Tropical region offshore marine meteorological fusion analysis method

Abstract

The invention discloses a tropical region offshore marine meteorological fusion analysis method. The method comprises the following steps: step 1, data collection and preprocessing: collecting offshore multi-platform collaborative networking observation data of a tropical region; 2, data quality control: checking the data, and analyzing error characteristics of marine meteorological observation data; 3, data fusion processing: establishing a tropical region offshore multi-source data fusion analysis system, performing intelligent fusion on the multi-source observation data subjected to quality control, and constructing a marine meteorological fusion analysis data set; 4, constructing a coupling analysis model: performing coupling analysis on marine meteorological elements by combining a physical model and a numerical simulation technology, and discovering an evolution rule and influence factors of a marine meteorological system by simulating and predicting an interaction process between the ocean and the atmosphere; and 5, dynamic application and evaluation: through real-time analysis of key meteorological elements, potential marine meteorological disaster risk affairs are discovered and early warned.

Description

Technical Field

[0001] The present invention relates to the field of offshore marine meteorological technology, and in particular to a method for fusion analysis of offshore marine meteorology in tropical regions. Background Art

[0002] Tropical coastal waters are a key area for ocean-atmosphere interaction. Especially under strong winds, the exchange process at the air-sea interface profoundly impacts weather, climate, and even the marine environment. Fusion analysis of tropical coastal marine meteorology can leverage multi-source observational data to improve the accuracy of forecast models and provide strong support for maritime operations, shipping safety, and fishery production.

[0003] The complexity and variability of marine meteorological systems are particularly pronounced in tropical coastal waters. This region is frequently influenced by multiple meteorological factors, including monsoons, typhoons, and abnormal sea temperatures, resulting in highly variable and unpredictable marine weather conditions. Traditional single-data source observation and analysis methods often struggle to fully and accurately reveal the evolution and influencing factors of this complex system, posing challenges to scientific decision-making. Therefore, a fusion analysis method for marine meteorological data in tropical coastal waters is proposed. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the existing technology and propose a tropical offshore marine meteorological fusion analysis method.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A tropical offshore marine meteorological fusion analysis method comprises the following steps:

[0007] Step 1: Data collection and preprocessing: Collect multi-platform collaborative network observation data from tropical coastal areas and process the collected data;

[0008] Step 2: Data quality control: Initiate high-frequency integrity testing using the typhoon impact domain spatiotemporal identification model, build a multi-platform dynamic weight allocation system, generate a three-dimensional error field distribution model based on the Kriging interpolation algorithm, and calculate a multidimensional quality index based on the output error field to quantitatively evaluate data quality.

[0009] Step 3: Data fusion processing: Establish a multi-source data fusion analysis system for tropical coastal areas, intelligently fuse the quality-controlled multi-source observation data, and construct a marine meteorological fusion analysis dataset;

[0010] Step 4: Constructing a coupled analysis model: Combining physical models with numerical simulation techniques, we conduct coupled analysis of marine meteorological elements. By simulating and predicting the interaction between the ocean and the atmosphere, we can discover the evolution patterns and influencing factors of the marine meteorological system.

[0011] Step 5: Dynamic application and evaluation: Based on the marine meteorological fusion analysis data set, a real-time monitoring and early warning system is constructed. Through real-time analysis of key meteorological elements, potential marine meteorological disaster risks are discovered and warned, providing scientific decision-making support and services for marine resource development, environmental protection, disaster prevention and mitigation, and other fields.

[0012] The above further includes:

[0013] Furthermore, in step one, the multi-platform collaborative networking observation data includes satellite remote sensing data and in-situ observation data. The satellite remote sensing data obtains the temperature and humidity information of the ocean surface by accessing the tropical ocean dedicated channel data of the Fengyun satellite, and combines the drone drop sounding data to obtain vertical profile information. At the same time, the ocean reanalysis data set is used to fill the data gaps in the open ocean area; and the ocean environmental parameters are obtained through in-situ observation.

