North pole sea fog model training method and device, storage medium and electronic equipment

Abstract

The invention provides a training method and device of a north pole sea fog model, a storage medium and electronic equipment. The electronic equipment obtains the sea fog visibility of the research area at a plurality of historical time points by using a pre-constructed PWRF model; for each historical time point, weighting the original meteorological data, the original sea surface temperature data and the original sea ice data of the plurality of meteorological models at the historical time point to obtain sample meteorological data, sample sea surface temperature data and sample sea ice data at the historical time point; and training the to-be-trained model through the synthesized data to obtain the north pole sea fog model. Therefore, the problem that sea fog visibility data with long time sequence, large range area and high spatial resolution cannot be obtained at present is solved, and input data formed by fusing multiple models has high confidence, so that the trained north pole sea fog model can effectively capture complex mechanisms such as humidity threshold effect, interaction of wind speed and sea ice and the like; and the sea fog prediction precision is improved.

Description

Technical Field

[0001] The present application relates to the field of meteorology, and more particularly, to a method, device, storage medium, and electronic device for training an Arctic sea fog model. Background Art

[0002] With global warming, the first ice-free day in the Arctic is expected to occur before 2030, and the navigability of the Arctic shipping route has been significantly improved. However, due to the unpredictability of weather and climate factors, the Arctic region is regarded as one of the most challenging environments for maritime navigation, and the safe navigability urgently needs to be further improved. The increase in temperature leads to enhanced sea surface evaporation, and sea fog frequently appears on the shipping route, seriously affecting navigation safety and becoming a major risk factor for maritime transportation. Although sea fog monitoring is crucial for ensuring Arctic shipping, there is a lack of sufficient observational data in the Arctic region, making it difficult to meet the needs of sea fog monitoring on the shipping route and becoming a bottleneck for navigation safety.

[0003] Existing Arctic sea fog monitoring technologies face many challenges, mainly including limited spatial resolution, insufficient coverage of the monitoring network, insufficient ability to distinguish sea fog from low clouds, poor real-time performance, high cost, difficult logistics support, lack of monitoring of the vertical structure of the fog layer, and low simulation accuracy of the sea ice-atmosphere interaction. For example, traditional global climate models (such as CMIP6) have a low resolution (greater than 50 km), making it difficult to accurately simulate the complex terrain and microscale meteorological processes in the Arctic, resulting in insufficient simulation accuracy of the sea fog generation and dissipation mechanisms. At the same time, meteorological observation stations in the Arctic region are sparse, especially in the sea ice-covered areas, resulting in insufficient observational data and making it difficult to support high-precision sea fog monitoring and forecasting. In addition, existing remote sensing technologies and numerical models have difficulties in distinguishing sea fog from low clouds, and the update frequency of satellite remote sensing and ship observation data is low, unable to meet the real-time safety needs of the Arctic shipping route. Moreover, the extreme environment limits the deployment and maintenance of meteorological observation equipment, resulting in high costs and increased difficulty in logistics support. Finally, existing technologies also have obvious accuracy problems in monitoring the vertical structure of the fog layer and simulating the sea ice-atmosphere interaction, mainly manifested as a vertical resolution greater than 50 m, making it difficult to meet the needs of sea fog monitoring on the shipping route.

[0004] Therefore, how to improve the accuracy of sea fog identification and the reliability of visibility prediction in the case of scarce data has become an urgent problem to be solved. Summary of the Invention

[0005] In order to overcome at least one deficiency in the prior art, the present application provides a method, device, storage medium, and electronic device for training an Arctic sea fog model, specifically including:

[0006] In a first aspect, the present application provides a method for training an Arctic sea fog model, the method including:

[0007] Obtain the sea fog visibility in the study area at multiple historical time points using a pre - constructed Polar Weather Research and Forecasting PWRF model;

[0008] For each of the historical time points, weight the original meteorological data, original sea surface temperature data, and original sea ice data of multiple meteorological models at the historical time point to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data at the historical time point;

[0009] Train the model to be trained using the meteorological data, sample sea surface temperature data, sample sea ice data, and sea fog visibility at the multiple historical time points to obtain an Arctic sea fog model. Among them, the sea fog visibility at each historical time point is used as a label during training, and the Arctic sea fog model is used to predict the future sea fog visibility in the study area.

[0010] In a second aspect, the present application provides a training device for an Arctic sea fog model. The device includes:

[0011] A sample preparation module for obtaining the sea fog visibility in the study area at multiple historical time points using a pre - constructed Polar Weather Research and Forecasting PWRF model;

[0012] The sample preparation module is further configured to, for each of the historical time points, weight the original meteorological data, original sea surface temperature data, and original sea ice data of multiple meteorological models at the historical time point to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data at the historical time point;

[0013] A model training module for training the model to be trained using the meteorological data, sample sea surface temperature data, sample sea ice data, and sea fog visibility at the multiple historical time points to obtain an Arctic sea fog model. Among them, the sea fog visibility at each historical time point is used as a label during training, and the Arctic sea fog model is used to predict the future sea fog visibility in the study area.

