A method for simulating soil relative humidity based on meteorological index

Abstract

The application discloses a soil relative humidity simulation method based on a meteorological index, relates to the technical field of soil moisture simulation application, and comprises the following steps: obtaining daily meteorological data of a target region and calculating a previous precipitation evapotranspiration index; based on the S-shaped change characteristic of a soil moisture absorption process, a logistic function relationship between the previous precipitation evapotranspiration index and soil relative humidity is constructed; parameters of the logistic function are fitted by using soil relative humidity observation samples, and a soil relative humidity meteorological simulation model for outputting soil relative humidity simulation results is obtained; and the application solves the problems that the existing SPEI, SAPEI and other drought indexes are difficult to accurately reflect daily dynamic changes of farmland soil moisture and the inconsistency between a drought grade and actual soil moisture conditions of agriculture.

Description

Technical Field

[0001] This invention relates to the field of soil moisture simulation application technology, and more specifically, to a method for simulating soil relative humidity based on meteorological indices. Background Technology

[0002] To enhance drought disaster monitoring and early warning capabilities, numerous studies and applications of drought indices have been conducted using readily available data such as temperature and precipitation. In global and regional drought monitoring and analysis, the Palmer Drought Severity Index (PDSI), the Standardized Precipitation Index (SPI), and the Standardized Precipitation Evapotranspiration Index (SPEI), which incorporates a potential evapotranspiration term into the SPI, are widely used. The SPEI index possesses multi-timescale characteristics and the ability to consider temperature effects. These indices are typically suitable for monthly (seasonal) drought assessments and may not fully reflect daily variations in drought and flood conditions. When significant abrupt shifts between drought and flood occur within a statistical period, if precipitation rapidly moves out of the statistical window, it may lead to certain biases in the drought and flood characterization results. To enhance the diurnal dynamic monitoring capability of drought meteorological indices, the precipitation evapotranspiration statistics in the SPEI index were improved by replacing them with accumulated precipitation evapotranspiration from previous periods based on diurnal processes. A three-parameter log-logistic probability distribution was then used for fitting, constructing a standardized antecedent precipitation evapotranspiration index (SAPEI) suitable for diurnal scales. However, indices such as SPEI and SAPEI rely entirely on climatic probabilities to classify drought levels, leading to inconsistencies between the drought levels they represent and agricultural drought levels in practical applications. Summary of the Invention

[0003] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a soil relative humidity simulation method based on meteorological indices. By constructing a logistic relative humidity simulation method based on the evapotranspiration index of previous precipitation and the S-shaped change characteristics of soil moisture absorption, the present invention solves the problems that existing drought indices such as SPEI and SAPEI cannot accurately reflect the daily dynamic changes of farmland soil moisture and that the drought level is inconsistent with the actual soil moisture in agriculture.

[0004] To achieve the above objectives, the present invention provides the following technical solution: This application provides a method for simulating soil relative humidity based on meteorological indices. The method includes: acquiring daily meteorological data of the target area and calculating the evapotranspiration index of the previous precipitation; constructing a logistic function relationship between the previous precipitation evapotranspiration index and soil relative humidity based on the S-shaped variation characteristics of the soil moisture absorption process; and fitting the parameters of the logistic function using soil relative humidity observation samples to obtain a soil relative humidity meteorological simulation model for outputting soil relative humidity simulation results.

[0005] In one embodiment, acquiring daily meteorological data of the target area and calculating the evapotranspiration index of the previous precipitation includes: acquiring daily meteorological data of the target area and performing data preprocessing; calculating reference evapotranspiration based on the preprocessed daily meteorological data; calculating potential crop evapotranspiration based on the crop coefficient and the reference evapotranspiration; and calculating the evapotranspiration index of the previous precipitation based on the daily precipitation and the potential crop evapotranspiration.

[0006] In one embodiment, based on the S-shaped variation characteristics of the soil moisture absorption process, a logistic function relationship between the anterior precipitation evapotranspiration index (APEI) and soil relative humidity is constructed, including: acquiring the anterior precipitation evapotranspiration index and the soil relative humidity observation data of the target area during the same period; selecting the logistic function as the nonlinear mapping model between APEI and soil relative humidity based on the S-shaped variation characteristics of the soil moisture absorption process; and constructing the logistic function relationship between the anterior precipitation evapotranspiration index and the soil relative humidity meteorological index as follows.

[0007] In the formula, The relative humidity of soil is a meteorological index. This represents the maximum relative humidity of the soil. These are parameters for adjusting the shape of the logistic function. The growth rate parameter of the logistic function. This represents the evapotranspiration index of previous precipitation.

[0008] In one embodiment, based on the S-shaped change characteristics of the soil moisture absorption process, the method further includes: acquiring the previous precipitation evapotranspiration index sequence and the corresponding soil relative humidity observation sequence of the target area; sorting the sequences according to the previous precipitation evapotranspiration index in ascending order to construct a response relationship sequence; calculating the soil relative humidity change rate corresponding to adjacent intervals in the response relationship sequence and performing smoothing; dividing the soil moisture absorption process into an initial slow water absorption stage, a rapid humidification stage, and a saturated stable stage based on the continuous change trend of the smoothed soil relative humidity change rate; calculating the average change rate of each stage respectively, and determining that the soil moisture change process exhibits an S-shaped change characteristic based on the fact that the average change rate of the initial slow water absorption stage is less than the average change rate of the rapid humidification stage and the average change rate of the rapid humidification stage is greater than the average change rate of the saturated stable stage.

[0009] In one embodiment, the parameters of the logistic function are fitted using soil relative humidity observation samples to obtain a soil relative humidity meteorological simulation model for outputting soil relative humidity simulation results. This includes: inputting samples of previous precipitation evapotranspiration index into the logistic function to obtain simulated soil relative humidity values; calculating the fitting residuals based on the simulated and observed soil relative humidity values, and identifying abnormal and normal samples based on the statistical characteristics of the residuals; assigning different fitting weights to abnormal and normal samples respectively, and constructing a weighted least squares objective function; using an iterative optimization method to solve for the logistic function parameters that minimize the weighted least squares objective function, thus obtaining the soil relative humidity meteorological simulation model; and inputting real-time or forecasted meteorological data into the soil relative humidity meteorological simulation model to obtain the soil relative humidity simulation results for the target area.

[0010] In one embodiment, different fitting weights are assigned to abnormal samples and normal samples respectively, including: calculating the absolute value of the residual corresponding to each sample based on the identified abnormal samples and normal samples; assigning a constant first fitting weight to normal samples; assigning a second fitting weight to abnormal samples that decreases continuously as the absolute value of the residual increases; and constructing a sample weight sequence corresponding one-to-one with the samples based on the first fitting weight and the second fitting weight, which is used to adjust the contribution of different samples in fitting the logistic function parameters.

[0011] In one embodiment, a quadratic weighted Kappa coefficient is used to verify the consistency between the simulated soil relative humidity results and the measured soil relative humidity levels. This includes: dividing soil moisture into multiple levels based on the simulated and measured data; encoding the simulated and measured data into levels and constructing a confusion matrix between the simulated and measured levels; calculating a quadratic weighted matrix based on the deviation between each level, where the weights are proportional to the square of the level deviation; and calculating the weighted observation consistency based on the confusion matrix and the quadratic weighted matrix. Based on the marginal probabilities of the measured and simulated levels, the weighted stochastic consistency is calculated; based on the weighted observational consistency and the weighted stochastic consistency, the quadratic weighted Kappa coefficient is calculated.

[0012] In one embodiment, the verification further includes: constructing a grade transition sequence between adjacent dates in the target area based on soil moisture levels; calculating the transition probability between each grade according to the grade transition sequence, and calculating a grade stability index to characterize the frequency of soil moisture level changes; dynamically adjusting a preset basic Kappa consistency evaluation interval based on the grade stability index to obtain an adjusted Kappa consistency evaluation threshold; and using the adjusted Kappa consistency evaluation threshold to evaluate the consistency between the simulation results reflected by the quadratic weighted Kappa coefficient and the measured grades.

