Good Morning everyone from very wet South Florida. This summer season started out very dry as June and part of July at least the first part of July started out very dry. In South Florida we were running a significant deficit in rainfall. By mid – June, we were over 10 inches below normal for year to date figures. However, since mid-July the rain has come with a vengeance. We are above normal for August and at least in Wellington a western suburb of West Palm Beach we are well above the normal. In fact this past week, I measured 6 inches of rain in my rain gauge. The tropics look busy and quiet at the same time. There is a flurry of activity from Danny to Erica, to Grace and now Henri but the wind shear that was predicted to affect the Atlantic Main Development Region that was predicted due to a strong El Nino is entrenched. These areas of high wind in the upper levels of the atmosphere are caused by thunderstorm development in the Eastern Pacific. These storms cause a diverging wind aloft and “vent” the air to the east affecting the Atlantic side of the tropics. Thus a flurry of activity but no significant development. To provide some insight of the potential energy in the atmosphere in south florida I will give you some facts. since March 1st in Wellington Fl. The temperature has been 90 degrees or better, if you combine a 70 degree dewpoint or better those are the ingredients for potential strong thunderstorm development. Now, not everyday there was a 70 degree dewpoint or better especially in March or April dewpoint didn’t reach 70 degrees but May onward the humidity and heat was in full swing. Strong to severe thunderstorms are like thoroughbreds or Ferraris because it is not easy for them to initializes, all the parameters must be met – heat 90+ degrees or better(esp in S. FL), surface moisture(dewpoint 70+ degrees), cool to cold pocket aloft, unny to start the day to create instability(cold aloft, warm at the surface) and some luck.
Ok, I don’t want to get off track of what I want to talk about today and that’s Computer Modeling Forecasting. Computer Model Forecasts are a great tool and support resource for all Meteorologists. This includes : Broadcast, Operational, Aviation, and military branches of meteorology. Forecasters can use these Models to predict where Low Pressure systems are going to be in the short range and the long range, in terms of the number of days to weeks ahead. Right now there are many Computer Model Forecast brands out there. I will discuss today the Names of the Models, Numerical Weather Models, and the Advantages and Disadvantages of these Computer Forecast Models.
1. COMPUTER FORECAST MODEL NAMES
A. GFS – The Global Forecast Model is a Numerical Weather Forecast model produced from the National Centers for Environmental Prediction. Dozens of Atmospheric and Land-Soil variables are available through this data set: From TEMPERATURE, WIND, PRECIPITATION, SOIL MOISTURE, AND ATMOSPHERIC OZONE. The GFS covers or forecasts the entire globe, with updated forecasts at different times based on Greenwich time or (GMT). The GFS uses a base horizontal resolution of 18 miles between grid points, which is used by operational weather forecasters to a period of 16 days out. The GFS is available to the public on the www.noaa.gov website. The GFS is a coupled model, composed of 4 separate models(atmospheric, ocean, land/soil, sea ice) which work together to provide accurate weather forecasts worldwide. Changes are regularly made to the GFS to provide accuracy and performance. The GFS just received an upgrade this year.
B. ECMWF – which is the leading global forecast model for medium range and one of the best models out there. It is now considered the Premier global computer forecast Model. THe ECMWF or the European Computer Forecast Model has ten times more computing ability than the National Weather Service GFS model. For example, Hurricane Sandy or Superstorm Sandy the Euro Forecast Model was the only one to accurately predict where it was going. The European Model is also the leading Model for hurricane or tropical weather prediction. The ECMWF uses a concept called 4D. 4D is an assimilation that basically allows the weather model to be updated or nudged with accurate space-time information as new satellite and other data input is available. It is akin to putting a beach ball in the stream or river and predicting where this ball might be in a few days. Remember folks, the atmospheric is fluid(first law in thermodynamics) so it is analogous to the river.
There are some computer models that the National Oceanic Atmospheric administration is working on called the:
C. WRF and HRRR – These initially seem very highly rated models. The HRRR is a very short range model that provides neighborhoods/town scale forecasts and is assimilating radar and other rapidly advancing data points. The WRF seems to be a good Tropical Cyclone model in conjunction with the GFS. The GFS as we speak are getting upgrades in Resolution and more grid point data.
