Snowfall Forecast Communication: We Need to do Better

I've written variations of this post many times over the last 5 or so years and it bears repeating. I rarely post snowfall maps/computer model projections further out than 2-3 days on social media platforms. Why?  Look at these long range computer model projections starting Monday--4 days prior--for the snow event Friday and Saturday, January 12 and 13th.

Sunday Evening 1/7 (Model #1. I forgot to save Model #2 for comparison)

Monday Morning 1/8 (Model #1 vs #2)

Monday Evening 1/8 7PM (Model #1 vs #2)

Tuesday Evening 1/9 7PM (Model #1 vs #2)

Wednesday Evening 1/10 7PM (Model #1 vs #2)

Look at the HUGE difference between each model.

Many people posted these images online and even showed them on the air. In most instances these were shown to convey the fact that we had a long way to go before an official forecast could be made with a high degree of certainty. Uncertainty was being conveyed. It has the best of intentions. I get it. I've done it. Makes sense, right?

Except in the real world, most humans rarely interpret this type of information--data--through a rational filter. Data makes us feel conflicted. We, whether we like it or not, overly simplify complex ideas. Computer model projections are a prime example. That's how we're wired. It's well documented with behavioral scientists.

We look at the numbers, pick one out--say 20" of snowfall--react impulsively to it without historical perspective (by the way that's never happened in 24 hours from a general snow event since record keeping began) or context and before you know it, hysteria ensues. We see it on social media every time. Unfortunately social media reactions spread at an exponential rate devolving as the comment thread lengthens. Before you know it, the benign, educational intent of the snowfall uncertainty post is lost. It involuntarily morphs into a totally different story that spreads like wildfire changing with every re tweet and re post. The old game where you sit in a circle an say a sentence to the person to your right and that person to their right applies here. When the message comes back around to the beginning, it's changed entirely. Do you see the problem?

All of us meteorologists here at WJW FOX 8 always choose to wait until we have a high degree of certainty before we post any snowfall numbers usually in some sort of range. The aim is to eliminate what I just described in the paragraph above. This was my post Thursday morning, roughly 36 hours before the event begins. (Note: Andre Bernier, our evening meteorologist had a more basic snowfall map the night before not shown here)

How do we combat our hard wired human nature? The answer is simple:  We don't. WE QUALITY CONTROL OUR INFORMATION!

Psychology plays a huge role in weather forecasts. As a meteorologist who reaches hundreds of thousands of people each morning on the top rated morning show in Cleveland, Ohio, I owe it to each viewer to quality control what I say and what I create visually to accompany it. The psychological element--how the viewer perceives what we say--is a critical weather cast element that is often overlooked by on air meteorologists. The words/numbers that we say/post on the air and online can be more powerful than the weather itself.  In this age of posts solely created for clicks with the hopes of a long, viral lifespan, quality is too often sacrificed. That's fine for a cute picture or a clever, humorous meme but when a major weather event is eminent over a large population area, high quality information created for the masses is of the utmost importance.


Recently the National Weather Service has implemented a program in the works for several years that streamlines their winter season watches, advisories and warnings in an effort to simplify the forecast message for the general public. The great folks at the NWS Office in Cleveland recently hosted a seminar for media personnel highlighting these changes. The information was invaluable.

As meteorologist we always need to remember the psychology behind what we do. As communicators of science (data) need to put ourselves in the minds of the public and ask: "How is the viewer going to interpret what I say?" So when I create a snowfall forecast range map, I take into account how people react to EACH WORD knowing that many will perceive these range to fit their location.

We must remember:

*  Public perception is very powerful

*  We need to be better communicators of information

*  Choice of words is of the utmost importance in conveying severity of the weather

*  Mass media is for the masses. Risk is personal to the public. Yet people want personal forecasts. Huge conundrum.

* Human beings dislike uncertainty.

*  Too much emphasis on uncertainty breeds confusion, inaction and ultimately apathy when the next snow or weather event of importance happens.  Basic psychology here.  We need to find a delicate balance between voicing uncertainty and sticking to a forecast geared to the general public without causing social media unrest

Mindset Behind My Snowfall Forecasts

Sandusky, Ohio snow on February 25, 2016.  Photo courtesy: Bryan Edwards (Twitter)

Snowfall forecasts are very difficult animals both in the meteorology that's used to create them and the public's perception of the numbers you display.

Tuesday morning, I posted a general low resolution hand drawn preliminary forecast for Thursday's snowfall. 

Preliminary Thursday Snow Forecast Issued Tuesday Morning

Yet, many people on social media interpreted this preliminary forecast too literally. Several on Twitter freaked out when they saw Lorain close to the 3-5" range contour.  I told them it probably didn't matter since its--A RANGE!  Sure, Lorain was closer to the potentially heavier snowfall but the overall difference using the range for Lorain was small, probably an inch or two. Yet many on social media see it differently. 