[0014] In a third aspect, the present application provides a storage medium storing a computer program, which, when executed by a processor, implements the training method of the Arctic sea fog model described above.

[0015] In a fourth aspect, the present application provides an electronic device, which includes a processor and a memory. The memory stores a computer program, which, when executed by the processor, implements the training method of the Arctic sea fog model described above.

[0016] Compared with the prior art, the present application has the following beneficial effects:

[0017] This application provides a training method, device, storage medium, and electronic device for an Arctic sea fog model. The electronic device uses a pre-constructed Polar Weather Research and Forecasting (PWRF) model to obtain the sea fog visibility in the study area at multiple historical time points. For each historical time point, the original meteorological data, original sea surface temperature data, and original sea ice data of multiple meteorological models at the historical time point are weighted to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data at the historical time point. Finally, the model to be trained is trained with the meteorological data, sample sea surface temperature data, sample sea ice data, and sea fog visibility at multiple historical time points to obtain the Arctic sea fog model. Among them, the sea fog visibility at each historical time point is used as the label during training, and the Arctic sea fog model is used to predict the future sea fog visibility in the study area. In this way, by synthesizing the sea fog visibility using the PWRF model as the label during model training and fusing the climate variables in multiple meteorological models to obtain the input data during model training, it not only solves the problem that long-time series, large-area, and high-spatial-resolution sea fog visibility data cannot be obtained at the present stage, but also the input data formed by the fusion of multiple models has a high confidence level, enabling the trained Arctic sea fog model to effectively capture complex mechanisms such as humidity threshold effects and the interaction between wind speed and sea ice, thereby significantly improving the prediction accuracy of sea fog. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the following will briefly introduce the drawings required for use in the embodiments. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0019] Figure 1 Schematic flowchart of the training method for the Arctic sea fog model provided by the embodiment of this application;

[0020] Figure 2 Schematic detailed view of the training method for the Arctic sea fog model provided by the embodiment of this application;

[0021] Figure 3 Complete principle schematic view of the training method for the Arctic sea fog model provided by the embodiment of this application;

[0022] Figure 4 Schematic structural view of the training device for the Arctic sea fog model provided by the embodiment of this application;

[0023] Figure 5 Schematic structural view of the electronic device provided by the embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] To make the objectives, technical solutions, and advantages of the embodiments of this application (hereinafter simply referred to as "these embodiments") clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are some, but not all, of the embodiments of this application. Components of the embodiments of this application that are generally described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0025] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of this application that is claimed, but merely represents selected embodiments of this application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of this application without creative efforts fall within the scope of protection of this application.

[0026] It should be noted that like reference numerals and letters denote like items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0027] In the description of this application, it should be noted that the terms "first", "second", "third", etc. are only used for descriptive distinction and cannot be construed as indicating or implying relative importance. In addition, the term "comprising", "including" or any other variation thereof is intended to cover a non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the said element.

[0028] Based on the above statements, as introduced in the background art, sea fog monitoring is crucial for ensuring Arctic shipping. However, there is a lack of sufficient observational data in the Arctic region, making it difficult to meet the needs of sea fog monitoring for shipping lanes. Therefore, how to improve the accuracy of sea fog identification and the reliability of visibility prediction in the case of scarce relevant data has become an urgent problem to be solved.

[0029] Based on the discovery of the above technical problems, the inventors have, through creative efforts, proposed the following technical solutions to solve or improve the above problems. It should be noted that the defects existing in the above prior art solutions are the results obtained by the inventors through practice and careful research. Therefore, the process of discovering the above problems and the solutions proposed by the embodiments of this application below for the above problems should be the contributions made by the inventors to this application during the invention creation process and should not be construed as well-known technical content to those skilled in the art.

[0030] In view of this, this embodiment provides a method for training an Arctic sea fog model. As Figure 1 shown, the method includes:

[0031] S1, obtaining the sea fog visibility of the research area at multiple historical time points by using a pre-constructed Polar Weather Research and Forecasting (PWRF) model.

[0032] S2, for each historical time point, weighting the original meteorological data, original sea surface temperature data, and original sea ice data of multiple meteorological models at the historical time point to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data at the historical time point.

[0033] S2, training the model to be trained with the sample meteorological data, sample sea surface temperature data, sample sea ice data, and sea fog visibility at multiple historical time points to obtain an Arctic sea fog model.

[0034] Among them, the sea fog visibility at each historical time point is used as a label during training, and the Arctic sea fog model is used to predict the future sea fog visibility of the research area.

[0035] In this way, in this embodiment, the sea fog visibility is synthesized by the PWRF model as the label during model training, and the climate variables in multiple meteorological models are fused to obtain the input data during model training, which not only solves the problem that long-time series, large-scale area, and high-spatial-resolution sea fog visibility data cannot be obtained at the present stage, but also the input data formed by multi-model fusion has a high confidence level, enabling the trained Arctic sea fog model to effectively capture complex mechanisms such as humidity threshold effects and interactions between wind speed and sea ice, thereby significantly improving the prediction accuracy of sea fog.

[0036] For the method for training the Arctic sea fog model provided in this embodiment, the electronic device implementing this method can be, but is not limited to, a mobile terminal, a tablet computer, a laptop computer, a desktop computer, a server, etc., as long as it can provide sufficient computing power for model training. In this embodiment, when the electronic device is a server, the server can be a single server or a server group. The server group can be centralized or distributed (for example, the server can be a distributed system). In some embodiments, the server can be local or remote relative to the user terminal. In some embodiments, the server can be implemented on a cloud platform; by way of example only, the cloud platform can include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an inter-cloud, a multi-cloud, etc., or any combination thereof. In some embodiments, the server can be implemented on an electronic device having one or more components.

[0037] To make the solution provided in this embodiment clearer, the following uses a server as the electronic device implementing this method to elaborate on each step in the Figure 1 shown method in detail. However, it should be understood that the operations in the flowchart may not be implemented in sequence, and steps without a logical context relationship can be reversed or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart or remove one or more operations from the flowchart under the guidance of the content of this application.

[0038] As Figure 1 shown, this method includes:

[0039] S1. Obtain the sea fog visibility of the study area at multiple historical time points using the pre-constructed Polar Weather Research and Forecasting (PWRF) model.

[0040] It can be understood that the training sample set applicable to training the model in this embodiment includes the meteorological sample information of the study area at different historical time points. Each piece of meteorological sample information includes the sample meteorological data, sample sea surface temperature data, sample sea ice data, and sea fog visibility as the label of the study area.

[0041] In addition, the study area in this embodiment can be the Russian Northern Sea Route section of the Northeast Passage in the Arctic. This area is a region with significantly improved navigability under the background of global warming, and it is also an important area where sea fog occurs frequently and affects navigation safety. Of course, this embodiment is applicable to any area where sea fog may occur, not limited to the Russian Northern Sea Route section of the Northeast Passage in the Arctic.

[0042] Since long-term, large-scale, and high-spatial-resolution sea fog visibility data cannot be obtained at the present stage, in this embodiment, the sea fog visibility of the study area at different historical time points is retrieved through the existing data that can affect the generation, maintenance, and dissipation of sea fog. Based on this inventive concept, as Figure 2 shown, the following alternative implementation manners of step S1 are provided in this embodiment:

[0043] S1-1. For each historical time point, obtain the cloud water mixing ratio, air temperature, and air pressure of the study area at the historical time point through the PWRF model.

[0044] In this embodiment, the server can obtain the original meteorological data, original sea surface temperature data, and original sea ice data of the research area at historical time points; process the original meteorological data, original sea surface temperature data, and original sea ice data at historical time points through a pre-constructed Polar Weather Research and Forecasting (PWRF) model to obtain the cloud water mixing ratio, air temperature, and air pressure at historical time points.

[0045] The above original meteorological data can be obtained through NCEP FNL reanalysis data (2000 - 2015), the original sea surface temperature data can be obtained by integrating GHRSST sea surface temperature data (2000 - 2015), and the original sea ice data can be obtained through NOAA / NSIDC sea ice data. Among them, the NCEP FNL data is a global reanalysis dataset released by the National Centers for Environmental Prediction of the United States, with a time span from 2000 to 2015. These data have undergone assimilation processing of multiple observational data and have high spatio-temporal resolution, capable of providing detailed meteorological data such as wind speed, wind direction, humidity, cloud cover, etc. The GHRSST (Group for High Resolution Sea Surface Temperature) sea surface temperature dataset is sea surface temperature data obtained based on satellite remote sensing technology, with the same time span from 2000 to 2015. These data have high spatio-temporal resolution and can provide accurate sea surface temperature information. The NOAA / NSIDC sea ice data is a sea ice dataset jointly released by the National Oceanic and Atmospheric Administration (NOAA) and the National Snow and Ice Data Center (NSIDC) of the United States. These data include information such as sea ice concentration and thickness, with a time span from 2000 to 2015.

[0046] In addition, it should be noted that the above PWRF model is a weather research and forecasting model specifically designed for polar regions, aiming to improve the problem of insufficient accuracy of traditional numerical weather prediction models in polar regions, especially in the Arctic region, due to the special geographical environment and scarce observational data. In this example, the microphysical parameters of the model are optimized according to the Arctic climate characteristics, that is, the PWRF model includes a single-phase six-class cloud microphysics scheme (WRF Single-Moment 6-Class Microphysics Scheme, WSM6) model, a rapid radiative transfer model (Rapid Radiative Transfer Model, RRTM), a convective parameterization scheme Grell-Freitas model, a land surface process model (Land Surface Model, Noah LSM) model, and a surface layer and planetary boundary layer scheme (Mellor-Yamada-Nakanishi-Niino Level 3, MYNN3) model. Next, each sub-model in the PWRF model will be described in more detail:

[0047] The WSM6 model is a six-class cloud microphysics scheme used to describe cloud microphysical processes, which can accurately simulate different types of cloud particles such as cloud droplets, ice crystals, snow grains, graupel, supercooled water droplets, and large raindrops. Therefore, using this model can well describe the formation, development, and dissipation processes of clouds. Especially in polar regions, the microphysical characteristics of clouds have an important impact on weather and climate.

[0048] The RRTM model is a fast and accurate radiative transfer model that can efficiently calculate the transmission processes of long-wave and short-wave radiation. In polar regions, due to the drastic changes in solar radiation, the impact of radiative transfer on weather and climate is particularly important, and the RRTM model can provide reliable radiative calculation results.

[0049] The Grell-Freitas model is an advanced convective parameterization scheme that can simulate the initiation, development, and dissipation processes of convective activities. In polar regions, although convective activities are not as frequent as in tropical regions, they still have a significant impact on weather under certain special conditions, and these processes can be better captured through the Grell-Freitas model.

[0050] The Noah LSM model is a land surface process model that can simulate the influence of factors such as soil temperature, soil moisture, and vegetation cover on the surface energy balance. In polar regions, land surface processes play an important role in the changes of weather and climate, and the Noah LSM model can provide detailed information on the land surface state.

[0051] The MYNN3 model is an advanced surface layer and planetary boundary layer scheme that can simulate the turbulent exchange process in the atmospheric boundary layer. In polar regions, the turbulent exchange of the atmospheric boundary layer has an important impact on weather and climate changes. The MYNN3 model can provide accurate information on the state of the atmospheric boundary layer.

[0052] In addition, when using the PWRF model, this embodiment sets a 1 km horizontal resolution, 38 vertical layers, and a nested grid covering the Arctic Northeast Passage (70°N-80°N, 30°E-180°E). Among them, the 1 km horizontal resolution can more finely capture the complex terrain and meteorological features of the polar region, such as glaciers, mountains, and ocean fronts, so as to accurately simulate the polar weather system. The 38 vertical layers can describe the changes in the atmospheric stratification in more detail, especially for the precise prediction of key areas such as the atmospheric boundary layer and the tropopause that are unique to the polar region.

[0053] Based on the above description of the cloud-water mixing ratio, temperature and air pressure in the study area at historical time points, we will continue to Figure 2 Step S1-2 in the following is explained:

[0054] S1-2, based on the cloud-water mixing ratio, temperature and air pressure at the historical time point, invert the sea fog visibility in the study area at the historical time point.

[0055] In this embodiment, the server can obtain the liquid water content at the historical time point based on the cloud-water mixing ratio, temperature and air pressure at the historical time point; and obtain the sea fog visibility at the historical time point based on the liquid water content at the historical time point. Specifically, the relationship between the cloud-water mixing ratio, temperature and air pressure at the historical time point and the liquid water content at the historical time point is:

[0056]

[0057] The relationship between the liquid water content at a historical point in time and the sea fog visibility at a historical point in time is:

[0058]

[0059] Where LWC represents liquid water content, VIS represents sea fog visibility, and Q cloud represents the cloud water mixing ratio, P represents the air pressure, and T represents the temperature.

[0060] Thus, in this embodiment, the server can obtain the original meteorological data, original sea surface temperature data, and original sea ice data at different historical time points in the study area; then, process these original data through a pre-constructed polar weather research and forecasting model (including the single-phase six-category cloud microphysics scheme WSM6 model, the rapid radiative transfer model RRTM, the convective parameterization scheme Grell-Freitas model, the land surface process model Noah LSM model, and the surface layer and planetary boundary layer scheme MYNN3 model) to obtain the cloud water mixing ratio, air temperature, and air pressure at each historical time point in the study area; and calculate the corresponding liquid water content at the historical time point according to these meteorological parameters using a specific relational formula; finally, invert the sea fog visibility at each historical time point in the study area according to the relational formula between the liquid water content and the sea fog visibility, so as to determine the sample meteorological data, sample sea surface temperature data, sample sea ice data, and sea fog visibility at this historical time point as meteorological sample information.

[0061] Therefore, in the above-described embodiment, considering that at the present stage, long-time series, large-scale area, and high-spatial-resolution sea fog visibility data cannot be obtained, it means that the sea fog visibility data used as labels cannot be simply obtained. Therefore, the PWRF model is used to simulate the sea fog visibility at multiple historical time points in the historical period (2000 - 2015).

[0062] Based on the description of the sea fog visibility inversion method in the above embodiment, the following will describe Figure 1 step S2 in

[0063] S2, for each historical time point, weight the original meteorological data, original sea surface temperature data, and original sea ice data of multiple meteorological models at the historical time point to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data at the historical time point.

[0064] It should be understood that the sample meteorological data in the above steps may include data such as the 2m air temperature, 925hPa air temperature, near-surface relative humidity, precipitation information, and wind speed information in the study area. Among them, the 2m air temperature represents the air temperature at a height of 2 meters above the ground and is an important indicator for measuring the near-surface thermal condition; the 925hPa air temperature represents the air temperature at an altitude of about 750 meters and reflects the temperature distribution of the upper atmosphere; the near-surface relative humidity represents the ratio of the actual water vapor content in the near-surface air to the saturated water vapor content and is one of the important conditions for the formation of sea fog; the precipitation information includes precipitation amount and precipitation type (such as rain, snow, etc.) and directly affects sea surface evaporation and air humidity; the wind speed information represents the speed of air flow and affects the diffusion and distribution of sea fog. Therefore, the combined action of these meteorological parameters affects the occurrence and development of sea fog. By collecting these parameters and conducting comprehensive analysis, the visibility of sea fog can be predicted more accurately.

[0065] At the same time, the above sample sea surface temperature data reflects the temperature condition of the sea surface in the study area and is an important indicator for measuring the heat exchange of seawater. It should be understood that the change in sea surface temperature will affect the seawater evaporation rate and thus affect the water vapor content in the air. The sample sea ice data includes sea ice concentration and thickness and is used to describe the sea ice coverage in the study area. Similarly, it should be understood that the presence of sea ice not only changes the thermal characteristics of the sea surface but may also be a potential triggering factor for the formation of sea fog.

[0066] For the above sample meteorological data, sample sea surface temperature data, and sample sea ice data, it was found during the research process that the sparse original observation data in the Arctic region led to the problem of incomplete input. In response to this, in this embodiment, the above sample meteorological data, sample sea surface temperature data, and sample sea ice data are obtained through the method of dynamic weight synthesis of the model, thereby improving the reliability of prediction and data coverage.

[0067] Specifically, the server can assign weights to multiple meteorological models according to the accuracy of the multiple meteorological models, where the weight of each meteorological model is positively correlated with its own accuracy; finally, according to the weights assigned to the multiple meteorological models, the original meteorological data, original sea surface temperature data, and original sea ice data at historical time points of the multiple meteorological models are weighted to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data in the study area at historical time points.

[0068] It can be understood that in this embodiment, the original meteorological data, original sea surface temperature data, and original sea ice data of multiple meteorological models at the same historical time point in the study area are obtained; then, according to the accuracies of these meteorological models, a weight that is positively correlated with the accuracy is assigned to each meteorological model; and using these assigned weights, the original data from different meteorological models are weighted to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data of the study area at this historical time point. In this way, the advantages of multiple meteorological models can be fully utilized, and the errors or biases that may be brought by a single model can be reduced. That is, by adjusting and optimizing the climate field synthesized by multiple models through the multi-model weight method, the systematic bias that may be caused by the traditional equal-weight method can be significantly reduced, thereby improving the accuracy and reliability of the finally obtained sample data.

[0069] Based on the above description of Figure 1 step S2 shown, continue to refer to Figure 1 , and next, Figure 1 step S3 in

[0070] S3. The model to be trained is trained with the meteorological data, sample sea surface temperature data, sample sea ice data, and sea fog visibility at multiple historical time points to obtain the Arctic sea fog model.

[0071] Among them, the sea fog visibility at each historical time point is used as the label during training, and the Arctic sea fog model is used to predict the sea fog visibility in the study area in the future.

[0072] In this embodiment, the random forest is used as the model to be trained. For this random forest, in the model training stage, the method of 10-fold cross-validation can be used to optimize the hyperparameters of the random forest. For example, the depth of the tree is set to 15, and an appropriate subset of features is selected. The training objective is to minimize the mean squared error (MSE), and it is required that the coefficient of determination R 2 of the test set is greater than 0.85 to ensure that the model has good fitting ability and prediction performance.

[0073] For the trained Arctic sea fog model, when predicting the future sea fog visibility, the CMIP6 climate variable data (time range from 2030 to 2060) based on two climate scenarios of SSP1-2.6 and SSP5-8.5 can be input into the trained random forest model; thereby evaluating the spatial distribution of sea fog visibility in the summer (from June to September) of the Arctic Northeast Passage from 2030 to 2060 in the future, and providing a probabilistic prediction result. Moreover, the study found that the formation of sea fog involves microphysical processes (1-10 km), and a 5-km resolution can capture micro processes such as the land-sea boundary effect and the thermal gradient in the sea ice edge area. Therefore, when evaluating the spatial distribution of sea fog visibility in the summer (from June to September) of the Arctic Northeast Passage from 2030 to 2060 in the future, the resolution was set to 5 km.

[0074] In addition, for the prediction results of the model, this embodiment is also divided into different risk levels, including high risk (visibility < 0.5 km), medium risk (0.5-1 km), and low risk (> 1 km). Moreover, warning information can also be released in real time through the Web platform to provide dynamic obstacle avoidance path planning for ships.

[0075] It should be understood that the above SSP1-2.6 and SSP5-8.5 refer to two different scenario settings in the Shared Socioeconomic Pathways (SSPs), which are used to describe different possibilities of future global climate change under different economic and greenhouse gas emission patterns. Among them, under the SSP1-2.6 scenario, it is assumed that the global socioeconomic development is relatively moderate, and active and effective greenhouse gas emission reduction measures are taken, resulting in a decrease in the concentration of carbon dioxide in the atmosphere in the future. Under the SSP5-8.5 scenario, it is assumed that global socioeconomic activities continue to grow rapidly, the use of fossil fuels increases significantly, resulting in a continued increase in greenhouse gas emissions, and thus more serious climate change.

[0076] In addition, this embodiment can also perform 1000 resampling operations on the random forest model through the Bootstrap resampling technique to generate a probability density function of sea fog visibility. This probability density function helps to understand the degree of uncertainty of the prediction results and can calculate the probability value at a specific confidence level. For example, at a 95% confidence level, the probability corresponding to a sea fog visibility less than 1 km is 70%, which can provide more comprehensive risk assessment information for decision-makers.

[0077] To make the above embodiments easier to understand, the following will further explain the above embodiments with a complete example:

[0078] As Figure 3As shown in the figure, the original meteorological data from NCEP FNL, the original sea surface temperature data from GHRSST, and the original sea ice data from NOAA / NSIDC are shown; the PWRF model is used to process the data at the same historical time point in these data to predict the cloud water mixing ratio, humidity, temperature, air pressure and other information at that historical time point; then, the cloud water mixing ratio, humidity, temperature, air pressure and other information are used to invert the sea fog visibility as the label in the meteorological sample information.

[0079] Then, the meteorological data, sea surface temperature data, and sea ice data generated by the CMIP6 model at that historical time point are weighted with the original meteorological data from NCEP FNL, the original sea surface temperature data from GHRSST, and the original sea ice data from NOAA / NSIDC, and the weighted fused data are used as the input data in the meteorological sample information.

[0080] Finally, the random forest model is trained using the meteorological sample information obtained above to obtain the Arctic sea fog model. The Arctic sea fog model is then used to process the CMIP6 synthesized meteorological data, surface temperature data, and sea ice data from 2030 to 2060 to obtain the sea fog conditions from 2030 to 2060.

[0081] Based on the same inventive concept as the training method of the Arctic sea fog model provided in this embodiment, this embodiment also provides a training device for the Arctic sea fog model, which includes at least one software function module that can be stored in a memory or fixed in an electronic device in the form of software. The processor in the electronic device is used to execute the executable module stored in the memory. For example, the software function module and computer program included in the device. Please refer to Figure 4 , functionally speaking, the device may include:

[0082] The sample preparation module 11 is used to obtain the sea fog visibility of the study area at multiple historical time points using a pre-built Polar Weather Research and Forecasting PWRF model;

[0083] The sample preparation module 11 is also used to weight the original meteorological data, original sea surface temperature data and original sea ice data of multiple meteorological models at each historical time point to obtain sample meteorological data, sample sea surface temperature data and sample sea ice data at the historical time point;

[0084] The model training module 12 is used to train the training model through meteorological data, sample sea surface temperature data, sample sea ice data and sea fog visibility at multiple historical time points to obtain an Arctic sea fog model, wherein the sea fog visibility at each historical time point is used as a label during training. The Arctic sea fog model is used to predict the future sea fog visibility in the study area.

[0085] In this embodiment, the sample preparation module 11 is used to implement Figure 1 steps S1 and S2 in Figure 1 , and the model training module 12 is used to implement

[0086] step S3 in

[0087] . Therefore, for the detailed description of each of the above modules, reference may be made to the specific implementation manners of the corresponding steps.

[0088] For each historical time point, obtain the cloud water mixing ratio, air temperature, and air pressure of the study area at the historical time point through the PWRF model;

[0089] According to the cloud water mixing ratio, air temperature, and air pressure at the historical time point, invert the sea fog visibility of the study area at the historical time point.

[0090] Optionally, the sample preparation module 11 is further specifically used for:

[0091] Obtain the original meteorological data, original sea surface temperature data, and original sea ice data of the study area at the historical time point;

[0092] Process the original meteorological data, original sea surface temperature data, and original sea ice data at the historical time point through the PWRF model to obtain the cloud water mixing ratio, air temperature, and air pressure at the historical time point.

[0093] Optionally, the sample preparation module 11 is further specifically used for;

[0094] Obtain the liquid water content at the historical time point according to the cloud water mixing ratio, air temperature, and air pressure at the historical time point;

[0095] Obtain the sea fog visibility at the historical time point according to the liquid water content at the historical time point.

[0096] Optionally, the sample preparation module 11 is further used for:

[0097] Assign weights to multiple meteorological models according to the accuracies of the multiple meteorological models, where the weight of each meteorological model is positively correlated with its own accuracy;

[0098] According to the weights assigned to the multiple meteorological models, weight the original meteorological data, original sea surface temperature data, and original sea ice data of the multiple meteorological models at the historical time point to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data of the study area at the historical time point.

[0099] In addition, in each embodiment of the present application, each functional module can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.

[0100] It should also be understood that if the above embodiments are implemented in the form of software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application.

[0101] Therefore, this embodiment also provides a storage medium, which is a computer-readable storage medium. This storage medium stores a computer program, and when the computer program is executed by a processor, it implements the training method of the Arctic sea fog model provided in this embodiment. Among them, the storage medium can be various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk, or an optical disc.

[0102] An electronic device for implementing the training of the Arctic sea fog model provided in this embodiment. As Figure 5 shown, the electronic device may include a processor 22 and a memory 21. Moreover, the memory 21 stores a computer program, and the processor reads and executes the computer program corresponding to the above embodiments in the memory 21 to implement the training method of the Arctic sea fog model provided in this embodiment.

[0103] Continue to refer to Figure 5 , the electronic device further includes a communication unit 23. Each element of the memory 21, the processor 22, and the communication unit 23 is directly or indirectly electrically connected to each other through a system bus 24 to achieve data transmission or interaction.

[0104] Among them, the memory 21 can be an information recording device based on any electronic, magnetic, optical, or other physical principles, and is used to record execution instructions, data, etc. In some embodiments, the memory 21 can be, but is not limited to, a volatile memory, a non-volatile memory, a storage drive, etc.

[0105] In some embodiments, the volatile memory may be a Random Access Memory (RAM); in some embodiments, the non-volatile memory may be a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electric Erasable Programmable Read-Only Memory (EEPROM), a flash memory, etc.; in some embodiments, the storage drive may be a disk drive, a solid state drive, any type of storage disk (such as an optical disk, a DVD, etc.), or a similar storage medium, or a combination thereof, etc.

[0106] The communication unit 23 is configured to transmit and receive data via a network. In some embodiments, the network may include a wired network, a wireless network, an optical fiber network, a telecommunication network, an intranet, the Internet, a Local Area Network (LAN), a Wide Area Network (WAN), a Wireless Local Area Networks (WLAN), a Metropolitan Area Network (MAN), a Wide Area Network (WAN), a Public Switched Telephone Network (PSTN), a Bluetooth network, a ZigBee network, or a Near Field Communication (NFC) network, etc., or any combination thereof. In some embodiments, the network may include one or more network access points. For example, the network may include a wired or wireless network access point, such as a base station and / or a network switching node, and one or more components of the service request processing system may be connected to the network via the access point to exchange data and / or information.

[0107] The processor 22 may be an integrated circuit chip with signal processing capabilities, and the processor may include one or more processing cores (e.g., a single-core processor or a multi-core processor). By way of example only, the above-mentioned processor may include a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), an Application Specific Instruction-set Processor (ASIP), a Graphics Processing Unit (GPU), a Physics Processing Unit (PPU), a Digital Signal Processor (DSP), a Field Programmable Gate Array (FPGA), a Programmable Logic Device (PLD), a controller, a microcontroller unit, a Reduced Instruction Set Computing (RISC), or a microprocessor, etc., or any combination thereof.

[0108] It can be understood that Figure 5 The structure shown is only illustrative. The electronic device 100 may also have more or fewer components than Figure 5 shown, or have a different configuration from Figure 5 that shown. Figure 5 Each of the components shown may be implemented in hardware, software, or a combination thereof.

