Sailonline.org - Online Sailing

Facebook

Login

Support Sailonline

If you haven't already - join the SAILONLINE YACHT CLUB!

Please also consider making a donation - all amounts are greatly appreciated!

Tweets by @sailonline

Board » Technical Discussion » The Weather Research & Forecasting Model (WRF)

Posted by ita10267 Posts 86 Registered at 2008-12-29 10:33:09	2015-02-10 21:30:23 - Re: The Weather Research & Forecasting Model (WRF)
	so, you are referring to the grid discretization. In SOL this changes from race to race. Usually we start from 0.5x0.5 deg GRIB file from NOAA which is then interpolated to the grid of our interest. In Leg4, for the first time, we are running the ocean on a 1x1 deg grid. The "pointy" edges are caused by the interpolation algorithm (bilinear), which is the same we use to get wind in between grid points. The WRF model we are using in the BOS-NEW race is evaluated on a much denser grid. This is not an interpolation process but a complete physical evaluation. The 0.5degx0.5deg GRIBS from NOAA are taken as boundary and starting conditions then the calculations are done on a powerful computer. The result is a 0.05X0.05 deg GRIB.
Posted by A2 Posts 347 Registered at 2009-10-26 00:39:42	2015-02-10 22:28:06 - Re: The Weather Research & Forecasting Model (WRF)
	Thank you for the additional information. Your terms are more precise, I am glad you understood my explanation. None so blind
Posted by ita10267 Posts 86 Registered at 2008-12-29 10:33:09	2015-02-11 14:37:14 - Re: The Weather Research & Forecasting Model (WRF)
	A small video on high resolution weather http://youtu.be/n5ThrmvMJxA
Posted by ita10267 Posts 86 Registered at 2008-12-29 10:33:09	2015-02-13 11:14:02 - Re: The Weather Research & Forecasting Model (WRF)
	forecast for sunday morning Attachments sunday_1.png (666.8 KB)
Posted by JB Posts 129 Registered at 2013-03-07 08:54:34	2015-02-16 11:03:51 - Re: The Weather Research & Forecasting Model (WRF)
	Viva Andrea. Great expectations for this race with the new “WRF” weather forecast system and corresponding GRIB files. The first and only routing I’ve made was Friday morning with qTVlm (file“weather_225_wrfems_20150213_1030.grb”). A small thought. The GRIB files for the SWR Leg 4 are around 1,3 Mb size, for the Boston test aprox. 2 Mb. If we linearly extrapolate the file size in function of the corresponding race field’s dimensions we get a big size GRIB file. Some pre-race facts. - Calculations for routing are “heavier” as expected - the GRIB as much more and also better info; - It’s easier to obtain a geometrical stable tacking path within the qTVlm minimum 5 minutes step calculations (I could have used the 10 minute time step for getting the same results); - The qTVlm routing optimization process was quicker - the “raw path” is already very good. The race “test”. I only used the Zappe Android app (again) for running this one. With the initial WF it went smoothly and the first South tack was accomplished normally. After the second WF I couldn’t load the app as it started crashing caused by memory issues. The result was an impossibility to make the second South tack (at the time I was running aprox. at #11). I tried several tweaks on the phone, as closing all the open apps, inputting minimum setting on the SOL apps, etc., but it didn’t work, so I’ve stayed for hours without the possibility for maneuvering the boat. At the same time I was running psail on the SWR and on Huelva TR without any issues. The future trend is more power on the “mobile” and less pc/laptop-ish. Hence this “test” made using the Android app - well, the message is delivered. I totally support and expect the trend will also be more and better GRIB’s, like the ones from WRF. Andrea we go in the right direction. Thank you. Fair winds to all. Attachments Boston_routing_13.Feb.2015_1_of_2.png (1.1 MB) Boston_routing_13.Feb.2015_2_of_2.png (875.9 KB)
Posted by ita10267 Posts 86 Registered at 2008-12-29 10:33:09	2015-02-17 07:01:29 - Re: The Weather Research & Forecasting Model (WRF)
	JB, at present time we have to allow for some WX data outside the race course (the purple line). This to allow predictor to work properly in the client. In this race we have used 0.5deg from each side. In the end we had more than 30% of datapoints "useless" for the race. In an optimal condition (no points outside race area), wxfile should have been around 4MB, which travels compressed now, so perfectly usable for the "desktop" client. Regarding mobile, probably it has some problems (depending on hardware) regarding memory and cache. Probably Zappe can look into that. Finally, the WRF is defined every hour, so dataset is 3X for the same resolution compared to GFS. As outlaw commented, we should push for even higher time resolution (1 frame every 15min). In this case probably we'll have denser frames for the first 6hrs and sparser afterwards. Ciao a.
Posted by kroppyer Posts 244 Registered at 2010-02-05 15:25:55	2015-02-19 11:29:52 - Re: The Weather Research & Forecasting Model (WRF)
	Hi ita, It seems the WRF changes a lot more with each Wx update compared to the GFS. Is that just because the changes are a lot more noticeable with the Higher resolution, or also because it's a lot harder to forecast accurately on this scale? (I guess they both play a big role)
Posted by ita10267 Posts 86 Registered at 2008-12-29 10:33:09	2015-02-19 16:46:03 - Re: The Weather Research & Forecasting Model (WRF)
	In synoptic conditions, i.e. wind given by large pressure systems, forecast are simpler and more reliable (see the Bos-New race with the heavy storm). In fluky conditions, with basically no pressure gradient, wind is given by "local"effect such as cloud coverage etc. So yes, on mesoscale, weather is less "consistent". As a comparison, GFS is "averaged" over a much bigger gridsize, so all the small scale variation are zeroed and you ave a more stable but less interesting windfield. On the other hand NAM model, on a 0.1deg scale, has the same behaviour.

Page: Previous 1 2

Please login to post a reply.

Races

Next Race: 00d 00h 00m

Current Races:

Garagiste's Test Lab

The first TIMED race of the northern hemisphere Summer takes us to the Gulf of Finland on a 165nm zig-zagging course beginning and ending in the seacoast village of Haukilahti and circling the island of Suursaari. It’s only appropriate that the boat for this race is Finland’s own Club Swan 50. This is a TIMEDrace so you may RE-REGISTER HEREto try again after finishing a run. You will have 13 days and 11 hours to test your skill and decision making after the race opens.
Race #1928
INFOby brainaid.de
PARTICULARS
WX Updates:
0430 / 1030 / 1630 / 2230
Ranking:
TRQ2 - TRCH - SUPSOL - SYC
RACE CLOSE: Saturday,
26 July at 23:00 UTC
Race starts: Jul 13th 12:00 Registration will open soon

Facebook

Login

Support Sailonline

Races

SYC Ranking

Series

Mobile Client