Tropical cyclone water vapor mass change calculation and evaluation method based on machine learning and fused with accurate water vapor pressure differential equation, medium and program product

Abstract

The invention discloses a tropical cyclone water vapor quality change calculation and evaluation method based on machine learning and fused with an accurate water vapor pressure differential equation, a medium and a program product, and belongs to the technical field of meteorological science and numerical calculation. Comprising the steps of constructing an accurate water vapor pressure differential equation considering the specific humidity nonlinear effect, establishing a unit area air column water vapor mass change calculation model, obtaining meteorological data of a tropical cyclone region, dynamically tracking the tropical cyclone and determining a calculation region. And constructing a multilayer convolutional neural network model about tropical cyclone water vapor quality calculation and evaluation, performing layered calculation and evaluation on water vapor quality changes of different height layers, and outputting a calculation and evaluation result of the water vapor quality changes. According to the method, by introducing the specific humidity nonlinear effect, the heterogeneity of water vapor distribution in the atmosphere can be reflected more accurately, so that the precision of water vapor mass change calculation is improved, and key technical support is provided for tropical cyclone path prediction, intensity simulation and numerical mode optimization.

Description

Technical Field

[0001] The present invention belongs to the field of meteorological science and numerical computing technology, and relates to the evaluation and analysis of tropical cyclone water vapor mass changes and numerical computing technology. Specifically, it is a tropical cyclone water vapor mass change calculation and evaluation method, medium and program product based on machine learning and integrated with accurate water vapor pressure differential equations, which can be used for quantitative analysis of the spatiotemporal evolution characteristics of tropical cyclone water vapor content and simulation and prediction of intensity changes. Background Art

[0002] Tropical cyclone (TC) is an extreme phenomenon in the global weather system, and its development and evolution are of great significance to the study of atmospheric dynamics and thermodynamics. The formation, evolution and extinction of tropical cyclones are closely related to the distribution and changes of atmospheric humidity. The latent heat release of water vapor phase change is the main energy source for the development of tropical cyclone intensity, while the horizontal and vertical transport of humidity field directly affects the intensity change of tropical cyclones and the internal water vapor redistribution. Therefore, accurately calculating and evaluating the changes in tropical cyclone water vapor quality is of great significance for revealing the mechanism of its intensity change, improving the simulation accuracy of numerical models, and improving the ability to forecast tropical cyclone intensity.

[0003] Atmospheric humidity is an important physical quantity that characterizes the water vapor content in the atmosphere. Its commonly used measurement indicators include water vapor pressure, water vapor density, mixing ratio, specific humidity, relative humidity and dew point temperature. Among them, water vapor pressure is the weight of water vapor in a unit area of ​​air column, which directly reflects the water content of the atmosphere. Specific humidity indicates the mass of water vapor in a unit mass of wet air and is a key parameter to describe the characteristics of wet air. Studies have shown that the rapid intensification of tropical cyclones is often accompanied by a significant increase in low-level humidity and a significant decrease in high-level humidity. This humidity distribution characteristic directly affects the latent heat release efficiency and intensity development process of the inner core of the cyclone. Under the background of global warming, the atmospheric water holding capacity continues to increase, the atmospheric humidity is on the rise, and the TC intensity is also showing an increasing trend. This correlation has been widely recognized on the climate scale. However, on the weather scale, the mechanism of the influence of atmospheric humidity on TC intensity is quite complex. Humidity and its changes at different heights, locations, and degrees will cause different responses in TC structure and intensity. Current studies have shown that the increase in humidity in the inner core of TC is conducive to its rapid intensification, but the response of TC intensity to atmospheric humidity is not linear, and there is a complex nonlinear feedback mechanism.

[0004] The calculation of water vapor mass change is the core link in humidity field analysis. The existing water vapor pressure differential equation is usually used to analyze water vapor mass change. of the form, in which p v is the water vapor pressure, p is the ambient atmospheric pressure, qv is the atmospheric specific humidity, t is time. This equation completely ignores the nonlinear effect of atmospheric specific humidity and cannot accurately reflect the non-uniform characteristics of water vapor distribution in the atmosphere. Specifically, the above equation only considers the zero-order term of specific humidity, while ignoring the contribution of first-order and higher-order nonlinear terms. This simplification leads to a large deviation between the equation calculation results and actual observations, especially in the critical stage of TC development. This deviation may affect the accurate assessment of TC intensity changes. For example, the humidification rate of water vapor is usually overestimated in the low layer, while the drying rate of water vapor may be underestimated in the middle and high layers.

[0005] In summary, the existing water vapor pressure differential equation has problems such as insufficient calculation accuracy, ignoring the non-uniformity of water vapor distribution, and not considering the nonlinear effect of specific humidity when calculating the change of TC water vapor mass. Therefore, it is a technical problem that needs to be solved urgently to establish a calculation and evaluation method for TC water vapor mass change that takes into account the nonlinear effect of specific humidity and can provide higher calculation accuracy, so as to more accurately calculate and evaluate the change of water vapor mass in the tropical cyclone area, and thus have a deeper understanding of the impact of water vapor on the change of tropical cyclone intensity. Summary of the invention

[0006] 1. Purpose of the invention Abstract: Most of the existing tropical cyclone water vapor mass change calculation methods are based on simplified water vapor pressure differential equations, ignoring the nonlinear effect of specific humidity, resulting in insufficient calculation accuracy and difficulty in accurately evaluating the impact of water vapor on tropical cyclone intensity changes. In order to solve at least one of the above and other technical problems in the prior art, the present invention aims to provide a tropical cyclone water vapor mass change calculation and evaluation method, medium and program product based on machine learning and integrated with accurate water vapor pressure differential equations. By introducing accurate water vapor pressure differential equations considering the nonlinear effect of specific humidity and combining dynamic tracking analysis technology, high-precision calculation and evaluation of water vapor mass changes at different vertical levels during the development of tropical cyclones can be achieved, and the accuracy of humidity change calculation and the adaptability to complex weather systems can be significantly improved, thereby providing key technical support for the intensity prediction, path simulation and development of related numerical models of tropical cyclones.