[0013] As can be seen from the above technical solutions, the embodiments of this application have the following advantages: By introducing daily meteorological data to construct the evapotranspiration index of previous precipitation and combining it with the S-shaped variation pattern of soil moisture absorption, a logistic nonlinear mapping relationship between the evapotranspiration index and soil relative humidity is established, enabling continuous simulation of the dynamic changes in farmland soil moisture. Compared with traditional methods such as SPEI and SAPEI, which only classify drought levels based on climate probability, this invention can not only comprehensively reflect the impact of precipitation replenishment, evapotranspiration water consumption, and the memory effect of previous moisture accumulation on soil moisture, but also characterize the nonlinear response characteristics of soil from slow water absorption, rapid wetting to saturation and stability, thereby improving the physical rationality and dynamic response capability of soil relative humidity simulation results. At the same time, by using measured soil relative humidity samples to fit the logistic function parameters, the model's adaptability to the characteristics of farmland moisture changes in different regions is enhanced, improving the accuracy and practicality of agricultural drought monitoring and soil moisture assessment. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of a soil relative humidity simulation method based on meteorological indices provided in an embodiment of this application.

[0016] Figure 2 The response diagram of soil relative humidity meteorological index based on previous precipitation evapotranspiration index to precipitation is provided for the embodiments of this application. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0018] Reference Figure 1 As shown, the present invention provides a schematic flowchart of a soil relative humidity simulation method based on meteorological indices, which includes the following steps: S1, acquire daily meteorological data for the target area, and calculate reference evapotranspiration based on the daily meteorological data, including: Acquire daily meteorological data for the target area. Daily meteorological data refers to meteorological observation data such as temperature, precipitation, wind speed, humidity, and sunshine duration acquired with a time resolution of "day" to characterize the weather conditions of the target area. The daily meteorological data includes daily average temperature, daily maximum temperature, daily minimum temperature, daily precipitation, daily sunshine duration, daily average wind speed, and daily average relative humidity. The daily meteorological data is preprocessed to obtain standardized daily meteorological data; The preprocessing includes outlier identification, physical rationality verification, and supplementing missing data using adjacent time interval interpolation or historical average method.

[0019] The outlier identification includes determining whether daily meteorological elements exceed the historical climate extreme range of the corresponding meteorological elements; if they do, the corresponding data is marked as outlier data; the physical rationality check includes verifying the consistency of the physical relationship between daily maximum temperature, daily minimum temperature, daily average relative humidity, and daily precipitation; when the daily minimum temperature is higher than the daily maximum temperature, the relative humidity is greater than 100% or less than 0%, or the precipitation is less than 0 mm, the corresponding data is marked as invalid data.

[0020] Based on the obtained standardized daily meteorological data, the daily total solar radiation of the target area is calculated. When solar radiation observation data is lacking for the target area, the total solar radiation is calculated using the daily sunshine hours according to the Angstrom empirical formula. The specific formula for calculating the total solar radiation is as follows:

[0021] In the formula, Total solar radiation. Astronomical radiation, This represents the actual hours of sunshine. To maximize the possible hours of sunshine, , This is an empirical coefficient.

[0022] Based on the obtained total solar radiation, the net radiation of the target area is further calculated; The net radiation is calculated using the following formula:

[0023] In the formula, Net radiation, Net shortwave radiation, This is net longwave radiation.

[0024] The net shortwave radiation is calculated based on total solar radiation and surface albedo, specifically as follows:

[0025] The net longwave radiation is calculated using the FAO-56 empirical formula for longwave radiation, based on daily maximum and minimum temperatures, actual vapor pressure, and solar radiation conditions. Specifically:

[0026] In the formula, It is the surface albedo. This is the Stefan-Boltzmann constant. , These are the daily maximum absolute temperature and the daily minimum absolute temperature, respectively. For clear sky solar radiation, This is the actual water vapor pressure.

[0027] It should be noted that the daily maximum and minimum absolute temperatures are obtained by adding 273 to the Celsius temperature.

[0028] Based on standardized daily meteorological data, the air saturation water vapor pressure is calculated, and the water vapor pressure deficit is calculated in combination with the actual water vapor pressure. The specific formula for calculating the saturated water vapor pressure of the air is as follows:

[0029] In the formula, The pressure of water vapor in saturated air. The daily maximum temperature, The daily minimum temperature, , These are the saturated water vapor pressures at the corresponding temperatures.

