Spatial gridding snow disaster early warning method, device and equipment and storage medium

Abstract

The invention belongs to the technical field of natural disasters, and provides a space gridding snow disaster early warning method and device, equipment and a storage medium. The method comprises the following steps: completing analysis of daily snow melting characteristics and laws by using hour-by-hour snow depth data, proposing accumulated snow melting indexes on this basis, screening sensitive factors and sensitive time periods of accumulated snow melting through analysis of a collaborative change relationship between snow depth and air temperature, ground temperature and the like, determining index factors of snow melting prediction, establishing a snow melting prediction model, and predicting the snow melting. Performing regression calculation on the snow depth of each grid point in the target area by using the snow melting prediction model based on the historical snow depth of accumulated snow and the gradual time meteorological data, extracting the unique slope and intercept parameter of each grid point, and constructing a space gridding model; and based on the space gridding model, snow disaster early warning is realized. The method provided by the invention solves the problems of inaccurate prediction and low spatial quantification degree caused by snow melting deficiency in the current snow disaster early warning process.

Description

Technical Field

[0001] The present application relates to the technical field of natural disaster prediction, and in particular to a spatial grid-based snow disaster warning method, device, equipment and storage medium. Background Art

[0002] In winter, snow accumulation is widespread, deep, and lasts for a long time. Excessive snow accumulation not only affects traffic roads and driving safety, but also often affects livestock feeding, causing snow disasters of varying degrees in a wide range of different fields. After snow accumulation occurs, the melting time and duration are key parameters for judging whether the snow disaster continues to worsen or is about to subside. Quantitative and accurate prediction of the snowmelt process is not only a basic technology for snow disaster monitoring and prediction, but also has important application needs in the field of disaster prevention and mitigation.

[0003] In addition, snow, as part of water resources, is an important source of water replenishment for rivers, lakes and ecological environment. Slowly melting snow water is conducive to the improvement of soil moisture conditions in the ecological environment, and promotes the restoration and protection of the ecological environment and the effectiveness of agricultural production. However, rapid melting of snow may cause snowmelt flood disasters. Therefore, the prediction of snow accumulation and its melting process is not only the actual demand for disaster prevention and control, the scientific understanding of natural laws, but also the key point of concern for various industries in ecological environment and social and economic development. The development and progress of snowmelt prediction technology has important technical promotion and scientific significance.

[0004] Existing snow disaster warning technology ideas can be divided into two main categories: one is mainly based on meteorological warnings of snow disasters based on weather and climate forecasts; the other is based on satellite remote sensing monitoring and evaluation of the current snow accumulation status to estimate future snow disaster scenarios or risks.

[0005] According to the forecast and warning method of future snowfall weather, whether a snow disaster will occur is mainly determined by the amount of snowfall in the future weather. It focuses on the forecast of weather scenarios and is weaker than the evaluation of snow accumulation conditions. It does not consider the melting of existing snow and the original surface snow stock. In addition, the forecast of future snowfall weather cannot provide grid-based quantitative forecast and warning. Snow disaster warning still uses county-level administrative areas as geographical units in spatial areas. Snow disaster warning based on satellite snow remote sensing monitoring focuses on monitoring and evaluating the current status of snow accumulation and risk identification, observing the existing snow accumulation and snow depth in the current area, but does not predict changes caused by melting under future weather and climate conditions. Snow disaster warning is mainly based on the current snow accumulation conditions and the duration of snow accumulation in the past.

[0006] The weather forecast method is based on the daily snow weather forecast for the next 1-7 days. It mainly predicts the amount of snow that may occur in the future, including the amount of snowfall (snow depth) and the area covered. There is a lack of judgment on the current amount and range of snow accumulation. Most importantly, it does not consider the amount of snow melting, and lacks the prediction of the melting process and changes of future snow accumulation, resulting in obvious errors in the judgment of the future snow accumulation range and snow depth. The snow disaster warning of this method is often based on medium- and long-term weather forecasts. The results of the snow disaster warning are based on days as time units, county-level administrative districts as basic geographical units, and one prediction result for each administrative district in space, which cannot achieve spatial grid-based quantitative prediction and warning.

[0007] Snow disaster warning based on satellite remote sensing monitoring focuses on the current and previous snow range and depth. According to the current monitored snow depth, area, and duration, it is judged whether a snow disaster is currently occurring. Although the snow disaster monitored by satellite remote sensing has grid quantitative capabilities, it does not take into account the changes in snow accumulation under future weather conditions.

[0008] The above two types of technical methods do not consider whether the existing snow will melt under future weather conditions, how much will be melted, how long the snow will last, etc., and the prediction models are established in days as time units, ignoring the daily changes of snow on an hourly basis. Therefore, the existing snow disaster warning technical methods lack the understanding of the daily changes of snow accumulation, the prediction of the current snowmelt process, and the quantitative prediction of the future surface snow cover and snow disaster warning based on spatial grids. Summary of the invention

[0009] The present application provides a spatial grid-based snow disaster warning method, device, equipment and storage medium. Based on the hourly daily change process of snow depth, a snow accumulation and snow disaster prediction and warning scheme that takes into account the snowmelt process is provided to solve the problems of inaccurate prediction and low spatial quantification caused by the lack of snowmelt in the current snow disaster warning process.