[0014] Furthermore, in step 1, the collected data is processed by using a spatiotemporal registration method of shipborne 3D laser scanning combined with a dynamic time warping (DTR) algorithm, including the following steps:

[0015] Data preprocessing: preprocess the collected data, and perform denoising, filtering and smoothing on the point cloud data for shipborne 3D laser scanning data;

[0016] Time synchronization: Using the time interpolation method, we interpolate missing or asynchronous data points based on the time series of known data points to obtain data at a unified time point. Assuming that satellite remote sensing data is updated every half hour, while in-situ observation data is updated every hour, we use the linear interpolation method to increase the time resolution of the in-situ observation data to every half hour, thus achieving time synchronization.

[0017] Spatial registration: Coordinate conversion methods are used to convert data from different observation devices into the same geographic coordinate system. For shipborne 3D laser scanning data, known geographic reference information (such as GPS coordinates, nautical charts, etc.) is used for coordinate conversion and calibration.

[0018] Dynamic Time Warping (DTR) algorithm: Based on time synchronization and spatial registration, the dynamic time warping (DTR) algorithm is used to construct a two-dimensional matrix M, where the rows represent the time series of satellite remote sensing data and the columns represent the time series of in-situ observation data. The elements in the matrix M represent the similarity between two data points (obtained by calculating the Euclidean distance or other distance metrics between them). The dynamic programming method is used to find a path from the upper left corner to the lower right corner in the matrix M so that the cumulative similarity of all elements on the path is maximized. The path is the optimal matching path between the two data sequences, which reflects the temporal asynchrony and spatial misalignment relationship between them. Through this path, the satellite remote sensing data and the in-situ observation data are aligned in time and space, thereby obtaining more accurate and reliable marine meteorological data.

[0019] Furthermore, the specific steps of the data quality control are:

[0020] Enhanced tropical feature inspection: A spatiotemporal identification model for the typhoon's impact domain is established, and a convolutional neural network (CNN) is used to identify typhoon eyewall characteristic data (such as the "sharp rise-sudden drop-stable" pattern of wind speed). Power outage data is interpolated and reconstructed based on the rate of change of the pressure gradient. Dynamic compensation formulas for temperature and humidity sensors are used to correct sensor drift errors caused by long-term exposure to high temperature and high humidity environments.

[0021] Multi-platform cross-validation optimization: Kriging interpolation is used to construct a three-dimensional (longitude, latitude, and altitude) error distribution field to quantify the spatial heterogeneity of wind speed errors in the typhoon area;

[0022] Intelligent correction and hierarchical decision-making: Collaborative correction of multi-source data and hierarchical handling of abnormal data are performed, and data sets with quality labels are output;

[0023] Quality assessment and feedback optimization: Build an adaptive iterative mechanism to drive continuous optimization of quality control algorithms through quantitative evaluation.

[0024] Furthermore, the establishment of a tropical offshore multi-source data fusion analysis system, which intelligently fuses the quality-controlled multi-source observation data to construct a marine meteorological fusion analysis data set, includes the following steps:

[0025] Establish a multi-source data fusion and analysis system for tropical coastal areas: The architecture of the multi-source data fusion and analysis system includes a data input module, a data fusion module, and a data output module. The data input module is responsible for receiving data from different observation platforms and performing format conversion and preprocessing. The data fusion module is responsible for realizing the intelligent fusion of multi-source data. The data output module is responsible for outputting the fused data.

[0026] Construction of marine meteorological fusion analysis dataset: Based on data fusion, a marine meteorological fusion analysis dataset is constructed. The dataset contains multiple marine meteorological elements such as ocean temperature, salinity, wind speed, wind direction, sea surface temperature, and waves. The structure of the marine meteorological fusion analysis dataset includes time dimension, space dimension, and element dimension. The time dimension represents the timestamp or time interval of the data, the space dimension represents the geographical location or spatial distribution of the data, and the element dimension represents the marine meteorological elements contained in the data.

