Fine agriculture suitability index prediction method and system

Abstract

This application provides a refined agricultural suitability index forecasting method and system. The method includes: uniformly interpolating multi-source heterogeneous meteorological data into a pre-constructed regular grid network to form initial gridded data; fusing station-observed precipitation and radar-monitored precipitation from the initial gridded data to obtain a precipitation analysis field, and fusing WRF model data and multi-source meteorological observation data to obtain a meteorological environment analysis field; using a recursive extrapolation algorithm and the WRF model to perform time-series extrapolation of the precipitation analysis field and the meteorological environment analysis field to generate a refined meteorological gridded data sequence; and inputting the refined meteorological gridded data sequence into a meteorological suitability model for adaptive calculation to obtain a gridded agricultural operational suitability index. This achieves refined forecasting of local agricultural meteorological environment and suitability with farmland-scale matching at the 100-meter grid level, improving the refinement and accuracy of gridded forecasts for farmland meteorological environment and suitability index.

Description

Technical Field

[0001] This application belongs to the field of agricultural data technology and relates to a method and system for predicting the precision agricultural suitability index. Background Technology

[0002] In modern precision agriculture, achieving refined forecasting of meteorological elements at the field level is a core requirement for supporting efficient and scientific agricultural decision-making. The urgency of this requirement stems primarily from the significant spatial variations in key meteorological elements such as temperature, precipitation, wind speed and direction, and relative humidity among different fields within the same agricultural region, strongly influenced by factors such as complex topography, local water body distribution, and underlying surface type. These microscale climate differences have a decisive impact on agricultural production activities. Furthermore, the key phenological periods of specific crops are extremely sensitive to changes in environmental parameters such as temperature and humidity, and their suitability directly affects yield and quality. Simultaneously, core agricultural operations such as irrigation scheduling, precision fertilization, pest and disease control, and determination of the optimal harvest window all heavily rely on accurate predictions of meteorological conditions for specific field locations within the next few hours to days. However, current weather forecasting technologies for agricultural applications are significantly insufficient in meeting the needs for refined and quantitative forecasting at the field level. The main reasons are as follows:

[0003] First, key data sources required for building refined forecasting models, such as meteorological observation data, agricultural soil moisture / growth data, and high-resolution land cover / use and soil data, suffer from significant barriers in format, standardization, and sharing. This data heterogeneity and access barriers severely hinder the effective construction and training of quantitative, integrated forecasting models that integrate multidisciplinary knowledge and are geared towards agricultural activity suitability. Second, existing operational agricultural meteorological forecasts generally rely directly on kilometer-level resolution numerical weather prediction products as input, such as national-level smart grid forecasts at approximately 5 kilometers, and regional scale models such as CMA-MESO and ZJWARRS with resolutions of 1-3 kilometers. This level of spatial resolution is far from sufficient to resolve microscale climate characteristics caused by complex topography, small water bodies, and local surface heterogeneity. Forecast results exhibit an over-smoothing effect spatially, failing to accurately reflect the actual meteorological gradients and differences between adjacent fields, resulting in excessively large representativeness errors in forecast locations. Furthermore, because the spatial resolution of the input forecast data does not match the field-level requirements for agricultural decision-making, the inherent forecast errors of the initial weather forecast models are propagated across scales and significantly amplified by the nonlinear operations of downstream models or empirical methods when macro-forecast information is downscaled for application to micro-field-level suitability assessments and agricultural guidance. This error amplification effect directly leads to a significant reduction in the reliability and confidence of the final output of key agricultural activity guidance information such as irrigation water demand, fertilization window, suitable conditions for pesticide application, and harvesting timing. This makes it difficult to meet the stringent requirements of precision agriculture for decision support accuracy and may even lead to resource misallocation or production losses. Summary of the Invention

[0004] The purpose of this application is to provide a refined agricultural suitability index forecasting method and system to address the problems in existing agricultural suitability forecasting technologies. These technologies rely on kilometer-level numerical forecast products with insufficient spatial resolution to resolve microscale climate characteristics caused by complex terrain and other factors. Furthermore, the spatial resolution of the input forecast data does not match the field-level resolution required for agricultural decision-making, resulting in a significant reduction in the reliability and confidence of key agricultural activity guidance information and making it difficult to meet the stringent requirements of precision agriculture for decision support accuracy.

[0005] In a first aspect, this application provides a method for predicting the suitability index of refined agriculture, the method comprising:

[0006] Acquire multi-source heterogeneous meteorological data of the target area, and uniformly interpolate the multi-source heterogeneous meteorological data into a pre-constructed regular grid network to form initial gridded data;

[0007] The precipitation analysis field is obtained by fusing the station observation precipitation and radar monitoring precipitation in the initial gridded data, and the meteorological environment analysis field is obtained by fusing WRF model data and multi-source meteorological observation data.

[0008] The precipitation analysis field and the meteorological environment analysis field are temporally extrapolated using a recursive extrapolation algorithm and a WRF model to generate a refined meteorological grid data sequence with the same resolution as the regular grid network for a future preset time period.

[0009] The refined meteorological gridded data sequence is input into the meteorological suitability model for adaptive calculation to obtain the gridded agricultural operation suitability index.

[0010] In one embodiment, the multi-source heterogeneous meteorological data includes at least the most recently acquired radar observation data, WRF numerical forecast data, automatic weather station data, agricultural micro-station data, and static geographic information data.

[0011] In one embodiment, before uniformly interpolating the multi-source heterogeneous meteorological data onto a pre-constructed regular grid network, the method further includes:

[0012] The acquired multi-source heterogeneous meteorological data are subjected to climate extreme value checks, spatial consistency checks, temporal consistency checks, and multi-source consistency checks.

[0013] The climate extreme value check includes: acquiring historical monitoring data of the target station, calculating the 0.1 percentile and 99.9 percentile of the historical monitoring data according to meteorological elements, setting the monitoring data corresponding to the 0.1 percentile as a first dynamic threshold, and setting the monitoring data corresponding to the 99.9 percentile as a second dynamic threshold; when the real-time monitored multi-source heterogeneous meteorological data is less than the first dynamic threshold or greater than the second dynamic threshold, the multi-source heterogeneous meteorological data is marked as a missing value;

[0014] The spatial consistency check includes: generating a continuous field across the entire region using bilinear interpolation based on all multi-source heterogeneous meteorological data from the target station; calculating the deviation between the multi-source heterogeneous meteorological data of each target station and the expected location value of the corresponding target station; and marking the multi-source heterogeneous meteorological data of the corresponding target station as missing values ​​when the deviation is greater than a preset spatial deviation threshold.

