Professional Forecast Tools

[Bill Tabor]

Question - particularly for SPC severe weather forecasters, but for others that know as well...

What tools do professional severe weather meteorologists use to help create a forecast such as Convective Outlooks? Is this still done primarily by hand analysis and drawing of maps with features, while also reviewing soundings, hodographs, model output? Are there any tools which help facilitate this task?

 

[Joey Ketcham]

I wish they would make some windows based software that would allow you to create easily create your own forecast mapes, 5-day forecasts, and what not for web use or whatever.

 

[rdale]

The GEMPAK/NMAP suite of products is utilized for analysis & forecasting (and printing for the coloring time) and then it's also the software that draws the outlook / meso / etc. maps. http://www.unidata.ucar.edu or http://skywatch.org

- Rob

 

[Robert Dewey]

Originally posted by jketcham
I wish they would make some windows based software that would allow you to create easily create your own forecast mapes, 5-day forecasts, and what not for web use or whatever.

You can setup a seperate Linux box with GEMPAK and hook it up to a 100Mbps (or a 1Gbps) hub... Then all you need to do is SSH your way in and you can run all of the apps on your Windows desktop pretty easily.

 

[Kiel Ortega]

Originally posted by Bill Tabor
Question - particularly for SPC severe weather forecasters, but for others that know as well...

What tools do professional severe weather meteorologists use to help create a forecast such as Convective Outlooks? Is this still done primarily by hand analysis and drawing of maps with features, while also reviewing soundings, hodographs, model output? Are there any tools which help facilitate this task?

To answer the other part of your question, yes they still do analysis by hand...I walk by the SPC usually during my lunch break and someone is usually hand analyzing a map. From what I have heard and been taught, there is nothing better than your own hand analysis (I suck at it, so I prefer to look at someone else's)

 

[Robert Dewey]

Originally posted by Kiel Ortega
From what I have heard and been taught, there is nothing better than your own hand analysis (I suck at it, so I prefer to look at someone else's)

That's true... I find that doing a hand analysis on a piece of paper puts you in some kind of a "zone". Your only focus is on the charts that you are drawing, and you often find features that you would have missed...

Generally though, I let the computer draw the isobars, isotherms, isodrosotherms, iso-whatevers, etc., so I am guilty :lol:

 

[Mike Gauldin]

Generally though, I let the computer draw the isobars, isotherms, isodrosotherms, iso-whatevers, etc., so I am guilty :lol:

I haven't gotten the iso-whatevers down yet, can you give an example?

 

[nickgrillo]

Originally posted by Mike Gauldin

Generally though, I let the computer draw the isobars, isotherms, isodrosotherms, iso-whatevers, etc., so I am guilty :lol:

I haven't gotten the iso-whatevers down yet, can you give an example?

"Isowhatevers" are the lines you draw around the big fat skull and cross bones (iso-mich) around the great lakes during a potential severe weather event. In N-AWIPS you'll find them easily by pulling up NMAP and simply just clicking on the state of MI...

 

[Rich Thompson]

Bill,

Different forecasters approach the convective outlooks in different ways. Most of my outlooks start with a set of surface and upper air charts that are analyzed by hand, and the hand analyses are augmented by 12-24 hour satellite loops. Then, I'm usually off into several variations of the numerical model output. I prefer a few basic fields (500 mb heights, vorticity, winds, temps) and MSL pressure to develop an overview of the situation, and then I look back at the observations to get a feel for moisture and lapse rates in the typical source regions. The short range ensemble forecasts, and comparisons amongst the operational NAM/GFS/etc., are used to solidify my general expectation for the weather pattern in the upcoming outlook period.

I follow this with a more detailed look at several composite "spaghetti" plots derived from the model output, and the ensemble mean/spread of the same ingredients. We post process much of the model output and reproduce forecast versions of many of our mesoanalysis parameters, and we also have the "NSHARP" sounding software to slice-and-dice soundings in a multitude of ways (including impacts of model convective schemes).

Last but not least, we consider climatology and our own pattern recognition skills as a way to check our forecast and perhaps make modifications. You can approach the problem in many ways, but we're ultimately trying to forecast the future distribution of moisture, lapse rates, vertical shear, lift, and resultant convective mode(s).

Rich T.

 

[Bill Tabor]

Originally posted by Rich Thompson
Bill,

Different forecasters approach the convective outlooks in different ways. Most of my outlooks start with a set of surface and upper air charts that are analyzed by hand, and the hand analyses are augmented by 12-24 hour satellite loops. Then, I'm usually off into several variations of the numerical model output. I prefer a few basic fields (500 mb heights, vorticity, winds, temps) and MSL pressure to develop an overview of the situation, and then I look back at the observations to get a feel for moisture and lapse rates in the typical source regions. The short range ensemble forecasts, and comparisons amongst the operational NAM/GFS/etc., are used to solidify my general expectation for the weather pattern in the upcoming outlook period.

I follow this with a more detailed look at several composite \"spaghetti\" plots derived from the model output, and the ensemble mean/spread of the same ingredients. We post process much of the model output and reproduce forecast versions of many of our mesoanalysis parameters, and we also have the \"NSHARP\" sounding software to slice-and-dice soundings in a multitude of ways (including impacts of model convective schemes).

Last but not least, we consider climatology and our own pattern recognition skills as a way to check our forecast and perhaps make modifications. You can approach the problem in many ways, but we're ultimately trying to forecast the future distribution of moisture, lapse rates, vertical shear, lift, and resultant convective mode(s).

Rich T.

Ok, thanks all and thank you Rich for specifics here. So I get the impression much of it is done manually and by hand, and forecaster specific based on their own techniques (although most know and are taught to recognize the same features / techniques). I say manually and by hand but of course complimented by much technical output from various products sources (as you mention). Is it true that outlooks and mesoscale is drawn using GEM, or a similar product? I suppose I need to get that and review just what it can do.

I follow a similar strategy, although not as broad since trying to focus typically on just a target where tornado potential is maximized (perhaps nearest me). I also of course don't get quite as involved in the number of formalized processes / procedures Rich mentions and tend to do most of it 'in my head' from memory rather than hand drawing of features even though I do sometimes draw these as well. I can see that is much more important if you are trying to determine the overall threat to all areas rather than just a maximized point of tornado or supercell potential.

Anyway I guess the point of all this is I wanted to get an idea for how the pro's do it now, because I am not satisfied with my procedure in many respects, and also had some ideas for 'next generation' forecast tools. Don't know if I will just throw these ideas out into the public domain to help hopefully see them eventually developed, or if I will first try and get a patent on some of them. I don't think I will develop them myself as too complicated and my programming experience is limited - primarily to COBOL, BASIC, and some html. The more time I spend on these concepts the more complicated they get - as forecasting as a science and in reality can be very complicated.

 

[Rich Thompson]

Bill,

We use NMAP software to draw the outlooks, and most of our graphics (model output, mesoanalysis, etc.) are generated via GEMPAK. Within NMAP, there are drawing tools that allow symbols, fronts, etc., that you see at times in the MD graphics.

Rich T.