[0027] Furthermore, the data fusion module includes a weighted average method, a Kalman filter method, and a particle filter method. When the satellite remote sensing data and the in-situ observation data are highly overlapped in time and space and the data quality is comparable, the weighted average method is used for fusion. The data is weighted averaged according to the weights of the satellite remote sensing data and the in-situ observation data. The weights are determined based on factors such as the error estimate of the data, the proximity of the timestamps, and the proximity of the spatial positions. When the state estimate is updated using the estimated value at the previous moment and the observed value at the current moment, the Kalman filter method is used for processing to establish a state space model to describe the dynamic changes of the ocean environment parameters over time. Then, the Kalman filter recursive algorithm is used to update the state estimate based on the estimated value at the previous moment and the observed value at the current moment. When the satellite remote sensing data is affected by nonlinear factors such as cloud cover and atmospheric interference, the particle filter method is used for processing to generate a group of particles (i.e., candidate states) through Monte Carlo simulation. Each particle represents a possible ocean environment parameter state. Then, the particles are weighted and resampled according to the observation data to obtain a state estimate.

[0028] Furthermore, the specific steps of constructing the coupling analysis model are:

[0029] Model selection and initialization: Select the ocean circulation model and the atmospheric circulation model. The ocean circulation model is expressed as

[0030]

[0031] Where u is the velocity vector, ρ0 is the seawater density, p is the pressure, g is the acceleration of gravity, v is the dynamic viscosity, and F is the external force (such as wind stress). The atmospheric circulation model is expressed as

[0032]

[0033] Where v is the wind speed vector, ρ is the air density, P is the air pressure, and C D is the drag coefficient, u ocean is the ocean surface velocity vector, v airIt is the aerodynamic viscosity. Based on the observation data and historical data, the model is initialized and the initial conditions (such as temperature, salinity, wind speed, wind direction, etc.) and boundary conditions (such as coastline, seabed topography, atmospheric circulation, etc.) are set.

[0034] Model coupling and data exchange: Establish a coupling interface between the ocean circulation model and the atmospheric circulation model to achieve data exchange and interaction between the ocean circulation model and the atmospheric circulation model;

[0035] Numerical simulation and result output: Use numerical simulation technology to solve the physical model and simulate the evolution process of the ocean circulation model and the atmospheric circulation model.

[0036] Furthermore, the specific steps of constructing the coupling analysis model are:

[0037] Model selection and initialization: Select the ocean circulation model and the atmospheric circulation model. The ocean circulation model is expressed as

[0038]

[0039] Where u is the velocity vector, ρ0 is the seawater density, p is the pressure, g is the acceleration of gravity, v is the dynamic viscosity, and F is the external force (such as wind stress). The atmospheric circulation model is expressed as

[0040]

[0041] Where v is the wind speed vector, ρ is the air density, P is the air pressure, C D is the drag coefficient, u ocean is the ocean surface velocity vector, v air It is the aerodynamic viscosity. Based on the observation data and historical data, the model is initialized and the initial conditions (such as temperature, salinity, wind speed, wind direction, etc.) and boundary conditions (such as coastline, seabed topography, atmospheric circulation, etc.) are set.

[0042] Model coupling and data exchange: Establish a coupling interface between the ocean circulation model and the atmospheric circulation model to achieve data exchange and interaction between the ocean circulation model and the atmospheric circulation model;

[0043] Numerical simulation and result output: Use numerical simulation technology to solve the physical model and simulate the evolution process of the ocean circulation model and the atmospheric circulation model.

[0044] The present invention has the following beneficial effects:

[0045] In the present invention, the coupled analysis model constructed combines physical models and numerical simulation technology to simulate and predict the interaction process between the ocean and the atmosphere, understand the internal mechanism of the marine meteorological system, reveal its evolution law and influencing factors, and more accurately predict the occurrence and development trend of marine meteorological disasters, providing early warning information for disaster prevention and mitigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a step diagram of a tropical offshore marine meteorological fusion analysis method proposed by the present invention. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0048] See also Figure 1 As shown, the present invention is a method for fusion analysis of tropical offshore marine meteorology, comprising the following steps:

[0049] Step 1: Data collection and preprocessing: Collect multi-platform collaborative network observation data from tropical coastal areas and process the collected data;

[0050] Step 2: Data quality control: Initiate high-frequency integrity testing using the typhoon impact domain spatiotemporal identification model, build a multi-platform dynamic weight allocation system, generate a three-dimensional error field distribution model based on the Kriging interpolation algorithm, and calculate a multidimensional quality index based on the output error field to quantitatively evaluate data quality.

