Method for constructing marine meteorological heat flux parameterization model under marine spray effect

Abstract

This invention discloses a method for constructing a parameterized model of marine meteorological heat flux under the marine droplet effect. This invention relates to the field of marine meteorological numerical forecasting technology. It integrates localized multi-source data preprocessing, subdivides marine droplets by particle size and establishes heat exchange correlations, models wind speed intervals, and constructs a heat flux equation with multiple correction terms. Through localized calibration, multi-scenario verification, and dynamic optimization, the computational adaptability is improved. The advantages of this invention are: by subdividing marine droplets by particle size and establishing their heat exchange characteristics with wind speed and wave period, it constructs a droplet generation rate model with sea ice adjustment terms specifically for wind speed intervals. Simultaneously, it incorporates localized measured data calibration and a quarterly dynamic optimization mechanism, solving the shortcomings of traditional heat flux parameterized models that do not fully consider droplet particle size differences, dynamic changes in sea state, and the influence of mid-to-high latitude sea ice, leading to large calculation deviations in complex weather scenarios.

Description

Technical Field

[0001] This invention relates to the field of marine meteorological numerical forecasting technology, specifically to a method for constructing a parameterized model of marine meteorological heat flux under the marine droplet effect. Background Technology

[0002] Marine meteorological heat flux is the core physical quantity of energy exchange in the process of ocean-atmosphere interaction. Its accurate quantification is of great significance for climate system simulation, marine weather forecasting, and safety assurance of maritime activities. The heat flux exchange at the ocean-atmosphere interface not only depends on basic factors such as the temperature and humidity difference between the sea surface and the near-sea atmosphere and wind speed, but is also significantly modulated by marine droplets. Under the action of wind, marine droplets formed by the breaking of seawater will participate in the ocean-atmosphere heat exchange through processes such as evaporation and heat transfer, becoming one of the key factors affecting the accuracy of heat flux calculation. Especially in complex scenarios such as strong winds, typhoons, and freezing periods in mid-to-high latitude sea areas, the impact of the droplet effect is more prominent. Existing parameterized models for marine meteorological heat flux mainly calculate air-sea heat flux by ignoring differences in ocean droplet size and using fixed parameters or a single wind field model. This approach has several drawbacks. First, it fails to adequately consider the differences in ocean droplet size, dynamic changes in sea state, and the impact of mid-to-high latitude sea ice on heat exchange. This leads to significant calculation deviations under complex weather scenarios such as ordinary strong winds, northward-moving typhoons, and periods of sea icing, resulting in insufficient consistency with actual air-sea exchange processes. Second, the model parameters are mostly fixed, making it difficult to respond to interannual sea state variations and the specificities of extreme weather events. This results in insufficient stability in long-term operational applications and poor reliability of heat flux forecasts under extreme weather scenarios. Therefore, we propose a method for constructing a parameterized model for marine meteorological heat flux under the ocean droplet effect. Summary of the Invention

[0003] The purpose of this invention is to provide a method for constructing a parameterized model of marine meteorological heat flux under the marine droplet effect.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for constructing a parameterized model of marine meteorological heat flux under the marine droplet effect, wherein the model construction method includes the following methods: Step 1: Collect localized observation data and sea state parameters of the target sea area and surrounding estuary areas, and perform missing value interpolation, outlier removal and format standardization on the data; Step 2: Divide marine droplets into three particle size ranges and establish a database of correspondences between particle size, wind speed, wave period, and heat exchange characteristics; Step 3: Construct a correlation model between droplet generation rate and localized sea state parameters for different wind speed ranges; Step 4: Derive the heat flux calculation equation that includes grain size weighting, wave period correction, and sea ice adjustment term; Step 5: Calibrate the model parameters using measured data from the target sea area; Step 6: Apply the calibrated model to different weather scenarios, calculate hourly heat flux, and verify accuracy. Step 7: Establish a dynamic optimization mechanism for the model to continuously correct systemic biases in the model.

[0005] As a further aspect of the present invention: In step one, the localized observation data includes buoy observation data of the target sea area, satellite remote sensing data, sea ice coverage data, and observation data from ground automatic stations and offshore platforms. The buoy observation data includes sea surface wind speed, significant wave height, wave period, sea surface temperature, and near-sea atmospheric temperature and humidity. The satellite remote sensing data includes radiation data from the geostationary interferometric infrared detector of Fengyun-4A satellite and atmospheric motion vector data from Kuihua-8 satellite. The sea ice coverage data is based on high-precision global coastline data and ice condition monitoring products. The time resolution of the preprocessed data is not less than 1 hour, and the spatial coverage is the entire target sea area. Missing values ​​are interpolated using the mean of adjacent time observations. Outlier removal is performed using the 3-times standard deviation method. The format is uniformly converted to a common network data format.

[0006] As a further aspect of the present invention: In step two, the three particle size ranges are 50μm-100μm split droplets, 10μm-50μm splash droplets, and <10μm foam droplets. The heat exchange coefficient is obtained by fitting observation data with a theoretical model, and the expression is as follows: ; in, No. The heat exchange coefficient of droplets of similar particle size , , For the fitting coefficients, the split droplet is: , =0.65、 =0.32, splashing: =0.15、 =0.58、 =0.28, Mote Droplet: =0.22、 =0.49、 =0.21, Wind speed at a height of 10m (unit: m / s). The wave period (unit: s) is given. The evaporation rate is derived from the energy balance equation, and the atmospheric residence time is calculated based on the coupled model of gravity settling and turbulent diffusion.

[0007] As a further aspect of the present invention: in step three, the wind speed intervals specifically refer to low to medium wind speeds (U < 30 m / s) and high wind speeds (U ≥ 30 m / s), and the droplet generation rate model under low to medium wind speeds is: ; in, =2.3×10 -5 , Sea ice coverage, 0 ≤ ≤1, the droplet formation rate model under high wind speed is: ; in, =1.8×10 -4 The model passed the significance test (R²). 2 ≥0.85) to verify validity.

[0008] As a further aspect of the present invention: in step four, the heat flux calculation equation is: ; in, Total heat flux (unit: W / m³) 2 ), This is the air-sea interface heat flux (calculated using the COARE algorithm built into the WRF mode). For the first Weighting coefficients for droplet size, split droplets =0.45, splashing =0.35, foam droplet =0.20, For the first The generation rate of droplets of similar size is taken at low to medium wind speeds. Take during high wind speed , For the first The heat exchange coefficient of droplets of similar particle size Sea surface temperature (unit: K). The near-sea surface atmospheric temperature (unit: K).

[0009] As a further aspect of the present invention: In step five, the measured data are the observed values ​​of buoy and offshore platform heat flux in the target sea area from 2020 to 2024. The root mean square error between the model calculated value and the measured value is minimized using the least squares method. The adjusted parameters include those from step two. , , With step three , After calibration, the root mean square error of heat flux calculation at low and medium wind speeds is ≤15W / m². 2Root mean square error ≤25W / m under high wind speed 2 , Introducing sea ice roughness correction term during the freezing period .

[0010] As a further aspect of the present invention: In step six, different weather scenarios include ordinary strong winds, northward-moving typhoons, and the ice-covered period. The input sea state parameters include wind speed, wave period, sea ice coverage, sea surface temperature, and atmospheric temperature and humidity. The accuracy verification indicators include root mean square error, mean absolute error, and correlation coefficient. Among them, the root mean square error is ≤12W / m in the medium and low wind speed scenario. 2 Mean absolute error ≤ 9W / m 2 The correlation coefficient is ≥0.90, and the root mean square error is ≤20W / m² in high wind speed scenarios. 2 Mean absolute error ≤15W / m 2 The correlation coefficient is ≥0.85, and the root mean square error during the freezing period is ≤18W / m. 2 Mean absolute error ≤13W / m 2 The correlation coefficient is ≥0.88.

[0011] As a further aspect of the present invention: in step seven, the optimization cycle of the model dynamic optimization mechanism is every quarter, collecting newly added buoy and satellite remote sensing update data to recalibrate parameters. , , , and A dedicated optimization database has been established for individual cases of extreme weather caused by northward-moving typhoons. The dynamic adjustment range has been expanded to 1.5×10. -4 Up to 2.1×10 -4 Based on real-time verification data from the integrated meteorological operational platform, the model system bias is continuously corrected to ensure that the model is stable and applicable to the calculation of marine meteorological heat flux in mid-to-high latitude sea areas in the long term.

