A method and system for predicting the flowering period of shrubs based on smart grid data

Abstract

The application discloses a shrub flowering period prediction method based on intelligent grid data, which comprises the following steps: step one, data preparation and preprocessing; step two, key meteorological factor set construction; step three, prediction factor screening and model establishment; and step four, flowering period prediction and product generation. The prediction system comprises a data acquisition and processing module, an accumulated temperature and factor calculation module, a flowering period prediction module, a prediction execution and product production module and a service interface module. The seamless grid data is used to significantly improve the spatial representativeness and accuracy of the flowering period prediction under complex terrain, the limitation of a single empirical model is overcome by establishing a special model library for different species and different phenological stages, the universality and scalability of the method are enhanced, and the model construction fully integrates ecological cognition such as the differentiated phenological driving mechanism of fast and slow flowering shrubs, so that the prediction has a solid scientific basis.

Description

Technical Field

[0001] This invention relates to the field of ecological meteorological forecasting and information technology, and in particular to a method and system for forecasting shrub flowering time based on intelligent grid data. Background Technology

[0002] Plant phenology is one of the most sensitive indicators of ecosystems to climate change. As an important vegetation type in arid and semi-arid regions, the forecasting of shrub flowering time is of great significance for guiding ecological restoration, agricultural production, tourism activities, and studying climate change response mechanisms.

[0003] Currently, flowering forecasts largely rely on observational data from fixed meteorological stations and statistical analysis of historical phenological data. For example, existing research, through long-term observation of several typical shrubs in specific regions, has revealed that temperature is a key factor dominating shrub phenology and established statistical prediction models based on meteorological station data. However, these methods have obvious limitations: First, the meteorological station data they rely on are spatially distributed in a point-like manner, making it difficult to accurately reflect the continuous distribution of meteorological elements in areas with complex terrain, resulting in low spatial resolution and insufficient representativeness in forecasts; second, the models built are mostly empirical statistical models for single species and specific locations, and the model factors and parameters are difficult to directly transfer to other species or regions, resulting in poor universality; finally, traditional methods have a low degree of automation, making it difficult to meet the needs of large-scale, high-frequency, and operational flowering forecasts.

[0004] Intelligent grid meteorological data is a type of meteorological analysis and forecast field data with high spatiotemporal resolution and seamless coverage generated by fusing and assimilating multi-source data (satellite, radar, station, numerical model, etc.). Applying this type of data to ecological models can realize the spatial continuity and refinement of meteorological driving fields, and can provide a new technical approach for flowering forecasting methods. Therefore, this invention proposes a shrub flowering forecasting method and system based on intelligent grid data to solve the problems existing in the prior art. Summary of the Invention

[0005] To address the aforementioned problems, the present invention aims to propose a method and system for forecasting shrub flowering time based on intelligent grid data. This method and system, by combining high spatiotemporal resolution intelligent grid meteorological data with differentiated shrub phenological prediction models, can achieve high-precision, spatially continuous, and automated flowering time forecasting.

[0006] To achieve the objectives of this invention, the invention is implemented through the following technical solution: a method for predicting the flowering period of shrubs based on intelligent grid data, comprising the following steps: Step 1: Acquire smart grid meteorological data for the target area and historical phenological observation data for the target shrub species; Step 2: Based on intelligent grid meteorological data, calculate multiple key meteorological factors related to shrub phenology; Step 3: Based on key meteorological factors and historical phenological observation data, construct flowering period forecast models for different shrub species and different phenological stages using statistical regression methods; Step 4: Based on real-time and forecast intelligent grid meteorological data, drive the flowering period forecast model to generate regionalized shrub flowering period forecast products.

[0007] The further improvement is that the key meteorological factors in step two include multiple threshold effective accumulated temperatures from the base date to the phenological occurrence date, the average temperature during and before the phenological occurrence period, the highest or lowest temperature during and before the phenological occurrence period, and the precipitation during and before the phenological occurrence period; among which multiple threshold effective accumulated temperatures include effective accumulated temperatures >0℃, effective accumulated temperatures >5℃, or effective accumulated temperatures >10℃.

[0008] Further improvements are made in the following aspects: the shrub phenological periods in step three include the greening-up period, the flowering period, and the yellowing-and-withering period; the flowering period forecast model is a differentiated forecast model constructed for the greening-up period, the flowering period, and the yellowing-and-withering period respectively; when constructing the flowering period forecast model, the key meteorological factors selected include the effective accumulated temperature >5℃ and the average minimum temperature during the flowering period.

