The weather guy I mean you have an outdoor thermometer a barometer a weather vane even a rain gauge what else do you need to declare that a little rain tomorrow weather forecasting is a much more involved technology and and involves getting observations over the entire planet and using supercomputers to integrate equations so it's probably the most complex thing our species does scientifically in technologically sure we can complain about the weather grumble about the rain when the forecast called for sunshine but thanks to sophisticated orbiting instruments weather forecasts are much more accurate than they were just 20 years ago satellites are our eyes in the sky but as good as they are when it comes to predicting the behavior of a large storm such as hurricanes they can disappoint says Cliff mass an atmospheric scientist at the University of Washington he says the u.s. Is falling behind in weather forecasting but could improve by adopting something called ensemble forecasting the idea here is that you take a number of scenarios and you average the result but for that you need complex modeling software to produce the scenarios which we don't have well the key to forecasting is having comprehensive global information and right now weather satellites are the main source of that information and 95 to 97 percent of all the observations we use for weather forecasting comes from satellites over all the oceans over over the entire atmosphere but can the satellites actually see down into the clouds I mean they're just looking at the cloud tops and they know these some of the satellites can actually look down into the clouds they can they can see temperature and humidity and other features 3 dimensionally it depends on which wavelength we use sort of like Star Trek except we can do it better than Star Trek in what sense there is not much weather in space no but in Star Trek when they orbited a new planet you know Spock would always do the scan of the planet with his wonderful structural information that they had. Have on their own and they told us a class implant in this atmosphere was like this and they could see the storms and things like that we can do better than that you know we have a large number of weather satellites up there they're giving us information about the detail structure of the atmosphere telling us what we're also where the clouds are what the temperature of the clouds are temperature the surface and saving talent telling us the waves and the winds near the surface over the ocean so we have amazing technology for determining what the atmosphere and the ocean surface is like well you're painting a pretty attractive picture here cliff and yet you know we still get it wrong certainly if you make a forecast for 2 weeks into the future I usually assume that's not terribly reliable the whole thing about forecasting is that weather of course is a very complex system and chaotic in some sense so even a small largely unpredictable change variation can have large scale consequences just a few days down the line is that not so I mean is it really a chaotic system it is a chaotic system a chaotic system is one in which differences in initial state can eventually produce large differences in the forecast atmosphere certainly like that but because of the satellites because of the other observational assets we have a much better description of what's happening in our commute computer models now have become extremely sophisticated you go back 20 years ago when we virtually got every one of the storms wrong move it when when they would come in the next day the forecast was bad and now we're pretty good we don't have any more surprises and virtually every storm is predicted the day of the few days before well can you describe the atmosphere can even oceanic forces that come together to form a hurricane you know and why are these storms so violent Where does the energy for these storms come from. Well hurricanes to run their energy from Waste year and that moisture originally comes from water evaporated off the tropical oceans so that's where the energy from hurricanes comes from it's water vapor that's different in mid-latitudes storms that hit here up in Seattle our storms get their energy from the differences in temperature between the north and south latitudes Can you give us an example of how our hurricane forecast was bungled to illustrate what's lacking in our weather forecasting hurricanes it's a mix situation our ability to forecast the track of hurricanes has become immensely better we're very good in describing and forecasting we're hoping it will go during the next few days on the other hand are believed to forecast the intensity of hurricanes is not good and has not improved substantially a great example of that would be Hurricane Matthew which hit Haiti and it went from a Category one to a Category 5 storm in one day and basically we didn't forecast that So track were great but intensity were not you've been somewhat critical in fact of our ability to forecast the weather you would think that the United States would lead the world in this capability doesn't seem to be the case that's right so you know my point is we've come a long way in weather forecasting globally you know we're not the only ones United States is not the only one doing forecasting but the problem is that the United States is formed behind some of the leaders like the European Center for medium range weather forecasting all the u.k. Met Office we're no longer the leaders in global weather prediction and we're not doing what we should for regional or local weather forecasting in which ways have we fallen behind which significant ways if I mean is it just the Quitman if we buy some more computers maybe invest in somebody else's software or is it just you know the total effort being given to the problem lack of funding lack of organization where's the problem. Well the big problem is probably our models our models are inferior to those use elsewhere that's the code that describes the physics of the atmosphere it's the code that describes how we combine the observations to describe what the atmosphere is like right now these are complex numerical models that are run by weather forecasting centers these are models that compas millions of lines of code that include everything from combining the observations to describe the state of the atmosphere to describing how clouds were cow the boundary layer works and the dragon the boundary layer Well these physical processes are described in code you've advocated for something called ensemble forecasting in which I gather you