[0007] (II) Technical solution In order to achieve the purpose of the invention and solve the technical problems, the present invention adopts the following technical solutions: The first invention object of the present invention is to provide a method for calculating and evaluating the change of water vapor mass in a tropical cyclone based on machine learning and integrating an accurate water vapor pressure differential equation, which is used to calculate and evaluate the change of water vapor mass at different vertical levels during the development of a tropical cyclone, and to improve the calculation accuracy of the non-uniformity of humidity distribution and the law of change. The method comprises at least the following steps when implemented: SS1. Establish an accurate water vapor pressure differential equation considering the nonlinear effect of specific humidity Combining the physical characteristics of the atmosphere and the relationship between specific humidity and water vapor pressure, the nonlinear term of specific humidity is introduced, and an accurate water vapor pressure differential equation including the nonlinear effect of specific humidity is established: in, p is the ambient atmospheric pressure, t For time, p v is the water vapor pressure, q v is the atmospheric specific humidity, γ v is the nonlinear coefficient of specific humidity, dp / dt is the rate of change of air pressure per unit time, dp v / dt is the rate of change of water vapor pressure per unit time, dq v / dt is the rate of change of specific humidity per unit time; SS2. Construct a calculation model for the change in water vapor mass per unit area of ​​air column Substitute the precise water vapor pressure differential equation established in step SS1 into the calculation formula for the water vapor mass per unit area of ​​air column: China and Israel build a calculation model for the change of water vapor mass per unit area of ​​air column , and in the constructed calculation model, the nonlinear term of specific humidity is expanded into an infinite series form , in order to fully consider the nonlinear effect of specific humidity on the water vapor distribution and improve the calculation accuracy of water vapor mass change to the nonlinear order, where m v is the water vapor mass per unit area of ​​the air column, g is the acceleration due to gravity, p b and p t are the bottom and top pressures per unit area of ​​the air column, k is the order of the infinite series expansion; SS3. Dynamically track tropical cyclones and determine calculation areas Obtain meteorological data for the tropical cyclone area to be evaluated and calculated, determine the intensity and position of the tropical cyclone center at each time according to the lowest average sea level pressure value and its position at each time in the meteorological data, dynamically track the tropical cyclone path, establish a dynamic coordinate system based on the tropical cyclone center position, use the tropical cyclone center position at each time as the origin of the dynamic coordinate system and the center of the dynamic calculation area, and map relevant meteorological physical quantities to the dynamic coordinate system. By adjusting the scope and position of the calculation area in real time, it is precisely matched with the moving path of the tropical cyclone center, so as to realize dynamic tracking of tropical cyclones and ensure the spatiotemporal calculation and evaluation accuracy of water vapor mass changes; SS4. Constructing a multi-layer convolutional neural network model for tropical cyclone water vapor quality calculation and assessment Taking the calculation model of the change of water vapor mass per unit area of ​​air column constructed in step SS2 as a benchmark, a multi-layer convolutional neural network model for calculating and evaluating the water vapor mass of tropical cyclones is constructed by using a machine learning algorithm, wherein the multi-layer convolutional neural network model includes an input layer, a hidden layer, and an output layer, wherein the model input layer is used to receive the meteorological data of each altitude layer in the tropical cyclone region in a dynamic coordinate system and the water vapor pressure change rate calculated according to the accurate water vapor pressure differential equation, the hidden layer uses multiple convolutional layers and pooling layers to extract the multidimensional features of the tropical cyclone water vapor distribution and uses a fully connected layer for feature fusion, and introduces an attention mechanism module to capture the correlation between the water vapor mass changes at different altitudes, and the output layer uses the water vapor mass changes calculated according to the calculation model results in step SS2 as the label of the training data, and respectively calculates and evaluates the water vapor mass changes per unit area of ​​the air column at each altitude; SS5. Output the calculation and evaluation results of water vapor mass change The multi-layer convolutional neural network model for tropical cyclone water vapor mass calculation and evaluation constructed in step SS4 is used, combined with the meteorological data of the calculation area and each altitude level dynamically determined in step SS3, to generate the spatiotemporal distribution data of the changes in water vapor mass at different altitude levels of the tropical cyclone.

[0008] The second inventive object of the present invention is to provide a computer program product, comprising computer instructions, to execute the above-mentioned method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating the accurate water vapor pressure differential equation.

[0009] The third inventive object of the present invention is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating an accurate water vapor pressure differential equation.

[0010] (III) Technical Effect Compared with the prior art, the tropical cyclone water vapor mass change calculation and evaluation method, medium and program product of the present invention based on machine learning and integrating accurate water vapor pressure differential equation have the following beneficial and significant technical effects: (1) The present invention significantly improves the calculation accuracy of water vapor mass changes by constructing an accurate water vapor pressure differential equation and introducing the nonlinear effect of specific humidity. The traditional water vapor pressure differential equation fails to fully consider the nonlinear characteristics of specific humidity. Based on the relationship between specific humidity and water vapor pressure, the present invention constructs an accurate water vapor pressure differential equation containing nonlinear terms of specific humidity, and uses an infinite series expansion method to accurately characterize the contribution of specific humidity to the non-uniformity of water vapor distribution. Compared with traditional methods, the present invention can more accurately reflect the real change characteristics of low-level humidification and mid- and high-level drying, especially during the development of tropical cyclones. The introduction of nonlinear terms reduces the calculation error and significantly reduces the average residual of water vapor mass changes.

