Hurricane Lorenzo

Dave100456 said:
Why is it when I go to
https://www.nhc.noaa.gov

to track Lorenzo, I'm told max sustained winds of 120knts, however, when I go to

https://www.windy.com/?19.878,-44.429,8,i:pressure,m:edSaeXO

I struggle to see anything much above 69knts?

Thanks
Click to expand...

The models used for large scale forecasts like gfs and Ecmwf are designed to predict winds over relatively large areas but the hurricane forecasts have different models and predict max winds at the pressure wall to the bottom of the hurricane. So as I believe windy will be accurate when looking out 50km or so away from eye but closer then more specific hurricane models will predict the speeds.

Just don’t get near the eye!
 
Dave100456 said:
Why is it when I go to
https://www.nhc.noaa.gov

to track Lorenzo, I'm told max sustained winds of 120knts, however, when I go to

https://www.windy.com/?19.878,-44.429,8,i:pressure,m:edSaeXO

I struggle to see anything much above 69knts?

Thanks
Click to expand...

It's mainly because of GFS (and ecmwf to a lesser extent) model physics: hurricanes are basically enormous heat pumps where convection plays a fundamental role, convection and related processes are rather poorly modeled by those type of models, as often integrated as parameter.

Read a few of the NHC "Discussions" to have an idea of the additional data forecaster use to integrate and "tweak" the various models output to provide a more accurate forecast.
Actual wind speeds are usually provided by scatterometry satellites.
 
Northcave and Roberto have got it broadly right. I hope that does not sound too condescending and would like to add a few comments. All numerical weather models use the laws of physics. Limitations to what models can do arise from several factors. The use of a finite grid determines the size of detail tat can be handled by models to about 5 grid lengths. The two best NWP global models, the ECMWF and UK, use grid lengths of 0.1 degree lat/long so can, at best only, predict detail of about 50 km size. Secondly, there are always uncertainties due to uncertainties in the initial data analysis. Other limitations are due to the difficulties in calculating how heat is transferred throughout the atmosphere.

As Roberto says, convection plays a large part on tropical storm development. Until recently, global models were "hydrostatic". that means that convection effects were dealt with crudely. Now, ECMWF, UK, DWD ICON and, perhaps, the GFS are all “non-hydrostatic." Modellers say that 10 km is about the limit of grid length below which models can be non-hydrostatic. So, even though some global models are non-hydro9static, I doubt that convection effects are realistically calculated
L
https://www.bas.ac.uk/data/our-data/publication/ice-cores-and-climate-change/ lists the various model inputs to the NHC and, as you will see, meteorology is truly international. In addition to global model data, NHC also uses finer scale , non-hydrostatic models fed by input from global models and more detailed data nalyses. The global models are used for the longer term evolution and track prediction. Fine scale models help with short term detail. However, in the final analysis, NHC, forecasts are a combination of model outputs plus real time data from aircraft and whatever else is available. The worded output and hurricane track predictions that you see are a man-machine mix. Even the fine scale models will under-estimate the strongest winds. Getting back to the OP, it is inevitable that strong winds generally will be under-predicted. Given the nature of hurricanes, that will be even more so.
 
Last edited:
Thanks Frank. Very interesting.

For the same reasons, I might also add, that this is why we often find wind speed greater around coastal areas. Often when windy says 20kts we might be getting 25-30. The models can’t accommodate for acceleration zones and local
Pressure differences.

Offshore however I find the models very accurate in terms of average wind speed.

However for the UK the paragliding RASP model is highly accurate. Go to the link below and selecting BL average wind. Compare it to the wind model on windy.

http://rasp.stratus.org.uk/app/viewer.php

A22E07BC-30B4-4EA3-B0BE-B0AAAAE3462B.jpg
 
Coming our way? Lookee here....

https://www.nhc.noaa.gov/graphics_at3.shtml?start#contents

48807388148_7c79b6dae7_b.jpg
 
Last edited:
northcave said:
Thanks Frank. Very interesting.

For the same reasons, I might also add, that this is why we often find wind speed greater around coastal areas. Often when windy says 20kts we might be getting 25-30. The models can’t accommodate for acceleration zones and local
Pressure differences.

