Dynamic science popularization pushing system based on real-time meteorological data

Abstract

The application relates to the field of meteorological popular science technology, in particular to a dynamic popular science pushing system based on real-time meteorological data. A perception threshold is determined by acquiring historical meteorological data and user feedback, a perceptible area is obtained by calculating a space-time gradient comparison based on a meteorological data sequence and a user position, an evolution stage sequence is calculated by analyzing a boundary moving direction, a physical process is mapped and popular science knowledge is organized, and the popular science knowledge is pushed when the user is located in the perceptible area. The method realizes accurate popular science pushing based on real-time meteorological changes.

Description

Technical Field

[0001] This invention relates to the field of meteorological science popularization technology, and in particular to a dynamic science popularization push system based on real-time meteorological data. Background Technology

[0002] Current meteorological science popularization push technologies mainly rely on fixed meteorological warning mechanisms and static content distribution models. Meteorological departments typically classify warning levels based on pre-set meteorological element thresholds (such as temperature, rainfall, wind speed, etc.) and broadcast warning information or popular science articles to users based on administrative regions or geographic grids. This popular science content is often pre-compiled general knowledge text, such as introductions to typhoon paths and explanations of the causes of heavy rain, lacking personalized adjustments based on the user's real-time location and dynamic weather changes. Users may receive push notifications unrelated to their actual meteorological experience; for example, a user far from precipitation areas may receive heavy rain defense information, or a user may receive early warning content even when the weather is in its late stages.

[0003] This conventional approach results in inaccurate timing and scope of push notifications, failing to differentiate whether users are in areas where weather changes are perceptible. Fixed thresholds and administrative boundaries ignore the spatiotemporal gradient differences of meteorological elements, leading to potentially drastically different experiences of the same weather event for users in different locations, yet the push notifications are applied uniformly, resulting in information redundancy or missing crucial information. Furthermore, the rigid organization of science popularization content lacks a dynamic connection to the evolution of meteorology. The science popularization units received by users are presented in isolation, failing to reflect the complete physical chain of meteorological elements from formation, development to dissipation, making it difficult for users to understand the causes and trends behind the weather phenomena they are currently experiencing, thus reducing the effectiveness of science education. Summary of the Invention

[0004] This invention provides a dynamic science popularization push system based on real-time meteorological data, which can solve the problems in the prior art.

[0005] A first aspect of this invention provides a method for dynamically pushing popular science information based on real-time meteorological data, comprising:

[0006] Acquire historical meteorological data and user meteorological perception feedback data, calculate the correspondence between the changes in meteorological elements in historical meteorological data and the perception response in user meteorological perception feedback data, and determine the meteorological element perception threshold based on the correspondence.

[0007] Acquire meteorological data sequences and user locations for the target area. The meteorological data sequences include meteorological element observations from multiple spatial locations. Calculate the spatiotemporal gradient of the meteorological element observations and compare the spatiotemporal gradient with the meteorological element perception threshold to determine the area of ​​perceptible meteorological changes.

[0008] Analyze the spatial distribution changes of meteorological element observations in meteorological data sequences, identify the boundary locations and movement directions of meteorological element observations, and deduce the meteorological evolution stages that the user's location will experience based on the spatial relationship between the user's location and the boundary locations and movement directions.

[0009] The meteorological evolution stages in the meteorological evolution stage sequence are mapped to meteorological physical processes. Popular science knowledge units corresponding to the meteorological physical processes are organized according to the temporal relationship of the meteorological evolution stage sequence to generate popular science content.

[0010] Determine if the user's location is within an area where weather changes can be perceived. If the user's location is within an area where weather changes can be perceived, push science popularization content to the user's terminal.

[0011] Calculate the correspondence between changes in meteorological elements in historical meteorological data and perceived responses in user meteorological perception feedback data, and determine the meteorological element perception thresholds based on this correspondence, including:

[0012] Identify the time abrupt change points and spatial gradient jump points of meteorological element observation values ​​from historical meteorological data, extract the time variation amplitude of the time abrupt change points and the spatial variation amplitude of the spatial gradient jump points, and combine the time variation amplitude and spatial variation amplitude into the meteorological element variation quantity.

[0013] Extract the perception response time from the user's meteorological perception feedback data, align the perception response time with the time series of historical meteorological data, and obtain the change in meteorological elements corresponding to the perception response time.

[0014] The number of matching samples between the change in meteorological elements and the sensing response time is counted. The distribution curve of the number of matching samples with the change in meteorological elements is calculated. The inflection point of the distribution curve is identified. The matching relationship between the change in meteorological elements and the inflection point is used as the correspondence. The meteorological element sensing threshold is determined based on the change in meteorological elements corresponding to the inflection point in the correspondence.

[0015] The user meteorological perception feedback data includes meteorological perception feedback information submitted by users and records of user interaction behavior.

[0016] Calculate the spatiotemporal gradient of meteorological element observations, compare the spatiotemporal gradient with the meteorological element perception threshold, and determine the areas where meteorological changes can be perceived, including:

[0017] The meteorological element observation values ​​at multiple spatial locations at consecutive times are extracted from the meteorological data sequence. The difference between the meteorological element observation values ​​at adjacent times is calculated to obtain the temporal gradient. The difference between the meteorological element observation values ​​at adjacent spatial locations is calculated to obtain the spatial gradient. The spatiotemporal gradient of the meteorological element observation values ​​is constructed based on the rate of change of the temporal gradient and the distribution density of the spatial gradient.

[0018] The spatiotemporal gradient is compared with the meteorological element perception threshold on a spatial basis to identify the spatial location where the spatiotemporal gradient continuously exceeds the meteorological element perception threshold and record the duration of exceeding the threshold. Spatial locations where the duration of exceeding the threshold is greater than the preset duration are extracted to form a candidate change location set.

[0019] Calculate the ratio of the temporal gradient to the spatial gradient magnitude of each spatial location in the candidate change location set, extract spatial locations whose ratio falls within a preset range as spatiotemporal coupling locations, extract connected components from the spatiotemporal coupling locations, and determine the area delineated by the connected components as the area of ​​perceptible meteorological changes.

[0020] Analyzing the spatial distribution changes of meteorological element observations in meteorological data sequences, identifying the boundary locations and directions of movement of these observations, and based on the spatial relationship between the user's location and these boundary locations and directions of movement, the sequence of meteorological evolution stages that the user's location will experience is estimated, including:

[0021] Spatial interpolation is performed on meteorological element observations at multiple spatial locations in a meteorological data sequence to generate a spatial distribution field. The spatial gradient of the meteorological element observations in the spatial distribution field is calculated, and the extreme value locations of the spatial gradient are identified. The extreme value locations are determined as the boundary locations formed by the meteorological element observations.

[0022] Extract the spatial coordinates of the boundary position at consecutive time steps, and calculate the displacement vector of the spatial coordinates at adjacent time steps to determine the direction of movement of the boundary position;

[0023] Based on the spatial coordinates of the boundary position and the direction of movement, the predicted coordinates of the boundary position at multiple future times are calculated by linear extrapolation. The spatial distance between the user position and the predicted coordinates is calculated, and the predicted coordinates and the corresponding arrival times of the spatial distances less than a preset distance threshold are identified. The spatial relationship between the user position and the boundary position and the direction of movement is established.

[0024] Meteorological element observation values ​​are extracted from both sides of the boundary position corresponding to the arrival time from the spatial distribution field. The numerical difference between the meteorological element observation values ​​on both sides of the boundary position is calculated. The boundary position is divided into different meteorological evolution stages according to the numerical difference. The meteorological evolution stages are arranged in chronological order according to the arrival time. The sequence of meteorological evolution stages that the user's position will experience is calculated and generated.

[0025] Based on the spatial coordinates of the boundary location and the direction of movement, linear extrapolation is used to calculate the predicted coordinates of the boundary location at multiple future times, including:

[0026] Extract the displacement vector magnitude of the boundary position according to the direction of movement, calculate the ratio of the displacement vector magnitude to the time interval, and obtain the movement rate of the boundary position along the direction of movement.

[0027] The displacement vector corresponding to the preset extrapolation time step is calculated based on the movement speed and direction. The current spatial coordinates of the boundary position are vector-superimposed with the displacement vector to obtain the predicted coordinates of the boundary position after an extrapolation time step.

[0028] Multiple extrapolation time steps are set, and the displacement vector corresponding to each extrapolation time step is calculated sequentially according to the movement speed and movement direction. The current spatial coordinates of the boundary position are vector superimposed with each displacement vector, and the predicted coordinates of the boundary position at multiple future times are obtained by linear extrapolation calculation based on the spatial coordinates of the boundary position and the movement direction.