[0010] In a first aspect, the present application provides a spatial grid-based snow disaster warning method, comprising:

[0011] Determine snowmelt index based on daily snowmelt characteristics and snowmelt patterns;

[0012] Based on the coordinated variation characteristics of snowmelt and heat and the temperature sensitive period, the sensitive factors of snowmelt are screened out from the snowmelt indicators;

[0013] Ridge regression and stepwise linear regression were used to introduce the order of sensitive factors, the partial correlation coefficients of sensitive factors and their t-test significance, and the factors retained in the final equation. The snow depth at the set time was taken as the dependent variable, and the sensitive factors at different times with correlation significance at the 0.05 test level were taken as independent variables. The dominant influencing factors of snow depth change and their occurrence time were identified through the regression process and regression analysis results.

[0014] A snowmelt prediction model is constructed based on the dominant influencing factors. The snowmelt prediction model is expressed as:

[0015]

[0016] In the formula, Sd i is the predicted snow depth on day i, a is the model slope, is the temperature at 12:00 on the ith day, is the relative humidity at 20:00 on the i-th day, is the relative humidity at 21:00 on the i-th day, d is the difference between the simulated snow depth peak and the intercept, Sd 0 It is the latest measured initial snow depth;

[0017] Based on the historical snow depth and hourly meteorological data, the snowmelt prediction model is used to perform regression calculation on the snow depth of each grid point in the target area, extract the unique slope and intercept parameters of each grid point, and construct a spatial grid model;

[0018] Snow disaster warning is achieved based on the spatial grid model.

[0019] In a second aspect, the present application provides a spatial grid snow disaster warning device, comprising:

[0020] The ablation index determination module determines the snow ablation index based on the daily snowmelt characteristics and snowmelt rules;

[0021] A sensitive factor determination module, based on the coordinated change characteristics of snowmelt and heat and the temperature sensitive period, screens out the sensitive factors of snowmelt from the snowmelt indicators;

[0022] The dominant factor determination module is configured to use ridge regression and stepwise linear regression. The order of sensitive factors, partial correlation coefficients of sensitive factors and their t-test significance, and the factors retained in the final equation are introduced into the equation. The snow depth at the set time is used as the dependent variable, and the sensitive factors at different times with correlation significance at the 0.05 test level are used as independent variables. Through the regression process and regression analysis results, the dominant influencing factors of snow depth changes and their occurrence time are identified;

[0023] The prediction model building module is configured to build a snowmelt prediction model according to the dominant influencing factors, and the snowmelt prediction model is expressed as:

[0024]

[0025] In the formula, Sd i is the predicted snow depth on day i, a is the model slope, is the temperature at 12:00 on the ith day, is the relative humidity at 20:00 on the i-th day, is the relative humidity at 21:00 on the i-th day, d is the difference between the simulated snow depth peak and the intercept, Sd 0 It is the latest measured initial snow depth;

[0026] The grid model building module is configured to perform regression calculation on the snow depth of each grid point in the target area based on historical snow depth and hourly meteorological data, and to extract unique slope and intercept parameters of each grid point to build a spatial grid model by using the snowmelt prediction model;

[0027] The snow disaster warning module is configured to implement snow disaster warning based on the spatial grid model.

[0028] In a third aspect, the present application provides an electronic device, comprising: a processor and a memory; the memory is used to store instructions; the processor is used to execute the instructions in the memory, so that the electronic device executes the method described in the first aspect.

[0029] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the method described in the first aspect.

[0030] In a fifth aspect, the present application provides a computer program product, including a computer program, which implements the method described in the first aspect when executed by a processor.

[0031] The spatial grid-based snow disaster warning method, device, equipment and storage medium provided by the present application have the following beneficial effects:

[0032] 1) This application aims at the melting of snow under future meteorological conditions when snowfall occurs, and realizes the prediction of the duration of snow accumulation and the change in snow depth caused by melting. That is, by revealing the laws of the dynamic process of snowmelt, and the most important meteorological factors and key periods that affect snowmelt within 1 day, a three-variable linear prediction model of multi-factor snowmelt amount and duration based on temperature and relative humidity is proposed.

[0033] 2) Based on historical data and the dominant factors of the snowmelt process, this application constructs a snow disaster prediction model for each grid, realizes the independent calculation and prediction of the snow melt volume for each 10km×10km grid in a certain area, and realizes spatial gridded spatiotemporal quantitative snow disaster warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0035] Figure 1 This is a flow chart of a spatial grid snow disaster warning method shown as an exemplary embodiment;

[0036] Figure 2 Schematic diagram of daily snow depth variation in Nyalam and Purang in the late stage of snowmelt, shown in an exemplary embodiment, where: a, Nyalam; b, Purang;

[0037] Figure 3 Schematic diagram of snow depth at the end of snowmelt in Nyalam and Purang and related temperature sequence shown in an exemplary embodiment, wherein: a, snow depth at the end of snowmelt in Nyalam, b, related temperature sequence in Nyalam, c, snow depth at the end of snowmelt in Purang, d, related temperature sequence in Purang;

[0038] Figure 4 This is a graph showing the variation of the snow depth at the end of snowmelt in Nyalam and Purang and the average temperature at different times over multiple days, as shown in an exemplary embodiment; wherein: a, Nyalam, b, Purang;

[0039] Figure 5 This is a graph showing the correlation between the snow depth at 21:00 in Nyalam and Purang and the temperature 12 hours ahead of synchronization, as shown in an exemplary embodiment, where: a, Nyalam; b, Purang;

[0040] Figure 6 This is a correlation coefficient diagram between the snow depth at 21:00 during the snowmelt period in Nyalam and Purang in 2019 and different meteorological factors 12 hours ahead, shown as an exemplary embodiment;

[0041] Figure 7 A schematic diagram of the temperature at 12:00 on each grid of the spatial grid model of the test area shown as an exemplary embodiment;