[0109] It should be understood that the devices and methods disclosed in the above embodiments can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0110] As described above, these are only various embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A training method for an Arctic sea fog model, characterized in that The method includes: Obtaining the sea fog visibility of the study area at multiple historical time points by using a pre-constructed Polar Weather Research and Forecasting (PWRF) model; For each of the historical time points, weighting the original meteorological data, original sea surface temperature data, and original sea ice data of multiple meteorological models at the historical time point to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data at the historical time point; Training a model to be trained with the meteorological data, sample sea surface temperature data, sample sea ice data, and sea fog visibility at the multiple historical time points to obtain an Arctic sea fog model, where the sea fog visibility at each historical time point is used as a label during training, and the Arctic sea fog model is used to predict the future sea fog visibility of the study area.

2. The training method of the Arctic sea fog model according to claim 1, characterized in that Obtaining the sea fog visibility of the study area at multiple historical time points by using a pre-constructed Polar Weather Research and Forecasting (PWRF) model, including: For each of the historical time points, obtaining the cloud water mixing ratio, air temperature, and air pressure of the study area at the historical time point through the PWRF model; Inverting the sea fog visibility of the study area at the historical time point based on the cloud water mixing ratio, air temperature, and air pressure at the historical time point.

3. The training method of the Arctic sea fog model according to claim 2, wherein Obtaining the cloud water mixing ratio, air temperature, and air pressure of the study area at the historical time point through the PWRF model, including: Obtaining the original meteorological data, original sea surface temperature data, and original sea ice data of the study area at the historical time point; Processing the original meteorological data, original sea surface temperature data, and original sea ice data at the historical time point through the PWRF model to obtain the cloud water mixing ratio, air temperature, and air pressure at the historical time point.

4. The training method of the Arctic sea fog model according to claim 3, wherein The PWRF model includes a single-phase six-category cloud microphysics scheme (WSM6) model, a rapid radiative transfer model (RRTM), a convective parameterization scheme (Grell-Freitas) model, a land surface process model (Noah LSM) model, and a surface layer and planetary boundary layer scheme (MYNN3) model.

5. The training method of the Arctic sea fog model according to claim 2, characterized in that, Inverting the sea fog visibility of the study area at the historical time point based on the cloud water mixing ratio, air temperature, and air pressure at the historical time point, including: Obtaining the liquid water content at the historical time point based on the cloud water mixing ratio, air temperature, and air pressure at the historical time point; Obtaining the sea fog visibility at the historical time point based on the liquid water content at the historical time point.

6. The training method of the Arctic sea fog model according to claim 5, characterized in that The relationship between the cloud water mixing ratio, air temperature, and air pressure at the historical time point and the liquid water content at the historical time point is: The relationship between the liquid water content at the historical time point and the sea fog visibility at the historical time point is: wherein, LWC represents the liquid water content, VIS represents the sea fog visibility, Q cloud represents the cloud water mixing ratio, P represents the air pressure, and T represents the air temperature.

7. The training method of the Arctic sea fog model according to claim 2, characterized in that Weighting the original meteorological data, original sea surface temperature data, and original sea ice data of multiple meteorological models at the historical time point to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data at the historical time point, including: Allocating weights to the multiple meteorological models according to the accuracy of the multiple meteorological models, where the weight of each meteorological model is positively correlated with its own accuracy; Weights are assigned according to the multiple meteorological models, and the original meteorological data, original sea surface temperature data, and original sea ice data of the multiple meteorological models at the historical time point are weighted to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data of the study area at the historical time point.

8. A training device for an Arctic sea fog model, characterized in that, The device includes: A sample preparation module for obtaining the sea fog visibility of the study area at multiple historical time points by using a pre-constructed Polar Weather Research and Forecasting (PWRF) model; The sample preparation module is further configured to, for each of the historical time points, weight the original meteorological data, original sea surface temperature data, and original sea ice data of the multiple meteorological models at the historical time point to obtain the sample meteorological data, sample sea surface temperature data, and sample sea ice data at the historical time point; A model training module for training a model to be trained through the meteorological data, sample sea surface temperature data, sample sea ice data, and sea fog visibility at the multiple historical time points to obtain an Arctic sea fog model, wherein the sea fog visibility at each historical time point is used as a label during training, and the Arctic sea fog model is used to predict the future sea fog visibility of the study area.

9. A storage medium, characterized in that, The storage medium stores a computer program, and the computer program, when executed by a processor, implements the training method of the Arctic sea fog model according to any one of claims 1-7.

10. An electronic device, characterized in that, The electronic device includes a processor and a memory, the memory stores a computer program, and the computer program, when executed by the processor, implements the training method of the Arctic sea fog model according to any one of claims 1-7.

Images