Connect with us


Most Dependable Hurricane Models [ Performance Wise ]



Hurricane Dorian : Dorian computer model has strengthened to a Category 2 hurricane and is anticipated to strengthen further as it approaches Florida this weekend.

According to a forecast issued at 11 p.m. Thursday by the National Hurricane Center, the storm has maximum sustained winds of 105 mph (169 kph), falling just shy of reaching a Category 3 major hurricane.

Dorian is expected to become a major hurricane on Friday and make landfall on Florida’s east coast on Monday night, according to forecasters.


According to their 2019 performance, these are the most dependable hurricane models

most reliable dorian model

If you’re trying to figure out which of the several hurricane computer forecast models to accept, the best advice is to believe none of them.

Put your faith in the forecast from the National Hurricane Center, or NHC.

While a single model may outperform the official NHC forecast in some cases – for example, the European model’s track forecasts for Hurricane Isaias outperformed the NHC forecast at all forecast time periods – the 2019 National Hurricane Center Forecast Verification Report found that it is extremely difficult for any single model to consistently beat the NHC forecasts for track and intensity.

The accuracy of NHC track forecasts was near the five-year average during the 2019 Atlantic hurricane season, with five-day track projections hitting a new high.


One- to three-day track forecast errors have fallen by 70–75 percent in the last 30 years, and four- and five-day track forecast errors have decreased by 60 percent in the last 15 years.

That is a remarkable achievement, one that will undoubtedly result in significant reductions in the number of lives lost, property damage, and emotional distress.

For three- to five-day forecasts, the 2019 NHC Atlantic track projections tended to have a south to southwest bias of 31 – 95 miles (i.e., the official forecast tended to fall to the south or southwest of the verifying position).

The NHC track estimates for Humberto and Lorenzo were exceptionally good, but the overall track forecast errors would have been much lower if it hadn’t been for one oddball storm – Tropical Storm Sebastien – which created huge problems for forecasters.


Sebastien’s average NHC three-day track forecast error was 390 miles, far higher than the standard NHC three-day track forecast error of 118 miles included in the cone of uncertainty.

The European is the best track model in 2019

The official NHC track forecasts for Atlantic storms were difficult to exceed in 2019, as they always were, while the European Center (ECMWF) model did outperform the official NHC forecast for one-day and two-day forecasts.

In 2019, the COAMPS-TC model was the second-best model, closely followed by the UKMET model.

The Navy model (NAVGEM) and NOAA’s GFS, HWRF, and HMON models all underperformed.


The following is a list of some of the key hurricane forecast models utilised by the National Hurricane Center:

Euro: The global prediction model of the European Center for Medium-Range Weather Forecasting (ECMWF).

The Global Forecast System (GFS) is a model developed by the National Oceanic and Atmospheric Administration (NOAA).

The worldwide forecast model of the United Kingdom Met Office is known as UKMET.


Hurricanes in a Multi-scale Ocean-coupled Non-hydrostatic Regional Model (HMON), with GFS data as initialization

HWRF: Hurricane Weather and Research Forecasting regional model, with data from the Global Forecast System (GFS).

COAMPS: COAMPS-TC regional model, with GFS data as a starting point.

Navy: Global Environmental Model for the Navy (NAVGEM).


When the track projections from three or more of these models are averaged together, the NHC official forecast rarely deviates much from it.

These seven models are fed into other “consensus” models, including as “TVCN,” which is frequently mentioned in NHC storm discussions.

Improved consensus modelling techniques are one of the main reasons for the dramatic improvement in NHC track projections over the last 20 years.

Intensity forecasts from the NHC for 2019: not much improvement

In the Atlantic, official NHC intensity forecast errors were similar to the five-year average for 12-hour, one-day, two-day, and three-day forecasts in 2019, but higher than the mean for four-day forecasts and far above the mean for five-day forecasts.


In 2019, the average intensity forecast error was 5 mph for 24 hours and 30 mph for five-day forecasts.

Official forecasts were biassed too low at all forecast times, with four-day and five-day forecasts showing the biggest underprediction.

The National Hurricane Center performed a poor job forecasting Hurricane Dorian’s rapid intensification, which bottomed out as a category 5 storm with 185 mph winds over the Bahamas, which contributed to the absence of stronger long-range intensity forecasts in 2019.

The COAMPS-TC is the best intensity model in 2019

In 2019, the official NHC intensity forecast surpassed the five top intensity models at all forecast times, with the exception of five-day projections, where the Naval Research Laboratory’s COAMPS-TC model fared best.


The regional/dynamical HWRF, HMON, and COAMPS-TC models (which divide the atmosphere into a 3-D grid surrounding the storm and solve the atmospheric equations of fluid flow at each point on the grid), as well as the statistics-based LGEM and DSHP models, make up the five primary intensity models (DSHP is the SHIPS model with inland decay of a storm factored in).

The European (ECMWF) and GFS models, two of the best global dynamical models for hurricane course, are often ignored by NHC forecasters when making intensity forecasts. In 2019, these models predicted low intensity.

For one-day to 1.5-day projections in 2019, the HMON model was the best-performing intensity model, with the other four primary intensity models close behind.

At 2-day, 3-day, 4-day, and 5-day forecasts, the COAMPS-TC model was the most accurate.


When it came to longer-term forecasting, the HMON model fared badly. When it came to forecasting intensity, all of the models exhibited a low bias.

Improvements to the model for 2020

For the 2020 hurricane season, many of the leading models used to forecast hurricanes have undergone major upgrades.

Unfortunately, model accuracy reductions as a result of reduced atmospheric sounding data being gathered by commercial aeroplanes due to the Covid-19 pandemic may be partly compensated by these changes.

(This past spring, an open-access research published in Geophysical Research Letters examined the data gap and its impact on global model performance.)


On June 30, the European model’s most recent upgrade went live.

More data from many satellites, particularly the latest radio occultation sounding data from the COSMIC-2 satellites released last year, will improve the HWRF model, which was modified at the end of July.

For the first time, the model will take in Doppler radar data from land-based radars.

NOAA’s Jim Yoe said in an interview with that he expects “improvements on the order of 5% for track and potentially as much as 8% for hurricane intensity in the North Atlantic and Pacific basins.


” The HWRF model employs three nested grids with resolutions of 13.5 km, 4.5 km, and 1.5 km that are zoomed in on hurricanes. has graphics for the legacy HWRF model as well as the new version, dubbed “HWRF-PARA.”

The GFS model received its most recent substantial modification in June of this year.

The GFSv16, the next major improvement, is anticipated for the winter of 2021.


The updated GFSv16 output, designated “GFS-Para,” can be examined at (for parallel GFS model).

Most of the top global models have ensemble model runs available.

An ensemble model is constructed by taking a forecast from a high-resolution version of a model, such as the GFS or European, and running many versions of the model with slightly varying initial circumstances to generate a set of prospective forecasts that indicate possible uncertainties.

To save computer time, some ensemble members are run at a lesser resolution.


The GFS has 21 ensemble members, while the European model has 51.

The experimental phase of a new version of the GFS ensemble model (dubbed GEFSv12) is underway. This is the first update to GEFS in five years.

The new GEFS boosts its members’ resolution to 25 km, grows to 31 members, and extends the 0Z run to Day 35.

(note: there is approximately a 24-hour delay for Days 17-35 to be recorded). For the time being, consider Days 17-35 ensemble forecasts with a grain of salt, but they may still be useful for tracking long-term or seasonal variations.


Tropical Tidbits offers plots of the experimental GEFS as GEFS-Para.

Hurricane Isaias forecasts

I plotted the model performance for Hurricane Isaias, 2020’s most major Atlantic tropical cyclone thus far, using data provided by Brian Tang of SUNY Albany.

In terms of track, the European model continued to shine, surpassing every other model as well as the official NHC forecast at every time interval.

The official NHC forecast, as well as the UKMET and Navy NAVGEM models, behaved admirably.


In terms of intensity, there was no difference between the official NHC forecast and several forecasts, while the LGEM model fared substantially worse than the other top models.

It’s worth noting that the greatest track model, the European, isn’t a top intensity model thus performed poorly for Isaias.

Forecasts for tropical cyclone formation

The National Hurricane Center (NHC) publishes a Tropical Weather Outlook four times a day, with two-day and five-day forecasts of tropical cyclone genesis.

For five-day genesis estimates of 40–70 percent for the Atlantic in 2019, these forecasts were quite reliable: So, when the National Hurricane Center predicted a tropical cyclone would emerge within five days, it was right around 40% of the time.


NHC’s genesis forecasts, on the other hand, were overly conservative at the lower and top ends of the distribution.

One hundred percent of the storms with an 80 percent or 90 percent likelihood of developing developed, and 20% of the systems with a 0% chance of developing developed.

Surprisingly, the worst NHC predictions were for systems with a 30% chance of forming – over 80% of these systems eventually developed.

You Might Read :


Continue Reading