[0012] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows: 1. This invention subdivides marine droplets by particle size and establishes their correlation with wind speed and wave period for heat exchange characteristics. It constructs droplet generation rate models with sea ice modifiers for wind speed ranges and incorporates localized measured data calibration and quarterly dynamic optimization mechanisms. This solves the problem that traditional heat flux parameterization models do not fully consider droplet particle size differences, dynamic changes in sea state, and the influence of mid-to-high latitude sea ice, resulting in large calculation deviations in complex weather scenarios. This invention not only makes heat flux calculation more consistent with the actual air-sea exchange physical process, but also significantly improves the accuracy and stability of heat flux calculation under different scenarios such as ordinary strong winds, northward typhoons, and freezing periods. 2. This invention establishes a quarterly dynamic optimization mechanism, combined with a dedicated optimization library for extreme weather cases, to dynamically adjust key parameters to adapt to interannual sea state changes in the sea area. At the same time, it links with the integrated meteorological business platform to continuously correct system deviations by verifying data in real time. This solves the defects of traditional heat flux parameterized models, such as fixed parameters, difficulty in responding to dynamic changes in sea state and the special characteristics of extreme weather processes, which leads to insufficient stability and poor reliability of extreme scenario forecasts in long-term business applications. It not only enables the model to continuously adapt to the complex and changeable marine meteorological environment in mid-to-high latitude sea areas, but also significantly enhances the adaptability of heat flux forecasts for extreme weather processes. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the method steps in an embodiment of the present invention. Detailed Implementation

[0014] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0015] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0016] Please see the appendix Figure 1 This invention provides a method for constructing a parameterized model of marine meteorological heat flux under the marine droplet effect. The model construction method includes the following steps: Step 1: Collect localized observation data and sea state parameters of the target sea area and surrounding estuary areas, and perform missing value interpolation, outlier removal and format standardization on the data; Step 2: Divide marine droplets into three particle size ranges and establish a database of correspondences between particle size, wind speed, wave period, and heat exchange characteristics; Step 3: Construct a correlation model between droplet generation rate and localized sea state parameters for different wind speed ranges; Step 4: Derive the heat flux calculation equation that includes grain size weighting, wave period correction, and sea ice adjustment term; Step 5: Calibrate the model parameters using measured data from the target sea area; Step 6: Apply the calibrated model to different weather scenarios, calculate hourly heat flux, and verify accuracy. Step 7: Establish a dynamic optimization mechanism for the model to continuously correct systemic biases in the model.

[0017] Example 1, please refer to the appendix. Figure 1 Marine droplets were further subdivided by particle size and their heat exchange characteristics were correlated. The specific implementation process is as follows: Observation preparation: Three typical observation stations were selected in the mid-to-high latitude target sea area. A buoy observation system, a laser particle size analyzer and a wave period monitoring device were deployed. The observation period was set to 12 months and the data acquisition frequency was 1 hour / time. Sea state parameters such as sea surface wind speed, wave period and sea surface temperature were collected simultaneously. Particle size detection and classification: The particle size of the collected marine droplet samples was detected by a laser particle size analyzer. After removing outlier data, the particle size data was classified by K-means clustering algorithm, and finally the three categories of split droplets, splash droplets and foam droplets were determined. Calculation of heat exchange characteristic parameters: Based on the energy balance equation and combined with the observed atmospheric temperature difference between the sea surface and near-sea surface, the evaporation rate of droplets of various particle sizes is derived. ,in This refers to the net radiation flux at the sea surface, ranging from 400 to 600 W / m². 2 , The value is the sensible heat flux, ranging from 20 to 80 W / m. 2 , The density of seawater is 1025 kg / m³. 3 , The latent heat of vaporization is 2.45 × 10⁻⁶. 6 J / kg, the calculated evaporation rates for splitting droplets, splashing droplets, and foaming droplets are 0.0008-0.0012 kg / (m²). 2 ·s), 0.0013-0.0018kg / (m 2 ·s), 0.0019-0.0025kg / (m 2 ·s); By coupling a gravity settling model with a turbulent diffusion model, the atmospheric residence time of various droplets under different wind speeds and wave periods was calculated. The coupled gravity settling and turbulent diffusion model was employed. ,in The boundary layer height is taken as 1000m. For gravity settling velocities, the velocity is 0.8-1.2 m / s for cracking droplets, 0.3-0.7 m / s for splashing droplets, and 0.05-0.2 m / s for foam droplets. Let be the turbulent diffusion coefficient, taken as 0.5m. 2 / s, The spatial variation rate of turbulent diffusion (characterizing the non-uniformity of turbulent diffusion in the vertical direction, with a value ranging from 0.001 to 0.003 m) -1 The atmospheric residence times of the three types of droplets were calculated to be 800-1200s, 1500-2500s, and 4000-6000s, respectively. The least squares method was used to fit the observed data and the theoretical model, and the fitting coefficients in the heat exchange coefficient formula were obtained. Substituting the observed wind speed, wave period, and measured heat exchange data, we obtain the fitting coefficients for the split droplet: , =0.65、 =0.32, splashing: =0.15、 =0.58、 =0.28, Mote Droplet: =0.22、 =0.49、 =0.21.

