Research On Nonlinear Prediction For Glacial Lake Outbursts In The Himalayas Area, Tibet

?In recent years, , the Himalaya Mountains area glacial lake disaster is been prominent day by day for global climate change's influence, the glacial lake which possibly bursts quantity to increase unceasingly, causes the personnel casualty and the massive property damage for the natives, the Tibet economic society develop is restricted seriously. Meanwhile, along with strategy implementation of the national west development, each infrastructure's construction consummates gradually, harmful levels of the glacial lake getting more and more serious. In summary, the forecast which the glacial lake burst to research necessarily.Research paper on the environment and conditions of the glaciers geological surveys and research relevant information, summarizes natural geographic and geological survey of the Himalaya mountain area, analyzed engineering geology environmental conditions of the Himalayan mountains area. Reference to the research and combination of the survey data, research the outburst condition and mode of glacial lake in the investigation area; Summarizes the influencing factor of the glacial lake outburst and research values and measure for forecast the glacial lake outburst.Disaster of glacial lake outburst is serious day by day in research area, with wide distribution the characteristics of large damage under the global temperature rises to the environment. At present, glacial lake burst for the study area is still not an effective disaster prediction method. The first use of paper-based support vector machine - extension theory of nonlinear prediction model glacial lake outburst, and applied to paper and make it rely on the glacial lake in the project survey, as glacial lake outburst with a high degree of nonlinearity and uncertainty. Studies have shown that glacial lake outburst based on nonlinear theory prediction model can be more objective to predict the risk of glacial lake, provide a scientific basis for disaster prevention for glacial lake outburst of Himalayan region.Articles by glacial lake system of the Himalayan region to obtain the following results and conclusions:1.To study the area of glacial lake outburst has occurred based on survey data, analysis of conditions of glacial lake outburst disasters; including Climatic background, catchment conditions, terrain conditions, hydrological conditions, earthquakes, etc. The glacial lake outburst model are considered both ways, the combination of the climatic background and the different failure modes of moraine, in which the former one is divided into 4 categories ,17 modes,as the latter one 3 modes:wave overflow,seepage deformation, tributaries ditch erosion . No clear boundaries could be seen between the latter three modes, transformation is also available under certain conditions as well.2.Factor in the analysis of glacial lake outburst, constructed by the glacier - glacier lakes - the main line of morained dam glacial lake outburst prediction index system. With supply glacier area, glacier snow area average longitudinal, tongue slope, the front tongue distance from the glacial lake, glacier fractured situation, glacial lakes area, development of cross-strait situation of collapse, defection slope degree,glacier crest width of the ditch by the side degree of erosion, water and heat combination of 11 factors such as glacial lake outburst forecast evaluation indicators.3. Risk of glacial lake outburst is the classification of various factors on its characteristics and relationship between the function approximation problem, namely the pattern recognition problem. Support vector machines for nonlinear system has to avoid the local minimum, over the advantages of learning defects. In this paper, based on the support vector machine algorithm and basic principles established risk classification model of glacial lake outburst, and successfully applied this model to the Himalayan region of 25 instances of glacial lake outburst prediction.4. In the theory of support vector machine based prediction models were tested and found, Learning samples obtained through training to learn the correct rate of 90% or more than one model. This shows that, sampling method for training by learning from the model is correct is not enough reliable.Studying samples of 20 randomly selected models were tested again, test results showed that more than 90% correct, leaving only one, finally ,identified as a reliable algorithm based on support vector machine prediction model ?glacial lake outburst.5.In the extenics prediction method of glacial lake outburst, the key to the evaluation result perfection is precise weighing. The science in the evaluation calls for weights of reason. This article introduces a weighing methods based on rough set theory. The weight coefficient of each evaluation indicators has been determined under the concept of attribute importance in rough set theory. Thus weights have been given more objectivity as the indicator priorities can be easily seen thought the values of weights. The weighing problem in glacial lake outburst prediction is therefore solved based on extenics theory . In the results, the most influential indicators to the Himalaya glacial lake outburst are glacier indicators, followed by moraine dam indicators and lake indicators. Analyzing on outburst-positive or negative, about 70% indicators are outburst-positive with 30% negative.. Given the danger ranks of glacial lakes follow uniform distribution, outburst danger exists in 70% glacial lakes in the Himalaya areas, while 30% are comparatively stable.6. More glacial lakes in Himalaya areas are highly dangerous, about 40% of those remaining to survey, while 20% with low danger. Highly dangerous lakes commonly appear with large glacier coverage on the trailing edge and glacier tongue into the lakes, supplying abundant water for the outburst. The positive climate condition for outburst are wet-hot and dry-hot on water-heat combination in the areas. Moreover, danger is seen when it is narrow and weak on the top of the moraine dam, as damage and breakdown can be easily produced when the lake is flooded.7. Compare support vector machine method to predict the results and evaluate the results of extension as well as the actual results, the results show that, both nonlinear method using glacial lake outburst compared to predict the results of similar, consistent with the actual situation. Visible by means of a combination of both linear comparison can be drawn is the result of reasonable and objective, this also explains the nonlinear theory for the nonlinear problems studied are the most effective way, once again proved the topic of this paper is scientific and practical significance...