Regional Landslide Hazard Assessment Based On TRIGRS Model

Rainfall is one of the main inducing factors of landslide,according to statistics,70%of landslides are induced by rainfall.The rainfall-induced landslide has the characteristics of wide distribution,large number and great harm.Therefore,in recent years,the research heat of rainfall-induced landslide has been high.At present,the research methods of rainfall-induced shallow landslides by domestic and foreign researchers are gradually shifting from qualitative analysis to quantitative analysis.TRIGRS（Transient Rainfall Infiltration and Grid-Based Regional Slope-Stability Analysis）model is one of the most widely used models.Such quantitative models are mostly used in the susceptibility and risk evaluation of loess shallow landslides,while few studies are carried out in the red-bed area.Therefore,it is of great scientific significance and practical value to carry out quantitative evaluation of landslide risk in red-bed area.Xide Town is located in the southwest Liangshan Prefecture of Sichuan Province.Red beds are widely distributed in the county,and geological disasters are frequent and concentrated,especially landslides.Therefore,on the basis of field investigation and data collection,this thesis selected the Luoha Town of Xide Town as the research area,took"8.31"mass geological disaster as the background,used remote sensing images to visually interpret and classify the landslide in the area,and analyzed the spatio-temporal distribution law of the most representative landslide types in the area.Then,TRIGRS model was used to calculate slope stability under different rainfall conditions in the study area,and hazard zoning maps under different rainfall conditions were obtained,so as to complete the risk evaluation of landslide in the study area.AUC value method and historical data demonstration method were used to verify the evaluation results and analyze the accuracy.The main conclusions of this thesis are as follows:（1）The temporal distribution characteristics of shallow landslides in Loha Town are as follows:landslides mostly occur in heavy rain in July and August,the frequency of landslides in July and August was 0.27 and 0.43.（2）The spatial distribution characteristics of shallow landslides in Luoha Town are as follows:the landslides are mainly distributed in the mountain canyon landform,the proportion of landslide distribution is 58.4%,and the point density is 3.7 per km²;In the range of slope,it is mainly distributed in 20-45;In elevation range,they are mainly distributed in 2400m-2800m.Landslides are mainly distributed along the two sides of the main river basin of Mishi River.The landslides are mainly distributed in the soft siltstone,fine sandstone and mudstone formations with medium to thick layers of Leidashu Formation（K₂l）and Xiaoba Formation（K₂x）of Upper Cretaceous.There are many landslides in the area with strong human engineering activities.（3）When TRIGRS model is used for shallow landslide risk assessment,attribute zoning based on lithology differences and setting multiple rainfall stages based on rainfall intensity changes can conform to the geological background and rainfall process of the study area to a large extent,thus reducing the error of model simulation results.（4）TRIGRS simulation results show that with the increase of rainfall intensity,the area of danger area keeps increasing,while the area of extremely low risk area keeps decreasing.Under the ten-year rainfall condition,the total area of extremely high and high risk areas in Loha Town accounted for 12.74%,mainly distributed along the main river basin of Mizhi River,and the area of extremely low risk areas accounted for 80.84%.Under the condition of one rainfall in 20 years,the total area of extremely high and high risk areas accounted for22.19%.In addition to the distribution along the main river basin of Mizhi River,the risk areas were also scattered in narrow gullies and strong catchment areas.Extremely low risk area accounted for 68.77%;In fifty years once rainfall condition,high and high danger zone with a total area of 28.92%,low of danger area ratio had fallen to 60.83%,extremely low zone or low risk area into high danger zone accelerated along with the increase of rainfall intensity,one in 50 years rainfall conditions can lead to unstable slope area to achieve high water content,The value of rainfall in 50 years can be used as the reference threshold of regional shallow landslide,debris flow and mountain flood.Under the condition of 100-year rainfall,the total area of extremely high and high risk areas increased to 46.28%,which was significantly increased compared with the above three conditions.The increase of extremely high risk areas was particularly obvious,and their distribution covered all catchment areas in the region.The proportion of the area in the extremely low risk area decreased to 51.35%,and the overall risk of the area increased significantly.（5）The AUC value method was used to evaluate the accuracy of the calculation results of the model under the condition of once-in-a-century rainfall,and the AUC value was 0.77,with good accuracy;At the same time,the prediction results were compared and verified by the method of historical data demonstration.By superimposing the distribution location of"8.31"mass landslide and the hazard zoning map under the 100-year rainfall intensity,the results showed that there were 820 historical landslides in the high risk area and extremely high risk area,accounting for 77.08%of the total landslide in the area.There were 181landslides in medium and low risk zones,accounting for 16.99%of the total number of landslides.The number of landslides distributed in the relatively stable area was the least,only63,accounting for 5.93%of the total number of landslides.The validation results of the two accuracy evaluation methods are good,indicating that the evaluation results of TRIGRS model are scientific and reasonable,which can provide scientific basis for the risk management and spatial planning in Loha Town.