|
Author
|
Topic: HRRR as a racing tool
|
Chgodave
|
posted July 06, 2023 09:16 PM
Hello. I would like to see how racers are using HRRR for navigation and strategy. I have never attended a pre-race weather briefing where HRRR has been treated differently from the other models. I'm thinking that some consideration must be given to the fact that is updated hourly while other models are updated less frequently. Also when underway, assuming you have access to data, how often should you be looking at HRRR and how far out distance wise? I have totally given up on the wind fields provided by Sirius. They may shed light on the big picture but are not reliable for decision making during a race. I'm sure I'm leaving something out, but I'd be grateful to hear what others are thinking. Thanks
From: Chicago
|
|
David Burch
|
posted July 07, 2023 08:10 AM
There is a usually a huge improvement for the starting days of a race or cruise using a regional model such as HRRR compared to a global model such as GFS. You can test this by choosing a buoy or two near shore that has live wind reports and then just compare the GFS to a regional model. qtVlm will do this comparison almost automatically. Let us know if that procedure is not familiar to you. See also other ways to evaluate a forecast.
The regional models are updated hourly and have higher resolution, plus they account in more detail to the effects of terrain.
The weather trainer has a summary of all models in G282.
A routing program such as qtVlm lets you load both the regional and global models and then it will use the one with higher resolution till that data runs out then switch to the global.
It is more work however to evaluate which of the local models might be best at the moment. You have HRRR, NBM Conus, and NDFD. Also compare NAM conus.
With that said, we can stand by to hear what our lead weather instructor Dave W has to say on this because he uses the regional forecasts to set up the racing buoys in his local waters every week and thus has good experience with this. He is also located (in Port Townsend) at a tough location to forecast being just around the corner from a large open channel.
From: Starpath, Seattle, WA
|
|
Dave Wilkinson
|
posted July 07, 2023 04:07 PM
Well there is more to say on using high resolution weather models in support of boat racing, but here are a few thoughts based mostly on local racing around the buoys here in Port Townsend.
One feature handy feature of qtVlm is the ability to compare models, e.g. HRRR and NBM Conus, by overlaying the two, both in the map display and in the meteogram. Then when you go out on the water, or even before if you have live data that is representative of your area, begin to verify the model that is working best.
If a model forecast appears to be “wrong”, look to see if the wind speed / direction might be OK but simple early or late in timing. This can help looking forward to the rest of the race. For example, the marine push has already arrived, so expect increased winds that are backed 20* for the rest of the evening.
One seeming benefit of viewing model data in a charting program is that potential courses can be set up. Then screen shots can be pasted into an email to the rest of the race committee. With the charting program running on a laptop and a GPS input, the helm knows where to go to drop off marks and where to set up the start line. This has been a big time saver and communication aid for the crew. Everyone is looking at the same model(s), can help evaluate the forecast vs. what is observed, and be prepared for course changes if needed.
Challenges remain even with high resolution models.
Afternoon and evening races can be subject to diurnal effects, such as a sea breeze, that are near their peak and then can fall off quickly. The timing of this will change through a racing season as days get longer in the spring and shorter after summer solstice.
Racing in areas with cold water compared to air temperature can also cause rapid onset of the mixed layer collapsing. Cold water sets up a surface based temperature inversion, creating a marine layer next to the water surface, which can isolate the surface wind from the pressure gradient flow above the inversion; which is what the models are largely based on. It is hard to tell if the high resolution models have sufficient vertical resolution to capture to early stages of the inversion.
Horizontal model spatial resolution still may not be small enough for some important terrain features. Even at 1.3 km resolution the model may not "see" a valley or bluff. Ideally several grid points are needed for the model to fully account for the changes. This leaves terrain features on the order of 1 nmi at the lower end for models to discern.
However, on the positive side, there are indications the HRRR model does allow the wind to veer as it leaves the shore and progresses over the water. It will also create corner effects downwind of large low points of land and show wind speed maxima at the outlet end of channels. A handy way to evaluate small changes in wind direction is to lay out a temporary course which is simply a straight line. Then compare the wind barbs against that line for small variations.
Input data is still vastly lacking. A shore-side weather station observations is what the model ingest for computations, but that one station is likely not representative of the wind away from shore and over the race course. To enhance the amount of data available, the Seattle NWS office monitors social media on Twitter, #wawx. They welcome a photo and an observations of wind speed/direction, wave height, sky cover, visibility. They have the ability to modify the NDFD grids with that additional data. This is an option for sailors fill in the gaps between weather buoys and help improve the forecasts.
Even with modern hi-resolution models, there is room to develop local knowledge for both your local winds and the models that work best under various conditions. One example that seems to be true here is that when wind speeds are light, say less than 5 kt, expect the forecast direction to be undependable. The practical adaptation has been to set up additional marks and position the committee boat / start line so that alternative courses can be set up through the afternoon / evening.
From: Starpath, Port Townsend WA
|
|
Chgodave
|
posted July 10, 2023 11:31 PM
All good stuff. I have a perfect race to look at land effects coming up this weekend. When I'm not at the helm I will pay attention to what you have discussed. The race includes going along the Lake Michigan shoreline through a large inlet into a fairly large bay. The winds in the inlet and bay are always challenging. Also, occasionally there is a wind driven current to deal with.
On two other fronts, I need to revisit David B's series on QtVlm as I work mostly with Expedition. I am pretty comfortable with XYGrib and lean on it when underway, but it has no routing capabilities. Also my 2007 version of the Weather Trainer no longer works. So I guess its time to go to the cloud version.
From: Chicago
|
|
David Burch
|
posted July 11, 2023 09:19 AM
May i encourage you to extend your webcard. That will get you back into the weather course where we can go over more discussion of local models, and your race in particular, and it will turn on your weather trainer that has a lot of valuable information in it.
See extension options at www.starpath.com/webcards .
From: Starpath, Seattle, WA
|
|
|
