With a population of over 250 million people, Indonesia is the fourth most populous country in the world (United Nations, 2013). Indonesia also experiences more earthquakes than any other country in the world (USGS, 2015). Its borders encompass one of the most active tectonic regions on Earth including over 18 000 km of major tectonic plate boundary, more than twice that of Japan or Papua New Guinea (Bird, 2003). The potential for this tectonic activity to impact large populations has been tragically demonstrated by the 20004 Sumatra earthquake and tsunami. In order to inform earthquake risk reduction in Indonesia, a new national earthquake hazard map was developed in 2010 (Irsyam et al., 2010). In this report historical records of damaging earthquakes from the 17th to 19th centuries are used to test our current understanding of earthquake hazard in Indonesia and identify areas where further research is needed. In this report we address the following questions: - How well does our current understanding of earthquake hazard in Indonesia reflect historical activity? - Can we associate major historical earthquakes with known active faults, and are these accounted for in current assessments of earthquake hazard? - Does the current earthquake hazard map predict a frequency and intensity of shaking commensurate with the historical record? - What would the impact of these historical earthquakes be if they were to reoccur today? To help answer questions like these, this report collates historical observations of eight large earthquakes from Java, Bali and Nusa Tenggara between 1699 and 1867. These observations are then used to: - Identify plausible sources for each event; - Develop ground shaking models using the OpenQuake Engine (GEM Foundation, 2015); - Assess the validity of the current national seismic hazard map; and - Estimate fatalities were the historical events to occur today using the InaSAFE (InaSAFE.org, 2015) software.

Many earthquakes in Indonesia have caused a large number of fatalities. Disaster risk-reduction of fatalities requires a representative fatality model derived from fatality data caused by historical earthquakes in Indonesia. We develop an empirical fatality model for Indonesia by relating macroseismic intensity to fatality rate using compiled subdistrict level fatality rate data and numerically simulated ground shaking intensity for four recent damaging events. The fatality rate data are compiled by collecting population and fatality statistics of the regions impacted by the selected events. The ground shaking intensity is numerically estimated by incorporating a finite fault model of each event and local site conditions approximated by topographically-based site amplifications. The macroseismic intensity distribution of each event is generated by using ShakeMap software with a selected pair of ground motion predictive equation (GMPE) and ground motion to intensity conversion equation (GMICE). The developed fatality model is a Bayesian generalized linear model where the fatality rate is assumed to follow a mixture of a Bernoulli and a gamma distribution. The probability of zero fatality rate and the mean non-zero fatality rate is linked to a linear function of shaking intensity by the logit and the log link functions, respectively. We estimate posterior distribution of the parameters of the model based on the Hamilton Monte Carlo algorithm. For validation of the developed model we calculate fatalities of the past events from the EXPO-CAT catalog and compare the estimates with the EXPO-CAT fatality records. While the developed fatality model can provide an estimate of the range of fatalities for future events it needs on-going refinement by incorporation of additional fatality rate data from past and future events.

Probabilistic seismic hazard map of Papua New Guinea, in terms of Peak Ground Acceleration, is developed for return period of 475 years. The calculations were performed for bedrock site conditions (Vs30=760 m/s). Logic-tree framework is applied to include epistemic uncertainty in seismic source as well as ground-motion modelling processes. In this regard two source models, using area source zones and smoothed seismicity, are developed. Based on available geological and seismological data, defined seismic sources are classified into 4 different tectonic environments. For each of the tectonic regimes three Ground Motion Prediction Equations are selected and used to estimate the ground motions at a grid of sites with spacing of 0.1 degree in latitude and longitude. Results show high level of hazard in the coastal areas of Huon Peninsula and New Britain/ Bougainville regions and relatively low level of hazard in the southern part of the New Guinea highlands block. In Huon Peninsula, as shown by seismic hazard disaggregation results, high level of hazard is caused by modelled frequent moderate to large earthquakes occurring at Ramu-Markham Fault zone. On the other hand in New Britain/Bougainville region, the geometry and distance to the subduction zone along New Britain Trench mainly controls the calculated level of hazard. It is also shown that estimated level of PGAs is very sensitive to the selection of GMPEs and overall the results are closer to the results from studies using more recent ground-motion models.

