Construction method and application of regional scale slope loose accumulation layer underground water level prediction model series

Abstract

The invention belongs to the technical field of geological disaster monitoring and early warning, and relates to a construction method and application of a regional scale slope loose accumulation layer underground water level prediction model series. According to the method, rainfall water supply intensity and saturation depth ratio are introduced, and an underground water saturation depth ratio prediction model of a slope loose accumulation layer of a monitoring point is established based on long-time sequence monitoring data; meanwhile, different types of slope units are considered, and an initial prediction model series of a corresponding drainage basin unit in the current early warning partition is established; correcting the initial prediction model series to obtain a corrected prediction model series of each drainage basin unit; and finally, a regional scale underground water level prediction model series is formed by taking the slope units as units, so that predicted underground water levels in related regions are obtained, and the problem that regional underground water level changes cannot be monitored timely and efficiently due to sparse mountain area underground water level monitoring stations is effectively avoided.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of geological disaster monitoring and early warning, and relates to a construction method and application of a series of regional scale slope loose accumulation layer groundwater level prediction models. BACKGROUND

[0002] Landslide is the most frequent and the most widely affected sudden geological disaster in China. Among them, the shallow landslide induced by rainfall has become the main disaster threatening the safety of people's life and property and social stability in mountainous areas due to its strong suddenness and easy group distribution characteristics.

[0003] The occurrence of shallow landslide is closely related to the saturation degree of slope loose accumulation layer. During the rainfall process, the dynamic fluctuation of groundwater level in slope loose accumulation layer is an indication of directly reflecting the saturation degree of slope loose accumulation layer, and the saturation degree of slope loose accumulation layer is directly related to the stability of slope. Therefore, accurately grasping the real-time groundwater level of slope loose accumulation layer in the regional scope is the core link to improve the accuracy of shallow landslide early warning and prediction.

[0004] However, due to the complex topographic environment in mountainous areas and high monitoring cost, the existing groundwater level monitoring means (such as borehole monitoring) is usually sparse in point distribution and limited in coverage, which is difficult to realize real-time monitoring of groundwater level in slope loose accumulation layer in the whole range of regional scale, resulting in a major technical bottleneck in the accurate regional landslide risk early warning based on groundwater level.

[0005] Therefore, the present application is proposed. SUMMARY

[0006] The present application aims to overcome the shortcomings of the prior art and provide a construction method and application of a series of regional scale slope loose accumulation layer groundwater level prediction models, which can realize indirect, efficient and real-time perception of the groundwater level of regional scale slope loose accumulation layer, and provide accurate and timely early warning and prediction for rainfall type shallow landslide.

[0007] To achieve the above-mentioned purpose, the present application provides the following technical solutions: On the one hand, the present application provides a construction method of a series of regional scale slope loose accumulation layer groundwater level prediction models, which takes the global coverage meteorological monitoring data as the input factor to predict the groundwater level fluctuation of each slope loose accumulation layer in the regional scale, so that the groundwater level prediction of slope loose accumulation layer realizes cluster prediction from point to plane. Specifically, the construction method comprises the following steps: Step 1, based on the long time series monitoring data of slope loose accumulation layer groundwater in the early warning subarea of the study area, a groundwater saturation depth ratio prediction model of slope loose accumulation layer of the monitoring point in the current study area is established; Step 2: Based on the groundwater saturation depth ratio prediction model, establish the initial model of the shady slope unit and the initial model of the sunny slope unit in each watershed unit within the warning zone. The initial models of the shady slope and the initial models of the sunny slope together constitute the series of initial prediction models for the corresponding watershed unit within the current warning zone. Step 3: Modify the initial prediction model series on a watershed unit basis to obtain the modified prediction model series for each watershed unit. Step 4: Using slope units as units, create a series of regional-scale groundwater level prediction models; details are as follows: For each slope unit within each watershed unit, the thickness of the loose deposit layer of each slope unit is input. Based on the saturation depth ratio model of the watershed unit where the current slope unit is located, a groundwater level prediction model for each slope unit is obtained, thereby forming a series of groundwater level prediction models for each slope unit at the regional scale.

