WRF 9 km / WRF 27 km resolution

zidewayz

I am hoping someone could help me with the following question. I use windguru to view forecast for wind. What is WRF 9KM resolution and WRF 27KM resolution and which would be more accurate for a wind forecast for a particular location? Is it that there is more input data computed in the 27KM than the 9KM forecast giving the 27KM a more accurate and stable forecast? Or am I looking at this the wrong way?
WRF 9 km

Forecast based on WRF model with 9 km resolution. Updates 4 times per day and offers forecast for 78 hours. Map shows the area covered by the model.


More info about weather models used by windguru


WRF 27 km

Forecast based on WRF model with 27 km resolution. Updates 4 times per day and offers forecast for 78 hours. Map shows the area covered by the model.


More info about weather models used by windguru

Harps

9km means that every cell computed by the model is 9km x 9km, likewise for the 27km. The higher the resolution (smaller grid), then theoretically the more accurate the forecast should be so I'd go with the 9km one first.

It doesn't always work that way though and sometimes the higher resolution one can 'over think' the situation while the lower resolution gives a less detailed but overall more accurate forecast so I'd still keep an eye on both.

Can't say I've ever used them for forecasting wind, is there ever much of a difference between the two?

zidewayz

Harps wrote: »

9km means that every cell computed by the model is 9km x 9km, likewise for the 27km. The higher the resolution (smaller grid), then theoretically the more accurate the forecast should be so I'd go with the 9km one first.

It doesn't always work that way though and sometimes the higher resolution one can 'over think' the situation while the lower resolution gives a less detailed but overall more accurate forecast so I'd still keep an eye on both.

Can't say I've ever used them for forecasting wind, is there ever much of a difference between the two?

[Quote=Harps;

I might not be thinking right here or thinking too hard. What the 9km/27km resolution physically/spacially represents and how it's forecast data is calculated with reference to the forecasted location is confusing me. Are they just a zoom into the forcasted data around the location?

A forecast can be made for a particular location. How do these resolutions differ in computing the forecast for the location?

The more I think into this the more I get confused so questions may become more irrelevant.

Harps

Just going by what I know about CFD models here, I'd assume weather models are much the same so someone can correct me if I'm wrong..

For the 9km WRF, think of the atmosphere as a grid of 9x9km squares over the entire area in the maps you posted above, basically the model mathematically solves the conditions in every one of these 9x9 squares (and its interactions with the squares around it) separately for the next 78 hours. If you're looking for a forecast for Dublin, it'll just reference whatever square Dublin is in and whatever the conditions in that square are at the time. 9x9 is reasonably small so you'll get a precise forecast for Dublin City.

For a 27km model though, the grid size is a lot bigger at 27x27km so the forecast for Dublin might be exactly the same as Drogheda as they're both in the same square. Whereas the 9x9km model might be showing rain in Dublin and dry in Drogheda, the 27x27km one will just show one weather type for both.

Not sure if that makes it any clearer, not quite sure what I'm trying to say myself!

zidewayz

Harps wrote: »

Just going by what I know about CFD models here, I'd assume weather models are much the same so someone can correct me if I'm wrong..

For the 9km WRF, think of the atmosphere as a grid of 9x9km squares over the entire area in the maps you posted above, basically the model mathematically solves the conditions in every one of these 9x9 squares (and its interactions with the squares around it) separately for the next 78 hours. If you're looking for a forecast for Dublin, it'll just reference whatever square Dublin is in and whatever the conditions in that square are at the time. 9x9 is reasonably small so you'll get a precise forecast for Dublin City.

For a 27km model though, the grid size is a lot bigger at 27x27km so the forecast for Dublin might be exactly the same as Drogheda as they're both in the same square. Whereas the 9x9km model might be showing rain in Dublin and dry in Drogheda, the 27x27km one will just show one weather type for both.

Not sure if that makes it any clearer, not quite sure what I'm trying to say myself!

Your explanation was very clear. I think I was misunderstanding that the grids forecast was set and defined for each location within the grid, thinking that the grid helped calculate the forecast for each location within that grid depending on its location within the grid.