Papua New Guinea (PNG) lies in a belt of intense tectonic activity that experiences high levels of seismicity. Although this seismicity poses significant risks to society, the Building Code of PNG and its underpinning seismic loading requirements have not been revised since 1982. This study aims to partially address this gap by updating the seismic zoning map on which the earthquake loading component of the building code is based. We performed a new probabilistic seismic hazard assessment for PNG using the OpenQuake software developed by the Global Earthquake Model Foundation (Pagani et al. 2014). Among other enhancements, for the first time together with background sources, individual fault sources are implemented to represent active major and microplate boundaries in the region to better constrain the earthquake-rate and seismic-source models. The seismic-source model also models intraslab, Wadati–Benioff zone seismicity in a more realistic way using a continuous slab volume to constrain the finite ruptures of such events. The results suggest a high level of hazard in the coastal areas of the Huon Peninsula and the New Britain – Bougainville region, and a relatively low level of hazard in the southwestern part of mainland PNG. In comparison with the seismic zonation map in the current design standard, it can be noted that the spatial distribution of seismic hazard used for building design does not match the bedrock hazard distribution of this study. In particular, the high seismic hazard of the Huon Peninsula in the revised assessment is not captured in the current building code of PNG. <b>Citation:</b> Ghasemi, H., Cummins, P., Weatherill, G. <i>et al.</i> Seismotectonic model and probabilistic seismic hazard assessment for Papua New Guinea. <i>Bull Earthquake Eng, </i><b>18</b>, 6571–6605 (2020). https://doi.org/10.1007/s10518-020-00966-1

On 6th July 2006, an intense swarm of earthquake activity began in the Sulu Range, Central New Britain, Papua New Guinea. The earthquakes were felt almost every one to two minutes, 24 hours a day, with modified Mercalli intensities of MM1 to MM4. They were accompanied by unusual vigorous activity in the hot springs southwest of the Sulu Range. Fearing a possible eruption and tsunami, about 1000 locals were evacuated.

Through Australian Department of Foreign Affairs and Trade, Geoscience Australia has been working closely with the Government of Papua New Guinea technical agencies (Rabaul Volcano Observatory, Port Moresby Geophysical Observatory, and Engineering Geology Branch) since September 2010 to enhance their capabilities to monitor and assess natural hazards. The objective of this program is to support the Government of Papua New Guinea in developing fundamental information and practices for the effective response and management of natural hazard events in PNG. Earthquakes as natural hazards are one of the key focus points of this project, as they continue to cause loss of life and widespread damage to buildings and infrastructure in Papua New Guinea. The country’s vulnerability to earthquakes is evident from the significant socio-economic consequences of recent major events in Papua New Guinea, e.g., a magnitude 7.5 earthquake that occurred in the Hela Province of Papua New Guinea in 2018. Earthquake risk is likely to increase significantly in the years to come due to the growth in population and urbanization in Papua New Guinea. However, earthquake risk, unlike hazard, can be managed and minimized. One obvious example would be minimizing earthquake risk by constructing earthquake-resistant structures following building standards. The high level of earthquake hazard of Papua New Guinea has been long recognised and the suite of building standards released in 1982 contained provisions to impart adequate resilience to buildings based on the best understanding of seismic hazard available at that time. However, the building standards and incorporated seismic hazard assessment for Papua New Guinea has not been updated since the 1980s. The integration of modern national seismic hazard models into national building codes and practices provides the most effective way that we can reduce human casualties and economic losses from future earthquakes. This report aims at partially fulfilling this task by performing a probabilistic seismic hazard assessment to underpin a revision of the earthquake loading component of the building standards of Papua New Guinea. The updated assessment offers many important advances over its predecessor. It is based on a modern probabilistic hazard framework and considers an earthquake catalogue augmented with an additional four decades-worth of data. The revised assessment considers advances in ground-motion modelling through the use of multiple ground-motion models. Also, for the first time, the individual fault sources representing active major and microplate boundaries are implemented in the input hazard model. Furthermore, the intraslab sources are represented realistically by using the continuous slab volume to constrain the finite ruptures of such events. This would better constrain the expected levels of ground motion at any given site in Papua New Guinea. The results suggest a high level of hazard in the coastal areas of the Huon Peninsula and the New Britain–Bougainville region, and a relatively low level of hazard in the southern part of the New Guinea Highlands Block. In comparison with the seismic zonation map in the current design standard, it can be noted that the spatial distribution used for building design does not match the bedrock hazard distribution of this study. In particular, the high seismic hazard of the Huon Peninsula in the revised assessment is not captured in the current seismic zoning map, leading to a significant under-estimation of hazard in PNG’s second-largest city, Lae. It can also be shown that in many other regions and community localities in PNG the hazard is higher than that regulated for the design of buildings having a range of natural periods. Thus, the need for an updated hazard map for building design has been confirmed from the results of this study, and a revised map is developed for consideration in a revised building standard of Papua New Guinea.