[0015] The time consistency check includes: calculating the time variability based on the multi-source heterogeneous meteorological data monitored at the current time, the multi-source heterogeneous meteorological data monitored at the previous time, and the multi-source heterogeneous meteorological data monitored at the next time; when the time variability is greater than a preset variability threshold, the multi-source heterogeneous meteorological data at the corresponding time is marked as a missing value.

[0016] The multi-source consistency check includes: comparing the absolute differences between the multi-source heterogeneous meteorological data; taking the maximum value of the absolute differences; when the maximum value is greater than a preset multi-source consistency threshold, it is determined that there is a significant conflict between the multi-source heterogeneous meteorological data, and a predefined error handling protocol is triggered.

[0017] In one embodiment, the regular grid is a regular grid in a spatial coordinate system; wherein the resolution of the regular grid is 30 meters to 500 meters.

[0018] In one embodiment, fusing the precipitation data observed at stations and the precipitation data monitored by radar in the grid network to obtain a precipitation analysis field includes:

[0019] The precipitation values ​​of the m nearest grid points of the target station are set to be consistent with the measured values ​​of the target station, and the precipitation in the non-target station area is estimated by radar.

[0020] The precipitation analysis field is obtained by fusing the observed precipitation data of stations in the grid network where the target station is located with the estimated precipitation data.

[0021] In one embodiment, the step of fusing WRF model data and multi-source meteorological observation data to obtain a meteorological environment analysis field includes:

[0022] The WRF forecast field is interpolated into the grid network using a three-dimensional interpolation method to obtain background field grid data.

[0023] Calculate the difference between the monitoring data of each target site and the background field grid data of the corresponding target site;

[0024] A difference field is generated by fusing the differences between n grid points surrounding each target site using an inverse distance weighting algorithm.

[0025] The difference field and the background field grid data are fused to obtain the meteorological environment analysis field.

[0026] In one embodiment, the step of using a recursive extrapolation algorithm and a WRF model to perform time-series extrapolation on the precipitation analysis field and the meteorological environment analysis field to generate a refined meteorological grid data sequence with the same resolution as the grid network for a future preset time period includes:

[0027] For 0-2 hour forecasts, the moving vector of precipitation echoes is tracked based on the cross-correlation algorithm to generate a precipitation extrapolation forecast field with the same resolution as the regular grid network for the next 0-2 hours.

[0028] For 2-6 hour forecasts, the precipitation extrapolation forecast field and the WRF model downscaling forecast field are dynamically weighted and fused, wherein the fusion weight function satisfies: ,in, To provide a forecast timeframe, The unit is hours. The weights for extrapolated forecast fields within 2 to 6 hours;

[0029] For forecasts longer than 6 hours, the WRF model is used to downscale the forecast results to the regular grid network.

[0030] In one embodiment, the expression for calculating the meteorological environment analysis field using a recursive extrapolation algorithm and a WRF model is as follows: ,in, For the first Forecast values ​​of elements at any given time, For the first WRF numerical prediction value at time 10:00. For the first Forecast values ​​of elements at any given time, For the first WRF numerical forecast value at time 10:00.

[0031] In one embodiment, the meteorological suitability model includes a wheat harvesting period suitability model, a bayberry picking model, a bayberry fertilization model, a rice heat damage model, and a Ganoderma lucidum mycelium growth and development period model.

[0032] Secondly, this application provides a precision agriculture suitability index forecasting system, the system comprising:

[0033] The gridded data generation module is configured to acquire multi-source heterogeneous meteorological data of the target area and interpolate the multi-source heterogeneous meteorological data into a pre-constructed regular grid network to form initial gridded data.

[0034] The data fusion module is configured to fuse the station-observed precipitation and radar-monitored precipitation in the initial gridded data to obtain a precipitation analysis field, and to fuse WRF model data and multi-source meteorological observation data to obtain a meteorological environment analysis field.

[0035] The short-term extrapolation module is configured to use a recursive extrapolation algorithm and WRF mode to perform time-series extrapolation on the precipitation analysis field and the meteorological environment analysis field, and generate a refined meteorological grid data sequence with the same resolution as the regular grid network for a future preset time period.

[0036] The index generation module is configured to input the refined meteorological grid data sequence into the meteorological suitability model for adaptive calculation to obtain the gridded agricultural operation suitability index.

[0037] Thirdly, this application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method.

[0038] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0039] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0040] As described above, the precision agriculture suitability index forecasting method, system, and equipment described in this application have the following beneficial effects:

[0041] By fully leveraging the value of multi-source heterogeneous meteorological data and breaking down data barriers, a multi-dimensional and high-precision perception of the farmland meteorological environment is achieved. This involves fusing multi-source heterogeneous meteorological data, obtaining a precipitation analysis field by integrating station-observed precipitation and radar-monitored precipitation in a gridded network, and obtaining a meteorological environment analysis field by integrating WRF model data and multi-source meteorological observation data. Recursive extrapolation techniques and the WRF model are then used to calculate the precipitation analysis field and the meteorological environment analysis field, resulting in a refined meteorological gridded forecast data sequence. This scales down kilometer-level forecasts to tens to hundreds of meters, enabling local agricultural meteorological environment forecasts with hundreds of-meter-level grid matching for farmland scale. This improves the accuracy and refinement of farmland gridded forecasts and solves the problem that existing agricultural suitability forecasts rely on kilometer-level numerical forecast products, which cannot characterize micro-scale precipitation differences between hundreds of-meter-level field plots, such as temperature and wind differences between slopes and valleys, and localized heavy rainfall at field edges. By using refined meteorological grid forecast data sequences to drive meteorological suitability models, a suitability forecast index product with a range of tens to hundreds of meters is generated, providing users with actionable and high-value decision-making support. Attached Figure Description

[0042] Figure 1 The diagram shown is a flowchart of a precision agriculture suitability index forecasting method provided in an embodiment of this application.

[0043] Figure 2 The diagram shows a flowchart of a precipitation analysis field generation method provided in an embodiment of this application.

[0044] Figure 3 The diagram shown is a flowchart of a meteorological environment analysis field generation method provided in an embodiment of this application.

[0045] Figure 4 This is a schematic diagram of a suitability index forecast product for the bayberry picking season in a certain county in Zhejiang Province, provided in an embodiment of this application.

