Technical Information on Wendy Windblows
What is Wendy?
The Wendy network of talking weather stations was
conceived with the idea that users could, purely by telephone,
get an accurate idea of the ACTUAL weather
conditions, RIGHT NOW, at a sporting site before deciding
to go there.
Not a Forecast
Note that this is NOT a
forecast. It is a "Now-cast", or even
an "After-cast". It is what you would
get if you had a convenient friend (with a mobile phone) standing
at the site, and telling you what the weather's doing.
The stations were originally set up with
Hang Gliders and Paragliders in
mind, as Rod Buck, her inventor, has been a Hang Glider pilot for
20+ years, and got fed up with driving 60 or so miles to fly,
only to find that, when he got to the site, the weather was
totally different to the forecast!
Although Wendy started life as a service for Hang and Paragliders, the Hilltop stations are now used by many other aviation types,
such as Sailplane Pilots,
Balloonists, Aeromodellers, Light Aircraft Pilots,
etc. Indeed, some Hill-Walkers also use the system, as do Farmers, who need to know the wind conditions before spraying crops,
and so on.
Sailors, Windsurfers, Kite Buggies etc
The system originally expanded on inland hilltop sites, as other Hang Gliding clubs requested a Wendy on their
local hills - but, in February 2000, Rod started erecting a chain of stations at coastal locations to service the
needs of all the marine and beach based sports. Many Paragliders, who knew about the inland stations, were also
Windsurfers, and suggested the coastal
chain. In the year 2000, seven coastal stations were added, 6 along the
South Coast, from Worthing
in the east, through Hayling Island,
and Portland Bill, to Exmouth in the west.
East-coast stations at Cromer near the Wash,
and Bridlington, were added, and
West-coast ones at Marazion (Penzance), Weston-Super-Mare,
and Thurstaston (West Kirby, Wirral)
to be used by the many
kite buggies and land-yachts and other beach sports that
use these areas. Rod hopes to add many more Coastal sites in future.
Subscribers Versus Pay-as-you-Call Users
Each Wendy station has a normal phone line connected to it. Subscribers to the
Wendy service ring this actual number, so the call is a normal BT -01 or -02 code,
and the call should cost around 5p or less, depending on which carrier you use. BT
has a minimum call charge of around 5p. Others phone carriers do not, and charge less
per minute. It is possible for a Subscriber to get a Wendy report for less than 3p.
So, as the subscriber is ringing the station direct, he gets the most up-to-date data,
which is never more than 30 seconds out-of-date, at a call cost of 5p or less.
These stations run 24 hours a day, and subscribers can call them anytime.
The Subscriber's access to the stations is controlled by a PIN-number system,
so only registered subscribers can get in to these lines. As they are ringing
direct to the site itself, that data is the most up-to-date of all.
Pay-as-you-Call Users & Web Data
Now, we need to get the station data back from the actual station site to both our
public lines, and our website here.
So, our main computer here in the office rings each station
periodically, and uplifts the data digitally. This occurs 24
hours a day. The interval varies with time of day and traffic,
but is currently about every half hour.
The data thus gathered is then placed on our website data page here, and also on our bank of
pay-as-you-call lines, which anyone can use.
Thus, the data on the web & pay-as-you-call lines
may be half an hour or so out-of-date. Or not, depending
on what time you ring in the uplift cycle. You'll
be told the time the data was taken on either the
pay-as-you-call line, or the website, so you can
tell how recent it is.
These technical notes are aimed principally at Hang Glider
and Paraglider pilots, who make up the bulk of the users of the Inland
stations. However, the information will be useful to all users, not matter
what their interest.
Differences between Met Office Forecasts & Wendy
Subscriber's Wendy data is updated every 30 seconds,
and even pay-as-you-call & web data is updated every
30 mins or so - so when you ring a Wendy station,
the information is pretty recent. Not always so
with Met Office data, I fear.
The Inland station data is also, as I've said, from hilltop
sites - Met Office sensors are usually in the flatlands, at
military airfields, and civilian airports. The weather in the
hills (particularly in plateaus, like the Peak District) is often
radically different from the flatland data. The
Wendy station at Shining Tor,
near Buxton, is only a few miles from Manchester Airport, but 1700 ft above Sea Level.
The difference between the Wendy data and the Airport readings are often
Data Collection and Processing
The wind data is gathered in a way which maximises
the meaningfulness from a pilots standpoint. Firstly,
pilots are not interested in short-term gusts,
so much, as in "thermal-induced" alterations in
Therefore, the pulses from the anemometer are integrated over a 30-second period.
This is long enough to average out short gusts, but long enough to
detect the increased wind caused by a thermal passing through the site.
As a thermal approaches from the front of the hill,
it causes a drop in windspeed for a minute or
two, and then, as it passes over, the "suck" of the
thermal is ADDED to the windspeed, causing an
increase in wind lasting maybe 1-5 minutes.
