Assessing risk from ballistic impacts through aerial hazard mapping, numeric modelling, and laboratory experiments to enhance risk management and risk communication

Summary

Rocks and lava thrown from volcanoes (ballistics) produce a severe hazard in areas close to the volcano summit. As volcano tourism continues to grow worldwide, ballistics have become one of the most common cause of incidents involving fatalities at volcanoes with at least 76 recorded deaths at six volcanoes since 1993, including Mt Yasur in Vanuatu. In New Zealand in 2006 one death was reported on Raoul Island, and since then a near miss on Ruapehu, a near miss at Tongariro which significantly affected infrastructure used by hikers on the high‐use Tongariro Alpine Crossing, and seven near misses from Whakaari (White Island). Here, we pioneer the quantification of ballistic hazard in New Zealand and Vanuatu.

Our adaptive mapping methodology allowed different aerial and ground based techniques to be employed in different scenarios. Mapping from the three volcanoes shows that the ballistic hazard varies considerably between eruptions styles. The distribution of ballistic blocks from the small 2016 phreatic eruption from Whakaari shows a small ballistic hazard zone but with extreme ballistic hazard almost anywhere within the zone (Kilgour et al., in review). At Red Crater on Mt Tongariro, we identified an age constrained eruption deposit (<500yrs) that shows that the hazard from ballistics from phreatic eruptions drops rapidly with distance ‐ from extreme hazard close to the crater to low hazard within a hundred meters (Gates et al., 2017). In contrast, the average individual magmatic eruption pulse at Yasur has a significantly lower hazard with most ballistics landing back into the crater (Fitzgerald et al., in prep). A decrease in hazard intensity with distance was also seen here, though is highly spatially variable ranging over several orders of magnitude with both proximity and azimuth to the vent. However, this assessment covers only a small time window (2 months) and further study is required to assess how ballistic hazard changes over longer periods of time.

The modified Ballista 3D numerical ballistic trajectory model was applied in two different ways at Whakaari and Ngauruhoe to assess ballistic hazard. Multiple model runs were compared to the mapped ballistic distribution and geophysical data at Whakaari to understand in greater detail the ballistic distribution and eruption dynamics. Ballista revealed that multiple low angle ballistic jets were likely coupled with surges which may change the traditional ballistic trajectory and subsequently the expected area of hazard (Kilgour et al., in review). On Ngauruhoe we coupled a single Ballista model with a RAMMs rockfall runout model to show how bouncing and rolling downhill after impact compounds the ballistic hazard, in this case increasing the magnitude of the hazard by over 400% (Kennedy et al., in prep). In addition, we are pioneering a new technique to systematically fit ballistic distributions to model runs (Tsunematsu et al., 2017, 2018).

Vulnerability of different roofing materials to ballistic impact was quantified using the experimental pneumatic ballistic cannon (Williams et al., 2017). Collaboration with the College of Engineering at the University of Canterbury on ballistic impacts to concrete roofs showed fibre reinforced plastics could eliminate the shrapnel hazard and that a 5 cm layer of ash or lapilli on the concrete could significantly dissipate the impact energy and reduce damage (Williams et al. 2017b; 2018).

Through a publication on ballistic hazard communication (Fitzgerald et al. 2017) and a ballistic hazard workshop we aimed to keep end users informing and directing the science discovery pathways throughout the project. We continue to develop better ways to inform end users. Discussion of this report with Harry Keys has already sparked improved communication methodologies with DOC.