Study On Forecast And Prediction Of Large Scale And Sudden Landslide Hazard

China is a country where landslide hazards are very active and bring a huge amount of damage to poverty and great threat to people's lives. Especially affected by global warming, lots of landslides occur when the climate change abnormally in recent years. For example, in Sichuan province, the Tiantai village landslide(large scale) occurred in 2004; a large crack had been discovered on the mountain behind Danba town in 2005; the Baishi village landslide(large scale) happened in 2006, Mianyang city, Beichuan county; Landslides take placed at Yanmen (in Qingning village) and Yangliuya substation, Dazhou county in 2007, and more, lots of landslides broke out after Wenchuan 8.0 earthquake in 2008. It is urgent that a theory and technique system of forecast and prediction of large scale and sudden landslides should be established during the emergency situation at large disasters which damage to both people's lives and property.Due to the complex triggering factors, landslide movement has lots of uncertainty and comprehensive influence elements, it is too difficult to predict landslide in a precise way which is still a global problem. Currently, the main method to predict landslide is according to the three developing stages of landslide in the displacement-time curve. When the deformation becomes close to or enters into the sliding stage, the prediction should be given combining the macro-deformation signs. However, the recent method only stays on the empirical aspect, and it is hard to be accurate to predict for no quantitative standard. The calculation of run out distance of landslide is also a big issue which has been concerned for a long time, because it is directly related with the disaster degree to peoples. Lots of scholars have provided several equations to calculate the speed and run out distance of landslide, but they are not available for other cases and still can not build a completed system to forecast it.Aiming to solve the problems in forecast and prediction of large scale and sudden landslides by gathering investigation and monitoring data from several cases, the paper studies detailly and deeply the evolution laws and characteristics in every deformation phases as well as possible run out distance after the fall of landslide, and provides some new results of research:(1) Based on lots of monitoring data, it is categorized the displacement-time curves into six types after thinking over carefully: stable, straight line, curve, stairs, up and down and Convergence. It is different among displacement-time curves according to causes and evolution stages of different landslides. To analyze landslides according to different types of curves is very important in describing the mechanisms of landslides, deformation stages and degree of disturbed by outer factors.(2) In order to solve the difficulties of change of displacement-time curves with the various coordinates, the paper shows the solution of finding out the specific tan angle of certain point of the curves through changing coordinates. Based on the adjusted tan angle characteristics, it is provided the standard criterions for distinguishing the different stages----the initialization deformation; equal speed deformation and accelerated deformation, of the deformation process after in-depth study. In order to realize the accurate prediction, it is given the criterions of recognizing the three sub stages----initial acceleration; middle acceleration and temporary slide, of the acceleration deformation. In the meanwhile, the max sliding angle is reaching to 89o before landslides fall。(3) For further research, the paper studies the characteristic and regulation of the acceleration and accelerated acceleration of displacement-time curves and provides the quantitative methods of how to recognize every deformation stage of slope and give the criterion of prediction of the sliding of landslide: Accumulate acceleration∑a >0 and keep on increasing; Accumulate accelerated acceleration∑? >0 and keep on increasing or stay at a high value.(4) The paper provides the different quantitative methods of prediction of the landslide with models according to the different deformation stages. For instance, it should be applied the artificial neuron network for the long term deformation to predict the deformation tendency; and applied the Verhulst methods to predict when the landslide enters into the sliding stage. The methods for predict the landslide in short period is also mentioned in this paper which is the parabolic curves modeling method as well as the acceleration method in sliding stage.(5) Based on the systematic researches on situations of different sliding planes of landslides and characteristics of cracks in various stages, it is summarized the laws and characteristics of cracks matching with various stage. According to the characteristics, it is judged and understood the deformation stages and the stability of landslide and applied to the forecast and prediction in empirical methods, based on the research of macro-deformation signs.(6) In order to analyze and assess the run out distance of landslide after it slides down, it is divided the types of sliding into 5 groups: whole sliding without colliding; whole sliding with colliding; sliding and depositing; flying, sliding at a arc plane of slope toe. It is also deduced the calculation equations of the landslide's run out distance with the theories of the law of conservation of energy and the law of conservation of momentum.(7) Based on the synthesis of characteristics of displacement-time curves, appropriate landslide prediction models of stages and characteristics of cracks matching with various stage, It is mentioned in the paper to divide the forecast and prediction stages into four levels depending on the degree of dangerous: blue-notice level; yellow-warning level; orange–dangerous level and red-alarming level. The synthetic methods of dealing with each level are provided in the table with lots of information including characteristics of deformation curves, macro-deformation signs, prediction criteria, response characteristics of landslide to the outer factors, appropriate prediction models and methods and disaster response measures.In the end, based on the researches of cases of landslide forecast and prediction, the paper shows some large scale and sudden landslide cases with the application of those methods and theories which are mentioned in the former of the paper.