[0051] Step 3: Data fusion processing: Establish a multi-source data fusion analysis system for tropical coastal areas, intelligently fuse the quality-controlled multi-source observation data, and construct a marine meteorological fusion analysis dataset;

[0052] Step 4: Constructing a coupled analysis model: Combining physical models with numerical simulation techniques, we conduct coupled analysis of marine meteorological elements. By simulating and predicting the interaction between the ocean and the atmosphere, we can discover the evolution patterns and influencing factors of the marine meteorological system.

[0053] Step 5: Dynamic application and evaluation: Based on the marine meteorological fusion analysis data set, a real-time monitoring and early warning system is constructed. Through real-time analysis of key meteorological elements, potential marine meteorological disaster risks are discovered and warned, providing scientific decision-making support and services for marine resource development, environmental protection, disaster prevention and mitigation, and other fields.

[0054] In one embodiment, in step one, the multi-platform collaborative networking observation data includes satellite remote sensing data and in-situ observation data. The satellite remote sensing data obtains ocean surface temperature and humidity information by accessing the tropical ocean dedicated channel data of the Fengyun satellite, and combines the drone drop sounding data to obtain vertical profile information. At the same time, the ocean reanalysis data set is used to fill the data gaps in the open ocean area; and the ocean environmental parameters are obtained through in-situ observation.

[0055] In one embodiment, in step 1, the collected data is processed by using a spatiotemporal registration method of shipborne 3D laser scanning combined with a dynamic time warping (DTR) algorithm, including the following steps:

[0056] Data preprocessing: preprocess the collected data, and perform denoising, filtering and smoothing on the point cloud data for shipborne 3D laser scanning data;

[0057] Time synchronization: Using the time interpolation method, we interpolate missing or asynchronous data points based on the time series of known data points to obtain data at a unified time point. Assuming that satellite remote sensing data is updated every half hour, while in-situ observation data is updated every hour, we use the linear interpolation method to increase the time resolution of the in-situ observation data to every half hour, thus achieving time synchronization.

[0058] Spatial registration: Coordinate conversion methods are used to convert data from different observation devices into the same geographic coordinate system. For shipborne 3D laser scanning data, known geographic reference information (such as GPS coordinates, nautical charts, etc.) is used for coordinate conversion and calibration.

[0059] Dynamic Time Warping (DTR) algorithm: Based on time synchronization and spatial registration, the dynamic time warping (DTR) algorithm is used to construct a two-dimensional matrix M, where the rows represent the time series of satellite remote sensing data and the columns represent the time series of in-situ observation data. The elements in the matrix M represent the similarity between two data points (obtained by calculating the Euclidean distance or other distance metrics between them). The dynamic programming method is used to find a path from the upper left corner to the lower right corner in the matrix M so that the cumulative similarity of all elements on the path is maximized. The path is the optimal matching path between the two data sequences, which reflects the temporal asynchrony and spatial misalignment relationship between them. Through this path, the satellite remote sensing data and the in-situ observation data are aligned in time and space, thereby obtaining more accurate and reliable marine meteorological data.

[0060] In one embodiment, the specific steps of data quality control are:

[0061] Enhanced tropical feature inspection: A spatiotemporal identification model for the typhoon's impact domain is established, and a convolutional neural network (CNN) is used to identify typhoon eyewall characteristic data (such as the "sharp rise-sudden drop-stable" pattern of wind speed). Power outage data is interpolated and reconstructed based on the rate of change of the pressure gradient. Dynamic compensation formulas for temperature and humidity sensors are used to correct sensor drift errors caused by long-term exposure to high temperature and high humidity environments.

