Agricultural drought prediction method and system based on assimilation of multi-source heterogeneous data

Abstract

The application provides an agricultural drought prediction method and system based on multi-source heterogeneous data assimilation, relates to the field of agricultural drought prediction, obtains and pre-processes multi-source heterogeneous agricultural environment monitoring information, identifies and confirms local sudden hydrological events based on specific difference patterns between different source information, determines the actual influence range through physical deduction after confirming the occurrence of the event, adjusts the credibility of different source information and performs fusion according to the influence range, generates a comprehensive agricultural environment condition distribution map, generates agricultural environment evaluation results and differentiated suggestions for different regions based on the distribution map, overcomes the shortcomings of the fixed data fusion strategy in the prior art, can dynamically adapt to complex and changeable agricultural environments, significantly improves the accuracy and pertinence of drought prediction, and thus effectively avoids unnecessary resource waste and crop loss.

Description

Technical Field

[0001] This application relates to the field of agricultural drought prediction, and more specifically, to an agricultural drought prediction method and system based on multi-source heterogeneous data assimilation. Background Technology

[0002] In modern, large-scale, and digitalized agricultural production and management, agricultural managers rely heavily on integrated monitoring platforms that combine satellite remote sensing, meteorological observation networks, and Internet of Things (IoT) sensors to obtain a macro-level picture of the farmland environment over large areas. This information supports water resource allocation, disaster early warning, and precision agricultural decision-making. Such systems aim to generate accurate and unified maps of agricultural environmental conditions (such as soil moisture) by assimilating multi-source, heterogeneous data.

[0003] However, due to inherent differences in spatiotemporal resolution, update frequency, observation principles, and coverage among different data sources, traditional data assimilation or fusion strategies with pre-defined weights exhibit limitations when faced with highly localized and sudden hydrological events. These events include, but are not limited to: localized heavy rainfall caused by thunderstorms, accidental ruptures in irrigation system pipes, or poor drainage and waterlogging caused by localized topography.

[0004] In such events, data streams from different sources can become sharply contradictory. For example, IoT soil moisture sensors deployed in fields may detect a rapid rise in humidity to saturation at a certain point in a short period of time, but weather radar or a sparsely distributed network of ground weather stations may fail to effectively capture the localized rainfall signal due to insufficient spatial resolution or data interpolation smoothing. At the same time, optical satellite remote sensing may be unable to provide effective observations at the time of the event due to cloud cover or transit time issues.

[0005] Faced with this contradiction, the existing system's rigid data fusion logic faces a dilemma: if local abnormal sensor readings are simply treated as "noise" or "fault data" and discarded or downweighted, key risk points (such as localized waterlogging) will be missed in the generated macro-level situation map, leading to overly optimistic assessments of the overall situation by managers and potentially delaying drainage and other response measures. Conversely, if the data from this local point is over-relied without discernment and its wetness status is extrapolated to an excessively large area based on simple spatial interpolation algorithms (such as inverse distance weighting), the affected area of ​​the event will be severely overestimated, masking surrounding farmland that is still experiencing water shortages, potentially leading to erroneous water resource allocation instructions.

[0006] The essence of this contradiction lies in the lack of dynamic event perception and interpretation capabilities in the relevant technologies. The system cannot intelligently identify "when a special event occurred" and, based on this, temporarily and reasonably adjust the confidence assessment and data fusion strategies for different data sources in different spatial locations. Therefore, in scenarios with frequent localized sudden hydrological events, the reliability of the comprehensive distribution maps generated by existing methods decreases, making it difficult to meet the high reliability requirements of large-scale agricultural management for precise regional water resource allocation, rapid disaster assessment, and differentiated agricultural guidance. Summary of the Invention

[0007] This application provides a method and system for agricultural drought forecasting based on multi-source heterogeneous data assimilation. It aims to address the technical problem that existing agricultural environmental monitoring systems, when facing localized sudden hydrological events, struggle to effectively reconcile conflicting data from multiple sources using fixed data fusion strategies, leading to decreased accuracy and reliability of the generated macro-environmental distribution maps. This application improves the accuracy and reliability of monitoring and forecasting agricultural drought and hydrological conditions under complex and variable environments by introducing an event-driven dynamic data assimilation mechanism, providing high-quality data support for upper-level agricultural management decision-making systems.

[0008] On the one hand, this application provides an agricultural drought prediction method based on multi-source heterogeneous data assimilation, including: Agricultural environmental monitoring information from different sources is acquired and preprocessed to obtain target agricultural environmental monitoring information. Based on specific difference patterns between different sources of information in the target agricultural environmental monitoring information, the occurrence of local sudden hydrological events is identified and confirmed; After confirming the occurrence of the localized sudden hydrological event, the actual impact range of the localized sudden hydrological event is determined through a physical deduction process, based on the geographical environment information of the event location and the event's characteristic information, thus obtaining the division result of the actual impact range. Based on the division results, the credibility of the information from different sources in different regions is adjusted, and the target agricultural environmental monitoring information is fused to generate a comprehensive agricultural environmental status distribution map; and based on the agricultural environmental status distribution map, agricultural environmental assessment results and differentiated recommendations are generated for different regions.

[0009] Optionally, the step of identifying and confirming the occurrence of a localized sudden hydrological event based on specific difference patterns between information from different sources in the target agricultural environmental monitoring information includes: When at least one soil moisture sensor is detected to have its soil moisture reading rise sharply to saturation within a preset short period of time, and at the same time period, the local radar rainfall intensity data from a preset spatial range is lower than a first threshold, the occurrence of a local sudden hydrological event is confirmed by combining the latest satellite remote sensing data.

[0010] Optionally, the step of determining the actual impact range of the local sudden hydrological event after confirming its occurrence, by combining the geographical environment information of the event location and the event's characteristic information, and through a physical deduction process, to obtain the delineation result of the actual impact range includes: Acquire micro-topographic data, soil stratification information, and real-time rainfall intensity data of the area where the event occurred; The system continuously receives data from stratified soil moisture sensors and infers and updates the local hydrological parameters required for the integrated hydrological model based on the data. The stratified soil moisture sensors are deployed at different depths in the same location to provide vertical profile information of soil moisture. An integrated hydrological model that updates local hydrological parameters is activated to simulate the movement of water in the soil, and the area where the soil moisture content reaches saturation is marked as the actual area of ​​influence.

[0011] Optionally, the step of adjusting the credibility of the information from different sources in different regions based on the division results, and fusing the target agricultural environmental monitoring information to generate a comprehensive agricultural environmental status distribution map includes: Continuously monitor all identified localized sudden hydrological events and obtain the classification results of each event's type, occurrence time, actual impact range, and suggested adjustments to the data's reliability. Identify the overlapping areas of the actual impact range of the event, prioritize all relevant events within the overlapping areas, and obtain the priority ranking result; Based on the suggested adjustment scheme for the data credibility corresponding to the highest priority event in the priority ranking results, the final credibility of information from different sources in the overlapping area is set; and based on the final credibility, the credibility of information from different sources in different areas is adjusted, and the information from different sources is fused to generate a comprehensive agricultural environmental status distribution map.

[0012] Optionally, the step of generating agricultural environmental assessment results and differentiated recommendations for different regions based on the agricultural environmental status distribution map includes: Obtain information on crop types, crop growth stages, soil nutrient levels, and agricultural machinery operation capabilities in the region; Based on the crop type information and the crop growth stage information, the crop water requirement characteristics database is queried to obtain the range and sensitivity of the crop's water requirement at the current growth stage, and the crop water requirement characteristics are obtained. Based on the crop type information and the crop growth stage information, the crop nutrient absorption characteristic database is queried to obtain the crop's demand for various nutrients and absorption efficiency at the current growth stage, thus obtaining the crop nutrient absorption characteristics. Based on the soil nutrient level information, analyze the soil test data, identify the content of major nutrients in the soil, and compare the content of major nutrients in the soil with the crop's requirements for various nutrients at the current growth stage to obtain the soil nutrient level. Based on the agricultural machinery operation capacity information, query the available agricultural machinery equipment types, operation efficiency, and operation cost to obtain the agricultural machinery operation capacity; Based on the moisture information in the agricultural environment distribution map, combined with crop water requirements, crop nutrient absorption characteristics, soil nutrient levels, and agricultural machinery operation capabilities, differentiated recommendations for irrigation, fertilization, and drainage operations are generated, and the expected resource consumption and potential benefit assessment of each recommendation are provided.

[0013] Optionally, the step of generating agricultural environmental assessment results and differentiated recommendations for different regions based on the agricultural environmental status distribution map includes: Historical hydrological event records for the region are obtained. Based on these records and the soil moisture dissipation patterns, historical hydrological events with residual effects are identified. The residual impact of these events on the current water conditions is calculated based on the type and intensity of the identified events. Identify newly occurring localized sudden hydrological events in the area, and calculate the immediate impact of the newly occurring localized sudden hydrological events on the current water conditions based on the type and intensity of the identified events. The residual effect and the immediate effect are superimposed or offset to obtain the comprehensive moisture status; Based on the comprehensive moisture conditions, the agricultural environmental assessment results and differentiated recommendations are generated.

[0014] Optionally, the step of superimposing or offsetting the residual effect and the immediate effect to obtain the comprehensive moisture status includes: The weight of the historical hydrological events on the current water conditions is determined based on the intensity and duration of the historical hydrological events. The weight of a newly occurring localized sudden hydrological event on the current water status is determined based on the intensity and duration of the event. Based on the weights of the historical hydrological events on the current water situation and the weights of the newly occurring localized sudden hydrological events on the current water situation, the residual impact and the immediate impact are weighted to obtain the comprehensive water situation.

[0015] Optionally, the step of determining the weight of the historical hydrological event on the current water condition based on the intensity and duration of the historical hydrological event includes: Acquire the type, intensity, duration of the historical hydrological events, as well as soil type and topographic features of the event locations; Select the initial decay mode based on the type of the historical hydrological events; Based on the intensity and duration of the historical hydrological events, and combined with soil type and topographic feature information of the event location, the parameters of the initial attenuation mode are adjusted to obtain the adjusted attenuation mode. Based on the adjusted attenuation pattern, calculate the attenuation coefficient of the historical hydrological event on the water status at the current time point; The weight of the historical hydrological events on the current water conditions is determined based on the attenuation coefficient.

[0016] Optionally, the step of determining the weight of the newly occurring localized sudden hydrological event on the current water situation based on the intensity and duration of the newly occurring localized sudden hydrological event includes: Obtain the sequence information of the newly occurring localized sudden hydrological events, the sequence information including the occurrence time, intensity, and duration of each event; Based on the sequence information, the spatiotemporal correlation between events in the event sequence is identified to obtain spatiotemporal correlation information; Based on the spatiotemporal correlation information, calculate the cumulative impact factor of each event in the event sequence; Based on the cumulative impact factor, the weight of newly occurring localized sudden hydrological events on the current water status is determined.

[0017] On the other hand, this application provides an agricultural drought prediction system based on multi-source heterogeneous data assimilation, the system comprising: The information receiving and processing module is used to acquire agricultural environmental monitoring information from different sources and preprocess the agricultural environmental monitoring information to obtain target agricultural environmental monitoring information. The event identification and confirmation module is used to identify and confirm the occurrence of local sudden hydrological events based on specific difference patterns between information from different sources in the target agricultural environmental monitoring information. The impact range determination module is used to determine the actual impact range of the local sudden hydrological event after confirming that the event has occurred, by combining the geographical environment information of the event location and the event's characteristic information, and through a physical deduction process, and to obtain the division result of the actual impact range. The assessment recommendation generation module is used to adjust the credibility of information from different sources in different regions based on the division results, and to integrate the target agricultural environmental monitoring information to generate a comprehensive agricultural environmental status distribution map; and based on the agricultural environmental status distribution map, to generate agricultural environmental assessment results and differentiated recommendations for different regions.

[0018] This application discloses an agricultural drought forecasting method and system based on multi-source heterogeneous data assimilation. By acquiring and preprocessing multi-source heterogeneous agricultural environmental monitoring information, it can comprehensively grasp the farmland condition. Furthermore, based on specific difference patterns between information from different sources, this method identifies and confirms localized sudden hydrological events, effectively solving the data inconsistency problem encountered by traditional methods when dealing with localized events. After confirming the event, the actual impact range is determined through physical deduction, combining geographical environment and event characteristics, avoiding misjudgments of the drought situation. Subsequently, the credibility of information from different sources is adjusted and fused according to the impact range to generate a comprehensive agricultural environmental condition distribution map, ensuring the accuracy of the forecast. Finally, based on this distribution map, agricultural environmental assessment results and differentiated recommendations are generated for different regions, providing precise guidance for agricultural production. This method significantly improves the accuracy and relevance of drought forecasting, thereby effectively avoiding unnecessary resource waste and crop losses. Attached Figure Description

[0019] To illustrate this application more clearly, the accompanying drawings used in the embodiments will be briefly described below. Obviously, those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0020] Figure 1 The diagram illustrates a flowchart of an agricultural drought prediction method based on multi-source heterogeneous data assimilation. Figure 2 The diagram illustrates the structure of an agricultural drought prediction system based on multi-source heterogeneous data assimilation.

[0021] Figure reference numerals: 100, Agricultural drought prediction system based on multi-source heterogeneous data assimilation; 10, Information receiving and processing module; 20, Event identification and confirmation module; 30, Impact range determination module; 40, Assessment suggestion generation module. Detailed Implementation

[0022] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0023] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0024] In modern, large-scale, and digitalized agricultural production management, managers rely on monitoring platforms that integrate multi-source heterogeneous data to obtain macro-level information. However, as described in the background section, when faced with sudden local hydrological events, traditional fixed-data fusion strategies can lead to data inconsistencies and distorted situation maps, thereby affecting the reliability of critical decisions such as water resource allocation and disaster assessment. The method and system provided in this application aim to solve this core problem.

[0025] like Figure 1 The diagram illustrates a flowchart of an agricultural drought prediction method based on multi-source heterogeneous data assimilation. This application proposes an agricultural drought prediction method based on multi-source heterogeneous data assimilation, comprising: S10: Obtain agricultural environmental monitoring information from different sources, and preprocess the agricultural environmental monitoring information to obtain target agricultural environmental monitoring information; Agricultural environmental monitoring information refers to various data used by modern agricultural digital management platforms to assess regional environmental conditions. These data come from a wide range of sources and are diverse in type. For example, this information can include macroscopic data such as vegetation indices, surface temperature, and soil moisture content obtained through satellite remote sensing technology; conventional meteorological data such as temperature, humidity, wind speed, and rainfall from ground-based weather stations; and microscopic data obtained from soil moisture sensors, soil temperature sensors, and soil nutrient sensors deployed in the fields. Preprocessing includes steps such as data cleaning, format standardization, and spatiotemporal alignment, aiming to provide a consistent and reliable data foundation for subsequent fusion analysis. These data collectively constitute the data foundation for digital agricultural management.

[0026] S20, Based on the specific difference patterns between different sources of information in the target agricultural environmental monitoring information, identify and confirm the occurrence of local sudden hydrological events; S30, after confirming the occurrence of the local sudden hydrological event, the actual impact range of the local sudden hydrological event is determined through a physical deduction process, combining the geographical environment information of the event occurrence area and the characteristic information of the event, and the division result of the actual impact range is obtained. The physical deduction process refers to a method that uses simulation and calculation based on known physical laws and models to predict or explain the evolution of natural phenomena. In this application, the physical deduction process is mainly used to simulate the movement, diffusion, and evaporation of water in the soil to determine the actual impact range of localized sudden hydrological events.

