The Multidisciplinary Nature of Risk Prediction
Many fields of science require, by their nature, a multidisciplinary approach. The field of catastrophic risk prediction—in which scientists combine information about forecasting, economics and both current and past climate conditions to help insurance companies deal with global unpredictability—is a perfect example. Mark Guishard, head of the Risk Prediction Initiative (RPI) at BIOS, appreciates the need for diverse expertise to address complex issues that cross international boundaries. Not only does he have a background that includes 15 years with the Bermuda Weather Service and time as a reinsurance analyst and risk modeler, he also mentored RPI interns this summer with three different academic careers and plans for the future.
Michael Johnston
Not many people know what career path they want to follow once they leave high school, but Michael Johnston decided his after Hurricane Fabian hit Bermuda in 2003. “After the hurricane I started saying ‘weatherman’ and it stuck,” he said. Johnston also gave a nod to the American Weather Channel’s meteorologist Jim Cantore for influencing his career choice. “Everyone wanted to be him,” Johnston said. “He was always chasing storms and doing reports from the middle of tornadoes and hurricanes, standing on sea walls with waves crashing behind him. It seemed very exciting.”
BIOS’s Bermuda Program fanned the flames of his interest in meteorology. The unique summer internship opportunity pairs Bermudian senior school and university students with BIOS scientists for paid, hands-on summer research experiences. This year marks the Bermuda Program’s 40th Anniversary and the fourth year that Johnston has participated in the program.
For three of these four years, Johnston has worked with Guishard studying extreme precipitation and severe weather in Bermuda. Meteorologists rely on a combination of weather observations, computer models and long-term climate data to produce reliable and accurate forecasts. Johnston’s current work focuses on combining data sets from multiple sources to better inform the existing computer models that predict extreme precipitation events.
“Michael’s work not only benefits our academic understanding of these hazards, but will also benefit Bermuda by bringing new insights to local flood risk,” Guishard said.
Johnston was first introduced to the Bermuda Program by his science teacher at Saltus Grammar School, where his experiences at BIOS guided his university preparations and focused his coursework on relevant subjects such as chemistry, physics, and math. Later, while working on his undergraduate degree in meteorology at Penn State in Pennsylvania, Johnston realized that his summers as a Bermuda Program intern were still paying dividends.
“The Bermuda Program gave me an edge when I started doing practical work in meteorology,” he said. “I didn’t struggle as much as some of my peers because I had the hands-on experience already.”
While he came into the program with a clear sense of purpose, Johnston recommends the experience for students interested in a variety of fields. “I thought there would be no place for me to work on meteorology in a place that focused on marine biology, but I quickly found out otherwise,” he said.
Johnston was recently accepted into a doctoral program at the University of Reading in the United Kingdom, where he intends to do his thesis research on island-induced convection in Bermuda. During this process, air warms over land and rises into the atmosphere, and is often a major feature in tropical meteorology. Bermuda and other islands serve as “pockets” of warmth that create instability and extreme weather events, including thunderstorms, hail, and tornados. Through his research, Johnston will work to improve forecast models and how they handle convection, using Bermuda as a model.
True to his nature, however, Johnston is already looking ahead. “With a Ph.D. there are lots of things I could do, such as become a forecaster or work for the military. But my main plan is to start working for one of the modeling companies in weather forecasting or risk management.”
Elaine Lu
As a child growing up in Xinjian in the Western province of China, Elaine Lu’s parents recognized her early talent for learning languages. Her father, a petroleum engineer, was particularly keen to ensure this talent didn’t go to waste and, when the time came for Lu to attend college, he encouraged her to apply abroad for the experience. Lu also wanted to see the world and in 2013 moved to Canada to attend the University of Toronto, a popular choice for students studying abroad.
With her father’s career as an early influence, it comes as no surprise that Lu began coursework in geology; more surprising might be her decision to also pursue a minor in economics. “I always wanted to get into the oil and gas field,” she said. “I thought a focus in economics would help me understand national economies and oil and gas prices. Plus companies are more likely to hire people who have this multi-field background.”
