 Good evening everybody. My name is Yaga Richter. I'm a climate scientist at NCAR and the lead organizer of the NCAR Explorer Series. It is my pleasure to welcome you today to NCAR's Mesa Laboratory and a second lecture in the NCAR Explorer Series. NCAR, or the National Center for Atmospheric Research, is a world-leading organization dedicated to the study of the atmosphere, the earth system, and the sun. The atmospheric and related sciences are of critical importance to society, especially in our rapidly changing environment. At NCAR, we recognize the importance of communicating the findings of our research that directly influence the world and the people around us. Our scientists and engineers in the NCAR Explorer Series present their research findings to you, the public, at a local level, and we extend the reach of these talks to a global audience through our live web stream and archive presentations on NCAR Connect. Thank you for taking the time this evening to join us and we look forward to answering your questions and hearing your comments and the Q&A directly following the lecture. Today's speaker is Dr. Bruce Carmichael. Dr. Carmichael studied mathematics as an undergraduate at the University of New Mexico. He received a master's in applied mathematics from Northwestern University and a PhD in computer science from University of Maryland. After completing his PhD, Dr. Carmichael started working for a Washington-based consulting company. Two years later, his wife encouraged him to get a hobby and Dr. Carmichael started taking flying lessons from the Baltimore Washington Airport. He proceeded to get a pilot's license and has flown for fun, never for pay, with his wife and children ever since. Dr. Carmichael's professional career over time evolved towards the aviation industry. He worked for a naval air test center building software and then continued as a software engineer for a company that worked with the Federal Aviation Administration or the FAA. While holding the job, he met the director of research applications laboratory at NCAR and shortly afterwards accepted an engineering director position. Currently, Bruce serves as a program director of the aviation applications program at NCAR overseeing the efforts of nearly 70 people. Please join me in welcoming Dr. Bruce Carmichael. Yaga, thank you very much. Thank all of you folks for coming up this evening. We had a session on Saturday afternoon. Some of you may remember Saturday. Surprisingly enough, we had a pretty good showing in the room. But there was a special bonding. All of us had sort of slogged up in the snow and it was sort of ugly. Tonight's a really beautiful night to be out. I'd like to, first of all, thank Yaga and her team for all the good work they've done in putting this together and making it work. They've made everything easy for those of us who are presenting. Question, how many of you in the audience are pilots? Raise your hand if you're a pilot. Okay. I'm guessing 15. That's about the same size of group we had of pilots when we did this on Saturday. So we appreciate the pilots being up here. The only jeopardy in having you guys here is that when I mess up, you're going to know right away and you're going to call me on it. So I'll try to be careful. So I direct a research group. I want to say, first of all, that what I'm presenting are the things that our group has done. So I'm representing the work of a number of researchers in the research applications lab. And I'm not representing my own work here, but I hope I will do them justice as I show you aviation. So first of all, the thing you see in the background there is what we call a microburst. Now, back in the late 70s, early 80s, that name was not part of the jargon. But there were airplanes that were falling out of the sky on takeoff and landing, big commercial airlines. And there was something happening that people didn't quite understand. So NCAR, along with other research organizations, were sponsored by the FAA to go in and do the research to try to figure out what the problem was, exactly what was happening in the atmosphere. Now, here is, let me play that clip. And you'll see at the end of the clip, sort of a real-time microburst watches as it unrolls for about 20 minutes here on the clock. And you look at that and you say, that looks really ugly, right? And what pilot would really like to fly through that? Now, what exactly what is this microburst thing that we're talking about? Let me explain to you what the issue is. So think of having a large bucket of water and you pour it out on the ground and it's just sort of, as it hits the ground, it goes every which way. Now, imagine that you're an airplane and you're on approach to the runway and you've got the wheels down, you've got the flaps down and you're going pretty slow. And now pilots are trained that when they're landing, they want to stay on the glide slope and they want to maintain a target airspeed. So everything is just fine. And then suddenly they encounter a headwind when they enter this microburst. And so suddenly their airspeed goes up and they may start to rise above the glide slope. So pilots are taught instinctively to reach over and pull the power back. And so that will slow it down and get it back on the glide slope. Now, about the time they get that done, they get to the middle of this thing and now there's no longer the headwind, now it's a downburst. And suddenly they lose airspeed and they start descending below the glide slope. And then as they're pushing the power up, that becomes a tailwind. And so now they really lose airspeed. And so they're pushing the power up but they're sinking on the glide slope. And if they can't get the engine spooled up quickly enough, then they pancake into the ground. So what did that look like? It looks like that. There are a number of those accidents that occurred in the late 70s, 80s. This one happens to be in 1988 in Dallas. So the research community figured out with the help of a fellow named Ted Pugeta from the University of Chicago and a lot of other researchers, what was the problem? And then the question becomes, how do we mitigate the problem? So one of the first steps in mitigation was the development of a system called the low-level wind shear alert system. This is the series of animometers of wind sensors that are placed on poles around the airport. And by looking at the variation of the wind across this network, one can deduce where a microburst might be. If you go to the airport in Denver and look around a little bit, you'll see 32 of these poles sticking up around the Denver Airport. So this is very good for figuring out where a microburst is, sort of within this network of sensors. Another piece of the puzzle was something called the terminal Doppler weather radar. This is a radar that is able to look beyond the network of wind sensors and actually figure out where the microbursts are further out away from the airport. This work was done by a collaboration jointly between NCAR, MIT Lincoln Labs, and others It was installed at 50 U.S. airports that happened to have a high danger of wind shear and a lot of traffic. And a lot of these systems have been deployed in one form or another internationally now. So these systems along with training have greatly reduced wind shear related accidents. Now this was an early decision support tool that we developed. And the tool, there were two versions of it. There was a graphical display which the supervisors and managers would look at and they could see on that tool where the microburst was. If you see the red ellipse there, there's a microburst there. But the controllers themselves who were talking to the pilots really didn't want to have to look at a graphical display and figure out what it meant. So for them, for the controllers talking to the pilots on the radio, a different display was developed. It looks very simple but actually developing it and figuring out how to translate what we knew into this message was not easy. So for example, if a United flight was arriving on runway 26 and this was happening, the controller would read this message. He would say United 235 microburst alert, expect a 35 knot loss on a three mile final wind 100 at four clear to land. I can see that makes a lot of sense to all y'all. What do you mean clear to land? I'm told that there was a case at Denver at one time where there was a 90 knot microburst and the controller gave the message to the pilot and said say intentions and the pilot said say again. 90 knot loss. Pilots I think would go around. Thank you very much. But that provoked some interesting discussion with the FAA about what controllers should and shouldn't do because the controllers weren't supposed to be making decisions for the pilots. They were just providing information and in this case the controller was nudging the pilot to say you better make a decision say intentions. So the controller got in a bit of hot water about that. We've evolved today so that a controller will never put a pilot onto an approach that has a microburst active on that approach. So they'll never have to do a go around. That's a whole lot safer situation. Was that 35 knots? Is that tail versus head or just head versus still there? Yeah. Yeah. Okay. Well, we've been we were successful. Here are the fatalities that people were seeing on wind shear accidents. And there were there were several things that helped. This was not all one one thing. There was of course the TWR and ELWAS. There was also before that pilot training. There was a wind shear training aid and there was extensive pilot training that was done by the airlines and we were involved in helping put that together. Now I would like to point out there are at least three people in the audience here that I know who were up to their necks in the wind shear program almost from day one. There's Cleon Bider sitting here on the aisle and Bill Mahoney and Bob Barron anybody else that I missed? I came in 91 and when I got here 91 this problem was was mostly a solved problem. So now let me let me turn the calendar forward to modern more modern times. The wind shear problem sort of a solved problem but there are other weather things that impact the system. Now the reason that weather is an issue is because for the most part it causes a hazard. In order to avoid the hazard we've learned to be very conservative in our decision-making. And by being conservative in decision-making what we do is we introduce into the system delays. We reduce our capacity because we're being conservative. Ten million minutes in 2013 of weather related delay in the system two-thirds of that we think is avoidable if we could do a better job of forecasting and using those forecasts in the system. So the federal government is designing something called the next generation air transportation system. This national plan involves multiple agencies. The thing is it can't work without better weather information. Not just better weather information but that information has to be integrated closely into the air traffic management decision process. So safety then is always the controlling objective in aviation. The problem that we face is how to achieve safety while at the same time reducing delays, increasing capacity, improving efficiency. So we need to do a better job of precision forecasting the various hazards to flight. So what I'm going to do now is we're going to step through a number of the hazard areas, another number of the things that we have been working to try to do a better job of forecasting. So these are some of the areas that we'll be looking at. We'll be looking at in-flight icing, turbulence, convective weather, much rain-induced turbulence, winter weather, ceiling visibility, starting out with icing. Now for large commercial aircraft, by and large, in-flight icing is not a huge issue. The airlines are very well protected. They run hot air off the jet engines, usually into the wing, keeps the ice melted off. We very seldom hear of an icing accident with a large commercial aircraft. But there was one with a commuter aircraft in Roseland, Indiana, which sort of opened everybody's eyes up to the fact that icing was not just a problem for those of us who fly small GA aircraft. And it turns out there have been over 800 deaths over 20 years in non-commercial aviation related to icing. So we have developed technology that has been deployed to National Weather Service and it helps pilots understand where the hazard is and helps them make decisions to stay out of the hazardous area. The FAA estimates that we're preventing eight accidents a year and saving 60 million dollars a year when one makes the evaluation of the cost of life, which is sort of an odd calculation the federal government likes to make. So what is this icing thing? Well first of all, let me tell you that it doesn't start out as ice. We have an aircraft wing and we have little water droplets. These are water droplets, but they happen to be, they happen to be lower than, the temperature happens to be lower than freezing, but they stay water until they impact the wing. And then these water droplets stick and they freeze as soon as they hit the aircraft. You've probably encountered this driving your car and it's raining and as soon as that rain gets your windshield you've seen it freeze on the windshield. Well this is exactly what's happening with structural icing on an airplane. Well why do we care? Well we care because two things happen. One thing that happens, the lift is generated on the wing by a smooth airflow over the wing. The ice disturbs that smooth airflow and we get less lift generated by the wing. At the same time it's rough and it generates drag so it gets us on two fronts, it reduces the lift and increases the drag. And so we may get to the point where the power on the aircraft is not sufficient to maintain the altitude and the airplane begins to sink. So that's why icing is a problem. Here's a video of a wing showing the icing build up on that wing. Now if you're flying along and you look at it and you see that happening you want to get out of dodge right away unless you have some way to get that eyes off like inflatable boots on the wing or some sort of a chemical system on the wing. It's not a good place to be. Now I want to shift gears and talk about something else that's called icing. Oh before I do that let me just show you the products. So we have a couple of products. We have something called the current icing product. The current icing product is basically a diagnosis of where the icing is right now and so what we do is you pick an altitude and you pick a time, well in the case of current icing it's now and you can draw a flight path in and then you can see a vertical cross section of icing along that flight path. Now this is a picture of icing severity. By sliding the bar at the bottom to a different time you can actually look at a forecast that looks the same except it's in the future and then another piece of this puzzle is the icing probability. So it's the same kind of depiction with the vertical and horizontal depiction and the time slider but now we can say not only how severe is the icing but how probable is the icing. All right shifting gears here is a modern jet engine flying at flight level. This jet engine is ingesting lots of ice crystals and as the ice crystals get into the engine they begin to adhere to the blades and pretty soon things get so gummed up inside the engine that it may just stop. Imagine if you're flying on a large jet liner at flight level and all the engines just stop disconcerting. So this is called high ice water content engine icing. This is a phenomena that has become much more prevalent as we have made aircraft engines more efficient. So the new more efficient engines are more susceptible to this problem than a lot of the older engines. In fact to the point that the FAA has had to issue special directives to people flying 787s and the new 747s with the newer engines that they're to stay as much as 50 miles away from any thunderstorms particularly on the downwind side. What happens is these ice crystals typically blow off the top of a thunderstorm in the anvil. This is optically may be optically totally clear so the pilot doesn't see it but those ice crystals those small crystals are there saturating the atmosphere. This is a global problem. It tends to occur a lot in a tropical air mass. So here are 67 engine icing events that Boeing analyzed just to sort of show you where they are. So this has been an international problem, international organizations are all over this and we have worked with NASA, the FAA, the Canadian folks, the Europeans, the Australians. Everybody is trying to deal with this problem. And we have developed an algorithm called alpha algorithm for prediction of high water areas. We now are running this over the North American domain as well as over Australia and the Indonesian area. The idea is for us as weather people to be able to predict areas where this kind of ice crystal behavior is likely to occur. Now there's another side of the research that's being carried out by the people who built engines to try to figure out how to modify engines so they're not susceptible to the problem. So we're coming at this solution from two different directions. Fix the engines and figure out where it's happening in the atmosphere so we can avoid the areas. So for the airlines this is a big deal. I mean they're being safe but to do that they're spending lots of extra fuel and lots of extra time going way around thunderstorms. And in some of the long oceanic flights that's an expensive proposition to avoid all of this convective weather. We ran several field programs. One of them we ran out of Darwin, Australia. And out of Darwin we had a Falcon 20 and so when this kind of weather was forecast we would send the Falcon 20 up and they would fly patterns in this area and try to diagnose whether or not we were getting it right with Alpha. So this field program again was a combination of the Europeans, Canadians, and the US. Shifting gears, a whole other set of problems for aviation. Ceiling and visibility. Now let me start out by saying a little bit about ceiling visibility as it relates to small aircraft. Now it's not the size of the aircraft so much that's important. It's the qualification of the pilot. In general aviation the majority of pilots who fly GA aircraft are not rated to fly an instrument meteorological conditions. They're they fly in what are called visual meteorological conditions. They fly visual flight rules which means you got to be able to look out the window and see what's going on. And once you get in a cloud you can't do that. And unless you're trained for that it's very easy to become spatially disoriented and to lose the aircraft in the cloud. So for a lot of GA pilots ceiling visibility is a killer kind of problem. So we're working hard to try to come up with solutions for pilots so that they'll make better decisions on when to stay on the ground or when to avoid areas that are going to likely cause them a CNV problem. Now on the other hand for the airlines ceiling visibility is not much of a safety issue. I mean these guys know how to fly with a low ceiling and and low visibility. Oftentimes they're equipped to fly right down to zero zero. The problem is that for the commercial world whenever we have a ceiling visibility problem it immediately reduces the throughput of the airport. The capacity is reduced. We begin to get delays at the airport. Now one of the reasons for this is that typically on a pretty day the the airline may approach an airport and the controller will say clear for the visual approach. Now once he does that the pilot then is totally free to space the aircraft in whatever manner they deem safe to land on that runway. They don't have the required kind of procedures in spacing but as soon as that weather becomes instrument conditions now the pilot has to follow instrument flight rules and that slows the system down. So almost invariably when we have a ceiling visibility problem at an airport we will see the traffic at that airport slow down. Here's another problem. You might recognize the geography there New York City. Thunderstorm moving through. Look at what's happening out to the west. You see all those planes going in circles out there? Why are they doing that? Didn't they know that they were going to have to do something? They couldn't just go in and land in New York? Well the answer is no they didn't know because we didn't do a good enough job of forecasting with precision where and when this thunderstorm was going to affect the New York airports. So one of the R&D activities that the FAA has been funding for a while now is the development of a very accurate zero to eight hour forecast and this was a joint effort of NCAR, MIT Lincoln Labs, and NOAA to put together this forecast which now is being implemented as part of the next gen by a company called Raytheon. So Raytheon is essentially building the operational version of this forecast which which we have prototyped for a number of years for the FAA and done statistics and verification on. Well here's another classic problem ice on the aircraft prior to take off. Now the FAA frowns on taking off in an airplane with ice on it. The reason is it's the same problem with getting ice on the aircraft in the air. It degrades the performance of the aircraft. So you've probably experienced this. You'll be at DIA in the winter time and it's a winter situation so before the aircraft can take off they'll go out to a deicing ramp and a truck will pull up several trucks in the case of Denver will pull up and they'll start spraying this glop all over the airplane. There's sort of two steps to this. They spray something that will melt the ice off and then they will spray something on that will inhibit the ice reforming. It's sort of like molasses. It's a thick gel kind of stuff and that will keep the ice from reforming for a while. But if you get enough snow falling on that wing and melting into that into that glop pretty soon the glop will freeze. So now what do you have to do? You have to go back and start over again. Rede ice reapply the anti-icing. So for this reason operating efficiently in the winter is much much harder than other times of the year whenever you have snow and freezing precipitation to deal with. So they need decision support tools and here's a tool that that we develop to help understand when they should deice and when they should refrain from deicing and just wait for the weather. This is a this is an analysis with a short forecast indicated by those little barbs, those are 30 minute barbs so they can kind of see what's happening, what's coming at them and sometimes it's just smarter to not do anything. Just wait until the thing clears before you deice and try to get out. Now for the people doing the deicing a lot of times all they need is a very simple depiction. Now let me explain to you that with snow you know when we when we go out and look at snow we sort of say oh we got we got 14 inches of snow. Well that's not exactly what's important for aviation. What's important for aviation is if you take that snow and you melt it how much liquid do you get when you melt the snow because it's that liquid equivalent that dilutes the glop on the wing. So so it's very important then that we keep track of how much liquid equivalent has fallen on the wing after we've done the anti-icing fluid because we may have to go back and do it again. That period of time that's that is predicted is called the holdover time so there's a certain amount of time that that you have to that you're allowed before you have to go do it over again. So the guys who are doing the application are looking very carefully at how much snow is falling now because they know then what it's doing to dilute their deicing fluid. Well and then you also have the problem in the wintertime of snow on the runways on the taxiways breaking action on the runway. So this is a whole other problem where the airlines are worried about deicing aircraft the airports are worried about keeping the runways and taxiways serviceable. Now it turns out that we got some leverage here because the U.S. Department of Transportation had funded NCAR for road weather purposes to develop something called the maintenance decision support system. This system is used by highway departments in a number of states and it helps them make decisions about when to apply chemicals and plow roads in the states. Well it turns out that a runway and a taxiway are not that much different than a highway. And so we've worked with airports for example the city of Denver to apply this same tool to to their interior operation inside the airport gates. So this is an example of what the decision support tool looks like as it's deployed at Denver where you see the the various runways each managed separately with with a history of treatment a history of snowfall and and what's happening on that runway. So each one is dealt with separately. Now if you go into the operation center at the airport you will see sitting side by side this MDSS system and beside that you will see an aircraft de-icing display. So the people in airport operations are sort of looking at both aspects of the problem. Here's another picture at Denver. You probably recognize the the famous bridge across to the ACON course. There's an interesting factoid about that bridge by the way. The idea was that airplanes would be able to go under that but they didn't really properly account for the tail height of all kinds of aircraft. So the controllers have to be really careful which aircraft they route down that pathway. There's some that don't fit sort of like driving down the highway and you see the the overpass and it says 16 foot you know one inch and the truckers have to kind of be conscious of that. We'll talk about lightning you know we talk about thunderstorms we're typically thinking about flight that is the airplane is flying and we're trying to avoid thunderstorms but at the airport lightning provides a different kind of problem. You have people on the ramp who are loading baggage unloading baggage fueling the airplane putting the food and drink on board so the ramp is just full of people doing stuff to that airplane. Imagine what a few lightning strikes on that ramp would do and so the idea is get people off the ramp before the lightning strike happens. Now in case in case you don't think there's ever lightning on the ramp watch this here's an airplane lightning strikes the tail now there's a mantle cover and it goes skittering across over here and you see people going every which way wonder why they're running let's do that again so we'll now watch right here's the mantle cover lightning strikes goes through the airplane mantle cover there it goes there it is right out there these guys probably are gun shy at this point they don't want to finish servicing that airplane so um the problem is that the FAA takes no responsibility for lightning on the ramp that's that's the airline's own personal uh asha uh worker safety problem so every airline does it differently every airline has a different lightning source they have different rules for when to get people off the ramp that is if the lightning is three miles from the field five miles from the field 10 miles from the field and once they get them off the ramp how long do they keep them off five minutes 10 minutes 15 minutes so if you go to an airport like denver and you watch every airline is doing different things there's no consistency across the airport and we've observed this at other airports so this is an area that we're working with the FAA on to to try to make a convincing argument that this is an area that they ought to be concerned about and interested in it also has to do with traffic management uh and delays and capacity because as soon as you pull people off the ramp suddenly everything comes to a halt so now uh that airport is part of a large network so now you've stopped that airport so people are coming in and landing but they can't get to a gate and uh and this ripples across the country so so their big air uh impacts not just a worker safety but also to capacity and efficiency of the system watch this what's that drink card and by the way the passenger who wasn't belted beside the drink card this is a this is a very realistic simulation of a turbulence event in an aircraft the um so so the first moral of this story is keep your seat belt fastened notice the guy on the left does a lot better than the one on the right unless it gets hit by the drink card um turbulence encounters have a significant safety efficiency and workload impact on the airlines um there are a number of pilot reported turbulence encounters 75 of all weather related accidents and incidents are turbulence there is a huge cost to the u.s. airlines due to injuries uh aircraft damage cabin damage flight delays time loss because they have to inspect the aircraft and tear it down and look at it when they have severe turbulence um interestingly enough who do you suppose in the aircraft is the most susceptible to a turbulence injury flight attendants that's exactly right the flight attendants you see them running up and down the aisle doing stuff when the seat belt sign is on and and they're just trying to do their job and and they're holding on to the seats as they're going down the the aisle for each of the major airlines they're on the order of maybe between 300 and 400 major flight attendant injuries per year by major i mean the flight attendant has broken something is hospitalized is off of work so this is a big uh safety issue for the cabin crew this is an area that that we are working with the f a and the airlines on to try to to improve so some of the things that we're doing in turbulence um airplanes are very good sensors for turbulence when they fly through it they know it and so uh we have equipped airlines united delta korean southwest with the ability to automatically report uh the condition of the air as they as they ride through the air uh it's important not only where's the turbulence but it's important to know where the turbulence isn't because where it isn't is where we'd like to be flying so um so these automated turbulence observations from the aircraft are extremely valuable to us in our forecasting process um your turbulence is not easy to detect in clear air we really don't have good sensors that can sit on the ground and tell you where the clear turbulence is we can tell you where the turbulence is in cloud but in the clear air the best sensor we have is the airplane that went in front of you now we take the aircraft data we take data from models we diagnose uh the information and we produce something called graphical turbulence guidance this is very much like the icing product here you see a picture where you pick a flight level uh say 30 flight level 340 and you pick a time and then you can draw a flight path through that and you can see the vertical cross section of turbulence across that flight path now this is an extremely good planning tool for the people who are planning flights it's also an extremely good tool if the pilot can look at it in the cockpit and so we're making inroads now and finally getting the pilot information that they can actually see for themselves notice this aircraft turbulence so um from the biggest to the smallest turbulence can be an issue you know drones are um a big thing right now but there are a lot of weather related things associated with drones that we still are learning about and one of them is how to deal with turbulence so uh we're doing work with NASA they're trying to develop an automated control system for drones and as part of that they want to know what the turbulent situation is in the first 400 feet of the atmosphere above the ground and and so um this is a new project we're excited about it and we think it has a lot of potential from the very smallest to the very largest scale uh we have turbulence all over the globe on any day one can see turbulence in a lot of places and so for flight planning for airlines they really would like to be able to flight plan someplace else other than where they're going to expect this turbulence here a good example is you know typically if you're going from san francisco to to asia you take off and you make a hard right turn and you fly right up the coast and around alaska and down the allusions to get to asia you may wonder why you would take such a circuitous route well the answer is it's not circuitous at all because the globe is round and the shortest great circle distance from the west coast to asia is actually around the alaska and the allusion islands but look at that turbulence up there that's not a place that they want to be flight planning probably so they're probably going to come in with another plan for those flights going to asia with this kind of a forecast here's another uh another kind of wind and turbulence related problem you may remember on my birthday in december of 2008 there was a continental airline that lined up on runway 34 ride at denver applied power and about halfway down the runway you can see from the picture here on the top that they departed the runway and started heading across the pasture skipped over a road jumped over another road landed in a uh pond drainage a drainage area at the airport and the aircraft caught on fire and was destroyed everybody got out there were injuries but there were no fatalities in that accident it's pretty amazing that there were no fatalities so what happened so here you see a modeled behavior of winds in the vicinity of the airport right here where my cursor is is where the airport is in relationship to the mountains now you'll notice on this day that a lot of the wind was was screaming from the west and a lot of these these high winds were periodically sort of dipping down to the surface and so so when the pilot started to roll down the runway there was probably no difficulty but on the roll they got a wind gust across that runway and if you look at the color scale here this is in meters per second a rough way to translate that into miles per hour is sort of double it so so we have a system here you know where maybe 40 50 mile an hour winds suddenly gusting right across the runway during the takeoff roll so there's no way the aircraft or the pilot could deal with that uh oddly enough we have instrumentation at the airport that uh saw that all those elwa sensors that i mentioned to you you know we could find elwa sensors that showed that the problem is that that data is not instantaneously available to the people who need to see it i mean the controller sees a roll-up of all of that the pilot doesn't doesn't see that data so one of the questions that we've been looking at is how could we do a better job of alerting for that kind of a condition here's another airport picture-esque airport anybody know where that one is juno alaska yeah juno is a cute little town with no roads into it so the airport is the is the way people get in and out of juno now this airport is sort of down in a bowl it has mountains kind of surrounding it and sometimes at this airport the winds are such that they have to take off sort of the way you're looking down that picture you know facing away from you and you would say well that doesn't look too bad you just take off and make a little right turn and go down that cute little channel there but imagine that the visibility and ceiling are are not too good and you can't really see very well that's a pretty narrow channel to navigate so when we first started work in juno the only way that they could get out of there you know on those kind of days was they would do what was called a turning departure now a turning departure you see where the airport is on this on this map and you see those two turns now the way this would work is so the pilots would get out of the runway full throttle down the runway take off and as soon as you get off the ground you got to you got to pull the power back you don't want to go too fast because you want to make a really tight turn and the faster you're going the larger your turn radius is so pull the power back and go slow and make that turn as close to the ground as you can get started with it now imagine you got 10 knots at the airport and 100 knots at mountaintop height and think about that airplane in that turn sort of climbing through that layer and imagine the airplane having a 95 degree roll three inner feet off the ground actually that kind of thing happened on one airplane it happened twice 95 degree roll recovery 95 degree roll the other way so probably not a whole lot of happy passengers on that flight so it turned out that the pilot on that flight got an award from the airline pilots association for meritorious airmanship and saving the airplane there was an FAA inspector in the back of the room who got curious what do you mean meritorious airmanship so the FAA intruded itself into the situation and started closing the airport down totally when the winds were adverse so sometimes the airport would be shut down for days at a time in june with no roads and the state legislature there further imagine that you have a u.s senator named ted stevens who was in control of appropriations on the hill so senator stevens sent the FAA a nice note and said fix this problem whatever it takes and take it out of your hide by the way no extra money so of course the FAA was overjoyed but came to NCAR and asked us to take a look at this because we had just finished doing the Hong Kong airport uh Hong Kong airport has somewhat similar problems except not quite as complex as you know it's sort of one mountain on one side of the runways causing most of the problems in Hong Kong but we had built a system custom designed for Hong Kong Bill Mahoney was the project manager on that project and we were successful in getting it done so the FAA said look at june and see what we can do so we flew a lot of aircraft in june we put in sensors we had anemometers at the top of the mountains we had wind profilers which are vertically pointing radars that can uh they can see the wind layers and the wind at different altitudes so with that system uh we were able to devise a set of alerts so that whenever the conditions were dangerous we would give the pilots alerts and they learned not to take off when they got an alert so i'd like to use june as an example sort of an end-to-end project where we started out doing a lot of science that is we had we were focused on turbulence we had to do a lot of field research together a lot of data and analyze it to understand the problem uh we collaborated with the FAA with the weather service universities alaska airlines uh in that field program we flew airplanes we flew both both the Wyoming king air the citation um from UND uh alaska airlines flew their 737 we had the doppler on wheels up there uh gathering data we had anemometers that were specially designed to work in that environment we had to test those anemometers a lot of engineering software development web applications maintaining the sensors and operating them in the field validating all of this with a sponsor they're in the middle you see a picture of one of our anemometer sites a sheet mountain this is an example of what it's like in the winter time up there so it was a real hassle keeping all those mountaintop anemometers working in the winter time in the bottom you see one of the wind profilers vertically pointing radars sitting in the middle of a muskeg swamp it's where it needed to be so that was an interesting exercise this involved tech transfer system hardware and software provided to the FAA so that they could build from our prototype the in-state system uh we had three generations of prototypes the reason we did was in hong kong it took us 44 months to do that project and you know it took us 16 years the problem really wasn't that much harder than hong kong it was that we were fighting the FAA at every turn of the road they didn't want the system they didn't want to maintain it they didn't want to deal with it and so they would stop the funding we would move all the people off the project then the funding would come back and we would move the people back on the project again so we restaffed that project I think three times on the 16 years but finally we turned the system over to the FAA in 2012 for operations and maintenance this is a picture of the display one of the things that happened while we were up there was the technology improved for guiding the aircraft they use a GPS high precision approach and they could go up and down that little channel and everybody thought well that's going to fix the problem we don't have to do the turning departures anymore the problem is when you're in that little channel uh there's a lot of turbulence in that little channel too enough sometimes to knock the autopilot offline and so now the pilot's flying in the clouds in that little channel autopilot kicks off so that's not a good place to be so it turned out that forecasting and alerting on the turbulence in the channel was just as important as for the turning departures here's another atmospheric problem anybody recognize that what was the atmospheric problem here birds flying in the atmosphere now you may wonder why we're we're interested in this well of course it's an aviation hazard there are a lot of people killed with bird strikes every year a high cost to the system but radar turns out to be a pretty good detector of birds if you use it right so we were working with a system called detect who had developed a bird radar system and this bird radar system was on a trailer and could be deployed and moved around the airport to try to understand where the birds are where they're going what areas do we need to worry about in terms of bird control around the airport so this is an ongoing problem not solved yet but that's how the atmospheric science community got involved in bird strikes the aviation digital data service this is a website that NCAR developed with f a funding it was probably the best source of aviation weather information for pilots dispatchers and others this system was prototyped and operated on an experimental basis here for many years and eventually we were able to complete a tech transfer of this system to the National Weather Service so now it runs totally in an operational mode at Kansas City as part of the aviation weather center we know that the number of hits per day last time we we looked was somewhere in the order of 20 million hits a day 500,000 visits a day and a lot of data 300 gigabytes of data spewed out at the site on a daily basis it's a great website now one of the the other things we did with the website was there there was a a rash of helicopter emergency management service helicopter accidents the helicopters got in trouble they were flying you know they they scud run that's the way they get around they try to stay under the ceiling and go slow enough to deal with the visibility so congress was all over the FH case and they came to us and said could you guys meet with the helicopter operators and come up with something in the weather world to improve their operation so after we did that and worked with the users we developed a tool that was part of the ads website for these operators to use that tool now is also running operationally been transferred to the weather service to run here's a picture of some dispatchers who are dispatching these emergency service helicopters and they're sitting there looking at their hems tool trying to make decisions on dispatching you may recognize this flight track from Rio de Janeiro to Paris this was Air France 447 this aircraft flew into this convective activity and lost the aircraft now there are a lot of problems with this but but i assert that if the aircraft had not flown into that system we wouldn't have lost the aircraft and other aircraft were diverting around it but the pilot had no good display in the cockpit other than the radar so they didn't they couldn't sort of look left and right and see what their options were so they went through the middle every day you see these convective systems all of the world and pilots need help to try to avoid this stuff so one of the the beauties of current technology is we can take an ipad we can put that on the internet using the wireless of the aircraft we can put weather data up on that ipad attached to the cockpit and the pilot can see whether that's just as good as the people on the ground can see this is the first time that pilots have actually known as much about the weather they're flying through as the people on the ground know about the weather we're doing this with lufthansa airlines they're flying our our product on their aircraft on the oceanic flights and we're getting a lot of good good feedback from lufthansa on this tablet um finally let me just mention to do all this weather stuff you got to have a lot of data and you've got to be able to move it where you need it you got to be able to access it efficiently and effectively so for a number of years we've been working with the f a on a program originally called next gen uh in new next gen network enabled weather now we change the name to hide you know you change the name oftentimes it's now called common support services weather the idea is to collect weather data from everywhere bring it into a common repository virtual repository and then be able to provide the needed information to anybody in the aviation business who needs the data this is now a completed prototype a harris corporation is under contract at the f a to build this prototype based on work that was done here at ncar uh at mit linkin labs and at noah so finally in closing we have a long history serving the avian community we've made fundamental improvements in the understanding of aviation weather hazards and the science behind those hazards and in the process we've developed a broad array of tools and systems that reduce our overall vulnerability to the weather we really appreciate your coming to see us today and uh we'll have a q and a session and be glad to to try to answer any questions that people may want to pose thank you thank you very much bruce are there any questions regarding clear turbulence i was under the impression that some commercial aircraft have the capability to detect turbulence out ahead of them i noticed that some of the commercial flights i'm on they'll slow down way ahead of time so um so the new radars have the ability to detect turbulence in cloud but in terms of clearer turbulence no the radars do not have the ability to detect the clearer turbulence there is a lie the lidar is able to to look ahead and detect turbulence a very short distance uh but the airlines are not implementing lidars on their aircraft and the distance that you see forward with the lidar is so short that it's not cost effective so there so we have to use other means for dealing with clearer turbulence Dr. Carmichael i have a brother that's a pilot for delta airlines i was actually texting him during your talk and uh he mentioned that automatic uh turbulence detection system you're talking about he says actually they have it now yes they should probably know he says he can't use it for two more weeks but he's excited about it um i asked him a couple days ago i said i was coming this talk i said do you have any questions and he kind of whimsically said well ask the guy when the system's gonna automatically turn the seatbelt sign on and off for me and it sounds like the clear turbulence problem still hasn't been licked are there any other issues there that you guys can help the guys up front figure out what to do to keep passengers and flight attendants safe well so uh one of the things that we have been doing and delta has been our guinea pig with this is we've been working with them to have a tablet in the cockpit that has specifically turbulence information on it so that so that they can help uh the tablet can help them in the decision making process for turbulence now we have a product that's going on to that cut onto that display that we're continually enhancing um we have a a new product called graphical turbulence guidance nowcast and by nowcast we mean that we're taking the current state of the world as we understand it and implementing it into that tablet so it's not just a forecast but what's going on in front of you right now and that includes the radar information from in-cloud radar reports so we get the in-cloud radar information we get satellite data we have other aircraft reports we have model data anything that you could think of that that we can use as an indicator of turbulence is built into that system now let me say another word about the notion of turning on the seatbelt sign um we are working on a project with the fAA called tactical turbulence and the notion behind that project is that when the pilot suddenly finds themselves in a situation where they're going to hit turbulence it's going to happen no matter what then there needs to be an action taken quickly and so this system would provide an alert just says turbulence batten down the hatches it's not designed for the pilot to try to avoid it it's saying you're going to hit it so button down the hatches get the flight attendant seated move quickly so we've been the first step on that is to simulate it in a simulator at the fAA technical center in Atlantic City prove whether or not it really works is this the kind of thing that pilots want and that they will use so uh so that system may be part of the answer to the idea turn on the seatbelt sign the aircraft manufacturers themselves modifying the planes so that they could be more stable or i guess less prone to those downdrafts or