[0018] Relationship Database Construction and Validation: Integrating particle size classification results, heat exchange coefficient, wind speed, and wave period data, a relationship database corresponding to particle size, wind speed, wave period, and heat exchange characteristics was established. Leave-one-out cross-validation was used to verify the accuracy of the relationship database and ensure goodness of fit. ≥0.92, meeting the application requirements under different sea conditions.

[0019] Example 2, please refer to the appendix. Figure 1 A droplet formation rate model with a sea ice moderating term was constructed for each wind speed range. The specific implementation process is as follows: Data collection: Collect buoy observation data (including wind speed of 0-50 m / s and wave period of 2-15 s) and sea ice coverage data retrieved from satellite remote sensing for the target sea area from 2020 to 2024 (based on the fusion processing of high-precision global coastline data and ice condition monitoring products, with values ​​ranging from 0 to 1). Divide the data into quarterly subsets to ensure that the data for medium and low wind speeds (<30 m / s) and high wind speeds (≥30 m / s) are evenly distributed.

[0020] Wind speed range division and data screening: Using the statistical breakpoint analysis method and combined with the frequency of strong winds in the target sea area, 30 m / s was determined as the wind speed division threshold. 1200 sets of valid data for medium and low wind speeds and 450 sets of valid data for high wind speeds were screened. Outliers in sea ice coverage (>1 or <0) were removed, and 1580 sets of valid data were retained.

[0021] Model Construction: In the low to medium wind speed range, wave period was used as a correction factor, and a negative feedback term of sea ice coverage was introduced. The model was obtained by fitting the data through multiple linear regression. ,in =2.3×10 -5 In the high-wind-speed range, considering the enhancing effect of sea ice on sea surface roughness, a nonlinear fitting method is used to construct the model. ,in =1.8×10 -4 .

[0022] Model validation: Measured data from January to March 2024 were selected as the validation set. The data were substituted into the model to calculate the droplet formation rate, and compared with the measured values. The results passed the significance test. (≥0.85) Validation of effectiveness: The model calculation error was reduced by 18%-25% for ice-covered data with sea ice coverage of 0.3-0.8, and it was adapted to the differences in sea conditions between ice-covered and non-ice-covered periods in the target sea area.

[0023] Example 3, please refer to the appendix. Figure 1 The dynamic optimization mechanism, which combines quarterly dynamic optimization with a dedicated optimization library for extreme weather, is implemented as follows: Data update and collection: At the end of each quarter, collect newly added buoy observation data (wind speed, wave period, heat flux), updated sea ice coverage and sea surface temperature data from satellite remote sensing in the target sea area, and simultaneously organize real-time verification data from the integrated meteorological business platform (including hourly heat flux calculation deviation records) to ensure that no less than 300 sets of new valid data are added each quarter.

[0024] Parameter calibration: Using the least squares method, with the goal of minimizing the root mean square error between the calculated and measured values, the fitting coefficients of the heat exchange coefficient were recalibrated. , , ) and droplet generation rate model coefficients ( , Generate a quarterly calibration parameter set to ensure that the root mean square error at low to medium wind speeds is ≤15W / m after calibration. 2 ≤25W / m at high wind speeds 2 .

