The
SPC Mesoanalysis is one of my
favorite tools to use when looking getting a glimpse of conditions out there. It is based on the RAP model which is updated hourly, making
over 140 products available. The page shows the 'zero-hour' (f0) forecast. Viewing previous imagery is very limited; maybe 2 or 3 images (4 to 6
hrs past/future). Interestingly enough, about 3 days worth of data are archived (the 78 previous hours), but the user's access to these are very
limited. The purpose of this tool/script is to give access to the user to this imagery. The primary use of this would be for archival/later-viewing
and research purposes. Like, when a storm system comes through, afterward (or during), you can run this tool, indicate how long of a back up you
want to do, pick the products you want, download them, and view them later. Being able to look at a regional view (mesoscale) allows some features
to stand out.
A snippet of the menu
* The program then creates a file directory structure that archives the images, first by region, then by product, then by date. The
advantage of this is that you can study the images without the included html files and you'd still be able to cycle through them in
order.
* This script is extremely advantageous because it utilizes the
HAniS
module (included with download). My script simultaneously produces an html file (which will be used for viewing), control file
(which tells the script how you want it to look and which options to include), and a filenames file that the HAniS program reads so
it knows what images to display (and corresponding overlays) according to the time.
This took quite a while as I had little experience with HAniS. Then, if you look at the python code, you'll also notice it is nearly 300
lines long. I also took a long break from working with it when I first got it functional. The thing is if you miss one little detail in
creating the config and filenames (get format wrong), it can throw the whole operation off. So I was really happy when I got it to work.
In the code I included error controls which should prevent 'malfunction' during use. I wasn't initially sure how to get some things to
work, so searching the web was trusty in that. I tried to include links to all my sources; generally they're found at the start of each
function.
* Included in the bundle is a small script called 'http_server.py.' This likely will be needed for the web browser to display your file
properly since we're operating locally. Follow the on-screen instructions. You'll notice if you do need it, that you won't be
directly opening the created html file. P.S. - You'll want to unzip the file's contents into their own folder. See the included readme
for terms of use. Contact me if you have any questions!
* NOTE: You'll likely run into missing images of the most-recent hour if your request is in the early
part of an hour. This is because it takes time to generate each model run. The filename scheme that I used is anticipatory of what the
filenames will be. So if you encounter some blank images, understand it generally has to do with the time you ran the script and not an
error in the program. I'm considering options on how to handle that.
* NOTE 2: Some images aren't very overlay friendly. This is due to that archived images have a thatched fill rather than a solid fill.
Play around with which products work the best with one another. Another option is to use the CONUS sector (s19). Things may be able to
be better-seen on that, especially synoptic features.
* NOTE 3: The SPC does archive the CONUS sector with solid-fill variables. I've been thinking about releasing a similar script with that
so working with older events works. Stay tuned!
RMDaV.zip