Ground motion prediction equations (GMPEs) relate ground motion intensitymeasures to variables describing earthquake source, path, and site effects. Frommany available GMPEs, we select those models recommended for use in seismichazard assessments in the Global Earthquake Model. We present a GMPE selec-tion procedure that evaluates multidimensional ground motion trends (e.g., withrespect to magnitude, distance, and structural period), examines functional forms,and evaluates published quantitative tests of GMPE performance against inde-pendent data. Our recommendations include: four models, based principallyon simulations, for stable continental regions; three empirical models for interfaceand in-slab subduction zone events; and three empirical models for active shallowcrustal regions. To approximately incorporate epistemic uncertainties, the selec-tion process accounts for alternate representations of key GMPE attributes, suchas the rate of distance attenuation, which are defensible from available data.Recommended models for each domain will change over time as additionalGMPEs are developed. [DOI: 10.1193/013013EQS017M]

Initiated by the OECD’s Global Science Forum, GEM was formed in 2009 as a non-profit foundation in Pavia, Italy, funded through a public-private sponsorship with the vision to create a world that is resilient to earthquakes. GEM’s mission is to become one of the world’s most complete sources of risk resources and a globally accepted standard for seismic risk assessment, whose products are applied in risk management worldwide.

Macroseismic shaking intensity is a fundamental parameter for the development, calibration, and use in a variety of hazard maps as well as in empirical (direct) and semi-empirical (indirect) earthquake shaking loss methodologies. Macroseismic data also quantify damage from past and present events and facilitate communicating ground motion levels in terms of human experiences and incurred losses. The aim of this report is to summarize and recommend “best practices” for the use of macroseismic intensity in conjunction with hazard maps (particularly ShakeMaps) and as input to associated loss models. The continued reliance on macroseismic intensity data dictates that ground motion prediction equations (GMPEs) alone are not always sufficient for estimating or constraining shaking hazards. Relations that allow direct estimation of intensity given an earthquake magnitude and distance, and those that convert ground motions to intensity (and vice versa) are required. Forward estimation of macroseismic intensities take two primary forms: 1) direct intensity prediction equations (IPEs), and 2) ground-motion-to-intensity conversion equations (GMICE). In addition, one can potentially better constrain historical ground motions at particular sites by employing intensity-to-ground-motion conversion equations (IGMCEs), though such equations are rare. Both the Global Earthquake Model (GEM) and Global ShakeMap (GSM) require advice and optimization in the state-of-the-art use of ground motion and intensity data. We provide background on the issues relating ground motions to intensities, directly predicting intensities, and offer insight into their uses. In the end, we recommend initial default IPE and GMICE selections for use in the immediate short term while additional research on these fronts continues and develops. A brief summary of highly related, current studies that help inform this report is also provided. Based on these ongoing analyses, and this report’s summary, we provide recommendations for further refinements in the form of continued research and development efforts.

The Global Earthquake Model (GEM) is a public/private partnership initiated and approved by the Global Science Forum of the Organisation for Economic Co-operation and Development (OECD-GSF). GEM aims to provide uniform, independent standards to calculate and communicate earthquake risk worldwide. With committed backing from academia, governments, and industry, GEM will contribute to improved modelling of earthquake risk worldwide. More information is available on the GEM website: www.globalquakemodel.org. As a first step in developing a global earthquake model, a focused pilot project named GEM1 was launched to generate GEM’s first products and develop GEM’s initial IT infrastructure. GEM1 formally started in January 2009 and ended on March 31st 2010, whilst ETH Zurich was appointed as the coordinator, with EUCENTRE (Italy), GFZ (Germany), NORSAR (Norway) and the USGS (USA) as contributing partners (and a number of other institutions and individuals, named in the companion reports, also provided models, data and feedback). The main objective of GEM1 was that it would provide a basis upon which the future development of the full GEM computing environment and product set could be built. The aim as spelled out in the GEM1 implementation plan was to largely use existing tools and datasets in hazard and risk, connected through a unified IT infrastructure. The GEM1 deliverables are therefore to be considered ‘proof-of-concept’ rather than final products, hence any resulting outputs should be conceived as conceptual only and are thus not suitable for application. This report briefly summarizes the achievements of GEM1.