Sorry I'l stop there..I think I have a tendency to over complicate things. I have checked the 27km for two locations within Dublin and they are indeed identical while there 9km is slightly different. Both resolutions do be pretty accurate when I use them. Will take more notice tommorrow as heading out on the water to take advantage of the high winds in the morn. Thanks for your reply harps.

Capillatus

When looking at wind, I would go with the highest resolution possible. Basically because the grid points are much more dense, and so the model has a much better idea of the terrain, which is of course important when it comes to wind.
And of course as Harps pointed out, higher resolution gives you much better spatial presentation of the parameter.

27km resolution is very nice for features on the meso-alpha scale, while higer resolution like 9km or 4km or lower, is recommended for the meso-beta and meso-gamma scale features.

Harps

Yeah that's pretty much what I was trying to say, if you're just looking for a general forecast along the lines of 'will it be windy?' then the 27km one is fine but if you're looking for a precise forecast for your local area then the smaller the resolution the better.

I think the YRNO model on weatheronline is something like 1.5km so that might be another one you'd want to keep an eye on, haven't used it much for wind but I find it to be the most reliable model around for temperature. On the phone now so can't link but I'll stick it up later if I get the chance!

dloob

I use it for windsurfing sometimes.
For coastal spots the 27km tends to give lower windspeeds than the 9km.
Understandable as the 27 will likely contain more inland areas which will bring the average down.

zidewayz

Capillatus wrote: »

When looking at wind, I would go with the highest resolution possible. Basically because the grid points are much more dense, and so the model has a much better idea of the terrain, which is of course important when it comes to wind.
And of course as Harps pointed out, higher resolution gives you much better spatial presentation of the parameter.

27km resolution is very nice for features on the meso-alpha scale, while higer resolution like 9km or 4km or lower, is recommended for the meso-beta and meso-gamma scale features.

ooooh.... plenty of jargon there for me to start googling.

zidewayz

Harps wrote: »

Yeah that's pretty much what I was trying to say, if you're just looking for a general forecast along the lines of 'will it be windy?' then the 27km one is fine but if you're looking for a precise forecast for your local area then the smaller the resolution the better.

I think the YRNO model on weatheronline is something like 1.5km so that might be another one you'd want to keep an eye on, haven't used it much for wind but I find it to be the most reliable model around for temperature. On the phone now so can't link but I'll stick it up later if I get the chance!

Bookmark going into the iphone for the YRNO model.

zidewayz

dloob wrote: »

I use it for windsurfing sometimes.
For coastal spots the 27km tends to give lower windspeeds than the 9km.
Understandable as the 27 will likely contain more inland areas which will bring the average down.

Windsurfing is indeed the activity for opening this thread and question. Both WRF 9 km and WRF 27 km do be fairly accurate but I do always give met eireann, windguru and Dub airport and various weather stations along the coast a browse before deciding what time is best to head out.

What I found interesting in todays wind forecast from WRF 9 km was that there was going to be a sudden lull in wind conditions while winds swung around from SE to W around 13-15hrs. ME forecast did not forecast this lull, prob because their forecasts are gennerally directed to sea faring vessels, while the WRF 9 km was nearly spot on. The lull was actually between 14-16hrs. If there was another few hours daylight I would of actually got wet today.

If we don't get any snow this winter I will be happy enough if the atlantic batters us. Enough of the recent cold days frosty nights.;)

Harps

Heres the yrno link for wind if you didn't find it

http://www.weatheronline.co.uk/cgi-bin/expertcharts?LANG=en&MENU=0000000000&CONT=ukuk&MODELL=yrno&MODELLTYP=1&BASE=-&VAR=uv10&HH=6&LOOP=1&ZOOM=1&ARCHIV=0&RES=0&WMO=

Forecast here for wave heights and other sea related stuff might be useful as well

http://www.weatheronline.co.uk/marine/weather?LEVEL=3&LANG=en&WIND=g049&CEL=C&SI=mph&MENU=0&MN=nae&MODELLTYP=uv10&TIME=0