Forecasting Techniques

[Andrew Clope]

I've been trying to make my own forecasts for a little more than a year now. I've had a little success, a lot of failure and I got to thinking, I imagine that everyone has their own way of making a forecast. A progression that you go through, and fine tune as the years go by and experience begins to grow. I was wondering if you guys would be willing to share that "process" with us beginners who are trying to come up with our "own" process?

What is the first thing you look at in trying to decide if an event will be chaseable? What are the deal breakers in a setup? What are the "must have's", if there are any? What are your favorite parameters/models/observations to use? Parameters that you thought were important when you started out, that turned out to be useless, and vice versa, parameters you didn't look at, that you really needed to pay closer attention to... How does it progress for you personally from a few days out, up to the hour before initiation? Etc...?

Not sure if this thread will fly, but it's something that I often wonder about, so I thought I'd throw the question out there!

There was a similar thread a few years back that I found, that I linked below, but given the gap in time, I figured it was fair to start another thread on this topic.

Original thread:

https://stormtrack.org/threads/my-forecasting-technique.15857/#post-187370

 

[Bob Schafer]

Andrew, buddy, with all due respect what you're asking is like asking a doctor "What did you learn in college?" Take one thing at a time.

 

[Royce Sheibal]

Storm Chasing in Nebraska:

1. Open up 16 day GFS set to north / south plains and pull up 3k EHI. See if anything shiny shows up.
2. <72 hours Pull up NAM to check DP/LI/850mb winds/250mb vort/ML LCL/ML CAPE as well as 1k and 3k EHI.
3. <60 hours Whip out NAM 4k (my baby) and compare to NAM, check for 0-1 Updraft Helicity spikes >100
4. <24 hours dig into some RAP and HRRR and look closer at helicity, cape, cape, sim reflectivity.
5. <12 hours stare at radar and visible sats, looking for outflows, boundaries, LLJ's, Q-fields, check surface data for dry lines and heavily capped / mixed areas to avoid
6. <2 hours Go find the Tallest Turkey Tower I can find and hope something drops.
7. Beer.

For all of my threshold values I consider tornadic atmosphere climatology, which in eastern Nebraska is typically >1200 Cape, >150 0-3k SRH, Bulk Shear >50, STP >2.0, ML LCL <1250. These wrap up the majority of requirements for a tor in some easy to find maps. On the downside, I tend to end up chasing way too many HP monsters and multi-supercell clusters, because storms mode in Nebraska tend to favor those setups given my set of conditions.

Some of my best chases have been with CAPEs right around 1000, so it's important to remember all of the dynamic factors that can boost or ruin your chances.

 

[Andrew Clope]

Andrew, buddy, with all due respect what you're asking is like asking a doctor "What did you learn in college?" Take one thing at a time.

Suppose I should clarify... I'm not asking for an incredibly in depth dynamical answer here. Didn't mean for that to be the assumption.

 

[Todd Rector]

I thought @Royce Sheibal 's answer was very helpful - Thanks!
Nebraska is an enigma to me. I witness the very photogenic Bradshaw/York tornadoes a few years ago, then get nothing but HP messes ever since.

 

[Bobbi Burrows]

Storm Chasing in Nebraska:

1. Open up 16 day GFS set to north / south plains and pull up 3k EHI. See if anything shiny shows up.
2. <72 hours Pull up NAM to check DP/LI/850mb winds/250mb vort/ML LCL/ML CAPE as well as 1k and 3k EHI.
3. <60 hours Whip out NAM 4k (my baby) and compare to NAM, check for 0-1 Updraft Helicity spikes >100
4. <24 hours dig into some RAP and HRRR and look closer at helicity, cape, cape, sim reflectivity.
5. <12 hours stare at radar and visible sats, looking for outflows, boundaries, LLJ's, Q-fields, check surface data for dry lines and heavily capped / mixed areas to avoid
6. <2 hours Go find the Tallest Turkey Tower I can find and hope something drops.
7. Beer.

For all of my threshold values I consider tornadic atmosphere climatology, which in eastern Nebraska is typically >1200 Cape, >150 0-3k SRH, Bulk Shear >50, STP >2.0, ML LCL <1250. These wrap up the majority of requirements for a tor in some easy to find maps. On the downside, I tend to end up chasing way too many HP monsters and multi-supercell clusters, because storms mode in Nebraska tend to favor those setups given my set of conditions.

Some of my best chases have been with CAPEs right around 1000, so it's important to remember all of the dynamic factors that can boost or ruin your chances.

Awesome answer. This is what I was looking for when I asked a similar question a while back. Thanks!

 

[Bobbi Burrows]

I thought @Royce Sheibal 's answer was very helpful - Thanks!
Nebraska is an enigma to me. I witness the very photogenic Bradshaw/York tornadoes a few years ago, then get nothing but HP messes ever since.

I thought so too! Certainly helped me... I like having things broken down into steps with actions I can take. I (along with many others) learn best by doing. It was a great answer.

 

[Andrew Clope]

Thank you Royce for that answer. That's pretty much what I was looking for.

I appreciate you taking the time!

 

[Shane Adams]

I use the GFS to get an idea if anything is even being hinted at in the Narnia realm (10-17 days out). This allows me to be able to tell my boss "Hey, around the so-and-soth there might be something to chase." He only asks for 24hrs notice, but I always try to give him as much heads up as possible. Regardless, any "Narnia" range event notifications are forgotten by him within a few days....but it gives me a platform to argue should a chase day arrive and he says "you didn't say anything about this."

Continuing with the GFS, I'll keep scanning for any Narnia events that make it into the "Shame on you for even considering this will happen" realm of 5-9 days out. If a Narnia event makes it into the "Shame on you" realm, it's on the board, and the boss is notified "Hey, we'll likely be chasing at least once next week."

Once a "Shame on you" event makes it to within the NAM's range ("Reality" realm), it's a green light, the boss is told "we're chasing this day/days" and plans are set into motion. I stop using the GFS completely, and go exclusively with the NAM. I usually get a decent indication of which way an event is trending by looking at several consecutive GFS runs, followed by several NAM runs. In the unlikely situation the two models are worlds apart on the "handoff" day, I may temporarily reference both to see if they start to agree more. Unlike a lot of chasers, I generally just don't like comparing different models day to day in the days leading up to an event, because it's just too much information for me to ingest. Less is more, as far as my brain and getting a feel for a forecast.

When an event gets within 24-36hrs inside the "Reality" realm, I'll start looking at convection-allowing models, to see how well they agree with each other as well as the NAM. I really don't put stock in placement of COM so much as consistency regarding storms/no storms in a given region.

This might not be what you were asking, but this is my overall forecast method, as far as how I work a chase into my day-to-day life plans regarding work, which is the biggest obstacle I have to deal with. As far as specific parameters, that's a whole other reply that is too long for me to write at this moment.

 

[Todd Rector]

@Shane Adams - I love your descriptions (Narnia realm; Reality realm)... Your method closely mirrors my own. There are SO many variables at play and my personal knowledge is limited so I don't do model run comparisons either. I consider myself lucky of I can decide on a target the day of an event.

 

[Andrew Clope]

Thanks Shane!