Meteorological early warning and forecast of landslide and debris flow geological disasters

Meteorological factors are the key factors to induce geological disasters such as landslides and mudslides. Developing a real-time early warning and forecasting system based on Web-GIS and real-time meteorological information, realizing real-time early warning and forecasting of geological disasters and a networked geological disaster reduction and prevention system, making real-time early warning and forecasting of possible geological disasters, and releasing early warning information in a timely and extensive manner are conducive to scientific, efficient and rapid disaster prevention, thus minimizing disaster losses, protecting people's lives and property safety, and changing passive prevention into active prevention and control of geological disasters.

First, the main basis for meteorological early warning and prediction of landslide and debris flow geological disasters

Spatial prediction of regional geological disasters (landslide, debris flow, etc.). ) mainly delineating the geological disaster-prone areas, that is, the risk assessment and zoning of geological disasters discussed above. Based on the spatial prediction of regional geological disasters, combined with real-time meteorological dynamic information, the main inducing factors of geological disasters such as landslides and mudslides are analyzed, and the statistical laws and internal mechanisms of geological disasters under different meteorological conditions in the same geological environment region are studied. By determining the critical thresholds of effective rainfall model, rainfall intensity model and rainfall process model, the spatio-temporal coupling relationship of regional geological disaster early warning and prediction based on real-time dynamic meteorological information is established, and the dangerous spatio-temporal early warning and prediction of regional geological disasters such as landslides and mudslides are carried out.

According to the geological conditions, disaster investigation and meteorological conditions in the study area, the grades of geological disaster-prone areas are divided, the correlation between geological disasters such as landslides and mudslides and effective rainfall and 24-hour rainfall intensity is counted, and the critical rainfall (I, II) of different grades in different disaster-prone areas is determined as the threshold for discriminant analysis, and the risk grade of rainfall is determined. Rainfall less than Grade I critical rainfall is low risk, rainfall between Grade I and Grade II critical rainfall is medium risk, and rainfall greater than Grade II critical rainfall is high risk.

The effective rainfall of each unit is compared with the critical effective rainfall, and the rainfall risk grade of each unit is determined, and the rainfall risk grade is superimposed with the grade of geological disaster-prone areas. The superposition results are shown in Table 3-4 and Figure 3-2. The early warning and forecasting grades of geological disasters are divided into five grades, corresponding to four different prone areas: Grade 3 and above are the early warning and forecasting grades, Grade 5 is the highest warning and forecasting area, and Grade 1 and Grade 2 are. The third-level early warning area refers to the area where the monitoring of disaster points should be strengthened; The four-level early warning area refers to the area where the monitoring of disaster points should be closely strengthened and certain preventive measures should be taken; Level-5 early warning zone refers to all-day monitoring of disaster points, and taking avoidance measures for direct victims, especially residents and personnel when necessary. In early warning and forecasting, level 3 is the attention level, level 4 is the early warning level, and level 5 is the alarm level.

Table 3-4 Classification Table of Geological Disaster Early Warning Areas

Figure 3-2 Schematic Diagram of Macro Early Warning Construction of Regional Geological Disasters

Since China started the national meteorological early warning and forecasting of geological disasters in 2003, some experts and scholars have devoted themselves to the research and exploration of early warning and forecasting models and methods, which have mainly gone through two stages.

In the first stage, from 2003 to 2006, the first generation early warning method, namely the critical rainfall standard method, was adopted. The main principle of this method is that according to the statistical analysis results of geomorphologic pattern, geological environment characteristics and their relationship with rainfall-induced landslides and geological disasters in China, the first-level division is carried out on the basis of national river basins, climatic zones, tectonic units and regional geological environment conditions; According to the distribution density, topographical features, stratigraphic lithology, geological structure and neotectonic movement, and annual average rainfall distribution of regional watersheds, historical landslides and debris flows, the secondary zoning is carried out. Divide the whole country into 7 early warning zones and 74 early warning zones; And the statistical relationship between historical geological disaster points and actual rainfall in different regions is carried out to determine the critical rainfall of landslide and debris flow disasters in each early warning region, and the early warning and forecasting criterion template is established (Figure 3-3); Using the national geological disaster database, the geological disaster samples in the county and city survey information system and the rainfall data provided by China Meteorological Bureau, through statistical analysis, the critical rainfall of 1, 2, 4, 7, 10, 15 days before the occurrence of geological disasters is determined as the criterion template, and the meteorological early warning and forecasting model of geological disasters is established and carried out.

