CN105225046A - A kind of Regional Landslide sensitivity assessment data sampling method - Google Patents

Abstract

The invention discloses a kind of Regional Landslide sensitivity assessment data sampling method, the method comprises the following steps: the particular location and the bounds that obtain original each landslide in study area; The sliding mass main body landing part on each landslide is outwards expanded certain distance and obtains buffer zone, border; According to hydrologic condition and the geographic and geomorphic conditions acquisition boundary line of study area, and binding buffer district generates landslide sample area; Region beyond sample area and landslide itself is considered as non-landslide areas, chooses certain proportion as non-landslide sampled data; Sample area is sampled to each factor of influence on landslide respectively, and generates landslide and non-Landslide Features sampled data set.The present invention with generation area of coming down for starting point, searching has landslide condition and not destroyed region is Landslide Features region, and complete the sampling work of Regional Landslide in sample region as a comparison with non-landslide characteristic area, for Regional Landslide sensitivity assessment provides a kind of effective data sampling method.

Description

A Data Sampling Method for Regional Landslide Susceptibility Assessment

technical field

The invention relates to the technical field of disaster prevention and mitigation, in particular to a data sampling method for regional landslide susceptibility assessment.

Background technique

Landslide disaster is one of the most important types of disasters in geological disasters, which has the characteristics of wide distribution area, high frequency of occurrence, fast movement speed and serious disaster loss. So far, landslide disasters have occurred in almost all mountainous areas with human habitation and engineering activities in the world, becoming the most frequent and costly geological disaster type among various disasters. According to the National Geological Hazards Bulletin released by the China Geological Environment Monitoring Institute, an average of more than 9,000 landslides occurred in my country every year from 2012 to 2014. In addition, landslides cause more economic losses than other natural disasters in many countries. Therefore, it is of great scientific and practical significance to deeply study the law of landslide occurrence and carry out landslide prediction. The key technology for the first step in landslide prediction is landslide susceptibility assessment, the result of which will directly affect the accuracy of follow-up prediction work.

Landslide susceptibility assessment is not only a method to predict the spatial location of future landslides, but the main idea of the current sensitivity assessment is to use the conditions of landslides in the past to predict the possibility of landslides in the future, that is, future landslides will be different from landslides that have occurred in the past. They have similar geological, landform and environmental conditions. As such, the main basis for this approach is to sample data from past and recent landslides to construct a reasonable sensitivity assessment model. Different sampling methods have great influence on the construction of sensitivity evaluation model. The present invention is completed in order to meet the above requirements, and its purpose is to design an effective data sampling method for regional landslide susceptibility assessment.

Contents of the invention

The technical problem to be solved by the present invention is to provide a regional landslide susceptibility assessment sampling method for the defects in the prior art, which can improve the landslide susceptibility assessment accuracy and provide prerequisites for subsequent landslide prediction and disaster prevention and mitigation work.

The technical solution adopted by the present invention to solve the technical problems is: a method for sampling regional landslide susceptibility assessment data, comprising the following steps:

1) Combining the existing historical data and available field and real-time data to obtain the specific location and boundary range of each original landslide in the study area;

2) On the basis of the original boundary of the landslide, the sliding part of the main body of the landslide is expanded to a buffer zone of a certain distance, which is regarded as an area that has landslide conditions and has not been damaged by the landslide;

3) Obtain the boundary line according to the hydrological conditions and topographic conditions of the study area, and combine the buffer zone to generate the landslide sampling area for landslide susceptibility assessment;

details as follows:

The hydrological conditions are micro-watershed dividing lines, river networks, etc., and the topographic and landform conditions are slope unit boundaries.

