Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/springer-journals/a-method-for-evaluating-population-and-infrastructure-exposed-to-Y79GH0qNWC
- Publisher
- Springer Journals
- Copyright
- Copyright © The Author(s) 2023
- eISSN
- 2197-8670
- DOI
- 10.1186/s40677-023-00235-8
- Publisher site
- See Article on Publisher Site
Abstract
Background Coastal communities are highly exposed to ocean- and -related hazards but often lack an accurate population and infrastructure database. On January 15, 2022 and for many days thereafter, the Kingdom of Tonga was cut off from the rest of the world by a destructive tsunami associated with the Hunga Tonga Hunga Ha’apai volcanic eruption. This situation was made worse by COVID-19-related lockdowns and no precise idea of the magnitude and pattern of destruction incurred, confirming Tonga’s position as second out of 172 countries ranked by the World Risk Index 2018. The occurrence of such events in remote island communities highlights the need for (1) precisely know- ing the distribution of buildings, and (2) evaluating what proportion of those would be vulnerable to a tsunami. Methods and Results A GIS-based dasymetric mapping method, previously tested in New Caledonia for assessing and calibrating population distribution at high resolution, is improved and implemented in less than a day to jointly map population clusters and critical elevation contours based on runup scenarios, and is tested against destruction patterns independently recorded in Tonga after the two recent tsunamis of 2009 and 2022. Results show that ~ 62% of the population of Tonga lives in well-defined clusters between sea level and the 15 m elevation contour. The patterns of vulnerability thus obtained for each island of the archipelago allow exposure and potential for cumulative damage to be ranked as a function of tsunami magnitude and source area. Conclusions By relying on low-cost tools and incomplete datasets for rapid implementation in the context of natural disasters, this approach works for all types of natural hazards, is easily transferable to other insular settings, can assist in guiding emergency rescue targets, and can help to elaborate future land-use planning priorities for disaster risk reduction purposes. Keywords Kingdom of Tonga, Tsunami, Impact assessment, Population inventory, GIS, Dasymetric mapping Introduction The 15 January 2022 tsunami event and its impact On January 15, 2022, the violent eruption of an underwa- ter volcano in the Southwestern Pacific Ocean (175.385° W, 20.565° S) entailed disastrous consequences on neigh- *Correspondence: boring islands of the Tonga archipelago (Fig. 1). In addi- Bruce Enki Oscar Thomas tion to the volcanic products (mostly ash) and to the bruce.thomas@gis.uni-stuttgart.de Institute of Geodesy (GIS), University of Stuttgart, Stuttgart, Germany massive airborne shockwave generated by the eruption Earth Structure and Processes, GNS Science, Lower Hutt, New Zealand and recorded around the world (Gusman et al. 2022), the Université Lumière Lyon 2, CNRS UMR 5600, Bron, France 4 eruption was followed by a tsunami that was promptly Data Science and Geohazards Monitoring, GNS Science, Lower Hutt, New Zealand recorded on nearby gauges and DART sensors (Gusman and Roger 2022). A Pacific-wide tsunami threat bulletin © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. Thomas et al. Geoenvironmental Disasters (2023) 10:4 Page 2 of 18 Fig. 1 Geological hazards in Tonga. Left panel (framed by red rectangle on globe): seismotectonics around the Tonga–Kermadec subduction zone. Right panel (framed by black rectangle in left panel): the three main island groups of Tonga ( Vava’u, Ha’apai and Tongatapu). Circles: epicentres of magnitude Mw > 7.5 earthquakes (circle sizes proportional to magnitude; USGS data from January 1970 to February 2022) with color as a function of focal depth: shallow (red), intermediate (orange), deep (yellow). Black lines: main tectonic features (subduction trench, spreading ridges). Black triangles: active Holocene volcanoes (https:// volca no. si. edu, accessed on February 13, 2022). Red triangle: Hunga Tonga-Hunga Ha’apai (HTHH) volcano. Red and black contours: tsunami travel times from the volcano, in minutes (calculated using Mirone Software; Luis 2007). Bathymetric data from GEBCO (2014). Map generated using Generic Mapping Tools ( Wessel et al. 2019) was issued, and coastal populations were evacuated on had been destroyed by the tsunami (OCHA 2022a). The New Zealand’s north island (Hunt and Piper 2022; NZ evidence was published 20 days later by UNOSAT (2022). Herald 2022), in New Caledonia (LNC 2022), and as far Based on the topographic elevation of those buildings, it as Japan, where 229,000 residents were moved to higher appears that the tsunami reached + 15 m on Tongatapu, ground (Imamura et al. 2022; Japan Times 2022). ‘Eua and Ha’apai islands (Government of Tonga 2022). The maximum amplitude of the tsunami wave was gen - This satellite-based analysis has so far revealed that more erally less than 1 m, but for some islands located only than 80% of Tonga was affected by the disaster, whether tens of kilometers away from the volcano, the waves were by thick ash fall or the tsunami (OCHA 2022a). much higher and destructive, and suspected to be asso- ciated with one or several submarine landslide events Context (Lynett et al. 2022). Several deep-sea telecommunications The Kingdom of Tonga is a southwest Pacific country cables were severed, therefore prohibiting easy commu- composed of 172 islands of surface area above 0.005 k m , nications (GFDRR 2022; Terry et al. 2022). Moreover, the 45 of them inhabited by a total population of 100,651 general pandemic situation in a COVID-19-free island (2016 census, TSD 2019). About 70% of this population is kingdom (BBC 2022; Vainikolo 2021) impeded opportu- concentrated on the main island, Tongatapu (~ 260 km ), nities for investigating the impact of the tsunami directly which hosts the capital city Nuku’alofa. The Tonga Sta - on the ground in the days following the event. Thus, tistics Department (TSD) indicates that 27% of the coun- the number of tsunami-impacted people and buildings try’s population is poor, living monthly on less than TOP$ remained unknown for many weeks and generated uncer- 970 (Tongan Pa’anga, equivalent to US$ 428 on February tainty around how to prioritize immediate international 13, 2022). According to World Bank assessments, ~ 3% of help to Tonga. The United Nations Institute for Training Tonga’s population lives in extreme poverty. This refers and Research (UNITAR) released preliminary informa- to people with a monthly income of less than TOP$ 3.10 tion only two days after the event, in a satellite-derived (= US$ 1.37 on February 13, 2022; Fifita et al. 2018). damage assessment report showing that many buildings T homas et al. Geoenvironmental Disasters (2023) 10:4 Page 3 of 18 These islands are volcanic edifices associated with relatively accurate assessments of the numbers of peo- the subduction of the Pacific Plate beneath the Austral - ple and buildings affected by a tsunami and, by exten - ian Plate. Some of the volcanoes are extinct and sub- sion, by other types of natural hazards within a defined sided/eroded (now mostly coral islands), but others are region. While showcasing its potential for two tsunamis still active (Bryan et al. 1972). The island population is recorded in 2009 and 2022, the results and discussion thus exposed to strong geological hazards like subduc- provide insights into the human and infrastructural vul- tion megathrust earthquakes exceeding magnitudes of nerabilities in Tonga. M 8.0, explosive volcanism, subaerial and submarine landslides originating on volcanic slopes, and tsuna- Data and methodology: implementation mis. Recent reminders of the destructive capacity of of a low‑cost toolbox for disaster vulnerability these processes include the September 9, 1946 eruption, mapping which led to a definitive evacuation of the population This study follows a dasymetric approach for mapping from Niuafo’ou (Rogers 1981); the June 23, 1977 M 7.2 settlements and estimating population numbers that are earthquake, which caused major damage on Tongatapu potentially vulnerable to natural hazards. Unlike the spa- and’Eua (Campbell et al. 1977); the May 3, 2006 M 8.0 tially-averaged approach commonly used for generating earthquake, which caused damages and a small tsunami choropleth map units, population data is redistributed in southern Tonga (Cummins et al. 2007; Heeszel et al. into dasymetric map units based on a combination of 2006; Tang et al. 2008); and the September 29, 2009 tsu- areal weighting and the estimated population densities. nami triggered by a M 8.1 and 8.0 earthquake doublet The spatial heterogeneity of variables such as building (Clark et al. 2011; Fritz et al. 2011; Lay et al. 2010). Traces and