Indicated wind speed on earth.nullschool

[Mirror16]

I'm looking at hurricane Matthew in the Caribbean and its current highest indicated wind speed on earth.nullschool is about 50 knots. But the news and the National Weather Service reports sustained speeds of 130 knots. Why this huge discrepancy?

[jamhass]

earth.nullschool, windty, etc etc are displaying results of global models (essentially GRIB files). These compute average winds within cells of some dimension (on the order of km's, exact size depends on the model.) Thus, they do not usually predict maximum velocities correctly, nor do they accurately predict highly localized structures. The numbers NHC report are often the results of specific measurements, sometimes satellite, sometimes buoys, often from hurricane chasing aircraft.

Plus, their predictions are based on more comprehensive hurricane-specific models.

[Mirror16]

Thanks for the education jamhass. How accurate are the GRIB-based wind speed models in less extreme weather?

[jamhass]

It depends ...

On the particular model, on the location, etc etc.

Some models do better than others in some cases, some not so. If you look at several different model results and they all agree, you can have pretty good confidence in the result. It is hard to predict which model works best under which conditions and/or locals. That is where experience helps, and makes really experienced weather gurus and navigators quite in demand.

Consider the NHC presentation of different hurricane track models. Sometimes they all agree nicely, sometimes not.

Location near land is often problematic. Local conditions impact winds and weather significantly and are not always well represented in the models. Newer models with finer grid structures tend to do a better job with local conditions, but even then they can vary a lot. That said, many locals have regional-specific models that do quite well. Example, SF Bay. Others, not so, example Sea of Cortez.

Even when models agree, and you are not near particular land features, you can figure on gusts being 50% higher (roughly) than steady wind predictions. Some models allow you to see gust predictions.

What we do (before passages especially) is watch several models for our route for a few days and compare them, making our route plans accordingly. Even then, sometimes we get surprised. For example, last year we crossed from La Paz to Mazatlan right between 2 hurricanes, neither of which was predicted to head towards the Sea of Cortez, both of which took unexpected turns in that direction. Fortunately, both died rather sudden deaths and did not affect us beyond our blood pressures, and the sailing proved magnificent!

[jimbunyard]

This buoy, NDBC - Station 42058 Recent Data, located about 60 miles to the northwest of the 8:00 AM reported center,
shows steady winds of 35 knots, gusts to 47, and peak wind in the last hour of 52 knots. The anemometer height is only 15 ft, and the current sea state is 24, so chances are that these numbers may not be representative, but the wave height does seem to correlate with the wind speed.

According to the 8:00 AM EDT advisory, hurricane force winds extend only 25 miles from the center.

Provided the buoy stays on line, we should be able to watch the wind shifts and pressure fluctuations to gain a small amount of insight, as it appears that the eye is about to pass directly over it.

[barnakiel]

Originally Posted by Mirror16

I'm looking at hurricane Matthew in the Caribbean and its current highest indicated wind speed on earth.nullschool is about 50 knots. But the news and the National Weather Service reports sustained speeds of 130 knots. Why this huge discrepancy?

Mathematical models vs. sensors.

b.

[knockabout]

sometimes it takes a while to place the curser ring on the highest wind strength / and that can change fast

[UNCIVILIZED]

Originally Posted by jamhass

It depends ...

On the particular model, on the location, etc etc.

Some models do better than others in some cases, some not so. If you look at several different model results and they all agree, you can have pretty good confidence in the result. It is hard to predict which model works best under which conditions and/or locals. That is where experience helps, and makes really experienced weather gurus and navigators quite in demand.

Consider the NHC presentation of different hurricane track models. Sometimes they all agree nicely, sometimes not.

Location near land is often problematic. Local conditions impact winds and weather significantly and are not always well represented in the models. Newer models with finer grid structures tend to do a better job with local conditions, but even then they can vary a lot. That said, many locals have regional-specific models that do quite well. Example, SF Bay. Others, not so, example Sea of Cortez.

Even when models agree, and you are not near particular land features, you can figure on gusts being 50% higher (roughly) than steady wind predictions. Some models allow you to see gust predictions.

What we do (before passages especially) is watch several models for our route for a few days and compare them, making our route plans accordingly. Even then, sometimes we get surprised. For example, last year we crossed from La Paz to Mazatlan right between 2 hurricanes, neither of which was predicted to head towards the Sea of Cortez, both of which took unexpected turns in that direction. Fortunately, both died rather sudden deaths and did not affect us beyond our blood pressures, and the sailing proved magnificent!

Thanks for the mini education. Might you be kind enough to share the addresses of the models that you use, along with some of your other favorite weather links please? I'd surely appreciate seeing them firsthand, & what as yet unknown weather education they may provide.

[StuM]

Originally Posted by UNCIVILIZED

Thanks for the mini education. Might you be kind enough to share the addresses of the models that you use, along with some of your other favorite weather links please? I'd surely appreciate seeing them firsthand, & what as yet unknown weather education they may provide.

Take a look at Model Guidance

It shows lots of models, their coverage areas and a few details such as update frequency.

[wannacat]

"Knockabout" is correct, you have to be patient and also hit the right spot.
Yesterday I recorded 93 kts at one instance but then couldn't duplicate it or get higher than 90 and 91 after that.

[jamhass]

Originally Posted by StuM

Take a look at Model Guidance

It shows lots of models, their coverage areas and a few details such as update frequency.

That, plus some portals (windty for example, and Airmails Grib portal) allow you to select from a small set of models. Well worth practicing when you have broadband rather than trying them out via SSB emails!