RASP Update

Or more like weather education. I've posted information I use to help figure out the day. Often I'll also post graphics with explanations of "why" the weather guessers think the day is good.

Moderator: Chip

Post Reply
AlC
Posts: 87
Joined: Sun Aug 12, 2007 9:02 am

RASP Update

Post by AlC »

I upgraded the RASP machine this fall so the models run in about half the time. With a little prodding from our new RD Dan, the run schedules have now been changed. Forecasts are available much more quickly which should improve the accuracy. See the new Run Schedule page http://www.soaringpredictor.info/raspta ... e_web.html for details. I run some of the models multiple times to provide the most current data as early as possible. New sets of data become available every 6 hours.

- Alan
User avatar
JD
Posts: 1682
Joined: Fri Apr 25, 2008 11:05 am

Re: RASP Update

Post by JD »

AlC wrote:I upgraded the RASP machine this fall so the models run in about half the time....
Mooch Ohs Grassy Ahhs !!
AlC
Posts: 87
Joined: Sun Aug 12, 2007 9:02 am

Soaring Predictor Skew-T Plots

Post by AlC »

If you have been having trouble getting the Soaring Predictor Skew-T plots from the table (note, I always recommend the data from RASPTABLE as it is a much better predictor, but...) then you have seen some of the Internet implications of the recent issues with Java security ( http://articles.latimes.com/2013/jan/11 ... y-20130111 ). For a brief time, it seemed impossible to run the java code in IE10. Anyway, thanks to the good folks at Certum http://www.certum.eu/certum/main.xml and their support to the Open Source community, the Java code is now 'signed' which at least gives you the option to trust it.

Note, you must have the Java Runtime Environment loaded to get the plot.

- Alan
User avatar
Vrezh
Posts: 230
Joined: Wed Aug 31, 2005 8:52 pm

Post by Vrezh »

Hi Alan,

First, I thank you for making this wonderful tool available for us.

I contacted you couple of months ago about issues we were having with

RASP, and was explained, that it was a temporary problem, and was

going to go away as soon as the migration to a new hardware was over.

I wonder if you are aware, that the same issue is persisting now, even

after above mentioned updates.

Mainly, I am talking about Skew-T plots just jumping to some random

dates and hours, when clicked for the "next" hour's graph.

For example, if you want to check Skew-T plots for Sunday, February 17,

it might even start out from totally different day and hour, or, in a better

case scenario, will start from 11:00AM, for example, and instead of

going to 12:00, 13:00 and so on with each "next" click, it might just go to

some 14:00PM, November 4, or something like that...

Does this sound familiar?
AlC
Posts: 87
Joined: Sun Aug 12, 2007 9:02 am

Actually Not

Post by AlC »

Vrezh:

First I've heard. I'll check, but I'm thinking time change from the description. The run times change by an hour at the spring and fall time change as they are all based on GMT. One of the scripts may not take that into account.

Are there other odd dates/times?

Thanks,

Alan
User avatar
Vrezh
Posts: 230
Joined: Wed Aug 31, 2005 8:52 pm

Post by Vrezh »

November 4 is what I remember the most.

In fact, I checked right now, Saturday, February 16, 11:00 PM, before I

wrote this, and it's doing the same thing. I check for tomorrow, it starts out

just fine, from 11:00 AM, and then, as I click "next", it goes to the next hour,

as it is supposed to. But after 15:00 PM it jumps to 16:00 of that proverbial

November 4th.
AlC
Posts: 87
Joined: Sun Aug 12, 2007 9:02 am

Progress

Post by AlC »

Vrezh:

Ok, so the infamous November 4 files are gone. They were leftover from the time change. But all is not perfect. If you click the "North Area" and current day, you will have no issues. But if you are using "Entire Area" and/or Day+1 or Day+2 you will have some "Image not found" frames after 15:00. This is because I do not run as many hours of forecast further out and for the bigger model. Right now the Skew-t viewer does not 'know' that and always tries to display the same set of hours. So, I need to spend some time making the viewer a bit 'smarter'.

- Alan
AlC
Posts: 87
Joined: Sun Aug 12, 2007 9:02 am

RASP Viewer Update

Post by AlC »

Major Upgrade to the RASP model viewer.

As some of you may know, GoogleMaps shifted to a paid service about a year ago. More recently they required a version upgrade. If you’ve been trying to use the RASP maps it’s been pretty ugly with a “development purposes only” banner.

These hot days have been good for one thing, I’ve implemented Darren’s OpenStreetMap replacement viewer. The model underneath hasn’t changed, but it’s free of any reliance on GoogleMaps.

http://www.soaringpredictor.info/raspviewer/index.php

Enjoy,

Alan
User avatar
Vrezh
Posts: 230
Joined: Wed Aug 31, 2005 8:52 pm

Re: RASP Update

Post by Vrezh »

Hello Alan,
Thanks again for your wonderful public service!

Ever since I resumed flying, I religiously follow the weather every day,
(even though I fly mostly weekends).
Checking out Windgrams and Rasp for Kagel is a given.
To give you a feedback, according to my observations, they undercall almost on regular basis.
That is, when we compare the predictions and how actually the conditions played out.
In comparison, the Blip is been overcalling noticeably, Windy seems to be more realistic, and
XC Skies is a mixed bag, I call it a 50/50.
Though, they all have their hits and misses- it's weather.
On Windgrams, switching to Meteograms mode gives a somewhat better readings,
but seems like overall they need some tweaking.

I thought you'd find this worth looking at.

Best regards.
AlC
Posts: 87
Joined: Sun Aug 12, 2007 9:02 am

Re: RASP Update

Post by AlC »

Vrezh:

Thank you for the feedback. Can you tell whether the under call is in the W* fpm up; or the calculated maximum altitude? I can adjust the sink rate to produce a higher maximum altitude fairly easily. But the W* forecast is baked into the modeling :-(

This was a pretty major upgrade, moving up a couple of generations of the WRF model and adding more detailed geography. Hopefully it is a useful guide. I see a discussion has started on the model website of another upgrade to the WRF 4.3 version, with some thought it may be available by Spring.

Anyway, glad to hear the model is being used. That is the entire point of the project.

Happy Holidays and Fly Safe!

Alan
Post Reply