Received: from hplms26.hpl.hp.com by opus.hpl.hp.com with SMTP (1.37.109.8/15.5+ECS 3.3+HPL1.1) id AA06284; Wed, 28 Sep 1994 12:17:47 -0700 Return-Path: <WindFinder@aol.com-DeleteThis> Received: from mail02.prod.aol.net by hplms26.hpl.hp.com with SMTP (1.36.108.4/15.5+ECS 3.3+HPL1.1S) id AA14775; Wed, 28 Sep 1994 12:18:24 -0700 Received: by mail02.prod.aol.net (1.38.193.5/16.2) id AA07467; Wed, 28 Sep 1994 15:14:33 -0400 From: WindFinder@aol.com-DeleteThis X-Mailer: America Online Mailer Sender: "WindFinder" <WindFinder@aol.com-DeleteThis> Message-Id: <9409281514.tn90782@aol.com-DeleteThis> To: wind_talk@opus.hpl.hp.com-DeleteThis Date: Wed, 28 Sep 94 15:14:17 EDT Subject: Re: Windsight forecasts stats
Ken Poulton asked:
"Mike, do you have a comparison of Windsight forecasts to actual winds?
I made this comparison for the NWS terminal forecasts for SFO and the
correlation with winds recorded at SFO each afternoon. The correlation
is a rather dismal .46. I'd like to know whether Windsight is hitting
better than that.
Clearly there is more value to Windsight than just forecasts, but this
is the kind of thing that can and should be quantified and tracked to
gauge and improve the accuracy of the service."
Ken
All of our forecasts and corrected actual wind from Penninsula sensors from
March 1, 94 to present is in our database. We still have to quantify and
enter subjective on-site verbal reports from Jeff at Candlestick, Dennis at
Coyote and Bill and Lisa at 3rd. (unfortuntly we do not have recordings of
on-site data for all days) Once the sailing season ends and I return to just
being a college professor we expect to spend the winter doing an analysis of
all this data. Casual inspection of the data and feedback from users seems to
indicate that that we are doing much better than NOAA. especially on those
days when they forecast 15 to 25 and in reality the wind in the Penninsula is
non-existent. I will put summary of results on the net in spring.
Unlike NOAA Windsights two meteorologists are making forecast specific
forecasts for17 Bay Area sailing sites. When NOAA forecasts 25 knots for the
Bay Area and some sites, say 3rd Ave., are dead while others like Berkeley
are cranking they could say their forecast was "accurate" overall. When we
blow a site forecast it is very conspicious and we work hard to figure out
how to improve. Incidentally we have updates plotted every 60 seconds from
most of our sensors. When it becomes clear that reality and our forecast do
not jibe we or our on-site sailors update the forecast.
For example last monday our noon forecast for Larkspur was 5 to 15. It was
wrong! At 1:16 p.m. our wind histogram shot upward as winds built rapidly to
20 to 25 knots. We promptly made an update report and made a note that when
there is a cut-off low SW of the Bay, fog on the Marin headlands but not at
Crissy and vorticity in the jetstream and SW winds develop atop Mt. Tam
Larkspur can get stronger winds than the local pressure gradint would
indicate. Hopefully next time this occurs we won't get caught with our pants
down.
One hint for those who like to do their own Penninsula forecasting. Windsight
has the bouy wind and swell reports updated HOURLY on option 59. Always in
the spring and often in the summer these ocean winds are as important as the
PRESSURE GRADIENT in determining Penninsula wind. We are currently working on
a Windsight Users Manual to help sailors understand and find bay wind.
( Ken: When the guys in Hood River have time to pull the 3rd. Ave. sensor
data out of the data base in tabular form I will send it to you. The laser
idea sounds interesting...it this just a concept or could a working prototype
be made easily? This would get around the need for a cellular/solar system.
Putting some Windsight data on the net at a lower cost because of lower phone
overhead is an interesting idea. But how could we charge anything for it to
cover our costs?)
Mike Godsey
This archive was generated by hypermail 2b29 : Mon Dec 10 2001 - 02:28:12 PST