Figure 3-3 Early Warning and Prediction Standard Template

The second stage is the second generation early warning method. In 2006 ~ 2007, the project "Research on the National Geological Disaster Meteorological Early Warning and Forecasting Technology and Method" was initiated to carry out the research on upgrading the national geological disaster meteorological early warning and forecasting method. Liu Jiaoshou put forward three methods of regional early warning theory of geological disasters, namely implicit statistical prediction method, explicit statistical prediction method and dynamic prediction method; The design idea of explicit statistical early warning method (called the second generation early warning method) is put forward. This method improves the limitation that the first generation early warning method only depends on the critical process rainfall, and realizes the coupling between the critical process rainfall criterion and the spatial analysis of geological environment. Preliminary research results were obtained in 2007. After improvement, it was officially used in the national flood warning work in 2008.

According to the principle of regional early warning of geological disasters and the design idea of explicit early warning system, the specific early warning model establishment process is as follows:

(1) Geological hazard early warning zoning. The whole country is divided into seven early warning regions, and an early warning model is established in each region.

(2) Compilation of meteorological early warning information layer of geological disasters. Fully considering the basic information of geological environment and the historical occurrence of geological disasters, * * * compiled 30 early warning information layers.

(3) Calculation of geological hazards. This paper explores a calculation method for calculating the potential degree of geological disasters. According to the distribution of historical geological disasters, the uncertainty coefficient method is used to calculate the CF value of geological environment, and the weight determination method proposed by the project team is used to determine the weight, thus calculating the potential degree of geological disasters.

(4) Establishment of statistical early warning model. Divide the grid of 10km× 10km, with the potential degree of geological disasters, current rainfall and past rainfall at historical disaster points as input factors and the actual occurrence of geological disasters as output factors. The early warning index calculation model is established by using multiple linear regression method, so as to determine the early warning level.

Meteorological early warning system for landslide and debris flow in San Francisco Bay, USA

At present, the meteorological early warning of landslide and debris flow disasters in the world is mainly based on the critical rainfall threshold method proposed by the early warning system of landslide and debris flow in San Francisco Bay, USA. The system ran from 1985 to 1995 for 10 years, and was forced to shut down for various reasons. It is the longest-running landslide and debris flow early warning system in the world, and its experience is worth learning.

Campbell began to study the mechanism of landslide in Los Angeles from 1969. 1975, he put forward the idea of establishing an early warning system of debris flow in Los Angeles based on the rainfall forecast of the National Weather Service (NWS) and the radar image before Doppler. Campbell pointed out that debris flow prediction is still possible. By monitoring the intensity and duration of rainfall and comparing it with the critical value established according to the relationship between rainfall and landslide probability, the debris flow disaster level can be predicted. Once the critical value is exceeded, residents living at the foot of the mountain need to be warned and evacuated from dangerous places to minimize disaster losses. Campbell debris flow early warning system consists of the following aspects: ① rain gauge observation system, which records hourly rainfall; (2) Having a meteorological mapping system capable of identifying the rainfall intensity center in the rainstorm area; Draw rainfall data on topographic (slope) map and related landslide influence map; ③ Real-time data acquisition and early warning management and communication network.

1982 1 At the beginning of the year, disastrous rainstorms hit the San Francisco Bay Area, causing thousands of mudslides and other types of shallow landslides. Economic losses amounted to millions of dollars and 25 people died. Although people in this area were told about the rainstorm forecast, they didn't get any warnings about landslides and mudslides. Although Campbell's suggestion was not implemented in the San Francisco Bay Area, the catastrophic event of 1982 made it very urgent and necessary to establish a debris flow early warning system.

