Can thousands of amateurs make an important contribution to climate science that is eluding experts in the field?
“Yes” is the resounding answer from Chris Hennon, UNC Asheville associate professor of atmospheric sciences, and he has 360,000 pieces of evidence.
Two years ago, Hennon and a team of distinguished scientists launched the online Cyclone Center – a website offering amateur volunteers a chance to help perform a task that is too large for a small group of scientists – estimating the strength of thousands of tropical cyclones that have been observed by satellites, but not measured by aircraft surveillance or ground-based instruments.
Storm images from cyclonecenter.comMore than 7,400 citizen-scientists around the world have now participated. Using the easy online tools to analyze the colorful satellite images of cyclones on the cyclonecenter.org website, they have completed more than 360,000 storm classifications. And, says Hennon, “the non-expert citizen-scientists, taken as a group, have shown skill in analyzing cyclone images comparable to that of experts.”
Analyzing the strength of past cyclones is far more than an academic exercise. Information about past cyclones is used in models that predict the risk of future catastrophic storms, which is why Hennon has just received a $25,000 grant for his work from the insurance industry’s Risk Prediction Initiative.
The project is significant enough that the Bulletin of the American Meteorological Society has accepted a paper about it authored by Hennon and his colleagues, who include Kenneth Knapp and James Kossin of the National Climatic Data Center (NCDC), along with Carl Schreck III and Scott Stevens of the Cooperative Institute for Climate and Satellites-North Carolina (CICS-NC). The article is available here in its early online release.
Hennon’s team turned to “crowdsourcing” because the number of cyclone analyses needed is so huge – nearly two million cyclone classifications are needed to complete the data collection portion of the project. It would take one person roughly 25 years working full-time, without any vacations or holidays, to accomplish that task.
But with thousands of citizen-scientists pitching in a few minutes a week, there is hope that the two million analyses will be completed within a couple of years. And although it seems counter-intuitive, using a large number of non-experts promises greater accuracy than employing a smaller number of experts. Hennon points out that expert agencies have often disagreed in their estimations of cyclone intensity, so reanalysis is needed.
“Instead of using a small number of experts, we tap into the scientific curiosity of thousands of ordinary people, using their enthusiasm, time and pattern-recognition skills,” says Hennon.
Classifying a storm online can be done easily in just a few minutes after reading the brief tutorial which uses illustrated examples of actual satellite images. Citizen-scientists look at the storms’ spiral patterns, and patterns of colors which represent the divergent temperatures found in different sections of each cyclone.
When the citizen-scientists disagree about a storm, its classification is given a low-degree of confidence in the Cyclone Center model. When many non-experts agree on a cyclone’s intensity, Hennon’s team places greater confidence in that assessment.
The project is entering its third year with a newly redesigned website. “It has been a very exciting time for us and the thousands of citizen-scientists from all over the world who are helping us resolve some of the biggest questions in the global tropical cyclone data record,” said Hennon. “With our new grant, website and continued support from UNC Asheville, CICS-NC and the NCDC, we are looking forward to an even more successful third year.”