[0046] Figure 5 The diagram shown is a schematic diagram of a suitability index forecast product for the expansion period of Dendrobium officinale in a certain county of Zhejiang Province, provided in an embodiment of this application.

[0047] Figure 6 The diagram shown is a structural schematic of a precision agriculture suitability index forecasting system provided in an embodiment of this application.

[0048] Figure 7 It is displayed as an electronic device.

[0049] Explanation of reference numerals in the accompanying drawings: 1. Grid division module; 2. Data fusion module; 3. Short-term extrapolation module; 4. Product generation module; 20. Electronic equipment; 21. Processor; 22. Memory; 23. Output interface; 24. Communication interface; 25. Antenna. Detailed Implementation

[0050] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0051] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0052] It should be noted that the background technology mentions complex terrain such as slope, aspect, and altitude; underlying surface types such as soil and vegetation cover; key phenological periods such as the flowering period of bayberry and the specific developmental stage of dendrobium officinale. Microscale refers to less than 1 kilometer, especially at the tens to hundreds of meters level; forecast errors such as temperature and humidity deviations.

[0053] To address the aforementioned technical challenges, this application provides a refined agricultural suitability index forecasting method, system, and equipment. This method fully leverages the value of multi-source heterogeneous meteorological data, breaking down data barriers and achieving multi-dimensional, high-precision perception of the farmland meteorological environment. It integrates multi-source heterogeneous meteorological data, obtaining a precipitation analysis field by fusing station-observed precipitation and radar-monitored precipitation in a gridded network, and a meteorological environment analysis field by fusing WRF model data and multi-source meteorological observation data. Recursive extrapolation techniques and the WRF model are used to calculate the precipitation and meteorological environment analysis fields, resulting in a refined meteorological gridded forecast data sequence. This scales down kilometer-level forecasts to tens to hundreds of meters, enabling local agricultural meteorological environment forecasts with hundreds-meter-level grid matching for farmland scale. This improves the reliability and confidence of hundreds-meter-level farmland grids and reduces errors. It solves the problem that existing agricultural suitability forecasts rely on kilometer-level numerical forecast products, which cannot characterize micro-scale precipitation differences between hundreds-meter-level field plots, such as temperature and wind differences between slopes and valleys, or localized heavy rainfall differences at field edges. By using refined meteorological grid forecast data sequences to drive meteorological suitability models, a suitability forecast index product with a range of tens to hundreds of meters is generated, providing users with actionable and high-value decision-making support.

[0054] The technical solutions in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0055] like Figure 1As shown in the figure, this embodiment provides a method for predicting the suitability index of precision agriculture. The method includes:

[0056] Step 100: Acquire multi-source heterogeneous meteorological data of the target area and interpolate the multi-source heterogeneous meteorological data into a pre-constructed regular grid network to form initial gridded data.

[0057] Specifically, the pre-constructed regular grid is a regular grid in the spatial coordinate system of the target area; wherein the resolution of the grid is 30 meters to 500 meters.

[0058] In some embodiments, the spatial coordinate system is a plane rectangular coordinate system or a latitude and longitude coordinate system.

[0059] In some embodiments, the grid resolution is matched to the spatial scale of a typical field management unit within the target area.

[0060] Specifically, the rule grid consists of multiple grid points, each ranging in size from tens to hundreds of meters.

[0061] In some embodiments, the size of each grid point can be configured in the range of 30 meters to 500 meters, and the geographical area covered by each grid point is equivalent to a rectangular area with a side length of 30 meters to 500 meters.

[0062] In this embodiment, the size of each grid point is set to 100 meters × 100 meters.

[0063] Specifically, using high-precision topographic data as a base map, GIS technology is employed to generate a regular grid network in the target agricultural area, with each grid point having a resolution ranging from 30 meters to 500 meters. This high-precision topographic data is a resolution digital elevation model.

[0064] Specifically, the multi-source heterogeneous meteorological data includes at least the most recently acquired radar observation data, WRF numerical forecast data, automatic weather station data, agricultural micro-station data, and static geographic information data.

[0065] Furthermore, the radar observation data includes at least radar quantitative precipitation estimates continuously monitored over a 10-minute period; the automatic weather station data and the agricultural micro-station data include at least temperature, humidity, precipitation, and wind direction and speed; the WRF numerical forecast data includes at least the most recent WRF model surface and near-surface multi-level forecast products. The agricultural micro-station observation data is acquired through micro-station sensors from micro-stations deployed by farmers and enterprises, and includes at least agricultural environmental data such as temperature, relative humidity, and precipitation. The static geographic information data includes at least digital elevation model data within 100 meters to obtain high-precision terrain data more accurately.

[0066] In some embodiments, the WRF numerical forecast data includes at least multi-source heterogeneous meteorological data for each grid point, including but not limited to forecast data for temperature, relative humidity, wind field U / V components, sea level pressure, and precipitation for the next 0-24 hours.

[0067] In some embodiments, the multi-source heterogeneous meteorological data is obtained through official channels, monitoring equipment, database access, file transfer, or other means, and is not limited to these. Examples include: government open data platform API interfaces, data collection using sensors, and data collection from surveillance cameras, etc.

[0068] In some embodiments, optimal interpolation techniques can be used to uniformly interpolate multi-source heterogeneous meteorological data into a pre-constructed regular grid network.

[0069] Specifically, before uniformly interpolating multi-source heterogeneous meteorological data onto a pre-constructed regular grid network, the process also includes:

[0070] The acquired multi-source heterogeneous meteorological data were subjected to climate extreme value checks, spatial consistency checks, temporal consistency checks, and multi-source consistency checks.

[0071] The climate extreme value check includes: acquiring historical monitoring data of the target station, calculating the 0.1 percentile and 99.9 percentile of the historical monitoring data according to meteorological elements, setting the monitoring data corresponding to the 0.1 percentile as the first dynamic threshold, and setting the monitoring data corresponding to the 99.9 percentile as the second dynamic threshold; when the real-time monitored multi-source heterogeneous meteorological data is less than the first dynamic threshold or greater than the second dynamic threshold, the multi-source heterogeneous meteorological data is marked as a missing value.

[0072] In some embodiments, the historical monitoring data includes historical monitoring data from at least the most recent three years.

[0073] For example, suppose the 0.1 percentile of the historical monitoring data of a certain summer farmland micro-station is -2℃. If the temperature of a certain summer farmland micro-station suddenly drops to -5℃, which is lower than -2℃ (i.e. the 0.1 percentile of the historical monitoring data), it is determined to be an equipment malfunction and the temperature data is marked as a missing value. This avoids erroneous data from contaminating the subsequent fusion process and improves the reliability of agricultural microclimate data.

[0074] This embodiment dynamically calculates the 0.1 percentile and 99.9 percentile thresholds based on independent historical observation data of each target station, accurately identifying outliers that may be caused by sensor failure, extreme weather, or local interference such as equipment getting wet in the rain or animal damage.

[0075] The spatial consistency check includes: generating a continuous field across the entire region based on all multi-source heterogeneous meteorological data from the target station using bilinear interpolation; calculating the deviation between the multi-source heterogeneous meteorological data of each target station and the expected location value of the corresponding target station; and marking the multi-source heterogeneous meteorological data of the corresponding target station as missing values ​​when the deviation is greater than a preset spatial deviation threshold.

[0076] Specifically, the continuous field across the entire region includes the expected location values ​​of all target sites.

[0077] Specifically, the expected value of this location is the interpolation result of the location of the target station in a spatially continuous field generated based on data from surrounding stations using bilinear interpolation.

[0078] In some embodiments, the bilinear interpolation method can be replaced by ordinary kriging interpolation and inverse distance weighted interpolation that takes elevation into account.

[0079] In some embodiments, the preset spatial deviation threshold can be set by those skilled in the art according to the actual application scenario.

[0080] In this embodiment, the preset spatial deviation threshold can be a temperature of 10°C and a specific humidity of 0.007 kg / kg (approximately 35% relative humidity at standard atmospheric pressure of 25°C).

[0081] For example, in area A of a hilly region at an altitude of 200 meters and a temperature of 30°C (corresponding to a relative humidity of approximately 26%), if the automatic weather station observes a relative humidity of 100% while the expected value is 65%, then the deviation of 35% between the humidity reading of 100% and the expected value of 65% obtained by the automatic weather station in this hilly region exceeds the preset spatial deviation threshold of 26%. Therefore, the humidity reading of 100% obtained by the automatic weather station in this hilly region is marked as a missing value. Only this missing value is removed without affecting the depiction of the real environment, ensuring the integrity of the micro-topographical meteorological characteristics of the farmland.

[0082] For example, the measured temperature of a target site B is 25.3℃, and the expected value of the target site B is 15.1℃. The preset spatial deviation threshold is 10℃. Since |25.3-15.1|=10.2℃>10℃, the measured temperature of the target site B of 25.3℃ is marked as a missing value and is removed in the subsequent calculation process.

[0083] This embodiment can effectively distinguish between real microclimate phenomena and observational noise, avoid erroneous data from contaminating subsequent fusion processes, and improve the reliability of agricultural microclimate data.

[0084] The time consistency check includes: calculating the time variability based on the multi-source heterogeneous meteorological data monitored at the current time, the multi-source heterogeneous meteorological data monitored at the previous time, and the multi-source heterogeneous meteorological data monitored at the next time; when the time variability is greater than the preset variability threshold, the multi-source heterogeneous meteorological data at the corresponding time is marked as missing value.

[0085] Specifically, the formula for calculating the time variability based on the multi-source heterogeneous meteorological data monitored at the current moment, the multi-source heterogeneous meteorological data monitored at the previous moment, and the multi-source heterogeneous meteorological data monitored at the next moment is as follows: ,in, The maximum absolute difference between multi-source heterogeneous meteorological data monitored at adjacent time points. To obtain the maximum value of the function or value within the parentheses, This refers to the multi-source heterogeneous meteorological data being monitored at the current moment. This refers to multi-source heterogeneous meteorological data monitored at the previous moment. This is multi-source heterogeneous meteorological data monitored at a later time.

[0086] Specifically, the preset variability threshold is set by those skilled in the art based on the local climate background.

[0087] For example, assuming the preset variability threshold is 15℃ / h, the temperature monitored by the agricultural micro-station at the previous time of a certain grid point is 20℃, the temperature monitored by the agricultural micro-station at the current time of a certain grid point is 0℃, and the precipitation monitored by the agricultural micro-station at the next time of a certain grid point is 15℃, if max(|20-0|,|0-15|)=20℃>15℃, then the temperature corresponding to that grid point is marked as a missing value.

[0088] This embodiment ensures the spatiotemporal continuity of short-term extrapolation input by calculating the variability of multi-source heterogeneous meteorological data monitored at the current time and the previous / next time.

[0089] The multi-source consistency check includes: comparing the absolute differences between the multi-source heterogeneous meteorological data; taking the maximum value of the absolute differences; when the maximum value is greater than the preset multi-source consistency threshold, it is determined that there is a significant conflict between the multi-source heterogeneous meteorological data, triggering a predefined error handling protocol.

[0090] Specifically, the predefined error handling protocol includes outputting detailed error log information and terminating the execution of the current forecast calculation process.

[0091] The error log information includes records of the location and time of the conflict, the data source involved, and the specific numerical differences.

[0092] Specifically, the target site is a site deployed within the target agricultural area.

[0093] Specifically, the significant conflicts among multi-source heterogeneous meteorological data indicate that at least one data source may be erroneous.

[0094] This embodiment sets up a multi-source consistency check to ensure that when there are irreconcilable inconsistencies in the monitoring data from different sources, unreliable forecast calculations are immediately stopped, thereby ensuring the reliability of subsequent forecast results.

[0095] This embodiment eliminates significant differences between data from different sources by performing climate extreme value checks, spatial consistency checks, temporal consistency checks, and multi-source consistency checks on the acquired multi-source heterogeneous meteorological data. This improves the reliability of the data sources and makes the subsequent data fusion and fitness index experimental results more accurate.

[0096] Step 200: The precipitation data obtained from station observations and radar monitoring in the initial gridded data are fused to obtain the precipitation analysis field, and the meteorological environment analysis field is obtained by fusing WRF model data and multi-source meteorological observation data.

[0097] In some embodiments, the multi-source meteorological observation data includes, but is not limited to, data from automatic weather stations, data from agricultural micro-stations, temperature, relative humidity, and wind.

[0098] In some embodiments, the multi-source meteorological observation data can be obtained by monitoring various types of sensors, such as temperature sensors, humidity sensors, and wind direction sensors.

[0099] Specifically, such as Figure 2 As shown, the precipitation analysis field is obtained by fusing the precipitation observed by stations and the precipitation monitored by radar in this grid, including:

[0100] Step 201: Set the precipitation of the m nearest grid points of the target station to be consistent with the measured value of the target station, and use radar to estimate the precipitation in areas other than the target station.