The are so many computer Models out there I will list as many as I can but I will here all week discussing all of them, maybe all month or all year for that matter. More notable models :
D. NAM – North American Model – Short Range Numerical Model forecasts to 3 days out.
E. UKMET – Great Britain Computer Forecast Model . Another World Class Model like the ECMWF, and GFS.
F. CMC – Canadian Model. Inferior to the ones listed above. This Model loves to overblow all storms. If this Model was the most accurate we would all be in a lot of trouble.
G. BAMS – used in tropical Meteorology.
H. NAVGEM – U.S Navy Environmental Model formerly was called NOGAPS. The U.S. Navy improved the NOGAPS big time.
These here are examples of computer models. I will keep this at a reasonable length to its easy to read.
blogs.trb.com
Here are what is called spaghetti computer plot models, which display where storms MAY move. The operative term is MAY, because 48 hours and out longer the degree of accuracy goes down big time. This was Hurricane Ike in 2008 and Ike made landfall in Houston not South Florida.
2. NUMERICAL WEATHER PREDICTION COMPUTER MODELS
The models uses the power of Computers to make accurate weather and climate forecasts. Complex Computer systems, also known as forecast models, run on supercomputers and provide predictions on many different atmospheric parameters such as: Temperature, wind, pressure, and precipitation. These numerical models work by taking current weather observations and processing this data with numeric computer models to make an accurate forecasts.
Like I pointed out above I will provide some advantages and disadvantages of Numeric based Computer Models like the ones listed above.
I. DISADVANTAGES
A. Computing Power – Remember what I stated that some computer forecast models have a lack of computing power well here it is in more details. My professors n school told me that we could run a computer model forecast for tomorrow with 100% accuracy, but it would not finish running until the day after tomorrow. As computing power increases, our forecasts won’t , though they could reach a point where they can’t get much better without more initialization of data. Right now because of lack of computing power, we have to pick between high resolution models, wide coverage areas, or forecast length. For example, it takes about the same time to run the GFS worldwide 15 day forecast run at poor resolution, as it takes the 4-km WRF over half of the U.S. at extremely high resolution.
B. Initialization of Data – In the latter part of the 20th Century, the computer models only took data from a sparse network of upper level weather stations. Which sent up balloons that transmit weather data back to Earth, and when put together, form a 3-D picture of the current state of the atmosphere. This is called Initialization. Then they apply algorithms to applied mathematical equations to predict how the atmosphere will behave. Initialization can be good or bad. Its bad or can ruin a forecast if bad or missing data is inputted. Its called the GIGO rule, Garbage in and garbage out. Sadly, the balloon procedures haven’t changed much in time, however computer modeling has improved some and I will discuss it in advantages. Lastly the more data you have the slower the initialization if more parameters are included in the data.
C. Bugs – Like many computer programs, the models are subjected to Bugs in their hundreds of thousands of lines of codes, created by humans bound to make mistakes by typos or math errors. Bugs can hide in algorithms as well.
D. Model Bias – Each Computer forecast model has a “bias” regarding certain weather storm systems or weather/climate situations. Like I stated before the Canadian Model has a Bias in developing every tropical sustem and turns them into CAT 5 Hurricanes. Source – (Accuweather)
II. ADVANTAGES
There are many advantages to Computer Forecast models.
A. The first one to come to mind is Speed of forecasting. These models can compute differential equations way, way faster than human beings.
B. They are a great Resource and Tool for any forecaster especially the ECMWF which is renown for its speed and accuracy medium to long range.
C. The third is Technology and Convenience – with increasing accuracy of computer forecast models Meteorologists can make accurate and precise forecasts for regions both local and distant using Computer Models.
noaa.gov
Here is a precipitation map showing isobars – lines of equal pressure at height and isohyet areas of rainfall and precipitation using GFS data going 5 days out. The trough in the middle of the country digging south and High Pressure anchored over the Northeast nosing southwestward. The green areas are the precipitation with the blue to red colors of increasing heavy precipitation.
In Short, Computer Forecast Models are a great tool for Meteorologists and Forecasters to make accurate predictions going forward 7 days out. However there are some issues stated above. If we look at Tropical Storm Erica all the Computer Models had a tough time forecasting Intensity and Forecast Track of Erica. Sometimes as Meteorologists or even Weather Enthusiast we need to go to the 500mb charts and look at the synoptic or large scale systems governing other synoptic scaled systems or mesoscale systems. I hope you like the blog. This is Brandon Brady now back to busy school schedule.