The sampling area for landslide susceptibility assessment forms the final landslide sampling area for landslide susceptibility assessment after subtracting the area enclosed by the boundary of the landslide itself from the minimum area enclosed by the above two boundary lines and the buffer zone boundary line of step 2);

4) The area outside the landslide sampling area and the landslide itself is regarded as a non-landslide area, and a certain proportion of the area range is selected as the non-landslide sampling data in this non-landslide area;

5) each impact factor of landslide is sampled with landslide and non-landslide sampling area respectively, and generate landslide characteristic sampling data set and non-landslide characteristic sampling data set;

The impact factors include disaster-causing factors and inducing factors.

According to the above scheme, the data in the step 1) includes on-site survey data, historical landslide data, and remote sensing data; the specific position and boundary range of the landslide are represented by a vector diagram with a coordinate system including the boundary of the landslide and the internal area.

According to the above scheme, the hazard-causing factors in step 5) include slope gradient, aspect, elevation, rock formation type, slope form, vegetation index, etc., and the inducing factors include earthquakes, rainfall, human engineering activities, etc.

The beneficial effects produced by the present invention are: the present invention provides an effective sampling method for regional landslide susceptibility assessment, takes the landslide occurrence area as the starting point, searches for landslide conditions and undamaged areas so as to form landslide characteristic areas, and uses non-destructive sampling methods The landslide characteristic area is used as a comparison sampling area to complete the sampling work of regional landslides, which is used as the input data for sensitivity assessment to complete the calculation of sensitivity index and classification. This sampling method can improve the accuracy of sensitivity classification.

Description of drawings

The present invention will be further described below in conjunction with accompanying drawing and embodiment, in the accompanying drawing:

Fig. 1 is the structural representation of the embodiment of the present invention;

Fig. 2 is a schematic diagram of the sampling area when the buffer zone is smaller than other conditions according to an embodiment of the present invention;

Fig. 3 is a schematic diagram of the sampling area when the buffer zone is larger than other conditions according to an embodiment of the present invention;

detailed description

In order to make the object, technical solution and advantages of the present invention more clear, the present invention will be further described in detail below in conjunction with the examples. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

Fig. 1 is a flowchart of a regional landslide susceptibility sampling method according to an embodiment of the present invention. In conjunction with Fig. 1, a kind of regional landslide susceptibility assessment sampling method of the present invention is described in detail with 4.20 Lushan earthquake Baosheng Township area landslide susceptibility assessment, comprises the following steps:

(1) Obtain the specific location and boundary range of each original landslide in the study area. Using UAV images and field surveys on the day of the earthquake, 226 earthquake-induced landslides were extracted, the largest of which was 45000m ² , the smallest was 100m ² , and the average landslide was 5000m ² ;

(2) As shown in the buffer zone in Figure 2 and Figure 3, on the basis of the original boundary of the landslide, the sliding part of the main body of the landslide is expanded to a buffer zone of a certain distance, which is regarded as an area that has landslide conditions and has not been damaged by the landslide .

The sliding part of the main body of the landslide refers to the buffer zone that takes the boundary of the back wall of the landslide as the top, and takes the boundary of the landslide body except the tongue and the back wall of the landslide as the side, and expands the buffer zone with a distance of m to the outside with the top and the side as the boundary. The value of the buffer distance m is set by those skilled in the art according to specific conditions, and in this embodiment, 100 meters is used as the buffer distance.

(3) Set specific rules for extracting sampling areas according to the hydrological conditions and topographic conditions of the study area.

In this embodiment, the dividing line of the micro-watershed is used as the hydrological condition, and the boundary line of the slope unit is used as the terrain condition. The landslide boundary line is taken as one of the boundary conditions of the sampling buffer zone, whichever is closer to the microwatershed dividing line and the slope unit boundary line. If the boundary exceeds the buffer zone extending 100 meters to the outside, the 100-meter buffer zone is used as the boundary line; The surrounding area of the sampling boundary of each landslide is subtracted from the area enclosed by the boundary of each landslide to form the landslide sampling area. The schematic diagrams of landslide sampling areas in the two cases are shown in Fig. 2 and Fig. 3.

