| Whoa! I checked in to my local weather office to see what they were thinking regarding the future Lee, and found this out-of-the-blue and in-depth discussion of the spaghetti models. Really great stuff here, from today's forecast discussion for the DFW area. Check it out: DUE TO THE INCREASED USE OF SPAGHETTI PLOTS FOR THE DISTURBANCE IN THE GULF...I THOUGHT IT WOULD BE OF USE TO PROVIDE A BRIEF DESCRIPTION OF THE VARIOUS MODELS AND THEIR FORECASTING METHODS. THIS INFORMATION IS PROVIDED AT LENGTH BELOW. THERE ARE FOUR TYPES OF MODELS USED...STATISTICAL...STATISTICAL- DYNAMICAL...DYNAMICAL...AND TRAJECTORY. ENSEMBLE AND CONSENSUS FORECAST METHODS ARE ALSO USED BUT THESE METHODS USE A COMBINATION FOUR TYPES OF MODELS LISTED ABOVE. STATISTICAL MODELS USE FACTORS SUCH AS LOCATION OF THE STORM...TIME OF YEAR...AND PAST STORM TRACK AND INTENSITY TO MAKE THEIR FORECASTS. AS SUCH...THEY PROVIDE A HISTORICAL BASELINE FOR A STORMS TRACK. THE AVERAGE TRACK ERROR FOR THE CLIPER5...THE MOST WIDELY USED STATISTICAL MODEL...WAS 200 N MI AT 96 HRS. THESE MODELS HAVE RECENTLY BEEN SURPASSED IN ACCURACY OF TRACK BY DYNAMICAL MODELS. DYNAMICAL-STATISTICAL MODELS...SUCH AS THE SHIPS AND LGEM...ARE BASED ON STATISTICAL RELATIONSHIPS BETWEEN CLIMATOLOGY...THE DYNAMICAL OUTPUT FROM THE GLOBAL MODELS...SUCH AS SEA SURFACE TEMPERATURES AND VERTICAL WIND SHEAR...AND STORM MOTION AND INTENSITY. THE SHIPS MODEL IS ONE OF THE MOST RELIABLE TROPICAL CYCLONE MODELS. DYNAMICAL MODELS...SUCH AS THE GFS...ECMWF...GEM...AND CMC...ARE THE MODELS TYPICALLY USED IN EVERYDAY FORECASTING. THERE IS A LONG LIST OF THESE TYPE OF MODELS AND THEY VARY IN DOMAIN FROM GLOBAL TO MESOSCALE. THE BEST PERFORMING MODELS IN 2010 WERE THE ECMWF...GFS..AND UKMET..IN THAT ORDER. ONE THING THAT SHOULD BE NOTED ABOUT DYNAMICAL MODELS IS THE DIFFERENCE BETWEEN EARLY-TRACK VERSUS LATE-TRACK MODEL GUIDANCE. LATE-TRACK GUIDANCE IS TYPICALLY WHAT WE THINK OF WITH COMPUTER MODELS. LATE-TRACK GUIDANCE INCLUDES THE ENTIRE FORECAST OF A PARTICULAR MODEL RUN. THE EARLY- TRACK GUIDANCE IS THE FORECAST DATA FROM THE PREVIOUS RUN OF MODEL IS INTERPOLATED SO THAT THE 6 HOUR FORECAST MATCHES THE OBSERVED DATA. EARLY-TRACK MODEL RUNS ARE NOT NEW MODEL RUNS. DYNAMICAL MODEL RUNS USUALLY TAKE THREE TO FOUR HOURS TO COMPLETE. FOR EXAMPLE...THE EARLY-TRACK 12Z IS NOT A NEW MODEL RUN. THE 12Z RUNS OF THE GLOBAL MODELS ARE USUALLY AVAILABLE AROUND 16Z WHILE THE REGIONAL AND MESOSCALE MODELS AROUND 15Z. TRAJECTORY MODELS...SUCH AS THE BAM...ARE VERY SIMPLE AND MOVE A TROPICAL CYCLONE BASED ON THE PREVAILING FLOW OUTPUT FROM DYNAMICAL MODELS. THEY INCLUDE A CORRECTION TERM FOR THE SPIN OF THE EARTH...KNOWN AS THE VARIABLE BETA...AND ARE SOMETIMES CALLED BETA-ADVECTION MODELS. IF THE FLOW IS WEAK AND THERE ARE NO MAJOR SYNOPTIC CHANCES...TROPICAL CYCLONE WILL FOLLOW THESE MODELS CLOSELY. ENSEMBLE AND CONSENSUS MODELS USUALLY PROVIDE THE BEST FORECAST. CONSENSUS MODELS FOR THE TRACK FORECAST BEGIN WITH THE LETTER T...SUCH AS THE TCON AND TVCN WHILE THE CONSENSUS MODELS FOR THE INTENSITY BEGIN WITH THE LETTER I...SUCH AS THE ICON AND IVCN. HOPEFULLY THIS INFORMATION WILL HELP SHED SOME LIGHT ON THE COMPLEX SPAGHETTI PLOTS. MOST OF THIS INFORMATION WAS FOUND IN TWO DOCUMENTS PRODUCED BY THE NHC. THESE DOCUMENTS CAN BE FOUND AT WWW.NHC.NOAA.GOV/VERIFICATION/PDFS/VERIFICATION_2010.PDF AND WWW.NHC.NOAA.GOV/PDFS/MODEL_SUMMARY_20090724.PDF |