generate numerous examples of possible scenarios maybe you could explain that a little bit why would be better well the future forecasting is definitely in what we call probabilistic prediction we can't run a model once and say you know here's the forecast because there's definite certainties that forecast uncertainties in initial state and uncertainties Mrs 6 others a much better way to forecast this is called in sambal forecasting and that is instead of forecasting once run the model once run the model 50 times each time starting slightly differently in terms of the initial state takes time using slightly different physics and what we get then is a whole collection of forecasts and we can use that to figure out what the uncertainty of the forecasts is and we can actually calculate probabilities for instance if you have 50 forecasts and half say rain and half say no rain well maybe the probability is 50 percent parade but you can do this with temperature and when you Miltie anything so we need to use an Sol bills to go probabilistic in our forecasts so this preference for ensemble forecasting I can understand that you run the model many many times and you look at what comes out are we only running the models once now I mean is that the situation. Well right now we are running some and saw goals but they're too low resolution and not enough of them that's true in the global side we are not running a high resolution and solve all of the United States and enough resolution to get the local features and thunderstorms and convection that we need to do so on a global scale we're doing a poor job at it on the local scale we're not doing it at all what do you think that these improved modeling techniques will do for forecasting phenomena such as snowstorms or tornadoes I mean tornadoes are pretty local right where we have to differentiate between the large scale type of weather features like snow storms and things like thunderstorms or convection. I'm very optimistic that we can get more forecasts skill by using these and sawbones high resolution better models better data simulation how we use the observations I think we can push the horizon of forecasts skill for the larger scale phenomena out you know several more days so I think we can have substantial skill in the 2nd week so I'm looking forward to that it's already happening. Thunderstorms tornadoes that's a much harder problem and that requires you know from Asian requires using in sawmills at extremely high resolution and these are areas that we have to work on so Sunder storms tornadoes that's a those are very hard problems I think we can make progress but the other type of phenomena the big storms I think those we can make very substantial progress and well what about even longer term predictions I'm thinking here maybe the Farmer's Almanac that it's never hesitated to say next year will be exceptionally dry you can count on the drought in California to continue or if I go to you know some medium and they look in their crystal ball if they only tell me what's going to happen in the next 2 weeks you know I probably don't feel like paying and there's anybody say to you Look what are you going to be able to tell me what next year's going to be like. Well I really can't do that the Farmer's Almanac has no skill so that's all throw that out right right there so I would listen to that I mean basically you know we have a lot of skill the 1st week marginal skill the 2nd week and our forecast skill fades very substantially after that long term predictions the only thing we really have that useful is the connection between the atmosphere and the ocean and that gives us some skill for things like understanding how El Nino and La niña call a would weather months ahead but other than that our skill long term is very very marginal cliff mass thank you so very much for speaking with us My pleasure thanks Cliff mass is a professor of atmospheric sciences at the University of Washington. Well up next why merely the suggestion that the weather could be predicted was laughable Lesson 2 centuries ago and what are those weird clouds doing in the Bermuda Triangle it's a weather vane on big picture science. Today we have sophisticated satellites and computer programs to provide data for weather forecasting truly astounding technology and as we've heard there's a desire to up our game but all this fancy equipment it's relatively new it wasn't all that long ago when the height of weather forecasting was to look at a rooftop metal rooster to see which way the wind blew org gaze around for other clues people would look at clouds and try and work out the shape of the clouds if that predicted a thing or the movements of animals if that was going to give them an indication of what was going to happen writer Peter Moore says that prior to 150 years ago the phenomena responsible for weather were still a mystery a space mice who just thought it was chaotic. As anything it was almost a bit like the mood music and a hedgehog film I think that a bit louder and it kind of went one way then it went to No way and it seemed to be connected sometimes but then sometimes it didn't give a damn save it was a strange mysterious phenomena then in London around the $850.00 s. Admiral Robert Fitzroy along with a few other daring scientists shocked the scientific status quo by suggesting that it was possible to produce a weather forecast this idea created a storm of controversy says Peter Moore the author of the weather experiment the word forecast is a particular word which was coined in the 860 s. And was coined for the specific reason just to divide it from the old ideas of prognostications o. Prophecies the word forecast was invented by really the 1st forecast he was called Fitzroy and he saw these forecasts as a kind of blend of of what you could see with your eyes and what you deny in terms of scientific theory so it was a blend of the. 2 things and before that it's had never had the scientific grounding that we would today consider to be part of a forecast say focus on You 150 is we've been having them before that they came under different names but surely even if they didn't have too much science behind it they could correlate things like the shapes of clouds with impending storms and stuff like that I mean in a way that's a forecast you're right and it's maybe a facility we've kind of lost a bit nowadays because we say reliant on the technology to tell us what's going to happen if we don't use our own sense of sites or whatever just to study the clouds or watch what the animals are doing and the start of the 8906 was one of the