[0042] Figure 8 A schematic diagram of relative humidity at 20:00 on a grid-by-grid basis in a spatial grid model of a test area shown as an exemplary embodiment;

[0043] Fig. 9 A schematic diagram of relative humidity at 21:00 on a grid-by-grid basis in a spatial grid model of a test area shown as an exemplary embodiment;

[0044] Fig.10 A schematic diagram of the slope of each grid of the spatial gridding model of the test area and the intercept of the prediction model is shown as an exemplary embodiment;

[0045] Fig.11 The actual snow depth map on March 19, 2024 is shown as an exemplary embodiment;

[0046] Fig.12 The model-predicted snow depth map for March 20, 2024 is shown as an exemplary embodiment;

[0047] Fig.13 The model predicted snow depth map for March 21, 2024 is shown as an exemplary embodiment;

[0048] Fig.14 The model predicted snow depth map for March 22, 2024 is shown as an exemplary embodiment;

[0049] Fig.15 The model-predicted snow depth map for March 23, 2024 is shown as an exemplary embodiment;

[0050] Fig.16 The model predicted snow depth map for March 24, 2024 is shown as an exemplary embodiment;

[0051] Fig.17 The model predicted snow depth map for March 25, 2024 is shown as an exemplary embodiment;

[0052] Fig.18 The model predicted snow depth map for March 26, 2024 is shown as an exemplary embodiment.

[0053] Fig.19 The figure is a schematic diagram of dynamic prediction of snow depth in Tibet based on snowmelt, showing an exemplary embodiment.

[0054] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0055] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0056] It should be noted that the links involved in this application and the related information in the links, platform-related information, etc. (including but not limited to data used for analysis, stored data, displayed data, etc.) are all information and data that are understood and authorized by the relevant users or fully authorized by all parties, and the collection, use, processing, transmission, provision, disclosure and application of the relevant data comply with the laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and comply with the principles of legality, legitimacy and necessity.

[0057] At present, the existing snow accumulation prediction methods mainly predict the amount of snow accumulation in a certain place based on the amount of future snowfall. They do not involve the prediction of the snow melting process, nor do they propose similar calculation models. In addition, the existing technologies generally have one prediction model for each administrative area, only one prediction value in an administrative area, or obtain grid analysis through interpolation. They do not use a single grid to independently establish a model and calculate each grid independently. There are also methods that use numerical weather forecasts to obtain gridded snow accumulation predictions for a certain area, but the prediction results of this method are only the amount of future snowfall, and ignore the actual amount of snow accumulation in the future.

[0058] Based on this, a spatial gridded snow disaster warning method, device, equipment, storage medium and computer program product are proposed in this embodiment. On the one hand, the duration of snow accumulation and the change in snow depth caused by melting are predicted for the melting of snow under future meteorological conditions under the condition of snowfall. That is, by revealing the law of the dynamic process of snowmelt, as well as the most important meteorological factors and key periods affecting snowmelt within 1 day, a ternary linear prediction model of multi-factor snowmelt amount and duration based on temperature and relative humidity is proposed. On the other hand, based on historical data and the dominant factors of the snowmelt process, a snow disaster prediction model for each grid is constructed, realizing the independent calculation and prediction of the snow melting amount of each 10km×10km grid in a certain area, and realizing spatial gridded spatiotemporal quantitative snow disaster warning.

[0059] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0060] The embodiment of the present application provides a spatial gridding snow disaster warning method. It should be noted that the spatial gridding snow disaster warning method provided by the present application can be implemented on any electronic device with data processing capabilities, or it can be a spatial gridding snow disaster warning system. It should be noted that the spatial gridding snow disaster warning system can be deployed separately on an electronic device in any environment (for example: separately deployed on an edge server in an edge environment), or it can be deployed entirely in a cloud environment, or it can be deployed in a distributed manner in different environments.

[0061] For example, the spatial grid snow disaster warning system can be logically divided into multiple parts, each with different functions. The various parts of the spatial grid snow disaster warning system can be deployed in any two or three of the electronic device (located on the user side, such as the client), the edge environment, and the cloud environment. The edge environment is an environment that includes a collection of edge electronic devices that are close to the electronic device. The edge electronic devices include: edge servers, edge stations with computing power, etc. The various parts of the spatial grid snow disaster warning system deployed in different environments or devices work together to realize the functions of the data processing platform.

[0062] It should be understood that this application does not restrictively divide which parts of the spatial grid snow disaster warning system are deployed in what specific environment. In actual application, it can be adaptively deployed according to the computing power of electronic devices, the resource possession of edge environment and cloud environment, or specific application requirements.

[0063] Figure 1 FIG. 1 is a flowchart of a spatial grid snow disaster warning method shown in an exemplary embodiment. Figure 1 As shown, the method includes steps S1 to S6, which are described in detail as follows:

[0064] S1: Determine snowmelt indicators based on daily snowmelt characteristics and snowmelt patterns.

[0065] In some embodiments, daily snowmelt characteristics include snow depth and daily snowmelt amount; wherein, the snow depth value at the first moment is used as the snow depth value for the day's cutoff snowmelt, and the difference between the snow depth value at the first moment and the snow depth value at the same moment of the previous day is used as the daily snowmelt amount, and the first moment is the starting moment when the snowmelt depth change value does not exceed the set threshold.

[0066] In some embodiments, the snow ablation index includes a terminal snow ablation index, which includes the snow depth and the ground temperature at 0 cm in a set period of time on the day. When the snow depth and the ground temperature at 0 cm in a set period of time on the day meet the first condition and the second condition respectively, it is determined that the snow will be completely melted within 2-3 days; wherein, the first condition is that the snow depth is less than 10 cm and the ground temperature at 0 cm in the set period of time on the day is greater than 0℃.

