When an earthquake affects an inhabited area, a need for information immediately arises among the population. In general, this need is not immediately fulfilled by official channels which usually release expert‐validated information with delays of many minutes. Seismology is among the research fields where citizen science projects succeeded in collecting useful scientific information. More recently, the ubiquity of smartphones is giving the opportunity to involve even more citizens. This paper focuses on seismic intensity reports collected through smartphone applications while an earthquake is occurring. The aim is to provide a framework for predicting and updating in near realtime earthquake parameters that are useful for assessing the effect of the earthquake. This is done by using a multivariate space–time model based on time‐varying coefficients and a spatial latent variable. As a case‐study, the model is applied to more than 200000 seismic reports globally collected over a period of around 4 years by the Earthquake Network citizen science project. It is shown how the time‐varying coefficients are needed to adapt the model to an information content that changes with time, and how the spatial latent variable can capture the local seismicity and the heterogeneity in the people's response across the globe.

中文翻译：

---

满足地震后的信息需求：公民科学地震报告的统计模型，用于近实时预测地震参数

当地震影响到居民区时，居民中立即需要信息。通常，官方渠道通常无法立即发布经过专家验证的信息，而延迟了很多分钟，因此并不能立即满足这种需求。地震学是公民科学项目成功收集有用科学信息的研究领域之一。最近，智能手机的普及为人们提供了参与更多机会的机会。本文着重于地震发生时通过智能手机应用程序收集的地震烈度报告。目的是提供一种用于预测和更新近实时地震参数的框架，这些参数可用于评估地震的影响。这是通过使用基于时变系数和空间潜在变量的多元时空模型来完成的。作为案例研究，该模型适用于地震网络公民科学项目在大约4年的时间内在全球范围内收集的200000多份地震报告。它显示了如何需要时变系数来使模型适应随时间变化的信息内容，以及空间潜变量如何捕获全球地震响应中的局部地震活动和异质性。