[0008] Specifically, in step 1, the functional relationship of the groundwater saturation depth ratio prediction model is as follows: Formula (1) In formula (1), This is a predicted value for the groundwater saturation depth ratio. For the influence factor vector, , Rainfall water supply intensity ( ), The average temperature This is the average wind speed. This represents the average air humidity. For rainfall, Slope; This is a bias term, and its value can be calculated based on the model building method and sample data.

[0009] Specifically, step 2 further includes the following sub-steps: Step 2.1: First, divide each early warning zone into multiple watershed units, then divide each watershed unit into multiple slope units, and then divide each slope unit into shady slope units and sunny slope units according to slope aspect. Step 2.2: Use the groundwater saturation depth ratio prediction model as the initial prediction model for groundwater in each slope unit, and establish the initial models for the shady slope and the sunny slope in each watershed unit; Step 2.3: The initial model of the shady slope and the initial model of the sunny slope together constitute the series of initial prediction models for the corresponding watershed units within the current early warning zone.

[0010] Furthermore, in step 2.2, the functional relationships of the initial model for the shady slope and the initial model for the sunny slope are as follows: Formula (3) In formula (3), Number the watershed units, , The total number of watershed units within the study area. For shady slopes, For the sunny side, This is the influence factor vector, which has the same meaning as in formula (1). This is a bias term, and its value can be calculated based on the model building method and sample data. Further, in step 2.3, the expression for the saturation depth ratio of the initial prediction model series is as follows: Formula (4) In formula (4), This represents the total number of watershed units within the current study area.

[0011] Specifically, step 3, modifying the initial prediction model series, includes: If there are actual samples for the watershed unit, the initial prediction model series is evaluated and its parameters are adjusted using the actual samples to obtain the corrected prediction model series for the corresponding watershed unit. If there are no actual samples for a watershed unit, the initial prediction model series will be used as the corrected prediction model series for the corresponding watershed unit.

[0012] Furthermore, the initial prediction model series is evaluated and its parameters are adjusted using actual samples to obtain the corrected prediction model series for the corresponding watershed unit. The specific process is as follows: Step 3.1: Obtain actual samples, taking watershed units as the unit, and form actual samples of shady slopes and sunny slopes within the watershed unit respectively; Step 3.2: Input the model influencing factors corresponding to the actual samples of the shady slope one by one into the initial model of the shady slope to obtain the initial predicted values ​​of the shady slope, thus forming the predicted samples of the shady slope. The model influence factors corresponding to the actual samples of the sunny slope are input one by one into the initial model of the sunny slope to obtain the initial predicted values ​​of the sunny slope, thus forming the predicted sample set of the sunny slope. ; For the predicted sample size of the shady slope, Predict the number of samples for the sunny slope. , ; Step 3.3: Compare the initial prediction results with actual samples from sunny and shady slopes of the same watershed unit, evaluate the model's prediction performance, and adjust parameters until the prediction results are close to the actual results. Then, establish corrected models for shady and sunny slopes using the corrected model parameters. Formula (5) In formula (5), Number the watershed units, The total number of watershed units within study area N. For shady slopes, For the sunny side, This is the influence factor vector, which has the same meaning as in formula (1). This is a bias term, and its value can be calculated based on the model building method and sample data.

[0013] Specifically, in step 4, the functional relationship of the groundwater level prediction model for each slope unit and the expressions of the regional-scale groundwater level prediction model series are as follows: Formula (8) Formula (9) In formula (8), Number the watershed units, , The total number of watershed units within the study area. Number the slope units. It is a slope unit type. This is the influence factor vector, which has the same meaning as in formula (1); For the first Total thickness of shallow deposits in each slope unit; In formula (9), For the first Groundwater level depth of loose deposits in each slope unit , This represents the total number of slope units within the study area. This is a series of regional-scale groundwater level prediction models.