Figure 3-4 Critical Line of Debris Flow Rainfall in Lahonda, California

Cannon and Ellen( 1985) established the critical line of debris flow rainfall in La Honda, California (Figure 3-4). They revised (standardized) the critical rainfall duration and critical rainfall intensity with the annual average rainfall (MAP), that is, the critical rainfall intensity was revised to the critical rainfall intensity/annual average rainfall (MAP). The critical value of landslide rainfall established by them is the basis of the early warning system of debris flow in San Francisco Bay Area. 1February, 1986, there was a continuous rainstorm in the San Francisco Bay Area. The US Geological Survey and the National Weather Service jointly launched the early warning system of debris flow disaster, and issued two public early warnings through the NWS radio system. This is the first time that the United States has issued a mudslide disaster warning. Heavy rain triggered hundreds of mudslides in the San Francisco Bay Area, resulting in 1 death and property loss of100000 USD. If it is not the accurate prediction of the early warning system, the loss will be more serious.

1986 debris flow disaster warning was issued according to the critical value of empirical rainfall determined by Cannon and Ellen (1985). From 65438 to 0989, Wilson et al. established the relationship curve of cumulative rainfall/rainfall duration on the basis of empirical rainfall critical value, and determined different critical rainfall for debris flows with different scales and frequencies. USGS Landslide Working Group predicted the debris flow disaster accordingly.

Since 1995, Wilson has been studying the problem that the critical value of debris flow rainfall, which puzzles the early landslide warning system, is strongly influenced by local precipitation conditions (topographic effects).

As mentioned earlier, the critical rainfall value of debris flow in the San Francisco Bay Area established by Cannon( 1985) attempts to correct the influence of terrain effect with long-term rainfall (map). MAP is the most commonly used parameter to describe the climatic conditions of long-term rainfall, which can be obtained from standard meteorological maps. The map standardization critical value established by Cannon is the main technical basis of landslide early warning system. However, as Cannon himself said, during the operation of the early landslide warning system, it was found that the rainfall data of the warning system in areas with less rainfall would produce "false alarms", which reflected the inconsistency of map standardization in low-map areas. Later, Wilson (1997) applied the standardized map method of the San Francisco Bay Area to Southern California and the Pacific Northwest, which obviously led to the problem of underestimating or overestimating the critical rainfall value.

As a parameter, rainfall actually reflects the two comprehensive processes of rainstorm scale and frequency. The frequency of rainfall in the Pacific Northwest of the United States is high, but the rainfall is small every time, resulting in a large annual rainfall; In Southern California, however, the frequency of rainfall is low, but it is heavy every time, resulting in small annual rainfall. The standardized method of annual average rainfall should identify those "extreme" rainfall events, that is, the rainfall far exceeds those with high frequency but small rainfall. Therefore, the scale of a single rainstorm is more important than the rainfall frequency to estimate the critical value of debris flow rainfall.

The long-term climate action makes the slope itself reach a state of gravity balance, that is, the balance between slope infiltration, evaporation and surface drainage. This long-term equilibrium process may contain countless known and unknown mechanisms. Geotechnical engineering properties of slope soil, surface drainage rate, distribution of water network and native vegetation may all affect the local climate. Wilson re-examined the inconsistency of standardized critical value of annual average precipitation by daily rainfall scale-frequency analysis. In the San Francisco Bay Area with little annual rainfall, the critical rainfall value of debris flow is higher than that predicted by map standardization. Wilson put forward the reference critical value of debris flow rainfall, which is helpful to study the interaction between rainfall and surface drainage. Wilson's research shows that the recurrence rate of rainstorm once every five years can represent the best combination relationship between rainfall frequency and erosion rate. According to three different regions with obviously different rainfall climate patterns (Los Angeles, San Francisco Bay and Northwest Pacific), the historical rainfall data that triggered the fatal debris flow disaster were collected, and the relationship curve between the 24-hour peak rainfall that triggered large-scale debris flow in history and the reference rainfall value (the recurrence value of the five-year rainstorm) was established (Figure 3-5). The relationship curve can be used to estimate the critical rainfall value of debris flow. Compared with Cannon's map standardized critical rainfall value, it can estimate the critical rainfall value of a specific location (especially the mountainous area affected by terrain) through interpolation in a more reliable range.

Figure 3-5 Relationship between 24-hour rainstorm peak and large-scale debris flow triggered by history

Although 1995 closed the meteorological early warning system of landslides and mudslides in the San Francisco Bay Area, the research on the critical value of rainfall/mudslides has not stopped since 1995. These studies have deepened the understanding of the interaction among rainfall, hydrological conditions of slopes, meteorological conditions of long-term rainfall and slope stability, and will lay a good scientific foundation for landslide meteorological early warning in the San Francisco Bay Area and even in other parts of the world.

