Consistent MMI area estimation for Australian earthquakes

Geoscience Australia has been providing estimates of felt and potential damage radiuses for all earthquakes above magnitude 3.0 since 2002. Similarly, over the last decade, using the hazard modelling software EQRM, GA has produced scenario MMI maps for most Australian cities and several cities in our region. The former uses Empirical relations developed from measuring MMI III, IV, V and VI radiuses from the isoseismal map of ~150 Australian earthquakes. The later using various GMPEs to generate the hazard field and PGA/PGV to MMI conversions to estimate MMI. These two approaches have not previously been directly compared.We have tested the fit with between the empirical MMI areas and the scenario models for several combinations of GMPEs and PGA/PGV to MMI conversions. We also investigate the possible importance of site effects in biasing the empirical data, for which only the minimum and maximum epicentral distance was measured, particularly at low MMI. A scenario model that more accurately reflects the empirical data should be more robust than the current method, for both real-time earthquake advice and scenarios. It should also enable the conversions used to estimate the magnitude of pre-instrumental earthquakes to be tested. Additionally the GMPEs that give the best fit to the empirical data might provide guidance when selecting GMPES for PSHAs and future scenario products.自2002年起，澳大利亚地球科学局（Geoscience Australia，以下简称GA）便一直为所有震级3.0级以上的地震提供有感范围与潜在破坏范围的估算值。同理，在过去十年间，GA借助灾害建模软件EQRM，为澳大利亚多数城市及本区域部分城市绘制了情景式麦卡利地震烈度（Modified Mercalli Intensity, MMI）分布图。前者依托从约150次澳大利亚地震的等震线图中测得的MMI Ⅲ、Ⅳ、Ⅴ、Ⅵ级半径所推导得到的经验关系开展计算；后者则通过多种地面运动预测方程（Ground Motion Prediction Equations, GMPEs）生成灾害场，并借助峰值地面加速度（Peak Ground Acceleration, PGA）与峰值地面速度（Peak Ground Velocity, PGV）到MMI的转换关系来估算MMI烈度。此前尚未对这两种方法开展过直接对比。我们针对多种GMPE与PGA/PGV到MMI的转换组合，检验了经验式MMI影响区域与情景模型之间的拟合程度。我们还探究了场地效应对经验数据的偏倚影响——这类经验数据仅记录了震中距的最小值与最大值，在低烈度等级下该问题尤为突出。相较于当前方法，能更精准匹配经验数据的情景模型，在地震实时预警与情景模拟两类场景中都将具备更强的鲁棒性。该模型还可用于检验用于估算无仪器记录时期地震震级的转换关系。此外，与经验数据拟合度最优的GMPEs，可为概率性地震危险性分析（Probabilistic Seismic Hazard Analysis, PSHA）及未来情景类产品的GMPE选择提供参考依据。