[0009] Further improvements are made in that: when constructing the flowering period forecast model, differentiated key factor combinations are used for shrub species with different reproductive strategies. For fast-flowering shrubs with a short interval between greening and flowering, the key factor combination focuses more on the short-term accumulated temperature after greening; for slow-flowering shrubs with a long interval between greening and flowering, the key factor combination focuses more on the heat accumulation in the early growing season.

[0010] The further improvement lies in the fact that the statistical regression method in step three is the stepwise regression method, which is specifically used to screen out the subset of forecast factors that contribute the most to the interpretation of phenological periods from key meteorological factors and establish a multiple linear regression model.

[0011] A further improvement is that the flowering forecast product in step four is a spatially continuous grid map, and the value of each grid cell represents the predicted phenological date of the target shrub species at that geographical location.

[0012] A forecasting system for shrub flowering period forecasting based on intelligent grid data includes a data acquisition and processing module, an accumulated temperature and factor calculation module, a flowering period prediction module, and a forecast execution and product production module. The data acquisition and processing module is used to acquire and preprocess intelligent grid meteorological data and historical phenological data of shrubs. The accumulated temperature and factor calculation module calculates multiple effective accumulated temperature sequences in parallel based on the accumulated temperature and factor calculation engine and extracts key meteorological factors. The flowering period prediction module stores and manages multiple regression prediction models established for different shrub species and different phenological stages based on a flowering period prediction model library. The forecast execution and product production module is used to call the models for calculation and generate thematic maps of flowering period forecasts.

[0013] Further improvements are made in that: the data acquisition and processing module specifically converts the phenological observation dates into Julian day format and matches and aligns the meteorological data with the phenological data in time and space; The accumulated temperature and factor calculation engine is specifically based on intelligent grid meteorological data. It calculates multiple effective accumulated temperature sequences for different shrub species and different temperature thresholds at each grid point in parallel, and extracts meteorological elements for specific time periods as key meteorological factors.

[0014] A further improvement is that the model input in the flowering period prediction model library is a screened key meteorological factor, and its output is the predicted Julian Day of the phenological period; The forecast execution and product creation module specifically calls the accumulated temperature and factor calculation engine and the flowering period prediction model library, and then drives the prediction model to perform calculations based on real-time and forecast intelligent grid meteorological data, and generates a gridded thematic map of flowering period forecast.

[0015] Further improvements include a service interface module, which receives query requests from users specifying forecast areas, shrub types, and forecast durations, and returns the results generated by the forecast execution and product creation module to the user.

[0016] The beneficial effects of this invention are as follows: This invention significantly improves the spatial representativeness and accuracy of flowering forecast under complex terrain by utilizing seamless grid data. By establishing a dedicated model library for different species and different phenological stages, it overcomes the limitations of a single empirical model, enhances the universality and scalability of the method, and fully integrates ecological knowledge such as the differentiated phenological driving mechanisms of fast- and slow-flowering shrubs into the model construction, giving the forecast a solid scientific basis. The entire process of this invention can achieve automated operation from data acquisition to product generation, providing dynamic and timely decision support for large-scale ecological restoration, tourism planning, and other projects. Attached Figure Description

[0017] Figure 1This is a flowchart of the forecasting method of the present invention.

[0018] Figure 2 This is a diagram of the forecasting system architecture of the present invention. Detailed Implementation

[0019] To enhance understanding of the present invention, the embodiments of the present invention will be further described in detail below with the example of shrub flowering period forecasting in northwestern Loess Plateau. This embodiment is only used to explain the present invention and does not constitute a limitation on the scope of protection of the present invention.

[0020] Example 1 according to Figure 1 As shown in the figure, this embodiment takes the flowering period forecast of Caragana narrow-leaved as an example to illustrate a method for forecasting the flowering period of shrubs based on intelligent grid data.

[0021] Step 1: Data Preparation and Preprocessing Acquiring intelligent grid meteorological data: Obtain intelligent grid data products for the target area (106°42′~111°27′E, 37°35′~40°51′N) from national meteorological operational centers (such as the National Meteorological Information Center), including: Historical reanalysis data (2004-2024): used for model training. The data is a daily grid field with a spatial resolution of 1km×1km. The elements include daily average temperature, daily maximum temperature, daily minimum temperature and daily precipitation.