Three. Rainfall monitoring and forecasting

During the ten years when the landslide warning system in San Francisco Bay Area was in operation, the weather forecast of local NWS mainly relied on the meteorological satellite GOE-7 launched in February/987 (/KOOC-0/997 was replaced by GOES-/KOOC-0/0). Every 30 minutes, GOES meteorological satellite transmits images of clouds covering the west coast of North America from gulf of alaska to Hawaii. According to these images, the local NWS can estimate the speed, direction and intensity of the rainstorm. The infrared spectrum image in the image can also indicate the temperature of the cloud cluster, which is an important information for estimating the rainfall intensity. In addition, the ground meteorological station can obtain the data of air pressure, wind speed, temperature and rainfall, combined with the long-term weather trend forecast information provided by NWS National Meteorological Center. The local NWS weather forecast office comprehensively analyzes these data, prepares and provides quantitative weather forecast (QPT), and publishes weather forecasts for northern California and southern California for the next 24 hours twice a day.

Rainfall monitoring (warning) system can automatically collect high-intensity rainfall observation data from a long distance and transmit the data to the local real-time weather forecasting center. By 1995, the alarm system in San Francisco Bay Area has established 60 rainfall observation stations (Figure 3-6). Although the establishment of each site is supported by NWS, the purchase, installation and maintenance of equipment at each site is the responsibility of other federal, state and local government agencies. During the operation of 1985 to 1995 landslide warning system, USGS was responsible for maintaining the early warning receiver and data processing microcomputer system in Menlo Park, California.

To evaluate whether the upcoming rainstorm will cause debris flow disasters, we need to consider two critical values: ① cumulative rainfall in the early stage (that is, soil water content); ② Comprehensive analysis of the intensity and duration of approaching rainstorm. Therefore, USGS Landslide Working Group installed a shallow pressure gauge in La Honda research area and monitored the soil. If the pressure gauge first shows a strong response to the rainstorm, it is considered that the previous critical value has been reached. It usually takes several weeks from winter to the future to make the soil moisture exceed the previous critical value, and then we should always pay attention to whether the intensity and duration of heavy rain are enough to cause mudslides.

Figure 3-6 1992 San Francisco Bay landslide early warning rainfall monitoring system-early warning

Four, issued a mudslide disaster warning.

When the rainstorm begins, monitor the rainfall intensity and estimate the speed of the rainstorm front. Quantitative rainfall forecast (QPF) based on observed rainfall and local NWS; Compared with the established critical value of debris flow rainfall, the type and scale of debris flow disaster are determined. The staff of NWS and USGS participated in this stage, and issued three levels of debris flow disaster warning to the public: ① flood consultation in cities and streams; (2) Mountain torrent/debris flow monitoring; ③ Early warning of mountain torrents/mudslides. During the period from 1986 to 1995, early warnings of different levels of debris flow disasters were issued many times.

Verb (abbreviation of verb) abstract

The risk prediction of landslide and debris flow disaster mainly predicts the possibility of landslide and debris flow disaster in a certain area or a certain slope in the future through the analysis of disaster conditions, and delineates the influence scope and activity intensity of landslide and debris flow disaster. The hierarchy of landslide and debris flow disaster prediction index system is shown in Figure 3-7. According to the different research objects of landslide and debris flow disaster prediction, the index system of landslide and debris flow disaster prediction can be established from three research scales.

Figure 3-7 Hierarchy Diagram of Spatial Prediction Index System of Geological Hazards

Regional landslide and debris flow disaster risk prediction is to define the relative risk area of landslide and debris flow disaster by analyzing the aggregation and regularity of landslide and debris flow disaster in regional space, thus providing basis for national land planning, disaster reduction and prevention, disaster management and decision-making. Different prediction scales correspond to different investigation stages and research accuracy. The risk zoning of landslide and debris flow disaster corresponds to the feasibility study stage, so it is necessary to make a preliminary comprehensive evaluation of the zoning law of engineering geological conditions in the proposed development zone to determine the possibility and sensitivity of landslide and debris flow disaster. The submitted results are comprehensive zoning map of regional engineering geological conditions and geological disaster prediction zoning map.