[0025] Construction of an extreme weather case database: Thirty sets of extreme weather cases, including northward-moving typhoons, were selected. Sea state parameters and heat flux data for high wind speeds above 30 m / s were extracted to establish a dedicated optimized database. The dynamic adjustment range has been expanded to 1.5×10. -4 Up to 2.1×10 -4 The parameter weights in the high wind speed range are optimized separately to adapt to the droplet generation characteristics under extreme sea conditions.

[0026] Deviation correction and stability verification: Set system deviation threshold (medium and low wind speed ≥12W / m) 2 High wind speed ≥20W / m 2 When real-time test data shows that the deviation exceeds the threshold, parameter correction is automatically triggered. Stability verification is carried out every six months. By comparing the model calculation accuracy in different quarters, it is ensured that the model is suitable for long-term operational calculations in mid-to-high latitude sea areas, with an annual stability improvement of more than 30%.

[0027] Specifically, the localized observation data covers the entire seasonal cycle of the target sea area, ensuring the capture of sea state changes in different seasons. The spatial coverage precisely matches the geographical boundaries of the target sea area and surrounding outlets, avoiding data redundancy or coverage gaps. In the preprocessing stage, the missing value interpolation and outlier removal methods are designed specifically for the spatiotemporal continuity of marine observation data. The format is standardized and unified to a common network data format, which not only ensures data compatibility but also provides a unified data foundation for subsequent multi-source data fusion modeling, ensuring that the data processing results can directly support subsequent model construction.

[0028] Specifically, droplet size classification is not based solely on particle size, but rather on a combination of measured sea conditions in the target sea area and reference to the physical mechanisms of droplet generation (such as wave breaking intensity and airflow shearing). This ensures that the classification results conform to the actual air-sea exchange scenario. During the quantification of heat exchange characteristics, the differences in the motion trajectories of droplets of different sizes in the atmosphere are fully considered. Through dynamic fitting of measured data and theoretical models, the established correlation database is not only applicable to normal sea conditions, but also adaptable to the complex and variable wind and wave field environment of the target sea area, providing accurate characteristic support for subsequent generation rate modeling.

[0029] Specifically, the wind speed intervals are divided based on the frequency of strong winds in the target sea area and the critical wind speed characteristics for droplet formation. This ensures that the interval division can accurately distinguish the differences in droplet formation under different wind field intensities. The design of the sea ice adjustment term fully considers the sea ice distribution characteristics during the freezing period in the target sea area, taking into account the influence of sea ice on sea surface roughness and wave motion. Differentiated function forms are used to adapt the inhibitory effect of sea ice on droplet formation under low and medium wind speeds and the enhancing effect under high wind speeds, enabling the model to accurately respond to the differences in sea state between the freezing period and the non-freezing period.

[0030] Specifically, each component of the heat flux calculation equation is designed to suit the sea state characteristics of the target sea area: the air-sea interface heat flux uses a mature algorithm to ensure the accuracy of the basic calculation; the particle size weight is determined according to the contribution ratio of droplets of different particle sizes in air-sea heat exchange; the wave period correction term responds to the impact of dynamic changes in sea state on heat exchange; the sea ice regulation term adapts to the special environment during the freezing period; and the equation comprehensively integrates the heat flux of direct air-sea interface exchange and indirect droplet exchange through the superposition of sub-terms, ensuring that it can fully characterize the complex process of air-sea heat exchange in the target sea area.

[0031] Specifically, the measured data used for calibration covers different seasons and weather scenarios (ordinary strong winds, northward typhoons, and the freezing period) in the target sea area, ensuring that the calibration samples are sufficiently representative and can support the model's accuracy performance in various scenarios. During the calibration process, the least squares method is used to focus on optimizing the fit between the core parameters and the measured values. At the same time, in response to the problem of sea ice cover changing sea surface characteristics during the freezing period, a sea ice roughness correction term is specially designed. By dynamically adjusting the parameters to adapt to the sea conditions during the freezing period, the calculation error during the freezing period is further reduced, and the reliability of the model in special scenarios is improved.

[0032] Specifically, the accuracy verification selected different weather scenarios to comprehensively cover high-impact weather types in the target sea area, ensuring that the verification results can reflect the model's comprehensive performance in actual operations. The selected verification indicators evaluate the model's accuracy from multiple dimensions, including error magnitude (root mean square error, mean absolute error) and correlation (correlation coefficient), avoiding the limitations of a single indicator. By comparing with measured data and existing solutions, not only can the model's computational accuracy be quantified, but the technical advantages of the model's localized optimization for the target sea area can also be highlighted, providing sufficient accuracy support for the model's operational application.

[0033] Specifically, the model's dynamic optimization mechanism has a quarterly optimization cycle, which is set with reference to the seasonal variation patterns of sea conditions in the target sea area. This ensures that the parameters can respond promptly to seasonal changes in the marine environment and avoids the decrease in accuracy caused by parameter stagnation. The extreme weather-specific optimization library focuses on weather types that have a significant impact on the target sea area, such as northward-moving typhoons. By collecting historical sea condition characteristic data, the core parameters in high-wind-speed ranges are optimized in a targeted manner. Combined with real-time verification data from the integrated meteorological business platform, a closed loop of "data update - parameter calibration - deviation correction" is formed to ensure that the model is adapted to the marine meteorological environment of the target sea area in the long term.

[0034] Working principle: First, localized observational data and sea state parameters, including buoy observations, satellite remote sensing, and sea ice coverage, were collected from the target sea area and surrounding estuary regions. Missing values ​​were filled using the mean interpolation method for adjacent observations, and outliers were removed using the three-times standard deviation method. The data was then uniformly converted to a common network data format for preprocessing. Next, marine droplets were divided into three particle size ranges: split droplets, splash droplets, and foam droplets. The heat exchange coefficient was obtained by fitting the observational data with a theoretical model, and a database of correspondences between particle size, wind speed, wave period, and heat exchange characteristics was established. Then, ranges were divided according to low-to-medium wind speed and high wind speed. A droplet generation rate correlation model was constructed based on sea ice coverage, and a sea ice adjustment term was introduced to adapt to sea states during freezing and non-freezing periods. Finally, based on the Moning-Obukhov similarity theory, a model including particle size weights, wave period corrections, and... The heat flux calculation equation for the sea ice moderating term integrates the air-sea interface heat flux and droplet heat flux. Then, using measured data from the target sea area from 2020 to 2024, the core parameters of the model are calibrated using the least squares method. During the freezing period, a sea ice roughness correction term is introduced to further reduce the error. Subsequently, the calibrated model is applied to different weather scenarios such as ordinary strong winds, northward typhoons, and the freezing period to calculate hourly heat flux. Multiple indicators such as root mean square error, mean absolute error, and correlation coefficient are used to verify the accuracy. Finally, a dynamic optimization mechanism for parameter calibration is established every quarter. New observation data is collected to recalibrate the parameters. A dedicated optimization library is established for northward typhoons to expand the adjustment range of key parameters. Combined with real-time verification data from the integrated meteorological business platform, the system deviation is continuously corrected. At this point, the entire workflow is completed.

[0035] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, any modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for constructing a parameterized model of marine meteorological heat flux under the marine spray effect, characterized in that, The model construction method includes the following methods: Step 1: Collect localized observation data and sea state parameters of the target sea area and surrounding estuary areas, and perform missing value interpolation, outlier removal and format standardization on the data; Step 2: Divide marine droplets into three particle size ranges and establish a database of correspondences between particle size, wind speed, wave period, and heat exchange characteristics; Step 3: Construct a correlation model between droplet generation rate and localized sea state parameters for different wind speed ranges; Step 4: Derive the heat flux calculation equation that includes grain size weighting, wave period correction, and sea ice adjustment term; Step 5: Calibrate the model parameters using measured data from the target sea area; Step 6: Apply the calibrated model to different weather scenarios, calculate hourly heat flux, and verify accuracy. Step 7: Establish a dynamic optimization mechanism for the model to continuously correct systemic biases in the model.

2. The method for constructing a parameterized model of marine meteorological heat flux under the marine spray effect according to claim 1, characterized in that: In step one, the localized observation data includes buoy observation data, satellite remote sensing data, sea ice coverage data, and observation data from ground automatic stations and offshore platforms in the target sea area. The buoy observation data includes sea surface wind speed, significant wave height, wave period, sea surface temperature, and near-sea atmospheric temperature and humidity. The satellite remote sensing data includes radiation data from the geostationary interferometric infrared detector of Fengyun-4A satellite and atmospheric motion vector data from the Kuihua-8 satellite. The sea ice coverage data is based on high-precision global coastline data and ice condition monitoring products. The time resolution of the preprocessed data is no less than 1 hour, and the spatial coverage is the entire target sea area. Missing values ​​are interpolated using the mean of adjacent time observations. Outlier removal is done using the 3x standard deviation method. The format is uniformly converted to a common network data format.

3. The method for constructing a parameterized model of marine meteorological heat flux under the marine spray effect according to claim 1, characterized in that: In step two, the three particle size ranges are 50μm-100μm split droplets, 10μm-50μm splash droplets, and <10μm foam droplets. The heat exchange coefficient is obtained by fitting observation data with a theoretical model, and the expression is as follows: ； in, No. The heat exchange coefficient of droplets of similar particle size , , For the fitting coefficients, the split droplet is: , =0.65、 =0.32, splashing: =0.15、 =0.58、 =0.28, Mote Droplet: =0.22、 =0.49、 =0.21, At a height of 10m, The wave period is given, the evaporation rate is derived from the energy balance equation, and the atmospheric residence time is calculated based on the coupled model of gravity settling and turbulent diffusion.

4. The method for constructing a parameterized model of marine meteorological heat flux under the marine spray effect according to claim 1, characterized in that: In step three, the wind speed intervals specifically refer to low to medium wind speeds (U < 30 m / s) and high wind speeds (U ≥ 30 m / s). The droplet generation rate model under low to medium wind speeds is as follows: ； in, =2.3×10 -5 , Sea ice coverage, 0 ≤ ≤1, the droplet formation rate model under high wind speed is: ； in, =1.8×10 -4 The model's effectiveness was verified through a significance test.

5. The method for constructing a parameterized model of marine meteorological heat flux under the marine spray effect according to claim 1, characterized in that: In step four, the equation for calculating heat flux is: ； in, For total heat flux, For air-sea interface heat flux, For the first Weighting coefficients for droplet size, split droplets =0.45, splashing =0.35, foam droplet =0.20, For the first The generation rate of droplets of similar size is taken at low to medium wind speeds. Take during high wind speed , For the first The heat exchange coefficient of droplets of similar particle size Sea surface temperature, This refers to the near-sea surface atmospheric temperature.

6. The method for constructing a parameterized model of marine meteorological heat flux under the marine spray effect according to claim 1, characterized in that: In step five, the measured data are the observed heat flux values ​​of buoys and offshore platforms in the target sea area from 2020 to 2024. The root mean square error between the model-calculated values ​​and the measured values ​​is minimized using the least squares method. The adjusted parameters include those from step two. , , With step three , After calibration, the root mean square error of heat flux calculation at low and medium wind speeds is ≤15W / m². 2 Root mean square error ≤25W / m under high wind speed 2 , Introducing a sea ice roughness correction term during the freezing period .

7. The method for constructing a parameterized model of marine meteorological heat flux under the marine spray effect according to claim 1, characterized in that: In step six, different weather scenarios include ordinary strong winds, northward-moving typhoons, and the ice-covered period. The input sea state parameters include wind speed, wave period, sea ice coverage, sea surface temperature, and atmospheric temperature and humidity. The accuracy verification indicators include root mean square error, mean absolute error, and correlation coefficient. Among them, the root mean square error is ≤12W / m in the medium and low wind speed scenario. 2 Mean absolute error ≤ 9W / m 2 The correlation coefficient is ≥0.90, and the root mean square error is ≤20W / m² in high wind speed scenarios. 2 Mean absolute error ≤15W / m 2 The correlation coefficient is ≥0.85, and the root mean square error during the freezing period is ≤18W / m. 2 Mean absolute error ≤13W / m 2 The correlation coefficient is ≥0.

88.

8. The method for constructing a parameterized model of marine meteorological heat flux under the marine spray effect according to claim 1, characterized in that: In step seven, the optimization cycle of the model dynamic optimization mechanism is every quarter, during which newly added buoy and satellite remote sensing update data are collected to recalibrate the parameters. , , , and A dedicated optimization database has been established for individual cases of extreme weather caused by northward-moving typhoons. The dynamic adjustment range has been expanded to 1.5×10. -4 Up to 2.1×10 -4 .

Images