SPC MesoA Bootcamp No. 1 - The I4 Shift
Kelton Halbert
Published: 2024-06-24
Contents
- 0: First 3 Months as a TDM
- 1: The I4 Shift
- 2: 1630Z General Thunder and 17Z Enhanced Thunder
- 3: Mesoscale Discussions
- 4: Thoughts So Far…
0: First 3 Months as a TDM
It’s been nearly 3 months since I’ve transitioned into my role as a Techniques Development Meteorologist (TDM) at the Storm Prediction Center, and I thought it could be fun to stick to my intention of blogging and documenting things more frequently… especially while I conduct my training as a fill-in Mesoscale Assistant/Fire Weather Forecaster (MesoA). My primary responsibilities in the Science Support Branch (SSB) lie in developing and maintaining new sounding and mesoanalysis tools, along with supporting operational software and data flows. Most of my time so far has been spent in developing an upgrade to the SPC Mesoscale Analysis for testing in the Hazardous Weather Testbed (HWT), using SHARPlib to post-process the 3-km scale mesoanalysis in parallel. However, the other responsibility of a TDM is to be able to fill-in on the forecast desk as a MesoA. I started training for these shifts last week, and training will continue every week through August, working a variety of shifts that each come with their own unique responsibilities, challenges, and nuances. To suggest that you are expected to come out of this training period knowing all of the answers would be preposterous - the only real way to gain experience and knowledge is time in the saddle. It is meant, however, to give you enough time in the saddle to learn the processes, workflows, and cadences of those shifts in order to help you hold your own. Even from the perspective of my primary responsibilities, there is no better way to understand the challenges and problems that need addressing in the forecast process than to experience them first hand. This is yet another reason why I thought it may be good to start a weekly log of the process, and maybe you’ll get to learn something along the way.
1: The I4 Shift
If you’re not already familiar with the way shift labels work within the National Weather Service (NWS), the start time of a shift is listed alphabetically, starting at midnight local (00) with A. An I shift, therefore, is an 8 hour shift that runs from 8 AM - 4 PM. The number (in this case, 4) represents which desk you’re working. The numbering system goes from 1, which is the Lead Forecaster desk, to 6, which is the social media desk (note: the responsibilities of this shift are changing soon to include more things). So, in this case, the 4 desk is the Mesoscale Assistant. In the future I will be working I4s, J5s, N6s, and R10s (R shift, 10-hours) - and I’ll go into what those are during the weeks I work them!
On the day shift, you usually have a Lead Forecaster, Outlook Forecaster, Mesoscale Forecaster, Mesoscale Assistant Forecaster, Fire Weather Forecaster, and then the social media shift doubles as an additional Mesoscale Forecaster as needed. The Lead Forecaster is responsible for the 1630Z Day 1 outlook update, coordinating and issuing convective watches, and QCing/approving forecast products from all of the other staff. The Mesoscale Forecaster coordinates directly with the Lead Forecaster in issuing MDs and prepping for watch issuance, and also generally has seniority on determining how to divide geographic regions for MDs. Finally, the MesoA assists in issuing MDs and coordinating with both the Lead and Meso Forecasters on products, as well as fulfilling duties as needed (answering telephones, AWIPS collaboration chat, etc). An additional responsibility of the MesoA is the 1630Z Day 1 General Thunder outlook lines, and the 17Z Enhanced Thunder outlook lines that are valid over two 4-hour periods, and one 8-hour period.
During the summer, the I4 shift can be quite busy. This is the time of year where there can be thunderstorm potential across much of the United States - it’s not uncommon to see outlooks that go from the Mexican to Canadian borders, or even the East Coast all the way to the Central Rockies. For this reason, usually the Meso, MesoA, and N6 forecasters will split up responsibilities geographically in order to manage the workload. Still, that can result in a single forecaster trying to manage a region of a half dozen states for severe potential! Additionally, since SPC is located in the Central Timezone, things can get started relatively early on in the shift on the East Coast… and it only accelerates with time and westward extent with diurnal heating. The shift can very quickly become a sprint, and there is no such thing as a “lunch break”! Training for a shift like this has little to do with learning more meteorology - you are expected to already have that part down (though questions are always encouraged)… training for this shift is an exercise in situational awareness, developing efficient workflows, learning to quickly pivot between geographic regions and meteorological regimes, learning time management, and learning how to take the failures on the chin, because they WILL happen.
2: 1630Z General Thunder and 17Z Enhanced Thunder
Most severe weather enthusiasts, storm chasers, and meteorologists are familiar with the SPC Day 1 Convective Outlook. What may be less obvious, however, is that this product is actually worked on by two different forecasters simultaneously: the Lead Forecaster, and the MesoA. The Lead Forecaster is responsible for determining the location and severity of any thunderstorm activity, from a Marginal risk onward to a High Risk, for the entire Continental US (CONUS), as well as the individual probabilities for tornadoes, hail, damaging winds, and the outlook discussion text. This can be a lot to manage, especially when there are multiple geographic regions with a marginal risk or greater… so the MesoA forecaster assists by conducting the analysis of the General Thunderstorm lines - the lightest shade of green on the SPC graphic. While I believe most people are familiar with the fact the categorical lines are based on underlying probabilities, I was surprised to learn the same is true of the General Thunder lines. These lines represent a 10% unconditional chance of a lightning strike, valid until 12Z the following calendar day.
The good news is, the Lead Forecaster and MesoA are not starting from scratch. We’ve already inherited an outlook from 13Z! We do, however, have newer guidance with the 12Z model suite and associated derived calibrated products we use internally. So, depending on the day and the number of unique general thunderstorm regions, sometimes significant changes are necessary. However, the general philosophy towards inheriting a forecast product is “do no harm”… so if you are going to make significant changes, make sure you have solid meteorological justification, and you have coordinated and conferred with the Lead Forecaster about those changes! The outlook is due at 1630Z, but depending on how active the day is anticipated to be, and how complicated the thunder lines are, these lines can be completed and sent to the Lead Forecaster as early as 15-1530Z.
While many types of observations and forecast data can go into creating and editing the thunder lines, we also have some internal calibrated guidance that is incredibly useful and accurate for this task. Developed by Dr. David Harrison with OU-CIWRO, we have calibrated thunderstorm guidance using the HREF that performs well as both a first guess and sanity check in the forecast process! If you’re interested in learning more about the science of that calibrated guidance, or are wanting some insight into how research-to-operations works at SPC, definitely give the paper a read.
3: Mesoscale Discussions
Once the thunder lines are done, the duty of the MesoA transitions to what most people expect and think of with Meso/MesoA forecasters, and arguably the most “fun” yet challenging part of the job: Mesoscale Convective Discussions (MCDs or MDs)! This is where you get to put on your short-term nowcaster/forecaster hat and really dig into the nitty-gritty of the observations of the environment, confer and collaborate with the other forecasters, and get to say what you think about the current state of the atmosphere. MCDs serve a multitude of purposes, as far as I understand currently: to alert the local NWS offices of potential watch issuance (or even lack thereof), to highlight specific corridors of enhanced severe weather threats (Meso-beta and Meso-gamma MDs), provide updates to the threat of any existing weather watches, and equally importantly, to help prime the Lead Forecaster on the environment and provide input on whether or not a watch is needed.
The process of writing an MD can actually be quite time intensive when you are first starting out. The workflows and “knobology” take some getting used to in drawing the lines, placing the text, any fronts, lows, and other helpful analysis graphics, and then shipping it over to a program called Prodgen that handles the rest. This includes the discussion text, the fancy graphics, and disseminating it from SPC to the wider world. It can also take some getting used to in order to develop a good writing style that strikes a balance between concise and detailed. You’re not writing a dissertation, and you’re not an observation/mesoanalysis printer just rattling off numbers about the environment. During the summer months, being timely with MD issuance can be especially crucial, since things not only develop and change rapidly, but as mentioned earlier, you’re often managing a large geographic area and cannot spend too much time focused on a single area! Often the goal is to get MDs out before any watches and warnings for the environments that are most likely to produce severe weather the soonest, and especially for the areas that are expected to have significant severe weather.
Once the pre-watch MDs are out, and then the watches are issued, you are then switching to a situational awareness mode that requires, once again, frequently changing geographic domains. You may have half a dozen radar sites up, and are constantly navigating between radar scans of ongoing storms and updated observations/mesoanalysis in order to assess the short-term risk of the watch areas. You also have to keep an eye out for any surprises that develop outside of watch areas, or even outside of the convective outlook bounds! It’s a game of constantly scanning, writing up a draft MD in case you might need it later, and then scanning again. As the MesoA, you’re kind of the last line of defense for the Lead Forecaster in terms of catching any issues before they become too problematic.
Also, since I get this question a lot…
There is no official directive or guidance on the meaning behind MD colors. It is completely up to the discretion of the forecaster. Some forecasters, for example, will use brown as the color for all of their MDs, regardless of severe hazard. Generally, though, a large number of forecasters will use red outlines to indicate possible tornado watch issuance, blue for possible severe watch issuance, brown for watch not likely, and purple for undecided between severe and tornado. Once again — this is not an official guideline, but a general rule of thumb. Always read the text if you want to know what’s going on.
4: Thoughts So Far…
So far this has largely been a technical discussion on the duties and responsibilities of the shift, as well as some of the challenges associated with those responsibilities. However, I also wanted to take some time to record my thoughts and how I’m feeling after my first week of working operational shifts. So far, I’ve only worked a total of 3 I4 shifts with direct supervision (a holiday in the middle of the week meant I got a nice break…), and this week I get to work another couple of I4s to really get into a rhythm before moving on to J5s. Even with just those 3 shifts, I feel like I have a different perspective and appreciation for the job than I did going into it.
For starters, I’ve been saying this for a long time, but I believe it even more now: When it comes to critiquing operational forecasts, you really have no idea what it is like, what challenges there are to face, and how hard it is to work under a time crunch and submit a product with YOUR NAME on it to the wider world, until you’ve sat down and actually done it. Without that experience of sitting in the seat yourself, it’s just nearly impossible to fully comprehend… and I say that as someone who regularly, even routinely, spends time listening and watching the operations floor. That doesn’t mean that we’re perfect, beyond correction or learning, or that we don’t critique ourselves, but I do understand now why so much input just becomes background noise. Often, the critiques completely miss the actual challenges and realities faced in issuing real time operational products. You also don’t realize just how quickly you become your own worst judge when it comes to product verification, until you’ve highlighted a region of prolonged supercell development in a favorable environment, hit send on an MD, and then have those supercells die within an hour. Nor does the feeling fully come across to the outside world when you’re just about to finish your thunder lines, and then storms develop completely outside of where you predicted, busting the forecast before you even submit it. The atmosphere is unforgiving and will humble you fast.
The firehose of data is real, and so is analysis paralysis. Not only do we have the full suite of operational model guidance and observations, but we also have a suite of experimental and next-generation guidance. This abundance of data is used to create some unique and powerful internal tools and products that are derived on top of it (e.g. Nadocast, HREF calibrated guidance). That sounds great, but it adds just one more data point to consider when issuing a forecast, and often, it doesn’t improve the clarity of the situation. It’s always easy with hindsight to say “ahh, this performed the best, so I’ll use it tomorrow”, and when tomorrow comes, it’s the worst performer. As great as the tools and guidance are, you never really know which one is going to be the best and most reliable on a given day. You still have to use your own meteorological knowledge, and more importantly, you still have to make a decision. Not only do you not have unlimited time, but unlike storm chasing, you have the whole CONUS to consider. You can’t just tunnel vision into the most interesting weather!
Personally, after working these shifts, I feel like I’ve barely been able to tread enough water to stay afloat, and I’m about a step and a half behind where I need to be at any given point. Even on the days where you’re doing relatively well and being timely, it still very much feels like a constant sprint from behind. Time management and the ability to quickly pivot are definitely some of the more important disciplines with this job, and as someone who is used to having as much time as he needs to solve problems in software land, it’s challenging. My first few MDs took upwards of 45 minutes, but my fastest was down to 15 minutes from start to finish. My goal is I’d like to bring my overall average to 20-30 minutes, with my fastest speed (as needed) down to 10 minutes.
Thankfully, I’ve had some excellent teachers to help me through all of this! Lead Forecasters Aaron Gleason and John Hart have been incredibly kind, patient, and helpful through this whole process. Jeremy Grams has also been helpful and patient from the Meso desk as I try to get the hang of things and learn how the process of dividing regions and internal collaboration works. Finally, Andrew Lyons and Harry Weinman have been excellent teachers so far! It’s still super cool to me to have been able to issue a product with Andrew after, nearly 10 years ago now, he called this very set of circumstances happening. It’s cool to see it come to fruition. I’m hoping that going into this next week, I can be a little faster, a little more nimble, and equally receptive to correction and input from my peers and mentors. Nothing quite humbles you like the atmosphere, but it helps when you have great colleagues supporting and helping you along.