[0062] Multi-platform cross-validation optimization: Kriging interpolation is used to construct a three-dimensional (longitude, latitude, and altitude) error distribution field to quantify the spatial heterogeneity of wind speed errors in the typhoon area;

[0063] Intelligent correction and hierarchical decision-making: Perform collaborative correction of multi-source data and hierarchical processing of abnormal data, and output data sets with quality labels;

[0064] Quality assessment and feedback optimization: Build an adaptive iterative mechanism to drive continuous optimization of quality control algorithms through quantitative evaluation.

[0065] In one embodiment, the establishment of a tropical offshore multi-source data fusion analysis system, which intelligently fuses quality-controlled multi-source observation data to construct a marine meteorological fusion analysis dataset, includes the following steps:

[0066] Establish a multi-source data fusion and analysis system for tropical coastal areas: The architecture of the multi-source data fusion and analysis system includes a data input module, a data fusion module, and a data output module. The data input module is responsible for receiving data from different observation platforms and performing format conversion and preprocessing. The data fusion module is responsible for realizing the intelligent fusion of multi-source data. The data output module is responsible for outputting the fused data.

[0067] Construction of marine meteorological fusion analysis dataset: Based on data fusion, a marine meteorological fusion analysis dataset is constructed. The dataset contains multiple marine meteorological elements such as ocean temperature, salinity, wind speed, wind direction, sea surface temperature, and waves. The structure of the marine meteorological fusion analysis dataset includes time dimension, space dimension, and element dimension. The time dimension represents the timestamp or time interval of the data, the space dimension represents the geographical location or spatial distribution of the data, and the element dimension represents the marine meteorological elements contained in the data.

[0068] In one embodiment, the data fusion module includes a weighted averaging method, a Kalman filter method, and a particle filter method. When the satellite remote sensing data and the in-situ observation data highly overlap in time and space and the data quality is comparable, the weighted averaging method is used for fusion. The data is weighted averaged according to the weights of the satellite remote sensing data and the in-situ observation data. The weights are determined based on factors such as the error estimate of the data, the proximity of the timestamps, and the proximity of the spatial positions. When the state estimate is updated using the estimated value at the previous moment and the observed value at the current moment, the Kalman filter method is used for processing to establish a state space model to describe the dynamic changes of the ocean environment parameters over time. Then, the Kalman filter recursive algorithm is used to update the state estimate based on the estimated value at the previous moment and the observed value at the current moment. When the satellite remote sensing data is affected by nonlinear factors such as cloud cover and atmospheric interference, the particle filter method is used for processing to generate a group of particles (i.e., candidate states) through Monte Carlo simulation. Each particle represents a possible state of the ocean environment parameter. Then, the particles are weighted and resampled according to the observation data to obtain the state estimate.

[0069] In one embodiment, the specific steps of constructing the coupling analysis model are:

[0070] Model selection and initialization: Select the ocean circulation model and the atmospheric circulation model. The ocean circulation model is expressed as

[0071]

[0072] Where u is the velocity vector, ρ0 is the seawater density, p is the pressure, g is the acceleration of gravity, v is the dynamic viscosity, and F is the external force (such as wind stress). The atmospheric circulation model is expressed as

[0073]

[0074] Where v is the wind speed vector, ρ is the air density, P is the air pressure, and C D is the drag coefficient, u ocean is the ocean surface velocity vector, v air It is the aerodynamic viscosity. Based on the observation data and historical data, the model is initialized and the initial conditions (such as temperature, salinity, wind speed, wind direction, etc.) and boundary conditions (such as coastline, seabed topography, atmospheric circulation, etc.) are set.

[0075] Model coupling and data exchange: Establish a coupling interface between the ocean circulation model and the atmospheric circulation model to achieve data exchange and interaction between the ocean circulation model and the atmospheric circulation model;

[0076] Numerical simulation and result output: Use numerical simulation technology to solve the physical model and simulate the evolution process of the ocean circulation model and the atmospheric circulation model.

