Disaster Planning Flashcards
Mayell, 2002
There are 457 volcanoes with cities that house 1 million people or more within 100km of them
What have we entered into an age of?
Mega-disasters in general, with many more people killed and affected (e.g. the 2004 South Asia Tsunami, which killed between 230-280 thousand people, and caused $19.9bn worth of damage)
Why are deaths from disasters increasing?
Population increase, increased urbanisation
Oppenheimer, 2011
Nearly 500 volcanic events in the 20th century impacted people, with up to 6 million people evacuated or left homeless
Sparks, 2003
It is not always possible to predict volcanic eruptions because they can become critical systems that move from one state to another with only very minor internal or external triggers
Complex and chaotic systems
Two examples where volcanic systems switched very quickly:
1) Lascar in Chile - benign lava effusions suddenly changed to hazardous explosive activity after many years of eruption (Mathews, 1997)
2) Mt St Helens - magma had been slowly intruding however eruption was actually triggered by an external magnitude 5 earthquake, and thus eruption timings could have been very different without this trigger (Lipman, 1981)
Outline what happened at Nevado del Ruiz (except why it ended up being a disaster)
Voight, 1996
1985, 23,000 killed due to Lahars (20,000 in just Armero)
This disaster was “exclusively caused by cumulative human error”
Town of Armero
Volcano provided ample warming (seismic activity, fumaroles activity, phreatic eruption) - it was being continuously monitored
An accurate hazard zonation map was available, explicitly demarcating Armero as squarely in the zone of highest hazard (the town was built on old Lahars)
Safety on higher ground was within walking distance
Why did Nevado del Ruiz end up being a disaster? (3)
BUT
1) Many simply did not understand the advance warning and the risk to them - the hazard map had been poorly distributed
2) An evacuation was ordered the day of the eruption, but then retracted when ash stopped being emitted - people were told to stay inside, and that the intermittent ash was nothing to worry about - one survivor reports going to the fire station and being told that the ash was “nothing”
3) When a final evacuation was called, very close to the time, it was difficult to distribute the message due to electrical problems caused by a storm - the storms heavy rain and constant thunder overpowered the noise of the volcano, and with no systematic warning efforts, the residents of Armero were completely unaware of the continuing activity
Fiske, 1984
The Tale of Two Volcanoes
Soufriere Hills volcano on Guadeloupe - 1975
People were angered because scientists made the “wrong” call - called for an evacuation of 74,000 people but an eruption never occurred
This caused significant and long-term impact not he economic prosperity of the island
The situation was caused by a number of things, including the fact that people were very anxious in the shadow of Mt Pelee, and over media involvement pitting two teams of scientists against each other
In contrast to St Vincents eruption 1979 - one team of scientists, long-term monitoring allowed for understanding of patterns - information given to media only after scientific agreement, and only presented facts of what had happened, no speculation - evacuated 22,000 citizens before eruption occurred
Outline examples of why forecasting is not always clear:
1) Mt St Helens (1980) - the volcano erupted during an apparent lull in activity - it was later shown that this can be a sign of pressurisation prior to a volcanic explosion
2) Mt Pinatubo - even though the eruption was predicted, the main impact wasn’t - rainfall added to ash loading and led to roof collapse that killed >300 people - the combination of the volcano with Typhoon Tanya contributed to this loading, exemplifying how it is important to think about cross-over between hazards (Newhall and Punongbayan, 1996)
Why is forecasting often expressed in probabilities?
In order to account for uncertainty
Volcanic eruptions involve the interactions between highly non-linear and complex kinetic and dynamic processes - this leads to a rich range of possible behaviour
This complexity makes forecasting inherently difficult, and thus probabilistic modelling can help to constrain this
Outline the use of geologic reconnaissance at Pinatubo
1991
Newhall et al., 1996; Punongbayan et al., 1996
Geologic reconnaissance in the weeks leading up to the eruption revealed expensive pyroclastic flow deposits
With radiocarbon age determinations, scientists were quickly able to prepare an accurate hazard zonation map recognising the volcano’s propensity for highly explosive activity - this led to the evacuation in a 40km radius of 85,000 people
This effort, in combination with extensive monitoring of the volcano and a good evacuation plan, was estimated to have saved 5000-20,000 lives, and prevent at least US$2540 million of property damage
Auippa et al., 2007
Used a quantitative saturation model to relate observed pre-eruptive increases in the ratio of CO2/SO2 to the refilling of Mt Etna’s shallow conduits with CO2 rich deep-reservoir magmas, leading to pressurisation and the triggering of an eruption
Case study Soufriere Hills, Montserrat
Began 1995 - still ongoing (intermittent)
Evacuated in 1995
19 people killed in a surge on the 25th of June, 1997 - they returned to the exclusion zone against official advice, due to apparent in activity
Posed a big problem because evacuation needed to be at least 25km away, but the island is only 9km long - fortunately the mountains provided a barrier between an evacuated population and the volcano - 2/3 are an exclusion zone - this highlights the importance of not just maths, but expert judgement
Expert elicitation method was developed here (Aspinall and Cooke, 1998) - which calibrates expert judgement in times of uncertainty
Also used event trees
Martí et al., 2008
Use of event trees at Tenerife to model possible hazards in life of an eruption
