Evaluating and ranking Southeast Asia's exposure to explosive volcanic hazards

Abstract

Abstract. Regional volcanic threat assessments provide a large-scale comparable vision of the threat posed by multiple volcanoes. They are useful for prioritising risk-mitigation actions and are required by local through international agencies, industries and governments to prioritise where further study and support could be focussed. Most regional volcanic threat studies have oversimplified volcanic hazards and their associated impacts by relying on concentric radii as proxies for hazard footprints and by focussing only on population exposure. We have developed and applied a new approach that quantifies and ranks exposure to multiple volcanic hazards for 40 high-threat volcanoes in Southeast Asia. For each of our 40 volcanoes, hazard spatial extent, and intensity where appropriate, was probabilistically modelled for four volcanic hazards across three eruption scenarios, giving 697 080 individual hazard footprints plus 15 240 probabilistic hazard outputs. These outputs were overlain with open-access datasets across five exposure categories using an open-source Python geographic information system (GIS) framework developed for this study (https://github.com/vharg/VolcGIS, last access: 5 April 2022). All study outputs – more than 6500 GeoTIFF files and 70 independent estimates of exposure to volcanic hazards across 40 volcanoes – are provided in the “Data availability” section in user-friendly format. Calculated exposure values were used to rank each of the 40 volcanoes in terms of the threat they pose to surrounding communities. Results highlight that the island of Java in Indonesia has the highest median exposure to volcanic hazards, with Merapi consistently ranking as the highest-threat volcano. Hazard seasonality, as a result of varying wind conditions affecting tephra dispersal, leads to increased exposure values during the peak rainy season (January, February) in Java but the dry season (January through April) in the Philippines. A key aim of our study was to highlight volcanoes that may have been overlooked perhaps because they have not been frequently or recently active but that have the potential to affect large numbers of people and assets. It is not intended to replace official hazard and risk information provided by the individual country or volcano organisations. Rather, this study and the tools developed provide a road map for future multi-source regional volcanic exposure assessments with the possibility to extend the assessment to other geographic regions and/or towards impact and loss.

Insights into the vulnerability of vegetation to tephra fallouts from interpretable machine learning and big Earth observation data

“…For example, the 2020 VEI 4 (volcanic explosivity index, Newhall andSelf, 1982) eruption of Taal Volcano (Philippines) affected ∼ 260 000 people and caused an estimated USD 63 million impact on agriculture (ReliefWeb, 2020), whereas the 2018 eruption of Fuego (Guatemala), also a VEI 4, indirectly affected ∼ 1.7 million people and caused USD ∼ 58 million impact on agriculture (The World Bank, 2018). By comparison, a recent study by Jenkins et al (2022) estimates that on the island of Java in Indonesia only, a VEI 4 eruption has a 50 % probability of directly affecting ≥ 5 million people and ∼ 700 km 2 of crops, which increases to ∼ 29 million people and 12 000 km 2 of crops for an eruption of VEI 5.…”

Grain size modulates volcanic ash retention on crop foliage and potential yield loss

“…Assuming a bulk density of 1 g cm −3 for the ash deposit (Eychenne et al, 2012), this corresponds to a relatively thin deposit of ∼ 0.6 mm (i.e. considering a bulk deposit density of 1 g cm −3 ; Eychenne et al, 2012), best representing accumulations encountered at distal sites (and over wide areas) affected by ash fallout from explosive eruptions (Fierstein and Nathenson, 1992;Jenkins et al, 2022). Pre-tests carried out with higher ash loads (≥ 1000 g m −2 ) already led to lodging of some tomato and chilli pepper plant specimens, a phenomenon that needed to be avoided in order to maximise the experiment's reproducibility.…”

“…These limitations are hindering the development of accurate process-based risk assessment models that can inform targeted strategies to build the resilience of agriculture-based communities in the case of an explosive eruption, for example in relation to aid allocation, land-use planning and insuring. Jenkins et al (2022) estimated that an explosive eruption of 4 on the Volcanic Explosivity Index (VEI, Newhall and Self, 1982) on the island of Java, Indonesia, has on average a 50 % probability of affecting ∼ 700 km 2 of crops with 5 kg m −2 of ash. The surface area potentially affected by ash fallout is ∼ 17 times larger for an eruption of VEI 5.…”

Identifying analogues for Melimoyu, a long-dormant and data-limited volcano in Chile, through hierarchical clustering

“…Analogue volcanoes have been typically defined based on location, tectonic setting, morphology, magma type, eruption style, or a combination of these factors for i) assessing local and regional volcanic hazards (e.g., Jenkins et al, 2012b;Lindsay and Robertson, 2018;Mastin et al, 2009;Newhall, 1982;Newhall and Pallister, 2015;Sandri et al, 2014Sandri et al, , 2012Tennant et al, 2021;Tierz et al, 2020); ii) estimating frequency-Magnitude (f-M) relationship (e.g., Hayes et al, 2022;Jenkins et al, 2012aJenkins et al, , 2022Rodado et al, 2011;Runge et al, 2014;Sheldrake and Caricchi, 2017;Solow, 2001;Whelley et al, 2015) s; iii) conducting probabilistic eruption forecasts (e.g., Bebbington, 2014;Bebbington and Jenkins, 2022;Marzocchi et al, 2004;Sheldrake, 2014), and iv) identifying unrest patterns (e.g., Acocella et al, 2015;Newhall et al, 2017).…”