Following what is perceived as a blown forecast of epic proportions for New York City, weather professionals are doing plenty of soul-searching. For a community that works tirelessly to get it right and takes enormous pride in its work, a failed prediction on the nation’s biggest stage is a penetrating blow to the gut.

But meteorologists are a passionate and resilient bunch, determined to bounce back up after getting knocked down.

A flurry of commentaries and comments – attempting to clarify and make sense of what happened and recommend what we can learn from it – are floating through the universe of digital media. The meteorologists and professional weather writers sharing these views make a lot of great points – which I thought I’d try to synthesize.

Weather forecasting has made enormous progress, despite this black eye

When I started my career, these types of storms were at times not even predicted a day in advance.

– Louis Uccellini, director of the National Weather Service

I can tell you story after story of using … models along with “rules of thumb” … in the 1970’s through the mid-80s and confidently forecasting “four to eight inches” and waking up the next day to absolutely dry streets and clear skies. We had no idea what a “dry slot” was. There were also heavy snow storms that went unforecast. What progress we have made!

The forecast, while bad along the southwest fringe – from New York City to Philadelphia – was excellent for southern New England

This was a glowing success for forecasters in Eastern New England. The models did well there for the most part, and we saw a ferocious monster of a storm, as was predicted.
Many journalists, bloggers, and meteorologists donned their rear-view goggles … to analyze what went wrong. Most of them … correctly observed that in the big picture, this wasn’t a bust at all. NWS computer models and public outlooks captured the nor’easter as a whole–including its extreme intensity–remarkably well. The devil lay in the details of predicting where the storm’s west edge would be, as a dramatic gradient in snowfall totals (which forecasters did anticipate) ended up being overlaid atop the nation’s largest metro area.

Let’s not play the blame game

Blame is a pretty harsh critique and implies that one individual or group caused an action. No one should blame the meteorologists, the media, the politicians, or the weather models. No one deserves blame for today because today’s snowtastrophe in New York City/NJ was a combination of a lot of issues we all have to contend with, regardless if we’re in the media, in the private sector, a viewer, a consumer, a decision maker, anyone.

Erring on the side of caution was better than the alternative

In the face of uncertainty, the right thing to do is almost always going to be to assume the worst-case scenarios. Maybe some decisions could have been different given what was known at the time (in particular, the subway shutdown might have been unnecessary even if the snow totals had come in as high as forecast). But it’s still way too easy for those of us with no responsibility to play Monday morning quarterback. And overall, given the uncertainty of the different models — and the fine line between the correct call and an inaccurate one — the authorities look to have taken the right approach in being proactive.
Assume for a moment that Manhattan received 9” of snow that was unforecast. Absolute gridlock would have resulted…
Our NYC forecast, while hardly perfect, was useful.

Still, there was too much focus on the worst case scenario

… there was plenty of evidence this storm wouldn’t be record-setting in New York…
Forecasters would have been better off averaging all the model data together, even the models that don’t have a stellar record. The Euro[pean model] is king, but it’s not so good that we should ignore all other forecasts.
When a forecast is so sensitive to small changes (eastern Long Island, not far away, received 30-plus inches), it is imperative to loudly convey the reality that small changes could have profound effects on what actually happens.
Unfortunately, the hype surrounding the historic possibility drowned out the very real scenario that the storm could underwhelm.

Overly strong language from forecasters overstated confidence in the threat

One of the bigger issues I had about this forecast is the choice of words [from the National Weather Service]. Strong words such as “historic” or “epic” or “crippling” shouldn’t be go-to words to talk about a huge weather event unless certainty is extremely high. Whenever NWS uses these words, it definitely gives fuel to media/journalists to use them to add on to the “hype.” Perhaps we should use “memorable” instead of “historic” for storms like this? I think many will remember this storm. However, historic should only be given to one out of 100 year storms such as the Storm of the Century (1993), Hurricane Katrina, or Sandy.

– Matt Daniel, broadcast meteorologist via email

We must interpret and communicate uncertainty information better

…if this storm points to a failure, it’s really a failure not of the specific forecast, but of the format and language that is used to communicate all forecasts, an approach that fails to clearly spell out uncertainties and the difficulty of picking one of a number of potential storm tracks.
…there was roughly a 70% chance that the official forecasts were too high. Instead of predicting 2-3 feet, consider if that forecasters had said there was a 30% chance of more than 2 feet, a 35% chance of 1-2 feet, and a 35% chance of less than a foot. No one would be writing critical headlines if they had done so and decision makers would have gotten far better information. And the fact that forecasters used terms such as “historic storm” was guaranteed to push the media into a feeding frenzy.
The real failure IMO was to appropriately interpret the information contained in the forecast ensemble…
Learn from your mistakes, understand uncertainty, and be prepared to state when you have low confidence in your prediction, or when there are two equally probable scenarios.  The NWS needs to get away from making deterministic forecasts

Communicating uncertain forecasts is hard

There is more risk and nuance in weather forecasts than the public is interested in consuming so it is a challenge to craft a message that gets attention, is not “hype”, yet has actionable information. We must continue to have the discussion about how to communicate uncertainty and risk effectively.
This is an imprecise science in a world that covets precision.

– New Jersey state climatologist David Robinson (via Climate Central)

The National Weather Service needs to evolve how they communicate forecasts

I think the NWS needs to break this regimented product setup. I think at times their hands get tied. They can’t be nimble. And they need to be. Products should target impacts and be able to highlight those impacts and risks above anything else.
….the NWS has to move to a much more probabilistic form of forecasting preparation and dissemination, one in which forecast uncertainties are made clear to users. The computer workstations used by NWS forecasters and NWS websites are not designed to facilitate probabilistic prediction. This needs to change.

Additional good reads:

Nick Wiltgen, Winter Storm Juno: Did We Get It Wrong?

Andrea Thompson, Climate Central: What Can We Learn from the NYC Forecast ‘Bust’?