The Greater Metro Manila Area is one of the world's megacities and is home to about 12 million people. It is located in a region at risk from earthquakes, volcanic eruptions, tropical cyclones, riverine flooding, landslides and other natural hazards. Major flooding affected the Greater Metro Manila Area in September 2009 following the passage of Typhoon Ketsana (known locally as Typhoon Ondoy). Following this event, the Australian Aid Program supported Geoscience Australia to undertake a capacity building project with its partner agencies in the Government of the Philippines. The output of this project has been a series of risk information products developed by agencies in the Collective Strengthening of Community Awareness for Natural Disasters (CSCAND) group. These products quantify the expected physical damage and economic loss to buildings caused by earthquakes, tropical cyclone severe wind and riverine flooding across the Greater Metro Manila Area. Spatial data is a key input to the development of hazard models and information on exposure, or the 'elements at risk'. The development of a spatially enabled exposure database was a crucial element in the construction of risk information products for the Greater Metro Manila Area. The database provides one central repository to host consistent information about the location, size, type, age, residential population and structural characteristics of buildings within the area of interest. Unique spatial analysis techniques were employed to quantify and record important aspects of the built environment, for inclusion in the database. The process of exposure data development within the Greater Metro Manila Area, including a new application developed by Geoscience Australia for estimating the geometric characteristics of buildings from high resolution elevation data and multi-spectral imagery, will be presented.

The Philippine Institute of Volcanology and Seismology (PHIVOLCS) and Geoscience Australia (GA) have developed a long-term partnership in order to better understand and reduce the risks associated with earthquake hazards in the Philippines. The Project discussed herein was supported by the Australian Agency for International Development (AusAID). Specifically, this partnership was designed to enhance the exposure and damage estimation capabilities of the Rapid Earthquake Damage Assessment System (REDAS), which has been designed and built by PHIVOLCS. Prior to the commencement of this Project, REDAS had the capability to model a range of potential earthquake hazards including ground shaking, tsunami inundation, liquefaction and landslides, as well as providing information about elements at risk (e.g., schools, bridges, etc.) from the aforementioned hazards. The current Project enhances the exposure and vulnerability modules in REDAS and enable it to estimate building damage and fatalities resulting from scenario earthquakes, and to provide critical information to first-responders on the likely impacts of an earthquake in near real-time. To investigate this emergent capability within PHIVOLCS, we have chosen the pilot community of Iloilo City, Western Visayas. A large component of this project has been the compilation of datasets to develop building exposure models, and subsequently, developing methodologies to make these datasets useful for natural hazard impact assessments. Collection of the exposure data was undertaken at two levels: national and local. The national exposure dataset was gathered from the Philippines National Statistics Office (NSO) and comprises basic information on wall type, roof type, and floor area for residential buildings. The NSO census dataset also comprises crucial information on the population distribution throughout the Philippines. The local exposure dataset gathered from the Iloilo City Assessors Office includes slightly more detailed information on the building type for all buildings (residential, commercial, government, etc.) and appears to provide more accurate information on the floor area. However, the local Iloilo City dataset does not provide any information on the number of people that occupy these buildings. Consequently, in order for the local data to be useful for our purposes, we must merge the population data from the NSO with the local Assessors Office data. Subsequent validation if the Iloilo City exposure database has been conducted through targeted foot-based building inventory surveys and has allowed us to generate statistical models to approximate the distribution of engineering structural systems aggregated at a barangay level using simple wall and roof-type information from the NSO census data. We present a comparison of the national and local exposure data and discuss how information assembled from the Iloilo City pilot study - and future study areas where detailed exposure assessments are conducted - could be extended to describe the distribution of building stock in other regions of the Philippines using only the first-order national-scale NSO data. We present exposure information gathered for Iloilo City at barangay level in a format that can be readily imported to REDAS for estimating earthquake impact.