population density is thus much more accurate than of past far-field events have also been highlighted by sev - in the case of choropleth maps, in which information is eral studies (e.g. Okal et al. 2004), including the contro- uniformly averaged across map units. The motivation to versial occurrence of tsunami-related boulder deposits use the dasymetric mapping methodology, detailed below (Frohlich et al. 2009; Lavigne et al. 2021). in several steps and implemented with the open-source Tonga is also situated along the path of devastating QGIS package, is that it is fully adapted to the context of tropical cyclones such as Isaac (March 1982; Reardon Tonga, where open-access datasets are available for rep- and Oliver 1983), Waka (December 2001; Hall 2004), Ian resenting the distribution of population and buildings (January 2014; Havealeta et al. 2017), and Gita (January before the 2022 tsunami. The data include (i) opensource 2018; Caritas 2018). The islands are also listed globally GIS layer of buildings, (ii) Google Earth images showing among the most vulnerable to sea-level rise (Magee et al. buildings, (iii) household information, and (iv) popula- 2016; Mimura 1999). tion numbers from the latest census. The vulnerability Given its levels of exposure to the aforementioned haz- maps and estimates of potential victim numbers were ards, the Kingdom of Tonga holds second place among 172 also tested against the results of a satellite-based post- countries covered by the World Risk Index 2018 (index tsunami damage assessment survey of Tonga generated value of 29.42: very high; World Risk Report 2018). With by UNOSAT (2022). the help of foreign partners, the government has accord- ingly been developing risk assessment and preparedness Step 1: operational definition of the coastal belt plans, including coastal development and emergency man- A band of terrain exposed to the hazard of interest and agement consolidation within the framework of interna- containing potentially vulnerable settlements needs to tional disaster risk reduction guidelines (Bolton et al. 2020; be delineated. Some studies focus on a loosely defined Fakhruddin et al. 2019; Jayavanth et al. 2009; Sattler et al. coastal belt a few kilometers wide (Andrew et al. 2019; 2020; Simpson et al. 2011). Freely available data providing Finkl 2004), but precise operational definitions in tropi - population numbers and building locations are embedded cal Pacific island contexts remain scarce (Dickinson in census and government reports, and these are exam- et al. 1994; Eliot et al. 2020; Nunn and Campbell 2020; ined below in the methodology and discussion sections. Nunn and McNamara 2019). In Tonga, one previous Although relatively incomplete and imprecise compared approach defined the coastal belt based on vegetation to datasets available in wealthier countries, the reliability criteria (Sykes 1981; Burley 2007). However, for assessing of the information they contain can be enhanced by data tsunami hazards such criteria are no substitute for run- cleansing and systematic cross-checking. up height as the best reference frame. In this paper, the coastal belt is defined between the highest astronomical Objectives of the study tide limit and a user-defined critical elevation contour. We propose a simple methodology dependent on a small The shoreline is defined using the Global Self-consist - quantity of open-access data for performing rapid and ent, Hierarchical, High-resolution Geography database Thomas et al. Geoenvironmental Disasters (2023) 10:4 Page 4 of 18 Fig. 2 Population density (pop./km ) in Tonga by villages ( TSD 2017) (GSHHG 2017). Given the small size of the islands and Step 3: inventory and distribution of buildings the highest run-ups reported (Clark et al. 2011; Fritz The vector-format GIS database (OCHA 2022c ) docu- et al. 2011; Government of Tonga 2022), the 0–30 m ele- ments most of the infrastructure in the country. It was vation band covers a comprehensive range of possibilities automatically extracted from satellite imagery and con- from small to potentially large tsunamis. In this study, firmed by ground-truth checks, some of them on the topographic data are extracted from the open-access main island. The 2016 census also reports household Shuttle Radar Topography Mission (SRTM) global data- counts per village, differentiating between private houses set (SRTM 2015), which has a horizontal resolution of 1 and institutional buildings (TSD 2019). However, an arc-second (~ 30 m at the equator) and a minimum verti- absence of data was apparent in 12 villages, most of them cal accuracy of 16 m with 90% confidence (Mukul et al. with reported residents (Table 1), and 16 areas were iden- 2017). The vertical reference datum for the SRTM dataset tified as containing buildings but not situated within the is mean sea level. boundaries of a village—thus lacking population figures (Table 2). Step 2: inventory of population distribution data The 2016 census provided crucial information about Step 4: identification and correction of errors population distribution at four administrative levels: by cross‑verification between datasets country, division, district, and village (levels 0 to 3) (TSD Given that some of the datasets contain gaps, it was on 2017, 2019; note that the 2021 census was still in its pre- occasions difficult to link a specific village with a popu - liminary stage in the aftermath of the HTHH eruption lation statistic to a spatial distribution of buildings. A and, therefore, not used in this study). The smaller census cross-verification between the four datasets was con - units are thus the 166 villages scattered across the archi- ducted in order to identify errors. pelago, with their boundaries identified in a GIS layer Table 1 shows 11 villages with reported inhabitants and dating from February 2018 (OCHA 2022b). The popula - buildings in the census but no buildings reported in the tion density of each village varies considerably, as shown GIS layer. For this study, the GIS layer of buildings was in Fig. 2. The largest island, Ha’atu’a (37.78 km ) hosts manually completed and updated through detailed visual 522 inhabitants; the smallest, Fata’ulua (0.09 k m ), 227; inspection of all the villages using imagery captured in and the most populated, Kolofo’ou (3.62 km ): 8265. 2018 and provided by Google Earth (Table 1). As a result, T homas et al. Geoenvironmental Disasters (2023) 10:4 Page 5 of 18 Table 1 Update of the GIS layer of buildings for 11 villages using confirms that all types of infrastructure, including make - Google Earth shift dwellings, were taken into account. Table 2 focuses on the 16 areas situated outside village Village Population No. of No. of buildings mapped in census buildings in from Google Earth in the perimeters but with buildings reported in the GIS layer. census GIS layer No link can be established with an existing village or pop- ulation number. In this study, these areas were termed Feletoa 363 58 92 undefined villages and considered as census units on the Hunga 178 39 90 same basis as other villages. The presence of buildings, Vaka’eitu 24 5 15 whether residential or not, in the GIS layer indicates that Makave 381 72 84 the population of these undefined villages needs to be Matamaka 79 23 46 included in the same way as the other villages (Table 2). Mounu 1 1 2 For all other Tonga islands, the census indicates an aver- Nuapapu 98 23 34 age household size of 5.5 people (Scott and Browne 1989; Okoa 261 48 68 TSD 2019). On that basis, the population of undefined Ovaka 97 20 31 villages was estimated as follows: no. of buildings × 5.5, Ta’anea 644 128 133 rounded to the nearest unit. Utungake 285 57 70 Euakafa 6 2 7 Step 5: delineation of built‑up areas The GIS layer of built-up areas was crafted directly from the GIS layer of buildings while excluding road networks. A process aimed at cutting away spaces between build- Table 2 Population estimates for undefined villages ings was implemented while maintaining a 50 m buffer Undefined villages No. of buildings Population calculated on area around the buildings, and was followed by merging in GIS layer the basis of 5.5 individuals/ the polygons obtained. A further 30 m band was shaved household off each resulting polygon, thus leaving a 20 m perimeter Futu 4 22 of land around each building or group of closely spaced Kao 1 6 buildings (typically < 100 m apart; Fig. 3a). These values Kelefesia 3 17 are consistent with methods previously used (Loriot and Lotuma 7 39 Di Salvo 2008), take account of the density and disparity Mafana 1 6 of buildings in Tonga, and are compatible with the out- Makaha’a 2 11 door lifestyle of Tonga’s inhabitants, who typically spend Mandala Resort 2 11 time outdoors within a 20 m radius of their residence Mu’omu’a 1 6 (Bolton et al. 2020; Jin et al. 2014). Niniva 2 11 Nuku 2 11 Step 6: delineation of populated areas