[0027] S40, based on the division results, adjust the credibility of the information from different sources in different regions, and fuse the target agricultural environment monitoring information to generate a comprehensive agricultural environment status distribution map; and based on the agricultural environment status distribution map, generate agricultural environment assessment results and differentiated recommendations for different regions.

[0028] Credibility refers to the quantitative assessment of the accuracy and reliability of agricultural environmental monitoring information from different sources under specific circumstances. Due to the differences in the characteristics of different data sources and the impact of local emergencies, credibility will change dynamically and needs to be adjusted according to the actual situation.

[0029] The agricultural environmental status distribution map is a graphical representation of the spatial distribution of key environmental parameters such as soil moisture, temperature, and nutrients in agricultural areas, generated by integrating multi-source heterogeneous data and considering the impact of localized sudden hydrological events. This map provides managers with a visual representation of the agricultural environmental conditions in different regions, offering high-quality decision-making support for subsequent assessments and recommendations.

[0030] To more clearly illustrate the core logic and steps of the method of this application, a simplified, principle-based embodiment is provided below. Those skilled in the art should understand that in practical applications, the following manual operation steps can be automatically executed by a computer system.

[0031] First, agricultural environmental monitoring information from various sources is acquired and preprocessed to obtain the target agricultural environmental monitoring information. For example, operators can download remote sensing images from satellite data providers' platforms, obtain weather station data from meteorological data service interfaces, and export field sensor data from IoT platforms. Subsequently, data cleaning tools are used to remove records with obvious errors or non-standard formats, and all data is uniformly converted to the same spatiotemporal coordinate system (such as WGS84 geographic coordinate system and UTC time) to ensure compatibility for subsequent analysis. In a practical system, this process can be automated by writing data crawlers, calling APIs, and running ETL (extract, transform, load) processes.

[0032] Secondly, based on specific discrepancy patterns between different sources of information in the target agricultural environmental monitoring data, the occurrence of localized sudden hydrological events can be identified and confirmed. For example, analysts can examine time-series data from soil moisture sensors. If they find that a sensor reading rises sharply from 30% to 95% (saturation state) within one hour, and simultaneously query satellite precipitation products or sparse weather station network data covering the area during the same time period show rainfall below 0.1 mm, and the latest satellite imagery shows no thick cloud cover in the area, then it can be preliminarily determined that a very localized heavy rainfall or non-rainfalling waterlogging event may have occurred at that location. In automated systems, this process can be achieved by setting up a rule engine or training a machine learning model to continuously monitor the data stream and automatically trigger alarms.

[0033] Secondly, after confirming the occurrence of a localized sudden hydrological event, the actual impact range is determined through a physical deduction process, combining the geographical environmental information of the event area with the event's characteristic information. For example, operators can import the digital elevation model (DEM) data and high-resolution soil type map of the area, and estimate an equivalent rainfall intensity based on the soil moisture rise curve as input. Using open-source or commercial hydrological modeling software (such as SWAT or HYDRUS), initial and boundary conditions are set to simulate the movement of water under the specific terrain and soil. By analyzing the spatial distribution of soil water saturation output by the model, the area where soil moisture reaches saturation is manually delineated, which is the actual impact range deduced from the physical model. The automated system can have a built-in or call-up hydrological model library, automatically run the simulation based on the input parameters, and use GIS tools for spatial analysis, outputting a digital extent layer.

[0034] Next, based on the segmentation results, the credibility of information from different sources in different regions is adjusted, and the target agricultural environmental monitoring information is fused to generate a comprehensive agricultural environmental status distribution map. For example, when drawing a comprehensive soil moisture distribution map, operators can use GIS software to set the fusion weight of soil moisture sensor data at the corresponding location to the highest level (e.g., 0.8) within the saturated area defined by the physical model, while reducing the weight of satellite-retrieved soil moisture data within that area (e.g., 0.2). Outside this area, conventional weight settings are used (e.g., 0.7 for satellite data, 0.3 for sensor data). Subsequently, spatial interpolation methods (e.g., Kriging) are used to fuse all weighted data to generate a comprehensive distribution map. The automated system will dynamically generate a spatialized weight matrix based on the event impact range layer and call data assimilation algorithms (e.g., ensemble Kalman filtering) for real-time fusion calculation.

[0035] Finally, based on the agricultural environmental condition distribution map, assessment results and recommendations are generated. For example, analysts can overlay crop planting distribution maps to identify crop types and growth stages in water-stressed (overly dry or overly wet) areas. Combining crop water requirement databases and farm irrigation equipment capabilities, zoning management recommendations are manually formulated: for example, drainage or suspension of irrigation is recommended for saturated areas, and drip irrigation of 20 mm is recommended for water-deficient areas within the next 3 days. The automated system can integrate farm management databases and decision rule bases to generate detailed reports containing operational items, expected resource consumption, and benefit assessments with a single click, and push them to the management terminal.

[0036] The agricultural drought forecasting method based on multi-source heterogeneous data assimilation proposed in this application demonstrates significant innovation and advantages in solving existing technical problems. Traditional agricultural drought forecasting often employs a fixed data fusion strategy when integrating multi-source heterogeneous data. This strategy, when faced with sudden local hydrological events, such as localized heavy rainfall, can lead to sharp contradictions between different data sources, resulting in misjudgments of drought conditions. For example, when local soil moisture sensor readings rapidly rise to saturation, while distant ground weather stations show zero rainfall, and satellite remote sensing data cannot reflect the latest conditions in a timely manner due to update frequency or cloud cover, the high soil moisture readings might traditionally be considered outliers and rejected, leading to an incorrect judgment that the area is still in a drought state and unnecessary irrigation. Conversely, over-reliance on local sensor data may incorrectly extrapolate moisture conditions to a large area, masking the continued drought in adjacent, un-rained fields.

[0037] The core innovation of this application lies in the introduction of mechanisms for "identifying and confirming the occurrence of local sudden hydrological events based on specific difference patterns between information from different sources in the target agricultural environmental monitoring information" and "adjusting the credibility of information from different sources in different regions according to the division results".

[0038] First, this application is capable of dynamically identifying localized sudden hydrological events. Traditional methods lack the ability to identify localized events and cannot distinguish between local anomalies and large-scale trends. This application, by monitoring specific patterns of difference between information from different sources, such as inconsistencies between soil moisture sensor readings and radar rainfall intensity data, can identify localized sudden hydrological events in a timely and accurate manner, which is not possible with existing technologies.

[0039] Secondly, this application introduces a physical extrapolation process to determine the actual impact range of localized sudden hydrological events. Traditional methods, when dealing with localized events, often employ simple statistical extrapolation, leading to inaccurate assessments of the impact range. This application, by combining geographical environmental information and event characteristic information, utilizes an integrated hydrological model to simulate the movement of water in the soil, enabling precise delineation of the actual impact range and avoiding overestimation or underestimation.

[0040] Furthermore, this application achieves dynamic adjustment of data credibility. In traditional methods, the credibility of data fusion is fixed and cannot be flexibly adjusted according to actual conditions. This application dynamically adjusts the credibility of information from different sources in different regions based on the actual impact range of local sudden hydrological events. For example, it increases the credibility of local sensor data in affected areas, thereby ensuring the accuracy and reliability of the fusion results.

[0041] Finally, this application can generate differentiated and refined agricultural environmental assessment results and recommendations. By comprehensively considering various factors such as the impact of localized sudden hydrological events, crop water requirements, soil nutrient levels, and agricultural machinery operation capabilities, this application can provide differentiated recommendations for agricultural operations including irrigation, fertilization, and drainage for different regions, and provide assessments of expected resource consumption and potential benefits.

[0042] In summary, this application effectively overcomes the challenges faced by existing technologies in processing multi-source heterogeneous data fusion by introducing innovative mechanisms such as intelligent identification of local sudden hydrological events, physical estimation of impact range, dynamic adjustment of data credibility, and refined assessment suggestions. It significantly improves the accuracy, timeliness, and pertinence of agricultural drought forecasting, and provides a more advanced and reliable technical solution for the refined management and sustainable development of modern agriculture.

[0043] In some embodiments, the step of identifying and confirming the occurrence of a localized sudden hydrological event based on specific difference patterns between information from different sources in the target agricultural environmental monitoring information includes: When at least one soil moisture sensor is detected to have its soil moisture reading rise sharply to saturation within a preset short period of time, and at the same time period, the local radar rainfall intensity data from a preset spatial range is lower than a first threshold, the occurrence of a local sudden hydrological event is confirmed by combining the latest satellite remote sensing data.

[0044] Specifically, soil moisture sensors are used to monitor the moisture content in the soil in real time. When the soil moisture sensor reading rises sharply and reaches saturation within a preset time period, such as several minutes to several hours, it indicates that the soil in that area has absorbed a large amount of water in a short period of time, and there may be local waterlogging or runoff accumulation. The soil moisture reading reaching saturation means that the soil pores are completely filled with water and can no longer absorb any more.

[0045] Simultaneously, localized radar rainfall intensity data from a preset spatial range is used to assess the rainfall situation in that area. This preset spatial range can be set according to actual monitoring needs and radar coverage. When the localized radar rainfall intensity data is below a first threshold, for example, below the intensity value of a light rainfall or no rainfall, the possibility of increased soil moisture directly caused by widespread heavy rainfall is ruled out. The first threshold can be calibrated based on local climate characteristics and hydrological models to distinguish between localized runoff and direct rainfall.

[0046] Furthermore, the status of the latest satellite remote sensing data is used to help confirm the occurrence of localized sudden hydrological events. Satellite remote sensing data can provide macroscopic ground cover information, such as changes in surface water bodies, abnormal vegetation indices, or cloud cover conditions. This information can supplement and verify ground sensor and radar data, especially in areas with sparse sensor coverage or uncertain data. For example, satellite remote sensing images may show surface runoff or small-scale water accumulation, thereby further supporting the identification of localized sudden hydrological events.

[0047] Through the aforementioned technical solution, this application can more accurately and timely identify localized sudden hydrological events. This multi-source data cross-validation mechanism effectively avoids misjudgments that may arise from a single data source, such as misclassifying large-scale rainfall as a localized event or sensor malfunction as a hydrological event. Especially in scenarios where radar rainfall intensity is low but soil moisture is abnormally high, this technical solution can accurately capture localized waterlogging or flooding caused by surface runoff, groundwater overflow, or other indirect rainfall factors, thereby providing more accurate event input for subsequent drought forecasting and agricultural management, significantly improving the reliability and relevance of forecasts.

[0048] In some embodiments, the step of determining the actual impact range of the local sudden hydrological event after confirming its occurrence, by combining the geographical environment information of the event location with the event's characteristic information, and through a physical deduction process, to obtain the delineation result of the actual impact range includes: Acquire micro-topographic data, soil stratification information, and real-time rainfall intensity data of the area where the event occurred; The system continuously receives data from stratified soil moisture sensors and infers and updates the local hydrological parameters required for the integrated hydrological model based on the data. The stratified soil moisture sensors are deployed at different depths in the same location to provide vertical profile information of soil moisture. An integrated hydrological model that updates local hydrological parameters is activated to simulate the movement of water in the soil, and the area where the soil moisture content reaches saturation is marked as the actual area of ​​influence.

[0049] To accurately determine the actual impact range of localized sudden hydrological events through physical simulation, it is necessary to acquire micro-topographic data, soil stratification information, and real-time rainfall intensity data of the event area. Specifically, micro-topographic data describes subtle surface undulations, which significantly influence surface runoff and water infiltration pathways; soil stratification information reveals the physicochemical properties of soil at different depths, such as porosity and permeability coefficient, which are key parameters for simulating the vertical movement of water in the soil; and real-time rainfall intensity data provides dynamic information on external water input during the event.

[0050] Furthermore, to more accurately capture the dynamic changes in soil moisture, the technical solution of this application continuously receives data from stratified soil moisture sensors. These stratified soil moisture sensors are deployed at different depths in the same location, thereby providing vertical profile information of soil moisture. Through this multi-depth data, the local hydrological parameters required for the integrated hydrological model, such as soil hydraulic conductivity and saturated hydraulic conductivity, can be more precisely inferred and updated, with the aim of ensuring that the model parameters are highly consistent with the actual soil conditions.

[0051] Based on this, an integrated hydrological model with updated local hydrological parameters was activated. This model was used to simulate the movement of water in the soil, including infiltration, vertical infiltration, lateral flow, and evaporation. Through simulation, changes in soil moisture content can be dynamically tracked. When the simulation results show that the soil moisture content in a certain area has reached saturation, that area is marked as the actual impact range of the localized sudden hydrological event. Soil saturation is a key indicator for judging the degree and extent of the impact of hydrological events, because saturated soil will lead to increased surface runoff or groundwater overflow, thereby affecting agricultural production.

[0052] The aforementioned technical solution overcomes the potential accuracy and lag issues of traditional methods in determining the impact range of hydrological events. Specifically, by incorporating micro-topographic data, soil stratification information, and real-time rainfall intensity data, combined with vertical soil moisture profile information provided by stratified soil moisture sensors, the integrated hydrological model can more precisely simulate the three-dimensional movement of water in the soil, thereby significantly improving the accuracy and real-time performance of determining the actual impact range of localized sudden hydrological events. This refined physical extrapolation process enables earlier and more accurate identification of affected areas, providing a solid data foundation for subsequent agricultural environmental assessments and differentiated recommendations.

[0053] In some embodiments, the step of adjusting the credibility of the information from different sources in different regions based on the division results, and fusing the target agricultural environmental monitoring information to generate a comprehensive agricultural environmental status distribution map includes: Continuously monitor all identified localized sudden hydrological events and obtain the classification results of each event's type, occurrence time, actual impact range, and suggested adjustments to the data's reliability. Identify the overlapping areas of the actual impact range of the event, prioritize all relevant events within the overlapping areas, and obtain the priority ranking result; Based on the suggested adjustment scheme for the data credibility corresponding to the highest priority event in the priority ranking results, the final credibility of information from different sources in the overlapping area is set; and based on the final credibility, the credibility of information from different sources in different areas is adjusted, and the information from different sources is fused to generate a comprehensive agricultural environmental status distribution map.

[0054] Specifically, all identified localized sudden hydrological events are continuously monitored, and information on these events that have occurred and been confirmed is tracked and updated without interruption. Each event type may include, but is not limited to, localized flooding, sudden heavy rainfall, irrigation system failure, or abnormal groundwater outflow. The event's occurrence time is used to assess its timeliness and potential attenuation impact. The actual impact range is defined as a geographical area determined through physical extrapolation, representing the extent of the event's direct impact on the agricultural environment. The suggested adjustments to data reliability refer to recommendations for revising the reliability of agricultural environmental monitoring information from different sources (e.g., satellite remote sensing data, ground sensor data, weather radar data, etc.) based on the nature and impact of a specific event. For example, in a localized flooding event, ground soil moisture sensor data may be considered more reliable in saturated areas, while satellite optical remote sensing data may have reduced reliability due to cloud cover.