[0029] The meteorological evolution stages in the meteorological evolution sequence are mapped to meteorological physical processes. Popular science knowledge units corresponding to these meteorological physical processes are organized according to the temporal relationship of the meteorological evolution stage sequence, generating popular science content including:

[0030] Extract the types and characteristics of meteorological elements in each stage of the meteorological evolution sequence, and retrieve matching meteorological physical processes from a pre-set meteorological physical process knowledge base based on the types and characteristics of meteorological elements.

[0031] Each meteorological evolution stage is mapped to a corresponding meteorological physical process to obtain a sequence of meteorological physical processes;

[0032] Extract the output state and input conditions of adjacent meteorological physical processes in a meteorological physical process sequence;

[0033] The causal relationship between adjacent meteorological and physical processes is determined based on the correspondence between output states and input conditions. Based on the meteorological and physical processes and the causal relationship, the corresponding popular science knowledge units are retrieved from the preset popular science knowledge base. The popular science knowledge units include the mechanism explanation of the meteorological and physical processes and the transitional statements of the causal relationship.

[0034] The popular science knowledge units are arranged according to the temporal relationship of the meteorological evolution stage sequence. Transitional statements are inserted between adjacent popular science knowledge units. The transitional statements connect the mechanism descriptions in the preceding popular science knowledge units with the mechanism descriptions in the subsequent popular science knowledge units, organize the popular science knowledge units corresponding to meteorological physical processes, and generate popular science content.

[0035] Determine whether the user's location is within an area where weather changes can be perceived. If the user's location is within an area where weather changes can be perceived, push science popularization content to the user's terminal, including:

[0036] Obtain the geographic coordinates of the user's location, calculate the shortest distance between the geographic coordinates of the user's location and the geographic boundary of the area where weather changes can be perceived, determine whether the shortest distance is less than a preset entry distance threshold, and determine that the user's location is in the area where weather changes can be perceived when the shortest distance is less than the preset entry distance threshold.

[0037] Extract the current meteorological evolution stage that the user's location will experience from the meteorological evolution stage sequence, select matching popular science knowledge units from popular science content based on the meteorological physical processes corresponding to the current meteorological evolution stage, and extract the mechanism explanations and transition sentences in the popular science knowledge units to combine them into popular science content to be pushed.

[0038] The urgency of the push is calculated based on the shortest distance, and a push message containing popular science content is generated based on the urgency. The push message is then sent to the user's terminal corresponding to the user's location.

[0039] A second aspect of this invention provides a dynamic science popularization push system based on real-time meteorological data, comprising:

[0040] The perception threshold unit is used to acquire historical meteorological data and user meteorological perception feedback data, calculate the correspondence between the changes in meteorological elements in historical meteorological data and the perception response in user meteorological perception feedback data, and determine the meteorological element perception threshold based on the correspondence.

[0041] The perceptible area unit is used to acquire meteorological data sequences and user locations for a target area. The meteorological data sequences include meteorological element observations from multiple spatial locations. The spatiotemporal gradient of the meteorological element observations is calculated, and the spatiotemporal gradient is compared with the meteorological element perception threshold to determine the perceptible meteorological change area.

[0042] The evolution stage unit is used to analyze the spatial distribution changes of meteorological element observations in meteorological data sequences, identify the boundary positions and movement directions of meteorological element observations, and calculate the meteorological evolution stage sequence that the user's position will experience based on the spatial relationship between the user's position and the boundary positions and movement directions.

[0043] The science popularization generation unit is used to map each meteorological evolution stage in the meteorological evolution stage sequence to meteorological physical processes, organize the science popularization knowledge units corresponding to the meteorological physical processes according to the temporal relationship of the meteorological evolution stage sequence, and generate science popularization content.

[0044] The sensing and push unit is used to determine whether the user's location is in an area where weather changes can be sensed. When the user's location is in an area where weather changes can be sensed, popular science content will be pushed to the user's terminal.

[0045] A third aspect of the present invention provides an electronic device, comprising:

[0046] processor;

[0047] Memory used to store processor-executable instructions;

[0048] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0049] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0050] This invention, based on in-depth analysis of real-time meteorological data to understand user perception responses and the changing patterns of meteorological elements, accurately determines individualized meteorological element perception thresholds, thus strictly limiting the scope of science popularization pushes to areas of meteorological change that users can actually perceive. This adaptation mechanism, tailored to user subjective experience, effectively avoids interference and cognitive burden caused by indiscriminate pushes, ensuring that each piece of science popularization content accurately corresponds to the weather changes that users are currently experiencing or will soon intuitively experience. This significantly improves the acceptability and practical value of science popularization information, transforming meteorological science popularization from passive reception to an active cognitive process closely linked to life experiences. By continuously tracking spatial gradients and boundary movements in meteorological data sequences, the invention can intelligently calculate the sequence of meteorological evolution stages that the user's location will experience, mapping each stage to a corresponding meteorological physical process. Science popularization knowledge units are organized according to the natural temporal sequence of spatiotemporal evolution, forming a coherent knowledge chain synchronized with the actual evolution of weather. Users receive not isolated, fragmented information, but a comprehensive interpretation covering the causes, evolution, and impacts of weather changes. This helps users understand the physical logic behind weather changes from a dynamic perspective, achieving a seamless connection between science popularization content and natural phenomena, greatly enhancing the systematic nature and comprehensibility of meteorological knowledge. By combining the identification of areas of perceptible weather change with the user's real-time location, the timing of push notifications is highly correlated with the user's spatial location, ensuring that science popularization content is delivered promptly when the user needs it most. When a user is about to enter a perceptible area, the push notification is automatically triggered, and the content is dynamically updated according to the user's location and weather evolution, forming a fluid, personalized science popularization service. This precise spatiotemporal push mechanism not only avoids cognitive misalignment caused by information lag or advancement, but also enhances the fun and immersion of science popularization through real-time interaction, effectively stimulating users' active attention to and desire to explore meteorological science. Attached Figure Description

[0051] Figure 1 This is a flowchart illustrating a dynamic science popularization push method based on real-time meteorological data.

[0052] Figure 2 This is a flowchart of the operation for extrapolating the sequence of meteorological evolution stages. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0055] Figure 1 This is a flowchart illustrating the dynamic science popularization push method based on real-time meteorological data according to an embodiment of the present invention.

[0056] Dynamic science popularization methods based on real-time meteorological data include:

[0057] Acquire historical meteorological data and user meteorological perception feedback data, calculate the correspondence between the changes in meteorological elements in historical meteorological data and the perception response in user meteorological perception feedback data, and determine the meteorological element perception threshold based on the correspondence.

[0058] Acquire meteorological data sequences and user locations for the target area. The meteorological data sequences include meteorological element observations from multiple spatial locations. Calculate the spatiotemporal gradient of the meteorological element observations and compare the spatiotemporal gradient with the meteorological element perception threshold to determine the area of ​​perceptible meteorological changes.

[0059] Analyze the spatial distribution changes of meteorological element observations in meteorological data sequences, identify the boundary locations and movement directions of meteorological element observations, and deduce the meteorological evolution stages that the user's location will experience based on the spatial relationship between the user's location and the boundary locations and movement directions.

[0060] The meteorological evolution stages in the meteorological evolution stage sequence are mapped to meteorological physical processes. Popular science knowledge units corresponding to the meteorological physical processes are organized according to the temporal relationship of the meteorological evolution stage sequence to generate popular science content.

[0061] Determine if the user's location is within an area where weather changes can be perceived. If the user's location is within an area where weather changes can be perceived, push science popularization content to the user's terminal.

[0062] In this invention, historical meteorological data is obtained through a standard interface connecting to a meteorological data service system. The interface returns structured data records containing timestamps, spatial coordinates, and observed values ​​of meteorological elements. Historical meteorological data refers to observation records of meteorological elements collected over a past period through meteorological observation stations, satellite remote sensing, radar detection, etc., including observed values ​​and changes in meteorological elements such as temperature, humidity, air pressure, wind speed, and precipitation at different times and spatial locations.

[0063] Calculate the correspondence between changes in meteorological elements in historical meteorological data and perceived responses in user meteorological perception feedback data, and determine the meteorological element perception thresholds based on this correspondence, including:

[0064] Identify the time abrupt change points and spatial gradient jump points of meteorological element observation values ​​from historical meteorological data, extract the time variation amplitude of the time abrupt change points and the spatial variation amplitude of the spatial gradient jump points, and combine the time variation amplitude and spatial variation amplitude into the meteorological element variation quantity.

[0065] Extract the perception response time from the user's meteorological perception feedback data, align the perception response time with the time series of historical meteorological data, and obtain the change in meteorological elements corresponding to the perception response time.

[0066] The number of matching samples between the change in meteorological elements and the sensing response time is counted. The distribution curve of the number of matching samples with the change in meteorological elements is calculated. The inflection point of the distribution curve is identified. The matching relationship between the change in meteorological elements and the inflection point is used as the correspondence. The meteorological element sensing threshold is determined based on the change in meteorological elements corresponding to the inflection point in the correspondence.

[0067] The user meteorological perception feedback data includes meteorological perception feedback information submitted by users and records of user interaction behavior.

[0068] This method identifies temporal abrupt changes in meteorological element observations from historical meteorological data. A sliding window differencing method is used to progressively scan the meteorological element time series, calculating the difference in the mean values ​​of meteorological elements within adjacent time windows. When the difference exceeds a preset detection sensitivity, the corresponding time point is marked as a candidate abrupt change point. Candidate abrupt changes are clustered and merged to remove pseudo-abrupt changes caused by data acquisition noise, retaining only physically meaningful true abrupt changes. For identifying spatial gradient jump points, the gradient field of meteorological elements between adjacent stations is calculated on the spatial observation grid at each time point. Second-order differencing is performed on the gradient field to identify spatial locations where significant jumps in gradient values ​​occur, marking these as spatial gradient jump points. The temporal variation amplitude at a temporal abrupt change point is defined as the absolute value of the difference between the mean values ​​of meteorological elements before and after the abrupt change. The spatial variation amplitude at a spatial gradient jump point is defined as the absolute value of the difference in the gradient values ​​of meteorological elements on both sides of the jump point. The temporal and spatial variation amplitudes are then combined using a weighted combination to form the meteorological element variation. The specific calculation method is as follows: ,in Indicates the magnitude of change over time. Indicates the magnitude of spatial variation. and These are the weighting coefficients for the time dimension and the spatial dimension, respectively, and the sum of the two is 1. The value of the weighting coefficient can be adjusted according to the physical characteristics of different meteorological elements. For example, the weight of the time dimension can be appropriately increased for the temperature element, and the weight of the spatial dimension can be appropriately increased for the precipitation intensity element.

[0069] User weather perception feedback data comprises two sources: one is weather perception feedback information actively submitted by users through applications, such as descriptions of their sensations when experiencing significant temperature drops, strong winds, or rainfall; the other is user interaction behavior records, including timestamps of when users open weather apps under specific weather conditions, the frequency of weather forecast inquiries, and browsing of meteorological science content. These two types of data together constitute the user weather perception feedback dataset, reflecting users' perceptual responses to weather changes from both active feedback and passive behavior dimensions.

[0070] When extracting the perception response time from user meteorological perception feedback data, for actively submitted perception feedback information, the timestamp of the user submission record is directly extracted as the perception response time; for user interaction behavior records, the moment when the frequency of user behavior significantly increases relative to the baseline level is identified, and this moment is used as a proxy indicator for the perception response time. When aligning the perception response time with the time series of historical meteorological data, a time alignment tolerance window is set considering the certain physiological and psychological delays in user perception. Before and after the perception response time Search for corresponding meteorological element change records within the time range. For multiple perception responses from the same user during the same meteorological event, the earliest perception response time is taken as the representative perception response time for that event to avoid duplicate counting interfering with the statistical results. After time alignment, each perception response record corresponds to a meteorological element change. The values ​​form a paired sample set of perceived responses and changes in meteorological elements.

[0071] Statistical changes in meteorological elements When matching the number of samples at the time of perception response, The value range is divided into several equally spaced intervals, and the number of matching samples in each interval is counted. ,in For interval numbering. The center value of the interval is the horizontal axis to match the number of samples. Plot the distribution curve with the vertical axis as the ordinate. This distribution curve physically reflects the frequency distribution of perceptible responses generated by users under different intensities of meteorological element changes. When the amount of change in meteorological elements is small, the number of matching samples is small, indicating that the intensity of meteorological change is insufficient to trigger a perceptible response from most users. As the amount of change in meteorological elements increases, the number of matching samples gradually increases. Near a certain specific amount of change, the growth rate of the number of matching samples changes significantly, forming an inflection point on the distribution curve.

[0072] When identifying the inflection point of the distribution curve, the curve is smoothed to eliminate the influence of statistical noise. A kernel density estimation method is used to smoothly fit the discrete sample counts, resulting in a continuous perceptual response density function. The second derivative of the smoothed distribution curve is calculated; the location where the absolute value of the second derivative reaches its maximum is the inflection point of the distribution curve. The inflection point physically corresponds to the meteorological element change with the highest rate of increase in perceived response frequency. That is, near this change, most users begin to generate a noticeable meteorological perception response. Therefore, the meteorological element change corresponding to this inflection point is determined as the meteorological element perception threshold. If the distribution curve has multiple inflection points, the first significant inflection point (i.e., the inflection point with the lowest meteorological element change) is taken as the perception threshold to ensure that the push strategy can cover users with high perception sensitivity.

[0073] The process of determining the perception thresholds for meteorological elements is performed separately for different meteorological elements. For example, temperature, wind speed, air pressure, and humidity each have independent perception thresholds. In practical applications, users in different geographical regions have significantly different meteorological perception thresholds due to differences in climate adaptability. For instance, users who live in high-temperature areas are generally less sensitive to temperature changes than users living in temperate areas. Therefore, when calculating the perception thresholds for meteorological elements, the perception feedback data is grouped according to the user's geographical affiliation, and the regional meteorological element perception thresholds for each region are calculated separately, making the perception thresholds regionally adaptable. Furthermore, the perception thresholds are not fixed. As new user perception feedback data accumulates, the distribution curves are periodically updated and recalculated, and the meteorological element perception thresholds are continuously revised to ensure that they always reflect the actual perception characteristics of the current user group, thereby providing an accurate quantitative basis for subsequent determination of areas where meteorological changes can be perceived.

[0074] Calculate the spatiotemporal gradient of meteorological element observations, compare the spatiotemporal gradient with the meteorological element perception threshold, and determine the areas where meteorological changes can be perceived, including:

[0075] The meteorological element observation values ​​at multiple spatial locations at consecutive times are extracted from the meteorological data sequence. The difference between the meteorological element observation values ​​at adjacent times is calculated to obtain the temporal gradient. The difference between the meteorological element observation values ​​at adjacent spatial locations is calculated to obtain the spatial gradient. The spatiotemporal gradient of the meteorological element observation values ​​is constructed based on the rate of change of the temporal gradient and the distribution density of the spatial gradient.

[0076] The spatiotemporal gradient is compared with the meteorological element perception threshold on a spatial basis to identify the spatial location where the spatiotemporal gradient continuously exceeds the meteorological element perception threshold and record the duration of exceeding the threshold. Spatial locations where the duration of exceeding the threshold is greater than the preset duration are extracted to form a candidate change location set.

[0077] Calculate the ratio of the temporal gradient to the spatial gradient magnitude of each spatial location in the candidate change location set, extract spatial locations whose ratio falls within a preset range as spatiotemporal coupling locations, extract connected components from the spatiotemporal coupling locations, and determine the area delineated by the connected components as the area of ​​perceptible meteorological changes.

[0078] Extracting meteorological element observations from multiple spatial locations at consecutive time points from a meteorological data sequence is a fundamental step in constructing a spatiotemporal gradient. Meteorological data sequences are typically stored in a gridded format, with each grid node corresponding to a spatial location and recording meteorological element observations at that location at several consecutive time points, such as temperature, air pressure, wind speed, and relative humidity. For any spatial location... Extract its time point and adjacent time Meteorological element observation values and The time gradient is obtained by calculating the difference between the two. This value reflects the rate of change of meteorological elements over time. For the spatial gradient, the selection is related to location. Adjacent spatial locations At the same time Calculation This value describes the spatial distribution differences of meteorological elements. The temporal gradient and spatial gradient characterize the local variation features of meteorological elements from the two dimensions of the time axis and the spatial axis, respectively.

[0079] Building upon the temporal and spatial gradients, a comprehensive spatiotemporal gradient is further constructed. The spatiotemporal gradient is not simply the superposition of the temporal and spatial gradients, but rather a joint representation based on the rate of change of the temporal gradient and the distribution density of the spatial gradient. The rate of change of the temporal gradient is reflected in the magnitude of the difference between the temporal gradient values ​​at adjacent time points, i.e. This quantity reflects whether changes in meteorological elements show an accelerating or decelerating trend; the distribution density of the spatial gradient is statistically analyzed by location. Mean of the absolute value of the spatial gradient in the neighborhood of the center As a measure, higher density indicates a more concentrated spatial distribution of meteorological elements in the vicinity of that location. (This is based on the combined spatiotemporal gradient.) based on and The weighted combination construction, i.e. ,in The weighting coefficient for the rate of change over time, with a value range of [value missing]. It can be adjusted according to the type of meteorological elements and the application scenario. When slow variables such as temperature are the main objects of analysis, Smaller values ​​result in a greater contribution from spatial distribution density; when analyzing rapidly changing weather such as severe convection, The value is set to be relatively large to highlight the acceleration characteristics in the time dimension.

[0080] spatiotemporal gradient Sensing threshold of meteorological elements Perform time-by-time comparisons of spatial locations to determine... Does it exceed For each spatial location Statistical analysis in continuous time series The duration of continuous validity is denoted as the overthreshold duration. The unit is consistent with the sampling interval of the time series (e.g., minutes or hours). Gradient overthreshold phenomena that occur only at a single moment are often caused by observational noise or transient disturbances and have no meteorological or physical significance; only when the overthreshold state is sustained over time does it indicate that the location is indeed undergoing a significant meteorological change. Therefore, a preset duration threshold is set. Extract to satisfy Spatial location, forming a set of candidate change locations . The specific value can be determined based on the physical characteristics of meteorological elements, such as temperature changes. It can be set to 30 minutes, to account for changes in air pressure. It can be set to 60 minutes to filter out false over-threshold positions caused by short-term fluctuations.

[0081] Candidate change location set Although all spatial locations satisfy the threshold persistence condition, not all locations exhibit spatiotemporal coupling characteristics. Spatiotemporal coupling refers to the consistent and significant changes in meteorological elements at a given location across both temporal and spatial dimensions, rather than being prominent only in one dimension. Therefore, calculations... Each spatial location temporal gradient magnitude With spatial gradient magnitude ratio ,in Take the mean of the absolute values ​​of the temporal gradient at that location during the time-out period. Take the mean of the absolute values ​​of the spatial gradients at that location during the time-out period. Ratio It reflects the relative intensity relationship between temporal and spatial changes: when When the value is too large, it indicates that the changes at that location are mainly reflected in the time dimension, with a relatively uniform spatial distribution and no obvious spatial frontal zone characteristics; when If the gradient is too small, it indicates that the spatial gradient is much larger than the temporal gradient, meaning the change is rapid spatially but slow temporally, which also does not conform to the characteristics of dynamic change perceptible to users. Only when... Falling into the preset range Only then is the location recognized as a spatiotemporal coupling location and included in the set of spatiotemporal coupling locations. The upper and lower boundaries of the preset range can be defined according to the actual meteorological scenario, such as for a frontal passage scenario. 0.3 is acceptable. A value of 3.0 can be used to ensure that the extracted locations have significant variation characteristics in both time and space dimensions.

[0082] For the set of spatiotemporal coupling locations Performing connected component extraction is a crucial step in integrating discrete spatiotemporally coupled locations into geographically significant contiguous regions. Connected component extraction is based on spatial adjacency, grouping adjacent spatiotemporally coupled locations (including 4-neighbor or 8-neighbor areas) on the grid into the same connected component. For each identified connected component, the geographical area it covers constitutes a candidate region for perceptible meteorological changes. To avoid introducing noise from a few isolated points or extremely small connected components, a minimum area threshold is set for the connected components. Only retain an area of ​​not less than The connected regions correspond to the areas defined by the area conditions. Ultimately, the areas defined by all connected regions that satisfy the area condition are identified as areas of perceptible meteorological change. These areas are geographically represented by several consecutive polygonal ranges, signifying zones within the target area where meteorological elements are undergoing significant and perceptible dynamic changes. In subsequent steps, the user's location will be spatially correlated with these areas of perceptible meteorological change to determine whether to trigger the push of science-related content.

[0083] like Figure 2 As shown, Figure 2 This is a flowchart illustrating the operation of meteorological evolution stage sequence estimation in an embodiment of the present invention.

[0084] Analyzing the spatial distribution changes of meteorological element observations in meteorological data sequences, identifying the boundary locations and directions of movement of these observations, and based on the spatial relationship between the user's location and these boundary locations and directions of movement, the sequence of meteorological evolution stages that the user's location will experience is estimated, including:

[0085] Spatial interpolation is performed on meteorological element observations at multiple spatial locations in a meteorological data sequence to generate a spatial distribution field. The spatial gradient of the meteorological element observations in the spatial distribution field is calculated, and the extreme value locations of the spatial gradient are identified. The extreme value locations are determined as the boundary locations formed by the meteorological element observations.

[0086] Extract the spatial coordinates of the boundary position at consecutive time steps, and calculate the displacement vector of the spatial coordinates at adjacent time steps to determine the direction of movement of the boundary position;

[0087] Based on the spatial coordinates of the boundary position and the direction of movement, the predicted coordinates of the boundary position at multiple future times are calculated by linear extrapolation. The spatial distance between the user position and the predicted coordinates is calculated, and the predicted coordinates and the corresponding arrival times of the spatial distances less than a preset distance threshold are identified. The spatial relationship between the user position and the boundary position and the direction of movement is established.

[0088] Meteorological element observation values ​​are extracted from both sides of the boundary position corresponding to the arrival time from the spatial distribution field. The numerical difference between the meteorological element observation values ​​on both sides of the boundary position is calculated. The boundary position is divided into different meteorological evolution stages according to the numerical difference. The meteorological evolution stages are arranged in chronological order according to the arrival time. The sequence of meteorological evolution stages that the user's position will experience is calculated and generated.

[0089] To analyze the spatial distribution variations of meteorological element observations in meteorological data sequences, it is necessary to construct a continuous spatial distribution field starting from discrete observation station data. Due to the uneven distribution of stations in actual meteorological observation networks, the meteorological element observations (such as temperature, air pressure, wind speed, and humidity) collected by each station only represent the local state of a discrete spatial location and cannot directly reflect the continuous spatial distribution characteristics of the entire target area. Therefore, spatial interpolation is performed on meteorological element observations at multiple spatial locations in the meteorological data sequence to generate a spatial distribution field covering the target area. Kriging interpolation or inverse distance weighted interpolation can be used, with the appropriate method selected based on station density and the spatial autocorrelation characteristics of meteorological elements. Kriging interpolation can use variograms to describe the spatial variation structure of meteorological elements and is suitable for scenarios with sparse station distribution but significant spatial correlation; inverse distance weighted interpolation has higher computational efficiency and is suitable for situations with relatively uniform station distribution. After interpolation, a spatial distribution field of meteorological element observations defined on a regular grid is obtained, with each grid node corresponding to an interpolated meteorological element value.

[0090] After the spatial distribution field is generated, the spatial gradient of meteorological element observations at each grid node within the field is calculated. The spatial gradient reflects the rate of change of meteorological elements in the horizontal direction. Regions with large gradient magnitudes indicate significant changes in meteorological elements over short distances, typically corresponding to meteorological boundary structures such as the boundary between cold and warm air masses, frontal systems, or the edge of precipitation belts. A global search is performed on the gradient magnitudes in the spatial distribution field to identify extreme locations of the gradient magnitudes, i.e., the set of spatial nodes where the gradient magnitude reaches a local maximum. These extreme locations physically correspond to interfaces where meteorological element observations undergo abrupt spatial changes, and are thus identified as the boundary locations where meteorological element observations form. To avoid false extremes caused by noise interference, connectivity filtering is performed on the extreme locations, retaining only spatially continuous chains of extreme points with a length exceeding a preset minimum boundary length as valid meteorological boundary locations.

[0091] Spatial coordinates of the identified boundary positions at multiple consecutive time points are extracted to construct a time series of the boundary positions. For boundary positions at adjacent time points, nearest neighbor matching or the Hungarian algorithm is used to correlate boundary points across time periods, ensuring that the positions of the same meteorological boundary at different time points can be correctly mapped. The displacement vector between corresponding boundary points at adjacent time points is calculated; the direction of the displacement vector is the direction of boundary position movement, and the magnitude of the displacement vector divided by the time interval is the boundary movement velocity. To improve the stability of the movement direction estimation, a weighted average of the displacement vectors at multiple adjacent time points is performed, with greater weight assigned to times with closer time intervals, thus obtaining the average movement direction and velocity of the boundary position at the current time.

[0092] Based on the current spatial coordinates, direction of movement, and speed of movement at the boundary position, a linear extrapolation method is used to calculate the predicted coordinates of the boundary position at multiple future time points. Let the spatial coordinates of the boundary position at the current time be... The moving speed vector is Then the future number time step Predicted coordinates at location satisfy ,in Number the prediction time steps. This represents a single time step. Linear extrapolation assumes the boundary maintains uniform linear motion over a short period, suitable for situations where meteorological boundary movement is relatively stable. If historical displacement vector sequences show an acceleration or turning trend in the boundary, a second-order extrapolation correction term can be introduced to improve prediction accuracy. The spatial distance between the user's location and each predicted coordinate is calculated, identifying locations where the spatial distance is less than a preset distance threshold. Predicted coordinates and their corresponding arrival times The arrival time indicates the estimated time when the meteorological boundary will reach the vicinity of the user's location, thus establishing the spatial relationship between the user's location, the boundary location, and the direction of movement. When multiple predicted coordinates satisfy the distance condition, they are arranged in order of arrival time from nearest to farthest, forming an ordered sequence of boundary arrival events.

[0093] Meteorological element observations at each arrival time are extracted from the spatial distribution field, corresponding to the boundary positions on both sides. The sampling range on both sides of the boundary is a strip-shaped area extending a certain distance (e.g., 50 kilometers) along the boundary normal direction. The mean values ​​of meteorological element observations in the areas on both sides are calculated to obtain the mean value on the front side of the boundary. and the mean behind the boundary Calculate the numerical differences in meteorological element observations on both sides of the boundary. The sign and absolute value of the numerical differences together reflect the degree of contrast in meteorological conditions on both sides of the boundary. Based on the magnitude and sign of the numerical differences, the boundary location is divided into different stages of meteorological evolution. For example, when... When the value is negative and its absolute value exceeds the temperature difference threshold, it indicates that the temperature behind the boundary is significantly lower than that in front, corresponding to the passage of a cold front; when When the value is positive and its absolute value exceeds the humidity difference threshold, it indicates that the humidity behind the boundary is significantly higher than that in front, corresponding to the warm and humid airflow transport stage. The numerical differences of different meteorological elements can be combined to form a multi-dimensional basis for dividing meteorological evolution stages, making the stage identification results more accurate.

[0094] According to each arrival time The sequence of meteorological evolution stages is arranged chronologically to generate a sequence of meteorological evolution stages that the user's location will experience. This sequence describes the changes in meteorological conditions the user will encounter as they sequentially encounter various meteorological boundaries over a future period, such as "clear weather stage → warm and humid stage before cold front → cold front passage stage → clear and cold stage after cold front". The temporal structure of the meteorological evolution stage sequence provides a physical basis for the organization of subsequent science popularization content, ensuring that the science popularization knowledge units pushed are consistent with the meteorological processes that the user will actually experience in terms of time and content, thereby making the science popularization push targeted and real-time.

[0095] Based on the spatial coordinates of the boundary location and the direction of movement, linear extrapolation is used to calculate the predicted coordinates of the boundary location at multiple future times, including:

[0096] Extract the displacement vector magnitude of the boundary position according to the direction of movement, calculate the ratio of the displacement vector magnitude to the time interval, and obtain the movement rate of the boundary position along the direction of movement.

[0097] The displacement vector corresponding to the preset extrapolation time step is calculated based on the movement speed and direction. The current spatial coordinates of the boundary position are vector-superimposed with the displacement vector to obtain the predicted coordinates of the boundary position after an extrapolation time step.

[0098] Multiple extrapolation time steps are set, and the displacement vector corresponding to each extrapolation time step is calculated sequentially according to the movement speed and movement direction. The current spatial coordinates of the boundary position are vector superimposed with each displacement vector, and the predicted coordinates of the boundary position at multiple future times are obtained by linear extrapolation calculation based on the spatial coordinates of the boundary position and the movement direction.

[0099] After obtaining the spatial coordinates and direction of movement of the boundary location, it is necessary to quantitatively predict the future location of the boundary in order to determine the meteorological evolution stages that the user's location will experience. Linear extrapolation is a method with low computational cost and strong applicability to short-term forecasts. Its core lies in extending the displacement characteristics of the boundary within historical observation periods to future moments.

[0100] When extracting the displacement vector magnitude at the boundary location, the Euclidean distance between the boundary space coordinates at two adjacent observation times is calculated based on the boundary space coordinates at those two times. This yields the actual displacement of the boundary along the direction of movement within that time interval. Let the time interval between two adjacent observation times be denoted as . The magnitude of the displacement vector of the boundary during this time interval is The rate of movement of the boundary position along the direction of movement. Determined by the following formula:

[0101]

[0102] in The units are consistent with the spatial and time units of the coordinate system; for example, in a geographic coordinate system, it can be expressed as kilometers per hour. The direction of movement is expressed as a unit direction vector. This means that the vector is obtained by normalizing the difference vector of the boundary coordinates at adjacent time points, ensuring that the direction information and the velocity information are independent of each other, which facilitates subsequent combined calculations.

[0103] In determining the movement rate and unit direction vector Then, for a preset single extrapolation time step Calculate the corresponding displacement vector :

[0104]

[0105] Spatial coordinates of the current boundary position With displacement vector By performing vector superposition, the predicted coordinates of the boundary position after one extrapolation time step can be obtained. :

[0106]

[0107] The predicted coordinates represent the boundary in the future. The estimated spatial location at any given time. The vector overlay operation is performed independently on the longitude and latitude components in a two-dimensional geographic coordinate space, thus ensuring that the predicted coordinates have a clear spatial meaning in the geographic reference system.

[0108] To obtain the boundary prediction coordinates at multiple future moments, multiple extrapolation time steps are set, with the step number denoted as . The value range is positive integers, and the corresponding extrapolated time step sequence is: ,in This is the preset maximum number of extrapolation steps. Under the equal step size setting, each step size takes the same value. , No. The cumulative time offset corresponding to each time step is The corresponding displacement vector for:

[0109]

[0110] Will With each displacement vector By performing vector superposition separately, we can obtain the future nth Predicted coordinates of the boundary position at each time step :

[0111]

[0112] In turn Performing the above calculations generates a complete sequence of predicted boundary coordinates. The sequence is uniformly distributed along the time axis, and the time interval between any two adjacent predicted coordinates is 1. .

[0113] In practical applications, the extrapolation time step The selection of the boundary needs to comprehensively consider both the movement speed of the meteorological boundary and the push response requirements of the user terminal. For severe convective weather boundaries with relatively fast movement speeds, it is recommended to [select the boundary]. Set to a smaller value (e.g., 5 to 10 minutes) to obtain a higher time resolution; for slow-moving frontal systems, you can... Increase the timeframe appropriately (e.g., from 30 to 60 minutes) to reduce unnecessary computational overhead. Maximum extrapolation steps. The coordinate error is usually determined based on the effective forecast lead time of the meteorological process; the error will increase if the forecast lead time exceeds the effective forecast lead time. As the confidence level of the predicted coordinates increases, the extrapolation sequence is terminated when the confidence level of the predicted coordinates is lower than the preset level. Therefore, in engineering implementation, an error tolerance can be set.

[0114] Movement speed The calculation can also use observation data from multiple historical moments for smoothing, rather than relying solely on the difference between two adjacent moments. For example, taking the most recent... For each consecutive moment of boundary coordinates, calculate the ratio of the displacement vector magnitude to the time interval between adjacent moments, and then... The smoothed moving speed is obtained by taking the arithmetic mean of the speed values. :

[0115]

[0116] in For the first Displacement vector magnitude over a time interval This corresponds to the time interval length. Smoothing of the direction of movement can also be achieved by normalizing the mean of the direction vector components, thereby reducing the impact of single observation errors on extrapolation results and improving the stability of the predicted coordinate sequence.

[0117] With non-uniform step size settings, the extrapolation time step sequence can be flexibly adjusted according to actual needs. For example, a smaller step size can be used in the early stages of forecasting to improve near-term forecast accuracy, while a larger step size can be used in the later stages of forecasting to reduce computational load. At this point... The cumulative time offset corresponding to each time step is The corresponding displacement vector is The calculation logic for predicted coordinates is exactly the same as that for the case of equal step size, except that the cumulative time offset is replaced with the sum of each step size.

[0118] After calculating the boundary prediction coordinate sequence, a spatial relationship analysis is performed between this sequence and the user's location coordinates to determine whether the user's location is within the influence range of the boundary's future movement path. This determines the expected arrival time of the boundary at the user's location, providing a time anchor for the subsequent calculation of meteorological evolution stages and the timely delivery of popular science content. The entire linear extrapolation process relies solely on historical boundary coordinate observations, without requiring the introduction of complex numerical weather prediction models. It features high computational efficiency and real-time updates, making it suitable for deployment in dynamic popular science delivery scenarios aimed at the public.