[0030] The specific calculation formula for the water vapor pressure deficit is as follows:

[0031] In the formula, This is a missing term due to water vapor pressure deficiency.

[0032] Based on the daily average temperature, net radiation, water vapor pressure deficit, and daily average wind speed, the Penman-Monteith method is used to calculate the daily reference evapotranspiration of the target area. The specific formula for calculating the reference evaporation rate is as follows:

[0033] In the formula, Evapotranspiration is a reference evapotranspiration rate, representing the amount of water lost through evapotranspiration per unit time under standard reference crop conditions. The slope of the saturated vapor pressure curve represents the rate of change of saturated vapor pressure with temperature, and G is the soil heat flux, representing the heat exchange flux between the soil and the land surface. The humidity constant is used to characterize the relationship between the thermodynamic properties of air and water vapor pressure. The daily average temperature The daily average wind speed is 2m above the ground.

[0034] S2, calculating the potential evapotranspiration of the crop based on the crop coefficient and the reference evapotranspiration, including: Obtain the types of crops planted in the target area, their growth stages, and the calculated daily reference evapotranspiration. The crop type includes at least one of winter wheat, summer corn, rice, soybean, or cotton; the growth stage includes the initial growth period, rapid growth period, mid-growth period, and late maturity period.

[0035] The crop coefficient is determined based on the crop type and corresponding growth stage of the target area; Different crop types and different growth stages correspond to different crop coefficients, which are determined based on FAO recommended crop coefficients, regional agro-meteorological experimental data, or historical farmland observation data.

[0036] Table 1 shows examples of crop coefficient values ​​for different crop types and different growth stages in this embodiment. As can be seen from Table 1, there are significant differences in the evapotranspiration water consumption capacity of different crops and different growth stages. Therefore, this invention improves the adaptability of the crop potential evapotranspiration calculation results to actual agricultural production conditions through a dynamic crop coefficient matching mechanism.

[0037] Table 1. Examples of crop coefficient values ​​for different crop types and growth stages.

[0038] Time series matching processing is performed on the determined crop coefficients; Specifically, the crop coefficient for the corresponding date is matched with the daily reference evapotranspiration for the corresponding date obtained in S1 to form a daily crop evapotranspiration calculation sequence. The daily matching includes the synchronous switching of the crop coefficient for the corresponding date to the crop coefficient for the corresponding growth stage when the crop growth stage changes, thereby forming a daily crop coefficient time series that dynamically changes with the crop growth process.

[0039] Based on the daily crop coefficient and daily reference evapotranspiration, calculate the daily potential evapotranspiration of the target area. The specific formula for calculating the daily potential crop evapotranspiration is as follows:

[0040] In the formula, Daily potential evapotranspiration of crops refers to the theoretical maximum water consumption due to evapotranspiration of a specific crop under conditions of sufficient water supply, determined jointly by crop growth and meteorological environment. For crop coefficients, This is a daily reference evaporation.

[0041] S3, based on daily precipitation and potential crop evapotranspiration, calculates the anterior precipitation evapotranspiration index, including: Obtain daily precipitation data for the target area and daily potential crop evapotranspiration calculated in step S2; The daily precipitation data and the daily potential crop evapotranspiration are matched one-to-one according to the date. Based on daily precipitation data and daily potential crop evapotranspiration, calculate the daily water surplus / deficit in the target area; The specific formula for calculating the daily moisture surplus / deficit is as follows:

[0042] In the formula, Let be the daily moisture surplus or deficit corresponding to the i-th leading day. Let be the daily precipitation corresponding to the i-th leading day. Let be the daily potential crop evapotranspiration corresponding to the i-th leading day.

[0043] Where: If A value greater than 0 indicates that the soil moisture in the target area is in surplus; conversely, a value less than 0 indicates that the soil moisture in the target area is in deficit.

[0044] A leading time series is constructed based on daily water balance. The preceding time series is arranged from most recent to oldest time, i=0 represents the current day, i gradually increases to indicate backtracking to historical dates, and m represents the maximum number of preceding days, which is 100 days.

[0045] Based on the obtained leading time series, time decay weights are applied to the daily water surplus and deficit for different leading dates. Among them: the closer the daily water surplus or deficit is to the current date, the greater the time decay weight corresponding to it; the further the daily water surplus or deficit is from the current date, the smaller the time decay weight corresponding to it. The time decay weight satisfies:

[0046] In the formula, The time decay weight corresponding to the i-th day-leader is... is the attenuation coefficient, and i is the preceding date sequence number.

[0047] The attenuation coefficient can be determined based on historical soil relative humidity observation data and daily meteorological data for the corresponding period, using a method of traversal search and correlation evaluation. Specifically, the attenuation coefficient is first set to a range of 0.90 to 0.99, with a step size of 0.001. Then, for each candidate attenuation coefficient, the corresponding anterior precipitation evapotranspiration index (APEI) is calculated based on daily precipitation and crop potential evapotranspiration, and the Pearson correlation coefficient between the APEI and the soil relative humidity observation value during the same period is calculated. Finally, the candidate attenuation coefficient with the largest correlation coefficient is selected as the final attenuation coefficient for the target area.

[0048] Based on the obtained time decay weight, the daily water surplus and deficit are weighted and accumulated to obtain the APEI (Advanced Precipitation Evapotranspiration Index) of the target area. The APEI is a meteorological index used to characterize the water balance of farmland, which is constructed based on the weighted cumulative water surplus and deficit relationship between the daily precipitation and the potential evapotranspiration of crops. The specific formula for calculating the APEI (Average Precipitation Evapotranspiration Index) is as follows:

[0049] In the formula, m is the number of preceding days.

[0050] It should be noted that the above-mentioned method for calculating the APEI (Advanced Precipitation Evapotranspiration Index) comprehensively considers the daily precipitation input, crop evapotranspiration, and the continuous impact of historical water processes. It uses daily water surplus / deficit to characterize the actual water balance of farmland and reflects the physical process of the gradual weakening of the impact of previous water processes over time through time decay weights. This results in meteorological water processes more recent to the current date having a higher contribution, while the influence of earlier historical processes gradually decreases. Compared with traditional fixed-timescale drought indices, this method can more realistically simulate the dynamic changes in farmland soil moisture, enhance the response to continuous drought, periodic precipitation, and rapid drought-flood transitions, and retain the cumulative memory effect of previous meteorological conditions on current soil moisture. This improves the accuracy and timeliness of agricultural drought and farmland moisture monitoring, and provides an input basis with clear physical meaning of farmland moisture for the subsequent construction of the SRMMI (Solar Relative Humidity Meteorological Index).

[0051] S4. Based on the S-shaped variation characteristics of soil moisture absorption, a logistic function relationship between the evapotranspiration index of previous precipitation and soil relative humidity is constructed. This logistic function is used to simulate the nonlinear variation process of soil moisture absorption, from slow growth to rapid growth and finally to saturation and stability.

[0052] In this embodiment, based on the S-shaped variation characteristics of the soil moisture absorption process, a logistic function relationship between the evapotranspiration index of previous precipitation and soil relative humidity is constructed, including: Obtain the APEI (Area of ​​Precipitation Evapotranspiration Index) calculated in step S3 and the soil relative humidity observation data of the target area during the same period, wherein the soil relative humidity observation data are the measured soil relative humidity values ​​of the soil layer corresponding to different dates in the target area; Analysis of soil moisture content variation characteristics based on soil moisture absorption process; Based on the characteristics of soil moisture content variation, the logistic function is selected as the nonlinear mapping model between the evapotranspiration index (APEI) of previous precipitation and soil relative humidity. The logistic function is used to describe the change pattern of soil relative humidity as it gradually increases and eventually stabilizes with the accumulation of farmland moisture.

[0053] Based on the selected logistic function, a functional relationship between the APEI (Air Evapotranspiration Index) and the SRMMI (Soil Relative Moisture Index) is constructed. The soil relative humidity meteorological index SRMMI satisfies:

[0054] In the formula, This represents the maximum relative humidity of the soil. These are parameters for adjusting the shape of the logistic function. This is the growth rate parameter for the logistic function.

[0055] The maximum relative humidity of the soil is set to 100% to ensure that the simulation results meet the physical range constraints of relative humidity of farmland soil.