Nuku’alofa Harbor 84 462 This component of the dasymetric mapping approach Oneata 4 22 involves defining populated areas, i.e., built-up areas Tofua 1 6 where the population of the census unit is effectively con - Tonumeia 2 11 centrated (rather than spatially averaged across the cen- Uoleva 15 83 sus unit, as would be shown in choropleth maps). Here, Uonukuhihifo 1 6 populated areas were generated as a GIS vector layer defined by the outlines of the built-up areas (see Step 5) and coupled with the census population values. Note that maps from the TSD were delivered after the 2022 erup- the number of buildings in the GIS layer does not exactly tion to locate ‘populated places’ (OCHA 2022d), which match the number given in the census but stays in the is a notion similar to the populated areas we are defin - same range. To explain this difference, we assumed (i) ing here. However, we performed compatibility tests that any new houses and infrastructures had been con- that show the data are incomplete and do not match structed since the 2016 census, and (ii) that our method the building data (for example, to the west of Fua’amotu of building delimitation differed from the one used in the International Airport, Fua’amotu village is not indicated). census (for example, it is likely that we would define a This study consequently ignored the populated places house and garage or outhouse as two buildings, instead of given in OCHA (2022d) and defined the populated areas one in the census). The population data were kept identi - using the built-up areas obtained at Step 5. cal. On the basis of the Google Earth search, this study Thomas et al. Geoenvironmental Disasters (2023) 10:4 Page 6 of 18 The boundaries of the populated areas are defined by a using visual interpretation of pre- and post-event high- 200 m spatial aggregation algorithm applied to the built- resolution satellite images shared by the USGS Hazards up areas generated at Step 5 to take into account building Data Distribution System (HDDS) portal (public satellite dispersal, then a 50 m erosion buffer is applied to achieve sources). The visual inspection of images highlighted the a tighter geographic fit to the buildings (i.e., each building damaged buildings, tsunami-related shoreline changes, or cluster of buildings is surrounded by a uniform band and the assessment of flooding extent (Fig. 3d). of terrain 150 m wide; Thomas et al. 2021). The intersec - tion of populated areas with census unit boundaries (in Results Tonga, census units are called villages) allows the popula- This rapid, low-cost methodology for assessing tsunami tion of the village to be directly associated with the corre- impacts on humans and infrastructure delivers a num- sponding unit(s) in the populated areas layer. Depending ber of dasymetric vulnerability maps. All the open-access on the distribution of buildings, a village can be com- shapefiles of the populated areas are accessible in greater posed of more than one populated area (see Fig. 3b). detail in the Additional file 1. Here we present some result highlights for later discussion. Step 7: delineation of coastal populated areas For each user-defined coastal elevation band and for The last task involves intersecting the populated areas each village, a precise estimation of population and build- with the coastal band defined at Step 1, thereby defin - ing numbers are given in Additional file 2. Those figures ing ‘coastal populated areas’, i.e., areas lying within reach serve as a quick reference tool for identifying and locat- of a given tsunami run-up magnitude. This process is ing where the highest disaster management challenges repeated for each elevation band up to + 30 m by incre- occur in the landscape. An extract of these data is pre- ments of 5 m. The population number is then allocated to sented in Table 3 and the full dataset for Tonga is plot- these coastal populated areas proportionally to polygon ted in Fig. 4. Strong contrasts in population distribution area, assuming thus a uniform population density within between each of the five island groups are highlighted. populated areas (Fig. 3c). The Ha’apai group is composed of 68, mostly low-lying ‘Coastal populated areas’ (step 7) and ‘populated areas’ islands with more than 50% of the global Tongan popula- (step 6) are thus directly generated by reference to the tion and more than 70% of the buildings located below ‘built-up areas’ (step 5), which are themselves upscaled the 10 m elevation contour. ‘Eua is hilly, with peaks above from the ‘buildings’ layer (step 3). The dasymetric map - 300 m and more than 85% of the population living above ping methodology thus operates at an aggregated level 30 m. Both island groups were struck by the 2022 tsu- rather than at the basic building footprint level because nami, with several buildings damaged. Figure 5 provides it allows buildings potentially missing from the data layer a close-up of some of the villages impacted by the tsu- (e.g., new, or unreported) to be included (thereby erring nami. In all villages, a jump in population and building on the side of caution in the risk assessment exercise). It numbers is observed between 5 and 10 m, revealing a also allows areas adjoining buildings (within the previ- concentration of human settlements below the 10 m ele- ously defined 150 m buffer belt) to be included as daily vation contour and thus greater vulnerability to coastal living space because life in Pacific island cultures includes hazards. a lot of time spent outdoors. An efficient way of understanding those datasets con - sists in mapping the entire population on each island, Step 8: model quality assessment from independent and then focusing on the populated areas that pinpoint in damage data each village where the population actually lives (Thomas The damage assessment data were retrieved from the EU’s et al. 2021). The result presented here focus on the Vava’u Copernicus Emergency Management Service (CEMS) group of islands, around the village of Ta’anea, in order and the United Nations Satellite Centre (UNOSAT), to illustrate the methodology (Fig. 6). The GIS layers are hosted by UNITAR. The data consist of different vector available in the Additional file 1 in order to zoom in. layers as ESRI shape files, Google Earth KML and GeoJ - This study estimates overall that 83% of the popula - SON formats (CEMS 2022; USGS HDDS 2022), prepared tion of Tonga lives within the 0–30 m elevation band, and (See figure on next page.) Fig. 3 Methodology describing the implementation of a low-cost toolbox for disaster vulnerability mapping. Components of the seven steps are displayed in the right panels. a Steps 3, 4 and 5 use the buildings inventory to define built-up areas. b Steps 2, 4 and 6 use the population inventory to define populated areas. c Steps 1 and 7 use an operational definition of the coastal belt to outline coastal populated areas. d Step 8 uses damage inventory to discuss the model quality. Green dots: buildings (OCHA 2022c). Yellow dots: buildings damaged by the tsunami in 2022 (CEMS 2022; USGS HDDS 2022). Black crosses: churches (churches are the only non-residential buildings described in the existing GIS layer: fire stations, police stations, schools and other public buildings are absent, thereby emphasizing the lacunar character of the information available) T homas et al. Geoenvironmental Disasters (2023) 10:4 Page 7 of 18 Fig. 3 (See legend on previous page.) Thomas et al. Geoenvironmental Disasters (2023) 10:4 Page 8 of 18 Table 3 Example of number of inhabitants and buildings by village, estimated for different coastal belt widths Village name < 1033 m (highest elevation) < 25 m elevation < 5 m elevation Population Buildings Population Buildings Population Buildings (No.) (No.) (No.) (%) (No.) (%) (No.) (%) (No.) (%) Pangaimotu 657 204 386 59 164 80 47 7 3 1 ‘Utulei 116 48 52 45 30 63 19 16 0 0 Nga’unoho 181 69 149 82 66 96 36 20 23 33 ‘Utungake 285 73 203 71 70 96 111 39 46 63 Tapana 3 6 3 100 6 100 1 33 1 17 42% lives below 10 m. Likewise, 87% of buildings occur of Tonga 2022; UNOSAT 2022). Figure 4 indicated that below the 30 m, and 53% below the 10 m contour. At least 62% of the population (n = 62,677) lives below the 15 m 14% of the population lives below 5 m. Finally, the graphs contour in all of Tonga (61% when restricting to Tonga- clearly highlight the large disparity in population and tapu, ‘Eua and Ha’apai islands). The CARE (2022) report building densities between two villages situated in con- recorded 84,776 affected people from the tsunami and trasting topographic settings. volcanic fallout combined. Our map-derived estima- tions indicate that 74% of buildings in Tonga are located Discussion in a coastal tsunami hazard area below 15 m, meaning all Here we test the predictive power of the methodol- those buildings may have suffered from tsunami run-ups ogy based on two real tsunamis in Tonga: the January in 2022. The additional impact of thick ash fall causing 15, 2022 and September 29, 2009 events. We also show roofs to collapse tallies with reports claiming that this the relevance of this methodology for other types of double disaster affected 80% of Tonga (OCHA 2022a) natural hazards such as storms, volcanic eruptions, and and caused almost 100% damage to buildings (GFDRR earthquakes. 