Laura Bliss is CityLab’s west coast bureau chief, covering transportation and technology. She also authors MapLab, a biweekly newsletter about maps (subscribe here). Her work has appeared in the New York Times, The Atlantic, Los Angeles magazine, and beyond.
A biweekly tour of the ever-expanding cartographic landscape.
Welcome to the latest edition of MapLab. Sign up to receive this newsletter in your inbox here.
Forecasting the worst
The east coast of the U.S. is bracing for Hurricane Florence, predicted to make landfall Friday as the most dangerous storm the Carolinas have seen in decades. Historic rainfall, storm surge, and winds threaten millions in the storm’s path. Leaders in North and South Carolina and Virginia, the areas expected to bear the worst impacts, are taking the storm’s threat seriously, with evacuation orders encompassing 1.7 million people across the region.
Emergency officials know when to sound the alarm thanks to researchers and mappers practicing the age-old science of weather prediction. Human beings have been watching and interpreting signs of rain and drought for millennia. Today, in the Anthropocene, the perilous consequences of climate change create a strong imperative for ever-more accurate weather models. In the age of big data, scientists seem to be making them.
In fact, next year, the U.S. government will officially switch over to a new model of global storm forecasting, called FV-3, in order to predict more complex storms with greater speed. Under development by NOAA scientists since 2012, FV-3 will replace the current model, called GFS, with a new level of accuracy and computational power to simulate complicated atmospheric processes both locally and globally.
Now, as this enormous storm bears down on the coast, all eyes in the weather prediction community are on how FV-3 performs compared to GFS. “Which will be right? You will know in a week,” wrote Clifford Mass, a professor of atmospheric sciences at the University of Washington, on his personal blog.
Earlier this week, Mass posted a side-by-side comparison of how the two models are charting Florence’s path. Statistically speaking, FV-3 has generally outperformed GFS on accuracy tests, Mass writes. But he highlights a huge difference in their pressure and intensity forecasts for Florence, which could have major implications for how the storm plays out. I followed up with him via email.
You wrote on your blog that Superstorm Sandy in 2012 alerted weather forecasters to the need for a new national model. Say more about that?
Sandy was a wake-up call to the U.S. that our nation had fallen behind in global (and hurricane) weather prediction. We are still behind. Replacing GFS with FV-3 is an attempt to catch up, but the model is only one component of the system. U.S. operational prediction lacks sufficient computer resources—NOAA and the National Weather Service could really use 100 times what they have today. For the price of a few fighter jets, the U.S. could have vastly improved weather forecasting.
Why do we see such a significant difference in estimated sea pressure between the two models for Florence? FV-3 shows much higher pressures, which would imply a much more intense storm.
FV-3 should be a more modern, superior model to GFS, but the two runs are being driven by the same data assimilation system and analysis, so that is why their two track forecasts are similar. But disturbingly, their forecast pressures have been very different. For the forecast initialized September 9, 12z, GFS has been much better—lower pressure (see graphic). Why? That is an important question.
What will you be looking for during and after the storm to assess the accuracy of the new model?
I will be comparing track forecast, intensity (lowest central pressure, highest winds), and precipitation totals.
Why build better weather models?
The stakes are huge. Better weather prediction saves lives and property, and has immense economic implications. In a period when we are worrying about more extreme weather impacts, the first line of defense is better weather prediction.
NOAA isn’t the only one working on better storm modeling. Oceanographers at the University of Rhode Island are improving forecasts for New England’s particularly complicated storms. And Columbia University researchers are developing systems that blend data from historic weather events with gigantic storms that haven’t happened yet but could—i.e., the storms of the climate change era.
An uneven impact
No one can know for sure how hard Florence will hit the coast until it does. But one thing is certain: the damage won’t be evenly distributed. Today on CityLab, I wrote about the particular challenges that come with responding to mega-storms in rural areas, while my colleague Nicole Javorsky highlighted a selection of maps showing the likely inequitable outcomes for those who don’t evacuate. She writes:
Location is an obvious differentiator—but not the only one. Factors like socioeconomic status, age, whether a person has a disability, whether or not they own a car, and what languages they speak will also determine how easy or difficult it is to survive and recover from disasters like Florence.
A dashboard of interactive maps from the emergency response nonprofit Direct Relief shows “the range of social vulnerability in Florence’s projected path,” she continues. In the map below, warmer shades represent populations that are more likely to need support in this emergency, due to the factors Javorsky lists; cooler shades indicate areas with a lower probability of need.
Maps are a helpful way to wrap one’s head around a storm of this magnitude, whether it’s NOAA’s rainbow bright, yet dire hurricane models, FEMA’s (dangerously un-updated) flood risk maps, or maps like this, that suggest how society’s existing inequalities are primed to become more deeply ingrained in the face of disaster.
What kinds of questions do you have ahead of a frightening storm like Florence? What types of maps would you want to see in the lead-up and aftermath? Let me know.
Is where you are, who you are? A writer sifts through his location data to find out. ♦ Why is public transit in America so bad? A map-heavy CityLab opus answers the question. ♦ Don’t trust that map: it could be a vehicle for fake news. ♦ Don’t breathe that air: Google Street View cars are gathering urban pollution data. ♦ It’s in the stars: restaurant reviews can predict gentrification down to the neighborhood. ♦ Not-so-next-door neighbors: Chicagoans on opposite sides of the city find their “map twins.” ♦ Jason Derulo knows what the girls want, New York to Haiti, London to Taiwan. But how would you map that?
Stay safe out there, MapLab readers. And if you’re not surviving a hurricane, share this newsletter with a friend—they can sign up here.