[0067] For example, take the Qinghai-Tibet Plateau region. Figure 2 FIG. 1 is a schematic diagram showing the daily snow depth variation in Nyalam and Purang in the late stage of snowmelt, according to an exemplary embodiment. Figure 2 As shown in the figure, the daily snowmelt process shows the basic characteristics of being flat first, then rapid and then slow. Specifically, when the temperature is high during the day, the snow depth decreases due to melting; when the temperature is low at night, the snow depth changes slowly and remains relatively stable. The daily variation of snow depth shows that the time point when the snow accumulation in Nyalam and Purang decreases significantly during the day generally occurs at 13:00, and 13:00-18:00 is the main period when the daily snow depth decreases rapidly. The snowmelt depth generally does not change significantly after 21:00. If the snow depth at this moment is taken as the snow depth value for the day's snowmelt, the difference between the snow depth at this moment and the same moment of the previous day can be regarded as the snowmelt amount for the day.

[0068] In the late stage of snow melting, in the last 2-3 days when the snow depth approaches 0 cm, the snow depth is generally less than 10 cm. When the snow depth is less than 10 cm, the rate of daily snow depth decreases significantly, and the snow generally melts completely within 2-3 days. Therefore, the snowmelt period of long-term snow accumulation in the Qinghai-Tibet Plateau can be roughly judged based on the 10 cm snow depth, that is, when the snow depth is ≤ 10 cm, the long-term snow may melt completely within the next 2-3 days.

[0069] The relationship between the average air temperature and 0cm ground temperature from 10:00 to 18:00 every day during the last 5 days of the melting period and the melting of snow was explored. Figure 3 As shown, from Figure 3 It can be found that on the second day of the final snowmelt period at Pulan Station, the daily maximum temperature had risen from below zero to above 0.0℃, and the snowmelt rate was significantly accelerated, from less than 3cm in the previous day to more than 4.5cm. This shows that the melting of snow at the end of the snowmelt period is directly related to the daytime temperature of the day. The higher the average temperature from 10 to 18 o'clock, the greater the daily snowmelt. The observation data of Nyalam also confirmed the correlation between the final snowmelt and the daytime temperature. Figure 3 Figures b and d show that the rapid melting of snow at the end of the snowmelt period is also significantly related to ground temperature. When the average 0cm ground temperature is >0℃ from 10 to 18 o'clock and the snow depth is <10cm, the snow cover enters the collapse stage and the snow will completely melt within 2-3 days. The above analysis shows that the judgment of the end of snowmelt in the southern Qinghai-Tibet Plateau can be based on the following two judgment criteria:

[0070] First condition: snow depth <10cm;

[0071] The second condition: 0cm ground temperature > 0℃ from 10:00 to 18:00 on the same day.

[0072] When the above two conditions are met, the snow may melt completely within 2-3 days.

[0073] S2: Based on the coordinated changes in snowmelt and heat and the temperature-sensitive periods, sensitive factors of snowmelt are screened out from the snowmelt indicators.

[0074] In some embodiments, the coordinated change characteristics of snowmelt and heat are the hourly change of multi-day average temperature and the hourly change of 0cm ground temperature, and the temperature sensitive period is 0-1h and 7-12h before the first moment: the snowmelt index also includes meteorological factors related to the snowmelt process, including hourly ground temperature, wind speed and relative humidity; based on the coordinated change characteristics of snowmelt and heat and the temperature sensitive period, the sensitive factors of snowmelt are screened out from the snowmelt index, including: according to the relationship between the change of snow depth and various meteorological factors at different time points, the meteorological factors whose correlation significance passes the significance level test are used as sensitive factors; the sensitive factors include hourly temperature Ta, W, relative humidity RH, 0cm ground temperature Ts, 20cm ground temperature T 20cm 、15cm ground temperature T 15cm 、10cm ground temperature T 10cm 、5cm ground temperature T 5cm .

[0075] For example, Figure 4 As shown, Figure 4 The figure is a schematic diagram of the change in snow depth and the average temperature at different times of the day at the end of snowmelt in Nyalam and Pulan. The snow depth in Nyalam and Pulan drops rapidly from 9:00 to 19:00 every day during the end of snowmelt. The hourly change in the average temperature (ΔTa) and the hourly change in the 0cm ground temperature (ΔTs) at the two stations over the day show that the hourly change in the air temperature and ground temperature during the period from 7:00 to 15:00 in Nyalam is positive, indicating the change characteristics of the continuous increase in temperature and the continuous accumulation of heat conditions during this period. The temperature change in Pulan also shows the same hourly change characteristics, but due to its western location, it shows a lag of 1 hour. In addition, the temperature changes of other ground temperatures at different depths are similar to those of the air temperature and 0cm ground temperature, except that the time periods with positive values ​​are more lagging. When the snow depth begins to decrease, the hourly changes in the temperature and 0cm ground temperature snowmelt factor quickly turn from negative to positive. The rapid and direct response characteristics between snow depth and temperature change intuitively demonstrate the heat accumulation conversion process, which directly corresponds to the decrease of snow depth. The time point when the snow depth decreases significantly corresponds to the period of significant positive temperature change. In addition, the above phenomenon occurs not only in the process of air temperature and 0cm ground temperature change, but also in ΔT 5cm , ΔT 10cm , ΔT 15cm , ΔT 20cm and ΔT 40cmThere is also a direct and obvious synergistic change in the temporal distribution characteristics. Although there is an obvious time lag in the correlation with snow depth due to the lag in the rise of ground temperature at different depths, the characteristics of this synergistic change relationship are still very obvious.