[0022] Real-time analysis field data: used to calculate current accumulated temperature and meteorological conditions; updated daily, providing real-time meteorological grid field data up to yesterday.

[0023] Smart grid forecast data: used to drive forecast models; to obtain gridded forecast fields of temperature and precipitation for the next 15 days.

[0024] Historical phenological data were acquired and processed: Phenological observation records of *Caragana angustifolia* from 2004 to 2024 were obtained from 8 ecological monitoring stations in the study area, with key phenological periods being the greening-up period, flowering period, and yellowing-withering period; the specific dates (month-day) of the records were converted to Julian days (January 1st of each year); the data from 2004 to 2022 were used as the training set for modeling, and the data from 2023 to 2024 were used as an independent validation set to test the accuracy of the model.

[0025] Step 2: Construction of the Key Meteorological Factor Set (Taking the Flowering Period as an Example) For each year and the geographical location of each ecological monitoring station (which can be mapped to the nearest grid point), based on the historical smart grid temperature data of that point, calculate the various effective accumulated temperatures from January 1 of that year to its historical flowering day (Julian Day).

[0026] Accumulated temperature: ΣN i=1T i T i Where N is the daily average temperature, and N is the number of days from January 1st to the flowering day; effective accumulated temperature > 0℃: ΣN i=1max(T i -0,0); >5℃ effective accumulated temperature: ΣN i=1max(T i -5,0); >10℃ effective accumulated temperature: ΣN i=1max(T i -10,0).

[0027] Simultaneously, short-term meteorological factors related to the flowering period were extracted: the average temperature and the average minimum temperature during the flowering period, and the total precipitation in the month before flowering.

[0028] Step 3: Predictor Factor Screening and Model Establishment 1. Correlation analysis and initial screening of factors: Perform correlation analysis on all meteorological factors calculated in step 2 with the flowering period, and screen out factors that pass the significance test (p<0.05), such as: accumulated temperature >5℃ (r=0.946), accumulated temperature >0℃ (r=0.929), average minimum temperature during the flowering period (r=0.842), monthly precipitation (r=0.566), etc.

[0029] 2. Stepwise Regression Modeling: Using the Julian date of flowering as the dependent variable (Y) and the factors screened above as independent variables, a predictive model is constructed using the stepwise regression method. The modeling process automatically removes factors with strong collinearity or insignificant contributions. The final model form is as follows: Greening period: y = 0.16·X1 - 0.12·X2 + 0.52·X3 + 0.38·X4 + 0.84·X5 + 70.06 Flowering period: y = 0.01·X6 - 0.03·X7 + 0.09·X8 ​​- 0.03·X9 + 1.32·X 10 +97.09 Yellowing and withering period: y = -2.58·X 11 +2.31·X 12 -1.62·X 13 -1.32·X 14 +326.11 Where X1 and X7 represent the accumulated temperature >0℃ during the greening and flowering periods, X2 and X8 represent the accumulated temperature >5℃ during the greening and flowering periods, and X3 and X 11 This represents the average temperature during the ten-day period of greening and withering, X4, X 10 and X 12 X5 represents the average minimum temperature of the ten-day period during the greening, flowering, and yellowing / withering stages; X6 represents the accumulated temperature during the flowering stage; X9 represents the accumulated temperature above 10℃ during the flowering stage; X... 13X represents the average temperature of the ten-day period preceding the dry season. 14 This represents the average minimum temperature of the previous ten-day period.

[0030] 3. Model Validation: The model accuracy was verified using an independent validation set (2023-2024). The mean absolute error (MAE), root mean square error (RMSE), and prediction accuracy were calculated between the predicted flowering date and the actual observed date. The results showed that the model's mean absolute error in predicting the flowering period could be controlled within 3 days, and the accuracy (error ≤ 3 days) exceeded 70%, proving the model's reliability.

[0031] 4. Extend to other species and phenological stages: Repeat steps two and three to establish models for the greening-up and yellowing-downing stages of *Caragana narrow-leavedis*, as well as for the phenological stages of five other shrub species (*Artemisia argyi*, *Caragana sinica*, etc.): For fast-flowering species (such as Caragana dwarf), the flowering period model may depend more on the short-term (such as 15-30 days) accumulated temperature after greening up; For slow-flowering species (such as shrubs like Artemisia argyi and Caragana korshinskii), the flowering period model emphasizes the long-term (3.5-5.5 months) heat accumulation from greening to flowering, and may incorporate spring precipitation as a moderating factor.

[0032] Step 4: Flowering Period Forecast and Product Generation Real-time data driven: Assuming the current date is April 1, 2025; the system calls the intelligent grid to analyze field data in real time and calculate factors such as the accumulated effective temperature of >5℃ at each grid point in the entire region from January 1 to April 1, 2025.

[0033] Combining forecast data: The system reads the intelligent grid temperature forecast for the next 15 days; it adds the accumulated temperature from the actual situation to the future forecast temperature to estimate the accumulated temperature value that may be reached on a future day (such as April 20).

[0034] Model calculation and spatial forecasting: The estimated accumulated temperature >5℃ and the predicted minimum temperature during the flowering period of each grid point are input into the flowering period model of *Caragana angustifolia*. The model outputs a predicted Julian date for flowering for each grid point. The system renders the results of all grid points in space to generate a forecast distribution map of the flowering period of *Caragana angustifolia* in northwestern Loess Plateau. Different colors can be used to represent the early or late flowering period. For example, red areas predict flowering before May 10, green areas predict flowering between May 10 and 20, and blue areas predict flowering after May 20.

[0035] Product Output: This forecast map can be provided to ecological management departments, tourism departments, or research institutions in image or GIS data format via a service interface.

[0036] Example 2 according to Figure 2 As shown, this embodiment provides a forecasting system for shrub flowering period forecasting based on intelligent grid data. The modules work together as follows: Data acquisition and processing module: Automatically acquires historical, current and forecast data from the intelligent grid at regular intervals from the meteorological data API, and decodes it into a standard spatiotemporal grid data format; reads shrub phenological records from the phenological observation database, performs quality control and Julian day conversion; spatially matches the geographical locations of phenological observation points with the intelligent grid, and extracts the meteorological sequences of the corresponding grid points.

[0037] Accumulated temperature and factor calculation module: After receiving meteorological grid data and time parameters, the accumulated temperature and factor calculation engine calculates multiple effective accumulated temperature sequences in parallel for each preset shrub species (such as Caragana axillaris, Artemisia argyi, etc.). For example, it calculates the diurnal variation grid field of effective accumulated temperature >0℃, >5℃, and >10℃ accumulated from January 1st for the entire region.

[0038] Based on preset time windows such as "ten-day period" and "month", the system quickly calculates short-term factor fields such as "average temperature of the current ten-day period" and "precipitation of the previous month" for each grid point.

[0039] Flowering period prediction module: Implemented through a flowering period prediction model library, which is a database that stores various prediction rules. Each rule corresponds to a model, including flowering period model, greening period model, yellowing and withering period model, and models for other species. The system calls the corresponding model according to the user's request.

[0040] Forecast Execution and Product Production Module: Upon receiving a forecast task instruction (such as "Forecast the flowering period of *Caragana korshinskii* in the next 15 days"), this module instructs the accumulated temperature and factor calculation engine to calculate the required factors (accumulated temperature >5℃ for *Caragana korshinskii*, etc.); it calls the "*Caragana korshinskii*-flowering period" model from the flowering period prediction model library; it substitutes the calculated gridded factor data into the model formula, performs full-network calculation, and obtains the predicted Julian day gridded field; it converts the Julian day field into calendar dates, performs visualization rendering, adds map elements such as legends and titles, and generates the final forecast thematic map.

[0041] Service Interface Module: Provides a Web API or graphical web interface, allowing users to select the forecast area (such as selecting administrative boundaries or drawing polygons), target shrub species (multiple selections are possible), target phenological period (greening / flowering / yellowing and withering), and forecast lead time via drop-down menus.

[0042] After the user submits a request, the interface passes the parameters to the forecast execution module and finally returns the generated forecast map or data file to the user.