[0101] Preferably, m is a positive integer, and in this embodiment it is set to 4.

[0102] For example, the precipitation at the four nearest grid points of the target station is set to be consistent with the measured value of the target station, and the precipitation in the non-target station area is estimated by radar.

[0103] In this embodiment, when the grid resolution is 100m×100m, the four nearest grid points are selected, which is about 200m×200m. This just covers the grid edges, slopes and other micro-areas around the target site, which can preserve the accurate measured values ​​of the site without over-correcting the range.

[0104] Step 202: Fuse the observed precipitation data of stations in the grid network where the target station is located with the estimated precipitation data of the radar to obtain the precipitation analysis field.

[0105] For example, suppose the target area is a 5×5 km bayberry planting area, and this target area is divided into a 50×50 grid, where each grid has a resolution of 100 m × 100 m. Assume there are two target stations in this bayberry planting area: an automatic weather station A and an agricultural micro-station B. The coordinates of the automatic weather station A are (2.3 km, 2.3 km), with a measured precipitation of 15 mm. The coordinates of the agricultural micro-station B are (3.7 km, 3.7 km), with a measured precipitation of 8 mm. The coordinates of the automatic weather station A are mapped to the grid at grid (23, 23), and its four nearest neighbors are (23, 22), (23, 24), (24, 23), and (22, 23). The coordinates of the agricultural micro-station B are mapped to the grid at grid (37, 37), and its four nearest neighbors are (37, 36), (38, 37), (37, 38), and (36, 37). The four nearest grid points of automatic weather station A were all set to the measured value of 15 mm; the four nearest grid points of agricultural micro-station B were all set to the measured value of 8 mm. Areas outside the target station were filled with radar estimated data, i.e., radar values ​​of 5 mm.

[0106] This embodiment improves the spatial resolution of the precipitation field by fusing kilometer-level radar data with target station data to generate a precipitation analysis field of tens to hundreds of meters. This embodiment also corrects the error of the radar data near the target station by using measured values ​​of the four nearest grid points, making the precipitation field closer to the actual farmland conditions.

[0107] Specifically, such as Figure 3 As shown, the meteorological environment analysis field is obtained by fusing WRF model data and multi-source meteorological observation data, including:

[0108] Step 211: Use three-dimensional interpolation to interpolate the WRF forecast field into the grid network to obtain background field grid data.

[0109] Step 212: Calculate the difference between the monitoring data of each target site and the background field grid data of the corresponding target site.

[0110] The formula for calculating the difference between the monitoring data of each target site and the background field grid data of the corresponding target site is as follows: , For the monitoring data of the kth station, This is background field grid data.

[0111] Step 213: Use the inverse distance weighting algorithm to fuse the differences between n grid points around each target site to generate a difference field.

[0112] For example, given the coordinates of a target station (i,j), the distance between the target station (i,j) and the kth automatic weather station is... Then the value difference of the (i,j)th target station is divided into: , Let be the difference field value of the target site (i,j), and n be the number of neighboring target sites involved in the calculation. For target site To the The horizontal distance between automatic weather stations The square of the distance is used for inverse distance weighting, where the weight decreases as the distance increases.

[0113] In this embodiment, n is set to 8 to balance the accuracy and efficiency of generating the difference field.

[0114] In some embodiments, for a target station (i,j), the distance between the target station (i,j) and the k-th automatic weather station is... Then the value difference of the (i,j)th target station is divided into: , The weights are denoted as 1. Assume an automatic weather station has a weight of 1, and an agricultural micro-station has a weight of 1.5 to 2. Since agricultural micro-stations are typically deployed directly in the crop canopy, their monitoring values ​​better reflect the microclimate of farmland. This embodiment increases the proportion of agricultural micro-stations in the differential field by increasing their weight, thereby reducing the data error in the differential field and ultimately improving the accuracy of the meteorological environment analysis field.

[0115] Step 214: Fuse the difference field with the background field grid data to obtain the meteorological environment analysis field.

[0116] Specifically, the calculation formula for this meteorological environment analysis field is as follows: ,in, For target site Background field grid data.

[0117] This embodiment utilizes WRF model data and multi-source meteorological observation data to fuse a meteorological environment analysis field, achieving a more refined correction of meteorological environment elements in the gridded network compared to traditional numerical model interpolation schemes. Objective verification shows that the analysis field effectively reduces errors compared to traditional numerical model interpolation schemes, especially in complex terrain areas with limited data, where the analysis field still provides reasonable and detailed meteorological information. This method calculates the difference between the monitoring data of each target station and the corresponding background gridded data, preserving the characteristics of meteorological abrupt changes while avoiding the over-smoothing phenomenon caused by traditional regional averaging methods.

[0118] This embodiment uses monitoring data from dense stations to correct the background field grid data, controlling the error of the meteorological environment analysis field results to a low level. Specifically, the average error of temperature is controlled at -0.12℃, the average error of wind speed is controlled at 0.01m / s, the average error of precipitation is controlled at -0.01mm, and the average error of relative humidity is controlled at -0.42%, achieving precise characterization of farmland microclimate at the 100-meter level.

[0119] Step 300: Use a recursive extrapolation algorithm and WRF model to perform time-series extrapolation of the precipitation analysis field and the meteorological environment analysis field to generate a refined meteorological grid data sequence with the same resolution as the grid network for a future preset time period.

[0120] Specifically, the refined meteorological grid forecast data sequence is a refined grid forecast data of various meteorological elements with a spatial resolution of 100 meters for the next 24 hours.

[0121] Specifically, the expression for calculating the meteorological environment analysis field using the recursive extrapolation method and the WRF model is as follows: ,in, For the first Forecast values ​​of elements at any given time, For the first WRF numerical prediction value at time 10:00. For the first Forecast values ​​of elements at any given time, For the first The WRF numerical forecast value at a given time. This forecast value can be the forecast value of humidity, wind, temperature, etc.

[0122] This embodiment uses the changing trends of meteorological environmental elements given by the WRF model to calculate the forecast values ​​of meteorological elements for each grid point using a recursive extrapolation method. While maintaining a spatial resolution of hundreds of meters, it achieves high-precision dynamic forecasts of meteorological environmental elements from 0 to 24 hours. In this embodiment, the improvement in each meteorological environmental element over the 1-6 hour period compared to the background field grid data is 21%-23%, thus making the forecast values ​​closer to farmland conditions. These meteorological environmental elements include, but are not limited to, temperature, relative humidity, and wind.