big embarrassments of science in a way this you'd get a bull in a farmer's field that would know it was going to rain before the scientists in his study of all these instruments and thermometers and barometers and things like this and it's true that the natural world does respond in advance to drops in pressure and things like say you've got that on one hand and on the shapes of clouds is another interesting one there's an old well lots of old sayings which are attributed to this but there's one I remember above all which is clouds like rocks and towers mean great showers and you could almost translate to and to meterological parlance of today and say that's acumen in the sky would you know rocks and tires tie was a Pave the atmosphere and so they kind of come to the same conclusion from a different direction if you like you say the Bulls were able Burdick the weather to I had heard that what is it that a bull knows that would allow them to predict the weather out there are they both find barometers or. Well I think that's nice way of putting it should be of our barometer I think this is I mean well this is a very contentious issue if you go out into the countryside and asked the farmers they'll tell you that for example they will lie down in the field before it rains that's a very common one but then you might have variations any lie down in this particular part of the field. When it's going to rain and so when you say the bovine barometer does exist now in the 19th century another accomplishment among the many of the Victorians was indeed to establish the weather forecasting as assignments. But they were considered rebels were they not I mean in what sense could they be considered a rebel after all weather prediction is something that's beneficial fairly benign nobody gets hurt why were they called rebels Well I think comes down to this issue a prediction within science and applying science before even stablished a theory and velocity different theories going on about the atmosphere in the 19th century that were kind of groundbreaking things people understanding why the wind would blow in a particular direction or why a storm would move around but it was very difficult to prove these things because they didn't have the satellites that we have today where you can look down and see everything going say they had an incomplete theory or shall we say a contentious theory and what they were trying to do is really save the lives of sailors see who would get caught in these storms but if you're trying to apply a theory which hasn't the backing of the entire scientific community or isn't stylish strongly enough you're going to have to sons and lots of people just thought that this was beyond the realms of science and it's only really when you get the Telegraph of the $84850.00 is a time that they start to be able to knit together observations from different places to give us weather maps and pitches the things that happened and yet raised a really interesting question of how much of a prediction is it wise to make because if you get it wrong it could have consequences and the consequences they were worried about was that science would lose its reputation and that was a big deal for them you know perhaps you could elaborate a bit on the basic science that needed to be understood there was Sir Francis both War Matthew Mari working out the mystery of the atmospheres behavior how did they do that well it was a few things 1st of all that to kind of codify the atmosphere in a way that different people could all relate to said it was things. The benefit scale which was a way of quantifying wind speed which was a big move forward there was another innovation from a guy called Lee coward he gave different names to clouds they have straw purchase Cumulus and Sirius and so on so this was a kind of numbing click show which people were able to use it was really to theorists in the 830 s. One called William Redfield devil one called James I speak who really began to understand how hurricanes moved and you know this idea that you know when we think of a big you know tropical cycling today you kind of think of this thing spiraling around in the sky and it's an idea that we have but they didn't have any of that before and and it really was the observations of these 2 men and it turned into a big scientific argument an 830 s. . But this understanding of storm physics was absolutely vital So if you're going to be forecasting storms you need to know how are they going to meet with going to be even around clockwise direction and so on here it was a real big American achievement I think at the heart of your book is the story of Captain Robert Fitzroy who wanted to fund a national weather prediction service in England what was driving him to do this and what sort of resistance did he meet. While Fitzroy is an absolutely fascinating character he is best known really as Charles Darwin's captain on the Beagle. Was a brilliant man but he had a terrible temper and he was I think what today we'd probably classify as someone who was almost bipolar and his personality was very high or use very low. But he was a very hard worker and he had these big ideals and one of his ideals was to decrease the number of fatalities and of sailors see often they'd be caught in storms and as a sailor himself in 850 s. He was around members of the British government and he decided he wants to come up with this you know kind of I did see issue weather warnings and storm predictions and so it's really like from this one very strong dynamic personality that we get the match office which is a weather bureau and that was for Troy's idea I understand that one of his allies in Parliament was nearly laughed out of the chamber when he suggested that we might be able to some day predict the weather 24 hours in advance Yeah that's absolutely correct I mean this was during the debates they were trying to decide how much money to give the new weather bureau if you like and this wonderful little exchange which is documented a bit like a screenplay or something today you can read it and say there's a scientific minded member of parliament stands up and says with the new weather office will soon be able to predict the weather in the metropolis 24 hours before it happens and then there's just a does like a pause and on the next line it just says in brackets laughter it really sums up how ridiculous an idea it was to these established politicians at the time but there was obviously clearly to do something in Fitzroy opinion because you know if you look back at the newspapers from them it was so often see a carrot s