[0076] The above observation data show that the snow depth in most areas generally does not change significantly after 21:00, indicating that the daily snowmelt process has ended before this time. The snow depth at this time and the difference between it and the previous snow depth can indicate the amount of snowmelt on that day. Therefore, the snow depth data at this time can be regarded as the final snow depth data after snowmelt on that day. Since the snow depth at different times during the day is significantly correlated with the leading temperature, it is difficult to identify the key moment when the temperature affects the change of snow depth. Therefore, based on the above analysis results, 21:00 snow depth is used as the key moment for analyzing the relationship with temperature.

[0077] The correlation between the snow depth at 21:00 on each day during the rapid snowmelt period and the synchronous and leading 12-hour temperature is shown by taking the calculation results of Nyalam and Pulan in 2019 as an example. Figure 5 As shown. Figure 5 It can be seen that there is a significant correlation between the snow depth at 21:00 in Nyalam and the temperature at each time, but the significant correlation value appears 7-11 hours ahead; the correlation between the snow depth at 21:00 in Pulan and the temperature at each time is not as obvious as that in Nyalam, but it can still be clearly seen that the significant correlation value appears 0-1h and 7-12h ahead. This shows that the snow depth at 21:00 in different places has different sensitive periods to the temperature.

[0078] Combined with the above analysis, the snow depth at 21:00 every day can be regarded as the final value of the snow depth after snowmelt on that day, and the amount of snowmelt on that day can be judged by the heat conditions during the local snowmelt sensitive period.

[0079] In addition to air temperature, it is also necessary to consider meteorological factors that may be related to the snowmelt process, such as hourly ground temperature, wind speed, and relative humidity, in order to explore the relationship between snow depth changes and various meteorological factors at different time points, and to screen out sensitive factors. Dynamic snowmelt is a multi-factor synergistic process. In order to further discuss the relationship between snow depth changes and different meteorological factors at different times, and to screen out snowmelt sensitive factors, the correlation coefficients between the snow depth at 21:00 and its synchronous and leading factors of 12 hours were calculated. Figure 6 The figure shows the correlation coefficient between the snow depth at 21:00 during the snowmelt period in Nyalam and Purang in 2019 and different meteorological factors 12 hours ahead. Figure 6It can be found that the snow melting in Nyalam and Pulan is directly related to the synchronous and leading temperatures 12 hours ahead, while the correlation between ground temperature at different depths and snow depth changes is quite different. During the snowmelt process in Nyalam, the temperature at each moment in the leading 12 hours was significantly correlated with the snow depth, while in Pulan, the significance of this correlation was somewhat weakened, but overall, the temperature at each moment in the leading 12 hours was significantly correlated with the snow depth, and the correlation significance was P<0.05. Compared with the temperature, the correlation significance of the ground temperature at each depth level in Nyalam and Pulan and the snow depth at 21:00 is very different. In Nyalam, the snow depth at 21:00 has a significant effect on Ts from 11:00 to 21:00 and T from 16:00 to 18:00. 5cm The snowmelt process in Pulan showed different sensitivities to the ground temperature at different depths and time periods: it was very sensitive to the ground temperature at 40 cm during all periods, and the relevant significance passed the 0.001 significance level test. 20cm 、T 15cm 、T 10cm 、T 5cm , Ts, etc. show a more obvious correlation with the increase of depth and lead time.

[0080] from Figure 6 It can also be seen that the snowmelt process in Nyalam and Pulan is not sensitive to the relative wind speed that is synchronous and 12 hours ahead. This shows that the change of snow depth at Nyalam and Pulan stations has no direct connection with wind speed. The change of snow depth caused by wind-blown snow is not obvious in this area, or it is difficult to observe the influence of wind-blown snow due to factors such as the location of meteorological stations. In addition, the snow depth at 21:00 in Nyalam and Pulan also has a significant correlation with the relative humidity from 17:00 to 21:00.

[0081] S3: Ridge regression and stepwise linear regression were used to introduce the order of sensitive factors, the partial correlation coefficients of sensitive factors and their t-test significance, as well as the factors retained in the final equation. The snow depth at a set time was taken as the dependent variable, and the sensitive factors at different times with correlation significance at the 0.05 test level were taken as independent variables. The dominant influencing factors of snow depth changes and their occurrence times were identified through the regression process and regression analysis results.

[0082] In some embodiments, ridge regression and stepwise linear regression are used, and the order of sensitive factors, partial correlation coefficients of sensitive factors and their t-test significance, as well as the factors retained in the final equation are introduced into the equation. The snow depth at a set time is taken as the dependent variable, and the sensitive factors at different times with correlation significance passing the 0.05 test level are taken as independent variables. The dominant influencing factors of snow depth changes and their occurrence times are identified through the regression process and regression analysis results, including: adding superscripts to each sensitive factor to indicate the time, and using the sensitive factors with superscripts as independent variables; using ridge regression to make a ridgeline diagram, excluding independent variables with collinearity through the k value, and introducing the independent variables that have passed the multicollinearity test into the stepwise linear regression; in the stepwise linear regression, the sensitive factors with superscripts are introduced one by one, and when the stepwise regression equation reaches the best fit, the dominant influencing factors and their occurrence times are determined based on the introduced sensitive factors with superscripts.