[0011] (2) The present invention establishes a calculation model for the change in water vapor mass per unit area of ​​air column and combines it with an accurate water vapor pressure differential equation to achieve a hierarchical calculation and evaluation of the change in water vapor mass at different altitudes of tropical cyclones. By combining the correction calculation of the accurate equation with the physical hierarchy division, the present invention can calculate the change in water vapor mass per unit area of ​​air column at the low, middle and high levels, respectively, and clarifies the hierarchical characteristics of water vapor distribution. In particular, the significant humidification rate at the low level and the drying rate at the middle and high levels are accurately quantified, providing a scientific basis for studying the impact of humidity distribution on the change in tropical cyclone intensity.

[0012] (3) The present invention dynamically tracks the path of tropical cyclones and establishes a dynamic coordinate system, thereby achieving real-time matching between the calculation area and the center position of the tropical cyclone. Traditional methods mostly use fixed-area calculations, which cannot accurately reflect the spatiotemporal dynamic characteristics of the humidity field of tropical cyclones, a weather system with significant mobility and non-uniformity. The present invention uses a dynamic tracking method to take the center position of the tropical cyclone as the origin of the dynamic coordinate system, and adjusts the scope and position of the calculation area in real time, so that the diagnostic results of water vapor quality changes can more accurately reflect the actual development process of the tropical cyclone. At the same time, the dynamic tracking method effectively solves the problem of distortion of humidity field diagnostic results and significantly improves the spatiotemporal adaptability of calculation evaluation.

[0013] (4) This paper further quantifies the regulatory effect of the nonlinear effect of specific humidity on the change of water vapor mass, providing a new perspective for studying the impact of humidity on the change of tropical cyclone intensity. By introducing the nonlinear term of specific humidity, this paper finds that the nonlinear effect of specific humidity can slow down the humidification rate in the low layer and the drying rate in the middle and high layers, and promote the concentration of water vapor to higher layers. This not only reveals the complex nonlinear law of humidity distribution, but also provides technical support for the numerical simulation and model optimization of humidity fields.

[0014] (5) The present invention constructs a multi-layer convolutional neural network model based on machine learning and combines it with precise water vapor pressure differential equations and dynamic tracking technology to effectively capture the complex nonlinear relationships and spatiotemporal variation characteristics in meteorological data. Compared with traditional calculation methods based on physical models, the present invention shows higher accuracy and generalization ability in predicting water vapor mass changes. It can more accurately predict the changes in water vapor mass of tropical cyclones at different development stages and different intensities, and provides more reliable technical support for the study of the mechanism of tropical cyclone intensity change, path simulation and numerical model optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. The embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein: Figure 1 Shown is an implementation flow chart of a method for calculating and evaluating changes in water vapor mass in a tropical cyclone based on machine learning and integrating an accurate water vapor pressure differential equation provided by an embodiment of the present invention; Figure 2 The diagram shows the change of (a) the central movement path and (b) the central mean sea level pressure and its 12-hour pressure variation of the tropical cyclone "Dusurui" from 00:00 on July 21 to 00:00 on July 29, 2023, where: (a) the blue, red and black dotted lines in the figure are the development, intensification and maturity stages of "Dusurui", respectively; (b) the blue and red dashed lines in the figure are the boundaries of the development and intensification stages and the boundaries of the intensification and maturity stages of "Dusurui", respectively, and the red dotted line is the central mean sea level pressure, unit: hPa, and the gray column is the 12-hour pressure variation, unit: hPa; Figure 3 Shown are the individual changes in the average unit area of ​​water vapor mass per air column of Tropical Cyclone Dusur in the lower, middle, and upper layers calculated based on (a) the exact water vapor pressure differential equation and (b) the first-order nonlinear term of specific humidity (unit: 10 -4 kg / (m 2 ·s)) and the 12-hour pressure variation of the mean sea level pressure at the center of Dusurui (grey column, unit: hPa), in which the black dotted line is the low layer, the red dotted line is the middle layer, and the blue dotted line is the high layer; Figure 4 Shown are the percentages (%) of the residuals of the precise water vapor pressure differential equation (solid line) and the commonly used water vapor pressure differential equation (dashed line) in the low (black dotted line), middle (red dotted line) and high (blue dotted line) layers of "Dusu Rui", as well as a schematic diagram of the 12-hour pressure variation of the average sea level pressure at the center of "Dusu Rui" (gray column, unit: hPa). DETAILED DESCRIPTION

[0016] In order to make the purpose, technical scheme and advantages of the implementation of the present invention clearer, the technical scheme in the embodiment of the present invention will be described in more detail below in conjunction with the drawings in the embodiment of the present invention. The described embodiments are part of the embodiments of the present invention, not all of the embodiments, and the described embodiments are exemplary and are intended to be used to explain the present invention, and cannot be understood as limiting the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0017] The present invention aims to provide a method, medium and program product for calculating and evaluating the change of water vapor mass in tropical cyclones based on machine learning and integrating the precise water vapor pressure differential equation. By introducing the precise water vapor pressure differential equation that takes into account the nonlinear effect of specific humidity and combining it with dynamic tracking analysis technology, high-precision calculation and evaluation of the change of water vapor mass at different vertical levels during the development of tropical cyclones can be achieved, which significantly improves the accuracy of humidity change calculation and the adaptability to complex weather systems, thereby providing key technical support for the intensity prediction, path simulation and development of related numerical models of tropical cyclones.