[0043] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for predicting shrub flowering time based on intelligent grid data, characterized in that, Includes the following steps: Step 1: Acquire smart grid meteorological data for the target area and historical phenological observation data for the target shrub species; Step 2: Based on intelligent grid meteorological data, calculate multiple key meteorological factors related to shrub phenology; Step 3: Based on key meteorological factors and historical phenological observation data, construct flowering period forecast models for different shrub species and different phenological stages using statistical regression methods; Step 4: Based on real-time and forecast intelligent grid meteorological data, drive the flowering period forecast model to generate regionalized shrub flowering period forecast products.

2. The shrub flowering period forecasting method based on intelligent grid data according to claim 1, characterized in that: The key meteorological factors in step two include multiple threshold effective accumulated temperatures from the base date to the phenological occurrence date, the average temperature during and before the phenological occurrence period, the highest or lowest temperature during and before the phenological occurrence period, and the precipitation during and before the phenological occurrence period; among which multiple threshold effective accumulated temperatures include effective accumulated temperatures >0℃, effective accumulated temperatures >5℃, or effective accumulated temperatures >10℃.

3. The shrub flowering period forecasting method based on intelligent grid data according to claim 1, characterized in that: The shrub phenological periods in step three include the greening-up period, flowering period, and yellowing-and-withering period; the flowering period forecast model is a differentiated forecast model constructed for the greening-up period, flowering period, and yellowing-and-withering period respectively; when constructing the flowering period forecast model, the key meteorological factors selected include the effective accumulated temperature >5℃ and the average minimum temperature during the flowering period.

4. The shrub flowering period forecasting method based on intelligent grid data according to claim 3, characterized in that: When constructing the flowering forecast model, differentiated key factor combinations were used for shrub species with different reproductive strategies. For fast-flowering shrubs with a short interval between greening and flowering, the key factor combination focused more on the short-term accumulated temperature after greening. For slow-flowering shrubs with a long interval between greening and flowering, the key factor combination focused more on the heat accumulation in the early growing season.

5. The shrub flowering period forecasting method based on intelligent grid data according to claim 1, characterized in that: The statistical regression method in step three is stepwise regression, which is used to select the subset of forecast factors that contribute the most to the interpretation of phenological periods from key meteorological factors and establish a multiple linear regression model.

6. The shrub flowering period forecasting method based on intelligent grid data according to claim 1, characterized in that: In step four, the flowering forecast product is a spatially continuous grid map, where the value of each grid cell represents the predicted phenological date of the target shrub species at that geographic location.

7. A forecasting system for a shrub flowering period forecasting method based on intelligent grid data according to any one of claims 1-6, characterized in that: The system includes a data acquisition and processing module, an accumulated temperature and factor calculation module, a flowering period prediction module, and a forecast execution and product creation module. The data acquisition and processing module is used to acquire and preprocess intelligent grid meteorological data and historical phenological data of shrubs. The accumulated temperature and factor calculation module calculates multiple effective accumulated temperature sequences in parallel based on the accumulated temperature and factor calculation engine and extracts key meteorological factors. The flowering period prediction module stores and manages multiple regression prediction models established for different shrub species and different phenological stages based on the flowering period prediction model library. The forecast execution and product creation module is used to call the models for calculation and generate thematic maps of flowering period forecasts.

8. The forecasting system for a shrub flowering period forecasting method based on intelligent grid data according to claim 7, characterized in that: The data acquisition and processing module specifically processes the data by converting the phenological observation dates into Julian day format and matching and aligning the meteorological data with the phenological data in time and space. The accumulated temperature and factor calculation engine is specifically based on intelligent grid meteorological data. It calculates multiple effective accumulated temperature sequences for different shrub species and different temperature thresholds at each grid point in parallel, and extracts meteorological elements for specific time periods as key meteorological factors.

9. The forecasting system for a shrub flowering period forecasting method based on intelligent grid data according to claim 7, characterized in that: The model inputs in the flowering period prediction model library are key meteorological factors that have been screened, and the output is the predicted Julian Day of the phenological period. The forecast execution and product creation module specifically calls the accumulated temperature and factor calculation engine and the flowering period prediction model library, and then drives the prediction model to perform calculations based on real-time and forecast intelligent grid meteorological data, and generates a gridded thematic map of flowering period forecast.

10. The forecasting system for a shrub flowering period forecasting method based on intelligent grid data according to claim 7, characterized in that: It also includes a service interface module, which is used to receive query requests from users for forecast areas, shrub species, and forecast time, and return the results generated by the forecast execution and product production module to the user.

Images