NDG wrote:Aric Dunn wrote:LarryWx wrote:
Part of the reasoning may be that even the GFS now suggests any possible genesis would likely be after 5 days.
current GFS and CMC are 3 to 4 days. they, of course, interpret everything and make a judgment. They also have access to other model outputs we do not. most notably the florida state model which by all indications is superior to the current model suit.
The FSU model is a super ensemble model, I believe they only use it on forecasting the track & intensity of an already developed system. Somebody correct me if I'm wrong.
Excerpts that came out from a nice article from a few days ago.
“It uses a combination of weighted, bias-corrected input models to create track and intensity forecasts for hurricanes and tropical storms,” said Richard Pasch, senior hurricane specialist at the National Hurricane Center. “This differs from a simple model consensus forecast (TCON or TVCN), which uses a simple average of the input models. The FSSE has been one of the most accurate numerical guidance tools for hurricane forecasting and warning at the National Hurricane Center.”
Mark Bourassa, associate director for the Center for Ocean-Atmospheric Prediction Studies and a meteorology professor at FSU, said the FSU invention is a combination or "ensemble" of other weather forecast models. Each of these models use knowledge of physical laws and the weather at the start of the forecast to estimate the weather in the future.
“NOAA routinely produces a simple average of such models,” he said. “The FSU Superensemble does more than that. It uses knowledge of model shortcomings, based on the proceeding weeks of forecasts, to adjust each of these forecasts before they are averaged. These corrections and the resulting improvement in the forecasts are what puts the 'super' in superensemble.“
Misra said the FSU Superensemble became known under Krishnamurti because of its success in forecasting the track and intensity of Atlantic hurricanes. In addition, he said, the model has outperformed other weather models in forecasting rainfall in the tropics and in forecasting climate a season in advance.
The uniqueness of the FSU Superensemble, said Misra, is in the weight given to each of the individual model forecasts entered into the superensemble. These weights are determined by statistical techniques on the performance of the individual model forecasts relative to the observations for past weather events.
“Superensemble uses a combination of forecast models,” he said. “It has usually shown itself to be superior to the best individual forecast model, which makes it quite unique. It is able to correct for systematic errors in the models and compensate for bad with good models through its weighted average technique.”
https://www.tallahassee.com/story/news/hurricane/2018/06/04/forecasting-model-created-florida-state-critical-tool-national-hurricane-center/661956002/