[0055] Furthermore, by utilizing the spatial analysis functions of a Geographic Information System (GIS), areas where the actual impact ranges of two or more localized sudden hydrological events overlap geographically can be detected and identified, thus enabling the identification of overlapping regions of the actual impact ranges of these events. All relevant events within these overlapping regions are then prioritized to obtain a priority ranking result. The purpose of this approach is to address the problem of determining the dominant impact when multiple events simultaneously affect the same area. Prioritization can be based on various factors, such as the intensity of the event (e.g., rainfall, water depth), the duration of the event, the time of occurrence of the event (more recent events have higher priority), the type of event (e.g., natural disasters are generally higher than human-caused failures), and its potential damage to agricultural production. For example, a scoring mechanism can be established to calculate the priority score for each event by comprehensively considering the above factors, with the event having the highest score receiving the highest priority.

[0056] Therefore, based on the suggested adjustment scheme for the data credibility of the highest-priority event in the priority ranking results, the final credibility of information from different sources within the overlapping area is set. This means that within the overlapping area, the data credibility adjustment scheme suggested by the highest-priority event will be adopted as the final credibility of all information from different sources within that area. This mechanism ensures the logical consistency and authority of data credibility adjustment in complex multi-event scenarios, avoiding conflicts between suggestions from different events. Finally, based on the final credibility, the credibility of the information from different sources in different areas is adjusted, and the information from different sources is fused to generate a comprehensive agricultural environmental status distribution map. The fusion process can employ techniques such as weighted averaging, Kalman filtering, and data assimilation. The weights of different sources will be dynamically allocated according to their adjusted credibility, thereby generating a more accurate and comprehensive agricultural environmental status distribution map.

[0057] Through the aforementioned technical solution, this application significantly improves the accuracy and robustness of agricultural environmental monitoring data fusion in complex scenarios where multiple localized sudden hydrological events may occur and their impact areas may overlap. By continuously monitoring, identifying overlapping areas, and prioritizing data, the conflicting issues related to data credibility adjustments can be effectively resolved, ensuring that data fusion is based on the most reasonable and authoritative event impact assessment in any region. This makes the generated comprehensive agricultural environmental status distribution map more accurate and reliable, thus providing a more solid data foundation for subsequent agricultural drought forecasting and differentiated recommendations. It effectively reduces decision-making risks caused by information uncertainty and improves the scientific and refined level of agricultural production management.

[0058] In some embodiments, the step of generating agricultural environmental assessment results and differentiated recommendations for different regions based on the agricultural environmental status distribution map includes: Obtain information on crop types, crop growth stages, soil nutrient levels, and agricultural machinery operation capabilities in the region; Based on the crop type information and the crop growth stage information, the crop water requirement characteristics database is queried to obtain the range and sensitivity of the crop's water requirement at the current growth stage, and the crop water requirement characteristics are obtained. Based on the crop type information and the crop growth stage information, the crop nutrient absorption characteristic database is queried to obtain the crop's demand for various nutrients and absorption efficiency at the current growth stage, thus obtaining the crop nutrient absorption characteristics. Based on the soil nutrient level information, analyze the soil test data, identify the content of major nutrients in the soil, and compare the content of major nutrients in the soil with the crop's requirements for various nutrients at the current growth stage to obtain the soil nutrient level. Based on the agricultural machinery operation capacity information, query the available agricultural machinery equipment types, operation efficiency, and operation cost to obtain the agricultural machinery operation capacity; Based on the moisture information in the agricultural environment distribution map, combined with crop water requirements, crop nutrient absorption characteristics, soil nutrient levels, and agricultural machinery operation capabilities, differentiated recommendations for irrigation, fertilization, and drainage operations are generated, and the expected resource consumption and potential benefit assessment of each recommendation are provided.

[0059] Specifically, this involves acquiring information on crop types, crop growth stages, soil nutrient levels, and agricultural machinery operational capabilities for a given region, collecting detailed data related to that specific agricultural area through various channels. Crop type information can include the main crop types grown in the region, such as rice, wheat, and corn. Crop growth stage information refers to the specific growth stage of the crop, such as sowing, seedling, jointing, flowering, and grain-filling stages; this information can be obtained through remote sensing monitoring, field surveys, or reports from farmers. Soil nutrient level information can be obtained through regular soil testing, including the content of major macroelements such as nitrogen, phosphorus, and potassium, as well as microelements. Agricultural machinery operational capability information covers the types of agricultural machinery available in the region, their operating efficiency, operating costs, and applicable scope; this information can be obtained from agricultural machinery management departments or agricultural machinery service providers.

[0060] Furthermore, based on crop species and growth stage information, the crop water requirement database is queried to obtain the range and sensitivity of the crop's water requirements at the current growth stage. The crop water requirement database is a pre-established knowledge base containing the water requirement patterns of different crops at different growth stages. For example, it can store data such as transpiration rates, optimal soil moisture content ranges, and yield loss rates under water stress for different crops at different growth stages. By querying this database, the crop water requirement characteristics can be obtained, providing a quantitative basis for subsequent water management.

[0061] Simultaneously, based on crop type and growth stage information, the crop nutrient absorption characteristics database is queried to obtain the crop's nutrient requirements and absorption efficiency at the current growth stage. Similar to the water requirement database, this database stores information on the absorption patterns of major nutrients such as nitrogen, phosphorus, and potassium by different crops at different growth stages, optimal nutrient supply levels, and typical symptoms of nutrient deficiency. By querying this database, crop nutrient absorption characteristics can be obtained, guiding precision fertilization.

[0062] Furthermore, based on soil nutrient level information, soil testing data is analyzed to identify the content of major nutrients in the soil. Then, the content of major nutrients in the soil is compared with the crop's theoretical nutrient requirements at the current growth stage to determine the soil nutrient level. This step, by comparing the actual soil nutrient content with the crop's theoretical requirements, accurately determines whether the soil nutrient supply meets the crop's growth needs, thus providing a scientific basis for fertilization decisions.

[0063] Furthermore, based on agricultural machinery operation capacity information, the types of available agricultural machinery, their operating efficiency, and operating costs are searched to determine the agricultural machinery operation capacity. This step aims to assess which agricultural machinery is available in a specific area, their operating efficiency, and the cost of using these machines, thereby ensuring that the proposed agricultural operation suggestions are feasible and economical in practice.

[0064] Finally, based on the moisture information in the agricultural environmental condition distribution map, combined with crop water requirements, crop nutrient absorption characteristics, soil nutrient levels, and agricultural machinery operation capabilities, differentiated recommendations for irrigation, fertilization, and drainage are generated, along with an assessment of the expected resource consumption and potential benefits for each recommendation. This step is the core of the entire technical solution; by comprehensively considering multiple factors, it can generate highly customized agricultural operation recommendations. For example, when the agricultural environmental condition distribution map shows insufficient soil moisture in a certain area, it will determine whether irrigation is needed based on crop water requirements and recommend suitable irrigation methods and equipment based on agricultural machinery operation capabilities. When soil nutrient analysis shows a deficiency of a certain nutrient, it will recommend a fertilization plan based on crop nutrient absorption characteristics and agricultural machinery operation capabilities. Simultaneously, each recommendation includes expected resource consumption (such as water consumption, fertilizer consumption, and operation time) and potential benefit assessments (such as expected yield increase and cost savings) to help farmers make more informed decisions.

[0065] Through the aforementioned technical solutions, this application significantly enhances the precision and practicality of the assessment results and recommendations provided by agricultural drought prediction methods. By fully considering key agricultural production factors such as crops, soil, and agricultural machinery, the generated differentiated recommendations not only accurately match the actual agricultural needs of the region, avoiding resource waste caused by indiscriminate irrigation or fertilization, but also provide actionable solutions based on the operational capabilities of agricultural machinery, thereby effectively guiding farmers in scientific agricultural management. This refined and personalized recommendation helps improve agricultural production efficiency, reduce production costs, and enhance agriculture's ability to cope with drought, ultimately achieving sustainable agricultural development.

[0066] In some embodiments, the step of generating agricultural environmental assessment results and differentiated recommendations for different regions based on the agricultural environmental status distribution map includes: Historical hydrological event records for the region are obtained. Based on these records and the soil moisture dissipation patterns, historical hydrological events with residual effects are identified. The residual impact of these events on the current water conditions is calculated based on the type and intensity of the identified events. Identify newly occurring localized sudden hydrological events in the area, and calculate the immediate impact of the newly occurring localized sudden hydrological events on the current water conditions based on the type and intensity of the identified events. The residual effect and the immediate effect are superimposed or offset to obtain the comprehensive moisture status; Based on the comprehensive moisture conditions, the agricultural environmental assessment results and differentiated recommendations are generated.

[0067] Specifically, historical hydrological event records for the region are acquired by collecting and storing data on all hydrological events that occurred in the region over a past period, such as rainfall, floods, and droughts, including their occurrence time, duration, intensity, and impact range. Soil moisture dissipation can be understood as the natural process of soil moisture decreasing over time, with its rate influenced by various factors such as soil type, evaporation, and plant uptake. By combining these historical records and dissipation patterns, we can identify historical hydrological events whose impacts have not yet been completely eliminated. Based on their type (e.g., torrential rain, continuous rainfall) and intensity (e.g., rainfall amount), we can calculate their persistent impact on current soil moisture conditions, i.e., the residual impact of historical hydrological events on current moisture levels.

[0068] By monitoring real-time data, such as soil moisture sensor data, radar rainfall data, and satellite remote sensing data, localized and sudden hydrological events occurring at the current time or within a very short period of time can be promptly detected and confirmed, such as localized heavy rainfall or localized waterlogging. Based on the type and intensity of these newly occurring localized sudden hydrological events, their immediate impact on the current soil moisture status can be calculated.

[0069] In practical applications, the residual and immediate impacts are superimposed or offset. Based on the residual impacts of historical hydrological events and the immediate impacts of newly occurring localized sudden hydrological events, appropriate mathematical models or algorithms are used to comprehensively consider both. For example, if a historical hydrological event leads to higher soil moisture levels, while a newly occurring localized sudden hydrological event results in localized drought, offsetting effects may occur; if both lead to higher moisture levels, superposition may occur. The aim is to obtain a more comprehensive and accurate overall water situation that reflects the true dynamics of water within the region.

[0070] Therefore, based on the comprehensive water situation, the agricultural environment assessment results and differentiated recommendations are generated. Using this refined comprehensive water situation, combined with other agricultural environment information such as crop water requirements and soil nutrient levels, a more accurate drought assessment is conducted, and more targeted agricultural operation recommendations are provided, such as irrigation amount, drainage measures, and fertilization plans.

[0071] Through the aforementioned technical solution, this application can more comprehensively and accurately assess the actual water conditions of a region, overcoming the assessment biases that may arise from traditional methods failing to adequately consider the residual effects of historical hydrological events and the immediate impact of newly occurring localized sudden hydrological events. Consequently, the generated agricultural drought forecasts and differentiated agricultural recommendations will be more timely and accurate, helping agricultural producers to take timely and effective countermeasures, optimize water resource management, reduce drought risks, and improve crop yields and water resource utilization efficiency. This refined water condition assessment method significantly enhances the reliability and practical value of agricultural drought forecasting.

[0072] For example, suppose an agricultural area experienced a moderate rainfall event in the past week. Based on historical hydrological records and the area's soil type (e.g., clay, which dissipates water slowly), a soil moisture dissipation model calculates that the rainfall event still has a 20% residual impact on current soil moisture. Simultaneously, in the last few hours, local soil moisture sensors and radar rainfall data detected a short-term heavy rainfall in a small plot within the area, causing a sharp increase in soil moisture. The immediate impact of this newly occurring localized sudden hydrological event on the current moisture status is calculated to be 30%. In this case, the 20% residual impact and the 30% immediate impact are combined to obtain a 50% overall improvement in moisture status. Based on this more accurate overall moisture status, combined with information such as crop growth stage and water requirements, more specific irrigation recommendations are generated. For example, for affected plots, it is recommended to suspend irrigation or reduce irrigation volume, while for unaffected or minimally affected plots, regular irrigation recommendations are given based on their overall moisture status. This approach ensures a dynamic and precise understanding of the water conditions within the region, thereby providing more practical agricultural guidance.

[0073] In some embodiments, the step of superimposing or offsetting the residual effect and the immediate effect to obtain a comprehensive moisture status includes: The weight of the historical hydrological events on the current water conditions is determined based on the intensity and duration of the historical hydrological events. The weight of a newly occurring localized sudden hydrological event on the current water status is determined based on the intensity and duration of the event. Based on the weights of the historical hydrological events on the current water situation and the weights of the newly occurring localized sudden hydrological events on the current water situation, the residual impact and the immediate impact are weighted to obtain the comprehensive water situation.

[0074] Specifically, the intensity of historical hydrological events can be understood as the degree of impact the event had on the regional water situation at the time of its occurrence, such as rainfall, flood levels, and drought persistence index, while the duration refers to the length of time from the event's occurrence to the point where its impact largely dissipates. The intensity and duration of newly occurring localized sudden hydrological events have similar definitions. Based on the intensity and duration of historical hydrological events, their relative importance or influence on the regional water situation at the current point in time is assessed, thus determining the weight of historical hydrological events on the current water situation. For example, historical events with greater intensity and longer duration may have a higher weight for their residual impact. Similarly, the weight of newly occurring localized sudden hydrological events on the current water situation is assessed based on their intensity and duration to evaluate the relative importance of their immediate impact. For example, a short-duration heavy rainfall event may have a high weight for its immediate impact. In practical applications, these weights can be calculated and adjusted using pre-set function models, expert experience, or machine learning algorithms to reflect the differences in the impact of different event characteristics on the water situation. Finally, by multiplying the residual impact and immediate impact by their corresponding weights and then summing or canceling them out, a more accurate comprehensive water situation is obtained.

[0075] Through the aforementioned technical solution, this application can more accurately assess the overall water situation of a region. Compared to simple superposition or offsetting without distinguishing event intensity and duration, this application introduces a weighting mechanism to reasonably weight the residual impact of historical hydrological events and the immediate impact of newly occurring localized sudden hydrological events according to their actual magnitude, thereby avoiding assessment bias caused by differences in event characteristics. As a result, the obtained overall water situation can more realistically reflect the actual hydrological state of the region, providing a more reliable and refined data foundation for subsequent agricultural drought forecasting, environmental assessment, and the generation of differentiated recommendations, significantly improving the accuracy and practicality of forecasts and recommendations.

[0076] In some embodiments, the step of determining the weight of the historical hydrological event on the current water condition based on the intensity and duration of the historical hydrological event includes: Acquire the type, intensity, duration of the historical hydrological events, as well as soil type and topographic features of the event locations; Select the initial decay mode based on the type of the historical hydrological events; Based on the intensity and duration of the historical hydrological events, and combined with soil type and topographic feature information of the event location, the parameters of the initial attenuation mode are adjusted to obtain the adjusted attenuation mode. Based on the adjusted attenuation pattern, calculate the attenuation coefficient of the historical hydrological event on the water status at the current time point; The weight of the historical hydrological events on the current water conditions is determined based on the attenuation coefficient.

[0077] Specifically, when acquiring information on the type, intensity, duration, soil type, and topographic features of historical hydrological events, the types of these events can include, but are not limited to, floods, localized waterlogging, and prolonged rainfall. Their intensity and duration can be obtained from historical monitoring data or event reports. Soil type information for the event area can refer to soil texture (e.g., sandy soil, loam, clay), permeability, and water-holding capacity. Topographic features can include slope, aspect, elevation, and drainage conditions. This information can be obtained from Geographic Information System (GIS) databases, soil survey data, or high-precision topographic maps.

[0078] Among these, the physical processes and timescales of water dissipation differ significantly for different types of hydrological events. For example, flood events may lead to widespread surface runoff and deep infiltration, and their attenuation pattern may focus more on slow, long-term dissipation; while localized waterlogging events may mainly affect the topsoil, and their attenuation pattern may exhibit faster initial dissipation. Therefore, multiple initial attenuation models can be preset, such as exponential attenuation models, power-law attenuation models, or piecewise linear attenuation models, and matched and selected according to the identified types of historical hydrological events.

[0079] In practical applications, the parameters of the initial decay model are adjusted based on the intensity and duration of historical hydrological events, combined with soil type and topographic features of the event area, to obtain the adjusted decay model. For example, for a selected initial decay model, its parameters (such as decay rate and half-life) can be initially adjusted based on the event's intensity (such as peak flow and water depth) and duration (such as inundation duration). Furthermore, these parameters are refined by considering soil type (e.g., clay has strong water retention capacity and slow decay; sandy soil infiltrates quickly and decays rapidly) and topographic features (e.g., areas with steep slopes drain quickly and decay rapidly; low-lying areas are prone to water accumulation and decay slowly), to more accurately simulate the actual dissipation process of water under specific conditions.

[0080] Therefore, based on the adjusted attenuation pattern, the attenuation coefficient of historical hydrological events on the current water situation is calculated. This attenuation coefficient is a value between 0 and 1, reflecting the degree of residual impact of the historical event on the water situation at the current moment. For example, if the attenuation coefficient is 0.2, it means that the residual impact of the historical event is 20% of the original impact. Finally, based on the attenuation coefficient, the weight of the historical hydrological event on the current water situation is determined. This weight can be simply equal to the attenuation coefficient, or the attenuation coefficient can be converted into a weight value more suitable for weighted processing through a mapping function.

[0081] Through the aforementioned technical solution, this application significantly improves the accuracy and precision of determining the weights of historical hydrological events on current water conditions. By incorporating event type, soil type, and topographic feature information, and employing an adjustable attenuation mode, the weight calculation is no longer a simple linear or fixed model, but rather dynamically adaptable to actual environmental conditions. This avoids weight biases caused by ignoring geographical differences and event type characteristics, thus making the assessment of the residual impacts of historical hydrological events more scientific and reasonable. Ultimately, this more accurate weight determination method provides a more reliable data foundation for the comprehensive assessment of water conditions, thereby improving the accuracy and reliability of the entire agricultural drought prediction method and providing more precise decision support for agricultural production.

[0082] In some embodiments, the step of determining the weight of the newly occurring localized sudden hydrological event on the current water condition based on the intensity and duration of the newly occurring localized sudden hydrological event includes: Obtain the sequence information of the newly occurring localized sudden hydrological events, the sequence information including the occurrence time, intensity, and duration of each event; Based on the sequence information, the spatiotemporal correlation between events in the event sequence is identified to obtain spatiotemporal correlation information; Based on the spatiotemporal correlation information, calculate the cumulative impact factor of each event in the event sequence; Based on the cumulative impact factor, the weight of newly occurring localized sudden hydrological events on the current water status is determined.

[0083] Specifically, the sequence information of newly occurring localized sudden hydrological events refers to the set of all identified localized sudden hydrological events within a specific monitoring period. This set includes detailed records of each event, such as its precise occurrence time, intensity level (e.g., rainfall, runoff velocity), and duration. Obtaining this sequence information aims to provide a comprehensive data foundation for subsequent analysis. Identifying the spatiotemporal correlations between events in the sequence can be understood as analyzing the occurrence time, geographical location, and impact range of events to determine whether there is temporal continuity, spatial proximity, or causal relationship between different events. For example, if two events occur consecutively in adjacent areas within a very short period, a strong spatiotemporal correlation may exist. This spatiotemporal correlation information aims to reveal the interaction patterns between events. In practical applications, calculating the cumulative impact factor of each event in the sequence specifically refers to quantifying the comprehensive impact of the superposition or interaction of multiple localized sudden hydrological events on water conditions based on the identified spatiotemporal correlation information. For example, if a previous event leads to soil saturation, even a subsequent event of low intensity may rapidly trigger runoff, significantly increasing its cumulative impact factor. The aim is to more accurately reflect the overall effect of a group of events. Therefore, based on the aforementioned cumulative impact factor, the weight of newly occurring localized sudden hydrological events on the current water situation is determined. This aims to quantify the comprehensive impact, which considers the interactions between events, into weights, so as to more accurately reflect the contribution of newly occurring events to the overall water situation in subsequent superposition or offsetting processes.

[0084] By employing the aforementioned technical solution, this application overcomes the limitations of traditional methods that simply determine weights based on event intensity and duration. This is particularly evident when dealing with complex and variable sequences of localized sudden hydrological events, enabling a more accurate assessment of their actual impact on agricultural water conditions. By introducing event sequence information, spatiotemporal correlation identification, and cumulative impact factor calculation, this application achieves a more comprehensive and in-depth understanding and quantification of the combined effects of multiple events. This allows for a more reasonable and precise determination of the weights of newly occurring localized sudden hydrological events on the current water situation. This improvement not only enhances the accuracy of comprehensive water condition assessment but also provides a more reliable data foundation for generating subsequent agricultural environmental assessment results and differentiated recommendations for different regions, thereby improving the overall accuracy and guiding value of agricultural drought forecasting.

[0085] This application also proposes an agricultural drought prediction system based on multi-source heterogeneous data assimilation, such as... Figure 2 As shown, an agricultural drought prediction system 100 based on multi-source heterogeneous data assimilation is provided. The system includes: The information receiving and processing module 10 is used to acquire agricultural environmental monitoring information from different sources and preprocess the agricultural environmental monitoring information to obtain target agricultural environmental monitoring information. The event identification and confirmation module 20 is used to identify and confirm the occurrence of local sudden hydrological events based on specific difference patterns between information from different sources in the target agricultural environmental monitoring information. The impact range determination module 30 is used to determine the actual impact range of the local sudden hydrological event after confirming that the event has occurred, by combining the geographical environment information of the event occurrence area and the event's characteristic information, and through a physical deduction process, and to obtain the division result of the actual impact range. The assessment recommendation generation module 40 is used to adjust the credibility of the information from different sources in different regions according to the division results, and to integrate the target agricultural environment monitoring information to generate a comprehensive agricultural environment status distribution map; and to generate agricultural environment assessment results and differentiated recommendations for different regions based on the agricultural environment status distribution map.

[0086] The system proposed in this application aims to address the contradictions and misjudgments that arise in traditional agricultural drought forecasting when dealing with multi-source heterogeneous data and localized sudden hydrological events. By decomposing the complex forecasting process into collaborative components such as an information receiving and processing module, an event identification and confirmation module, an impact range determination module, and an assessment and recommendation generation module, the system achieves intelligent assimilation and dynamic fusion of agricultural environmental monitoring information. The close cooperation among these modules ensures efficient operation throughout the entire chain, from raw data acquisition to final recommendation generation, thereby significantly improving the accuracy, timeliness, and relevance of agricultural drought forecasting and providing more reliable decision support for agricultural production.

[0087] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for predicting agricultural drought based on multi-source heterogeneous data assimilation, characterized in that, include: Agricultural environmental monitoring information from different sources is acquired and preprocessed to obtain target agricultural environmental monitoring information. Based on specific difference patterns between different sources of information in the target agricultural environmental monitoring information, the occurrence of local sudden hydrological events is identified and confirmed; After confirming the occurrence of the localized sudden hydrological event, the actual impact range of the localized sudden hydrological event is determined through a physical deduction process, based on the geographical environment information of the event location and the event's characteristic information, thus obtaining the division result of the actual impact range. Based on the division results, the credibility of the information from different sources in different regions is adjusted, and the target agricultural environmental monitoring information is fused to generate a comprehensive agricultural environmental status distribution map. Based on the aforementioned agricultural environment distribution map, agricultural environment assessment results and differentiated recommendations are generated for different regions.