Now in her third year at university, Lu is developing more field skills as a summer intern with RPI. Under the guidance of Guishard, Lu is using the city of Houston, Texas as a model to develop a methodology for studying flood events caused by storm surge and extreme precipitation events.“The soil found in Houston is very bad for filtration, so historically the city has had a lot of flood events due to precipitation and hurricanes,” said Lu. “This makes it a great case study for this type of research.”
By integrating data from a variety of sources—including water gauges and rainfall values—Lu plans to create a model to determine the threshold that will trigger flooding during storms. If successful, this model could be utilized by scientists and urban planners in other cities worldwide, using data specific to their respective regions.
Guishard said that Lu’s work, in conjunction with Johnston’s, fulfills his long-term interest in working on flood hazards and how extreme rainfall events vary in time, both locally and in other jurisdictions. “We are in a position for the first time to talk analytically about flood risk in Bermuda, because of this work,” he said.
For Lu, the experience with RPI has given her a newfound desire to pursue a career in research and academia. Her hands-on experience acquiring and working with multiple data sets to create predictive models will also be of benefit while she works on her senior thesis, which will focus on modeling slab fluid chemistry in subduction zones of oceanic plates. Scientists like Lu are interested in characterizing the processes that take place at subduction zones for a variety of reasons, including the potential for the development of tsunamis and earthquakes.
“Working with RPI has helped me understand the close linkage between science and society,” said Lu. “My research is something that might prove to be informative and beneficial to the wider disaster risk assessment society, which makes it rewarding.”
Thomas Perkins
Thomas Perkins, a fourth-year student at Cambridge University in England, is working with Guishard conducting research on the overlapping topics of risk prediction and analysis, meteorology, and the earth sciences. Specifically, Perkins is working on developing a model—unique to Bermuda—that relates the risk of hurricanes to damage on the island
“Other places like Florida and Mexico have risk models but Bermuda is unique in a number of ways: the small size of the island, the building codes that prevent more significant property damage, not to mention its location in the middle of the Atlantic,” he said.
However, unlike many “cat models” out there, Perkins is approaching this from a unusual perspective that takes advantage of his background in Earth science.
To overcome a relative paucity of data points—hurricane records extend back about 150 years–Perkins is turning to sediment cores taken from lagoons in Bermuda to extrapolate hurricane activity back to A.D. 950. Sediment cores are collected using hollow metal tubes, and the resulting samples are often saved in repositories where they can provide scientists with evidence of past climate change, fossils, oceanic and sedimentary processes—and hurricane activity.
“Using sediments, we can say more about how hurricane risk has changed for Bermuda,” said Perkins of this approach.
But Perkins is more focused on the future than the past. “All the time we are trying to consider how the risk of damage might change in the future,” he said. “Right now I’m looking at future climate predictions for the next 100 years, including predicted changes in sea surface temperature and carbon dioxide levels, both of which would affect how quickly and frequently hurricanes form in the Atlantic.”
Perkins’ research may serve as a framework for other people to build upon by incorporating other regional data sets, such as historical climatology, property values, and even other sediment core records. Guishard said the research has far exceeded initial project expectations, which started as developing a simple catastrophe model for Bermuda on hurricane risk.
“Despite the lack of data we’ve been able to push through and get some nice results, which isn’t always the case in science,” Perkins said.
Perkins said he also appreciates the time devoted to just one topic, noting that this semi-independent research is good preparation for his graduate research project, which will focus on understanding the geological carbon cycle. By analyzing sediment cores collected from the Mekong River in Asia, Perkins will trace the path of carbon in sediments created by the erosion of terrestrial rocks. Eventually these sediments will wash down the river and into the ocean, where they will settle to the bottom and become integrated back into Earth’s crust through subduction.
After college, Perkins said he plans to explore climate-based research. “I haven’t decided yet if my next step is a doctorate or if I’d like to be involved with something like RPI,” he said. “There’s something so reassuring about being able to do science that is directly applicable to the real world.”
For more information on these and other internships at BIOS visit: