Weather forecast method and device fusing multi-source data, equipment and storage medium

Abstract

The invention relates to the technical field of data processing, in particular to a weather forecast method and device fusing multi-source data, equipment and a storage medium. Secondly, calculating the prediction deviation level of each meteorological index according to the prediction results of the data sources and the actual observation data, then dynamically generating a data fusion weight based on the deviation level of each index in combination with the inherent characteristics of the data sources, and carrying out the weighted fusion of the prediction information of the plurality of data sources about each meteorological index. And outputting a fused weather prediction result. According to the method, the limitation influence of a single type of data source is effectively reduced through differential deviation calculation in combination with data source characteristics, dynamic weight distribution is performed in combination with historical performance of the data source, and the stability of a final prediction result and the data adaptability to meteorological conditions of a special region are improved; and the overall accuracy of weather forecast is obviously improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to weather forecasting methods, apparatus, equipment and storage media that integrate multi-source data. Background Technology

[0002] Weather forecasting plays a crucial role in various fields, including agricultural production, transportation, and disaster prevention. With the intensification of global climate change and the increasing frequency of extreme weather events, accurate weather forecasts not only ensure the orderly conduct of economic activities but also effectively mitigate the losses caused by natural disasters. However, the chaotic nature of atmospheric systems and the complex and variable climate patterns mean that scientific research and engineering applications in weather forecasting still face significant challenges.

[0003] In recent years, meteorological observation technology has developed rapidly, and various data sources, such as large-scale weather forecasting models, meteorological satellite remote sensing, ground observation stations, and radar systems, have provided abundant input information for meteorological forecasting. However, different data sources vary significantly in terms of spatiotemporal resolution, physical models, and equipment errors. Furthermore, different data sources exhibit varying levels of meteorological forecasting performance when dealing with different types of meteorological elements, leading to systematic biases in the forecast results of single-source data.

[0004] Therefore, how to scientifically integrate multi-source meteorological data to overcome the limitations of single-source meteorological data in forecasting and achieve high-precision, adaptive weather forecasting has become a technical problem that urgently needs to be solved in this industry. Summary of the Invention

[0005] The main objective of this invention is to provide a weather forecasting method, apparatus, device, and storage medium that integrates multi-source data, aiming to solve the technical problem of how to scientifically integrate multi-source meteorological data to overcome the limitations of single-source meteorological data prediction results and achieve high-precision, adaptive weather forecasting.

[0006] To achieve the above objectives, the present invention provides a weather forecasting method that integrates multi-source data, the method comprising the following steps: Obtain meteorological forecast information for the target location, wherein the meteorological forecast information includes forecast information from at least two different meteorological data sources; Based on the historical meteorological forecast results recorded by the meteorological data source and the meteorological observation data records of the target location, the prediction deviation level of the meteorological data source on each meteorological indicator is obtained; Based on the prediction deviation level of the meteorological data source on each meteorological indicator, calculate the data fusion weight of the meteorological data source on each meteorological indicator; Based on the data fusion weights of the meteorological data sources for each meteorological indicator, multi-source data fusion is performed on each meteorological indicator in the meteorological forecast information to obtain the fused weather forecast result.

[0007] Optionally, obtaining the prediction deviation level of the meteorological data source for each meteorological indicator based on the historical meteorological forecast results recorded by the meteorological data source and the meteorological observation data records of the target location includes: Based on the historical meteorological forecast results, extract the forecast results of each meteorological indicator for the target location from the meteorological data source within a preset time window; Based on the meteorological observation data records, extract the actual observation results that are aligned with the preset time window from the prediction results of each meteorological indicator; Based on the prediction results, the actual observation results, and the time span of the prediction behavior, the deviation is calculated to obtain the prediction deviation level of the meteorological data source on each meteorological indicator.

[0008] Optionally, the meteorological indicators include one or more of precipitation probability, precipitation amount, wind speed, wind direction, and temperature; The step of calculating the prediction deviation level of the meteorological data source on various meteorological indicators based on the prediction results, the actual observation results, and the time span of the prediction behavior includes: Based on the physical dimension type of the meteorological indicators, determine the deviation calculation rules; Based on the predicted behavior time span, a baseline coefficient is determined; The prediction deviation rate is determined based on the prediction results, the actual observation results, and the deviation calculation rules. The prediction deviation level of the meteorological data source on the meteorological index is obtained based on the prediction deviation rate and the benchmark coefficient.

[0009] Optionally, determining the deviation calculation rule based on the physical dimension type of the meteorological indicator includes: If the meteorological indicator belongs to the precipitation category, the first type of deviation calculation rule shall be used. The first type of deviation calculation rule includes the relative error calculation method and / or the logarithmic transformation error calculation method. If the meteorological indicator belongs to the temperature category, the second type of deviation calculation rule shall be used. The second type of deviation calculation rule includes the absolute error calculation method and / or the mean square error calculation method. If the meteorological indicator belongs to the wind field category, the third type of deviation calculation rule shall be used. The third type of deviation calculation rule includes the ring difference error calculation method and / or the vector error calculation method.

[0010] Optionally, calculating the data fusion weights of the meteorological data source on each meteorological indicator based on the prediction deviation level of the meteorological data source on each meteorological indicator includes: Based on the prediction deviation level of each meteorological indicator at the target location within a preset time window, calculate the average prediction deviation level of each meteorological indicator. Based on the predicted deviation level, the basic weight of the meteorological data source on the corresponding meteorological index is determined; Based on the prediction deviation level sequence within the preset time window, the trend of prediction deviation level changes for each meteorological indicator is obtained. Based on the rate of change of the trend of the prediction deviation level, the dynamic weight of the meteorological data source on the corresponding meteorological indicators is determined; Based on the basic weight and dynamic weight of the meteorological data source on the corresponding meteorological indicators, the data fusion weight of the meteorological data source on the corresponding meteorological indicators is obtained.

[0011] Optionally, the weather forecasting method that integrates multi-source data further includes: Obtain the geographic region classification label of the target location; Based on the geographical region classification labels, the target meteorological data source is determined; Based on the target meteorological data source, a data source queue is obtained by filtering multiple meteorological data sources to be selected. Based on the data source queue, the step of obtaining meteorological forecast information for the target location is performed.

[0012] Optionally, determining the target meteorological data source based on the geographic region classification label includes: If the geographical region is categorized as mountainous, then meteorological data sources that collect meteorological data using radar technology are selected as the target meteorological data source. If the geographical region is categorized as ocean, then meteorological data sources collected via satellite remote sensing technology are selected as the target meteorological data source. If the geographical region is categorized as a city, then meteorological data sources collected through a network of ground observation stations are selected as the target meteorological data source. If the geographical region is classified as a plain, then at least one type of meteorological data source from data collected through satellite remote sensing technology and ground observation station networks will be selected as the target meteorological data source.

[0013] Furthermore, to achieve the above objectives, the present invention also proposes a weather forecasting device that integrates multi-source data, the weather forecasting device integrating multi-source data comprising: The data acquisition module is used to acquire meteorological forecast information for the target location, wherein the meteorological forecast information includes forecast information from at least two different meteorological data sources; The deviation analysis module is used to obtain the prediction deviation level of the meteorological data source on each meteorological indicator based on the historical meteorological forecast results recorded by the meteorological data source and the meteorological observation data records of the target location. The weight allocation module is used to calculate the data fusion weight of the meteorological data source on each meteorological indicator based on the prediction deviation level of the meteorological data source on each meteorological indicator. The fusion prediction module is used to perform multi-source data fusion on each meteorological indicator in the meteorological prediction information based on the data fusion weight of the meteorological data source on each meteorological indicator, so as to obtain the fused weather prediction result.

[0014] Furthermore, to achieve the above objectives, the present invention also proposes a weather forecasting device that integrates multi-source data. The weather forecasting device that integrates multi-source data includes: a memory, a processor, and a weather forecasting program that integrates multi-source data stored in the memory and can run on the processor. The weather forecasting program that integrates multi-source data is configured to implement the steps of the weather forecasting method that integrates multi-source data as described above.

[0015] In addition, to achieve the above objectives, the present invention also proposes a storage medium storing a weather forecasting program that integrates multi-source data. When the weather forecasting program that integrates multi-source data is executed by a processor, it implements the steps of the weather forecasting method that integrates multi-source data as described above.

[0016] The present application proposes one or more technical solutions, which have at least the following technical effects: This solution obtains forecast information of the target location from at least two different meteorological data sources, then calculates the prediction deviation level of each meteorological indicator based on the past prediction results and actual observation data of these data sources. Subsequently, based on the deviation level of each indicator and the inherent characteristics of the data sources, the data fusion weights are dynamically allocated. For each meteorological indicator, higher weights are assigned to meteorological indicators with higher prediction accuracy from the data sources, while the weights of meteorological indicators with poor prediction performance from the data sources are suppressed. Finally, the optimal fusion prediction result is output based on a complementary weight strategy. This solution dynamically allocates weights based on the actual prediction levels of different data sources for each meteorological indicator, which can automatically strengthen the contribution of reliable data sources and suppress interference from abnormal data. In addition, the regional adaptation mechanism combined with geographical features can optimize the data fusion strategy under complex terrain, further reducing the forecast deviation in local areas, and enabling the system to have both global stability and local adaptability. This invention significantly improves the overall accuracy and robustness of weather forecasts through dynamic fusion and regional optimization, especially improving the local prediction capability in complex environments. At the same time, it provides a scalable technical framework for the intelligent fusion of multi-source meteorological data and has broad application value. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the first embodiment of the weather forecasting method that integrates multi-source data according to the present invention. Figure 2 This is a flowchart illustrating the second embodiment of the weather forecasting method that integrates multi-source data according to the present invention. Figure 3 This is a flowchart illustrating the third embodiment of the weather forecasting method that integrates multi-source data according to the present invention. Figure 4 This is a structural block diagram of the first embodiment of the weather forecasting device that integrates multi-source data according to the present invention; Figure 5 This is a schematic diagram of the structure of a weather forecasting device that integrates multi-source data in the hardware operating environment involved in the embodiments of the present invention.

[0020] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0022] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution of this application embodiment is as follows: Obtain meteorological forecast information for a target location, wherein the meteorological forecast information includes forecast information from at least two different meteorological data sources; based on historical meteorological forecast results recorded by the meteorological data sources and meteorological observation data records of the target location, obtain the prediction deviation level of the meteorological data sources on each meteorological indicator; calculate the data fusion weight of the meteorological data sources on each meteorological indicator based on the prediction deviation level of the meteorological data sources; and perform multi-source data fusion on each meteorological indicator in the meteorological forecast information based on the data fusion weight of the meteorological data sources on each meteorological indicator to obtain a fused weather forecast result.

[0024] Currently, meteorological observation technology is developing rapidly, and various data sources, such as large-scale weather forecasting models, meteorological satellite remote sensing, ground observation stations, and radar systems, provide abundant input information for weather forecasting. However, different data sources exhibit significant differences in spatiotemporal resolution, physical models, and equipment errors. Furthermore, different data sources demonstrate varying levels of weather forecasting performance when dealing with different types of meteorological elements, leading to systematic biases in the prediction results of single-source data. Therefore, how to scientifically integrate multi-source meteorological data to overcome the limitations of single-source meteorological data prediction results and achieve high-precision, adaptive weather forecasting is a pressing technical problem that needs to be solved.

[0025] This application obtains forecast information for a target location from at least two different meteorological data sources. Then, it calculates the forecast deviation level of each meteorological indicator based on the forecast results from these data sources and actual observation data. Subsequently, it dynamically generates data fusion weights based on the deviation levels of each indicator and the inherent characteristics of the data sources. For each meteorological indicator, data sources with higher forecast accuracy are assigned higher weights, while the weights of data sources with poor performance are suppressed. Finally, the optimal fusion forecast result is output based on a complementary weighting strategy. This scheme dynamically allocates weights based on the actual forecast levels of different data sources for each meteorological indicator, automatically strengthening the contribution of reliable data sources while suppressing interference from anomalous data. Furthermore, the regional adaptation mechanism combined with geographical features optimizes the data fusion strategy under complex terrain, further reducing forecast deviations in local areas, making the system both globally stable and locally adaptable. This invention significantly improves the overall accuracy and robustness of weather forecasts through dynamic fusion and regional optimization, especially improving local forecasting capabilities in complex environments. It also provides a scalable technical framework for the intelligent fusion of multi-source meteorological data and has broad application value.

[0026] It should be noted that the executing entity of this invention can be a weather forecasting device that integrates multi-source data, or a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or a thermal management device capable of implementing the above-mentioned functions of a weather forecasting device that integrates multi-source data. This embodiment does not specifically limit it in this way. The following uses a weather forecasting device that integrates multi-source data as the executing entity as an example to describe this embodiment and the following embodiments.

[0027] Based on this, embodiments of this application provide a weather forecasting method that fuses multi-source data, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the weather forecasting method that integrates multi-source data according to this application.

[0028] In this embodiment, the weather forecasting method that integrates multi-source data includes steps S10 to S40: Step S10: Obtain meteorological forecast information for the target location.

[0029] It should be noted that the target location in this embodiment is not limited to the user's real-time location. It can include any specific area where the user wants to receive weather forecasts. For example, outdoor event organizers need to monitor the weather conditions at the event venue, business people need to obtain the weather conditions of their destination for the next few days in advance, and freight companies need to track the weather conditions along the cargo transportation route. Therefore, this embodiment does not further limit the target location. At the same time, it supports users to add and manage multiple locations of interest, and each location independently executes the subsequent multi-source data fusion prediction process.