It is this alteration that the users are keen to detect, so the sample
period of 30 seconds is ideally suited to sort out thermal-induced changes
from simple turbulent gusts.
The station keeps 60 such 30-second "snapshots"
in its memory at any one time - that is, 30 minutes
worth. Each time it takes a new snapshot of conditions,
it over-writes the data received 30 mins & 30 seconds
ago, and looks back over the last 30 minutes to
find the lowest, highest, and average windspeed.
Thus, what you are getting is a 30-minute "rolling
window" that advances through the
day in 30-second jumps. Why 30 minutes? Because the
average cumulus cloud (of interest to pilots!)
takes 20 minutes-odd to form, pass over the site,
and decay. Keeping data for 30 minutes enables
"cloud cycle" to be more than covered.
Remember, cumulus clouds have lift under them - they suck upwards.
At least, most of 'em do. Therefore, when one is in front of the hill, it causes
a drop in windspeed, as the suction is subtracted from the general wind.
When the cloud passes over the back of the site, the
suction is added to the windspeed.Therefore, if
you take the min windspeed away from the max windspeed
in the last 30 minutes, and divide by 2, you get
the amount of "suck" generated by the clouds
It also gives you a good idea how turbulent conditions are: A day where the windspeed
is varying from 14-18mph, average 16mph is pretty smooth.
The same average speed of 16mph, but with the wind varying from 5-30mph
is a very different proposition, if you're planning to fly in it!
So, the station gives you the max, min, and average windspeed, and average direction
in the last 30 minutes.
It then gives you the average speed and direction 30, 60, and 90 minutes ago.
This is useful for forward projection. Is the wind getting up, dropping off,
backing, veering? The history thus revealed is very useful for tracking
the passage of fronts in real-time.
Front Tracking in Real Time
A wind that has switched from WSW to NW, and suddenly increased, in the last 30 minutes,
say, with a corresponding drop in temperature, some rainfall
perhaps, and a change in sky conditions, surely indicates that a cold front
has just passed through.
Similarly, as wind that is backing from W to SSW and steadily increasing,
with a darkening sky, but mild temperatures, indicates an approaching warm front.
Sky State & Cloud Conditions
This brings us to one of the sneakiest things about the Wendy stations:
their assessment of cloud conditions.
Each station contains an IR photometer, which measures the intensity of Infrared light
in the water-absorption band. Once the sun is, say, 10 degrees above the horizon,
the IR light is really only attenuated (cut down) by water between the sun and the station. That is, cloud.
- The thicker the cloud, the more it cuts down the IR light.
- Light cirrus or haze will reduce the normal blue-sky 100% reading down to maybe 90%.
- Light cumulus, may reduce the light level to to 70%.
- Normal mid-grey cauliflower cumulus reduce it to 50% or so.
- Heavy dark cumulus or cu-nim reduce it to 25%
- and thick stratus to maybe 8-10%
- Overdeveloping cu's and thunderheads reduce it to zero
Now, the same system of 30-second "snapshots" is
used. The station takes a snapshot of the IR sky
every 30 seconds. 60 readings (30 mins worth) are
kept in memory, and the station reads out the max,
min, and average light percentage when asked. What
does this tell you?
Well, if there has been no cloud between the sun and station in that
half-hour, then all snapshots will read 100%. So, the station will say:
Daylight varying between 100% and 100%, average 100%
Pretty obvious, really - a blue sky.
Now, the opposite - dark thick stratus, post warm front, probably raining. Each of the 60 snapshots will be between probably 8% and 15% light value.
Therefore, in the last half hour there have been no gaps, the sky is stable cloud. You will therefore hear:
Daylight varying between 8% and 15%, average 11%
What about a nice post-cold-front day? Well, when the sun is out, the station snapshot will be 100%. When moderate cu covers the sun, it will drop to maybe 50%.
If the station has taken 30 snapshots at 50%, and 30 at 100%, the
average will be 75%. So, the report will be:
Daylight varying between 50% and 100%, average 75%.
If it took nearly all cloud snapshots, and only a few of sun, then
the average would be near the low figure:
Daylight Varying Between 50% to 100%, average 55%
If most of the snapshots were sunny, then the average would be near the top:
Daylight Varying Between 50% to 100%, average 92%
In other words, the oktas are given by WHERE THE AVERAGE LIES WITH RESPECT TO THE MAX AND MIN VALUES!
And, remember, the DEPTH of cloud is given by how low the figures drop - so, these three numbers max, min, and average daylight (IR) give you a
complete picture of the sky cloud state, both in thickness and oktas.
The thing to remember is, just as the difference between max and min windspeed gives you the turbulence, the difference between max and min
daylight gives you a good idea of the instability.
Imagine the sky immediately after a strong cold front has passed through - as cumis develop and overdevelop, with brilliant blue sky between,
turning into black thunderheads, perhaps. The lowest light figure will drop
and drop, whilst the top figure stays at 100%, until the overdevelopment
spreads enough to cut out the sun altogether for 30 minutes. Then it will
rapidly drop to a low figure, as old data is overwritten.
In the most recent version of the stations, an algorithm converts these IR light values to direct Sky States,
and reads it out to you. You hear directly, a report such as:
The Sky is Mainly Sunny, with some Scattered Moderate Cloud
We hope to convert all the lines to this method shortly. Of course, you can't work this system
unless the sun is at LEAST 15 degrees or so above the horizon. Therefore, the skystate conversion
algorithm only operates between certain times of day, depending on the time of year, when the sun
should be up. Outside that time window, the report reverts to the raw daylight readings, as above.
Subscribers using the telephone stations can opt for
the raw daylight percentages instead of the pre-figured
skystate by entering a "0" digit
immediately after the station says:
This is (station name) at (time)
You have about 1.5 seconds to enter the control digit. This will cause the
report to revert to the more informative raw daylight percentages, if that
is what you prefer.
Other commands for phone readouts are given on the
Subscribers Help Page
Anticipated Height of Cloudbase
The height of cloudbase is given in feet, calculated on the formula
developed by that gliding met guru, Tom Bradbury. Lowest temperature
last night taken away from highest temperature today, times 400.
It works pretty accurately for most of the year, as the formula is
really highest temperature minus dewpoint x 400. However, for most of
the flying year (March-October) the lowest temperature last night IS the
dewpoint, as the release of latent heat by dew formation puts a floor on
the temperature fall.
The system breaks down in winter sometimes, for obvious reasons,
but generally, it is pretty reliable. Thinking about it, if there is
a cloud blanket and high humidity, then temperature at night doesn't fall
very much. Equally, in the morning, the cloud blanket prevents temperature
rise due to solar heating, and evaporation of surface water absorbs a lot
of heat that would otherwise cause a temperature rise. Therefore small
temperature difference from night to day = low cloudbase.
In high pressure times, with clear skies, though, night temperature
falls quickly, due to radiation. In the morning, clear skies and dry
conditions allow the temperature to rise rapidly. Large temperature difference =
high cloudbase (if there is enough moisture to form cloud at all!)
If there is blue sky, the station will still estimate a base height,
but obviously, the daylight sensor will tell you that there really isn't
any cloud around!
The rain gauge is a tipping-bucket type, graduated to 1/100
of an inch (0.25mm). The only snag with it is that in winter, it fills
with ice and snow, and then, on a lovely blue day, warm sun causes the
snow to melt - it runs through the gauge, which reads lots of rain!
Not having sufficient electrical power available on most remote sites to heat
the gauge above freezing point,
there is really no cure for this one.
The station records each tip of the bucket, and gives out the
total rain since midnight, and the time of the last rainfall.
After all, if you're looking for thermal activity, then knowing when it last
rained is vital, as wet ground doesn't give off much thermal activity at all!
If it last rained at 2:30am, then maybe no problem. If it last rained 2 minutes ago,
then you might as well stay home.....
Speed Readout in Knots/MPH and Beaufort Force
The Hilltop stations read windspeed out in MPH, currently.
They have the capability to do knots, km/hour, metres/sec, or Beaufort, but as most footlaunch
pilots think in mph, that's what the stations give out at present.
Hang and Paragliders represent about 80-90% of users of these Hilltop stations,
although there is a growing number of other users, such as GA pilots,
Balloonists, Rock climbers, Walkers, etc, etc. However, with the exception of
GA guys, who are Knot-orientated, most of these other users understand MPH better,
so we'll stick with that.
Coastal stations seem to be primarily used by Sailors, Windsurfers, etc,
and they tend to think in Knots and/or Beaufort Force. Therefore these coastal stations
give out the windspeeds in both of these, rather than in MPH.
Getting Data from Remote Sites
Obviously, it is not always possible to have mains power and phone
line on top of remote hills in the middle of nowhere. Therefore, the Wendy Windblows stations
can be split in two. The wind and other sensors can be installed on the hilltop
(to get the best data) and powered by a solar panel. The data is then periodically sent by a low-power
UHF radio telemetry link down to civilisation nearby, where there IS power and a phone line. The other half
of Wendy, ie. the computer and phone interface, are installed there.
A good example of this is the
Bradwell (Camphill) station near Sheffield. The wind sensors are on the Derby and Lancs Gliding Club airfield
at Great Hucklow, but the data is transmitted by radio to the Woodbine Cafe, in Hope, near Castleton,
where the main computer that runs the service for Bradwell is situated, as is the phone line into the station.
A display screen in the cafe shows the current weather, and Paraglider and Hang Glider pilots can often be
seen in the cafe, slurping coffee, and keeping an eye on the weather data till things improve!