Research On The Materials Evolution Characteristics Of Debris Flows Based On RS In Earthquake Area

On14August2010, a total of23debris flows in Minjiang River DrainageSystem were triggered by heavy rainfall around the town of Yingxiu where near theepicenter of the Wenchuan earthquake, especially Hongchun trench which has thelargest scale, and the debris flows caused great losses to the disaster area. Moreover,on July10,2013, the debris flow flooded again around the mid Minjiang River thatspreading from Yingxiu town to Wenchuan city which locates near the epicenter of theWenchuan earthquake, especially Qipan trench which has the largest scale. Debrisflow has become one of the largest geological disaster threats after the earthquake.The earthquske caused great injuries and deaths to the disaster area, at the sametime it triggered lots of landslides, collapsesand other geological disasters. The highresolution remote sensing images and field survey showed that, the total area of23debris flowin the study area with a area of225km2is59.82km2.After the earthquakeof May12,2008, the total number of landslidesdeposits in the basin was857, with atotal area of958.18×104m2, and the total amount of depositsincreased to52.3×106m3.The landslidesdeposits in the basin mainly distributed at altitude from800m to2000m, on slope between40°and50°, and among200m distance within therange of the trench.It is in the middle altitude region and in steep area near river channel as well asin the channel. The deposits could quickly join into the debris flows activities in theprocess of rainfall. Landslide deposits in the basin exists mainlyin three forms ofcollapses, landslides, unstabilizedslope, part of whichhas slided, there is a part of thelandslide“hang” on the steep wall which may fall at any time under the external force;part of which has already apart from the motherbody, lots of detritus piled up andjamed in the channel; part of which still not slide, but has broken structure. Under therainfall conditions, changed effective stress inside and reducedshearing resistance,landslide is likely to "resurrection" or generate new landslide which could join intothe debris flow activity by the help of the eroding, washing and carrying effect of flood.On August14,2010, the study area broke out large-scale debris flows. Throughhigh-precision remote sensingimage interpretation, comparative analysis and fieldsurvey, thevariationcharacteristics of rainstorm-triggered landslides are analyzed.From the Statistics data, we found that the total number of landslidesafter heavy rainwas1285, increased by49.94%compared with the number after earthquakes and thetotal area of the landslidedepositswas1127.01×104m2, increased by17.6%comparedwith the area after earthquake. The landslidestriggered by Stormrain majoritydistributed at altitude from2000m to3000m, on slope between40°and50°, andgreater than400m distance to the trench.Which were very developed at a higherelevation and in steep landslide zone,by comparing the datas, the landslides triggeredby Stormrain are mainly located in the rear edge of the basin and both sides of thechannel, with bedrock emerged in the rear edge. Broken structure after the earthquake,thin soil cover, changed effective stressafter rainfall and reducedshearing resistance,so that the landslides were the most developed in this region; the landslide near thechannel, debris flow erosion the end of the landslide, and moved into debris flowactivity. In this process of debris flows, the deposits supplycame from the shallowlandslides or the surface of a large landslide accumulation or newly generatedlandslides. Under the rainfall conditions, landslide partial "resurrectioned", landslideexpanded to the rear edge, the leading edge collapse after erosion; runoff emerged onslope surface in the flood, detritus on the slope surface were carried into the debrisflow activity; or newly generated landslides, which depend on the eroding, washingand carrying effect of stormrain and flood, carrying huge amounts of energy, andaddedinto the debris flow activity.This study selected5typical debris flow gullies near the town of Yingxiu,referenced predecessors’ researches to analysis the debris flows trend activities byconsulting2008,2010,2013three periods of high resolution remote sensing imagesand the real change data. In2010, the first rainstorm year that occured large scaledebris flows after the earthquake, the landslide area increased by16.4%comparedwith that of year2008in the study area. Then in2013, the second rainstorm year, thelandslide area reduced11.5%from that in2010. After the earthquake, the debris flowcumulative rainfall and critical rainfall has been lower than that before theearthquake, with the rainfall intensity in2013is higher than that in2010, so thereduced amount between2010and2013could be used to calculate the debris flowtrend in the future. The debris flow active time is17years near Yingxiu town of Wenchuan earthquake and the trend will continue until2025. Comparing withChi-Chi earthquake in Taiwan, the landslide source materials occupied area in thebasin of study area would show the attenuation trend every year, whereas on theheavy rainstorm year the landslide area would be rise, but it would present attenuationoverall, until return to the state before the earthquake.This research took45debris flow gullies in the study area as the sample data topredict the maximum run out amount of a debris flow in the study area by building amodel with three factors as landslide area of material source, watershed area and basinelevation difference based on the characteristics of debris flows in earthquakearea,and13debris flow gullies in Qingping are used to validated. The chosen factorsin the prediction model not only included material source that could impact theoccurrence of debris flow, but also contained natural characteristics of the basin.Watershed area determines the size of the catchment area, basin elevation differencetransforms the flood potential energy to kinetic energy which could provide energy fordebris flow active, and material source provides debris for debris flow runout source,they all have a direct relationship with the debris flow activity and the model isbilievable. However, multivariate regression modeling requires a lot of samples anddebris flow activity is restricted by many conditions including space, time and otherconditions of the starting form. This prediction model needs to be defined by thedebris flow characteristics; it could not accurately predict the maximum runoutamounts in all conditions. The results of this research provide scientific basis fordebris flow risk assessment and help with the evaluation of geologicalengineering-related fields, as well as offer developmental status estimations forsurvey and design of debris flows.