[0030] It is understandable that meteorological forecast information refers to meteorological forecast data released from various independent data sources. Currently, meteorological data sources can be divided into several categories based on different meteorological scientific methods, such as large-scale weather forecast models, meteorological satellite remote sensing, ground observation stations, and radar systems. Each type of data source has its own characteristics in terms of temporal resolution, spatial coverage, observation elements, and applicable scenarios.

[0031] It should be understood that large-scale numerical weather prediction models are supercomputing simulations based on mathematical equations. While these methods provide global weather forecasts, they suffer from systematic biases. Meteorological satellite remote sensing offers all-weather, wide-area monitoring, providing cloud images and temperature and humidity profiles, but its vertical accuracy is limited. Ground-based observation stations provide high-precision forecasts based on measured data, but spatial coverage is uneven, especially in oceanic and plateau regions where stations are scarce, resulting in lower data accuracy and reliability. Weather radar systems can track precipitation and severe convection in real time, but their detection range is limited and obstructed by terrain, thus only providing relatively accurate weather forecasts for areas close to the radar location. Sounding balloons, through a standard number of daily balloon probes, can provide accurate predictive data on vertical profiles within their coverage area, particularly upper-level temperature, humidity, and pressure data; however, the limited number of daily probes leads to delayed data updates. In summary, these data sources are complementary in terms of spatiotemporal resolution, accuracy, and coverage; their fusion optimizes forecast reliability. In this embodiment, to ensure the robustness and accuracy of weather forecasts, at least two different types of weather data sources must be selected when fusing multi-source weather forecast data.

[0032] Step S20: Based on the historical meteorological forecast results recorded by the meteorological data source and the meteorological observation data records of the target location, obtain the prediction deviation level of the meteorological data source on each meteorological indicator.

[0033] It should be noted that after obtaining preliminary meteorological forecasts for the target location from multiple meteorological data sources, it is necessary to analyze the reliability of each data source. Data reliability is measured by the discrepancy between the forecasts from each data source and the actual observation data recorded by meteorological observation departments over a past period. Specifically, once the target location is determined, the system sends meteorological forecast request commands to multiple different data sources. If a meteorological data source's coverage includes the target location, the data source will send its forecasts for the next few days and the past week (the specific timeframe is not limited) to the system. The system can then perform real-time evaluation of each data source based on historical forecasts and meteorological observation data for the target location.

[0034] Understandably, to ensure the reliability of the fusion results, the system first acquires historical forecast records from each meteorological data source and concurrent observation data of the target location, and then quantifies the forecast accuracy of the data sources through spatiotemporal alignment and bias calculation models. For key indicators such as temperature, precipitation, and wind speed, the system calculates statistics such as mean absolute error, root mean square error, and bias ratio to assess the systematic bias and random error of each data source.

[0035] In one feasible implementation, step S20 may include steps A10 to A30: Step A10: Based on the historical meteorological forecast results, extract the forecast results of each meteorological indicator of the target location from the meteorological data source within a preset time window.

[0036] Step A20: Based on the meteorological observation data records, extract the actual observation results that are aligned with the preset time window from the prediction results of each meteorological indicator.

[0037] Step A30: Calculate the deviation based on the prediction results, the actual observation results, and the time span of the prediction behavior to obtain the prediction deviation level of the meteorological data source on each meteorological indicator.

[0038] Understandably, a preset time window refers to a historical period used by the system to assess data source bias, such as the past 3 days or 7 days. Its range needs to be dynamically adjusted as needed. For short-term forecasts, such as precipitation, the window should be short (3-7 days) to capture the latest error trends as weather systems change rapidly. For long-term forecasts (such as monthly temperature), the window can be extended to 30 days to smooth out random fluctuations and reflect systematic biases. Secondly, the temporal resolution and release time of meteorological forecasts and observation data may be inconsistent. Since meteorological forecast data generally predicts weather conditions for a specific day or time, the time span from information release to the actual forecast timestamp is not fixed. Therefore, it needs to be aligned with the actual meteorological observation data to the same timestamp. The forecast behavior time span refers to the time difference from the forecast release time to the target forecast time, such as different time spans for meteorological forecasts like "24-hour advance forecast" and "6-hour advance forecast."

[0039] It should be noted that different data sources have inherent advantages and disadvantages in predicting different meteorological indicators due to their inherent characteristics. For example, the advantage of large-scale weather forecast models lies in predicting large-scale precipitation events, providing a high level of precipitation trend prediction for the next 3 to 7 days. However, they tend to underestimate the precipitation probability of localized severe convection (such as thunderstorms), and are prone to missing sudden rainfall in short-term forecasts. Meteorological satellites retrieve cloud top temperature through infrared / microwave inversion, so their advantage lies in identifying potential precipitation areas. However, they have difficulty distinguishing precipitation types (rain / snow / hail), and are prone to misreporting precipitation types. Weather radar makes meteorological predictions based on echo intensity, so its short-term precipitation probability calculation is relatively accurate and suitable for severe convection warnings. However, due to attenuation, the precipitation probability in areas far from the base station will gradually become distorted, and it cannot predict precipitation outside the radar coverage area.

[0040] It should be understood that commonly used meteorological indicators include, but are not limited to, precipitation probability, precipitation amount, wind speed, wind direction, and temperature. However, not all data sources can predict these meteorological indicators. Some data sources can only provide predictions for a few of them. In addition, the prediction accuracy of various data sources for different meteorological indicators varies. Therefore, in the evaluation process, it is necessary to combine historical data to determine the deviation level of each data source for the published multiple meteorological indicators. Thus, it is possible that data source A has a high prediction level for precipitation-related meteorological indicators but an inaccurate temperature prediction, while data source B has an inaccurate prediction level for precipitation but a high prediction level for indicators such as wind speed and wind direction. Therefore, in the subsequent weight allocation process, the weights assigned to different meteorological indicators of the same data source are not the same.

[0041] Step S30: Calculate the data fusion weight of the meteorological data source on each meteorological indicator based on the prediction deviation level of the meteorological data source on each meteorological indicator.

[0042] It should be noted that the calculation of data fusion weights converts the prediction deviation levels of different meteorological data sources into quantifiable and comparable values, thereby providing an objective basis for subsequent multi-source information fusion. The basic principle is that data sources with lower prediction deviation levels should be assigned higher fusion weights to fully reflect the dominant role of high-quality data in the final forecast results. At the same time, different weight calculation methods may be required for different types of meteorological indicators. For example, linear inverse weight models can be used for continuously changing physical quantities such as temperature and precipitation, while classification probability weight mechanisms are more suitable for discrete indicators such as visibility or weather phenomena, ensuring that the fusion process of various meteorological elements conforms to their inherent physical characteristics and operational application needs.

[0043] Understandably, the accuracy of weight calculation directly impacts the quality of the final fusion forecast. In practice, it's necessary to combine the strengths of each data source to maximize prediction performance. This means that even with the same data source, different meteorological indicators will have different weights. Weight allocation requires ranking the prediction results from all data sources based on the deviation level of the same meteorological indicator, and then determining the weight allocated to each data source for that indicator based on the ranking. Data sources with excessively high deviation levels can be completely ignored. Furthermore, considering that data source performance may vary significantly under different weather conditions—for example, some data sources may perform well under stable weather conditions but have larger errors under severe convective weather—the weight calculation module should also be dynamically adaptable, capable of scenario-based adjustments to the basic weights according to real-time weather conditions, thus outputting a reasonable weight allocation scheme under various meteorological conditions.

[0044] It should be understood that the weighting calculation process needs to maintain the relative independence of each meteorological indicator while also considering the synergy of business applications. For example, when producing integrated weather forecast products, the weights of each element can be calculated separately, but physical consistency must be maintained to avoid abnormal combinations such as extremely high weights for temperature and extremely low weights for precipitation. At the same time, in order to ensure the stable operation of the business system, reasonable upper and lower limits need to be set for the calculated weights to prevent a single data source from obtaining excessively high or low weights due to accidental errors. If necessary, a manual intervention mechanism can be introduced to check the rationality of the automatically calculated weights and make fine adjustments to ensure that the weight distribution that finally enters the fusion stage can objectively reflect the performance of the data source and meet the actual needs of forecasting business.

[0045] Step S40: Based on the data fusion weights of the meteorological data source on each meteorological indicator, perform multi-source data fusion on each meteorological indicator in the meteorological forecast information to obtain the fused weather forecast result.

[0046] It should be noted that multi-source data fusion is a key step in organically integrating forecast information from different meteorological data sources based on weight calculation results. Its core objective is to leverage the advantages of each data source to compensate for the shortcomings of a single source, thereby generating more accurate and reliable comprehensive forecast products. The fusion algorithm needs to select an appropriate mathematical model based on the characteristics of meteorological elements. For example, a weighted average method can be used for continuous variables such as temperature, while Bayesian fusion technology based on probability density functions may be needed for special indicators such as precipitation probability. This ensures that meteorological data with different attributes can be scientifically and rationally integrated and processed. At the same time, the fusion process must also consider the influence of regional characteristics. For example, a spatial variable weight strategy may be needed for complex terrain areas to improve the precision of local weather forecasts.

[0047] It should be understood that after fusion, the resulting product must undergo comprehensive testing and evaluation, including comparative analysis with independent observation data and expert judgment by forecasters. Only through quality control throughout the entire process can we ensure that the final fused forecast not only possesses the scientific validity of objective algorithms but also conforms to the rationality of meteorological principles, providing a reliable data foundation for various meteorological services.

[0048] This embodiment obtains the prediction information of the target location from at least two different meteorological data sources, then calculates the prediction deviation level of each meteorological indicator based on the prediction results of these data sources and the actual observation data, and then dynamically generates data fusion weights based on the deviation level of each indicator and the inherent characteristics of the data sources. By weighted fusion of the prediction information of multiple data sources for each meteorological indicator, the fused weather prediction result is output.

[0049] In summary, this technical solution dynamically allocates weights based on the actual forecast levels of various meteorological indicators from different data sources. This automatically strengthens the contribution of reliable data sources while suppressing interference from anomalous data. Furthermore, the regional adaptation mechanism combined with geographical features optimizes data fusion strategies under complex terrain, further reducing forecast bias in local areas and enabling the system to possess both global stability and local adaptability.

[0050] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 In the weather forecasting method that integrates multi-source data, step A30 includes steps B10 to B40: Step B10: Determine the deviation calculation rules based on the physical dimension type of the meteorological indicators.

[0051] It should be noted that in the field of meteorological forecasting, indicators of different physical dimensions have drastically different data distribution characteristics, which directly determines that different types of meteorological indicators require appropriate statistical deviation calculation methods.

[0052] It is understandable that, taking precipitation indicators in the hydrological dimension as an example, meteorological indicators under this dimension include precipitation amount, precipitation probability, humidity, etc. These indicators have the following statistical characteristics: zero expansion problem: that is, precipitation data often have a large number of zero values ​​(when there is no precipitation) and right-skewed distribution (when there is precipitation), and there is significant heteroscedasticity. That is, the greater the precipitation intensity, the observation error usually grows exponentially. Therefore, when dealing with precipitation indicators in the hydrological dimension, it is more suitable to use relative error calculation method and / or logarithmic transformation error calculation method. The reason is that the zero expansion characteristic of precipitation data requires that the deviation level calculation must be compatible with zero values, while the right-skewed distribution and heteroscedasticity determine that scale-invariant relative error or nonlinear compression logarithmic error method should be used to reasonably characterize the deviation level.

[0053] It is understandable that, taking temperature-related indicators with thermodynamic dimensions as an example, meteorological indicators under this dimension include air temperature, surface temperature, dew point temperature, etc. These indicators have the following statistical characteristics: approximately normal distribution, because temperature data usually conforms to a Gaussian distribution; homoscedasticity, that is, the range of values ​​that produce deviations is independent of the magnitude of the temperature value. Therefore, when dealing with temperature-related indicators with thermodynamic dimensions, it is more suitable to use the absolute error calculation method and / or the mean square error calculation method. The reason is that the approximately normal distribution characteristic of temperature data allows the absolute error to directly reflect the physical magnitude of the predicted deviation from the true value, and homoscedasticity ensures that the error is comparable in different temperature ranges. The Gaussian distribution characteristic of the mean square error makes large errors have a higher weight, which conforms to the statistical law of temperature data, and is especially suitable for the stringent accuracy requirements of extreme temperature warnings (such as high temperature heat waves and cold waves).

[0054] Understandably, taking wind field indicators with dynamic dimensions as an example, meteorological indicators under this dimension include wind speed, wind direction, and pressure gradient. These indicators have the following statistical characteristics: non-negative right-skewed distribution, meaning that wind speed usually follows a log-normal distribution, exhibiting a clear right-skewed characteristic, i.e., low wind speeds are highly probable, but there are a few extreme wind events (such as typhoons and squall lines); directional periodicity (wind direction), meaning that wind direction data is distributed in a circular pattern (0°~360°), which makes traditional linear statistical methods unsuitable and requires circular statistics, such as average wind direction angle and circular variance. In addition, wind field indicators also have vector characteristics. In summary, the strong nonlinearity, directional periodicity, and vector characteristics of wind field indicators require a hybrid error measurement strategy. Therefore, it is suitable to use the circular difference error calculation method and / or the vector error calculation method to calculate the prediction deviation level. This is because the circular difference error and vector error calculation methods can effectively handle the periodicity of wind direction and the vector characteristics of wind field, respectively, ensuring that the error measurement conforms to the statistical characteristics of dynamic indicators.

[0055] It should be understood that the differences in the physical dimensions of meteorological indicators directly determine their data distribution characteristics and applicable error calculation methods. Therefore, when assessing forecast bias, it is necessary to strictly match statistical characteristics and calculation rules to ensure the scientific and practical nature of error representation.

[0056] In one feasible implementation, determining the deviation calculation rule based on the physical dimension type of the meteorological indicator includes: if the meteorological indicator belongs to the precipitation category, a first type of deviation calculation rule is used, which includes a relative error calculation method and / or a logarithmic transformation error calculation method; if the meteorological indicator belongs to the temperature category, a second type of deviation calculation rule is used, which includes an absolute error calculation method and / or a mean square error calculation method; if the meteorological indicator belongs to the wind field category, a third type of deviation calculation rule is used, which includes a ring difference error calculation method and / or a vector error calculation method.

[0057] Step B20: Determine the baseline coefficient based on the predicted behavior time span.

[0058] It should be noted that different data sources have different prediction time spans for the same meteorological indicator (such as precipitation and wind speed). For example, some data sources predict the meteorological results 6 hours later, while others predict the meteorological results 24 hours later. Due to the different time spans, the prediction error will accumulate over time, and a benchmark coefficient needs to be introduced to correct the deviation calculation.

[0059] Understandably, the core function of the benchmark coefficient is to correct the difference in the level of deviation of different prediction time spans through time-dependent weights (such as linear, exponential or piecewise functions), so that the deviations of short-term and long-term predictions are comparable, and to optimize the weight allocation when fusing multi-source data. Its parameters need to be based on physical laws (how the error accumulates over time) and data statistics (historical error growth curves). The ultimate goal is to achieve fair evaluation and fusion across time and data sources.

[0060] Step B30: Determine the prediction deviation rate based on the prediction results, the actual observation results, and the deviation calculation rules.

[0061] It should be noted that the prediction deviation rate is a key indicator for evaluating the accuracy of meteorological data source predictions. By comparing the difference between the predicted value and the actual observed value and normalizing it with a specific benchmark value, the relative magnitude of the prediction error can be quantified. The selection of the benchmark value directly affects the rationality of the result. It is usually determined based on the climate mean, historical extreme thresholds, or industry standards to ensure that the calculation of different meteorological indicators is comparable. For example, in temperature prediction, the long-term climate average is used as the denominator, while in precipitation forecasting, a certain percentile threshold may be used to avoid extreme precipitation events causing the denominator to be too small and distorting the deviation rate.

[0062] Understandably, the calculation of prediction deviation rate needs to take into account the real-time characteristics of meteorological events. For example, the calculation method for the deviation rate of short-term heavy precipitation or typhoon track prediction may be different from that of conventional weather systems. In actual operation, it is usually necessary to configure appropriate deviation rate calculation methods for different types of meteorological processes to ensure the relevance of the evaluation results. At the same time, in order to avoid numerical abrupt changes in extreme cases, reasonable upper and lower limit constraints need to be set. For example, when the observed value is close to zero or abnormally high, a correction coefficient needs to be introduced to ensure that the deviation rate can objectively reflect the prediction performance, rather than being dominated by a few extreme cases in the overall evaluation results.

[0063] It should be understood that the prediction bias rate is not only used for performance evaluation of a single data source, but also an important basis for subsequent multi-source data fusion and weight allocation. Therefore, its statistical stability needs to be fully considered. The common practice is to perform a moving average or percentile statistics on multiple prediction results within a certain time window to reduce the impact of random fluctuations. At the same time, the physical rationality of the bias rate should be checked by combining meteorological principles. For example, if a certain data source continues to show a systematically high bias rate in a certain area, it is necessary to further analyze the reasons behind it. It may be that the initial field error of the model or the terrain of the local area has not been fully considered, thereby guiding the subsequent optimization of dynamic weight allocation or improvement of the bias level algorithm.

[0064] Step B40: Based on the prediction deviation rate and the benchmark coefficient, obtain the prediction deviation level of the meteorological data source on the meteorological index.

[0065] It should be noted that the forecast deviation level is a further correction to the forecast deviation rate. Its core lies in introducing a time base coefficient to reflect the impact of different forecast lead times on error accumulation. Since the uncertainty of meteorological forecasts usually increases with the increase of forecast lead time, the error of short-term forecasts changes less, while the error of medium- and long-term forecasts may show non-linear growth. Therefore, it is necessary to use time-dependent weighting coefficients for correction. For example, for the same meteorological indicator, the forecast error of 24 hours may need to bear a lower penalty weight than the same numerical error of a 6-hour forecast, so that forecasts with different lead times can be reasonably compared and integrated under the same standard.

[0066] It should be understood that the ultimate purpose of the prediction bias level is to serve the fusion decision of multi-source data. Therefore, in operational applications, it is necessary to set clear classification standards. For example, the prediction performance of the data source can be divided into high, medium and low confidence levels according to the numerical range of the bias level, and its weight in the fusion algorithm can be allocated accordingly. At the same time, in order to adapt to the dynamic changes of the data source, such as numerical model upgrades or observation system improvements, it is also necessary to re-evaluate the prediction bias level regularly to ensure that it can always reflect the latest data quality status, and to continuously optimize the performance and stability of the fusion forecast accordingly.

[0067] This embodiment proposes a differentiated deviation calculation method based on the physical characteristics of meteorological indicators. For different statistical characteristics of precipitation, temperature and wind field indicators, corresponding statistical processing strategies are designed respectively. Combined with the prediction time span factor, a benchmark coefficient is set to realize the comparability evaluation of prediction results at different time scales, and finally a series of evaluation indicators for prediction deviation rate and prediction deviation level are formed.

[0068] In summary, this embodiment, by adopting error calculation methods matched to indicators with different physical characteristics, can more accurately reflect the actual forecasting capabilities of various meteorological data sources. Simultaneously, the introduction of a time span correction coefficient effectively solves the comparability issue between short-term and long-term forecasts, ensuring the fairness of weight allocation during multi-source data fusion. Overall, this scheme significantly improves the targeting of meteorological forecast assessment, providing a more scientific and reliable basis for deviation quantification in subsequent data fusion, thereby effectively improving the accuracy and stability of the final fused forecast product.

[0069] Based on the first embodiment of this application, in the third embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 3 In the weather forecasting method that integrates multi-source data, before step S10, steps C10 to C40 are further included: Step C10: Obtain the geographic region classification label of the target location.

[0070] It should be noted that geographic region classification labels are a standardized identification of the environmental type of a target location and serve as the basis for subsequent meteorological data sources. The classification system is usually constructed based on comprehensive factors such as topographic features, climate zoning, and intensity of human activities. Common classifications include typical geographic types such as mountains, oceans, cities, and plains. These labels can be obtained through automatic identification by geographic information systems or manual annotation. To ensure classification accuracy, cross-validation is usually performed using multi-source geographic information such as online electronic maps and land use data. At the same time, considering the spatial heterogeneity of meteorological elements, a hybrid classification mechanism is introduced for boundary transition areas to more accurately reflect the actual geographic characteristics.

[0071] Understandably, the accuracy of geographic regional classification should not be too coarse, which would fail to reflect key differences, nor should it be too fine, which would increase unnecessary computational complexity. In practice, the appropriate classification level is usually determined based on the spatial resolution of various meteorological data sources. For example, a more refined classification system is needed for fine forecasts with a kilometer-level resolution, while a relatively simplified classification method can be used for forecasts of large areas.

[0072] It should be understood that the application of geographic region classification labels is not limited to the data source screening stage, but can also be extended to multiple stages such as subsequent forecast fusion and error correction. Different forecast bias assessment standards may be adopted for different geographic types. For example, urban areas have poor ventilation due to dense buildings, and their heat island effect and pollutant diffusion patterns have obvious special characteristics. Therefore, the allowable error range should be appropriately relaxed when assessing forecasts of indicators such as temperature and visibility. In mountainous areas, precipitation forecasts are often affected by orographic uplift, and precipitation forecasts often need to be superimposed with topographic correction coefficients on the numerical model. Their assessment standards should also be treated differently from those for traditional plain areas.

[0073] Step C20: Determine the target meteorological data source based on the geographical region classification labels.

[0074] It should be noted that the target meteorological data source, which is the preferred meteorological data source among various types of data sources, is an intelligent selection based on the applicability differences of various observation technologies in different geographical environments. This selection fully considers the advantages and limitations of different data acquisition methods in different geographical environments. For example, in mountainous areas, radar technology can effectively capture the cloud and rain formation process caused by topographic uplift, overcoming the problem of insufficient spatial coverage caused by terrain limitations of ground observation stations. For vast ocean areas, satellite remote sensing becomes the most important data source due to its large-scale synchronous observation capability. This location-specific data source selection strategy can significantly improve the quality and representativeness of the initial data.

[0075] Understandably, the formulation of optimal selection criteria needs to be based on extensive statistical analysis and objective technical evaluation, and the actual performance of various data sources in different geographical environments needs to be determined through retrospective testing of historical data. Furthermore, practical operational factors such as the spatiotemporal resolution and data latency of the data acquisition system must be considered to ensure that the optimal selection results are not only theoretically sound but also engineering feasible.

[0076] It should be understood that the process of selecting target meteorological data sources is not a fixed rule. With the development of new observation technologies and the improvement of data assimilation methods, the traditional selection criteria may need to be adjusted in a timely manner. For example, with the promotion of Internet of Things technology in meteorological observation, the target meteorological data sources in urban areas may gradually expand from traditional ground observation stations to hybrid data sources that include numerous micro-sensor networks. This requires the selection system to have sufficient scalability and adaptability to absorb new technological achievements and service needs in a timely manner.

[0077] In one feasible implementation, determining the target meteorological data source based on the geographic region classification label includes: if the geographic region classification label is mountainous, then selecting meteorological data sources that collect meteorological data using radar technology as the target meteorological data source; if the geographic region classification label is ocean, then selecting meteorological data sources that collect data using satellite remote sensing technology as the target meteorological data source; if the geographic region classification label is urban, then selecting meteorological data sources that collect data using a ground observation station network as the target meteorological data source; if the geographic region classification label is plain, then selecting at least one type of meteorological data source from satellite remote sensing technology and ground observation station network as the target meteorological data source.

[0078] Step C30: Based on the target meteorological data source, filter the multiple meteorological data sources to be selected to obtain a data source queue.

[0079] It should be noted that the number of meteorological data sources covered by different meteorological data sources varies for different geographical locations. Some areas have a relatively dense collection of meteorological data sources. For example, urban areas may simultaneously cover multiple data sources such as ground observation stations, radar, satellites, and numerical models. However, remote or special terrain areas such as plateaus, oceans, or deserts have relatively limited available data sources. Therefore, when determining the target meteorological data source, it is also necessary to adaptively filter data based on the data abundance of the target location to avoid insufficient data abundance during subsequent data fusion, which could lead to the fusion results being easily affected by the distortion of a single data source.

[0080] Understandably, the selection criteria for the data source queue should be transparent and traceable. A clear priority scoring mechanism should be established to objectively quantify the applicability of each candidate data source. For example, a comprehensive evaluation system should be constructed by considering factors such as data quality indicators, spatiotemporal resolution, and historical performance to avoid the systemic risks caused by over-reliance on a single data source. A certain number of alternative data sources are usually reserved in the queue to form a flexible structure that combines primary and backup. When the primary data source is abnormal, it can be switched to an alternative solution in a timely manner to ensure the stable operation of the business.

[0081] Step C40: Based on the data source queue, perform the step of obtaining meteorological forecast information for the target location.

[0082] This embodiment establishes a standardized geographic region classification and labeling system, and dynamically selects the most suitable data source combination for the meteorological monitoring needs of the region by combining the observation technology characteristics under different geographic environments. Specifically, this includes: first, accurately identifying the geographic type based on factors such as terrain features and climate characteristics of the target location; then, matching different target meteorological data source configuration schemes for different classification labels such as mountains, oceans, and cities; and finally, establishing a multi-source complementary data source queue based on actual data coverage to provide a reliable data foundation for the subsequent generation of high-quality meteorological forecast information. This scheme particularly emphasizes the geographic adaptability and technical rationality of data source selection, ensuring that the selected data can truly reflect the meteorological characteristics of the region.

[0083] In summary, this embodiment significantly improves the targeting and scientific rigor of data selection by establishing an intelligent geographic region matching mechanism, maximizing the advantages of various observation technologies. Secondly, the differentiated data source selection strategy based on geographic characteristics effectively addresses the problem of insufficient monitoring data in complex terrain areas, helping to improve forecast quality in traditionally difficult-to-monitor areas such as mountains and oceans. Finally, the dynamically constructed data source queue structure ensures the reliability of core data sources while enhancing the system's fault tolerance through alternative data sources, providing a solid guarantee for stable and continuous meteorological services. Overall, this scheme, through a geographically driven data source selection mechanism, ensures the high quality and representativeness of meteorological forecast data from the source, laying a solid foundation for subsequent accurate forecasts.

[0084] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the weather forecasting method of fusing multi-source data in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0085] This application also provides a weather forecasting device that integrates multi-source data; please refer to... Figure 4 The weather forecasting device that integrates multi-source data includes: The data acquisition module 10 is used to acquire meteorological forecast information for the target location, wherein the meteorological forecast information includes forecast information from at least two different meteorological data sources.

[0086] The deviation analysis module 20 is used to obtain the prediction deviation level of the meteorological data source on each meteorological indicator based on the historical meteorological forecast results recorded by the meteorological data source and the meteorological observation data records of the target location.

