Development Characteristics And Susceptibility Evaluation Of Ground Fissures In Weinan

The crisscrossing active faults resulting from the complex basement tectonic activities have partitioned the basin area of Weinan City into distinct structural fault blocks.The unique tectonic pattern has created complex and fragile shallow geological bodies and conditions within the research area,such as multi-level distribution of topographic scarps and various geomorphological types.In recent years,the increasing intensity of human activities has led to zonal cluster development of ground fissure disasters in Weinan City.The super destructive power of these fissures has caused significant losses to the safety of local people’s lives and property,hindering the local economic and social development and long-term planning.Facing the increasing occurrence of ground fissure disasters,it is of great theoretical and practical significance to investigate the latest development status of ground fissures in Weinan City,unveil its formation process and main influencing factors,and conduct susceptibility assessment for ground fissures in Weinan City.With the continuous expansion of the ground fissure disaster database and advancements in digitizing geological information,it has become feasible to employ quantitative assessment methods such as data-driven and machine learning techniques for evaluating the susceptibility of ground fissure disasters.Through detailed field investigation,UAV mapping,literature collection,trenching,drilling and other methods,the geological background of the study area is comprehensively understood,and the latest data on ground fissure development in the study area are obtained.Based on this,the Python programming language was used to study ground fissure susceptibility evaluation and prediction methods.The optimal evaluation model was used to identify the potential high-risk areas of ground fissure disasters in the study area.The main research work and achievements of this paper are as follows :(1)The latest database of ground fissure disasters in Weinan City,including 214 disaster points,was established,and the genetic types,plane morphology and activity rules of Weinan ground fissures are summarized.It is concluded that ground fissures have significant spatial consistency with geomorphology,lithology,faults,ground subsidence and other factors.The disaster-causing characteristics of Weinan ground fissures are elaborated in detail.(2)Respectively revealed the disaster-causing characteristics and genetic mechanism of fracture creep,earthquake,loess collapse and other genetic types of typical ground fissures,and summarized the main influencing factors of Weinan ground fissures development: landform,lithology,active faults,historical earthquakes,Quaternary thickness,human activities and land subsidence,rainfall and other factors.(3)The certainty factor method was used to analyze the classification characteristics of each factor,and the high-sensitivity factors were screened by using the ROC curve method to establish the susceptibility assessment index system."Equal spacing method" and "controlled spacing method" were used to extract disaster samples and non-disaster samples,respectively,to build a high-quality dataset containing 984 samples,which solves the problems that seriously affect the quality of sample data sets,such as inadequate expression of disaster data,too few positive samples,and non-disaster samples in high-prone areas.(4)The improved certainty factor model,CF-LR and CF-XGBoost machine learning models were selected to carry out experiments,and the performance of the three models was tested by ROC,AUC,Kappa coefficient,Confusion Matrix,Accuracy evaluation index,and comparison of prediction results.The performance ranking of the three models is CF-XGBoost >CF-LR > L-CF,and the machine learning model has higher performance and prediction accuracy.Finally,The optimal evaluation model was selected to evaluate the risk and susceptibility of ground fissures in Weinan,and it is identified that the potential high-prone areas of ground fissure disasters in the study area are mainly the overlapping areas of the faults and the second and third river terraces covered by loess in the north of Dali County.