Statistical feature of epicenter accuracy of historical strong earthquakes and its effect on seismic hazard study

Based on Catalogue of Chinese Historical Strong Earthquakes (23rd century B.C. to 1911 A.D.), we statistically analyzed the epicenter accuracy parameters given by the catalogue in different time and magnitude intervals, and gave geographical distribution of the earthquakes with different categories of accuracy. The result shows that percentage of the historical strong earthquakes with the category 3 accuracy or above is 58.4%, with corresponding epicenter uncertainty being greater than 25 km. Generally speaking, the historical strong earthquakes have a low epicenter accuracy while they have a longer history. The percentage of low accuracy historical earthquakes is not small even after 1500 A.D. For the historical MSge;6.0 earthquakes the percentage of category 3, 4 and 5 events is 60.1%. The geographical distribution of historical strong earthquakes with different epicenter accuracies shows different characteristics. For example, Yunnan Province has the largest number of category 1 and 2 earthquakes, while most historical earthquakes in Xinjiang and Tibet regions are low accuracy events. We also illustrated that the epicenter uncertainty of historical strong earthquakes may affect the results of potential seismic source delineation, near-field seismicity study and seismogeological environment assessment. So, in seismic hazard evaluation and seismogeological study, it is necessary to make textual criticism on the historical earthquakes with low epicenter accuracy, or appropriately reject some of them.