| |
ZL3AI > APRS 29.10.00 15:51l 166 Lines 7408 Bytes #999 (0) @ WW
BID : 34186_ZL3AI
Read: GUEST
Subj: TAPR Digest, Oct 26, 8/9
Path: DB0MRW<OK0PKL<OK0PPR<OK0PPL<OK0POK<9A0YRB<HI8GN<KP4IG<7M3TJZ<HA3PG<
VE1DRG<WA7V<KB7FRV<VK4WIT<ZS6RO<ZL3VML
Sent: 001027/1132Z @:ZL3VML.#80.NZL.OC #:45156 [Chch-NZ] FBB7.00g25 $:34186_ZL3
From: ZL3AI@ZL3VML.#80.NZL.OC
To : APRS@WW
----------------------------------------------------------------------
Subject: Notes From Dr Arnold.
From: David VanHorn <dvanhorn@cedar.net>
Date: Thu, 26 Oct 2000 21:40:20 -0500
X-Message-Number: 31
I had Dr Arnold come over and give me his thoughts on using WinAPRS WRT
stormchasing, and weather in general. He is, I think, pretty well
qualified in this area.
http://www.bsu.edu/geog/html/98sct1.html
http://msuinfo.ur.msstate.edu/alumnus/summer.96/10storm.htm
The BSU team mentioned above is separate from, and in addition to our ham
radio and "civilian" spotters here in Delaware county.
We went over all the basic program issues I emailed you about previously.
(Un handled exception errors, window positions that jump around.)
Some of these are "broken features" some are features that we would like to
see..
He agrees, that absolutely, placed weather objects, if given a vector,
should move. (Not the current system with the moving blue dot (that looks
like the blue dots for cities on the PM maps, when it decides to show up at
all), with the "radar rings" that lock up the system when they display with
dead-reckoning turned on.) He brought a colleague over to see this today,
but I was unable to place a weather object at all!. Despite repeated
attempts, I was not able to place any weather object on the map. Obviously
shutting down the program and restarting it would likely have solved the
problem, but again, this is not practical in actual operation.
He did have some further comments about a few things:
Weather objects:
Delete: Haze, Snowflake, Sunny, and Partly Cloudy (No threat)
Add: Snow, Freezing Drizzle, Hail, Fog, Smoke
I know that some of these can be placed through a different menu, but
that's pretty procrustean.
He added that it might be useful to be able to select "families" of
objects, as things like hurricanes aren't of much use in most areas of the
country. Some work could be done to split those out, and to put
non-hazardous objects into a separate group.
Also, deleting these objects should be done by clicking on the object, and
hitting the delete key, not by digging through a 7000 item list.
(Procrustes rides again)
This assumes that we were able to connect the station to the net without
greatly increasing the UHX error problems.
Actually, as I said before, we would really like to make that station our
igate, but we cannot tolerate the crash frequency, even with Afilter in
line. The EOC station is hardened to survive an F-4, and has a two week
power/food supply, and it's own well. My house (where the Xastir Igate
currently is) has a 30 minute ups, and is guaranteed to withstand a good cough.
He thought that the wind vector display (as Bob does in dos APRS) would be
quite useful.
Date and time stamp on wx station observations. As is, you know when you
got the packet, but you don't know when the observations were made. (I
realize this is a protocol issue, but I'm just the middleman here)
There are more basic functional problems:
Unhandled exception errors. Worse than an annoyance, they also cause any
data from remote wx stations (trend data) to be lost.
Windows that loose their screen position are a big distraction, and a
time-waster.
The inability to display county warnings along with PM maps.
Overlay files that disappear when you zoom in or out. (but are still
"turned on")
Use of the PM maps is very important. WinAPRS maps (at least any I have
seen) are missing a tremendous amount of detail. Sending someone down a
dead end during storm spotting could take on a very literal meaning.
Of course with the PM maps up, the county warning overlays don't show up.
The local weather display is a long vertical list that can't be displayed
all on the screen, at least not in any combo of screen resolution we can
get to, on our 14" screens. There is, however, plenty of room to lay that
list out in a more rectangular fashion. I don't eve know how long the list
is, I've never seen the end.
The wind display is nice,(Averaged and gust in polar form) but it should be
labelled so there is no guessing as to what is what. Also, it should be
documented what averaging methods are used to create the averages.
Rain display from other stations: no units.
The color (whatever that is) is un viewable, and he initially thought the
window was blank.
In the wind graph, he liked what you did to display wind direction and
speed, but thought there should be some labeling as to what is peak,
average, or ??. You can figure it out eventually, but in real use, you
don't have time.
Map clutter is also an issue. I realize that there are conflicting
requirements between wanting to "see everything" and how much can actually
be fit onto a page. One possibility is a chooser for "level of detail" or
maybe better, an expansion of the "WX mode" on the right side of the
screen, where you can select what you are going to be shown.
A selectable audio alarm for the appearance of certain icons (tornado, wall
cloud..) this should probably be selectable on a per-item basis.
We discussed the idea some more of mobiles reporting temperatures over an
area, WRT plotting isotherms, and looking for thermal boundaries, which are
KEY indicators of where a tornado will be generated. What needs to happen
is to equip our vehicles with temperature sensors, and to plot the readings
as isotherms, sorted in time order. (Covers cases where someone else drives
through the area, or you drive back through later) Older data is better
than no data, but more recent data takes precedence. I will stress again,
that the technique of locating outflow boundaries by temperature is proving
to be a very reliable predictor of where tornadoes will form, based on when
a suitable cell crosses the boundary. You can literally plot the
intersection, drive there and wait, and very often, watch a tornado
form. Knowing where these boundaries are takes a lot of temperature
measurements, something that a vehicle equipped with APRS is very well
suited to do. Steve Bragg and I have discussed this feature for the HHIII,
and we plan to implement it. In the meantime, KPC3+'s can be used.
There are some issues with the weather station data as well. We noted that
there is a wx station about 30 miles away from us, reporting a pressure
about 30mB higher than ours (Which IS correctly calibrated), which would
translate to roughly hurricane force winds. It's a nice calm day here..
He suggested that Wx station operators get the baro value from the nearest
airport, on a day with not much pressure gradient. For the meteo-impaired,
that translates to a nice calm day, with little or no wind. Wind means that
there's high over here, and low over there, and a gradient across where you
are. Obviously the farther the airport, the more important this becomes.
Some weather stations display wind, baro, or temp at 000 000.. Does this
mean that they don't have sensors, or that they are active, and reading
zeroes? We had a good laugh at a station near us displaying a baro press of
000mB, Dr A said that obviously this guy is in outer space. (Zero mB being
a hard vaccum)
Temperature sensors are another issue, and again, this is mostly an issue
for station implementation. They need to be mounted 5' above ground level,
Read previous mail | Read next mail
| |