microbursts or the outflows yeah the uh the aircraft flight systems are designed to um to do some mitigation of the turbulence to to react to the turbulence as they encounter it so the aircraft behaves better in turbulence nasa has done work actually using lidars to look at a very short distance ahead of the aircraft and have active controls uh that that lidar is is essentially providing the sensing data for where the controls can quickly react to the turbulence just before you hit it so there there is work going on on the aircraft side to try to mitigate the turbulence to some extent but i think it's going to be a long time before we get to the point where we can just turn any airplane uh through any kind of turbulence and expect a good outcome so i've got two questions if i may one can you talk a little bit about the process where you've collectively the community identified the microburst you've talked about once you had the concept that the accidents occurred because of a microburst and how you measure but it seems to me somebody took a lot of unrelated data and said yeah in a moment aha now let's prove it so let me uh let me talk a little bit about a fellow named Ted Fujita Fujita was a professor at university chicago he was a scientist but he was um an observational scientist that is he liked to just look at at nature and sort of in his mind try to deduce what he was seeing and he would go out and look at these accident sites and he would see patterns in the grass uh in the in the vegetation and he was looking at this saying you know it looks to me like what happened here is this kind of a thing um he was uh he was a japanese fellow who survived the atomic bomb attacks in japan and he recalls seeing patterns similar to this and so in his brain this is all clicking you know so a lot of the community said he was crazy i mean what's this guy going on looking at grass that's been blown over and thinking you know that that he knows what's going on but that was really the aha moment was Fujita saying this is what i see when i observe nature after the fact and later on the scientists finally kind of figured it out that yeah he was right so that was it was kind of the process second question somewhat unrelated you had talked about i think the alpha program that had highly accurate forecasting one to eight hours out no that's that's cospa okay yes it's it's it's convective weather a thunderstorm forecast so question is how accurate is highly accurate um and two uh how do you know what kind of feedback says you've talked about everything but sort of feedback and data gathering one of the great things about thunderstorms is that it's really easy to do verification because you can use radar data of what actually happens to verify the forecast you made about what you thought was going to happen so that's one of the one of the nicest verification problems that we have so it's easy relatively easy for us to tell you uh accurately how good a job we're doing now like all other kinds of forecast uh i guarantee you that our five minute forecast is better than our eight hour forecast right so the further out we go in time the less accurate the forecast is um but compared to uh other information that operators have now it's a big improvement so we've done cost benefit studies in the community uh independent organizations have done cost benefit studies and they show a significant dollar improvement from using the more accurate forecast so um not all the way there uh but it's an open-ended architecture so that as we get smarter we can make the forecast smarter i just thought something for you it's this i can't quite see what it is these spots on the wings are the cold snow that is sticking to it and these little spots over here inside the engine is where the cold spots are going inside the propeller over here all right so so we've captured icing here we have icing on the wings and we have icing in the engine that's very good so you you heard all that when i was talking about it so we have icing on the wing and we have ice inside the engine good thank you are there any other questions and this might be a really stupid question there's no stupid question do microbursts only occur low to the ground or do they only affect airplanes low to the ground so so microburst what happens is you you have a thunderstorm and and out of the bottom of that thunderstorm you have the microburst so the microburst is occurring aloft and falling toward the ground now the thing is it really only is a major hazard to airplanes as it nears the ground if you're flying in an airplane through a microburst at a higher altitude it's not that much of a hazard but close to the ground typically when you're landing or taking off you have the wheels down you have the flaps out the airplane is going slow and so you don't have much margin for error when you're taking off or when you're landing at altitude you're you know you're flying quite a bit faster and you're not getting the spreading of the microburst up aloft it's just almost a straight down kind of phenomena yes sir is there is there a way to shorten the delay in radar information that we get through adsb and xm weather yes and i'll tell you what it is so what you see is you take the radar data and it goes through a fair amount of processing and that processing induces latency perhaps 10 minutes let's just use 10 minutes as an example and then you you look at it now you've probably heard people say you know you got to be really careful in using that onboard radar data because it's got a latency in it so don't make tactical decisions based on that data well there's another solution we are really really good at very short term forecasting five minutes ten minutes so what if we send the data to the aircraft so that we display the forecast at the time the pilot is viewing the data so if the pilot's viewing the data we essentially use the extrapolation that corresponds to the viewing time now we take the latency out and those forecasts are very accurate they're a whole lot more accurate than looking at 10 minute old data so yes there is a solution we have made a proposal to the FAA to implement such a solution and we're waiting to hear whether or not they want to do it in terms of the the four manufacturers or the sorry the four airline companies do you get the data from the flight instrumentation or do you get data from your own sensor infrastructure now you're talking about the turbulence data correct so here's the way that works the aircraft already is well equipped to sense the winds that it's flying through and so what we do is we can take that wind data that's already on the aircraft and we can run it through an algorithm that resides on the aircraft that converts that to turbulence so there's no extra sensors required there's no extra there's nothing except some software that gets loaded on the aircraft and then that information is downlinked via whatever communications mechanism the airline is using for downlinking information and goes directly to the weather service any other questions well if there's no more questions thank you again for coming