Offshore however I find the models very accurate in terms of average wind speed.

However for the UK the paragliding RASP model is highly accurate. Go to the link below and selecting BL average wind. Compare it to the wind model on windy.

http://rasp.stratus.org.uk/app/viewer.php

View attachment 80657
Click to expand...

Agree about average wind forecasts over the sea. They are generally pretty good. I always assume that I will get spells about one force stronger and one force less. Occasionally, they can still go wrong. A couple of years ago on passage Ile de Brehat to St Peter Port we were expecting F5-6, possibly a touch of 7. From west of Roches Douvres to St Martins we had a sustained F8-9.

The kind of detailed forecasts that you show are alway fascinating both for the detail that they show and what they do not show. Most “unofficial” models that I see have unnaturally uniform wind patterns. This one looks better than most. I find it galling that we cannot see the Met Office output from their detailed model. It really should be the best given their modelling capability and data analysis schemes. You can see their 4km model but only with a degraded resolution and rather too late to be useful via Jersey Met for the English Channel. Of course, one problem with such detailed models is that the details predicted have such short lifetimes. For use such as gliding they are clearly useful. As a Cruising sailor, I do not find them any more help than global models. They may be indicative but not better in a deterministic sense,
 
Last edited:
A freak hurricane making landfall tomorrow on the Azores islands, probably hitting Ireland by the end of the week! I believe it's the strongest most Easterly hurricane ever recorded in the Atlantic by a long margin. Pretty scary! Seems like it's going to be affecting a lot of sailors! :/
 
Interestingly its drawing a big influx of air across the Canaries from Africa and giving shore temperatures of up to 35c.

Was sailing out of Puerto Calero on the east coast of Lanzarote this morning and it felt a bit like a hairdryer.

Best of luck to all those on the beautiful Azorean Island of Flores, they look like taking a pasting.
 
Hurricane Roberto - impacting the western Azores about now, with huge seas.

NOAA are now 'suggesting' this disturbance will likely hook right and plaster the west of Ireland..... then our own north-west. The hurricane winds may have dissipated by then... maybe... but the hurricane's rainfall won't.

And there are hints that the path could be.... might be.... a bit further south. As always, a published forecast is but one of a number of statistical possibilities the big computers churn out.

See here: Advisory 38: 5:00pm AST Tue Oct 01
 
zoidberg said:
Hurricane Roberto - impacting the western Azores about now, with huge seas.

NOAA are now 'suggesting' this disturbance will likely hook right and plaster the west of Ireland..... then our own north-west. The hurricane winds may have dissipated by then... maybe... but the hurricane's rainfall won't.

And there are hints that the path could be.... might be.... a bit further south. As always, a published forecast is but one of a number of statistical possibilities the big computers churn out.

See here: Advisory 38: 5:00pm AST Tue Oct 01
Click to expand...

There is always some uncertainty in any computer model forecast usually in the detail, occasionally, on the large scale. That is so for several reasons. Forecasters who have to present forecasts to the public and warnings to authorities have to make decisions on how to interpret the various model outputs available including ensemble forecasts. A few days ago they were saying that there seemed to be two posable tracks for Lorenzo: one towards SW U.K. and one towards the NW. Later forecasts have come down to the latter. In my time as a senior forecaster at Bracknell, it was my personal policy to say what I knew - or thought that I knew - and always encouraged other forecasters using my guidance to do likewise. In other words, if you are not reasonably sure then you should say so.

There have been massive improvement in forecasting over the 60+ years of my direct experience. The ability to predict change of type of weather situation, the ability to predict those horrendously difficult occasions when rain turns to snow, amounts of heavy rain or snow , outlooks for a week ahead etc. But, inevitably, sometimes it will still go wrong. It is a measure of the improvement in weather forecasting that it comes as a surprise when it is badly wrong. One example was my fairly recent experience of being out in a F8/9 when the forecasts wee saying F5/6 with a hint from one centre out of three of F7. Such occasions are rare but, inevitably, will occur from time to time.
 
You must log in or register to reply here.

Share this page

Top