Skip to content

Mapped geologic features used to forecast long-term likelihood of future eruption locations: applications during an ongoing series of frequent, differing-impact eruptions on the Reykjanes Peninsula, Iceland

Melissa Anne Pfeffer 1, Ásta Rut Hjartardóttir1, Sara Barsotti1, Bergrún Arna Óladóttir1, Bogi Brynjar Björnsson1, Gro B. M. Pedersen1

Abstract

Since 2020, magma intrusions and eruptions have occurred on the Reykjanes Peninsula. The current activity varies from eruptions constrained in remote valleys to eruptions affecting critical infrastructure, such as the town of Grindavík, power lines, water pipes, a power plant and a popular tourist attraction.  For long-term hazard assessment, the Matlab code MatHaz was applied to calculate the spatial probability of future eruptions on the peninsula. The input data are mapped geologic features: past eruption sites, fractures, geothermal areas, and the plate boundary. Input data changes over time: for example, new eruption sites are appended to the previous ones. Lava flows from younger vents can erase surficial record of an older vent, therefore the input of short-lived vent locations must be coarsened to be consistent with the geologic record of older eruptions. Scientific choices such as how to update critical input data generate additional, ongoing sources of uncertainty. Discussing long-term hazard results during an on-going event is particularly sensitive, both in terms of conveying the uncertainty of the results as well as ensuring that assessments made for a long-term purpose are not misapplied. During events that cause societal upheaval, non-ambiguous answers about an ongoing crisis are often desired. It is important to reemphasize that unlikely events can happen. It is a challenge to optimize communication of the breadth of results made possible through this approach of quantifying likelihood based on mapped geologic features. Making decisions about when it may be beneficial/harmful to emphasize worst-case scenarios is an on-going conversation.