[0077] In one embodiment, the real-time monitoring and early warning system is constructed based on the marine meteorological fusion analysis data set. By analyzing key meteorological elements in real time, the system discovers and warns of potential marine meteorological disaster risks, providing scientific decision-making support and services for marine resource development, environmental protection, disaster prevention and mitigation, and other fields, including the following steps:

[0078] Step 1: Extraction of key meteorological elements: Extract key meteorological elements from the fused dataset;

[0079] Step 2: Real-time monitoring model construction: SARIMA is used to build a real-time monitoring model, and key meteorological element data are used to monitor the marine meteorological status in real time. The basic formula of SARIMA is expressed as Among them, Y t is the value of the time series data at time t, Φ(B) is a non-seasonal autoregressive polynomial, which is in the form of Φ(B) = 1-φ1B-φ2B 2 -…-φ p B p , where B is the backshift operator, φ1, φ2, ..., φ p is the non-seasonal autoregressive coefficient, Φ s (B s ) is a seasonal autoregressive polynomial of the form Φ s (B s )=1-Φ1B s -Φ2B 2s -…-Φ P B Ps , where s is the seasonal period, Φ1, Φ2, ..., Φ P is the seasonal autoregressive coefficient, Is a non-seasonal difference operator, which means that the time series is non-seasonal differentiated d times. is the seasonal difference operator, which means that the time series is seasonally differenciated D times with a period of s. Θ(B) is the non-seasonal moving average polynomial, which is in the form of Θ(B) = 1 + θ1B + θ2B 2 +…+θ q B q , among them, θ1, θ2,..., θ q is the non-seasonal moving average coefficient, Θ s (B s ) is the seasonal moving average polynomial, which has the form Θ s (B s )=1+Θ1B s +Θ2B 2s +…+Θ Q B Qs , where Θ1, Θ2, ..., ΘQ is the seasonal moving average coefficient, ∈ t is a white noise process, representing the random error term at time t;

[0080] Step 3: Setting warning rules: Setting warning rules based on historical data of marine meteorological disasters and expert experience;

[0081] Step 4: Warning information release: Once the meteorological conditions that meet the warning rules are monitored, the system automatically releases warning information, which can be sent to relevant departments and personnel via SMS, email, APP push, etc.

[0082] Step 5: Disaster risk assessment: Based on real-time monitoring data and early warning information, assess the potential risk of marine meteorological disasters;

[0083] Step 6: Generate a decision support report: Based on the disaster risk assessment results, generate a decision support report that includes the disaster risk level, impact range, possible losses, and recommended disaster prevention and mitigation measures.

[0084] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for fusion analysis of offshore marine meteorology in tropical regions, characterized in that: The following steps are involved: Step 1: Data collection and preprocessing: Collect multi-platform collaborative network observation data from tropical coastal areas and process the collected data; Step 2: Data quality control: Initiate high-frequency integrity testing using the typhoon impact domain spatiotemporal identification model, build a multi-platform dynamic weight allocation system, generate a three-dimensional error field distribution model based on the Kriging interpolation algorithm, and calculate a multidimensional quality index based on the output error field to quantitatively evaluate data quality. Step 3: Data fusion processing: Establish a multi-source data fusion analysis system for tropical coastal areas, intelligently fuse the quality-controlled multi-source observation data, and construct a marine meteorological fusion analysis dataset; Step 4: Constructing a coupled analysis model: Combining physical models with numerical simulation techniques, we conduct coupled analysis of marine meteorological elements. By simulating and predicting the interaction between the ocean and the atmosphere, we can discover the evolution patterns and influencing factors of the marine meteorological system. Step 5: Dynamic application and evaluation: Based on the marine meteorological fusion analysis data set, a real-time monitoring and early warning system is built. Through real-time analysis of key meteorological elements, potential marine meteorological disaster risks can be discovered and warned, providing scientific decision-making support and services.