[0087] The weight allocation module 30 is used to calculate the data fusion weight of the meteorological data source on each meteorological indicator based on the prediction deviation level of the meteorological data source on each meteorological indicator.

[0088] The fusion prediction module 40 is used to perform multi-source data fusion on each meteorological indicator in the meteorological prediction information based on the data fusion weight of the meteorological data source on each meteorological indicator, so as to obtain the fused weather prediction result.

[0089] In one embodiment, the deviation analysis module 20 is further configured to: extract the prediction results of the meteorological data source for each meteorological indicator of the target location within a preset time window based on the historical meteorological forecast results record; extract the actual observation results that are aligned with the preset time window time from the prediction results of each meteorological indicator based on the meteorological observation data record; and calculate the deviation based on the prediction results, the actual observation results, and the time span of the prediction behavior to obtain the prediction deviation level of the meteorological data source on each meteorological indicator.

[0090] In one embodiment, the deviation analysis module 20 is further configured to determine the deviation calculation rule based on the physical dimension type of the meteorological indicator; determine the benchmark coefficient based on the time span of the prediction behavior; determine the prediction deviation rate based on the prediction result, the actual observation result, and the deviation calculation rule; and obtain the prediction deviation level of the meteorological data source on the meteorological indicator based on the prediction deviation rate and the benchmark coefficient.

[0091] In one embodiment, the deviation analysis module 20 is further configured to: calculate the average predicted deviation level of each meteorological indicator based on the predicted deviation level of each meteorological indicator at the target location within a preset time window; determine the basic weight of the meteorological data source on the corresponding meteorological indicator based on the average predicted deviation level; obtain the trend of the predicted deviation level of each meteorological indicator based on the predicted deviation level sequence within the preset time window; determine the dynamic weight of the meteorological data source on the corresponding meteorological indicator based on the rate of change of the predicted deviation level trend; and obtain the data fusion weight of the meteorological data source on the corresponding meteorological indicator based on the basic weight and dynamic weight of the meteorological data source.

[0092] In one embodiment, the weight allocation module 30 is further configured to obtain the geographic region classification label of the target location; determine the target meteorological data source based on the geographic region classification label; filter multiple meteorological data sources to be selected based on the target meteorological data source to obtain a data source queue; and execute the step of obtaining meteorological forecast information of the target location based on the data source queue.

[0093] In one embodiment, the fusion prediction module 40 is further configured to: if the geographical region classification label is mountainous, then select a meteorological data source that collects meteorological data using radar technology as the target meteorological data source; if the geographical region classification label is ocean, then select a meteorological data source that collects data using satellite remote sensing technology as the target meteorological data source; if the geographical region classification label is urban, then select a meteorological data source that collects data using a ground observation station network as the target meteorological data source; if the geographical region classification label is plain, then select at least one type of meteorological data source that collects data using satellite remote sensing technology and a ground observation station network as the target meteorological data source.

[0094] This embodiment acquires forecast information for a target location from at least two different meteorological data sources. Then, based on past forecasts and actual observation data from these data sources, it calculates the forecast deviation levels for each meteorological indicator. Subsequently, based on the deviation levels of each indicator and the inherent characteristics of the data sources, it dynamically allocates data fusion weights. For each meteorological indicator, higher weights are assigned to indicators with higher forecast accuracy from the data sources, while the weights of indicators with poor forecast performance are suppressed. Finally, based on a complementary weighting strategy, the optimal fusion forecast result is output. This scheme dynamically allocates weights based on the actual forecast levels of different data sources for each meteorological indicator, automatically strengthening the contribution of reliable data sources while suppressing interference from anomalous data. Furthermore, the regional adaptation mechanism combined with geographical features optimizes the data fusion strategy under complex terrain, further reducing forecast deviations in local areas, enabling the system to possess both global stability and local adaptability. This invention, through dynamic fusion and regional optimization, significantly improves the overall accuracy and robustness of weather forecasts, especially enhancing local forecasting capabilities in complex environments. It also provides a scalable technical framework for the intelligent fusion of multi-source meteorological data, possessing broad application value.

[0095] The weather forecasting device fusion of multi-source data provided in this application, employing the weather forecasting method fusion of multi-source data in the above embodiments, can solve the technical problem of how to scientifically fuse multi-source meteorological data to overcome the limitations of single-source meteorological data prediction results and achieve high-precision, adaptive weather forecasting. Compared with the prior art, the beneficial effects of the weather forecasting device fusion of multi-source data provided in this application are the same as those of the weather forecasting method fusion of multi-source data provided in the above embodiments, and other technical features in the weather forecasting device fusion of multi-source data are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0096] This application provides a weather forecasting device that integrates multi-source data. The weather forecasting device that integrates multi-source data includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the weather forecasting method that integrates multi-source data in the above embodiment 1.

[0097] The following is for reference. Figure 5 This document illustrates a structural schematic diagram of a weather forecasting device suitable for implementing embodiments of this application, which integrates multi-source data. The weather forecasting device integrating multi-source data in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 5 The weather forecasting device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0098] like Figure 5As shown, a weather forecasting device that integrates multi-source data may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1002 or a program loaded from storage device 1003 into random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the weather forecasting device that integrates multi-source data. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the weather forecasting device fusing multi-source data to wirelessly or wiredly communicate with other devices to exchange data. Although the figure shows a weather forecasting device fusing multi-source data with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0099] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0100] The weather forecasting device that integrates multi-source data provided in this application, employing the weather forecasting method that integrates multi-source data in the above embodiments, can solve the technical problem of how to scientifically integrate multi-source meteorological data to overcome the limitations of single-source meteorological data prediction results and achieve high-precision, adaptive weather forecasting. Compared with the prior art, the beneficial effects of the weather forecasting device that integrates multi-source data provided in this application are the same as those of the weather forecasting method that integrates multi-source data provided in the above embodiments, and other technical features in this weather forecasting device that integrates multi-source data are the same as those disclosed in the method of the previous embodiment, and will not be repeated here.

[0101] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0102] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0103] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the weather forecasting method for fusing multi-source data in the above embodiments.

[0104] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0105] The aforementioned computer-readable storage medium may be included in a weather forecasting device that integrates multi-source data; or it may exist independently and not be assembled into a weather forecasting device that integrates multi-source data.

[0106] The aforementioned computer-readable storage medium carries one or more programs that, when executed by a weather forecasting device that integrates multi-source data, cause the weather forecasting device to: acquire meteorological forecast information for a target location, the meteorological forecast information including forecast information from at least two different meteorological data sources; obtain the prediction deviation level of the meteorological data source on each meteorological indicator based on historical meteorological forecast results records of the meteorological data source and meteorological observation data records of the target location; calculate the data fusion weight of the meteorological data source on each meteorological indicator based on the prediction deviation level of the meteorological data source; and perform multi-source data fusion on each meteorological indicator in the meteorological forecast information based on the data fusion weight of the meteorological data source on each meteorological indicator to obtain a fused weather forecast result.

