It’s a popular opinion among meteorologists: the less chaotic the weather, the more boring the forecast. Fortunately, these guys strongly disagree.
Emerging from the difficulties of planning good sunset photos, Jake DeFlitch and Ben Reppert, recent graduates of Penn State University, and current student Steve Hallett have developed a tool to make it all easier. Now, you don’t have to be a meteorologist to tell whether clouds will be just right for a beautiful sunset. You can leave it up to the model.
“While I was photographing, I liked to grab some landscape shots around the campus,” said DeFlitch, who graduated with a B.S. in meteorology in 2015. “I usually tried to do sunrise or sunsets, and it didn’t always work out as I hoped and planned.”
So they put their heads together to try and solve this problem. “I had the idea in the back of my mind for a while,” Deflitch said. “We connected with Steve Hallett, and we knew he was one of the best computer programmers we knew, especially in the area, and it just kind of took off from there.”
The result is a model-based forecasting tool which displays the chances of seeing a vivid, average, or poor sunset in a specific area. Using data incorporated from the high resolution version of the NAM they devised an algorithm which incorporates some of the most important aspects which lead to a colorful sunset.
The variable carrying the most weight in the model is moisture, and since clouds are a direct result of moisture in the atmosphere, analyzing the cloud cover is particularly important. High clouds (such as cirrus or altocumulus) are typically composed of smaller ice particles and refract and scatter the sunlight to produce a more vivid spectrum of colors, differently than larger water particles associated with lower clouds.
Clouds closer to the surface are considered less in the model, because they limit the amount of sunlight which can be refracted to produce vivid colors, and often times these clouds create an overcast sky which blocks the sunlight outright.
The results of the project have led to discovering some unanticipated confusion in how people interpret weather models. In particular, people don’t understand Z-time, or UTC, which all weather models and data follow. After some iteration, the team made it easier for people to tell what hour sunset would be occurring in their location.
Social media has also been a major benefit. Early in the project, they would have to drive to the actual location they were forecasting a vivid sunset, to see if their forecast was right or wrong. Now that they have launched, Twitter has made the verification process so much simpler and become a useful tool for them to incorporate further in the future.
“Hopefully we can then take it further, and get the public more involved,” DeFlitch said, “having a way for the public to send us photos and print them out on a map-type platform so they can see each other’s photos taken across the country and across the world.”
Ultimately, the group views the project as a citizen science and larger social science project which incorporates aspects of the physical science behind weather forecasting to better engage the public. “We hope to offer something they can use to get away from their electronics or maybe take a minute and step away from their work and enjoy what the weather has to offer,” DeFlitch said.
Adam M. Rainear is currently a Ph.D. student at the University of Connecticut studying the social science side of weather and climate. He is researching how social media messages are communicated in the weather community, and studying new technology — such as robotics and virtual reality — to communicate environmental risk messages.