2022). Here we gather information restricted to tsu- nami damage and discuss links with the results obtained Case study #1: 2022 tsunami through the methodology. Satellite imagery has revealed considerable damage at UNOSAT (2022) provides an overview of the dam- locations on Tongatapu, ‘Eua and Ha’apai islands, suggest- age to buildings for several districts through satellite ing that run-ups reached the 15 m contour (Government imagery analysis. An example of this information is given Fig. 4 Cumulative distribution of population and buildings for all Tonga and for each of the 5 divisions (group of islands), paired by color as a function of elevation. Full lines: population distribution. Dotted lines: buildings distribution T homas et al. Geoenvironmental Disasters (2023) 10:4 Page 9 of 18 Fig. 5 Cumulative distribution of population and buildings in four selected villages damaged by the tsunami in 2022. Full lines: population distribution. Dotted lines: buildings distribution. Nomuka and Mango belong to the Ha’apai island group. ‘Ohonua and Ta’anga are in ‘Eua group Fig. 6 Example of dasymetric maps produced for locating populated areas in Tonga. a Buildings layer on Google Earth imagery background. b Coastal belts below 30 m elevation. c Coastal populated areas in Table 4, with direct links to the number of buildings, ‘Eua and Ha’apai shows sharp contrasts in topography, populated areas and inhabitants for each district ana- but the tsunami had similar impacts on structures stand- lyzed. As shown, the information supplied is imprecise ing below the 15 m contour on both islands. but can be advantageously refined by using the dasym - UNOSAT Maps (2022) provides an atlas of disaster- etric maps produced in this study for obtaining a rapid related damages on Tonga based on satellite imagery. assessment of the situation and a first-order, reasonably The assessment focuses on a selection of villages highly accurate estimation of vulnerable buildings and number affected by the tsunami, with a confirmed number of of residents in population clusters. As explained in Fig. 5, damaged buildings. The methodology implemented Thomas et al. Geoenvironmental Disasters (2023) 10:4 Page 10 of 18 Table 4 Comparison of damages to buildings between UNOSAT Table 6 Estimation of damages to buildings in specific (2022) data and results from this study populated areas: a comparison between three datasets Information provided by Information gained from this study Location No. of buildings Coefficient c UNOSAT (2022) Damaged bd In the In the Island District Buildings No. of Existing Estimated populated village b c damaged populated buildings no. of area bpa bv area units inhabitants Tungua 11 82 83 0.01 ‘Eua Fo’ou 1 13 776 2150 Fonoifua 30 30 30 1 Prope 139 21 1058 2795 ‘Atata 72 89 104 0.47 Ha’apai Lulunga 11 7 325 923 ‘Ohonua 48 118 500 0.85 Mu’omu’a 145 4 305 432 Ta’anga 6 6 95 1 ‘Uiha 4 5 407 695 Futu 1 7 99 0.94 Mango 26 26 26 1 Nomuka1 61 184 249 0.41 here shows an improvement in detecting the number Nomuka2 4 4 249 1 of potential buildings impacted, while also linking the a b c From UNOSAT Maps (2022); this study; from TSD (2019) geographic information directly to an estimated maxi- mum number of vulnerable inhabitants (Table 5). Our estimation of buildings damaged is often close to, the actual damage estimated by UNOSAT. As c although occasionally much higher, than the imagery- approaches 0, the estimation obtained stays close to the based numbers reported by the UNOSAT survey. TSD data, with little improvement. Among the examples Compared to the existing dataset of buildings per vil- compiled in Table 6, improvement in the precise knowl- lage, this overestimation clearly highlights an improve- edge of vulnerable buildings is gained for 4 locations, ment in the results (Table 6). A comparison coefficient where c > 0.85 (‘Ohonua, Ta’anga, Futu, and Nomuka2). c, defined below, is calculated to quantify the gain in Tungua, ‘Atata and Nomuka1 perform less well, perhaps accuracy when estimating vulnerable buildings using because our approach only considers the run-up distance the three datasets: (i) the number of buildings in the of the tsunami, i.e., the elevation range of the impact, and village, bv, based on the TSD (2019) data; (ii) the num- not the form of tsunami propagation or disparities in the ber of buildings in the populated area, bpa, identified in intrinsic mechanical strength of the buildings exposed this study; and (iii) the number of effectively damaged to it (our estimation is based on the assumption that all buildings, bd, using the UNOSAT Maps (2022). the buildings in the relevant coastal band are damaged). Fonoifua and Mango were completely destroyed by the (bv − bpa) c = tsunami. This is confirmed by numbers from the TSD (bv − bd) dataset as well as by this study’s dataset, and provides in itself a validation of the methodology. Furthermore, Coefficient c ranges between 0 and 1. The closer c gets to any overprediction of building damage in this study may 1, the closer the estimation from this study approximates Table 5 Comparison of damages to buildings in specific populated areas Information provided by UNOSAT Maps (2022) Information gained from this study Location No. of buildings Populated area Elevation (m) No. of buildings No. of damaged inhabitants Tungua 11 Tungua < 10 82 178 Fonoifua (Mu’omu’a) 30 Fonoifua < 10 30 25 ‘Atata 72 ‘Atata < 10 89 122 ‘Eua 48 ‘Ohonua (Prope) < 15 118 254 6 Ta’anga (Prope) < 15 6 42 1 Futu (Fo’ou) < 20 7 23 Mango (Mu’omu’a) 26 Mango < 10 26 27 Nomuka (Mu’omu’a) 61 Nomuka1 < 10 184 227 4 Nomuka2 < 15 4 28 T homas et al. Geoenvironmental Disasters (2023) 10:4 Page 11 of 18 Fig. 7 Detailed map of individual buildings damaged by the 2022 tsunami on ‘Eua island. Gray polygons: buildings included in this study (OCHA 2022c). Yellow dots: buildings confirmed damaged by UNOSAT Maps (2022) advantageously compensate for likely errors in the defini - in Nomuka, and the shoreline retreated by up to 10 m on tion of the coastal belts when based on SRTM elevation Nomuka and 30 m on Mango (CEMS 2022; GFDRR 2022; data, which have generated cases of underprediction in Pleasance 2022; UNOSAT Maps 2022). other studies (Kulp and Strauss 2016). In the context of a future tsunami event, maps such The good match between the UNOSAT data and our as Figs. 7 and 8 can be used to precisely locate in each mapping results presented in Tables 5 and 6 highlights village how many inhabitants were potentially vulner- the value of the methodology presented here. Geographic able to that particular hazard and could have been precision is additionally provided by the GIS maps of affected by its estimated magnitude. Such prior knowl - populated areas, here shown for ‘Eua in Fig. 7, and for the edge can serve as an important decision-making tool. three islands of the Ha’apai group in Fig. 8, which were For example, by coupling the population cluster maps severely affected by the 2022 tsunami in all of Tonga. with tsunami hazard scenarios, this would help to focus Entire villages on Mango and Fonoifua were swept away, attention on villages most exposed to a given tsunami 13 houses were flooded between the coast and the lake type and propagation pattern. It would also help to Thomas et al. Geoenvironmental Disasters (2023) 10:4 Page 12 of 18 Fig. 8 Overview of buildings damaged by the 2022 tsunami on Nomuka, Mango and Fonoifua islands. Gray polygons: buildings included in this study (OCHA 2022c). Yellow dots: buildings confirmed damaged by UNOSAT Maps (2022). Green line: shoreline before the tsunami (GSHHG 2017). Yellow line: shoreline retreat after the tsunami by UNOSAT Maps (2022) calibrate the logistics of rescue operations (e.g., quan- claiming 9 lives on Niuatoputapu and Tahafi islands, both tities of freshwater, first-aid equipment and food to be situated in the northeast of Tonga (Lay et al. 2010; World delivered), including passenger capacity on vessels or Bank 2022). aircraft for the temporary relocation of disaster vic- On Tahafi, run-ups reached 15 to 22 m on the south - tims. Despite more abundant first-response and health - western side, damaging fishing boats and one house care staff and facilities on the main island, residents on (Clark et al. 2011; Fritz et al. 2011; Okal et al. 2010; Wil- its west coast were nonetheless highly impacted by the son et al. 2009). Despite these high values, Tahafi is a 2022 tsunami. The maps and tables showcased in this steep-sided