[0119] The meteorological evolution stages in the meteorological evolution sequence are mapped to meteorological physical processes. Popular science knowledge units corresponding to these meteorological physical processes are organized according to the temporal relationship of the meteorological evolution stage sequence, generating popular science content including:

[0120] Extract the types and characteristics of meteorological elements in each stage of the meteorological evolution sequence, and retrieve matching meteorological physical processes from a pre-set meteorological physical process knowledge base based on the types and characteristics of meteorological elements.

[0121] Each meteorological evolution stage is mapped to a corresponding meteorological physical process to obtain a sequence of meteorological physical processes;

[0122] Extract the output state and input conditions of adjacent meteorological physical processes in a meteorological physical process sequence;

[0123] The causal relationship between adjacent meteorological and physical processes is determined based on the correspondence between output states and input conditions. Based on the meteorological and physical processes and the causal relationship, the corresponding popular science knowledge units are retrieved from the preset popular science knowledge base. The popular science knowledge units include the mechanism explanation of the meteorological and physical processes and the transitional statements of the causal relationship.

[0124] The popular science knowledge units are arranged according to the temporal relationship of the meteorological evolution stage sequence. Transitional statements are inserted between adjacent popular science knowledge units. The transitional statements connect the mechanism descriptions in the preceding popular science knowledge units with the mechanism descriptions in the subsequent popular science knowledge units, organize the popular science knowledge units corresponding to meteorological physical processes, and generate popular science content.

[0125] After determining the sequence of meteorological evolution stages, it is necessary to establish a mapping relationship between each meteorological evolution stage in the sequence and specific meteorological physical processes, and organize popular science knowledge units accordingly to form coherent popular science content. To achieve this goal, the meteorological element types and meteorological element change characteristics of each stage are extracted one by one from the meteorological evolution stage sequence. Meteorological element types include basic meteorological variables such as temperature, humidity, air pressure, wind speed, and precipitation intensity, while meteorological element change characteristics describe attributes such as the direction of change (increase or decrease), the magnitude of change (weak, moderate, or drastic), and the rate of change (slow or rapid) of meteorological elements within that stage. These attribute information together constitute a structured description of the meteorological state evolution pattern of that stage, serving as query conditions for subsequent knowledge base retrieval.

[0126] The pre-defined meteorological physical process knowledge base stores a large number of standardized meteorological physical process entries. Each entry includes the types of meteorological elements involved in the physical process, the typical range of meteorological element changes, and a description of the physical mechanism of the process. During retrieval, the types of meteorological elements and the characteristics of meteorological element changes in each meteorological evolution stage are matched with the corresponding attributes of each entry in the knowledge base to select the meteorological physical process that best matches the description of that stage. For example, when a meteorological evolution stage is characterized by a continuous decrease in air pressure, a significant increase in humidity, and a gradual increase in wind speed, the search result may match the meteorological physical process of "approaching low-pressure system"; when the stage characteristics are characterized by a sharp increase in short-term precipitation intensity and a sudden drop in temperature, it matches the process of "convective precipitation triggering". By performing the above search operation on each stage in the meteorological evolution stage sequence, meteorological physical processes corresponding to each stage are obtained. These physical processes are arranged chronologically according to the meteorological evolution stage sequence to form a meteorological physical process sequence.

[0127] After establishing the sequence of meteorological physical processes, it is necessary to analyze the connection between adjacent meteorological physical processes in the sequence. Each meteorological physical process is defined in the knowledge base with its output state and input conditions: the output state describes the state of meteorological elements in the atmospheric system at the end of the process, and the input conditions describe the preceding atmospheric state required for the process to occur. The output state of the preceding process and the input conditions of the following process are extracted from two adjacent meteorological physical processes, and their correspondence is analyzed. If there is a significant connection between the output state of the preceding process and the input conditions of the following process in terms of meteorological element type and numerical range, a direct causal relationship is determined between them; if there is an intermediate transitional state, an auxiliary process that can bridge this intermediate state is further retrieved from the knowledge base and included in the description of the causal relationship. Determining the causal relationship allows popular science content to reveal the inherent logical connections between meteorological phenomena, rather than introducing individual physical processes in isolation.

[0128] After clarifying the meteorological physical processes and their causal relationships, the system retrieves corresponding science knowledge units for each meteorological physical process and transitional statements for each causal relationship from a pre-defined science knowledge base. Each science knowledge unit in the knowledge base contains two parts: first, a mechanism explanation, describing the causes, physical mechanisms, and typical manifestations of the meteorological physical process in easy-to-understand language, suitable for explaining the essence of meteorological phenomena to ordinary users; second, transitional statements required to connect with adjacent science knowledge units. The purpose of these transitional statements is to naturally connect the result state described in the mechanism explanation of the preceding science knowledge unit with the starting conditions described in the mechanism explanation of the subsequent science knowledge unit, ensuring logical coherence of the overall science content. The retrieval of transitional statements uses the causal relationship as the index key, matches standardized statement templates describing the relationship from the knowledge base, and fills in parameters according to the specific meteorological element type and numerical characteristics to generate specific transitional statements for the current scenario.

[0129] After all popular science knowledge units and transitional statements have been retrieved, the units are arranged sequentially according to the chronological order of meteorological evolution stages, with corresponding transitional statements inserted between adjacent units. Specifically, the arrangement strictly follows the chronological order of meteorological physical processes, with the unit corresponding to the first meteorological physical process placed first and the unit corresponding to the last process placed last. A transitional statement is inserted between every two adjacent units, connecting the conclusive explanation of the mechanism in the preceding unit with the initial description in the following unit, forming a causal narrative structure of "the preceding phenomenon leads to… which in turn triggers…". Through this arrangement and insertion process, multiple popular science knowledge units and transitional statements alternately combine to form a complete piece of popular science content with both chronological and causal logic.

[0130] The specific operation of the above process is illustrated using a typical cold front passage scenario as an example. The meteorological evolution stage sequence may sequentially include four stages: "warm and humid air mass control stage before the cold front," "cold front approach stage," "frontal precipitation stage," and "cold and dry air mass control stage after the cold front." After extracting the meteorological element types and change characteristics for each stage, a search and matching process is performed from the meteorological physical process knowledge base to obtain four meteorological physical processes: "warm and humid advection transport," "frontal lifting," "frontal precipitation formation," and "cold advection diffusion," forming a meteorological physical process sequence. By analyzing the output state and input conditions between adjacent processes, a direct causal relationship is determined between the output state of "warm and humid advection transport" (increased humidity in the lower atmosphere and enhanced instability) and the input conditions of "frontal lifting" (sufficient water vapor and lifting triggering mechanism). This process is repeated to determine the causal relationships between all adjacent processes. The system retrieves explanations of the mechanisms for each process and transitional statements between adjacent processes from a popular science knowledge base. Arranged chronologically, the generated popular science content first explains how warm, moist air masses provide the conditions for water vapor accumulation for precipitation. Then, through transitional statements, it introduces how a cold front forcibly lifts warm, moist air, explaining the physical mechanism by which this lifting and cooling leads to precipitation. Finally, through transitional statements, it connects to the description of how a cold, dry air mass replaces the warm, moist air mass after the cold front passes, forming a complete and logically coherent popular science narrative. This method of organizing popular science knowledge units based on the sequence of meteorological physical processes and causal relationships ensures that the popular science content delivered to users not only accurately reflects the current meteorological evolution but also helps users understand the intrinsic connections between various meteorological phenomena, enhancing the educational value of meteorological popular science and improving the user experience.

[0131] Determine whether the user's location is within an area where weather changes can be perceived. If the user's location is within an area where weather changes can be perceived, push science popularization content to the user's terminal, including:

[0132] Obtain the geographic coordinates of the user's location, calculate the shortest distance between the geographic coordinates of the user's location and the geographic boundary of the area where weather changes can be perceived, determine whether the shortest distance is less than a preset entry distance threshold, and determine that the user's location is in the area where weather changes can be perceived when the shortest distance is less than the preset entry distance threshold.

[0133] Extract the current meteorological evolution stage that the user's location will experience from the meteorological evolution stage sequence, select matching popular science knowledge units from popular science content based on the meteorological physical processes corresponding to the current meteorological evolution stage, and extract the mechanism explanations and transition sentences in the popular science knowledge units to combine them into popular science content to be pushed.

[0134] The urgency of the push is calculated based on the shortest distance, and a push message containing popular science content is generated based on the urgency. The push message is then sent to the user's terminal corresponding to the user's location.

[0135] The system acquires the geographic coordinates reported by the user terminal, expressed in latitude and longitude with a precision of at least five decimal places to ensure the accuracy of subsequent spatial determinations. After acquiring the coordinates, it calculates the spatial distance between them and the geographic boundary of the area susceptible to weather change. The geographic boundary of the area susceptible to weather change is formed by the outer contours of raster pixels that meet the criteria of exceeding the comprehensive spatiotemporal gradient threshold and having a sufficient connected domain area, and is stored as a polygon vector. The system calculates the perpendicular distance and endpoint distance from the user's geographic coordinates to all line segments on this polygon boundary, and takes the minimum value as the shortest distance. When the user's location is already inside the polygon, Taking a zero value directly determines that the user is in an area where weather changes are perceptible. Less than the preset entry distance threshold Similarly, if the user's location is determined to be within an area where weather changes can be perceived, the subsequent push notification process will be triggered; when Greater than or equal to In such cases, push notifications will not be triggered, but the user can be added to a periodic review queue to address situations where the area affected by weather changes dynamically expands.

[0136] The value can be configured differently according to the type of meteorological element. For precipitation-type meteorological events, because their boundaries move quickly and have a direct impact on users, The range is typically set between 500 and 2000 meters; for events involving sudden temperature drops, the impact spreads more slowly. This range can be appropriately reduced to 200 to 800 meters. In practical implementation, It can be read from the meteorological element perception threshold configuration table, forming a one-to-one correspondence with the meteorological element type, avoiding the imbalance of push frequency caused by using a uniform threshold for all meteorological types.

[0137] After confirming that the user's location is within a region of perceptible meteorological change, the meteorological evolution stage that best matches the current time and the user's location is identified from the meteorological evolution stage sequence and designated as the current meteorological evolution stage. The meteorological evolution stage sequence is a set of stages arranged chronologically, with each stage corresponding to a meteorological physical process description, such as "warm and humid air convergence stage before a cold front," "frontal passage precipitation stage," and "cold air mass control stage after a front." The current meteorological evolution stage is determined by combining the spatial relationship between the user's location and the boundary location with the direction of boundary movement to determine whether the user is currently in front of, passing through, or behind the boundary, and then corresponding to the specific stage number in the meteorological evolution stage sequence.

[0138] Based on the meteorological physical processes corresponding to the current meteorological evolution stage, matching popular science knowledge units are selected from the existing popular science content. Popular science content consists of multiple knowledge units, each containing a tag field. The tag field records the applicable meteorological physical process type, the applicable meteorological evolution stage number, and the content level (basic explanation layer, mechanism description layer, and extended layer). During selection, the meteorological physical process type of the current meteorological evolution stage is used as the primary key, and the content level as the secondary key. Knowledge units in the mechanism description layer are extracted first, followed by transitional statements related to the connection between previous and subsequent stages. The mechanism description section explains the physical causes of the current meteorological phenomenon, such as the process of water vapor condensation caused by frontal lifting when cold and warm air masses converge. The transitional statements connect the evolutionary logic between the current stage and the next stage, helping users understand the continuity of meteorological changes, such as the subsequent trend of rising air pressure and falling temperature after a front passes. The mechanism description and transitional statements are concatenated in the logical order of "phenomenon explanation—cause explanation—subsequent expectations" to form the popular science content text to be pushed.

[0139] The length of the science popularization content pushed out should be controlled between 150 and 400 characters to adapt to the display limitations of mobile device notification bars. Abbreviations for professional meteorological terms should be avoided in the text, and all professional terms should be accompanied by a popular explanation in parentheses when they first appear. For example, "front (the interface between cold and warm air)" and "pressure gradient force (the force that drives air from high pressure to low pressure)." This approach ensures that the science popularization content is readable for non-specialist users while retaining sufficient knowledge density.

[0140] Based on the shortest distance Calculate push urgency The push notification urgency level reflects how close a user is to the area experiencing weather changes, and is used to determine the priority and notification method of push messages. When When the value is zero (the user is already within the region), Take the highest value; when In zero to In between, Follow It increases and monotonically decreases; specifically, a linear mapping method can be used to... From zero to The interval mapping is The interval from 1 to 0, i.e. Urgency of push notifications Divided into three levels: high, medium, and low. Due to the high degree of urgency, As medium urgency, The urgency level is low.

[0141] Different urgency levels correspond to different push notification generation strategies. At high urgency, push notifications include strong alerts, vibrate or sound when displayed in the notification bar, the title directly states the type of weather change and its estimated arrival time, the body contains complete educational content, and includes behavioral suggestions at the end (e.g., "We recommend carrying rain gear" or "We recommend reducing outdoor activities"). At medium urgency, push notifications are sent as regular notifications, the title describes the weather trend, the body includes an explanation of the mechanism, and transitional sentences can be selectively omitted to control text length. At low urgency, push notifications are sent as silent notifications, only displayed when the user actively opens the application, without triggering a system notification pop-up, the content focuses on educational information, and the urgency description is downplayed.

[0142] After the push message is generated, it is sent to the target user terminal via the push channel based on the user terminal identifier corresponding to the user's location. The user terminal identifier is obtained through the binding relationship between the user account and the device registration information. The push channel supports concurrent pushes from multiple terminals. When the same user account is bound to multiple terminals, the push is prioritized to the terminal device that was most recently active. Push records are written to a log, which includes the user identifier, push time, push urgency level, a summary of the science popularization content pushed, and push result status. This log is used for subsequent push effect analysis and continuous collection of user weather perception feedback data.

[0143] A second aspect of this invention provides a dynamic science popularization push system based on real-time meteorological data, comprising:

[0144] The perception threshold unit is used to acquire historical meteorological data and user meteorological perception feedback data, calculate the correspondence between the changes in meteorological elements in historical meteorological data and the perception response in user meteorological perception feedback data, and determine the meteorological element perception threshold based on the correspondence.

[0145] The perceptible area unit is used to acquire meteorological data sequences and user locations for a target area. The meteorological data sequences include meteorological element observations from multiple spatial locations. The spatiotemporal gradient of the meteorological element observations is calculated, and the spatiotemporal gradient is compared with the meteorological element perception threshold to determine the perceptible meteorological change area.

[0146] The evolution stage unit is used to analyze the spatial distribution changes of meteorological element observations in meteorological data sequences, identify the boundary positions and movement directions of meteorological element observations, and calculate the meteorological evolution stage sequence that the user's position will experience based on the spatial relationship between the user's position and the boundary positions and movement directions.

[0147] The science popularization generation unit is used to map each meteorological evolution stage in the meteorological evolution stage sequence to meteorological physical processes, organize the science popularization knowledge units corresponding to the meteorological physical processes according to the temporal relationship of the meteorological evolution stage sequence, and generate science popularization content.

[0148] The sensing and push unit is used to determine whether the user's location is in an area where weather changes can be sensed. When the user's location is in an area where weather changes can be sensed, popular science content will be pushed to the user's terminal.

[0149] A third aspect of the present invention provides an electronic device, comprising:

[0150] processor;

[0151] Memory used to store processor-executable instructions;

[0152] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0153] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0154] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0155] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dynamic science popularization push method based on real-time meteorological data, characterized in that, include: Acquire historical meteorological data and user meteorological perception feedback data, calculate the correspondence between the changes in meteorological elements in historical meteorological data and the perception response in user meteorological perception feedback data, and determine the meteorological element perception threshold based on the correspondence. Acquire meteorological data sequences and user locations for the target area. The meteorological data sequences include meteorological element observations from multiple spatial locations. Calculate the spatiotemporal gradient of the meteorological element observations and compare the spatiotemporal gradient with the meteorological element perception threshold to determine the area of ​​perceptible meteorological changes. Analyze the spatial distribution changes of meteorological element observations in meteorological data sequences, identify the boundary locations and movement directions of meteorological element observations, and deduce the meteorological evolution stages that the user's location will experience based on the spatial relationship between the user's location and the boundary locations and movement directions. The meteorological evolution stages in the meteorological evolution stage sequence are mapped to meteorological physical processes. Popular science knowledge units corresponding to the meteorological physical processes are organized according to the temporal relationship of the meteorological evolution stage sequence to generate popular science content. Determine if the user's location is within an area where weather changes can be perceived. If the user's location is within an area where weather changes can be perceived, push science popularization content to the user's terminal.