[0123] Specifically, the precipitation analysis field is calculated using recursive extrapolation techniques and the WRF model, including:

[0124] For 2-6 hour forecasts, based on the time-dependent dynamic fusion of cross-correlation algorithm extrapolation results and WRF model downscaling forecasts, the extrapolation weights for 3-6 hour forecasts are 0.8, 0.6, 0.4, and 0.2, respectively. That is, the extrapolation weight for 3 hours is 0.8, for 4 hours it is 0.6, for 5 hours it is 0.4, and for 6 hours it is 0.2. The weighting function formula satisfies... T.

[0125] Specifically, the expression for the extrapolation results of the time-dependent dynamic fusion cross-correlation algorithm and the downscaling forecast of the WRF model is as follows: ,in For forecast lead time, the unit is hours. For the WRF model's downscaled precipitation forecast, This is the extrapolation result of the cross-correlation algorithm.

[0126] For 0-2 hour forecasts, the movement vector of precipitation echoes is tracked using the cross-correlation algorithm (tracking radar echoes by correlation, TREC) to obtain the location and intensity of precipitation within 0-2 hours.

[0127] Specifically, first at time t-Δt, using grid points Define a dimension centered at the center. Extrapolation template window. Centered on this point, define a window of size [value missing] on the image at time t. The search region. Then, within the search region at time t, all possible displacements of the template window at times t-Δt are calculated. Spatial cross-correlation was performed on corresponding candidate windows of the same size to calculate the correlation coefficient. Then, find the correlation coefficient. Displacement corresponding to the maximum value The displacement Vectors are grid points from The optimal displacement vector up to time t. Finally, proceed with the future... Precipitation field forecast at any time: ,in, For the interval of precipitation, To perform bilinear interpolation on all multi-source heterogeneous meteorological data of the target station, non-integer grid points are used to obtain radar quantitative precipitation estimation data.

[0128] In this embodiment, a is set to 50, M is set to 100, and the interval is... Set the time to 10 minutes. 'a' and 'M' can be adjusted based on the grid resolution and search area size.

[0129] In this embodiment, the extrapolation weight is set to 1.0 for the first 2 hours and 0.8 for the third hour to preserve the cross-correlation algorithm's ability to capture sudden strong convection events. The extrapolation weight for the 6-hour period is set to 0.2 to rely on the WRF model for large-scale circulation forecasts and avoid accumulated extrapolation errors. By reducing the weight from 0.8 to 0.2 for the 3-6 hour periods, the timeliness of short-term extrapolation is balanced with the stability of the fitness index forecast.

[0130] Forecasts 6 hours later were based entirely on the WRF model, with optimal interpolation downscaling to a 100m x 100m grid.

[0131] This embodiment improves the 1-3 hour forecast accuracy of grid points in the target agricultural area under heavy rain scenarios by 140% compared with the TS score of the single WRF model through a three-stage design of cross-correlation algorithm displacement tracking, weight function fusion, and WRF model downscaling. (TS score is one of the most important classification scoring indicators in the meteorological field, especially in the verification of weather event forecasts such as precipitation and convection, and is used to measure the accuracy of whether the forecast event has occurred.)

[0132] Step 400: Input the refined meteorological gridded data sequence into the meteorological suitability model for adaptive calculation to obtain the gridded agricultural operation suitability index.

[0133] Specifically, the meteorological suitability model includes, but is not limited to: wheat harvesting period suitability model, bayberry picking model, bayberry fertilization model, rice heat damage model, Ganoderma lucidum mycelium growth and development period model, etc., which are suitability evaluation models based on temperature, humidity, wind, precipitation, and topography.

[0134] For example, consider the forecast of the bayberry picking season in a county in Zhejiang Province. Suppose the forecast starts at 14:00 on June 18, 2024, with a lead time of 4 hours, predicting the suitability for bayberry picking in a high-altitude bayberry planting area (within a 10km × 10km range) in the county. The 10km × 10km area is divided into 100 grid points with a resolution of 100m × 100m. Data is obtained from the following sources: 3km resolution WRF model data for the bayberry planting area; radar observation data at 13:50; automatic weather station data for temperature, relative humidity, wind, and precipitation at 13:30, 13:40, and 13:50 on the same day; agricultural micro-station data for temperature, relative humidity, wind, and precipitation at 13:30, 13:40, and 13:50 on the same day; and a 90-meter resolution digital elevation model. The acquired WRF model data, radar observation data, automatic weather station data, agricultural micro-station data, and digital elevation model will undergo climate extreme value checks, spatial consistency checks, temporal consistency checks, and multi-source consistency checks. These data will then be converted to NetCDF format. Data fusion and short-term extrapolation will be performed on the acquired WRF model data, radar observation data, automatic weather station data, agricultural micro-station data, and digital elevation model to generate a refined meteorological gridded forecast data sequence. This refined meteorological gridded forecast data sequence will be input into the suitability model for the bayberry picking season. The expression for the suitability model for the bayberry picking season is as follows: Where q(R) is the precipitation suitability function, It is a humidity suitability function, depending only on relative humidity. It is a wind speed fitness function that depends only on wind speed. It is a function of temperature comfort and depends only on temperature. , , , The weighting coefficient is determined by local climate characteristics and crop growth characteristics. Let's assume it's determined based on the characteristics of a certain bayberry harvesting period. , , , =0.4. The refined meteorological grid forecast data sequence is input into the bayberry harvesting period suitability model for adaptive calculation, yielding the bayberry harvesting period suitability index for each grid point. A schematic diagram of the suitability index forecast product for the bayberry harvesting period in a county in Zhejiang Province is shown below. Figure 4 As shown. Figure 4 BJT in Chinese refers to Beijing time, and each colored block represents a 100-meter by 100-meter grid. Figure 4 Each color block displays a color gradient from dark green (100) to brown (0), corresponding to a suitability index from high to low. When the suitability index is greater than or equal to 80, the color block is dark green, indicating that the risk of picking bayberries is low and the quality is good, and the corresponding suitability index is classified as "very suitable"; when the suitability index is between 55 and 80, the color block is light green, indicating that attention should be paid to weather changes when picking bayberries, and the corresponding suitability index is classified as "moderately suitable"; when the suitability index is between 25 and 55, the color block is light yellow, indicating that picking bayberries may affect picking efficiency or fruit quality, and the corresponding suitability index is classified as "less suitable"; when the suitability index is less than 25, the color block is brown, indicating that extreme weather such as heavy rain and strong winds may occur, and picking is prohibited, and the corresponding suitability index is classified as "unsuitable". Figure 4 By generating a quantitative suitability index of 0 to 100, the system clearly marks suitable and unsuitable areas for harvesting, providing farmers with actionable decision-making support. Farmers can use the gridded results to precisely plan harvesting routes, avoid risky areas, and reduce fruit losses due to weather changes.