[0107] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0108] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0109] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0110] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described weather forecasting method that fuses multi-source data. This solves the technical problem of how to scientifically fuse multi-source meteorological data to overcome the limitations of single-source meteorological data prediction results and achieve high-precision, adaptive weather forecasting. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the weather forecasting method that fuses multi-source data provided in the above embodiments, and will not be repeated here.

[0111] The computer program product provided in this application can solve the technical problem of weather forecasting by fusing multi-source data. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the weather forecasting method for fusing multi-source data provided in the above embodiments, and will not be repeated here.

[0112] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A weather forecasting method that integrates multi-source data, characterized in that, The weather forecasting method that integrates multi-source data includes: Obtain meteorological forecast information for the target location, wherein the meteorological forecast information includes forecast information from at least two different meteorological data sources; Based on the historical meteorological forecast results recorded by the meteorological data source and the meteorological observation data records of the target location, the prediction deviation level of the meteorological data source on each meteorological indicator is obtained; Based on the prediction deviation level of the meteorological data source on each meteorological indicator, calculate the data fusion weight of the meteorological data source on each meteorological indicator; Based on the data fusion weights of the meteorological data sources for each meteorological indicator, multi-source data fusion is performed on each meteorological indicator in the meteorological forecast information to obtain the fused weather forecast result.

2. The weather forecasting method based on multi-source data according to claim 1, characterized in that, The step of obtaining the prediction deviation level of the meteorological data source for each meteorological indicator based on the historical meteorological forecast results recorded by the meteorological data source and the meteorological observation data records of the target location includes: Based on the historical meteorological forecast results, extract the forecast results of each meteorological indicator for the target location from the meteorological data source within a preset time window; Based on the meteorological observation data records, extract the actual observation results that are aligned with the preset time window from the prediction results of each meteorological indicator; Based on the prediction results, the actual observation results, and the time span of the prediction behavior, the deviation is calculated to obtain the prediction deviation level of the meteorological data source on each meteorological indicator.

3. The weather forecasting method based on multi-source data according to claim 2, characterized in that, The meteorological indicators include one or more of the following: precipitation probability, precipitation amount, wind speed, wind direction, and temperature; The step of calculating the prediction deviation level of the meteorological data source on various meteorological indicators based on the prediction results, the actual observation results, and the time span of the prediction behavior includes: Based on the physical dimension type of the meteorological indicators, determine the deviation calculation rules; Based on the predicted behavior time span, a baseline coefficient is determined; The prediction deviation rate is determined based on the prediction results, the actual observation results, and the deviation calculation rules. The prediction deviation level of the meteorological data source on the meteorological index is obtained based on the prediction deviation rate and the benchmark coefficient.

4. The weather forecasting method based on multi-source data fusion according to claim 3, characterized in that, The step of determining the deviation calculation rules based on the physical dimension type of the meteorological indicators includes: If the meteorological indicator belongs to the precipitation category, the first type of deviation calculation rule shall be used. The first type of deviation calculation rule includes the relative error calculation method and / or the logarithmic transformation error calculation method. If the meteorological indicator belongs to the temperature category, the second type of deviation calculation rule shall be used. The second type of deviation calculation rule includes the absolute error calculation method and / or the mean square error calculation method. If the meteorological indicator belongs to the wind field category, the third type of deviation calculation rule shall be used. The third type of deviation calculation rule includes the ring difference error calculation method and / or the vector error calculation method.

5. The weather forecasting method based on multi-source data according to claim 1, characterized in that, The step of calculating the data fusion weight of the meteorological data source on each meteorological indicator based on the prediction deviation level of the meteorological data source includes: Based on the prediction deviation level of each meteorological indicator at the target location within a preset time window, calculate the average prediction deviation level of each meteorological indicator. Based on the predicted deviation level, the basic weight of the meteorological data source on the corresponding meteorological index is determined; Based on the prediction deviation level sequence within the preset time window, the trend of prediction deviation level changes for each meteorological indicator is obtained. Based on the rate of change of the trend of the prediction deviation level, the dynamic weight of the meteorological data source on the corresponding meteorological indicators is determined; Based on the basic weight and dynamic weight of the meteorological data source on the corresponding meteorological indicators, the data fusion weight of the meteorological data source on the corresponding meteorological indicators is obtained.

6. The weather forecasting method based on the fusion of multi-source data according to any one of claims 1 to 5, characterized in that, The weather forecasting method that integrates multi-source data also includes: Obtain the geographic region classification label of the target location; Based on the geographical region classification labels, the target meteorological data source is determined; Based on the target meteorological data source, a data source queue is obtained by filtering multiple meteorological data sources to be selected. Based on the data source queue, the step of obtaining meteorological forecast information for the target location is performed.

7. The weather forecasting method based on multi-source data according to claim 6, characterized in that, The step of determining the target meteorological data source based on the geographical region classification label includes: If the geographical region is categorized as mountainous, then meteorological data sources that collect meteorological data using radar technology are selected as the target meteorological data source. If the geographical region is categorized as ocean, then meteorological data sources collected via satellite remote sensing technology are selected as the target meteorological data source. If the geographical region is categorized as a city, then meteorological data sources collected through a network of ground observation stations are selected as the target meteorological data source. If the geographical region is classified as a plain, then at least one type of meteorological data source from data collected through satellite remote sensing technology and ground observation station networks will be selected as the target meteorological data source.

8. A weather forecasting device that integrates multi-source data, characterized in that, The weather forecasting device that integrates multi-source data includes: The data acquisition module is used to acquire meteorological forecast information for the target location, wherein the meteorological forecast information includes forecast information from at least two different meteorological data sources; The deviation analysis module is used to obtain the prediction deviation level of the meteorological data source on each meteorological indicator based on the historical meteorological forecast results recorded by the meteorological data source and the meteorological observation data records of the target location. The weight allocation module is used to calculate the data fusion weight of the meteorological data source on each meteorological indicator based on the prediction deviation level of the meteorological data source on each meteorological indicator. The fusion prediction module is used to perform multi-source data fusion on each meteorological indicator in the meteorological prediction information based on the data fusion weight of the meteorological data source on each meteorological indicator, so as to obtain the fused weather prediction result.

9. A weather forecasting device that integrates multi-source data, characterized in that, The weather forecasting device that integrates multi-source data includes: a memory, a processor, and a weather forecasting program that integrates multi-source data stored in the memory and executable on the processor, wherein the weather forecasting program that integrates multi-source data is configured to implement the steps of the weather forecasting method that integrates multi-source data as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores a weather forecasting program that integrates multi-source data. When the weather forecasting program that integrates multi-source data is executed by a processor, it implements the steps of the weather forecasting method that integrates multi-source data as described in any one of claims 1 to 7.

Images