March Snow Post Mortem and The Great Expectations

As I have mentioned previously, there are some people I cannot reach.  People who have made up their minds about weather, media, hype, and me.  We have reached an age where everyone appears entitled to their own truth.  If you want to badmouth me, you have the right.

But I think we did a pretty good job.  At least locally.  But the nature of this business reminds me that I may not be so lucky next time.

Great Expectations

Media hype builds because weather drives the news. That is not going to change any time soon. It is the one thing that affects everyone.

In Virginia, people really want to know when it will snow. There are huge practical and emotional consequences that go with accumulating snow in the Commonwealth.  

Sometimes a big snowstorm is an option 2-3 days in advance, but with the available data, it would be just as likely that the worst of the storm misses by 50 miles.  Both options can be equally valid.

But with little exception, if a meteorologist gives a best guess of 4-8″ of snow a few days ahead of time, the message the public takes away, and yes, the one most of the newsrooms across the country sends is: Up to 8 inches of snow is expected!… which then goes viral from the water cooler to the internet: They say we’re gonna get 8 inches of snow!

The Democratization of Weather Data

As a result, like most of my colleagues, I hate the idea of giving a snowfall range of amounts 4 days before the onset of a storm.  Hate it.  But we have reached an era where we have little choice.  As my friend, Matt Lanza pointed out in his blog, the democratization of weather data has changed the game entirely.

Certainly, it is great to have so much of the numerical guidance online.  It is wonderful for both professionals and hobbyists to have such open access: GFS, NAM/WRF, CMC, UKMET, ECMWF, NOGAPS, JMA, HRRR, FIM, SREF.  But when I see someone post the models say such-and-such, or some of the reliable models say deedle-deedle-dee, it is hard to take the post seriously.

So this open fire hose of data does have its drawbacks.  Lots of people are tweeting and starting their own facebook pages dedicated to weather, issuing their own forecasts.  But here is the next question: Who do you trust?

Sure, it’s fun beating the professionals. And there are a number of weather hobbyists that command my respect.  Keith Huffman runs a weather blog here in Lynchburg, and he is one of them. But that respect is earned.  Throwing a piece of weather data out there without some analysis reminds me of the old adage, “A little knowledge is a dangerous thing.”

I would encourage you to find a few internet weather sources and stick with them.  There is no shortage of opinion out there.

I often get questions about what so-and-so says on his blog, why The Weather Channel says what it does, or what’s the problem with the National Weather Service.

Or my favorite:  What are they saying for this storm?  Who are they, anyway?  I cannot speak to why another outlet has made the decisions they have made.  I’m not in their circle.

He Said, She Said

I usually have colleagues posting as well, and I may have a different analysis than them. But I do not want to get in the habit of sophomoric arguing about who is going to be right. We will all find out when the weather event has finished.

In storms like this past one, there were huge uncertainties for largely populated areas. And don’t confuse uncertainty with ignorance.  The entire meteorological community knew northwestern Virginia was going to get drilled with snow.

But like it or not, sometimes the data does not point to an obvious solution, even up to a few hours before the predicted event gets started.

That is when it is up to the individual meteorologist (or team) to use their experience to sort through the data and make a call.  Forecasting is a fluid process.

There is an analogy to medical diagnosis.  Get the data, make the best decision you can from it. Sometimes the answer is obvious, sometimes it is not.  When more data becomes available, add that to your analysis.

For the user: Don’t like what you hear?  Go get another opinion.

The Trouble with The Method

I am getting concerned that there is too much jumping around on forecasts every time a new model run comes out.  Heaven knows that I am guilty of it at times, but I made a conscious decision not fall into that hole with this event.

Four days before the first flake fell, the storm was still in the Gulf of Alaska, but there was good evidence that a plowable snow would fall in our part of the state.  And there was also strong evidence that there would be a sharp cutoff from north to south across the western half of the state.

So, faced with the escalating questions from social media, I thought it was best to make a guess.  Four days out, I set out to minimize errors.  I said 3-6″ for Greater Lynchburg and held it despite several shifts in model data through the weekend.  Finally backed off to 2-4″ about 24 hours before the snow began.

There will always be fluctuations in snow depth, even over distances as little as a mile; my hope is to have the public prepared, even if it does not work out precisely the way we forecast (ask the meteorologists in Richmond and Washington about that… they took it on the chin way more than I did).

Lynchburg City received 3 inches. I measured several times in my yard Wednesday morning (near Timbrook Park in Campbell County), and we reached 3.5 inches before the snow began to compact and melt.

I started the blog so that I had a record of the forecast process.  I am continually surprised at people telling me that I said something that, in reality, never came from my keyboard or my lips. The blog offers me an opportunity to document the process, and quite frankly, defend myself.

Meteorologists are like umpires; if we do our job well, no one notices.  Mistakes are magnified.  Believe me, the day after a missed forecast, I feel it in my gut the rest of the day.

Judgement Day

Below are two maps. The first was the forecast of record at 6pm Tuesday evening.  Snow began in Lynchburg about 4 hours later.  The second graphic is a rough estimate of snow reports.  The reports came from social media and the National Weather Service offices in Blacksburg, Sterling, and Wakefield.  There is some smoothing, but the map does fit the reports pretty well.  Color tables are the same for both maps.

Some places I did well.  Some not. Your mileage may vary.

SnowForecast6pm

Final forecast before the storm hit. 6pm Tuesday evening.

Verify

Snow reports. Values indicate the maximum amount of snow on the ground before melting began.

Advertisements

About seansublette

Meteorologist at Climate Central. Broadcast meteorologist in Virginia from 1995 to 2015. Born and raised in Richmond, VA. Penn State alumnus. Loves baseball and the rock band Rush. Views are independent of my employer. Past performance does not guarantee future results.
This entry was posted in Sociology, Weather Communications and tagged . Bookmark the permalink.

7 Responses to March Snow Post Mortem and The Great Expectations

  1. lynn says:

    Thank you so much for all you do. People need to understand that sometimes the slightest shift can make a very big difference in the weather.
    I love how you referenced Richmond. When talk of the storm first started, I messaged my friend in Richmond about the storm coming. She said it was just supposed to rain there. Funny, when all was said and done, she had some beautiful snowy pictures. Me? The ground wasn’t even completely covered where I live.
    Thank you.

  2. Diana says:

    Mr. Sublette, thanks for writing this. We love following your weather. I really believe you guys do the best you can with the information you have at the current time. You do a great job and we appreciate your expertise!! It’s so true that media/social networks builds things way beyond the reality. It’s fun to become engaged with it, excited and expectant of how the newest “weather drama” will affect us. It’s great to have the interest but it’s too easy (and unfair) for interest to turn into treating the Weatherman’s educated predictions as God’s Divine Ruling, pointing fingers when our reality doesn’t end the way we wanted it to. If people want to be guaranteed large amounts of snowy weather, VA would not be my top choice of places to live! (If only people were as sold on Jesus’ words as they were on the weather predictions …. 😉 )

  3. Elliot Broyles says:

    Didn’t know you had a blog Sean. I always tell people you’re the most reliable in town because of what you used to tell us on the radio…the process that went into the forecast…Like you put all of your experience and research into each forecast and If it comes out wrong it’s because of some fluke of nature. Now I’ve got my favorite two weather blogs, Kevin Myatt and yours.

  4. Doug G. says:

    Sean, I follow weather avidly and have lots of experience with it, being an ancient 61 years old. I live in the Roanoke area, and comment very often on Kevin Myatt’s blog. The big thing that I have noticed, especially in recent years, is that when a group of people see a model run 3+ days in advance that is what they want, they sort of “hook” onto it. It is what they want, tarnation, and they will endorse all those forecasts, forecasters, and models that further endorse it. What Kevin has taught me is that 1) to follow the trends of the models. (Before 12-18-09, ALL the models kept increasing the expected snowfall amounts). With this storm, the vast majority had the snow moving away from SW Virginia. (2) Pay huge attention to 3 things for a snowfall forecast …. A. moisture levels, both before the event and the expected amount during it. B. Temperatures at the surface. and C. Temperatures aloft. Unless all 3 can support snow, there may be real issues that would limit how much falls and sticks. (3) BE REALISTIC!!! Focus on what is happening, not what we want to happen. I think you are a fabulous meteorologist, and plan to start watching “13” a lot more now for your weather outlooks.

  5. wdbrand says:

    Agree. My son watches 13 a lot since he works at the lake. Says you are right a lot more than Roanoke. Super job on the forecast.

  6. Pingback: Snow Recap - RVANews

Comments are closed.