Yesterday morning (Wednesday), I adjusted the area of heavier snow further west but KEPT THE RANGES THE SAME. Once again, people on social media interpreted this as a huge alteration in the snowfall forecast to which I replied, "Use the ranges NOT absolute numbers".  In reality, the overall forecast within the confines of our ranges remained the pretty much the same aside from shifting the heavier amounts west a bit. Yet the perception by the some was that Lorain went from five inches to one inch.  

Most meteorologists on television, the public sector (National Weather Service) and in private industry use snowfall ranges in their forecasts. This works best because it takes into account the variance in movement and intensity of the many individual bands of snowfall within snowfall systems. The problem for the meteorologist is that most people want A SPECIFIC SNOWFALL NUMBER for their backyard.  Yet many people see a snowfall forecast during the morning newscast and believe that this holds for the entire day. Nothing is further from the truth. These forecasts are not one-and-done. They evolve throughout the day as the event unfolds. Forecasts change as the conditions change. People see individual snowfall amounts on their phone apps and believe these over human derived forecasts. They want absolute specifics yet more often than not, one snowfall amount number for any location will not work.

For example, here was my official snowfall forecast issued Thursday morning at 4am for the entire day using ranges:

One specific (WRF 4km) high resolution computer model early in the morning cranked out these amounts for Thursday's snowfall.  While these were in line with our official snowfall forecast above, people fixated on some of the individual numbers and were confused. 

How do we combat this confusion on social and broadcast media? Good question. 

My soft policy is that I rarely post computer model snowfall output unless we are inside 24 hours before a snowfall event. If my intention is to highlight the general outlook for the week ahead, I will post snowfall accumulations without numbers.  A quality controlled "hand drawn" map is another option. My ultimate goal is to present information that describes the weather without creating confusion.  

Situations like this reminds me of the psychology behind the weather forecast. Most people want exactness in there forecasts. Remember that we are all hard wired to simplify uncertainty. So when a snowfall forecast range is posted on a map, people immediately find a number within that range that best fits. I'd like to say that I make a forecast with a cold, rational eye but I don't. I take into account how people with react to EACH WORD knowing that many people with perceive selective elements to fit their location.  I learned that real quick after my first major lake effect event twenty years ago.

If I could make a poster with bullet points for meteorologists, it would list these five at the top

*  Public perception is very powerful

*  We need to be better communicators of information

*  Choice of words is of the utmost importance in conveying severity of the weather

*  Risk is personal (public).  Mass media is for the masses. Yet people want personal forecasts. Huge conundrum.

*  Too much emphasis on uncertainty breeds confusion, inaction and ultimately apathy when the next snow or weather event or importance happens. This is basic psychology that's been well documented over the years. We need to find a delicate balance between voicing uncertainty and sticking to a forecast.

Human Nature and the Uncertainty in this Weekend Forecast

A hurricane off the east coast, strong cold front, gulf moisture, Atlantic moisture. These are just a few of the factors that will drive the rain and wind this weekend. In my opinion, one of the most difficult forecasts in recent memory.

The various computer projections are taking Hurricane Joaquin and the other tropical disturbances in all sorts of directions as it begins its interaction with the other variables above. Each line represents a different computer model representation of each outcome.  Are you confused yet? 

Our minds don't easily handle probability. We especially hate probabilities in our weather forecasts especially when confronted with a map like this.

Why is this?

For our minds to grasp probabilities, we need to be able to handle multiple possible outcomes at once. Just our luck. Weather has many, many outcomes over a large area over a significant period of time. Change the initial weather conditions (humidity, wind flow, frontal position, upper level energy, etc) and you create more uncertainty. Factor in time and the probability becomes significantly higher.

Typically, our brains work much better with a theme that is linear: A story that has a beginning, middle and an end. We want to visualize a line of showers that moves in at a specific time, stays for a select amount of time and then moves out without fanfare. Unfortunately, rain events rarely behave in this manner.

Much to the chagrin of the general public—probabilities are the only way to tell the weather story. We use 90% chance of rain, 40% chance of rain, etc. Yet if it doesn't rain over their house when the probability is 90% chance of rain, the forecaster is wrong even if the rest of the area was hit with a good downpour. We want to know if it will rain or not; a black and white scenario without caveats. Yet the behavior of some small scale weather events like warm frontal rain/storms can behave semi-independently of the overall large scale pattern. Meteorologist try to convey this idea on the air. Most of the time this falls on deaf ears.

It all goes back to basic human nature. A good weather narrative (a feel-good forecast with some folklore) is desired versus something data/science driven. Nebulous weather data and science makes most of us feel uncomfortable even if the on-air meteorologist has the best of intentions.

In recent years,  some highly sophisticated models of the atmosphere have been developed that can make some very good “probabilistic” outcomes for weather events. Yet a level of uncertainty still remains and we humans don’t like it!  We try to rationalize the irrational. Our biases quickly dismiss the probabilistic science as irrelevant or at the very worst, an excuse.

