Goded, T.; Moratalla, J.M.; Canessa, S.; Huso, R.; Coomer, M.A.; Hanson, J.; Potter, S.H. 2021 Can GeoNet’s Felt RAPID reports be reliably used to produce rapid earthquake intensity maps?. Lower Hutt, N.Z.: GNS Science. GNS Science report 2021/17. 143 p.; doi: 10.21420/EQGT-EN57
Abstract:
During the response to a moderate-to-large earthquake in New Zealand, there is a great need for information on the geographical distribution of the damage. Emergency responders and decision makers use this information to prioritise mitigation actions, such as evacuation procedures and building interventions. At present, GeoNet produces only maps with the strong-motion data (in Peak Ground Velocity, Peak Ground Acceleration or Modified Mercalli intensity [MMI]) from the last hour. In addition, for each earthquake, currently GeoNet produces only a map with the locations of highest MMI per mesh block. However, there are currently no available maps with a geographical distribution of the damage per earthquake. GeoNet is working on producing these shaking layer maps per event to be used by emergency responders and the public. The potential use of GeoNet’s ‘Felt RAPID’ reports to produce shaking maps and complement strong-motion data is being analysed in this project. ‘Felt RAPID’ is a database of crowdsourced data where the public chooses from a set of six cartoons depicting their experience of the earthquake. This is a large dataset that produces a great amount of data within minutes ofthe earthquake, with its current record being more than 52,000 reports submitted for one single event within a few minutes after the event. Using our experience with a recent method to group the more robust ‘Felt Detailed’ reports from GeoNet, we will explore the possibility of usingFelt RAPID data within the first 24–48 hours after a damaging earthquake. The felt reports have the potential to ‘fill in the gaps’ far away from strong-motion stations, where otherwise results fully depend on attenuation models and thus increase their uncertainty. Results from this project show that intensities derived from Felt RAPID reports are consistent with those derived from Felt Detailed, as long as enough reports are being submitted for each community. A minimum of 10 felt reports, and ideally more, is needed to produce stable results. However, analysis using the same respondent filling in both Felt RAPID and Felt Detailed questionnaires indicates a tendency for Felt RAPID to overestimate low intensities (MMI2–4) and underestimate high intensities (MMI>=5) compared to Felt Detailed. This project has been developed in parallel to GeoNet’s ‘Shaking Layers’ project, which has started this year, with the aim of developing automatised shaking layer maps within minutes of a moderate-to-large earthquake in New Zealand. Data from the stakeholder engagements carried out during this project will be used in GeoNet’s ‘Shaking Layers’ project to develop a product that fulfils the end users’ needs. Until now, the Felt RAPID database has been used only for public communication purposes. Its use in emergency response has not been analysed in detail.The project is largely focused on stakeholder engagements to understand the end-user needs for hazard information following a damaging earthquake. We carried out empathy interviews with key stakeholders, as well as two half-day workshops (one in person and one online) with a wider community of end users. During these engagements, we gathered key information on end-users’ needs for hazard information following a damaging earthquake in New Zealand. The workshops caused great interest and positive feedback and showed the willingness for participants to keep informed and participate in the two projects. Results from these stakeholder engagements are provided in this report. The ‘Shaking Layers’ project from GeoNet will continue engaging with stakeholders (aut