2. The method for fusion analysis of tropical offshore marine meteorology according to claim 1, characterized in that: In step one, the multi-platform collaborative networking observation data includes satellite remote sensing data and in-situ observation data. The satellite remote sensing data obtains ocean surface temperature and humidity information by accessing the tropical ocean dedicated channel data of the Fengyun satellite, and combines the drone drop sounding data to obtain vertical profile information. At the same time, the ocean reanalysis data set is used to fill the data gaps in the open ocean area; and the ocean environmental parameters are obtained through in-situ observation.

3. The method for fusion analysis of tropical offshore marine meteorology according to claim 1, characterized in that: In step 1, the collected data is processed by using a spatiotemporal registration method of shipborne 3D laser scanning combined with a dynamic time warping algorithm, including the following steps: Data preprocessing: preprocess the collected data, and perform denoising, filtering and smoothing on the point cloud data for shipborne 3D laser scanning data; Time synchronization: Using the time interpolation method, we interpolate missing or asynchronous data points based on the time series of known data points to obtain data at a unified time point. Spatial registration: Use coordinate conversion methods to convert data from different observation devices into the same geographic coordinate system. For shipborne 3D laser scanning data, use known geographic reference information for coordinate conversion and calibration. Dynamic Time Warping Algorithm: Based on time synchronization and spatial registration, a dynamic time warping algorithm is used to construct a two-dimensional matrix M, where the rows represent the time series of satellite remote sensing data and the columns represent the time series of in-situ observation data. The elements in the matrix M represent the similarity between two data points. A dynamic programming method is used to find a path from the upper left corner to the lower right corner in the matrix M that maximizes the cumulative similarity of all elements on the path. This path is the optimal matching path between the two data sequences. Through this path, the satellite remote sensing data and the in-situ observation data are aligned in time and space.

4. The method for fusion analysis of tropical offshore marine meteorology according to claim 1, characterized in that: The specific steps of the data quality control are: Enhanced tropical feature inspection: Establish a spatiotemporal identification model for the typhoon's impact domain, use a convolutional neural network to identify typhoon eyewall feature data, interpolate and reconstruct power outage data based on the pressure gradient change rate, and correct sensor drift errors using dynamic compensation formulas for temperature and humidity sensors. Multi-platform cross-validation optimization: Kriging interpolation is used to construct a three-dimensional error distribution field to quantify the spatial heterogeneity of wind speed errors in the typhoon area; Intelligent correction and hierarchical decision-making: Collaborative correction of multi-source data and hierarchical handling of abnormal data are performed, and data sets with quality labels are output; Quality assessment and feedback optimization: Build an adaptive iterative mechanism to drive continuous optimization of quality control algorithms through quantitative evaluation.

5. The method for fusion analysis of tropical offshore marine meteorology according to claim 1, characterized in that: The establishment of a tropical offshore multi-source data fusion analysis system, which intelligently fuses quality-controlled multi-source observation data to construct a marine meteorological fusion analysis dataset, includes the following steps: Establish a multi-source data fusion and analysis system for tropical coastal areas: The architecture of the multi-source data fusion and analysis system includes a data input module, a data fusion module, and a data output module. The data input module is responsible for receiving data from different observation platforms and performing format conversion and preprocessing. The data fusion module is responsible for realizing the intelligent fusion of multi-source data. The data output module is responsible for outputting the fused data. Construction of marine meteorological fusion analysis dataset: Based on data fusion, a marine meteorological fusion analysis dataset is constructed. The structure of the marine meteorological fusion analysis dataset includes time dimension, space dimension and element dimension. The time dimension represents the timestamp or time interval of the data, the space dimension represents the geographical location or spatial distribution of the data, and the element dimension represents the marine meteorological elements contained in the data.