Instead, we favor more simplified stories even though that story might gloss over important details. Our minds involuntarily cherry-pick elements of the story so that it fits our biases. Think of a time when someone told you a weather fact or forecast which you didn’t believe. You felt uneasy. Your mind shrugged it aside only to be replaced by a story, forecast or explanation that made you feel better…accuracy be damned.

* Narratives (straight forward simple weather forecasts) are about hitting emotional buttons making the reader feel good by focusing on less qualitative aspects (weather science and probability) of an issue.

* Narratives (weather forecast) are/is about the outcome not the process (explanation of the science and probability)

* The process (weather science) is important in developing solid results

So remember the psychology. How you react when you hear a weather forecast?  Do you dismiss the science? How do you handle probability?  Do you like hearing an explanation to why the weather does what it does? Do you overly simplify the weather? Are you aware of your biases?

Why Are We Skeptical About Weather Forecasts? How To Fix It

(2nd Edition)

Weather is just as much psychology as it is science. Call it "Behavioral Meteorology".

Most people find it very difficult to grasp the fact that the weather is one big approximation. Not surprisingly, we humans hate approximations and probability. Why? For our minds to grasp probabilities and randomness, we need to be able to handle multiple possible outcomes at once.  The problem is that we are all wired to simplify uncertainty. We want life to be basic and easy to understand. Weather is no different. We all want a forecast that fits a nice and neat one-size-fits-all package.  Unfortunately, weather has many, many outcomes over a large area over a significant period of time. Change the initial weather conditions (humidity, wind flow, frontal position, upper level energy, etc.) and you create more uncertainty. Factor in length of time and the probability becomes significantly higher. We envision an area of rain approaching as a uniform “blob” which moves over our house at say 5:20pm and leaves at 7:15pm. Unfortunately, the actual rain area (or lake effect snow stream) rarely evolves into a tidy, uniform entity.  Instead, it has jagged edges, dry pockets and other random protrusions that impact local forecasts in a few minutes time.  See the problem?

I'd like to say that I make a forecast, short or long term, with a cold, rational, scientific eye but I don't.  I take into account how the general public will react to EVERY word knowing that most people selectively perceive the weather to fit their "sphere of reality". The more we selectively perceive the weather to fit our negative connotation (cognitive bias), the more hyper critical our reaction and the more rigid our bias becomes. It’s a vicious circle that feeds on itself. Throw in the thousands of weather apps out there that claim to provide the forecast for YOUR location along with the Old Farmers' Almanac and the laundry list of cognitive biases (some mentioned above) and you have the confluence of many psychological elements that are difficult to overcome with rational discussion. 

How do we navigate these rough waters of perception and weather bias? 

It all goes back to basic human nature: We all simplify complex, probabilistic themes. We all love a good story with pictures. It’s hardwired in our DNA. Bullet points and/or data without strong associative elements only trigger the Broca's Area and Wernicke's Area of the brain. This area is involved in language processing ONLY.  Brain scans show that if you incorporate emotional elements into your story telling, it will active multiple sensory parts of the brain like the Motor Cortex (body movements) and the Insular Cortex (emotional region) all at once producing chemicals that make us feel good in response specifically Cortisol, Dopamine and Oxytocin. Each one responsible for our involuntary reactions to a good story. We instinctively turn the story into our own personal experience! Given that personal stories make up more than 65% of our conversations, this makes perfect sense.  Here's a great quote that summarizes this up very well:

"THE POWER OF ANECDOTE IS SO GREAT THAT IT HAS A MOMENTUM IN AND OF ITSELF."..."NO MATTER HOW BORING THE FACTS ARE," WITH A WELL-TOLD STORY, "YOU FEEL INHERENTLY AS IF YOU ARE ON A TRAIN THAT HAS A DESTINATION."

A weather forecast is no different. Your brain wants a good story not boring data. This is why we use colorful, pictorial driven graphics like this one. 

Often times if the weather presentation becomes too data intensive, our brain doesn't produce the chemical that make us feel good. Our biases can surface quickly which can alter our perceptions fast. All of this happens in seconds and you've lost the viewer even if the on-air meteorologist has the best of intentions. !

Spaghetti Plot showing forecast uncertainty

Bottom line is a narrative or story is desired versus something that is solely data/science driven. 

Each day, I analyze the science and remember the psychology. I try to tell a compelling, relatable weather story with a dash or two of data, some description of probability and a bit of historical perspective. Human nature is a powerful beast. Each person is different. Sometimes it works for the viewer. Sometimes it doesn’t. 

How do you react when you hear a weather forecast? Do you dismiss the science? Do you like the story? How do you handle probability? Do you like hearing an explanation to why the weather does what it does? Do you overly simplify the weather?

Let me know what you think.