Predictwind for the Chesapeake

[Hard A Lee]

For those using PredictWind on the Chesapeake, are there any of the four forecasts that seem better for that area than others?

Maybe one is better for wind and another for waves, etc?

Thanks!

[TheOffice]

Short answer: no. I look for consensus. I don't find the European to be terribly accurate for the Bay. I use WindAlert to see if the forecast matches reality.

[Kenomac]

I've been very disappointed with Predictwind over the past three months after having promoted it heavily on this forum. The downloading takes forever and manytimes it even times out. Has become a waste of money.

Have now switched back to using Weather4D pro and pocket gribs.

[Pete O Static]

As previously stated, I also look for consensus between the models. There seems to be only two models really, the GFS and Euro along with the Predict Wind higher resolution version of the same models, hence PWG for their higher resolution GFA and PWE for their higher resolution version of the Euro model. Supposedly, the higher res will take into account the effect of land and sea breezes, land acceleration and temperature delta closer to shore, whereas the GFA and Euro are strictly gradient based. I only have the standard subscription, not pro and so far, I have seen no evidence of any land effect taken into account.

One thing I have found to be a frustration with "Route Planning" is that all computed data is given as a "True wind angle" and speed. Even with the boat's polars uploaded to the server. Therefore, it will calculate a route for me that at first glance looks quite benign. A nice reach with a 60 degree TWA, a wind speed of 12 knots true, giving a boat speed of 8 knots. Because the wind angle is greater than 55 degrees, it declares the leg as a "reach". Sounds great but when you actually convert all that to apparent, which is what the boat will "actually experience" under sail, it is in fact a blasting beat to weather.

I have not been able to find any setting that will fix this. I have a question in to tech support because it could very well be operator error. If anyone has any input, I would be interested to hear.

[Hard A Lee]

Thanks! I didn't realize that PWE and PWG were just enhanced by PredictWind (PW) but now it totally makes sense...

[thinwater]

I use Windfinder, Windy TV, and NOAA, and compare the three.

But particularity on lighter wind days, when wind is strongly influenced by local thermals, nothing beats observation of cloud patterns, the surface of the water, and just watching how things are developing. Just the unpredictable movements of cloud bands and their effects on thermals make meaningful forecasting impractical. You've just got to look, and when the prediction is off, figure out where the wind went.

[nickolson]

Originally Posted by Hard A Lee

Thanks! I didn't realize that PWE and PWG were just enhanced by PredictWind (PW) but now it totally makes sense...

That is not actually true. The PWG and PWE are a full model run of their own model from initial conditions point.
Have a look on their website for a full explanation.

[Auspicious]

There are two problems with Predictwind that are shared by all grib-based products such as Windy. First they don't show the effects of discontinuities like fronts. Second they don't show the topographic effects that the land around estuaries bring. The best product for the Chesapeake is CBOFS https://tidesandcurrents.noaa.gov/of...=wind_forecast . They have a wave model as well as the wind model I linked.

[Suijin]

I use buoyweather.com over PredictWind. While it's not graphical, it's extremely accurate. Also provides dew point, RH, lifted index and convective potential which is very useful in interpreting what's going to happen in the forecast at the local level.

For the Chesapeake, I generally use Windfinder as there are lots of stations in the Bay. That said, the bay being so small with so many rivers there is a lot of land effect...both channelling and convective...that has a profound affect on the overall forecast. Factoring those into your plans pays dividends.

[justakiss]

Originally Posted by TheOffice

Short answer: no. I look for consensus. I don't find the European to be terribly accurate for the Bay. I use WindAlert to see if the forecast matches reality.

We use WindAlert as well since it does ping the various buoys in the Chesapeake Bay. We usually look at markers out in the middle of the bay under the assumption that they're most likely to be free of land effect.

Thanks for mentioning other apps/ sites though, as we will check those out as well.

[nickolson]

Originally Posted by Auspicious

There are two problems with Predictwind that are shared by all grib-based products such as Windy. First they don't show the effects of discontinuities like fronts. Second they don't show the topographic effects that the land around estuaries bring. The best product for the Chesapeake is CBOFS https://tidesandcurrents.noaa.gov/of...=wind_forecast . They have a wave model as well as the wind model I linked.

Predictwind has 1km res modelling covering most of Chesapeake, so does take in to account land effects in the PWE and PWG models.
The map in the link is terrible to read

[nickolson]

Originally Posted by Hard A Lee

For those using PredictWind on the Chesapeake, are there any of the four forecasts that seem better for that area than others?

Maybe one is better for wind and another for waves, etc?

Thanks!

I work for them so I can answer the actual post question for you. The PWE and PWG will generally be better as they are a 1km resolution for this location, so a lot more reflective of the confined area, the 1km modelling is for 36 hours from forecast run, then it switches of to the 8km res for 7 days. The ECMWF is at 8km resolution so wont get the land effects as much as the PW models. The GFS is 27km resolution so wont cover the land effects, or more importantly it does not take in to account the water, wind is 30% less on land vs water as a rule of thumb in weather models.
Wave models fro the PWG and PWE are at 50km resolution so not overly helpful in this body of water, the ECMWF would have the best chance but is still at 14km resolution. I would imagine it is only wind wave with the real factor being when you have wind vs tide.

[Auspicious]

Originally Posted by nickolson

Predictwind has 1km res modelling covering most of Chesapeake, so does take in to account land effects in the PWE and PWG models.

I think you are confusing precision and accuracy. Wrong data at finer granularity is still wrong. If a cold front is crossing the Bay then all the grib-based products will be wrong.

That's quite different of course from the "nowcast" data that comes from measured information (buoys and shore stations).

[nickolson]

Originally Posted by Auspicious

I think you are confusing precision and accuracy. Wrong data at finer granularity is still wrong. If a cold front is crossing the Bay then all the grib-based products will be wrong.

That's quite different of course from the "nowcast" data that comes from measured information (buoys and shore stations).

Not sure why you think that fronts are not shown in weather models? They are.
Timing of fronts in models is pretty good. The difference between a forecast and a nowcast is a forecast is for planning what is coming. Nowcast tells you what is there "now" of course, which can help to validate a forecast.

[mdelatte]

Hello,

have you also compared with NAM, HRRR, RAP, NAM Nest Conus forecasts in the Chesapeake Bay ? they are probably more accurate than GFS or european model as they are hi-res models from NOAA itself.

Those models are amongst others solutions available with Squid.