SAMSI Special Guest Lecture – Bruce Pitman: Where Are You Gonna Go When the Volcano Blow?

Pitman Headshot

Bruce Pitman is a Professor in the Department of Materials Design and Innovation School of Engineering and Applied Science at the University of Buffalo.

One of the major hazards of volcanic activity is inundation by debris flows, block and ash flows, and hot pyroclastic flows. These can run out many kilometers, travelling at speeds of 50 m/sec or more. A challenge for volcanology is to predict the chances of a flow inundating a specific location – that is, what areas are at risk of suffering a hazardous flow event.

A group of scientists from geology, engineering, mathematics, and statistics has been studying this problem for more than 15 years. Several of the scientists came together during the 2006-2007 SAMSI program on “Computer Models”, and have been collaborating ever since.

After lying dormant for more than 300 years, the Soufriere Hills Volcano on the island of Montserrat began an eruptive phase in 1996. Hundreds of mass flows have occurred during the last 17 years, ranging in size from O(104) m3 of material

to 200 × 106 m3 of material. Several of the largest of these flows have caused tremendous damage to population centers on the island, to the extent that today more than half the island has been evacuated.

Using mathematics to derive model equations describing the flowing mass, and the sophisticated computer solver TITAN2D, scientists can simulate mass flows at Soufriere Hills, assuming certain input parameters. To calibrate the effort, each computer simulation currently takes about 20-60 minutes on a 16-processor parallel cluster.


This figure shows contour from flow thickness computations, showing safe and hazardous zones along the Belham valley on Montserrat.

By combining flow data and expert opinion, one can develop a probabilistic model of the severity and frequency of flow events. The naïve approximation to assessing the hazard probability, namely sampling this distribution to generate inputs and running those inputs through the simulator to estimate the percentage of these

runs that ‘hit’ that location, is not feasible because of the rarity of catastrophic events and the expense in running the simulator.  Furthermore, this approach ties the expensive simulator runs to a specific probabilistic model which may prove antiquated as new data and information become available.

Our group introduced a new twist: a statistical emulator — a computationally cheap response surface approximating the output of simulations — is constructed, based on carefully chosen computer model runs. The speed of the emulator then allows us to ‘solve the inverse problem’: determine regions of inputs values (characteristics of the flow) which result in a catastrophic event. Then the probability that a catastrophic event will occur at a particular location can be quickly calculated under any probability distribution of inputs. With a careful arrangement of computer simulations and emulator constructions, we can calculate the probability of a catastrophe at many locations simultaneously, producing a hazard map like that shown in the figure. This map shows the hazard regions near the Belham valley, under four different, potential eruption scenarios. Civil protection officials have set “trigger points”, sites at which flows of a certain thickness initiate evacuation of particular neighborhoods. Our hazard maps can provide additional information to these officials, and help refine the trigger and evacuation workflow.


Pitman taking questions from the group at the conclusion of his lecture on how statistics is being used to address potential hazards from volcanic activity.

By the way, the Volcano song by Jimmy Buffett was written in the late 1970s, about the Soufriere Hills Volcano. The volcano had been dormant since the 17th century, and erupted in 1996. Volcano was recorded at AIR Studios on Montserrat, a recording facility designed by Sir George Martin and opened in 1979. AIR Studios were destroyed by Hurricane Hugo in 1989.


SAMSI Postdoctoral Research on Ocean leads to Academic Career in Statistics


Mikael Kuusela, former Postdoctoral Researcher at SAMSI. Currently serving on a tenure-track as an assistant professor in the Department of Statistics and Data Science at Carnegie Mellon University.

Life has been good to Mikael Kuusela. A native of Helsinki, Finland, his has been a long journey to becoming a scientific researcher.

His curiosity and passion for new things, new people and new environments, led him to pursue Bachelor’s and Master’s Degrees in Engineering Physics and Mathematics from Aalto University in Finland, and then a Ph.D in statistics from École Polytechnic Fédérale de Lausanne (EPFL) in Switzerland.

Kuusela is one of a long line of postdoctoral researchers who have attended SAMSI. Like his experience here, others have found a direction to their dream of one day working in a field of study involving applied mathematics, statistics or many other disciplines of data science.

“Mikael’s performance was superb during his time at SAMSI, and he has contributed to the active and lively research atmosphere at SAMSI by his participation at the seminars, workshops, working groups and other research and outreach activities,” said Elvan Ceyhan, Deputy Director of SAMSI. “He was not just focused in his research but also striving for SAMSI to be a center of activity and interaction.”