[0018] Example 1 As a specific example, Figure 1 As shown, the tropical cyclone water vapor mass change calculation and evaluation method based on machine learning and integrating the accurate water vapor pressure differential equation of the present invention includes the following steps: SS1. Establish an accurate water vapor pressure differential equation considering the nonlinear effect of specific humidity Combining the physical characteristics of the atmosphere and the relationship between specific humidity and water vapor pressure, the nonlinear term of specific humidity is introduced, and an accurate water vapor pressure differential equation including the nonlinear effect of specific humidity is established: in, p is the ambient atmospheric pressure, t For time, p v is the water vapor pressure, q v is the atmospheric specific humidity, γ v is the nonlinear coefficient of specific humidity, dp / dt is the rate of change of air pressure per unit time, dq v / dt is the rate of change of specific humidity per unit time, dp v / dt is the rate of change of water vapor pressure per unit time.

[0019] Preferably, the derivation process of the accurate water vapor pressure differential equation includes: firstly, based on the ideal gas state equation and Dalton's law of partial pressure, construct the functional relationship between water vapor partial pressure and specific humidity; secondly, consider the nonlinear effect caused by the mass difference between water vapor molecules and dry air molecules in the atmosphere, and introduce the quadratic term of specific humidity; finally, by performing a total differential operation on the relationship between water vapor pressure and specific humidity, an accurate water vapor pressure differential equation including the nonlinear effect of specific humidity is obtained. In addition, in the accurate water vapor pressure differential equation established by the present invention, the nonlinear coefficient γ v The calculation formula is , M d is the molar mass of dry air, M v is the molar mass of water vapor, and by introducing the nonlinear coefficient γ v , to accurately reflect the nonlinear relationship between water vapor pressure and specific humidity. In addition, the molar mass of dry air M d The preferred value is 28.96 g / mol, the molar mass of water vapor M v The preferred value is 18.02 g / mol, so the nonlinear coefficient γ v The preferred value of is 0.61.

[0020] More specifically, the embodiment of the present invention constructs an accurate water vapor pressure differential equation in the following manner. First, water vapor pressure is the weight of water vapor in a unit area of ​​an air column, which directly reflects the water content of the atmosphere. When the water vapor pressure in the air reaches its maximum value, it is called saturated humid air, and the water vapor pressure at this time is the saturated water vapor pressure. The ratio of water vapor pressure to saturated water vapor pressure reflects the relative humidity of the atmosphere, that is, the water vapor saturation ratio U v : in, U v Relative humidity, also known as water vapor saturation ratio, is the water vapor supersaturation and is a second-order small quantity relative to 1, p v and p vs are water vapor pressure and saturated water vapor pressure respectively.

[0021] Secondly, suppose p is the ambient atmospheric pressure, water vapor pressure accounts for χ v Saturated water vapor pressure χ vs The algorithm formulas are , From the above definition, it can be seen that the saturated water vapor pressure ratio can also be regarded as a molar ratio, and its value is calculated by a high-precision empirical formula (such as the Goff-Gratch formula or the Tetens formula). Under normal atmospheric conditions, the water vapor pressure ratio is χ v The value of is usually no more than 0.06. Further combining the water vapor pressure and specific humidity q v Combining the above equations, we can get: in, q v is the atmospheric specific humidity, γ v is the nonlinear coefficient of specific humidity. By performing total differential operation on the above equation, the accurate water vapor pressure differential equation established by the present invention can be obtained: Compared with the water vapor pressure differential equation commonly used in the prior art In contrast, the accurate water vapor pressure differential equation established in the present invention takes into account the atmospheric specific humidity by introducing the nonlinear term of specific humidity. q v The nonlinear effect of the water vapor pressure differential equation is improved.

[0022] SS2. Construct a calculation model for the change in water vapor mass per unit area of ​​air column The exact water vapor pressure differential equation constructed in step SS1 is Substitute into the calculation formula of water vapor mass per unit area of ​​air column , construct a calculation model for the change of water vapor mass per unit area of ​​air column , and the nonlinear term of specific humidity in the calculation model is expanded into an infinite series form , in order to improve the calculation accuracy to the nonlinear order, and fully consider the regulatory effect of the nonlinear effect of specific humidity on water vapor distribution, where m v is the water vapor mass per unit area of ​​the air column, g is the acceleration due to gravity, p b and p t are the bottom and top pressures per unit area of ​​the air column, k is the order of the infinite series expansion.

[0023] Preferably, when calculating the nonlinear term of the specific humidity in the form of infinite series, the specific humidity under actual atmospheric conditions is q v Typical magnitude and nonlinear coefficient of specific humidity γ vThe value range of can be selected by only selecting the first several order terms as the approximate values ​​of the calculation, so as to ensure the calculation accuracy while maintaining the high-precision description ability of the model for the nonlinear effect of specific humidity.

[0024] SS3. Dynamically track tropical cyclones and determine calculation areas Obtain meteorological data for the tropical cyclone area to be evaluated and calculated, determine the intensity and position of the tropical cyclone center at each time according to the lowest average sea level pressure value and its position at each time in the meteorological data, dynamically track the tropical cyclone path, establish a dynamic coordinate system based on the tropical cyclone center position, use the tropical cyclone center position at each time as the origin of the moving coordinates and the center of the dynamic calculation area, map relevant meteorological physical quantities to the dynamic coordinate system, adjust the scope and position of the calculation area in real time to make it accurately match the moving path of the tropical cyclone center, realize dynamic tracking of tropical cyclones, and ensure the accuracy of spatiotemporal calculation and evaluation of water vapor mass changes.

[0025] Preferably, meteorological data of the tropical cyclone to be evaluated and calculated are obtained based on reanalysis data, satellite remote sensing observation data and / or ground meteorological observation data. The meteorological data at least include horizontal wind field data, vertical velocity data, potential height data, specific humidity data, relative humidity data and mean sea level pressure data with a time resolution of not less than 6 hours and a spatial resolution of not less than 1°×1° in the vertical direction of 1000 hPa to 50 hPa. The time range of the meteorological data covers the complete life history of tropical cyclones from generation to dissipation, and meteorological data from different sources are preprocessed through quality control and data assimilation methods to ensure the spatiotemporal continuity and physical consistency of the data. The acquisition and integration of the above-mentioned meteorological data can provide comprehensive support for the dynamic tracking of tropical cyclone areas, the real-time adjustment of the calculation area, and the evaluation of water vapor quality changes, while improving the applicability and accuracy of the calculation model at different levels.

