculated in paths from grid space to grid space. And so when the Earth is divided up into grid, we’re able to make sense of those Multivariate Calculus equations that we saw on the previous slide. Computer models follow a Forecast Cycle. Just because the weather is always changing the computer models have to keep up with that by issuing forecasts and updating it every few hours. And depending on the computer model, it could be hourly updates; it could be 6-hourly updates. A Global, the Global Forecast System, which is the Premier United States Global Forecast Model operates on a 6-hourly basis, in terms of what the operational forecasters have access to. The model begins with a Background “First Guess”. What that means is, it takes the previous run of the model, say 6 hours before, it takes it’s own forecast, for now, from 6 hours before and maps it onto the current data. And that process is called Data Assimilation. When the current data is brought into the model, it compares it to it’s forecast for now and analyzes it to make sure, to basically, it’s their, it’s, it’s the way of understanding what’s going on right now in the atmosphere. Then we apply the equations, to that information and then integrate it into the future. So the model comes up with a prediction. At that point, the numbers are Post-Processed into maps that humans can understand. This cycle is better illustrated perhaps using this figure here: the red line shows the actual observation. It could be temperature; it could be pressure, whatever variable you would like. Let’s say temperature. So as the Sun sets, you can see that the temperature goes down and then as it goes up the next day, you see that the temperature follows a cycle. So the forecaster’s task and what the computer model is attempting to do, is to understand and to predict this change So the blue lines show the attempted forecasts, right. So as time goes on into the future, you notice the blue lines will diverge from the red line and that indicates that forecast error will increase and that indicates that forecast error will increase into the future because the farther you get into the future by common sense, the less accurate the forecast becomes. So the forecast cycle is such that every few hours, whenever there’s a new computer model run, whenever a new forecast needs to be issued, the forecaster attempts to adjust, it nudges that forecast back to reality, back to the red line. Or you notice how it’s never quite exact, there’s a small gap between the blue line and the red line and that’s because there’s sampling error. We don’t know exactly what the atmosphere is doing at any given point. And so as that error is adjusted for, we attempt to make a new forecast. It can never be zero. We can never understand the state of the atmosphere perfectly and because of that, the errors will always grow. This is the, this is the essence of Chaos Theory. I think we’ve all heard of “The Butterfly Effect”. Ed Lorenz, MIT Professor, pioneered this idea, saying that weather is the perfect example of chaos and what that means is that, weather forecasting is inherently non-linear, there are so many different factors going into it, we can never say that A causes B exactly et cetera, et cetera. However, slight differences in initial conditions can result in widely different forecasts, even though the system is deterministic. What that means is, weather is predictable. And if we were to understand the initial snapshot perfectly in theory, we should be able to come up with a perfect forecast. But we never, we can never do that; we don’t know initial conditions perfectly. The atmosphere is not random, But there is still a limit to predictability governed by how well we know what’s going on right now. So, these fundamental principles must be considered in risk analysis when we talk about understanding risk and how the uncertainty of a weather forecast can play into that decision making process. We have to remember that there are fundamental limitations to weather forecasting. We’ve done an excellent job of improving forecast accuracy over the years, over the past few decades. But there will always be an inherent limit to our limit to our ability to predict. Here’s an example of Ensemble Forecasting. This is computer model simulation of, or prediction of where it thinks the jet stream will be 16 days from now. And you see the days counting down, as the event gets closer. Notice how within about a week, you see there are two dips, two troughs in a jet stream, one on the West Coast, one on the East Coast, those reflect storms. sites of mid-latitude cyclones. But notice how two weeks before that, all the computer models had really no consensus. There’s no, there’s no commonality between them. It isn’t until about a week in advance, do they begin to suggest that there might be something happening on the West and East Coast. And it isn’t until 4 or 5 days before hand, does it lock into place and finally all the models agree. So this is an excellent way to show how an Ensemble of Models can help us understand the level of uncertainty with an event. So when they’re going crazy like a spaghetti diagram, at that point, there’s very little certainty that a weather event would occur. But, within about 5 days, we start to get a better idea that something may occur on the West and East Coast in this situation. Of course, there are, are many different types of numerical models, you have operational vs. research. The operational ones are the ones used by forecasters. On a regular basis, they’ve been highly tested to make, to ensure consistency. Research models are always being constantly tweaked, behind the scenes, by researchers, to try to better improve the operational models. There are of course global models vs. regional models, you have weather vs. climate models. Lots of different purposes and they’re all made with physics packages. A lot of the models are managed out the, out of our, National Center for Weather and Climate Prediction in College Park, Maryland, but the computing centers are of course, actually located in other locations. Here’s an example of the GFS Surface Pressure and Precipitation Output. And you see how smooth the lines look. This is the computer models way of being able to predict the future of movement of highs and low-pressure systems across North America. It can also do the same for jet streams. So you see this wavy pattern around the level at which jet aircraft fly. And understanding how the wave pattern changes can help us better understand the, the physics of the atmosphere and whether or not there’ll be a chance of severe weather or winter weather or any type of hazardous weather in the days into the future. So looking ahead, technology of the near future, there’s more and more emphasis now on an “Earth-System” Approach to computer modeling. So rather than just looking at a meteorology model or an atmospheric model, we want to include ocean currents, effects from land, from the cryosphere or from, from the ice in, in the polar regions. The Dual Pole Doppler radar upgrades which we discussed earlier, have recently been deployed across the United States via the National Weather Service. So that’s something that is beginning to help us tremendously in terms of understanding the type of precipitation that’s falling from the sky. NOAA has been investing in the Warn-on-Forecast Program and what that means is an effort to try to integrate short term modeling, high resolution models into making tornado warnings so that the hope is someday we may be able to extend warning lead times to beyond just minutes to about an hour or so. Of course, this gets into other social science concerns, such as would people take action if they’re given more lead-time. These are things that are being discussed at great lengths right now as our research continues into this area. There are different crowdsourcing mechanisms being in development nowadays. There’s an app called NOAA mPING, which allows users to download onto their phones. And whenever it starts snowing or sleeting or freezing raining, you can, you can insert that information into the app and then NOAA is able to take that information and integrate it into a database which can help us fine tune and adjust computer models to make sure that we’re representing precipitation type properly because because again as we discussed earlier radar can’t really right at the ground. But people are right at the ground. So how do we use people to help us collect data on weather events? That can be very helpful. So we conclude this unit with a, with a “Nowcasting” Activity, where we’ll actually go through and investigate some of the current websites for official sources of weather information to see, to get a good snapshot of what’s going on across the country.