[0056] Based on the constructed logistic function relationship, a nonlinear transformation of the APEI (Air Evapotranspiration Index) to the SRMMI (Soil Relative Humidity Index) is realized: as the APEI increases, the SRMMI shows a continuous nonlinear growth. When the APEI continues to increase, the growth rate of the SRMMI first increases and then decreases. When the APEI tends to a larger value, the SRMMI gradually approaches the maximum value of the soil relative humidity.

[0057] It should be noted that by constructing a logistic function relationship between the evapotranspiration index (APEI) of previous precipitation and soil relative humidity based on the S-shaped change characteristics of the soil moisture absorption process, the logistic function can be used to nonlinearly characterize the continuous moisture response process of farmland soil from initial slow water absorption, rapid humidification to gradual saturation and stabilization. This allows the soil relative humidity simulation results to reflect the combined impact of previous precipitation accumulation and evapotranspiration on farmland moisture changes, while also satisfying the objective law that soil moisture content has a physical upper limit. This improves the physical rationality, dynamic response capability, and accuracy of agricultural drought monitoring in soil moisture simulation.

[0058] Furthermore, the characteristics of soil moisture content changes are analyzed based on the soil moisture absorption process, including: Obtain the historical APEI (Average Precipitation Evapotranspiration Index) sequence for the target area and the corresponding soil relative humidity observation sequence for the period; The soil relative humidity observation sequence is used to characterize the actual soil moisture status of farmland in the target area.

[0059] Based on the obtained APEI sequence and soil relative humidity observation sequence, the APEI values ​​were sorted from smallest to largest to construct the APEI-soil relative humidity response sequence. The APEI value, from small to large, corresponds to the process of farmland moisture gradually changing from a deficit state to a moist state.

[0060] Based on the APEI-soil relative humidity response sequence, the soil relative humidity change rate corresponding to adjacent APEI intervals was calculated. The specific formula for calculating the soil relative humidity change rate is as follows:

[0061] In the formula, Let be the rate of change of soil relative humidity corresponding to the i-th interval. , These are the soil relative humidity observation values ​​corresponding to adjacent APEI intervals. , These are the evapotranspiration indices of previous precipitation corresponding to adjacent APEI intervals.

[0062] The soil relative humidity change rate is smoothed to obtain the smoothed soil relative humidity change rate. Based on the smoothed soil relative humidity change rate, the APEI-soil relative humidity response process is divided into stages. Among them, if the rate of change of relative soil humidity after smoothing increases continuously, it corresponds to the rapid humidification stage; if the rate of change of relative soil humidity after smoothing reaches the stage before the local maximum value, it corresponds to the initial slow water absorption stage; if the rate of change of relative soil humidity after smoothing reaches the local maximum value and then decreases continuously, it corresponds to the saturated and stable stage; the local maximum value corresponds to the peak growth rate of the soil moisture absorption process.

[0063] Based on the obtained stage division results, the average rate of change corresponding to the initial slow water absorption stage, the rapid humidification stage and the saturated stable stage are calculated respectively. The average rate of change is calculated using the following formula:

[0064] In the formula, The average rate of change This represents the number of intervals included in the corresponding stage.

[0065] Based on the average rate of change at each stage, the S-shaped characteristics of the soil moisture change process are determined. Among them, when the average rate of change in the initial slow water absorption stage is less than the average rate of change in the rapid humidification stage, and the average rate of change in the rapid humidification stage is greater than the average rate of change in the saturated stable stage, it is determined that there is an S-shaped change characteristic between APEI and soil relative humidity. The S-shaped change characteristic is used to characterize the continuous change process of farmland soil moisture from initial water absorption, rapid wetting to gradual saturation and stabilization.

[0066] It should be noted that the S-shaped feature construction method based on soil moisture stage division described above can automatically identify the continuous moisture response process of farmland soil from initial slow water absorption, rapid humidification to gradual saturation and stabilization based on the actual change relationship between APEI and soil relative humidity. By calculating the rate of change, smoothing, and dividing the stages, the interference of random meteorological fluctuations and abnormal samples on soil moisture response analysis is reduced, improving the stability, objectivity, and repeatability of S-shaped feature identification. At the same time, this method avoids the problem of strong subjectivity in traditional experience-based judgment methods, giving the nonlinear relationship between the previous precipitation evapotranspiration index and soil relative humidity a clear data analysis basis and physical meaning of farmland moisture, thereby improving the accuracy and regional adaptability of subsequent soil relative humidity meteorological index model construction.