island with currently 31 inhabitants and 38 study also highlight the value of precise and regularly buildings counted, all situated above the 30 m contour. updated census data in geographically isolated islands, On Niuatoputapu, the villages of Falehau, Vaipoa and particularly given that urban growth and population Hihifo are situated on the northwest shore, where run- migration between islands occur continuously and are ups reached the 5 m contour (Clark et al. 2011; Wilson unlikely to decline in the near future (Lolohea 2016). et al. 2009). Several reports indicate around 135 to 145 For example, more than 40% of the population has buildings damaged out of a reported total of 225 to 228, moved away from the Ha’apai island group since 2011, mostly in the main village (Hihifo) and all occurring mostly settling on Tongatapu (GFDRR 2022). below the 5 m contour (Fritz et al. 2011; Government of Tonga 2009a, b; WHO 2009; World Bank 2022). Case study #2: 2009 tsunami Using the 2016 census data, Fig. 9 displays an overlay The second dasymetric mapping test case is the tsunami of Hihifo village on the post-tsunami survey of building triggered in Samoa by the September 2009 earthquakes, locations in 2009 (Clark et al. 2011). Only 7 buildings T homas et al. Geoenvironmental Disasters (2023) 10:4 Page 13 of 18 Fig. 9 Hihifo village layout with overlay of destroyed buildings during the 2009 tsunami. Black polygons: buildings included in this study (OCHA 2022c). Yellow polygons: buildings damaged during the 2009 tsunami (Clark et al. 2011) below the 5 m contour were spared among the 44 exist- from the coral reef surrounding the island, which pro- ing before the 2009 tsunami. Whereas 38 inhabitants tected the villages from excessive run-up magnitudes on are currently recorded as residing below the 5 m con- its northwest coast (Fritz et al. 2011). Although not uni- tour, population estimates at the time were closer to 242. versally verified, tsunami amplitudes can be mitigated by These figures highlight the magnitude (57%) of post-tsu - the presence of healthy coral reefs (Fernando et al. 2005; nami emigration, not just in temporary buildings in Fale- Ferrario et al. 2014; Hardy and Young 1996; Harris et al. hau (Clark et al. 2011; WHO 2009), but by a long-term 2018; Karim and Nandasena 2022; Kunkel et al. 2006; decision to definitively relocate housing away from areas Monismith et al. 2015; Roger et al. 2014). This unique exposed to coastal hazards—generally to higher eleva- ecological asset should be integrated in future risk man- tions. Niuatoputapu island nonetheless appears to benefit agement studies on Tonga. Thomas et al. Geoenvironmental Disasters (2023) 10:4 Page 14 of 18 Higher run-ups occurred at the southern tip of sites, like Italy around Mt. Vesuvius (Gugg 2022), Hawai’i Niuatoputapu along a near-shore fringing reef, with on Kīlauea (Meredith et al. 2022) and Java around Mt. observed flow depths of up to 6 m (Wilson et al. 2009). Merapi (Garcia-Fry et al. 2022) are densely inhabited and This is consistent with the damage sustained by the air - require regular updates of populated areas. For example, strip, which is situated below 10 m and was partially when the Niuafo’ou volcano erupted in September 1946, inundated at the southern end of the runway (Clark et al. lava flows and ash clouds destroyed infrastructures and 2011). This is the only air connection to other islands in vegetation all over the island (Rogers 1981; Taylor 1991). Tonga and highlights the dangers of placing communica- The entire population (1300 inhabitants) was relocated tions infrastructures at low elevations. The only undam - permanently on other islands. The population data from aged public building on the entire island was the high the last census show that 517 inhabitants returned per- school, which stands above the 5 m contour, emphasiz- manently to the island among 8 different villages despite ing here also the critical importance of choosing elevated the volcano still being highly active (more than 10 erup- ground for primary infrastructure whenever possible. tions since 1814; Taylor 1991). With 247 buildings scat- Today, the spatial distribution of population densities and tered along the east coast of the island, the population is the spatial distribution of buildings appears to indicate still highly vulnerable to volcanic hazards, and getting to that the hospital (completely destroyed in 2009), the pri- know their precise location is a necessary feature for risk mary schools, and the churches have been rebuilt above management in case of an eruption. the 10 m contour. This is confirmed by several reports By its geographic position, Tonga is also exposed to also indicating that water supplies, the police station, strong earthquakes, such as the previously mentioned and 73 houses were built out of cyclone-resistant mate- M 7.2 event in June 1977 causing serious damage on rial and on safer and higher ground, testimony to a grow- ‘Eua and Tongatapu islands. A similar event today would ing awareness of natural hazards in land-use planning in directly impact ~ 80,000 inhabitants and could damage Tonga since 2009 (Government of Tonga 2009a; World up to 33,268 service facilities and houses. Thus, a com - Bank 2022). plementary feature to the dasymetric mapping toolkit would be an updated GIS of Tonga, with all essential Application to other natural hazards contexts infrastructures explicitly positioned: civil security cent- In the aftermath of the 2009 and 2022 tsunamis, the fore- ers, hospitals, schools, food and freshwater supply cent- most concern was to relocate residents, to clean up, and ers. The building data could be enriched by an indication to stay on alert during the cyclone season. An accumu- of their function in order to differentiate between resi - lation of hazards hitting Tonga would significantly dam - dential areas and workplaces (which do not share the age the infrastructure of the islands and strongly increase same population balance during day and night). A sim- population vulnerability. The impact of the 2022 tsu - ple search on Google Maps and Maps.me confirmed how nami, for example, was compounded by the ash fall from currently difficult (often impossible) it is to find accurate the eruption, with damaging consequences beyond the data about healthcare facilities on islands in Tonga. Thus, tsunami-exposed coastal belt. Adding climate-change- an open-source GIS stands out as an essential tool for related impacts, Tonga is thus exposed to a cumulative risk management. Similar issues arise in the context of litany of disasters, highlighting the urgent need for dis- a cyclonic event such as the strongest and dramatic cat- aster management strategies (Ammann 2013; GFDRR egory-5 cyclone (Ian) of January 2014 (Havealeta et al. 2022; UNISDR 2005, 2009). 2017; Johnston 2015). Damages across the Ha’apai island Although focused on coastal inundation hazard on group were considerable, with 18 villages affected, 1094 the basis of an elevation/run-up criterion, with neces- buildings destroyed, and 2335 people relocated (Govern- sary adjustments of parameters and suitable informa- ment of Tonga 2014). This study documents 6125 cur - tion sources, the methodology presented in this study rently vulnerable residents on Ha’apai, and indicates that can address any type of natural hazard. As shown by along the path taken by the cyclone Ian, 2202 inhabitants the latest eruption of Hunga Tonga-Hunga Ha’apai, ash are now located among 1158 buildings. Finally, reports fall, for example, can destroy buildings even kilometers indicate that 17 schools were destroyed, impacting 1293 away from the eruption site. By overlapping tracking children (Government of Tonga 2014), while at the same models of volcanic ash clouds with the mapping toolkit time many infrastructures already meeting cyclone- of populated areas provided here (Carn et al. 2009; Fil- resistant building design were saved (World Bank 2022). izzola et al 2007; Searcy et al. 1998; Webley and Mastin These examples further highlight the urgent need for a 2009), the identification of vulnerable inhabitants can full GIS database capable of specifying precise building be easily assessed. Moreover, many hazardous volcanic functions. T homas et al. Geoenvironmental Disasters (2023) 10:4 Page 15 of 18 Abbreviations Conclusion CEMS Copernicus Emergency Management Service This study focused on a rapid, low-cost approach for HTHH Hunga Tonga-Hunga Ha’apai locating and quantifying vulnerable residents and infra- SRTM Shuttle Radar Topography Mission TSD Tonga Statistics Department structures as precisely as possible. It involves aggregating UNITAR United Nations Institute for Training and Research open-access datasets, cross-checking for consistency, and UNOSAT United Nations Satellite Centre generating dasymetric population maps using an open- USGS HDDS United States Geological Survey Hazards Data Distri bution System access GIS package. Its application to the population of Tonga, which is scattered over 45 islands and lives mostly Supplementary Information in coastal areas (~ 62% of inhabitants reside below 15 m The online version contains supplementary material available at https:// doi. elevation), was tested on a coastal belt defined by the org/ 10. 