6. The method for fusion analysis of tropical offshore marine meteorology according to claim 1, characterized in that: The data fusion module includes weighted averaging, Kalman filtering and particle filtering. When satellite remote sensing data and in-situ observation data highly overlap in time and space and the data quality is comparable, weighted averaging is used for fusion, and the data are weighted averaged according to the weights of the satellite remote sensing data and the in-situ observation data. When the estimated value at the previous moment and the observed value at the current moment are used to update the state estimate, the Kalman filtering method is used for processing to establish a state space model to describe the dynamic changes of the ocean environment parameters over time. Then, the Kalman filtering recursive algorithm is used to update the state estimate according to the estimated value at the previous moment and the observed value at the current moment. When the satellite remote sensing data is affected by nonlinear factors, the particle filtering method is used for processing to generate a group of particles through Monte Carlo simulation, each particle representing a possible state of the ocean environment parameter, and then the particles are weighted updated and resampled according to the observation data to obtain the state estimate.

7. The method for fusion analysis of tropical offshore marine meteorology according to claim 1, characterized in that: The specific steps of constructing the coupling analysis model are as follows: Model selection and initialization: Select the ocean circulation model and the atmospheric circulation model. The ocean circulation model is expressed as Where u is the velocity vector, ρ0 is the seawater density, p is the pressure, g is the acceleration of gravity, v is the dynamic viscosity, and F is the external force. The atmospheric circulation model is expressed as Where v is the wind speed vector, ρ is the air density, P is the air pressure, C D is the drag coefficient, u ocean is the ocean surface velocity vector, v air is the aerodynamic viscosity. Based on the observation data and historical data, the model is initialized and the initial conditions and boundary conditions are set; Model coupling and data exchange: Establish a coupling interface between the ocean circulation model and the atmospheric circulation model to achieve data exchange and interaction between the ocean circulation model and the atmospheric circulation model; Numerical simulation and result output: Use numerical simulation technology to solve the physical model and simulate the evolution process of the ocean circulation model and the atmospheric circulation model.

8. The method for fusion analysis of tropical offshore marine meteorology according to claim 1, characterized in that: The real-time monitoring and early warning system is constructed based on the marine meteorological fusion analysis data set. Through real-time analysis of key meteorological elements, potential marine meteorological disaster risks are discovered and warned, and scientific decision-making support and services are provided. The following steps are included: Step 1: Extraction of key meteorological elements: Extract key meteorological elements from the fused dataset; Step 2: Real-time monitoring model construction: SARIMA is used to build a real-time monitoring model, and key meteorological element data are used to monitor the marine meteorological status in real time. The basic formula of SARIMA is expressed as Among them, Y t is the value of the time series data at time t, Φ(B) is a non-seasonal autoregressive polynomial, which is in the form of Φ(B) = 1-φ1B-φ2B 2 -…-φ p B p , where B is the backshift operator, φ1, φ2, ..., φ p is the non-seasonal autoregressive coefficient, Φ s (B s ) is a seasonal autoregressive polynomial, which has the form where s is the seasonal period, Φ1, Φ2, ..., Φ P is the seasonal autoregressive coefficient, Is a non-seasonal difference operator, which means that the time series is non-seasonal differentiated d times. is the seasonal difference operator, which means that the time series is seasonally differenciated D times with a period of s. Θ(B) is the non-seasonal moving average polynomial, which is in the form of Θ(B) = 1 + θ1B + θ2B 2 +…+θ q B q , among them, θ1, θ2,..., θ q is the non-seasonal moving average coefficient, Θ s (B s ) is the seasonal moving average polynomial, which has the form Θ s (B s )=1+Θ1B s +Θ2B 2s +…+Θ Q B Qs , where Θ1, Θ2, ..., Θ Q is the seasonal moving average coefficient, ∈ t is a white noise process, representing the random error term at time t; Step 3: Setting warning rules: Setting warning rules based on historical data of marine meteorological disasters and expert experience; Step 4: Warning information release: Once the meteorological conditions that meet the warning rules are monitored, the system automatically releases the warning information; Step 5: Disaster risk assessment: Based on real-time monitoring data and early warning information, assess the potential risk of marine meteorological disasters; Step 6: Generate a decision support report: Based on the disaster risk assessment results, generate a decision support report that includes the disaster risk level, impact range, possible losses, and recommended disaster prevention and mitigation measures.

Images