Kuusela attended SAMSI through a joint, climate-based program through the Statistical Methods for Atmospheric and Ocean Sciences (STATMOS) research network and SAMSI. After spending one year at the University of Chicago as part of his postdoctoral experience, he arrived at SAMSI in 2017. Kuusela continued research he began at the University of Chicago, which focused on improving statistical methods for analyzing Argo float data.

“SAMSI enables postdocs to form a nationwide network of professional connections and to gain experience on cutting-edge interdisciplinary research.”

“The main reason I chose this position was that it provided an opportunity to work on an extremely interesting application: the measurement of global ocean temperatures and salinities using data from Argo floats,” said Kuusela.

Argo Float data is used to measure temperature and salinity in the upper 2000 meters of the global ocean. By finding better ways to analyze these data, Kuusela and his colleagues hoped to show the prevalent evidence of environmental change in our world. This research required statisticians and oceanographers to work together in order to identify the best statistical techniques for studying these environmental changes in the ocean.

In support of the SAMSI Program on Mathematical and Statistical Methods for Climate and the Earth System (CLIM), Kuusela helped set up and coordinate the Statistical Oceanography Working Group, comprised of some of the world’s leading oceanographers and statisticians. One of their aims was to study and estimate the amount of heat in the global ocean and develop time series models to graphically depict where these trends occur and thus, where efforts need to be focused in order to mitigate environmental change.


Mikael Kuusela gives a talk during the Postdoctoral Fellow 2018 Spring Session. Kuusela was a second-year postdoc researcher at SAMSI, using his time to research better ways to analyze data involving ocean temperature variables.

“We are especially interested in the rate at which the global ocean is warming up and in the month-to-month variations in the heat content. Accurate estimation of these quantities is incredibly important since almost all of the heat that is trapped in the Earth’s Climate System will eventually end up in the ocean where it shows up as increased ocean heat content,” said Kuusela.

Kuusela confesses each day at SAMSI was new and exciting and often led him to collaborating with colleagues all over the world, sometimes in a single day!

“Some days are full of meetings. Sometimes I’ve had morning meetings with researchers in Europe, lunch meetings with people on the East Coast and afternoon meetings with collaborators on the West Coast. Since SAMSI work is almost by definition collaborative, it tends to involve a lot of travel. Sometimes I feel like I’ve spent as much time on the road as I did at SAMSI,” he said.

In all, Kuusela pointed to the numerous contacts he developed over his time working at SAMSI as being some of the most valuable results of his success. “Thanks to all of these interactions, you not only get to work on interesting interdisciplinary research projects, but also end up with a valuable network of professional connections. The possibility of networking and collaborating with researchers from various institutes across the country is indeed one of the great benefits of being a SAMSI postdoc,” he said.

Kuusela noted that he was comforted by the fact that he ran into numerous former SAMSI postdocs who are now faculty members at some of the nation’s leading institutions during his interview process for academic positions.

“A SAMSI postdoc can expect plenty of opportunities for engaging and collaborating with leading researchers in their field and beyond. In many ways, SAMSI is ‘the CERN of Statistics,’ a central hub of activity that enables researchers to interact in ways that would not otherwise be possible,” said Kuusela. “SAMSI enables postdocs to form a nationwide network of professional connections and to gain experience on cutting-edge interdisciplinary research.”

He also talked about what postdocs who attend SAMSI in the future gain from the postdoctoral research experience.

“SAMSI postdoctoral experience is clearly highly valued in today’s academic job market and it is easy to see why: SAMSI postdocs gain experience in working on complex applied and computational problems in statistics and mathematics in a collaborative and interdisciplinary environment.”


Mikael Kuusela leads an instruction tutorial on R Software for a SAMSI-sponsored undergraduate workshop. During Kuusela’s second year as a postdoc, he routinely gave talks and participated in several workshops, where he had the opportunity to collaborate and mentor undergraduate and graduate students and also gain valuable contacts for future collaborations with fellow researchers in his field.

Kuusela has now joined the Department of Statistics and Data Science at Carnegie Mellon University in Pittsburgh, PA, as a tenure-track assistant professor. While teaching, he will also continue to focus on his research on statistical data analysis in the physical sciences.

SAMSI is one of eight mathematics and science institutes, funded by the National Science Foundation, whose aim is to inspire, enhance and prepare the next generation of applied mathematicians, statisticians, and computer and data scientists for the future.

Mikael Kuusela is now enjoying the fruits of his labor and getting to do what he is most passionate about – helping to develop better data analysis methods to solve scientific problems that impact our planet and society.

To see postdoctoral research opportunities at SAMSI, visit: