Abstract: At present, with the wide application of the Newmark method, various Newmark empirical formulas with different ground motion parameters have been fitted by many researchers based on global strong-motion records. However, the existing study about the Wenchuan earthquake does not quantitatively evaluate the applicability of different Newmark models based on the actual landslides distribution. The aim of this paper is to present a comparison between observed landslides from the 2008 Wenchuan earthquake and predicted landslides using Newmark displacement method based on different ground motion parameters. The factor-of-safety map and critical acceleration(a_c)map in the study area are obtained by using the terrain data and geological data. The distribution of Arias intensity(I_a)and PGA in the study area is obtained by using the attenuation formulas of Arias intensity(I_a)and PGA, which is regressed by Wenchuan ground motion records. Based on the distribution of Arias intensity(I_a)and PGA parameters, we obtained the predicted locations of landslide using Newmark regression equations which are generated using global strong-motion records. The results shows that the assessment results can better reflect the macroscopic distribution characteristics of co-seismic landslides, most predicted landslide cells are distributed on the two sides of the Beichuan-Yingxiu Fault, especially the Pengguan complex rock mass in the hanging wall. The abilities to predict landslide occurrence of the two Newmark simplified models are different. On the whole, the evaluated result of simplified model based on parameter I_a is better than that based on PGA parameter. The GFC values obtained by the Newmark model of I_a and PGA parameters are 65.7% and 34.9%respectively. The evaluated result based on I_a can better reflect the macro distribution of coseismic landslides. The Ls_Pred value based on the Newmark model of parameter I_a is 26.5%, and the Ls_Pred value based on the Newmark model of PGA parameter is 10.3%. However the total area of predicted landslides accounts for 2.4% of the study area, which indicates that the predicted landslide cells are greater than the observed landslide cells. This reminds us that depending on the current input of shear strength and ground-motion parameters, we can only conduct landslide hazard assessment in macro areas, the ability to predict landslide can be improved using more accurate topographic data and input parameters.

Key words: Wenchuan earthquake, earthquake-triggered landslides, Newmark cumulative displacement models, Arias intensity(I_a), PGA

摘要： 本研究的目的是利用真实的汶川同震滑坡数据分别对利用I_a和PGA参数的2种简易New￣mark滑坡预测模型进行定量评估分析。根据研究区的地形数据和区域地质资料，获得了研究区的临界加速度（a_c）分布，利用基于全球地震动记录拟合得到的Newmark经验公式，结合汶川地震的I_a和PGA参数分布，获得了基于2个地震动参数的滑坡位移分布图。大体上，基于I_a参数的滑坡危险性评估结果略好于基于PGA参数简易模型的评估结果。基于I_a的评估结果能够更好地反映同震滑坡的宏观分布特征。然而，2种Newmark模型反演得到的滑坡面积要明显大于实际滑坡物源区的面积。这提醒我们基于当前岩性和地震动数据只能获得宏观区域的滑坡相对危险性。而通过更为准确的地形数据和输入参数，才能进一步提高预测位移结果的精确度。

关键词: 汶川地震, 地震滑坡, Newmark累积位移模型, Arias烈度(I_a), 地震动峰值加速度(PGA)