[0083] For example, the sensitive factors obtained by the above screening are all heat factors, so there must be high autocorrelation and multicollinearity problems between the factors. The size of the linear correlation coefficient between a single factor and snow depth cannot fully determine the actual relationship between it and snowmelt. In order to screen out the dominant influencing factors of snow depth changes under the condition of eliminating multicollinearity, this project uses ridge regression and stepwise linear regression to introduce the order of factors into the equation, the partial correlation coefficient of the factors and the significance of the t test, and the factors retained in the final equation. The snow depth at 21:00 is the dependent variable, and the meteorological factors at different times with correlation significance tested at the 0.05 level are the independent variables. Through the regression process and regression analysis results, the dominant influencing factors of snow depth changes and their occurrence time are identified.

[0084] In order to distinguish different meteorological factors at different times, superscripts are added to the original meteorological factor identifiers to indicate the time, such as The temperature at 6:00 is That is, the ground temperature at 0 cm at 6:00. Before the stepwise regression, the ridge line diagram was first made using ridge regression, and the independent variables with collinearity were excluded by the k value. Finally, the independent variables that had passed the multicollinearity test were introduced into the stepwise linear regression.

[0085] Table 1 Stepwise regression process and related parameters of snow depth and sensitive factors at 21:00 during snowmelt in Nyalam and Purang in 2019

[0086]

[0087] Table 1 shows the results of stepwise regression calculation. Among all the potential meteorological factors affecting the sensitive period of snowmelt in Nyalam, the first step is to introduce This indicates that the final snow depth change in the region significantly responds to the temperature at 12:00 noon. The second and third steps introduce RH21 With RH 20 After that, the stepwise regression equation reaches the best fit, at which point R 2 The maximum value reached 0.825. The stepwise regression results show that the daily snowmelt in Nyalam is closely related to the temperature at noon at 12:00, the relative humidity at 21:00, and the relative humidity at 20:00. This factor is the dominant factor affecting and restricting the final snowmelt in Nyalam.

[0088] Among all the potential meteorological factors in the sensitive period of snowmelt in Pulan, the first step is to introduce This indicates that the final snow depth change in the area significantly responds to the 40 cm ground temperature at 21:00. The second and third steps introduced RH 17 and After that, the stepwise regression equation reaches the best fit, at which point R 2 The maximum value is 0.865, which indicates that the ground temperature at 40 cm at 21:00, the relative humidity at 17:00, and the air temperature at 20:00 are closely related, and these three factors are the dominant factors affecting and restricting the final snowmelt in the Pulan area.

[0089] The above analysis shows that in Nyalam and Purang areas, heat factor is the most important driving factor for snow melting, but with the change of longitude and latitude, different regions respond differently to heat factor. The air temperature at 12:00 noon is the most important driving factor for snow melting in Nyalam, while the ground temperature at 40cm at 21:00 is the most important driving factor for snow melting in Purang.

[0090] S4: Construct a snowmelt prediction model based on the dominant influencing factors.

[0091] In this embodiment, the snowmelt prediction model is expressed as:

[0092]

[0093] In the formula, Sd i is the predicted snow depth on day i, a is the model slope, is the temperature at 12:00 on the ith day, is the relative humidity at 20:00 on the i-th day, is the relative humidity at 21:00 on the i-th day, d is the difference between the simulated snow depth peak and the intercept, Sd 0 This is the latest measured initial snow depth.

[0094] Exemplarily, the dominant influencing factors of Nyalam and Pulan stations were used to construct snow depth prediction models, and the prediction models constructed by the two stations were tested. The snow depth prediction model of Nyalam station was better than that of Pulan station, so the temperature at 12:00, relative humidity at 20:00 and relative humidity at 21:00 of Nyalam station were used as independent variables to construct the station forecast model. Using a single station to construct the station forecast model simplifies the complexity of the model and provides a feasible and practical method for snow disaster risk assessment. In the process of model construction, the temperature at 12:00, relative humidity at 20:00 and relative humidity at 21:00 of each station during the snowmelt period in the past few years and the snow depth data at 21:00 of the same day were first collected, and the regression coefficients and intercepts were determined by multivariate linear regression analysis. These coefficients provide the contribution weights of each independent variable to the snow depth at the prediction time point. The regression model also introduces the daily actual snow depth Sd0 as a correction parameter to adjust the prediction and consider the impact of initial conditions on snow depth changes, ensuring that the model can adapt to the actual daily meteorological and snow depth conditions. The obtained model equation is shown in Equation (1).

[0095] The construction and application of the model allows the prediction of the dynamic changes of snow depth at different sites under specific weather conditions. By regularly updating the model parameters, the accuracy of the model can be maintained and new information brought by actual observation data can be reflected in a timely manner.

[0096] S5: Based on the historical snow depth and hourly meteorological data, the snowmelt prediction model is used to regress the snow depth of each grid point in the target area, extract the unique slope and intercept parameters of each grid point, and construct a spatial grid model.

[0097] For example, based on the ERA5 real-time historical snow depth and hourly meteorological data of 10km*10km grid from March 1 to 31, 2001 to 2023, Formula 1 is used to regress the snow depth of each grid point in Tibet, and finally extract the unique slope and intercept parameters of each grid point. Finally, a model for spatial gridded snow depth prediction covering characteristic areas is constructed. Based on historical data, the model can predict the changes in snow depth caused by changes in ambient temperature and relative humidity. The model parameters for spatial gridding in Tibet are shown in Figures 7 to 10 .

[0098] S6: Realize snow disaster warning based on spatial grid model.

[0099] In some embodiments, snow disaster warning is implemented based on a spatial grid model, including:

[0100] Using the 2m temperature forecast every three hours starting from 00:00 UTC and the relative humidity and snow depth forecast every three hours starting from 12:00 UTC, combined with the snowmelt prediction model constructed in formula (1) and the spatial gridding parameters, the snow depth in the future period is predicted by dividing the snow depth into five levels with 5 cm snow depth.

[0101] For example, the atmospheric state and forecast data for the next few hours to days provided by ECMWF, including snow depth, snow density, 2m temperature, relative humidity, etc., are used to calculate and predict the future snow depth grid by grid using formula (1). The 2m temperature forecast values ​​every three hours from 00:00 UTC and the relative humidity and snow depth forecast values ​​every three hours from 12:00 UTC provided by the ECMWF dataset are combined with the snowmelt prediction model constructed in formula 1 and Figures 7 to 10 The spatial gridding parameters obtained in the forecast are divided into five levels with a snow depth of 5 cm, and the snow depth for the week from March 20 to March 26, 2024 is predicted. The temporal resolution of ECMWF data is 3 hours, and the 2m temperature at 12:00, relative humidity at 20:00, and relative humidity at 21:00 are extracted every day during the forecast period. Fig.11 This is the actual snow depth map on March 19, 2024. Figures 12 to 18 This is the model predicted snow depth map from March 20, 2024 to March 26, 2024. Based on the snow depth numerical band of the prediction model within a certain range, the model predicted snow depth results are graded for business applications. It can be seen from the figure that the predicted snow depth is decreasing day by day, and the snow cover area is also decreasing day by day.

[0102] The spatial gridding model constructed according to this embodiment realizes the quantitative prediction of spatial gridding snow depth.

[0103] Referring to the national standard GB / T 20482-2017 "Grade of Snow Disaster in Pastoral Areas", assuming that the average height of winter grass in a certain area is 10 cm, based on the meteorological elements and their changing trends of future numerical weather prevention, and considering the two conditions of duration and snow depth, a snow disaster forecast and warning was conducted in the area. Fig.19 .

[0104] The present application embodiment provides a space grid snow disaster warning device, the space grid snow disaster warning device comprising:

[0105] The ablation index determination module determines the snow ablation index based on the daily snowmelt characteristics and snowmelt rules;

[0106] A sensitive factor determination module, based on the coordinated change characteristics of snowmelt and heat and the temperature sensitive period, screens out the sensitive factors of snowmelt from the snowmelt indicators;

[0107] The dominant factor determination module is configured to use ridge regression and stepwise linear regression. The order of sensitive factors, partial correlation coefficients of sensitive factors and their t-test significance, and the factors retained in the final equation are introduced into the equation. The snow depth at the set time is used as the dependent variable, and the sensitive factors at different times with correlation significance at the 0.05 test level are used as independent variables. Through the regression process and regression analysis results, the dominant influencing factors of snow depth changes and their occurrence time are identified;

[0108] The prediction model building module is configured to build a snowmelt prediction model according to the dominant influencing factors, and the snowmelt prediction model is expressed as:

[0109]

[0110] In the formula, Sd i is the predicted snow depth on day i, a is the model slope, is the temperature at 12:00 on the ith day, is the relative humidity at 20:00 on the i-th day, is the relative humidity at 21:00 on the i-th day, d is the difference between the simulated snow depth peak and the intercept, Sd 0 It is the latest measured initial snow depth;

[0111] The grid model building module is configured to perform regression calculation on the snow depth of each grid point in the target area based on historical snow depth and hourly meteorological data, and to extract unique slope and intercept parameters of each grid point to build a spatial grid model by using the snowmelt prediction model;

[0112] The snow disaster warning module is configured to implement snow disaster warning based on the spatial grid model.

[0113] The spatial grid-based snow disaster warning device provided in this embodiment can be used to execute the above-mentioned spatial grid-based snow disaster warning method. Its implementation principle and technical effects are similar, and this embodiment will not be repeated here.

[0114] An embodiment of the present application provides an electronic device, which may include: a processor and a memory, wherein the processor and the memory can communicate; exemplarily, the processor and the memory communicate via a communication bus, the memory is used to store computer execution instructions, and the processor is used to call the computer execution instructions in the memory to execute the spatial gridded snow disaster warning method shown in any of the above method embodiments.

[0115] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the present application may be directly implemented as being executed by a hardware processor, or may be implemented by a combination of hardware and software modules in the processor.

[0116] The present application provides a computer-readable storage medium, on which computer-executable instructions are stored; when the computer-executable instructions are executed by a processor, they are used to implement a spatial gridded snow disaster warning method as in any of the above-mentioned embodiments.

[0117] An embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the above-mentioned spatial grid snow disaster warning method is implemented.

[0118] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0119] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A spatial grid-based snow disaster warning method, characterized in that: include: Determine snowmelt index based on daily snowmelt characteristics and snowmelt patterns; Based on the coordinated variation characteristics of snowmelt and heat and the temperature sensitive period, the sensitive factors of snowmelt are screened out from the snowmelt indicators; Ridge regression and stepwise linear regression were used to introduce the order of sensitive factors, the partial correlation coefficients of sensitive factors and their t-test significance, and the factors retained in the final equation. The snow depth at the set time was taken as the dependent variable, and the sensitive factors at different times with correlation significance at the 0.05 test level were taken as independent variables. The dominant influencing factors of snow depth change and their occurrence time were identified through the regression process and regression analysis results. A snowmelt prediction model is constructed based on the dominant influencing factors. The snowmelt prediction model is expressed as: In the formula, Sd i is the predicted snow depth on day i, a is the model slope, is the temperature at 12:00 on the ith day, is the relative humidity at 20:00 on the i-th day, is the relative humidity at 21:00 on the i-th day, d is the difference between the simulated snow depth peak and the intercept, and Sd0 is the latest measured initial snow depth; Based on the historical snow depth and hourly meteorological data, the snowmelt prediction model is used to perform regression calculation on the snow depth of each grid point in the target area, extract the unique slope and intercept parameters of each grid point, and construct a spatial grid model; Snow disaster warning is achieved based on the spatial grid model.

2. The method according to claim 1, characterized in that: The daily snowmelt characteristics include snow depth and daily snowmelt amount; wherein the snow depth value at the first moment is used as the snow depth value for the day's snowmelt cutoff, and the difference between the snow depth value at the first moment and the snow depth value at the same moment of the previous day is used as the daily snowmelt amount, and the first moment is the starting moment when the snowmelt snow depth change value does not exceed the set threshold.

3. The method according to claim 1, characterized in that The snow melting index includes a terminal snow melting index, and the terminal snow melting index includes snow depth and 0 cm ground temperature during a set period of time on the day. When the snow depth and 0 cm ground temperature during a set period of time on the day meet a first condition and a second condition respectively, it is determined that the snow will be completely melted within 2-3 days; wherein, the first condition is that the snow depth is less than 10 cm and the 0 cm ground temperature during the set period of time on the day is greater than 0°C.

4. The method according to claim 3, characterized in that The snowmelt and heat coordinated change characteristics are the hourly change of multi-day average temperature and the hourly change of 0cm ground temperature, and the temperature sensitive period is 0-1h and 7-12h before the first moment: the snowmelt index also includes meteorological factors related to the snowmelt process, including hourly ground temperature, wind speed and relative humidity; Based on the coordinated change characteristics of snowmelt and heat and the temperature sensitive period, the sensitive factors of snowmelt are screened out from the snowmelt indicators, including: According to the relationship between snow depth change and various meteorological factors at different time points, the meteorological factors whose correlation significance passes the significance level test are taken as sensitive factors. The sensitive factors include hourly air temperature Ta, W, relative humidity RH, 0cm ground temperature Ts, 20cm ground temperature T 20cm 、15cm ground temperature T 15cm 、10cm ground temperature T 10cm 、5cm ground temperature T 5cm .

5. The method according to claim 1, characterized in that Ridge regression and stepwise linear regression were used to introduce the order of sensitive factors, the partial correlation coefficients of sensitive factors and their t-test significance, and the factors retained in the final equation. The snow depth at the set time was taken as the dependent variable, and the sensitive factors at different times with correlation significance at the 0.05 test level were taken as independent variables. Through the regression process and regression analysis results, the dominant influencing factors of snow depth changes and their occurrence times were identified, including: Add superscripts to each sensitive factor to indicate the time, and use the sensitive factors with superscripts as independent variables; Ridge regression was used to make a ridge line diagram, and the independent variables with collinearity were excluded by the k value. The independent variables that had passed the multicollinearity test were introduced into the stepwise linear regression. In stepwise linear regression, sensitive factors with superscripts are introduced one by one. When the stepwise regression equation reaches the best fit, the dominant influencing factors and their occurrence time are determined based on the introduced sensitive factors with superscripts.

6. The method according to claim 5, characterized in that Based on the spatial grid model, snow disaster warning is realized, including: Using the 2m temperature forecast every three hours starting from 00:00 UTC and the relative humidity and snow depth forecast every three hours starting from 12:00 UTC, combined with the snowmelt prediction model constructed in formula (1) and the spatial gridding parameters, the snow depth in the future period is predicted by dividing the snow depth into five levels with 5 cm snow depth.

7. A spatial grid snow disaster warning device, characterized in that: include: The ablation index determination module determines the snow ablation index based on the daily snowmelt characteristics and snowmelt rules; A sensitive factor determination module, based on the coordinated change characteristics of snowmelt and heat and the temperature sensitive period, screens out the sensitive factors of snowmelt from the snowmelt indicators; The dominant factor determination module is configured to use ridge regression and stepwise linear regression. The order of sensitive factors, partial correlation coefficients of sensitive factors and their t-test significance, and the factors retained in the final equation are introduced into the equation. The snow depth at the set time is used as the dependent variable, and the sensitive factors at different times with correlation significance at the 0.05 test level are used as independent variables. Through the regression process and regression analysis results, the dominant influencing factors of snow depth changes and their occurrence time are identified; The prediction model building module is configured to build a snowmelt prediction model according to the dominant influencing factors, and the snowmelt prediction model is expressed as: In the formula, Sd i is the predicted snow depth on day i, a is the model slope, is the temperature at 12:00 on the ith day, is the relative humidity at 20:00 on the i-th day, is the relative humidity at 21:00 on the i-th day, d is the difference between the simulated snow depth peak and the intercept, and Sd0 is the latest measured initial snow depth; The grid model building module is configured to perform regression calculation on the snow depth of each grid point in the target area based on historical snow depth and hourly meteorological data, and to extract unique slope and intercept parameters of each grid point to build a spatial grid model by using the snowmelt prediction model; The snow disaster warning module is configured to implement snow disaster warning based on the spatial grid model.

8. An electronic device, characterized in that: The electronic device comprises: a processor and a memory; the memory is used to store instructions; the processor is used to execute the instructions in the memory, so that the electronic device executes the method as claimed in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 6 when executed by a processor.

10. A computer program product, characterized in that The method comprises a computer program, which implements the method according to any one of claims 1 to 6 when being executed by a processor.

Images