(4) The landslide sampling and areas other than the landslide itself are regarded as non-landslide data, and a certain proportion of the area is selected as the non-landslide sampling data. In this embodiment, the principle of random selection is chosen, and an area with the same size as the landslide sampling area is taken as the non-landslide sampling data.

(5) Sampling the landslide and non-landslide sampling areas for each impact factor of the landslide, and generating a landslide characteristic sampling data set and a non-landslide characteristic sampling data set. Ground elevation, slope, aspect, stratum, slope structure, average distance to fault, average distance to water system, and normalized difference vegetation index are selected as hazard-causing factors, and peak earthquake acceleration and seismic intensity are selected as inducing factors. During specific implementation, those skilled in the art use related software to carry out data sampling for these multiple factors in the landslide and non-landslide sampling areas of steps (3) and (4).

(6) Input the landslide and non-landslide sampling data into the sensitivity assessment model to calculate the landslide hazard susceptibility index. The 70% data in the landslide and non-landslide areas are input into the sensitivity assessment model as the input data, and the sensitivity assessment model adopts the support vector machine. The sensitivity index value is calculated by this evaluation model.

⑺According to the value of the sensitivity index, use the grading method to realize the division of sensitivity grades. The landslide susceptibility is graded by the natural break point classification method, and the landslide susceptibility grade is generated, including five grades of extremely high sensitivity, high sensitivity, medium sensitivity, low sensitivity and extremely low sensitivity.

By statistics (Table 1 and Table 2), it can be known that the correct rate of this method to calculate landslides (comprising the sum of the landslide proportions of very high risk areas, high risk areas and medium risk areas) is 92.22%, and the receiver operating characteristic curve (ROC ) is 99.3%, the correct rate of the landslide obtained by the conventional sampling method (that is, the area of the landslide itself) is 82.23%, and the area under the receiver operating characteristic curve is 91.6%. This shows that the sampling method can improve the accuracy of regional landslide susceptibility classification.

Table 1 The experimental statistical results of this sampling method

Table 2 Statistical results of conventional sampling method experiments

It should be understood that those skilled in the art can make improvements or changes based on the above description, and all these improvements and changes should belong to the protection scope of the appended claims of the present invention.

Claims (3)

1. a Regional Landslide sensitivity assessment data sampling method, is characterized in that, comprises the following steps:

1) particular location and the bounds on original each landslide in study area is obtained in conjunction with different pieces of information;

2) on the basis on original border, landslide, the landing part of sliding mass main body is outwards expanded the buffer zone of certain distance, be considered as there is landslide condition again not by the region of landslide destruction;

3) according to hydrologic condition and the geographic and geomorphic conditions acquisition boundary line of study area, and binding buffer district generates the landslide sample area of landslide sensitivity assessment;

Specific as follows:

Described hydrologic condition is miniflow territory dividing line, and described geographic and geomorphic conditions is ramp unit boundary line;

The landslide sample area of final landslide sensitivity assessment is formed after the sample area of landslide sensitivity assessment deducts by the Minimum Area that above-mentioned two boundary lines and these three boundary lines, boundary line, buffer zone surround the region that border, landslide own surrounds;

4) region beyond landslide sample area and landslide itself is considered as non-landslide areas, chooses a certain proportion of regional extent in this non-landslide areas as non-landslide sampled data;

5) sampled to each factor of influence on landslide respectively in landslide and non-landslide sample area, and generate Landslide Features sampled data set and non-Landslide Features sampled data set;

Described factor of influence comprises and comprises Flood inducing factors and incitant.

2. method according to claim 1, is characterized in that, in described step 1), data comprise Field Research data, history land slide data, remotely-sensed data; The particular location on described landslide and bounds are represented by the polar plot of the band coordinate system on the border and interior zone that comprise landslide.

3. method according to claim 1, is characterized in that, in described step 5), Flood inducing factors comprises the gradient, slope aspect, ground elevation, rock stratum type, slope form, vegetation index, and incitant comprises earthquake, rainfall, Human dried bloodstains.