1186/ s40677- 023- 00235-8. limits of the highest tsunami run-up (+ 30 m) obser ved in the built-up areas of Tonga after the tsunami of Janu- Additional file 1. Shapefiles of the populated areas in Tong, all details ary 15, 2022. High concentrations of built infrastructure are provided in the associated Word document (Supplementary Material Dataset PA.docx). in the tsunami exposure zone (~ 74% below the 15 m Additional file 2. Excel file providing a precise estimation of population elevation contour) illustrate the high level of vulner- and building numbers for each user-defined coastal elevation band and ability of Tonga to ocean-related hazards. The mapping for each village. approach highlights the large disparity in population and building densities from one island to another, with vil- Acknowledgements lages positioned in a diversity of topographic settings. We are grateful to the editor and to two anonymous reviewers, who helped to These results improve existing datasets from the Tonga improve the paper. Statistics Department by providing more accurate geo- Author contributions graphic limits for population and buildings through the Study conception and design: BT, JR, and YG. Material preparation, data collec- coastal populated areas. In addition to documenting an tion, and data analysis: BT, SA, and JR. First draft was co-written by BT and JR, edited by YG, with all authors critiquing subsequent versions and approv- entire population of scattered islands from the angle of ing the final manuscript thereafter. All authors read and approved the final its exposure to tsunamis (in particular the 2009 and 2022 manuscript. tsunamis), we developed a dasymetric population map- Funding ping methodology—achieved in a short time span (less Open Access funding enabled and organized by Projekt DEAL. This work was than two days after the 2022 tsunami) and using a small partly funded by New Zealand’s Strategic Science Investment Fund (SSIF). number of open-access datasets—to obtain a maximum Availability of data and materials number of vulnerable residents and buildings potentially All data generated and analyzed during this study are included in this pub- affected by coastal hazards. Using the population data of lished article and its supplementary information files. 2016 (the only reliable dataset at the time of writing), this paper has aimed to provide a snapshot of population dis- Declarations tribution and possible rapid decision-making actions that Competing interests could have been taken following the 2022 tsunami event. The authors declare that they have no known competing financial interests By estimating a fraction of vulnerable population per or personal relationships that could have appeared to influence the work populated area and affording precise visualization tools, reported in this paper. this mapping-focused methodology is likely to appeal to a number of academic and operational stakeholders for Received: 11 August 2022 Accepted: 1 February 2023 its transferability to coastal zones more generally, and particularly to insular settings where first responders and risk management organizations need to acquire and ana- lyze complex but reliable primary datasets. If enhanced References Ammann WJ (2013) Disaster risk reduction. In: Bobrowsky PT (ed) Encyclope- by (1) regular updates of annual census data, by (2) dia of natural hazards, encyclopedia of Earth sciences series. Springer, monitoring of inter-island migratory movements, by (3) Dordrecht, pp 170–175. https:// doi. org/ 10. 1007/ 978-1- 4020- 4399-4_ 92 tracking patterns of relocation and reconstruction after Andrew NL, Bright P, de la Rua L, Teoh SJ, Vickers M (2019) Coastal proximity of populations in 22 Pacific Island countries and territories. PLoS ONE disastrous events, and by (4) analyzing vulnerability pat- 14:e0223249. https:// doi. org/ 10. 1371/ journ al. pone. 02232 49 terns on remote islands with few communication links, BBC (2022) Tonga: how aid deliveries try to avoid bringing in Covid. BBC News the primary datasets and the methodology presented Online. https:// www. bbc. com/ news/ 60066 470. Accessed 2 Feb 2022 Bolton A, Ashworth M, Akolo Fau S, Folau Fusi SK, Williamson J, Ofanoa R here can gain in accuracy, thereby assisting in planning (2020) Strengthening adaptation planning and action to climate-related increasingly precise and proportionate emergency rescue health impacts in Pacific islands countries: Tonga. Glob J Health Sci 12:55. targets in the future. https:// doi. org/ 10. 5539/ gjhs. v12n3 p55 Thomas et al. Geoenvironmental Disasters (2023) 10:4 Page 16 of 18 Bryan WB, Stice GD, Ewart A (1972) Geology, petrography, and geochemistry eruptions. Transp Res D Transp Environ 104:103–189. https:// doi. org/ 10. of the volcanic islands of Tonga. J Geophys Res 77:1566–1585. https:// doi. 1016/j. trd. 2022. 103189 org/ 10. 1029/ JB077 i008p 01566 GEBCO (2014) GEBCO 2014 grid. https:// www. gebco. net/ news_ and_ media/ Burley DV (2007) In search of Lapita and Polynesian Plainware settlements gebco_ 2014_ grid. html. Accessed 2 Feb 2022 in Vava’u, Kingdom of Tonga. Oceanic explorations: Lapita and Western GFDRR (2022) The January 15, 2022, Hunga Tonga-Hunga Ha’api Eruption and Pacific settlement, pp 187–198. https:// www. jstor. org/ stable/ j. ctt24 h9sg. Tsunami, Tonga: Global Rapid Post Disaster Damage Estimation (GRADE) 13 Report. Global Facility for Disaster Reduction and Recovery, Government Campbell MD, McKay GR, Williams RL (1977) The Tonga earthquake of 23 June, of Tonga and World Bank. https:// relie fweb. int/ report/ tonga/ janua ry- 15- 1977. Some initial observations. Bull N Z Soc Earthq Eng 10(4):208–2182022- hunga- tonga- hunga- haapi- erupt ion- and- tsuna mi- tonga- global- CARE (2022) Emergency situation report—Tonga Volcano and Tsunami rapid- post. Accessed 29 March 2022 Response. CARE International. https:// relie fweb. int/ report/ tonga/ emerg Government of Tonga (2009a) Tonga: operation Niuatoputapu—day 6, 5 Oct ency- situa tion- report- tonga- volca no- and- tsuna mi- respo nse- 28- febru ary- 2009a. Government of Tonga. Published 5 October 2009a. https:// relie 2022. Accessed 29 March 2022fweb. int/ report/ tonga/ tonga- opera tion- niuat oputa pu- day-6- 5- oct- Caritas (2018) Being prepared helps Tonga recover from cyclone Gita. https:// 2009a. Accessed 29 March 2022 www. carit as. org/ 2018/ 03/ being- prepa red- helps- tonga- recov er- from- Government of Tonga (2009b) Tonga: operation Niuatoputapu—day 7, 6 cyclo ne- gita/. Accessed 13 Feb 2022 Oct 2009b. Government of Tonga. Published 6 October 2009b. https:// Carn SA, Krueger AJ, Krotkov NA, Yang K, Evans K (2009) Tracking volcanic sul-relie fweb. int/ report/ tonga/ tonga- opera tion- niuat oputa pu- day-7- 6- oct- fur dioxide clouds for aviation hazard mitigation. Nat Hazards 51:325–343. 2009b. Accessed 29 March 2022 https:// doi. org/ 10. 1007/ s11069- 008- 9228-4 Government of Tonga (2014) Tonga: tropical cyclone Ian Situation Report No. CEMS (2022) Copernicus emergency management service. https:// emerg ency. 3. Government of Tonga. Published 23 January 2014. https:// relie fweb. int/ coper nicus. eu/. Accessed 29 June 2022report/ tonga/ tonga- tropi cal- cyclo ne- ian- situa tion- report- no-3- 23- janua Clark K, Power W, Nishimura Y, Kautoke R’A, Vaiomo’unga R, Pongi ‘A, Fifita M ry- 2014. Accessed 20 Apr 2022 (2011) Characteristics of the 29th September 2009 South Pacific tsunami Government of Tonga (2022) First official update following the volcanic erup - as observed at Niuatoputapu Island, Tonga. Earth-Sci Rev 107:52–65. tion. Press Release, 18th January, 2022, https:// www. gov. to/ press- relea https:// doi. org/ 10. 1016/j. earsc irev. 2010. 12. 001se/ first- offic ial- update- follo wing- the- volca nic- erupt ion/. Accessed 18 Cummins P, Whatman J, Lahtinen AL, Beavan J, Walace L, Wiens D, Shore P, Jan 2022 Heeszel D, Bevis MG, Kendrick E, Taylor FW, Malolo T, Mafi K, Nonu S, Fatai GSHHG (2017) Global self-consistent, hierarchical, high-resolution geography T, Moala ‘A (2007) Interim report on studies of the Mw ~7.9 earthquake of database. Version 2.3.7 of June 15, 2017. https:// www. ngdc. noaa. gov/ 3 May 2006, Kingdom of Tonga. University of Texas, Institute for Geophys-mgg/ shore lines/. Accessed 18 Jan 2022 ics, Technical Report 204. http:// www- udc. ig. utexas. edu/ exter nal/ facil Gugg G (2022) Ordinary life in the shadow of Vesuvius: surviving the ities/ tech_ repor ts/ tech_ repor ts/ UTIGTR_ 0204. pdf. Accessed 14 Feb 2022 announced catastrophe. In: Świtek B, Abramson A, Swee H (eds) Extraor- Dickinson WR, Burley DV, Shutler R Jr (1994) Impact of hydro-isostatic holo- dinary risks, ordinary lives. Springer, Cham, pp 249–275. https:// doi. org/ cene sea-level change on the geologic context of Island archaeologi-10. 