| DC 欄位 |
值 |
語言 |
| DC.contributor | 統計研究所 | zh_TW |
| DC.creator | 謝侑霖 | zh_TW |
| DC.creator | Yu-Lin Hsieh | en_US |
| dc.date.accessioned | 2020-6-30T07:39:07Z | |
| dc.date.available | 2020-6-30T07:39:07Z | |
| dc.date.issued | 2020 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:444/thesis/view_etd.asp?URN=107225019 | |
| dc.contributor.department | 統計研究所 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 強烈主震後經常引起大量餘震，短時間內發生的大規模餘震可能會破壞震區已經弱化的建物或結構，因此餘震風險評估對於主震後短時間內的救援行動極為重要。本文分別利用混合二元常態分布與混合迦馬結合條件常態分布描述餘震的空間分布，並且結合餘震規模時間風險模型(Reasenberg and Jones, 1898, 1994)，建立餘震規模時空風險模型，分別記為BNSRJ與GNSRJ模型，分析2019年規模7.1的美國加州里奇克萊斯特地震之餘震風險。根據在里奇克萊斯特主震後一段時間內的完整餘震，應用BNSRJ與GNSRJ模型或是RJ模型結合點格法(GRJ)計算研究中點格的餘震相對風險，建立相對餘震風險圖用於預警未來等長時間強餘震發生的可能區域。最後藉由接收者操作特徵(ROC)曲線相關的準則評估上述相對餘震風險圖在預警未來大餘震的表現。根據ROC曲線下面積(AUC)評估，應用BNSRJ及GNSRJ模型所建立的相對餘震風險圖在主震後3及6小時預警規模4.0以上餘震的效果皆優於GRJ，在12小時預警未來規模4.0以上餘震區域時，則以GNSRJ最優。但是，在主震後24小時因為餘震數量多，研究中的三種相對餘震風險圖之餘震效果無差異。最後，Youden 指數建議GNSRJ的相對風險門檻值設在0.05與0.15之間，才能在里奇克萊斯特主震後的3~12小時內有效預警未來規模4.0以上餘震區域。 | zh_TW |
| dc.description.abstract | A strong earthquake often brings up a lot of aftershocks. Large aftershocks that occur in a short time after the main shock may damage the weakened buildings or structures. Therefore, assessment of aftershock hazard is extremely important for rescue work within a short time after the main shock. In this article, we use the mixture bivariate normal distribution and the mixture gamma along with conditional normal distribution, respectively, to describe the spatial distribution of aftershocks. By combining the spatial distribution with the magnitude-time hazard model of aftershocks (Reasenberg and Jones, 1898, 1994), the two space-time-magnitude hazard models of aftershocks are then obtained, denoted as BNSRJ and GNSRJ models, respectively. The BNSRJ or GNSRJ are then applied to analyze the hazard of aftershocks of the 2019 Ridgecrest earthquake in the United States. Based on the completely recorded aftershocks in a short period after the main shock, relative hazard of aftershocks in each period under study is calculated by using the BNSRJ, GNSRJ or RJ models with the gridding method (GRJ). The relative aftershock hazard (RAH) maps are then obtained for depicting possible areas of future large aftershocks. Finally, the performance of RAH maps for alarming large aftershocks in a short time after the main shock is evaluated based on criterion related to the receiver operating characteristic (ROC) curve. Based on the area under the ROC curve (AUC), RAH maps based on the BNSRJ and GNSRJ models are better than the GRJ map for alarming 4.0 or larger aftershocks in 3 to 6 hours after the main shock. The GNSRJ map is superior to the other two when alarming such aftershocks in 12 hours after the main shock. However, the three RAH maps are of no difference when alarming future aftershocks in one day after the main shock. Finally, Youden’s index suggests threshold between 0.05 and 0.15 for the relative aftershock hazard to effectively depict the possible area of future 4.0 aftershocks. | en_US |
| DC.subject | 餘震時空規模風險模式 | zh_TW |
| DC.subject | 混合分布 | zh_TW |
| DC.subject | 接收者操作特徵曲線 | zh_TW |
| DC.subject | 相對餘震風險圖 | zh_TW |
| DC.subject | Space-time-magnitude aftershock hazard model | en_US |
| DC.subject | mixture distribution | en_US |
| DC.subject | receiver operating characteristic curve | en_US |
| DC.subject | relative aftershock hazard map | en_US |
| DC.title | 2019 年美國加州規模 7.1 里奇克萊斯特地震之餘震風險統計評估 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |