Case Study 1 – Island of Hawaii’s Kīlauea Volcano Eruption
The Kīlauea Volcano’s Eastern Rift Zone (Puʻu ʻŌʻō) has been actively erupting since 1983, making it the longest rift-zone eruption in 500 years. From May to June, the large eruption by the Big Island’s second largest volcano destroyed 716 structures – more than three times as many as it had in the first 30 plus years of eruption (215 in total from 1983 through 2017). The Puna Geothermal Venture power plant was shut down indefinitely from the eruption, cutting off more than 20% of the island’s electrical power. During May 3 to August 9, more than 13 square miles were covered by destructive lava and 876 acres of land were created. On 4 May, a Mw6.9 earthquake produced large aftershocks and ash clouds in subsequent months, but no significant reports of damage from the intense ground shaking.
Figure 1: U.S. Geological Survey – Hawaiian Volcano Survey map of recent Kīlauea Volcano lava flows (in shaded pink area) as of August 9, 2018. (Source: U.S. Geological Survey Hawaiian Volcano Observatory)
Insurance implications resulting from the recent volcanic activity were multi-layered in 2018. Typical U.S. homeowners’ policies cover ‘volcanic eruption’ as a named peril for Coverage C, but exclude losses from earthquakes caused by volcanic activity. The houses destroyed by the eruptive fissures were likely attributed to ‘volcanic eruption,’ and recent exclusions for ‘lava’ perils were largely disregarded for policyholders following filed lawsuits. At the end of 2018, there was an agreement to pay submitted claims by policyholders with ‘completely destroyed homes’ despite specific policy language excluding all ‘direct or indirect’ damage from lava inundation.
Figure 2: Active lava flows around the Puna Geothermal Venture power plant in early July 2018 when there were still active lava flows from the Kīlauea Volcano. (Source: Kish Kim / Sipa USA via © 2018 The Associated Press)
The U.S. Geological Survey estimates that the U.S. has 10% of the world’s potentially active volcanoes, of which Hawaii’s Kīlauea Volcano was considered the highest risk in 2018. Despite efforts to apply lava exclusions to insurance policies in high-risk areas (e.g., Hawaii, Washington, Alaska), damage from fire due to lava is likely to be paid in future events. Residents in such areas have typically found coverage through non-admitted insurers, but following the recent eruption activity, this coverage may no longer be available. From a global perspective, it is estimated that more than 800 million people live within 100 kilometers of an active volcano in 86 different countries. More recent eruptions of ash clouds, especially those from the 2010 Eyjafjallajökull volcano in Iceland, demonstrate that close proximity to active volcanoes does not necessarily indicate event impact.
Case Study 2 – Hokkaido, Japan’s Enhanced Earthquake Hazards
A strong Mw6.6 earthquake struck the northern island of Japan, Hokkaido, on September 6 at 03:12 local time. In the days following the earthquake, more than 41 fatalities were attributed to hundreds of landslides triggered by the event. The large number of landslides in the region was exacerbated by the remnants of Typhoon Jebi, which damaged large portions of Honshu only days before and moved past Hokkaido just before the earthquake occurred. Excessive rainfall in the region over the previous summer had also contributed to saturated ground conditions in the mountainous region.
Figure 3: Numerous landslides near the town of Atsuma, and reportedly hundreds more, were triggered by the September 6, 2018 earthquake in Hokkaido, Japan. (Source: Kyodo via © 2018 The Associated Press)
At a basic level, slope failures occur when the slope angle of the ground is greater than the friction angle (i.e., soil strength). As the soil is more saturated, it loses strength and therefore can become unstable, especially when coupled with intense ground shaking. The widespread rainfall in Japan and a recently passing typhoon increased the susceptibility to earthquake-triggered landslides in Hokkaido on 6 September. This particular kind of ‘land movement’ peril is covered by standard earthquake insurance policies. More saturated soil conditions are not directly considered in catastrophe modeling, but could present an alternative/conservative view when analyzing high risk exposures.
Case Study 3 – Sulawesi, Indonesia Earthquake & Submarine Landslide
On September 28, a Mw7.5 earthquake struck Donggala Regency of Central Sulawesi in Indonesia. Based on the location and depth of the event (relatively shallow at 10 kilometers), the earthquake occurred on the strike-slip Palo Karu Fault. A tsunami with a reported maximum height of four to seven meters (13 to 23 feet) followed the earthquake, and tragically killed more than 2,100 people. Tsunamigenic earthquakes in Indonesia occur from large magnitude (M8.0+) subduction interface events, like that in 2004. The tsunami wave height was also amplified by the long and narrow shape of the bay leading into Palu City.
Figure 4: Damage from the September 28 earthquake and tsunami in Central Sulawesi, Indonesia. (Source: © 2018 The Associated Press / Tatan Syuflana, File)
This tsunami was unusual in that it was triggered by a different mechanism; Indonesia’s Meteorology, Climatology, and Geophysical Agency, BMKG, confirmed that the tsunami was caused by a submarine or ‘underwater’ landslide. This is different from what is captured in traditional catastrophe modeling for tsunami – i.e. earthquake-induced sea-floor displacement. Underwater landslides have the potential to produce damaging tsunami despite this kind of hazard not explicitly considered in the traditional catastrophe models. Also, submarine landslide-triggered tsunami is not unique to Indonesia or Southeast Asia. Tsunami catalogs for various regions like the East Pacific, Artic and East Atlantic, and the Caribbean record previous events’ triggers, including submarine landslides. Studies for nuclear reactor safety in the U.S. have mapped the East Atlantic seafloor and found multiple examples of large submarine landslides.
Insurance applicability will vary across the globe as the tsunami peril is not always considered part of earthquake coverage, like in the U.S. where it is covered by flood insurance. Submarine landslides can be triggered by a variety of factors like earthquakes, glacial cycles, human/industrial activity, and volcanic flank collapse. Locations of some historical submarine landslides in the Atlantic are pictured in Figure 5. Further complicating the estimation of these hazards is the potential influence of climate change, especially closer to the Arctic regions. Human activity attribution can introduce third-party liability issues in managing submarine landslide and tsunami risk, similar to recent litigation around induced seismicity in the U.S.
Figure 5: Historical submarine landslides in the Atlantic Ocean near the Eastern U.S. (Source: USGS)