1007/ 978-3- 030- 83962-8_ 10 cal sites, Northern Ha’apai group, Kingdom of Tonga. Geoarchaeology Gusman AR, Roger J (2022) Hunga Tonga—Hunga Ha’apai volcano-induced 9:85–111 sea level oscillations and tsunami simulations. GNS Science webpage. Eliot I, Kumar L, Eliot M et al (2020) Downscaling from whole-island to an https:// doi. org/ 10. 21420/ DYKJ- RK41 island-coast assessment of coastal landform susceptibility to Metocean Gusman AR, Roger J, Noble C, Wang X, Power W, Burbidge D (2022) The change in the Pacific Ocean. In: Kumar L (ed) Climate change and 2022 Hunga Tonga—Hunga Ha’apai volcano air-wave generated impacts in the Pacific. Springer, Cham, pp 225–250. https:// doi. org/ 10. tsunami. Pure Appl Geophys 179:3511–3525. https:// doi. org/ 10. 1007/ 1007/ 978-3- 030- 32878-8_5s00024- 022- 03154-1 Fakhruddin B, Reinen-Hamill R, Robertson R (2019) Extent and evaluation of Hall JD (2004) The South Pacific and southeast Indian Ocean tropical cyclone vulnerability for disaster risk reduction of urban Nuku’alofa, Tonga. Prog in season 2001–02. Aust Meteorol Mag 53(4):285–304 Disaster Sci 2:100017. https:// doi. org/ 10. 1016/j. pdisas. 2019. 100017 Hardy TA, Young IR (1996) Field study of wave attenuation on an offshore coral Fernando HJS, McCulley JL, Mendis SG, Perera K (2005) Coral poaching wors- reef. J Geophys Res Oceans 101:14311–14326. https:// doi. org/ 10. 1029/ ens tsunami destruction in Sri Lanka. Eos 86:301. https:// doi. org/ 10. 1029/ 96JC0 0202 2005E O3300 02 Harris DL, Rovere A, Casella E, Power H, Canavesio R, Collin A, Pomeroy A, Ferrario F, Beck MW, Storlazzi CD, Micheli F, Shepard CC, Airoldi L (2014) The Webster JM, Parravicini V (2018) Coral reef structural complexity provides effectiveness of coral reefs for coastal hazard risk reduction and adapta- important coastal protection from waves under rising sea levels. Sci Adv tion. Nat Commun 5:3794. https:// doi. org/ 10. 1038/ ncomm s4794 4:eaa04350. https:// doi. org/ 10. 1126/ sciadv. aao43 50 Fifita VK, Sanchez AL, Catalan HN, Gordon D (2018) Assessing progress towards Havealeta M, Kaho F, Tuipulotu AA (2017) Tonga: cyclone Ian, Ha’apai Islands. the eradication of poverty in the Kingdom of Tonga. Statistics Depart- In: Bonito S, Minami H (eds) The role of nurses in disaster management in ment Tonga. https:// www. pover ty. ac. uk/ sites/ defau lt/ files/ attac hments/ Asia Pacific. Springer, Cham, pp 105–109. https:// doi. org/ 10. 1007/ 978-3- Asses sing- progr ess- towar ds- the- eradi cation- of- pover ty- Tonga- 2018. pdf319- 41309-9_ 11 Filizzola C, Lacava T, Marchese F, Pergola N, Scaffidi I, Tramutoli V (2007) Assess- Heeszel DS, Wiens DA, Cummins PR, Lahtinen A, Inoue H (2006) The May 3, ing RAT (Robust AVHRR Techniques) performances for volcanic ash cloud 2006 (Mw 7.9) Tonga earthquake: shallow thrust or slab tear? Evidence detection and monitoring in near real-time: the 2002 eruption of Mt. Etna from locally recorded aftershocks. AGU, Fall Meeting 2006, Abstract (Italy). Remote Sens Environ 107:440–454. https:// doi. org/ 10. 1016/j. rse. T21F-05 2006. 09. 020 Hunt T, Piper D (2022) Fishers run for lives as Tongan tsunami hits New Zea- Finkl CW (2004) Coastal classification: systematic approaches to consider in land, warnings continue. Stuff. https:// www. stuff. co. nz/ natio nal/ 12751 the development of a comprehensive scheme. J Coast Res 201:166–213. 2707/ fishe rs- run- for- lives- as- tongan- tsuna mi- hits- new- zeala nd- warni ngs- https:// doi. org/ 10. 2112/ 1551- 5036(2004) 20[166: CCSATC] 2.0. CO;2conti nue. Accessed 16 Jan 2022 Fritz HM, Borrero JC, Synolakis CE, Okal EA, Weiss R, Titov VV, Jaffe BE, Foteinis S, Imamura F, Suppasri A, Arikawa T et al (2022) Preliminary observations and Lynett PJ, Chan I-C, Liu PL-F (2011) Insights on the 2009 South Pacific tsu- impact in Japan of the Tsunami caused by the Tonga volcanic eruption nami in Samoa and Tonga from field surveys and numerical simulations. on January 15, 2022. Pure Appl Geophys 179:1549–1560. https:// doi. org/ Earth Sci Rev 107:66–75. https:// doi. org/ 10. 1016/j. earsc irev. 2011. 03. 00410. 1007/ s00024- 022- 03058-0 Frohlich C, Hornbach MJ, Taylor FW, Shen C-C, Moala ’A, Morton AE, Kruger J Japan Times (2022) Japan sees meter-high waves and tsunami warning after (2009) Huge erratic boulders in Tonga deposited by a prehistoric tsunami. massive Tonga eruption. https:// www. japan times. co. jp/ news/ 2022/ 01/ Geology 37:131–134. https:// doi. org/ 10. 1130/ G2527 7A.116/ natio nal/ japan- tsuna mi- tonga- volca no/. Accessed 16 Jan 2022 Garcia-Fry M, Murao O, Bachri S, Moya LA (2022) Land-use microsimula- Jayavanth P, Takai M, Akau’ola S (2009) Disaster and emergency preparedness tion model for livelihood diversification after the 2010 Merapi volcano in Tonga. Southeast Asian J Trop Med Public Health 40(1):31–40 T homas et al. Geoenvironmental Disasters (2023) 10:4 Page 17 of 18 Jin S, Huang H, Yang Z (2014) Research on Tonga traditional houses adaptive www. nzher ald. co. nz/ nz/ volca nic- erupt ion- new- zeala nd- ready- to- help- to environment. In: Kao JCM, Sung W-P, Chen R (eds) Green building, tonga- tsuna mi- surge- destr oys- boats- closes- beach es- in- nz/ QNI5G QKFSC materials and civil engineering. CRC Press, pp 891–894. https:// doi. org/ K52KN HXAT5 VAA6DE/. Accessed 16 Jan 2022 10. 1201/ b17568 OCHA (2022a) Daily noon briefing highlights: Tonga. https:// www. unocha. org/ Johnston I (2015) Traditional warning signs of cyclones on remote islands story/ daily- noon- brie fing- highl ights- tonga. Accessed 19 Jan 2022a in Fiji and Tonga. Environ Hazards 14:210–223. https:// doi. org/ 10. 1080/ OCHA (2022b) Data and Resources. https:// data. humda ta. org/ datas et/ cod- ab- 17477 891. 2015. 10461 56 ton. Accessed 1 Feb 2022b Karim F, Nandasena NAK (2022) Eec ff tiveness of coral reefs for marine flood OCHA (2022c) HOTOSM Tonga Buildings (OpenStreetMap Export). https:// data. reduction. In: 2022 Advances in science and engineering technology humda ta. org/ datas et/ hotosm_ ton_ build ings. Accessed 19 Jan 2022c international conferences (ASET ). IEEE, Dubai, United Arab Emirates, pp OCHA (2022d) Kingdom of Tonga: population—Tongatapu, ’Eua, Ha’apai, 1–4. https:// doi. org/ 10. 1109/ ASET5 3988. 2022. 97348 97 Vava’u, Niuas (As of 17 January 2022d). https:// relie fweb. int/ map/ tonga/ Kulp S, Strauss BH (2016) Global DEM errors underpredict coastal vulnerability kingd om- tonga- popul ation- tonga tapu- eua- haapai- vavau- niuas- 17- janua to sea level rise and flooding. Front Earth Sci. https:// doi. org/ 10. 3389/ ry- 2022d. Accessed 23 Jan 2022d feart. 2016. 00036 Okal EA, Borrero J, Synolakis CE (2004) The earthquake and tsunami of 1865 Kunkel CM, Hallberg RW, Oppenheimer M (2006) Coral reefs reduce tsunami November 17: evidence for far-field tsunami hazard from Tonga. Geophys impact in model simulations. Geophys Res Lett 33:L23612. https:// doi. J Int 157:164–174. https:// doi. org/ 10. 1111/j. 1365- 246X. 2004. 02177.x org/ 10. 1029/ 2006G L0278 92 Okal EA, Fritz HM, Borrero JC (2010) The Samoa tsunami of 29 September 2009: Lavigne F, Morin J, Wassmer P, Weller O, Kula T, Maea AV, Kelfoun K, Mokadem F, field survey in Tonga and preliminary modeling. In: EGU general assembly Paris R, Malawani MN, Faral A, Benbakkar M, Saulnier-Copard S, Vidal CM, conference abstracts, p 7008 Tu’I’afitu T, Kitekei’aho F, Trautmann M, Gomez C (2021) Bridging legends Pleasance C (2022) Tonga reveals 50 FOOT tsunami destroyed ALL houses and science: field evidence of a large tsunami that affected the Kingdom on one island while just two are left on another in ’unprecedented of Tonga in the 15th century. Front Earth Sci 9:748755. https:// doi. org/ 10. disaster’ as death toll climbs to three and new pictures reveal extent of 3389/ feart. 2021. 748755 devastation. Mailonline and Reuters. https:// www. daily mail. co. uk/ news/ Lay T, Ammon CJ, Kanamori H, Rivera L, Koper KD, Hutko AR (2010) The 2009 artic le- 10413 485/ Tonga- erupt ion- Aid- planes- land- island- damage. html. Samoa-Tonga great earthquake triggered doublet. Nature 466:964–968. Accessed 19 Jan 2022 https:// doi. org/ 10. 1038/ natur e09214 Reardon GF, Oliver J (1983) The impact of cyclone Isaac on buildings on Tonga. LNC (2022) Après une éruption aux Tonga, des mouvements anormaux J Wind Eng Ind Aerodyn 14:67–78. https:// doi. org/ 10. 