2. The agricultural drought prediction method based on multi-source heterogeneous data assimilation according to claim 1, characterized in that, The steps for identifying and confirming the occurrence of localized sudden hydrological events based on specific difference patterns between different sources of information in the target agricultural environmental monitoring information include: When at least one soil moisture sensor is detected to have its soil moisture reading rise sharply to saturation within a preset short period of time, and at the same time period, the local radar rainfall intensity data from a preset spatial range is lower than a first threshold, the occurrence of a local sudden hydrological event is confirmed by combining the latest satellite remote sensing data.

3. The agricultural drought prediction method based on multi-source heterogeneous data assimilation according to claim 1, characterized in that, The steps of determining the actual impact range of the local sudden hydrological event after confirming its occurrence, by combining the geographical environment information of the event location with the event's characteristic information, and through a physical deduction process, to obtain the delineation result of the actual impact range include: Acquire micro-topographic data, soil stratification information, and real-time rainfall intensity data of the area where the event occurred; The system continuously receives data from stratified soil moisture sensors and infers and updates the local hydrological parameters required for the integrated hydrological model based on the data. The stratified soil moisture sensors are deployed at different depths in the same location to provide vertical profile information of soil moisture. An integrated hydrological model that updates local hydrological parameters is activated to simulate the movement of water in the soil, and the area where the soil moisture content reaches saturation is marked as the actual area of ​​influence.

4. The agricultural drought prediction method based on multi-source heterogeneous data assimilation according to claim 1, characterized in that, The steps of adjusting the credibility of information from different sources in different regions based on the division results, and fusing the target agricultural environmental monitoring information to generate a comprehensive agricultural environmental status distribution map include: Continuously monitor all identified localized sudden hydrological events and obtain the classification results of each event's type, occurrence time, actual impact range, and suggested adjustments to the data's reliability. Identify the overlapping areas of the actual impact range of the event, prioritize all relevant events within the overlapping areas, and obtain the priority ranking result; Based on the suggested adjustment scheme for the data credibility corresponding to the highest priority event in the priority ranking results, the final credibility of information from different sources in the overlapping area is set; and based on the final credibility, the credibility of information from different sources in different areas is adjusted, and the information from different sources is fused to generate a comprehensive agricultural environmental status distribution map.

5. The agricultural drought prediction method based on multi-source heterogeneous data assimilation according to claim 1, characterized in that, The steps for generating agricultural environment assessment results and differentiated recommendations for different regions based on the agricultural environment status distribution map include: Obtain information on crop types, crop growth stages, soil nutrient levels, and agricultural machinery operation capabilities in the region; Based on the crop type information and the crop growth stage information, the crop water requirement characteristics database is queried to obtain the range and sensitivity of the crop's water requirement at the current growth stage, and the crop water requirement characteristics are obtained. Based on the crop type information and the crop growth stage information, the crop nutrient absorption characteristic database is queried to obtain the crop's demand for various nutrients and absorption efficiency at the current growth stage, thus obtaining the crop nutrient absorption characteristics. Based on the soil nutrient level information, analyze the soil test data, identify the content of major nutrients in the soil, and compare the content of major nutrients in the soil with the crop's requirements for various nutrients at the current growth stage to obtain the soil nutrient level. Based on the agricultural machinery operation capacity information, query the available agricultural machinery equipment types, operation efficiency, and operation cost to obtain the agricultural machinery operation capacity; Based on the moisture information in the agricultural environment distribution map, combined with crop water requirements, crop nutrient absorption characteristics, soil nutrient levels, and agricultural machinery operation capabilities, differentiated recommendations for irrigation, fertilization, and drainage operations are generated, and the expected resource consumption and potential benefit assessment of each recommendation are provided.

6. The agricultural drought prediction method based on multi-source heterogeneous data assimilation according to claim 1, characterized in that, The steps for generating agricultural environment assessment results and differentiated recommendations for different regions based on the agricultural environment status distribution map include: Historical hydrological event records for the region are obtained. Based on these records and the soil moisture dissipation patterns, historical hydrological events with residual effects are identified. The residual impact of these events on the current water conditions is calculated based on the type and intensity of the identified events. Identify newly occurring localized sudden hydrological events in the area, and calculate the immediate impact of the newly occurring localized sudden hydrological events on the current water conditions based on the type and intensity of the identified events. The residual effect and the immediate effect are superimposed or offset to obtain the comprehensive moisture status; Based on the comprehensive moisture conditions, the agricultural environmental assessment results and differentiated recommendations are generated.

7. The agricultural drought prediction method based on multi-source heterogeneous data assimilation according to claim 6, characterized in that, The step of superimposing or offsetting the residual effect and the immediate effect to obtain the comprehensive moisture status includes: The weight of the historical hydrological events on the current water conditions is determined based on the intensity and duration of the historical hydrological events. The weight of a newly occurring localized sudden hydrological event on the current water status is determined based on the intensity and duration of the event. Based on the weights of the historical hydrological events on the current water situation and the weights of the newly occurring localized sudden hydrological events on the current water situation, the residual impact and the immediate impact are weighted to obtain the comprehensive water situation.

8. The agricultural drought prediction method based on multi-source heterogeneous data assimilation according to claim 7, characterized in that, The step of determining the weight of the historical hydrological event on the current water condition based on the intensity and duration of the historical hydrological event includes: Acquire the type, intensity, duration of the historical hydrological events, as well as soil type and topographic features of the event locations; Select the initial decay mode based on the type of the historical hydrological events; Based on the intensity and duration of the historical hydrological events, and combined with soil type and topographic feature information of the event location, the parameters of the initial attenuation mode are adjusted to obtain the adjusted attenuation mode. Based on the adjusted attenuation pattern, calculate the attenuation coefficient of the historical hydrological event on the water status at the current time point; The weight of the historical hydrological events on the current water conditions is determined based on the attenuation coefficient.

9. The agricultural drought prediction method based on multi-source heterogeneous data assimilation according to claim 7, characterized in that, The step of determining the weight of a newly occurring localized sudden hydrological event on the current water situation based on the intensity and duration of the event includes: Obtain the sequence information of the newly occurring localized sudden hydrological events, the sequence information including the occurrence time, intensity, and duration of each event; Based on the sequence information, the spatiotemporal correlation between events in the event sequence is identified to obtain spatiotemporal correlation information; Based on the spatiotemporal correlation information, calculate the cumulative impact factor of each event in the event sequence; Based on the cumulative impact factor, the weight of newly occurring localized sudden hydrological events on the current water status is determined.

10. An agricultural drought prediction system based on multi-source heterogeneous data assimilation, characterized in that, The system includes: The information receiving and processing module is used to acquire agricultural environmental monitoring information from different sources and preprocess the agricultural environmental monitoring information to obtain target agricultural environmental monitoring information. The event identification and confirmation module is used to identify and confirm the occurrence of local sudden hydrological events based on specific difference patterns between information from different sources in the target agricultural environmental monitoring information. The impact range determination module is used to determine the actual impact range of the local sudden hydrological event after confirming its occurrence, by combining the geographical environment information of the event location and the event's characteristic information, and through a physical deduction process, thus obtaining the actual impact range division result. The assessment suggestion generation module is used to adjust the credibility of information from different sources in different regions based on the division result, and to fuse the target agricultural environmental monitoring information to generate a comprehensive agricultural environmental status distribution map; and based on the agricultural environmental status distribution map, to generate agricultural environmental assessment results and differentiated suggestions for different regions.

Images