[0026] Preferably, the specific steps of dynamically tracking tropical cyclones may include: determining the intensity and position of the center of the tropical cyclone at each time using the lowest mean sea level pressure value and its position at each time, constructing the movement trajectory of the tropical cyclone through the time series of the center position of the tropical cyclone, establishing a dynamic coordinate system with the center of the tropical cyclone as the origin, and constructing a dynamic calculation area (for example, 5°×5° in size) with the longitude and latitude coordinates of the center of the tropical cyclone as the center, and adjusting the position of the calculation area in real time with the movement of the center of the tropical cyclone to ensure that the calculation area always covers the core area of ​​the tropical cyclone. In addition, the dynamic tracking of tropical cyclones further includes mapping the relevant meteorological physical quantities (including horizontal wind field, vertical velocity, specific humidity, relative humidity, etc.) at each time to the dynamic coordinate system using an interpolation method, and updating the spatial distribution characteristics of the physical quantities in real time in combination with the changes in the dynamic tracking path to ensure the dynamic consistency of the water vapor quality change assessment with the actual physical field.

[0027] SS4. Constructing a multi-layer convolutional neural network model for tropical cyclone water vapor quality calculation and assessment Based on the calculation model of the change of water vapor mass per unit area of ​​air column constructed in step SS2, a multi-layer convolutional neural network model for calculating and evaluating the water vapor mass of tropical cyclones is constructed by using a machine learning algorithm. The multi-layer convolutional neural network model includes an input layer, a hidden layer, and an output layer. The model input layer is used to receive the meteorological data of each altitude layer in the tropical cyclone region in a dynamic coordinate system and the water vapor pressure change rate calculated according to the precise water vapor pressure differential equation. The hidden layer uses multiple convolutional layers and pooling layers to extract the multidimensional features of the tropical cyclone water vapor distribution and uses a fully connected layer for feature fusion. At the same time, an attention mechanism module is introduced to capture the correlation between the water vapor mass changes between different altitude layers. The output layer uses the water vapor mass changes calculated according to the calculation model results in step SS2 as the label of the training data, and calculates and evaluates the water vapor mass changes per unit area of ​​the air column at each altitude layer.

[0028] Preferably, in the multi-layer convolutional neural network model of the present invention, its input layer divides the dynamic calculation area into three height levels according to the vertical direction of the atmosphere for data reception, wherein the meteorological element field within the range of 1000-700hPa is taken at the lower layer to analyze the water vapor transport and humidification process of the lower atmosphere; the meteorological element field within the range of 700-500hPa is taken at the middle layer to capture the water vapor fluctuation characteristics of the middle atmosphere; the meteorological element field within the range of 500-50hPa is taken at the high layer to evaluate the drying effect of the upper atmosphere. In addition, the output layer of the multi-layer convolutional neural network model uses a mean square error loss function to measure the difference between the predicted value of the multi-layer convolutional neural network model and the water vapor mass change calculated according to the calculation model result in step SS2, and introduces the absolute value or square of the difference between the two as an additional loss term to optimize and adjust the loss function, so as to improve the model prediction accuracy and consistency with the calculation model result.

[0029] Preferably, in step SS4, according to the development stages of the tropical cyclone's life cycle, including the development stage, the intensification stage and the mature stage, the water vapor mass change characteristics at different altitudes in different stages are calculated and dynamically evaluated in a hierarchical manner to reveal the humidity change patterns of tropical cyclones at different development stages, and to provide key data support and theoretical basis for cyclone intensity evolution and path prediction.

[0030] In addition, the layered calculation to evaluate the water vapor mass changes at different altitudes can include the analysis of the convergence effect of the water vapor mass changes at the lower layer, the calculation of the horizontal convergence rate through the humidification rate of the water vapor mass changes at the lower layer, the evaluation of the contribution of the humidity changes at the lower layer to the tropical cyclone intensity enhancement, and the analysis of the impact of the humidity accumulation at the lower layer on the overall humidity distribution in combination with the changes at the middle and upper layers. At the same time, the layered calculation to evaluate the water vapor mass changes at different altitudes can also include the analysis of the fluctuation characteristics of the water vapor mass changes at the middle layer, the evaluation of the impact of the humidity fluctuations at the middle layer on the cyclone structure and convective activity by calculating the fluctuation intensity of the water vapor mass changes at the middle layer per unit area of ​​the air column, and the analysis of the spatiotemporal coupling relationship between the humidity fluctuations at the middle layer and the humidification at the lower layer.

[0031] SS5. Output the calculation and evaluation results of water vapor mass change The multi-layer convolutional neural network model for tropical cyclone water vapor mass calculation and evaluation constructed in step SS4 is used, combined with the meteorological data of the calculation area and each altitude level dynamically determined in step SS3, to generate the spatiotemporal distribution data of the changes in water vapor mass at different altitude levels of the tropical cyclone.

[0032] As a preference, the calculation and evaluation results of the water vapor mass change may further include a contribution rate analysis of the nonlinear effect of specific humidity, specifically by comparing the exact water vapor pressure differential equation based on the introduction of the nonlinear term of specific humidity Compared with the water vapor pressure differential equation based on the nonlinear term The contribution of the nonlinear term of specific humidity to the humidification rate in the low layer, the fluctuation intensity in the middle layer and the drying rate in the high layer is calculated based on the calculation results of the water vapor mass change, and the influence of the nonlinear effect of specific humidity on the overall humidity distribution regulation is comprehensively evaluated.

[0033] Through the above steps, the present invention realizes high-precision calculation and evaluation of tropical cyclone water vapor mass changes on the basis of considering the nonlinear effect of specific humidity, improves the diagnostic accuracy of humidity distribution, and provides technical support for intensity simulation, path prediction and numerical model optimization of tropical cyclones.

[0034] Example 2 Based on the above-mentioned Example 1, as a more specific and detailed example, this Example 2 takes Typhoon No. 5 "Doksuri" in 2023 as the research object, and combines the NCEP reanalysis data (FNL data) to calculate and evaluate the changes in water vapor mass at different vertical levels during the development of the typhoon, which fully verifies the effectiveness and applicability of the tropical cyclone water vapor mass change calculation and evaluation method based on machine learning and integrated with the precise water vapor pressure differential equation described in the present invention.

[0035] This embodiment uses the 6-hour global analysis field data provided by NCEP. The data includes the following meteorological elements: horizontal wind field, vertical speed, potential height, specific humidity, relative humidity and mean sea level pressure. The horizontal resolution is 1°×1°, and the vertical direction is 1000-50 hPa. The time covers 00:00 on July 21, 2023 to 00:00 on July 29, 2023, and fully records the entire process of Typhoon "Dusurui" from its formation to its dissipation. According to the lowest mean sea level pressure value and its position at each time, the intensity and path changes of Typhoon "Dusurui" are determined, as shown in the following figure. Figure 2 shown. Figure 2 (a) shows the central movement path of "Dusurui". The typhoon was generated in the sea east of the Philippines at 00:00 on July 21, 2023, and then continued to move westward and strengthen, reaching super typhoon intensity at 12:00 on July 24. At 21:00 on July 26, "Dusurui" weakened to a strong typhoon, and its path changed from west to north. At 02:00 on July 28, it strengthened again to a super typhoon and landed on the coast of Fujian, my country, and finally stopped being numbered at 03:00 on July 29. Figure 2 (b) shows the average sea level pressure at the center of Typhoon Dusurui and its 12-hour pressure variation. Its intensity variation can be divided into three stages: the first 60 hours (00:00 on July 21-12:00 on July 23) is the development stage, the middle 60 hours (12:00 on July 23-00:00 on July 26) is the strengthening stage, and the last 72 hours (00:00 on July 26-00:00 on July 29) is the mature stage.

[0036] according to Figure 2 The typhoon path and intensity changes shown in the figure are based on the lowest average sea level pressure center position of "Dusurui". A dynamic coordinate system is established through dynamic tracking method, the typhoon center position is used as the origin of the moving coordinates, and a 5°×5° area centered on the typhoon center is selected as the dynamic calculation area. The specific steps include: determining the longitude and latitude position of the typhoon center at each time, and constructing a dynamic coordinate system with this position as the origin; mapping the physical quantities in the meteorological data (such as specific humidity, relative humidity, horizontal wind field, etc.) to the dynamic calculation area; adjusting the position of the dynamic calculation area in real time as the typhoon center moves, so that it always covers the typhoon core area.

[0037] The dynamic calculation area is divided into three altitude layers according to the vertical direction of the atmosphere: low layer (1000-700hPa), middle layer (700-500hPa), and high layer (500-50hPa). The water vapor mass change calculation model constructed by the present invention is used to calculate the water vapor mass change per unit area of ​​air column at different altitude layers, and analyze its change trend. The results are as follows: Figure 3 shown. Figure 3(a) shows the evolution trend of the water vapor mass change per unit area of ​​the air column in the low, middle and high layers: for the low layer, in the development stage and the strengthening stage, the water vapor mass in the low layer always remains positive, indicating continuous humidification, and the humidification rate continues to accelerate with the increase in typhoon intensity; in the mature stage (especially after landing), the humidification rate drops significantly, indicating that the humidification process is closely related to the typhoon intensity. For the middle layer, the change of water vapor mass in the middle layer shows a fluctuating characteristic, with short-term humidification in the development stage and mainly drying in the rest of the time. For the high layer, the water vapor mass in the high layer is always negative, indicating continuous drying, and the drying rate is in anti-phase with the humidification rate in the low layer. Figure 3 (b) shows the regulating effect of the nonlinear term of specific humidity on the changes in water vapor mass in the low, middle and high layers: for the low layer, the nonlinear term of specific humidity slows down the trend of humidification, and its effect increases with the increase of typhoon intensity. For the middle and high layers, the nonlinear term of specific humidity slows down the trend of drying, especially in the strengthening stage, its effect increases significantly.

[0038] On the basis of the above-mentioned hierarchical calculation, the multi-layer convolutional neural network model for tropical cyclone water vapor mass calculation and evaluation constructed in step SS4 of Example 1 is used to further evaluate the water vapor mass change of the "Dusurui" typhoon. For the model input layer, the meteorological data of each altitude layer in the tropical cyclone region under the dynamic coordinate system (such as horizontal wind field, vertical velocity, specific humidity, relative humidity, etc.) and the water vapor pressure change rate calculated by the accurate water vapor pressure differential equation are input. The hidden layer uses 4 convolutional layers and 3 maximum pooling layers to extract features, the convolution kernel size is 3×3, and the number of channels is 32, 64, 128, and 256 respectively. The spatial attention mechanism and channel attention mechanism are introduced after the last convolutional layer to capture the correlation of water vapor mass changes between different altitude layers, and the features of the low, middle and high layers are fused through the fully connected layer to generate comprehensive features. The model output layer outputs the water vapor mass changes in the air column per unit area at each altitude layer, and generates the spatiotemporal distribution data of the low-level humidification rate, the middle-level fluctuation intensity and the high-level drying rate. The multi-layer convolutional neural network model uses the Adam optimizer during training. The initial value of the learning rate is set to 0.001, which decays to the original 0.1 every 50 epochs. The batch size is set to 32. The number of training rounds is 200. The loss function uses the mean square error function, and the L2 regularization term is introduced to prevent overfitting. The model performance is evaluated through 5-fold cross validation, and the average validation set loss converges to below 0.015.

[0039] In order to verify the calculation accuracy of the tropical cyclone water vapor mass change calculation and evaluation method based on machine learning and integrating the accurate water vapor pressure differential equation described in the present invention, it is compared with the commonly used water vapor pressure differential equation. The results are as follows: Figure 4 shown. Figure 4The calculation residual ratios of the accurate equation and the common equation in the low, middle and high layers are shown: for the low layer, the residual ratio of the accurate equation is 1 / 7 of that of the common equation on average, indicating that the introduction of the nonlinear term of specific humidity significantly improves the calculation accuracy of the low-layer humidification process. For the middle and high layers, the residual ratio of the accurate equation is 1 / 2 of that of the common equation on average, especially in the humidification process, while the common equation slightly overestimates the water vapor content in the drying process.

[0040] Finally, based on the stratified calculation results, the temporal and spatial distribution data of water vapor mass changes in different altitude layers during the evolution of "Dusurui" were generated, including: distribution maps and change curves of water vapor mass changes in different altitude layers; analysis of the contribution rate of nonlinear effects of specific humidity to the low-level humidification rate, middle-level fluctuation intensity and high-level drying rate; evaluation of the impact of water vapor mass changes on typhoon intensity changes, etc.

[0041] Through the above implementation process, the present invention has verified its applicability and advantages to complex weather systems in the "Dusurui" case, significantly improved the accuracy of water vapor mass change calculations, and provided reliable support for tropical cyclone intensity simulation and path prediction.

[0042] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention is described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for calculating and evaluating the change of water vapor mass in tropical cyclones based on machine learning and integrating the accurate water vapor pressure differential equation, characterized in that: The method comprises at least the following steps when implemented: SS1. Combining the physical characteristics of the atmosphere and the relationship between specific humidity and water vapor pressure, the nonlinear term of specific humidity is introduced, and the accurate water vapor pressure differential equation including the nonlinear effect of specific humidity is established: in, p is the ambient atmospheric pressure, t For time, p v is the water vapor pressure, q v is the atmospheric specific humidity, γ v is the nonlinear coefficient of specific humidity, dp / dt is the rate of change of air pressure per unit time, dp v / dt is the rate of change of water vapor pressure per unit time, dq v / dt is the rate of change of specific humidity per unit time; SS2. Substitute the accurate water vapor pressure differential equation established in step SS1 into the calculation formula for the water vapor mass per unit area of ​​air column: In this paper, a calculation model for the change of water vapor mass per unit area of ​​air column is constructed. , and the nonlinear term of specific humidity in the calculation model is expanded into an infinite series form ,in m v is the water vapor mass per unit area of ​​the air column, g is the acceleration due to gravity, p b and p t are the bottom and top pressures per unit area of ​​the air column, k is the order of the infinite series expansion; SS3. Obtain meteorological data of the tropical cyclone area to be evaluated and calculated, determine the intensity and position of the tropical cyclone center at each time according to the lowest average sea level pressure value and its position at each time in the meteorological data, dynamically track the tropical cyclone path, establish a dynamic coordinate system based on the tropical cyclone center position, use the tropical cyclone center position at each time as the origin of the moving coordinates and the center of the dynamic calculation area, and map the relevant meteorological physical quantities to the dynamic coordinate system. By adjusting the scope and position of the calculation area in real time, it can accurately match the moving path of the tropical cyclone center, realize dynamic tracking of tropical cyclones, and ensure the spatiotemporal calculation and evaluation accuracy of water vapor mass changes; SS4. Based on the calculation model of the change of water vapor mass per unit area of ​​air column constructed in step SS2, a multi-layer convolutional neural network model for calculating and evaluating the water vapor mass of tropical cyclones is constructed by using a machine learning algorithm. The multi-layer convolutional neural network model includes an input layer, a hidden layer, and an output layer. The model input layer is used to receive the meteorological data of each altitude layer in the tropical cyclone region in a dynamic coordinate system and the water vapor pressure change rate calculated according to the accurate water vapor pressure differential equation. The hidden layer uses multiple convolutional layers and pooling layers to extract the multi-dimensional features of the tropical cyclone water vapor distribution and uses a fully connected layer for feature fusion. At the same time, an attention mechanism module is introduced to capture the correlation between the water vapor mass changes between different altitude layers. The output layer uses the water vapor mass changes calculated according to the calculation model results in step SS2 as the label of the training data, and calculates and evaluates the water vapor mass changes per unit area of ​​air column at each altitude layer respectively. SS5. Using the multi-layer convolutional neural network model for tropical cyclone water vapor mass calculation and evaluation constructed in step SS4, combined with the meteorological data of the calculation area and each altitude level dynamically determined in step SS3, the spatiotemporal distribution data of the changes in water vapor mass at different altitude levels of the tropical cyclone are generated.

2. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 1, characterized in that: In step SS1, the derivation process of the precise water vapor pressure differential equation includes: first, based on the ideal gas state equation and Dalton's law of partial pressure, construct the functional relationship between water vapor partial pressure and specific humidity; second, consider the nonlinear effect caused by the mass difference between water vapor molecules and dry air molecules in the atmosphere, and introduce the quadratic term of specific humidity; finally, through the total differential operation of the relationship between water vapor pressure and specific humidity, obtain the precise water vapor pressure differential equation including the nonlinear effect of specific humidity.

3. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 1, characterized in that: In step SS1, the nonlinear coefficient of the specific humidity in the exact water vapor pressure differential equation is γ v , and its calculation formula is ,in M d is the molar mass of dry air, M v is the molar mass of water vapor, and by introducing the nonlinear coefficient γ v , in order to accurately reflect the nonlinear relationship between water vapor pressure and specific humidity.

4. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 1, characterized in that: In step SS2, when calculating the nonlinear term of the specific humidity in the form of an infinite series, the specific humidity under actual atmospheric conditions is q v Typical magnitude and nonlinear coefficient of specific humidity γ v The range of values ​​is selected, and the first several order terms are selected as the approximate values ​​of the calculation, so as to ensure the calculation accuracy while maintaining the high-precision description ability of the model for the nonlinear effect of specific humidity.

5. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 1, characterized in that: In step SS3, meteorological data of tropical cyclones are obtained based on reanalysis data, satellite remote sensing observation data and / or ground meteorological observation data, wherein the meteorological data include wind field data, vertical velocity data, potential height data, specific humidity data, relative humidity data and mean sea level pressure data with a time resolution of not less than 6 hours and a spatial resolution of not less than 1°×1° in the vertical direction of 1000 hPa to 50 hPa, and the time range covers the complete life history of tropical cyclones from generation to dissipation, and the meteorological data from different sources are integrated through quality control and data assimilation to ensure the spatiotemporal continuity and physical consistency of the data.

6. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 5, characterized in that: In step SS3, the specific steps of dynamically tracking the tropical cyclone include: determining the intensity and position of the tropical cyclone center at each time using the lowest average sea level pressure value and its position at each time, constructing the tropical cyclone movement trajectory through the time series of the tropical cyclone center position, establishing a dynamic coordinate system with the tropical cyclone center as the origin, and constructing a dynamic calculation area with the longitude and latitude coordinates of the tropical cyclone center as the center, adjusting the position of the calculation area in real time as the tropical cyclone center moves, and ensuring that the calculation area always covers the core area of ​​the tropical cyclone.

7. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 6, characterized in that: In step SS3, the dynamic tracking of tropical cyclones further includes mapping the relevant meteorological physical quantities at each time into the dynamic coordinate system using the interpolation method, and updating the spatial distribution characteristics of the physical quantities in real time in combination with the changes in the dynamic tracking path to ensure the dynamic consistency between the water vapor quality change assessment and the actual physical field.

8. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 7, characterized in that: In step SS4, the input layer of the multi-layer convolutional neural network model divides the dynamic calculation area into three height levels according to the vertical direction of the atmosphere for data reception, wherein the meteorological element field within the range of 1000-700 hPa is taken in the low layer to analyze the water vapor transport and humidification process of the bottom atmosphere; the meteorological element field within the range of 700-500 hPa is taken in the middle layer to capture the water vapor fluctuation characteristics of the middle atmosphere; The meteorological element field in the range of 500-50hPa is used to evaluate the drying effect of the upper atmosphere.

9. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 8, characterized in that: In step SS4, the output layer of the multi-layer convolutional neural network model uses the mean square error loss function to measure the difference between the predicted value of the multi-layer convolutional neural network model and the water vapor mass change calculated according to the calculation model results in step SS2, and introduces the absolute value or square of the difference between the two as an additional loss term to optimize the loss function to improve the model prediction accuracy and consistency with the calculation model results.

10. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 9, characterized in that: In step SS4, according to the development stage of the tropical cyclone life cycle, the water vapor mass change characteristics at different altitudes in the development stage, intensification stage and mature stage of the tropical cyclone are calculated and dynamically evaluated in a hierarchical manner to reveal the humidity change patterns of the tropical cyclone in different development stages.

11. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 10, characterized in that: In step SS4, the layered calculation and evaluation of water vapor mass changes at different altitudes further includes analyzing the convergence effect of low-level water vapor mass changes, calculating the horizontal convergence rate through the humidification rate of low-level water vapor mass changes, evaluating the contribution of low-level humidity changes to the enhancement of tropical cyclone intensity, and analyzing the impact of low-level humidity accumulation on the overall humidity distribution in combination with changes in middle and high levels.

12. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 11, characterized in that: In step SS4, the layered calculation and evaluation of water vapor mass changes at different altitudes further includes an analysis of the fluctuation characteristics of the water vapor mass changes in the middle layer. By calculating the fluctuation intensity of the water vapor mass changes in the middle layer per unit area of ​​the air column, the impact of the humidity fluctuation in the middle layer on the cyclone structure and convective activity is evaluated, and the spatiotemporal coupling relationship between the humidity fluctuation in the middle layer and the humidification in the low layer is analyzed.

13. The method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations according to claim 1, characterized in that: In step SS5, the calculation and evaluation results of the water vapor mass change further include the contribution rate analysis of the nonlinear effect of specific humidity, by comparing the accurate water vapor pressure differential equation based on the introduction of the nonlinear term of specific humidity Compared with the water vapor pressure differential equation based on the nonlinear term The contribution of the nonlinear term of specific humidity to the humidification rate in the low layer, the fluctuation intensity in the middle layer and the drying rate in the high layer is calculated based on the calculation results of the water vapor mass change, and the influence of the nonlinear effect of specific humidity on the overall humidity distribution regulation is comprehensively evaluated.

14. A computer program product comprising computer instructions, characterized in that The computer instructions are used to execute the tropical cyclone water vapor mass change calculation and evaluation method based on machine learning and integrated with the accurate water vapor pressure differential equation as described in any one of claims 1 to 13.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for calculating and evaluating changes in water vapor mass in tropical cyclones based on machine learning and integrating accurate water vapor pressure differential equations as described in any one of claims 1 to 13 is implemented.

Images