This document provides an overview of the PSHA input models collected during GEM1, of the engine used to perform PSHA calculations, and the methods and criteria adopted for computing a proof-of-concept global hazard map. The GEM1 PSHA input repository contains seventeen national or regional models and one global model based on a smoothed seismicity approach. The oldest models were developed in the context of the GSHAP project, ended at the end of the 1990s; the most recent models are the ones prepared by the USGS-NSHM project for South America and a global smoothed seismicity model specifically produced for GEM1. In terms of geographical coverage the gathered models cover almost all the globe; the only missing regions are the Caribbean, the area around Papua-New Guinea and the Pacific Islands. These areas will hopefully be updated soon in the context of Regional Initiatives. In terms of information content, a relevant part of the PSHA input models is based on area sources while a minority uses fault sources. All the models but the Japanese and the model for the New Madrid Zone in the eastern US incorporate a time independent Poissonian model. Least but not last, epistemic uncertainties are taken into account by just some models, usually the most recent ones, and frequently treated as aleatory in the calculations.

The present report reviews a set of probabilistic seismic hazard analysis (PSHA) computer programs (CRISIS, EQRM, NSHMP, OpenSHA, and SEISHAZ) proposed as possible platforms for the development of the GEM1’s hazard engine. The analysis is supplemented by including a number of additional software (FRISK88M, MoCaHAZ, MRS, OHAZ, and SEISRISK IIIM) considered important to obtain a more comprehensive understanding of the current state-of-the-art in PSHA. The report is organized into two sections. The first describes the main properties of each PSHA program on the technical (e.g. the programming language) and scientific level (e.g. the PSHA source typologies supported). The second illustrates, for a subset of the selected software, a simple benchmarking exercise aimed at understanding the behaviour of the programs, and to compare the results provided for very simple cases. The review of the selected PSHA software proved to be a very useful exercise to delineate the desirable properties for the GEM1 seismic hazard engine and shows that OpenSHA can accommodate the GEM1 IT and hazard specifications better than the other evaluated software.

This report provides a summary of the GEM1 Risk activities; it is divided into two parts, the first relating to a study of existing seismic risk software and previous global risk initiatives, and the second concentrating on the development of a computational engine for global risk assessment, based on those lessons learned.

The aim of this document is to provide an overview of the design of the OpenGEM system, within the scope of the GEM1 project. OpenGEM is the name that was chosen for the IT platform which will allow calculation and communication of earthquake risk on a global scale. One of the goals of the GEM1 project was to design an initial model building structure and this report hence describes the various components of the design. The IT infrastructure described in this report has been reviewed during an intensive IT-review after which a strategy has been determined for further development of GEMs IT architecture, which makes use of the work done during GEM1, but has taken a different approach.

TheaimofthisreportistoprovideacomprehensivedocumentationoftheGED4GEMprojectanditsresults, the Global Exposure Database (or GED). Although most (if not all) of the information provided in this document is available in the deliverables of the GED4GEM project (http://www.nexus.globalquakemodel.org/ged4gem/posts/ged4gem-deliverables/), this document offers a unique possibility to look at the various scientific aspects of the project. It also attempts to provide clear answerstothemainquestionsthatmightbeposedbypotentialusersoftheGED,specificallyaboutthedata setsused,themethodologiesemployedandthecompletenessofthedatabase.

In any probabilistic seismic‐hazard model, the earthquake activity that cannot be associated with well‐characterized fault structures is taken into account as seismicity distributed over a geographical region. Ground‐motion prediction equations (GMPEs) are generally based on predictor variables describing the spatial extension of a rupture. The approach taken to model rupture finiteness can therefore bias the estimation of seismic hazard. We study the effect of rupture finiteness in modeling distributed seismicity using the OpenQuake (OQ) engine, the open‐source software for seismic hazard and risk assessment promoted by the Global Earthquake Model initiative. For a simple test case we show how the inclusion of rupture finiteness, with respect to the point‐rupture approximation, leads to a significant increase in the probabilities of exceedance for a given level of motion. We then compare the OQ engine with the calculation software developed by the U.S. Geological Survey‐National Seismic Hazard Mapping Project. By considering a gridded seismicity model for California, we show how different approaches for modeling finite ruptures affect seismic‐hazard estimates. We show how sensitivity to rupture finiteness depends not only on the spatial distribution of activity rates but also on the GMPE model. Considering two sites in Los Angeles and San Francisco, we show that for a return period of 475 years, the percent difference in the associated ground‐motion levels when using point and finite ruptures ranges from 19% to 46%; for a return period of 2475 years the difference ranges from 29% to 58%.