1016/ 0167- 6105(83) du niveau de la mer sont à craindre au pays. https:// www. lnc. nc/ artic 90011-9 le- direct/ faits- divers/ mer/ socie te/ nouve lle- caled onie/ apres- une- erupt Roger J, Dudon B, Krien Y, Zahibo N (2014) Discussion about tsunami interac- ion- aux- tonga- des- mouve ments- anorm aux- du- niveau- de- la- mer- sont-a- tion with fringing coral reef. In: Kontar YA, Santiago-Fandiño V, Takahashi crain dre- au- pays. Accessed 16 Jan 2022 T (eds) Tsunami events and lessons learned. Springer, Dordrecht, pp Lolohea SF (2016) Internal Migration in Tonga, 2001–2011: a review of migrant 161–176. https:// doi. org/ 10. 1007/ 978- 94- 007- 7269-4_8 flows and characteristics. MA Thesis (Social Sciences), University of Wai- Rogers G (1981) The evacuation of Niuafo’ ou, an outlier in the kingdom of kato, Hamilton, New Zealand. https:// hdl. handle. net/ 10289/ 10741 Tonga∗. J Pac Hist 16:149–163. https:// doi. org/ 10. 1080/ 00223 34810 85724 Loriot P, Di Salvo M (2008) Détermination d’un MOS et calcul d’une tache 21 urbaine à partir de la BD TOPO de l’IGN: étude expérimentale. CERTU. Sattler DN, Lousi U, Graham JM, Latu V, Johnson J, Helu SL (2020) Climate https:// hal- lara. archi ves- ouver tes. fr/ hal- 02150 554 change in Tonga: risk perception and behavioral adaptation. In: Leal Luis JF (2007) Mirone: a multi-purpose tool for exploring grid data. Comput Filho W (ed) Managing climate change adaptation in the Pacific region. Geosci 33:31–41. https:// doi. org/ 10. 1016/j. cageo. 2006. 05. 005 Springer, Cham, pp 273–292. https:// doi. org/ 10. 1007/ 978-3- 030- 40552- Lynett P, McCann M, Zhou Z, Renteria W, Borrero J, Greer D, Fa’anunu O, 6_ 14 Bosserelle C, Jaffe B, La Selle S, Ritchie A, Snyder A, Nasr B, Bott J, Graehl Scott DA, Browne C (1989) Tonga. In: Scott DA, Browne C (eds) Economic N, Synolakis C, Ebrahimi B, Cinar GE (2022) Diverse tsunami genesis trig- development in seven Pacific Island countries. International Monetary gered by the Hunga Tonga-Hunga Ha’apai eruption. Nature 609:728–733. Fund, pp 135–156. https:// doi. org/ 10. 5089/ 97815 57750 358. 071 https:// doi. org/ 10. 1038/ s41586- 022- 05170-6 Searcy C, Dean K, Stringer W (1998) PUFF: a high-resolution volcanic ash track- Magee AD, Verdon-Kidd DC, Kiem AS, Royle SA (2016) Tropical cyclone ing model. J Volcanol Geotherm Res 80:1–16. https:// doi. org/ 10. 1016/ perceptions, impacts and adaptation in the Southwest Pacific: an urban S0377- 0273(97) 00037-1 perspective from Fiji, Vanuatu and Tonga. Nat Hazards Earth Syst Sci Simpson A, Johnson RW, Cummins P (2011) Volcanic threat in developing 16:1091–1105. https:// doi. org/ 10. 5194/ nhess- 16- 1091- 2016 countries of the Asia-Pacific region: probabilistic hazard assessment, Meredith ES, Jenkins SF, Hayes JL, Deligne NI, Lallemant D, Patrick M, Neal C population risks, and information gaps. Nat Hazards 57:151–165. https:// (2022) Damage assessment for the 2018 lower East Rift Zone lava flows doi. org/ 10. 1007/ s11069- 010- 9601-y of Kīlauea volcano. Hawaiʻi Bull Volcanol 84:65. https:// doi. org/ 10. 1007/ SRTM (2015) NASA Shuttle Radar Topography Mission (SRTM) Version 3.0 s00445- 022- 01568-2 Global 1 arc second data released over Asia and Australia. https:// www. Mimura N (1999) Vulnerability of island countries in the South Pacific to sea earth data. nasa. gov/ news/ nasa- shutt le- radar- topog raphy- missi on- srtm- level rise and climate change. Clim Res 12(2–3):137–143versi on-3- 0- global- 1- arc- second- data- relea sed- over- asia- and- austr alia. Monismith SG, Rogers JS, Koweek D, Dunbar RB (2015) Frictional wave dissipa- Accessed 26 Jan 2022 tion on a remarkably rough reef. Geophys Res Lett 42:4063–4071. https:// Sykes WR (1981) The vegetation of late, Tonga. Allertonia, pp 323–353 doi. org/ 10. 1002/ 2015G L0638 04 Tang L, Titov VV, Wei Y, Mofjeld HO, Spillane M, Arcas D, Bernard EN, Chamber - Mukul M, Srivastava V, Jade S, Mukul M (2017) Uncertainties in the Shuttle lin C, Gica E, Newman J (2008) Tsunami forecast analysis for the May 2006 Radar Topography Mission (SRTM) heights: insights from the Indian Hima- Tonga tsunami. J Geophys Res. https:// doi. org/ 10. 1029/ 2008J C0049 22 laya and Peninsula. Sci Rep 7:41672. https:// doi. org/ 10. 1038/ srep4 1672 Taylor PW (1991) The geology and petrology of Niuafo’ou Island, Tonga: suba- Nunn PD, Campbell JR (2020) Rediscovering the past to negotiate the future: erial volcanism in an active back-arc basin. Macquarie University. M.Sc. How knowledge about settlement history on high tropical Pacific Islands Thesis. https:// doi. org/ 10. 25949/ 19440 851. V1 might facilitate future relocations. Environ Dev 35:100546. https:// doi. Terry JP, Goff J, Winspear N, Bongolan VP, Fisher S (2022) Tonga volcanic org/ 10. 1016/j. envdev. 2020. 100546 eruption and tsunami, January 2022: globally the most significant Nunn PD, McNamara KE (2019) Failing adaptation in island contexts: the grow- opportunity to observe an explosive and tsunamigenic submarine erup- ing need for transformational change. In: Klöck C, Fink M (eds) Dealing tion since AD 1883 Krakatau. Geosci Lett 9:24. https:// doi. org/ 10. 1186/ with climate change on small islands: towards effective and sustainable s40562- 022- 00232-z adaptation. Göttingen University Press, Göttingen, pp 19–44. https:// doi. Thomas BEO, Roger J, Gunnell Y, Sabinot C, Aucan J (2021) A low-cost toolbox org/ 10. 17875/ gup20 19- 1210 for high-resolution vulnerability and hazard-perception mapping in view NZ Herald (2022) Volcanic eruption: New Zealand ready to help Tonga; of tsunami risk mitigation: application to New Caledonia. Int J Disaster tsunami surge destroys boats, closes beaches in NZ. NZ Herald. https:// Risk Reduct 62:102350. https:// doi. org/ 10. 1016/j. ijdrr. 2021. 102350 Thomas et al. Geoenvironmental Disasters (2023) 10:4 Page 18 of 18 TSD (2017) Tonga 2016 Census of population and housing Volume 1 Basic tables and administrative reports Second edition. Tonga Statistics Depart- ment, https:// tonga stats. gov. to/ census/ popul ation- stati stics/# 22- 60- wpfd- 2016- 14987 07296. Accessed 23 Jan 2022 TSD (2019) Tonga Census 2016 Census of population and housing Volume 2 Analytical report. Tonga Statistics Department, https:// tonga stats. gov. to/ census/ popul ation- stati stics/# 22- 60- wpfd- 2016- 14987 07296. Accessed 23 Jan 2022 UNISDR (2005) Hyogo framework for action 2005–2015: building the resilience of nations and communities to disasters UNISDR (2009) Reducing disaster risks through science: issues and actions UNOSAT (2022) 15 January 2022 volcanic eruption and tsunami. Preliminary satellite-based comprehensive damage assessment report. Tongatapu, Eua, and Ha’apai divisions of the Kingdom of Tonga. United Nations Satel- lite Centre (UNOSAT ) report. https:// relie fweb. int/ sites/ relie fweb. int/ files/ resou rces/ UNOSAT_ Compr ehens iveDa mageA ssess ment_ Tonga_ 20220 204. pdf. Accessed 23 Jan 2022 UNOSAT Maps (2022) Damage assessment in Tonga islands as of 20 January 2022. https:// exper ience. arcgis. com/ exper ience/ 5be5d 28e6f e740a e9ea8 d940f 6ef11 2b/? draft= true. Accessed 23 Jan 2022 USGS HDDS (2022) USGS Hazards Data Distribution System. https:// hddse xplor er. usgs. gov/. Accessed 29 June 2022 Vainikolo LC (2021) Small and Isolated Vava’u, Tonga: from weakness to strength during COVID-19? In: Campbell Y, Connell J (eds) COVID in the Islands: a comparative perspective on the Caribbean and the Pacific. Springer, Singapore, pp 207–218. https:// doi. org/ 10. 1007/ 978- 981- 16- 5285-1_ 11 Webley P, Mastin L (2009) Improved prediction and tracking of volcanic ash clouds. J Volcanol Geotherm 186:1–9. https:// doi. org/ 10. 1016/j. jvolg eores. 2008. 10. 022 Wessel P, Luis JF, Uieda L, Scharroo R, Wobbe F, Smith WHF, Tian D (2019) The Generic Mapping Tools version 6. Geochem Geophys Geosyst 20:5556–5564. https:// doi. org/ 10. 1029/ 2019G C0085 15 WHO (2009) Tonga: tsunami situation report 1. World Health Organization. Published 2 October 2009. https:// relie fweb. int/ report/ tonga/ tonga- tsuna mi- situa tion- report-1. Accessed 29 March 2022 Wilson KJ, Power WL, Nishimura Y, ‘Atelea Kautoke R, Vaiomo’unga R, Mori H, Pongi ‘A, Fifita M, Vaoahi M, Teukava S (2009) Post-tsunami survey of Niuatoputapu Island, Tonga, following the 30th September 2009, South Pacific tsunami, GNS Science Report 2009/71. http:// itic. ioc- unesco. org/ images/ docs/ SR- 2009- 71. pdf. Accessed 29 March 2022 World Bank (2022) Rebuilding tsunami-affected homes in remote islands of Tonga. The World Bank. Published 10 April 2014. https:// www. world bank. org/ en/ resul ts/ 2014/ 04/ 10/ rebui lding- tsuna mi- affec ted- homes- in- remote- islan ds- of- tonga. Accessed 29 March 2022 World Risk Report (2018) https:// weltr isiko beric ht. de/ wp- conte nt/ uploa ds/ 2019/ 03/ 190318_ WRR_ 2018_ EN_ RZonl ine_1. pdf. Accessed 13 Feb 2022 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations.
Journal
Geoenvironmental Disasters – Springer Journals
Published: Feb 16, 2023
Keywords: Kingdom of Tonga; Tsunami; Impact assessment; Population inventory; GIS; Dasymetric mapping