[0135] For example, taking the environmental suitability index forecast for the bulking period of Dendrobium officinale in a certain county in Zhejiang Province as an example, the agricultural planting area of ​​Dendrobium officinale in this county is 10km×10km. Assuming the report starts from 08:00 on December 18, 2025, with a lead time of 4 hours, the following data is obtained: 3km resolution WRF model data of the agricultural planting area of ​​Dendrobium officinale in the county at the most recent time (05:00 on December 18, 2025), automatic weather station data of temperature and relative humidity of the agricultural planting area of ​​Dendrobium officinale in the county at 07:50 on the same day, agricultural micro-station data of temperature and relative humidity of the agricultural planting area of ​​Dendrobium officinale in the county at 07:50 on the same day, and a 90-meter resolution digital elevation model. The acquired WRF model data, automatic weather station data, agricultural micro-station data, and digital elevation model were subjected to climate extreme value checks, spatial consistency checks, temporal consistency checks, and multi-source consistency checks. These data were then converted to NetCDF format. Data fusion and short-term extrapolation were performed on the acquired WRF model data, automatic weather station data, agricultural micro-station data, and digital elevation model to generate a refined meteorological gridded forecast data sequence. This refined meteorological gridded forecast data sequence was input into an environmental suitability model for the expansion stage of *Dendrobium officinale*. The expression for this environmental suitability model is Q(R,S,U,K)=a×q(S)+b×q(K), assumed to be determined by the climate and growth characteristics of a certain county. , A gridded suitability index forecast product for the bulking stage of Dendrobium officinale was generated with a resolution of 100m×100m. A schematic diagram of the suitability index forecast product for the bulking stage of Dendrobium officinale in a certain county is shown below. Figure 5 As shown. Figure 5 BJT in Chinese refers to Beijing time, and each colored block represents a 100-meter by 100-meter grid. Figure 5 Each color block in the chart displays a color gradient from dark green (100) to brown (0), corresponding to a suitability index from high to low. When the suitability index is greater than or equal to 80, the color block is dark green, indicating that environmental conditions such as temperature and humidity meet the requirements of Dendrobium officinale during its growth period, resulting in excellent quality; the corresponding suitability index is classified as "very suitable." When the suitability index is between 55 and 80, the color block is light green, indicating that slight adjustments to field management are needed, such as temperature and humidity control; the corresponding suitability index is classified as "moderately suitable." When the suitability index is between 25 and 55, the color block is pale yellow, indicating that meteorological conditions may affect the growth rate, requiring enhanced monitoring; the corresponding suitability index is classified as "less suitable." When the suitability index is less than 25, the color block is brown, indicating that temperatures may be too low or humidity too high; the corresponding suitability index is classified as "unsuitable." Dendrobium officinale is extremely sensitive to temperature and humidity during its growth period. Generating a quantitative suitability index forecast product ranging from 0 to 100 can accurately reflect the matching degree between environmental conditions and crop needs.

[0136] This embodiment converts refined meteorological elements such as rainfall, humidity, wind speed, and temperature into quantifiable suitability indices, providing farmers with actionable and high-value decision-making support, directly guiding agricultural activities, and solving the problem that existing technical solutions cannot achieve quantitative short-term forecasting.

[0137] The scope of protection for the precision agriculture suitability index forecasting method described in this application is not limited to the order of steps listed in this embodiment. Any scheme implemented by adding, subtracting, or replacing steps in existing methods based on the principles of this application is included within the scope of protection of this application.

[0138] This application also provides a precision agricultural suitability index forecasting system, which can implement the precision agricultural suitability index forecasting method described in this application. However, the implementation device of the precision agricultural suitability index forecasting method described in this application includes, but is not limited to, the structure of the precision agricultural suitability index forecasting system listed in this embodiment. All structural modifications and substitutions of the prior art made based on the principles of this application are included within the protection scope of this application.

[0139] like Figure 6 This embodiment provides a refined agricultural suitability index forecasting system, which includes: a gridded data generation module 1, configured to acquire multi-source heterogeneous meteorological data of a target area and uniformly interpolate the multi-source heterogeneous meteorological data into a pre-constructed regular grid network to form initial gridded data; a data fusion module 2, configured to fuse the station-observed precipitation and radar-monitored precipitation in the initial gridded data to obtain a precipitation analysis field, and to fuse WRF model data and multi-source meteorological observation data to obtain a meteorological environment analysis field; a short-term extrapolation module 3, configured to use recursive extrapolation technology and WRF model to perform time-series extrapolation of the precipitation analysis field and the meteorological environment analysis field to generate a refined meteorological gridded data sequence with the same resolution as the regular grid network for a future preset time period; and an index generation module 4, configured to input the refined meteorological gridded data sequence into a meteorological suitability model for adaptive calculation to obtain a gridded agricultural operation suitability index.

[0140] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0141] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of this application, depending on actual needs. For example, the functional modules / units in the various embodiments of this application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0142] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0143] In one exemplary embodiment, this embodiment also provides an electronic device, including a memory and a processor, wherein the memory stores a program that can run on the processor, and when the program is executed by the processor, the electronic device performs any of the methods described in the above embodiments.

[0144] In one possible embodiment, such as Figure 7 As shown, the electronic device 20 also includes: an output interface 23 for outputting results; a communication interface 24 for transmitting communication signals; and an antenna 25 for transmitting or receiving signals.

[0145] It should be noted that the processor 21 in this embodiment can be an image processing chip or an integrated circuit chip, capable of processing image signals. In implementation, each step of the above method embodiment can be completed by the integrated logic circuitry in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices. It can implement or execute the methods, steps, and logic block diagrams disclosed in this embodiment. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this embodiment can be directly implemented by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0146] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0147] In one exemplary embodiment, this embodiment also provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps in the above-described method embodiments.

[0148] In one exemplary embodiment, this embodiment also provides a computer program product, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0149] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0150] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0151] The workflow package initialization method, system, electronic device, and computer program product provided in this application can achieve efficient and intelligent workflow package initialization, enable batch processing of multiple business tasks, improve the management efficiency and collaboration of job tasks, optimize data processing flow, and meet the needs of enterprises for efficient operation of robotic process automation systems in complex business scenarios.