2. The method according to claim 1, characterized in that, Calculate the correspondence between changes in meteorological elements in historical meteorological data and perceived responses in user meteorological perception feedback data, and determine the meteorological element perception thresholds based on this correspondence, including: Identify the time abrupt change points and spatial gradient jump points of meteorological element observation values ​​from historical meteorological data, extract the time variation amplitude of the time abrupt change points and the spatial variation amplitude of the spatial gradient jump points, and combine the time variation amplitude and spatial variation amplitude into the meteorological element variation quantity. Extract the perception response time from the user's meteorological perception feedback data, align the perception response time with the time series of historical meteorological data, and obtain the change in meteorological elements corresponding to the perception response time. The number of matching samples between the change in meteorological elements and the sensing response time is counted. The distribution curve of the number of matching samples with the change in meteorological elements is calculated. The inflection point of the distribution curve is identified. The matching relationship between the change in meteorological elements and the inflection point is used as the correspondence. The meteorological element sensing threshold is determined based on the change in meteorological elements corresponding to the inflection point in the correspondence. The user meteorological perception feedback data includes meteorological perception feedback information submitted by users and records of user interaction behavior.

3. The method according to claim 1, characterized in that, Calculate the spatiotemporal gradient of meteorological element observations, compare the spatiotemporal gradient with the meteorological element perception threshold, and determine the areas where meteorological changes can be perceived, including: The meteorological element observation values ​​at multiple spatial locations at consecutive times are extracted from the meteorological data sequence. The difference between the meteorological element observation values ​​at adjacent times is calculated to obtain the temporal gradient. The difference between the meteorological element observation values ​​at adjacent spatial locations is calculated to obtain the spatial gradient. The spatiotemporal gradient of the meteorological element observation values ​​is constructed based on the rate of change of the temporal gradient and the distribution density of the spatial gradient. The spatiotemporal gradient is compared with the meteorological element perception threshold on a spatial basis to identify the spatial location where the spatiotemporal gradient continuously exceeds the meteorological element perception threshold and record the duration of exceeding the threshold. Spatial locations where the duration of exceeding the threshold is greater than the preset duration are extracted to form a candidate change location set. Calculate the ratio of the temporal gradient to the spatial gradient magnitude of each spatial location in the candidate change location set, extract spatial locations whose ratio falls within a preset range as spatiotemporal coupling locations, extract connected components from the spatiotemporal coupling locations, and determine the area delineated by the connected components as the area of ​​perceptible meteorological changes.

4. The method according to claim 1, characterized in that, Analyzing the spatial distribution changes of meteorological element observations in meteorological data sequences, identifying the boundary locations and directions of movement of these observations, and based on the spatial relationship between the user's location and these boundary locations and directions of movement, the sequence of meteorological evolution stages that the user's location will experience is estimated, including: Spatial interpolation is performed on meteorological element observations at multiple spatial locations in a meteorological data sequence to generate a spatial distribution field. The spatial gradient of the meteorological element observations in the spatial distribution field is calculated, and the extreme value locations of the spatial gradient are identified. The extreme value locations are determined as the boundary locations formed by the meteorological element observations. Extract the spatial coordinates of the boundary position at consecutive time steps, and calculate the displacement vector of the spatial coordinates at adjacent time steps to determine the direction of movement of the boundary position; Based on the spatial coordinates of the boundary position and the direction of movement, the predicted coordinates of the boundary position at multiple future times are calculated by linear extrapolation. The spatial distance between the user position and the predicted coordinates is calculated, and the predicted coordinates and the corresponding arrival times of the spatial distances less than a preset distance threshold are identified. The spatial relationship between the user position and the boundary position and the direction of movement is established. Meteorological element observation values ​​are extracted from both sides of the boundary position corresponding to the arrival time from the spatial distribution field. The numerical difference between the meteorological element observation values ​​on both sides of the boundary position is calculated. The boundary position is divided into different meteorological evolution stages according to the numerical difference. The meteorological evolution stages are arranged in chronological order according to the arrival time. The sequence of meteorological evolution stages that the user's position will experience is calculated and generated.

5. The method according to claim 4, characterized in that, Based on the spatial coordinates of the boundary location and the direction of movement, linear extrapolation is used to calculate the predicted coordinates of the boundary location at multiple future times, including: Extract the displacement vector magnitude of the boundary position according to the direction of movement, calculate the ratio of the displacement vector magnitude to the time interval, and obtain the movement rate of the boundary position along the direction of movement. The displacement vector corresponding to the preset extrapolation time step is calculated based on the movement speed and direction. The current spatial coordinates of the boundary position are vector-superimposed with the displacement vector to obtain the predicted coordinates of the boundary position after an extrapolation time step. Multiple extrapolation time steps are set, and the displacement vector corresponding to each extrapolation time step is calculated sequentially according to the movement speed and movement direction. The current spatial coordinates of the boundary position are vector superimposed with each displacement vector, and the predicted coordinates of the boundary position at multiple future times are obtained by linear extrapolation calculation based on the spatial coordinates of the boundary position and the movement direction.

6. The method according to claim 1, characterized in that, The meteorological evolution stages in the meteorological evolution sequence are mapped to meteorological physical processes. Popular science knowledge units corresponding to these meteorological physical processes are organized according to the temporal relationship of the meteorological evolution stage sequence, generating popular science content including: Extract the types and characteristics of meteorological elements in each stage of the meteorological evolution sequence, and retrieve matching meteorological physical processes from a pre-set meteorological physical process knowledge base based on the types and characteristics of meteorological elements. Each meteorological evolution stage is mapped to a corresponding meteorological physical process to obtain a sequence of meteorological physical processes; Extract the output state and input conditions of adjacent meteorological physical processes in a meteorological physical process sequence; The causal relationship between adjacent meteorological and physical processes is determined based on the correspondence between output states and input conditions. Based on the meteorological and physical processes and the causal relationship, the corresponding popular science knowledge units are retrieved from the preset popular science knowledge base. The popular science knowledge units include the mechanism explanation of the meteorological and physical processes and the transitional statements of the causal relationship. The popular science knowledge units are arranged according to the temporal relationship of the meteorological evolution stage sequence. Transitional statements are inserted between adjacent popular science knowledge units. The transitional statements connect the mechanism descriptions in the preceding popular science knowledge units with the mechanism descriptions in the subsequent popular science knowledge units, organize the popular science knowledge units corresponding to meteorological physical processes, and generate popular science content.

7. The method according to claim 1, characterized in that, Determine whether the user's location is within an area where weather changes can be perceived. If the user's location is within an area where weather changes can be perceived, push science popularization content to the user's terminal, including: Obtain the geographic coordinates of the user's location, calculate the shortest distance between the geographic coordinates of the user's location and the geographic boundary of the area where weather changes can be perceived, determine whether the shortest distance is less than a preset entry distance threshold, and determine that the user's location is in the area where weather changes can be perceived when the shortest distance is less than the preset entry distance threshold. Extract the current meteorological evolution stage that the user's location will experience from the meteorological evolution stage sequence, select matching popular science knowledge units from popular science content based on the meteorological physical processes corresponding to the current meteorological evolution stage, and extract the mechanism explanations and transition sentences in the popular science knowledge units to combine them into popular science content for push. The urgency of the push is calculated based on the shortest distance, and a push message containing popular science content is generated based on the urgency. The push message is then sent to the user's terminal corresponding to the user's location.

8. A dynamic science popularization push system based on real-time meteorological data, used to implement the method as described in any one of claims 1-7, characterized in that, include: The perception threshold unit is used to acquire historical meteorological data and user meteorological perception feedback data, calculate the correspondence between the changes in meteorological elements in historical meteorological data and the perception response in user meteorological perception feedback data, and determine the meteorological element perception threshold based on the correspondence. The perceptible area unit is used to acquire meteorological data sequences and user locations for a target area. The meteorological data sequences include meteorological element observations from multiple spatial locations. The spatiotemporal gradient of the meteorological element observations is calculated, and the spatiotemporal gradient is compared with the meteorological element perception threshold to determine the perceptible meteorological change area. The evolution stage unit is used to analyze the spatial distribution changes of meteorological element observations in meteorological data sequences, identify the boundary positions and movement directions of meteorological element observations, and calculate the meteorological evolution stage sequence that the user's position will experience based on the spatial relationship between the user's position and the boundary positions and movement directions. The science popularization generation unit is used to map each meteorological evolution stage in the meteorological evolution stage sequence to meteorological physical processes, organize the science popularization knowledge units corresponding to the meteorological physical processes according to the temporal relationship of the meteorological evolution stage sequence, and generate science popularization content. The sensing and push unit is used to determine whether the user's location is in an area where weather changes can be sensed. When the user's location is in an area where weather changes can be sensed, popular science content will be pushed to the user's terminal.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images