[0014] On the other hand, the present invention also provides a method for constructing a series of regional-scale slope loose deposit groundwater level prediction models as described in part or all of the above, and its application in the field of geological disasters (such as shallow landslides). Specifically, the series of groundwater level prediction models constructed using the technical solution provided by the present invention can be used for real-time prediction and assessment of regional groundwater levels on a standalone basis, or can be nested in a software system to predict and assess regional groundwater levels in real time by calling the model.

[0015] Compared with the prior art, the technical solution provided by the present invention has the following beneficial effects: Because existing direct monitoring data on loose deposits on slopes is extremely limited and requires dedicated monitoring equipment, resulting in high acquisition costs, comprehensive on-site monitoring in vast mountainous areas is not feasible. Furthermore, since the monitoring targets involve different slopes, each with varying slope gradients, soil thicknesses, and soil properties, they are independent of each other. Therefore, each slope requires a separate prediction model for real-time groundwater level assessment and prediction. The construction method provided by this invention uses meteorological monitoring data covering the entire region as input factors to predict groundwater level fluctuations in loose deposits on slopes at a regional scale. This enables cluster prediction of groundwater levels in loose deposits on slopes from point to area, achieving the assessment and prediction of regional groundwater levels from single regional meteorological precipitation data. Specifically: This invention introduces rainfall water supply intensity and saturation depth ratio to establish a groundwater saturation depth ratio prediction model for loose deposits on slopes at monitoring points based on long-term monitoring data. Simultaneously, considering different types of slope units, it establishes an initial prediction model series for corresponding watershed units within the current warning zone. The initial prediction model series is then modified to obtain a modified prediction model series for each watershed unit. Finally, using slope units as units, a regional-scale groundwater level prediction model series is formed, thereby obtaining the predicted groundwater level in the relevant area. This effectively avoids the problem of not being able to monitor regional groundwater level changes in a timely and efficient manner due to the sparse number of groundwater level monitoring stations in mountainous areas. Attached Figure Description

[0016] The accompanying drawings are incorporated in and form part of this specification, and together with the description serve to explain the principles of the invention.

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating the construction method of a series of regional-scale slope loose deposit groundwater level prediction models provided by this invention; Figure 2 This is a schematic diagram illustrating the division of early warning zones and watershed units within the research area provided by the present invention. Figure 3 A schematic diagram illustrating the principle of saturation depth ratio provided by this invention; Figure 4 This is a schematic diagram illustrating the division of watershed units and slope units in a certain early warning zone provided by the present invention. Detailed Implementation

[0019] Exemplary embodiments will now be described in detail. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples consistent with some aspects of the invention as detailed in the appended claims.

[0020] This invention provides a method for constructing a series of regional-scale groundwater level prediction models for loose deposits on slopes. Using comprehensive meteorological monitoring data as input, it predicts groundwater level fluctuations in loose deposits on slopes at the regional scale, enabling clustered prediction of groundwater levels from point to area. Furthermore, it achieves indirect, efficient, and real-time sensing of groundwater levels in regional-scale loose deposits on slopes, providing accurate and timely early warnings for rainfall-induced shallow landslides. It is important to emphasize that the groundwater level prediction model series constructed in this embodiment can be used for standalone real-time prediction and assessment of regional groundwater levels, or nested within a software system to evaluate regional groundwater levels in real time.

[0021] To enable those skilled in the art to better understand the technical solutions of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0022] Example This implementation provides a method for constructing a series of regional-scale slope loose deposit groundwater level prediction models, combined with... Figure 1 Specifically, it includes the following steps: Step 1: Based on long-term monitoring data of groundwater in loose deposits on slopes within the early warning zones of the study area, establish a prediction model for the groundwater saturation depth ratio of loose deposits on slopes at the current monitoring points in the study area; specifically, Combination Figure 2 The study area comprises four early warning zones (early warning zone 1, early warning zone 2, early warning zone 3, and early warning zone 4) and 18 watershed units. A long-term monitoring point is deployed within early warning zone 3 to collect long-term monitoring data (including groundwater level depth). average temperature Average wind speed Average air humidity Rainfall (and other influencing factors). By introducing rainfall water supply intensity and saturation depth ratio, a prediction model for the groundwater saturation depth ratio of the loose deposits on the slope at the monitoring points is established based on long-term monitoring data from monitoring points within warning zone 3: Formula (1) In formula (1), This is a predicted value for the groundwater saturation depth ratio. The number of training samples, For the influence factor vector, For rainfall water supply intensity, The average temperature This is the average wind speed. This represents the average air humidity. For rainfall, Slope; , As dual variables; It should be noted that long-term monitoring data is obtained by monitoring long-term monitoring points deployed in the early warning zones. Ideally, each slope unit in each early warning zone should have a long-term monitoring point. However, due to the complex terrain and high monitoring costs in mountainous areas, it is only necessary to ensure that at least one long-term monitoring point is deployed in at least one early warning zone. Furthermore, the groundwater saturation depth ratio is defined as the ratio of the saturation depth of the loose deposits on the slope to the total depth. (See [reference needed]). Figure 3 The calculation formula is as follows: Formula (2) In formula (2), The saturation depth ratio This represents the total depth of the loose deposits on the slope. The depth of the groundwater level in the loose deposits on the slope. This represents the saturation depth of the loose deposits on the slope.

[0023] Step 2: Based on the groundwater saturation depth ratio prediction model, establish initial prediction models for the groundwater saturation depth ratio of shady slope units (hereinafter referred to as "shady slope initial model") and sunny slope units (hereinafter referred to as "sunny slope initial model") within each watershed unit in the early warning zone. The shady slope initial model and the sunny slope initial model together constitute the initial prediction model series for the corresponding watershed unit in the current early warning zone; specifically including the following sub-steps: Step 2.1: First, divide each early warning zone into multiple watershed units, then divide each watershed unit into multiple slope units, and finally divide each slope unit into shaded slope units and sunny slope units according to slope aspect; see [link to relevant documentation]. Figure 4 Taking early warning zone 2 as an example, early warning zone 2 is divided into watershed unit C and watershed unit D. Watershed unit C is divided into 51 slope units. According to the slope aspect, the 51 slope units are divided into two categories: shady slope units and sunny slope units, namely 25 shady slope units and 26 sunny slope units. Step 2.2: Using the groundwater saturation depth ratio prediction model as the initial prediction model for groundwater in each slope unit, establish the initial model for the shady slope unit and the initial model for the sunny slope unit in each watershed unit; wherein, the functional relationships of the initial models for the shady slope and the sunny slope corresponding to watershed unit C are as follows: Formula (3-1) Formula (3-1), For shady slopes, For the sunny side, The number of training samples, The model architecture, model parameters, and selection of impact factors are the same as in step 1, which is the impact factor vector. The functional relationships of the initial models for the shady slope and the sunny slope corresponding to watershed unit D are as follows: Formula (3-2) In formula (3-2), the meanings of each parameter are the same as those in formula (3-1); Step 2.3: The initial model for the shady slope and the initial model for the sunny slope together constitute the initial prediction model series for the corresponding watershed unit within the current early warning zone, and their expressions are as follows: Formula (4) In formula (4), This represents the total number of watershed units within the current study area.

[0024] Step 3: Modify the initial prediction model series on a watershed unit basis to obtain a modified prediction model series for each watershed unit; specifically including: If actual samples are available for the watershed unit, the initial prediction model series can be evaluated and its parameters tuned using these actual samples (appropriate evaluation methods can be selected based on the model, such as hypothesis testing, goodness-of-fit testing, likelihood assessment, or R-squared). 2 Conventional methods, such as error index verification methods (ROC, etc.), are used to obtain a series of corrected prediction models for the corresponding watershed units. The specific process is as follows: Step 3.1: Obtain actual samples. Taking watershed units as the unit, actual samples of shady slopes and shady slopes are formed separately within the watershed unit. It should be noted that there are two ways to obtain actual samples: First, screen cases of rainfall-induced shallow landslides in the study area and invert the saturation depth ratio at the time of landslide occurrence through critical stability state; Second, screen the limit saturation depth ratio in the period when heavy rainfall triggers groundwater response based on monitoring data from other monitoring points (data with insufficient data such as short-term or discontinuous monitoring to directly establish a prediction model).

[0025] Specifically, in this embodiment, watershed unit C has two short-term monitoring points, one located on the sunny slope and the other on the shady slope. Rainfall data from the monitoring data is filtered, and the maximum saturation depth ratio of groundwater response triggered by each heavy rainfall is summarized. A total of 3251 data points of maximum saturation depth ratio from the shady slope monitoring points are summarized and divided into a training set of actual samples from the shady slope of watershed unit C. and test set A total of 1049 maximum saturation depth ratios were collected from monitoring points on sunny slopes, and these were compiled and divided into a training set of actual samples from sunny slopes for watershed unit C. and test set ; Step 3.2: Input the model influencing factors (daily rainfall, daily average temperature, daily average wind speed, daily air humidity, evaporation, and other available meteorological monitoring data) corresponding to the actual sample training set of the shady slope into the initial model of the shady slope to obtain the initial predicted value of the shady slope; input the model influencing factors (daily rainfall, daily average temperature, daily average wind speed, daily air humidity, evaporation, and other available meteorological monitoring data) corresponding to the actual sample of the sunny slope into the initial model of the sunny slope to obtain the initial predicted value of the sunny slope; Step 3.3: Evaluate the initial prediction results (initial predicted values ​​for sunny slopes and initial predicted values ​​for shady slopes) based on actual samples of sunny and shady slopes of watershed unit C, respectively. Adjust the parameters of the corresponding models (initial model for sunny slopes and initial model for shady slopes) according to the existing problems until the prediction results are close to the actual results. After evaluating the model performance using the test set, establish the corrected prediction model for the saturation depth ratio of shady slopes (hereinafter referred to as the "shady slope correction model") and the corrected prediction model for the saturation depth ratio of sunny slopes (hereinafter referred to as the "sunny slope correction model") of watershed unit C with the corrected model parameters. Step 3.4: Test and evaluate the performance of the shady slope correction model and the sunny slope correction model using the shady slope test set and the sunny slope test set respectively. If the prediction accuracy is not met, return to step 3.3 to adjust the parameters until the ideal prediction accuracy is achieved.

[0026] Similarly, using actual samples, the prediction model of watershed unit D is modified according to steps 3.1, 3.2 and 3.3 to obtain the modified model for the shaded slope and the modified model for the sunny slope of watershed unit D.

[0027] in: The functional relationships of the shady slope correction model and the sunny slope correction model for watershed unit C are as follows: Formula (5-1) Formula (5-2) In formula (5-1) and formula (5-2), For shady slopes, For the sunny side, The number of training samples, This is the impact factor vector; the selection of impact factors is the same as in step 1. The functional relationships of the shady slope correction model and the sunny slope correction model for watershed unit D are as follows: Formula (6-1) Formula (6-2) In formula (6-1) and formula (6-2), For shady slopes, For the sunny side, The number of training samples, This is the impact factor vector; the selection of impact factors is the same as in step 1. It should be noted that the modified models mentioned above have similar model structures to the initial models, but they are built on different datasets, and the specific parameters of each model are determined through evaluation and parameter tuning.

[0028] The aforementioned corrected models for shaded and sunny slopes of watershed units C and D together constitute the saturation depth ratio correction prediction model series for early warning zone 2, and their expressions are as follows: Formula (7) Repeat steps 2 and 3 to form a series of saturation depth ratio correction prediction models for other warning zones 1, 3, and 4. In this embodiment, since there are no actual samples in the watershed units of warning zone 4, if there are no actual samples in the watershed units, the initial saturation depth ratio prediction model is used as the saturation depth ratio correction prediction model for the corresponding watershed units.

[0029] Step 4: Using slope units as units, create a series of regional-scale groundwater level prediction models; details are as follows: There are currently 423 slope units in the study area. For each slope unit within each watershed unit, the thickness of the loose deposit layer of each slope unit is input. Based on the saturation depth ratio model of the watershed unit to which the current slope unit belongs, see [reference needed]. Figure 3 Formula (2) yields the groundwater level prediction model for each slope unit, thus forming a series of groundwater level prediction models for each slope unit at the regional scale. The functional relationship of the groundwater level depth prediction model for each slope unit and the expressions of the regional-scale groundwater level depth prediction model series are as follows: Formula (8) Formula (9) In formula (8), Number the watershed units, , The total number of watershed units within the study area. Number the slope units. It is a slope unit type. This is an impact factor vector; the selection of impact factors is the same as in step 1. For the first Total thickness of shallow deposits in each slope unit ;In formula (9), For the first The depth of groundwater level in the loose deposits of each slope unit. , This represents the total number of slope units within the study area. This is a series of regional-scale groundwater level prediction models.

[0030] In summary, the construction method provided by this invention establishes a single-point groundwater level prediction model based on the measured groundwater level at monitoring points located in the loose deposit layer of a slope. By utilizing timely acquired rainfall data and considering different types of slope units, the predicted groundwater level in the relevant area can be obtained through the regional-scale groundwater level prediction model series built by this invention. This simultaneously achieves cluster prediction from point to area and evaluation and prediction of regional groundwater level data from meteorological precipitation data in a single area. It effectively avoids the problem of not being able to monitor regional groundwater level changes in a timely and efficient manner due to the scarcity of groundwater level monitoring stations in mountainous areas, and can be widely applied in the field of monitoring geological disasters such as shallow landslides.

[0031] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention.

[0032] It should be understood that the present invention is not limited to the content already described above, and various modifications and changes can be made without departing from its scope. The scope of the present invention is limited only by the appended claims.

Claims

1. A method for constructing a series of regional-scale slope loose deposit groundwater level prediction models, characterized in that, Using meteorological monitoring data covering the entire region as input factors, the groundwater level fluctuations of loose deposits on slopes at the regional scale are predicted, enabling the prediction of groundwater levels in loose deposits on slopes to achieve cluster prediction from point to area.

2. The method for constructing the regional-scale slope loose deposit groundwater level prediction model series according to claim 1, characterized in that, Includes the following steps: Step 1: Based on the long-term monitoring data of groundwater in the loose deposits of slopes in the early warning zones within the study area, establish a prediction model for the groundwater saturation depth ratio of the loose deposits of slopes at the current monitoring points in the study area. Step 2: Based on the groundwater saturation depth ratio prediction model, establish the initial model of the shady slope unit and the initial model of the sunny slope unit in each watershed unit within the warning zone. The initial models of the shady slope and the initial models of the sunny slope together constitute the series of initial prediction models for the corresponding watershed unit within the current warning zone. Step 3: Modify the initial prediction model series on a watershed unit basis to obtain the modified prediction model series for each watershed unit. Step 4: Using slope units as units, create a series of regional-scale groundwater level prediction models; details are as follows: For each slope unit within each watershed unit, the thickness of the loose deposit layer of each slope unit is input. Based on the saturation depth ratio model of the watershed unit where the current slope unit is located, a groundwater level prediction model for each slope unit is obtained, thereby forming a series of groundwater level prediction models for each slope unit at the regional scale.

3. The method for constructing the regional-scale slope loose deposit groundwater level prediction model series according to claim 2, characterized in that, In step 1, the functional relationship of the groundwater saturation depth ratio prediction model is as follows: Official (1) In formula (1), This is a predicted value for the groundwater saturation depth ratio. For the influence factor vector, This is a bias term.

4. The method for constructing the regional-scale slope loose deposit groundwater level prediction model series according to claim 2, characterized in that, Step 2 also includes the following sub-steps: Step 2.1: First, divide each early warning zone into multiple watershed units, then divide each watershed unit into multiple slope units, and then divide each slope unit into shady slope units and sunny slope units according to slope aspect. Step 2.2: Use the groundwater saturation depth ratio prediction model as the initial prediction model for groundwater in each slope unit, and establish the initial models for the shady slope and the sunny slope in each watershed unit; Step 2.3: The initial model of the shady slope and the initial model of the sunny slope together constitute the series of initial prediction models for the corresponding watershed units within the current early warning zone.

5. The method for constructing the regional-scale slope loose deposit groundwater level prediction model series according to claim 4, characterized in that, In step 2.2, the functional relationships of the initial model for the shady slope and the initial model for the sunny slope are as follows: Official (3) In formula (3), Number the watershed units, , The total number of watershed units within the study area. For shady slopes, For the sunny side, For the influence factor vector, This is a bias term.

6. The method for constructing the regional-scale slope loose deposit groundwater level prediction model series according to claim 4, characterized in that, In step 2.3, the expression for the saturation depth ratio of the initial prediction model series is as follows: Official (4) In formula (4), This represents the total number of watershed units within the current study area.

7. The method for constructing the regional-scale slope loose deposit groundwater level prediction model series according to claim 2, characterized in that, Step 3, specifically revising the initial prediction model series, includes: If there are actual samples for the watershed unit, the initial prediction model series is evaluated and its parameters are adjusted using the actual samples to obtain the corrected prediction model series for the corresponding watershed unit. If there are no actual samples for a watershed unit, the initial prediction model series will be used as the corrected prediction model series for the corresponding watershed unit.

8. The method for constructing the regional-scale slope loose deposit groundwater level prediction model series according to claim 7, characterized in that, The initial prediction model series was evaluated and its parameters were adjusted using actual samples to obtain the corrected prediction model series for the corresponding watershed unit. The specific process is as follows: Step 3.1: Obtain actual samples, taking watershed units as the unit, and form actual samples of shady slopes and sunny slopes within the watershed unit respectively; Step 3.2: Input the model influence factors corresponding to the actual samples of the shady slope one by one into the initial model of the shady slope to obtain the initial predicted values ​​of the shady slope, thus forming the predicted samples of the shady slope. ; The model influencing factors corresponding to the actual samples of the sunny slope are input one by one into the initial model of the sunny slope to obtain the initial predicted values ​​of the sunny slope, forming a predicted sample set of the sunny slope. ; For the predicted sample size of the shady slope, Predict the number of samples for the sunny slope. , ; Step 3.3: Compare the initial prediction results with actual samples from sunny and shady slopes of the same watershed unit, evaluate the model's prediction performance, and adjust parameters until the prediction results are close to the actual results. Then, establish corrected models for shady and sunny slopes using the corrected model parameters. Official (5) In formula (5), Number the watershed units, The total number of watershed units within study area N. For shady slopes, For the sunny side, For the influence factor vector, This is a bias term.

9. The method for constructing the regional-scale slope loose deposit groundwater level prediction model series according to claim 2, characterized in that, In step 4, the functional relationship of the groundwater level prediction model for each slope unit and the expressions of the regional-scale groundwater level prediction model series are as follows: Official (8) Official (9) In formula (8), Number the watershed units, , The total number of watershed units within the study area. Number the slope units. It is a slope unit type. This is an impact factor vector; For the first Total thickness of shallow deposits in each slope unit; In formula (9), For the first Groundwater level depth of loose deposits in each slope unit , This represents the total number of slope units within the study area. This is a series of regional-scale groundwater level prediction models.

10. The application of a method for constructing a series of regional-scale slope loose deposit groundwater level prediction models as described in any one of claims 1 to 9 in the field of geological hazards, characterized in that, The geological hazards mentioned include shallow landslides.

Images