[0152] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A method for predicting the suitability index of refined agriculture, characterized in that, The method includes: Acquire multi-source heterogeneous meteorological data of the target area, and uniformly interpolate the multi-source heterogeneous meteorological data into a pre-constructed regular grid network to form initial gridded data; The precipitation analysis field is obtained by fusing the station observation precipitation and radar monitoring precipitation in the initial gridded data, and the meteorological environment analysis field is obtained by fusing WRF model data and multi-source meteorological observation data. The precipitation analysis field and the meteorological environment analysis field are temporally extrapolated using a recursive extrapolation algorithm and a WRF model to generate a refined meteorological grid data sequence with the same resolution as the regular grid network for a future preset time period. The refined meteorological gridded data sequence is input into the meteorological suitability model for adaptive calculation to obtain the gridded agricultural operation suitability index.

2. The method according to claim 1, characterized in that, The multi-source heterogeneous meteorological data includes at least the most recently acquired radar observation data, WRF numerical forecast data, automatic weather station data, agricultural micro-station data, and static geographic information data.

3. The method according to claim 1, characterized in that, Before uniformly interpolating the multi-source heterogeneous meteorological data onto a pre-constructed regular grid network, the process also includes: The acquired multi-source heterogeneous meteorological data are subjected to climate extreme value checks, spatial consistency checks, temporal consistency checks, and multi-source consistency checks. The climate extreme value check includes: acquiring historical monitoring data of the target station, calculating the 0.1 percentile and 99.9 percentile of the historical monitoring data according to meteorological elements, setting the monitoring data corresponding to the 0.1 percentile as a first dynamic threshold, and setting the monitoring data corresponding to the 99.9 percentile as a second dynamic threshold; when the real-time monitored multi-source heterogeneous meteorological data is less than the first dynamic threshold or greater than the second dynamic threshold, the multi-source heterogeneous meteorological data is marked as a missing value; The spatial consistency check includes: generating a continuous field across the entire region using bilinear interpolation based on all multi-source heterogeneous meteorological data from the target station; calculating the deviation between the multi-source heterogeneous meteorological data of each target station and the expected location value of the corresponding target station; and marking the multi-source heterogeneous meteorological data of the corresponding target station as missing values ​​when the deviation is greater than a preset spatial deviation threshold. The time consistency check includes: calculating the time variability based on the multi-source heterogeneous meteorological data monitored at the current time, the multi-source heterogeneous meteorological data monitored at the previous time, and the multi-source heterogeneous meteorological data monitored at the next time; when the time variability is greater than a preset variability threshold, the multi-source heterogeneous meteorological data at the corresponding time is marked as a missing value. The multi-source consistency check includes: comparing the absolute differences between the multi-source heterogeneous meteorological data; taking the maximum value of the absolute differences; when the maximum value is greater than a preset multi-source consistency threshold, it is determined that there is a significant conflict between the multi-source heterogeneous meteorological data, and a predefined error handling protocol is triggered.

4. The method according to claim 1, characterized in that, The regular grid is a regular grid in the spatial coordinate system of the target area; wherein, the resolution of the regular grid is 30 meters to 500 meters.

5. The method according to claim 2, characterized in that, The process of fusing station-observed precipitation and radar-monitored precipitation in the initial gridded data to obtain a precipitation analysis field includes: The precipitation values ​​of the m nearest grid points of the target station are set to be consistent with the measured values ​​of the target station, and the precipitation in the non-target station area is estimated by radar. The precipitation analysis field is obtained by fusing the observed precipitation data from stations in the grid network where the target station is located with the radar-estimated precipitation data.

6. The method according to claim 2, characterized in that, The meteorological environment analysis field obtained by fusing WRF model data and multi-source meteorological observation data includes: The WRF forecast field is interpolated into the grid network using a three-dimensional interpolation method to obtain background field grid data. Calculate the difference between the monitoring data of each target site and the background field grid data of the corresponding target site; A difference field is generated by fusing the differences between n grid points surrounding each target site using an inverse distance weighting algorithm. The difference field and the background field grid data are fused to obtain the meteorological environment analysis field.

7. The method according to claim 1, characterized in that, The recursive extrapolation algorithm and WRF model are used to perform time-series extrapolation of the precipitation analysis field and the meteorological environment analysis field to generate a refined meteorological grid data sequence with the same resolution as the grid network for a future preset time period, including: For 0-2 hour forecasts, the moving vector of precipitation echoes is tracked based on the cross-correlation algorithm to generate a precipitation extrapolation forecast field with the same resolution as the regular grid network for the next 0-2 hours. For 2-6 hour forecasts, the precipitation extrapolation forecast field and the WRF model downscaling forecast field are dynamically weighted and fused, wherein the fusion weight function satisfies: ,in, To provide a forecast timeframe, The unit is hours. The weights for extrapolated forecast fields within 2 to 6 hours; For forecasts longer than 6 hours, the WRF model is used to downscale the forecast results to the regular grid network.

8. The method according to claim 1, characterized in that, The expression for calculating the meteorological environment analysis field using the recursive extrapolation algorithm and WRF model is as follows: ,in, For the first Forecast values ​​of elements at any given time, For the first WRF numerical prediction value at time 10:

00. For the first Forecast values ​​of elements at any given time, For the first WRF numerical forecast value at time 10:

00.

9. The method according to claim 1, characterized in that, The meteorological suitability models include a wheat harvesting period suitability model, a bayberry picking model, a bayberry fertilization model, a rice heat damage model, and a Ganoderma lucidum mycelium growth and development period model.

10. A precision agricultural suitability index forecasting system, characterized in that, The system includes: The gridded data generation module is configured to acquire multi-source heterogeneous meteorological data of the target area and interpolate the multi-source heterogeneous meteorological data into a pre-constructed regular grid network to form initial gridded data. The data fusion module is configured to fuse the station-observed precipitation and radar-monitored precipitation in the initial gridded data to obtain a precipitation analysis field, and to fuse WRF model data and multi-source meteorological observation data to obtain a meteorological environment analysis field. The short-term extrapolation module is configured to use a recursive extrapolation algorithm and WRF mode to perform time-series extrapolation on the precipitation analysis field and the meteorological environment analysis field, and generate a refined meteorological grid data sequence with the same resolution as the regular grid network for a future preset time period. The index generation module is configured to input the refined meteorological grid data sequence into the meteorological suitability model for adaptive calculation to obtain the gridded agricultural operation suitability index.

Images