Atm Ocn 100 1st Edition Lecture 8 Outline of Last Lecture I Weather of the Day notes II Weather Models III Basic Types of Models IV Finite Difference Model V Vertical Grid VI Spherical Projection VII Topography Representation VIII Key Prediction models for this class Outline of Current Lecture II Weather of the day notes III Key Forces that Accelerate the Air IV Force Balances V Pressure Gradient Acceleration Formula VI Friction Forces Current Lecture Weather of the day Back to http tropic ssec wisc edu Possibly a typhoon There is east Pacific tropical cyclone activity also Several have hit Baja and some affects in San Diego and strong affects in Arizona Mostly just rain affects Moisture inland this is what has happened to Arizona Different satellite view of cyclone developing by infrared image that we can see at night We can see a lot more of the same picture a lot more stuff Shows polar tops as clouds move up high they form blue colors on the color map which suggest a colder top The temperature of the clouds are around 245 degrees Kelvin which is about 28 degrees Celsius and somewhere 10 degrees below 0 Fahrenheit Cloud tops when they get really high you get down to below 200 which is 70 80 degrees Weak storm right now Not considered a tropical storm yet Now we can look at forecasts of this http cup aos wisc edu will HFIP AOS page Shows what we are doing in predictions Predicting storms in East Pacific and the Atlantic Deterministic forecast shows the red which is our forecast And blue is GFDL forecast And the green is a separate forecast It shows we all sort of agree until the end Wisconsin model goes to the left But the GFDL model goes toward Baja a little bit Intensity we can also look it the intensity is very low Doesn t even get passed tropical storm stage We are the only ones who get it to tropical storm status other models say it is not as intense Another output here too Ensemble forecast run model again with slightly different physics Can make approximations of physic calculations Different initial conditions almost the same but a little different Most numbers of ensemble are on here Black line is the average Even with the same model there is some differences When time goes by these numbers start to converge Effect of nonlinear chaos that makes solutions become less related to each other Still there is a tight ensemble tight packing of members Forecast based on this and other models we just looked at are pretty certain it is going to go on this track Book goes into this a little bit Deterministic forecast just one member like this last thing we saw 3 different models or 4 different models if you look at one model run our best shot at what we think physics are at initial conditions and that is the forecast Given these initial conditions this is the final result that is the deterministic forecast Initial conditions determine final result Could be wrong but still deterministic forecast Before all deterministic forecast initial condition find weather sometime later by integrating them Now factors that influence this Butterfly affect small difference in initial condition can make a huge difference later on Butterfly effect in South America can change New York weather in a week Small effects work their way up Example Look at water in facet water small doesn t come out straight Water wiggles when it comes out Impossible to predict those wiggles Could you predict those Difficult thing to predict Weather is wiggles We can try to observe those wiggles Someone probe the water coming out at very high resolution down to fractions of a millimeter you would get a ton of data Can see what wiggles will look like a 10 th of a second later or 2 seconds later If you go a short enough time and see big wiggle moving down stream of water can observe that well you may be able to predict what wiggle will be 10 th of a second from now when it moves down Next wiggle may be effect of two wiggles into each other Way up in pipe other things going on way before it gets to faucet If there was a small imperfection maybe hardly feel it can eventually affect water coming out of pipe Two imperfections could affect things may interfere with each other If you knew pipe perfectly and knew all imperfections could predict with initial conditions No other things going on that we can t predict Other things or imperfections that can go on in water upstream Maybe truck drove over Maybe something happened upstream We don t know So many things going on Affect of initial condition wears off just left trying to guess what can happen What can you do with wiggle With water coming out One thing we can predict is that if we have faucet handle at certain condition water will fall straight down not across room 5 ft away from faucet will not feel splash can say with a lot of certainty Can define envelope wiggles allowed to occur Will contain all the possible places where wiggle will be Forecast after time will deteriorate define cylinder where wiggles will be This is what happens with tropical cyclone Know it starts here it could change later Could go different way later can define region where solutions most likely to exist Maybe something like 99 probability that tropical cyclone within certain region that is what we can do with these model runs Definitely sometime later this is the forecast That is a deterministic prediction Ensemble the many lines describing different statistical forecasts we have predicted The lines are different colors The black line is the average between all of them Many things can happen even within constraints of what we have observed We know tropical cyclone isn t going to be random be within ensemble Define probability of feeling tropical cyclone someone hit according to where you are according to ensemble Probability low if you are way far away from ensemble If you are in the region of ensemble maybe 50 or 20 that tropical cyclone is going to hit you here Most members up but some low so not impossible for tropical cyclone to go lower by one line About 20 change tropical cyclone will be If right on black line chances of direct hit from tropical cyclone are as much as 60 7080 Not 100 other solutions can happen with initial conditions More that goes into it looking at all models and many ensembles tropical condition centers may work with almost 150 models at tropical cyclone But look at different models some you believe more than others Statistics can show how good different models are Human