In June 2012 Geoscience Australia was commissioned by Commonwealth Scientific and Industrial Research Organisation (CSIRO) to undertake detailed wind hazard assessments for 14 Pacific Island countries and East Timor as part of the Pacific-Australia Climate Change Science and Adaptation Planning (PACCSAP) program. PACCSAP program follows on from work Geoscience Australia did for the Pacific Climate Change Science Program (PCCSP) looking at CMIP3 generation of climate models. The objective of this study is to improve scientific knowledge by examining past climate trends and variability to provide regional and national climate projections. This document presents results from current and future climate projections of severe wind hazard from tropical cyclones for the 15 PACCSAP partner countries describing the data and methods used for the analysis. The severe wind hazard was estimated for current (1981 to 2000) and future (2081 to 2100) climate scenarios. Tropical-cyclone like vortices from climate simulations conducted by CSIRO using six Coupled Model Intercomparison Project phase 5 (CMIP5) models (BCC-CSM1.1, NorESM1-M, CSIRO-Mk3.6, IPSL-CM5A, MRI-CGM3 and GFDL-ESM2M) as well as the International Best Track Archive for Climate Stewardship were used as input to the Geoscience Australia's Tropical Cyclone Risk Model to generate return period wind speeds for the 15 PACCSAP partner countries. The Tropical Cyclone Risk Model is a statistical-parametric model of tropical cyclone behaviour, enabling users to generate synthetic records of tropical cyclones representing many thousands of years of activity. The 500-year return period wind speed is analysed and discussed into more details in this report, since it is used as a benchmark for the design loads on residential buildings. Results indicate that there is not a consistent spatial trend for the changes in 500-year cyclonic wind speed return period when CMIP5 models are compared individually. BCC-CSM1M and IPSL-CM5A presented an increase in the annual TC frequency for East Timor, northern hemisphere and southern hemisphere. On the other hand, NorESM1M showed a decrease in the annual TC frequency for the same areas. The other three models showed a mixed of increase and decrease in their annual TC frequency. When CMIP5 models were analysed by partner county capitals for the 500-year cyclonic wind speed return period, IPSL-CM5A and GFDL-ESM2M models presented an increase in the cyclonic wind speed intensity for almost all capitals analysed with exception of Funafuti (GFDL-ESM2M), which presented a decrease of 0.7% and Honiara (IPSL-CM5A) with a decrease of 1.6%. The tropical cyclone annual frequency ensemble mean indicates an increase in the tropical cyclone frequency within all three regions considered in this study. When looking at individual capitals, a slight increase in the 500-year return period cyclonic wind speed ensemble mean varying between 0.8% (Port Vila) to 9.1% (Majuro) is noticed. A decline around 2.4% on average in the 500-year return period cyclonic wind speed ensemble mean is observed in Dili, Suva, Nukualofa and Ngerulmud. The ensemble spatial relative change did not show any particular consistency for the 500-year cyclonic wind speed. Areas where Marshall Islands and Niue are located presented an increase in the 500-year cyclonic wind speed while a decrease is observed in areas around South of Vanuatu, East of Solomon Islands, South of Fiji and some areas in Tonga. The information from the evaluation of severe wind hazard from tropical cyclones, together with other PACCSAP program outputs, will be used to build partner country capacity to effectively adapt and plan for the future and overcome challenges from climate change.

Probabilistic earthquake hazard maps were prepared for the Fiji Islands. Damage has been caused by Fiji earthquakes around 1850, in 1884, 1902, 1919, 1932 (twice), 1953 and 1979. No previous assessment had produced a comprehensive description of the earthquake hazard in Fiji and the present study was initiated in 1990 when the author was attached to the Mineral Resources Department, Fiji. Collection and analysis of data continued at MRD until 1992 and the study was completed at the Australian Geological Survey Organisation in 1993-1997. The aim of the study was to produce probabilistic earthquake hazard maps which can be used in the National Building Code for Fiji, for design of special structures, for planning, for emergency management and for risk management. Few, if any, similar studies have been undertaken in the seismically active Southwest Pacific.