[0067] S5, using soil relative humidity observation samples to fit the parameters in the logistic function, a meteorological simulation model of soil relative humidity for the target area is obtained. Real-time or forecast meteorological data is input into the soil relative humidity meteorological simulation model to obtain the simulation results of soil relative humidity for the target area, including: Based on the logistic function relationship, the simulated value of soil relative humidity for each sample is calculated. The APEI sample is input into the logistic function to obtain the simulated soil relative humidity value for the corresponding sample. .

[0068] Calculate the fitting residuals for each sample based on the simulated and observed values ​​of soil relative humidity. The specific formula for calculating the fitting residual is as follows:

[0069] In the formula, To fit the residuals, Let be the soil relative humidity observation value corresponding to the i-th sample.

[0070] Based on the fitted residual sequence, calculate the mean and standard deviation of the residuals for all samples; The specific formula for calculating the mean residual is as follows:

[0071] In the formula, Let y be the mean of the residuals and y be the total number of samples.

[0072] The specific formula for calculating the residual standard deviation is as follows:

[0073] In the formula, This represents the standard deviation of the residuals.

[0074] Outlier samples are identified based on the residual mean and residual standard deviation. Wherein, the fitting residual corresponding to a certain sample satisfies: When the condition is met, the corresponding sample is identified as an abnormal sample; the remaining samples are identified as normal samples. Based on the identified abnormal and normal samples, different fitting weights are assigned to different samples to obtain the weights of each sample. Based on the sample weights, a weighted least squares objective function is constructed;

[0075] In the formula, For the weighted sum of squared errors, Let be the weight corresponding to the i-th sample.

[0076] Based on the constructed weighted least squares objective function, the logistic function parameters are... and Perform optimization and solution; Among them, an iterative optimization method is adopted, with the goal of minimizing the weighted sum of squared errors Q, to optimize the parameters. and Perform optimization fitting.

[0077] Based on the parameters obtained from the optimization fit and A meteorological simulation model of soil relative humidity in the target area is constructed, wherein the meteorological simulation model of soil relative humidity is used to realize the nonlinear conversion of the APEI (Air Evapotranspiration Index) of the previous precipitation to the SRMMI (Soil Relative Humidity Meteorological Index). By inputting real-time meteorological data or meteorological forecast data into the soil relative humidity meteorological simulation model, the simulation results of soil relative humidity in the target area are obtained.

[0078] The soil relative humidity simulation result is the obtained soil relative humidity meteorological index (SRMMI).

[0079] like Figure 2It can be seen that when there is no precipitation, the soil relative humidity meteorological index shows a slow downward trend. When precipitation occurs to varying degrees, the simulated soil relative humidity increases accordingly, and its response curve conforms to the water balance process of farmland. When there is no precipitation, the downward slope of the curve is gentler in winter than in summer, indicating that this index retains the advantages of the previous precipitation evapotranspiration index and can characterize the influence of temperature and evapotranspiration on water consumption.

[0080] It should be noted that by constructing an anomaly sample identification and dynamic weighted least squares parameter fitting mechanism based on residual statistical characteristics, adaptive suppression of the impact of anomaly observation samples is achieved. While preserving the true soil moisture change pattern and extreme weather response characteristics, the stability and robustness of logistic function parameter fitting and the accuracy of soil relative humidity meteorological simulation results are improved.

[0081] Furthermore, based on the identified abnormal and normal samples, different fitting weights are assigned to different samples to obtain the weights of each sample, including: Based on the obtained abnormal and normal samples, calculate the absolute value of the residual for each sample. The absolute value of the residual is calculated using the following formula:

[0082] In the formula, Let be the absolute value of the residual corresponding to the i-th sample.

[0083] Based on the absolute value of the residuals, the normal samples are assigned the first fitting weight; Wherein, the first fitting weights satisfy: This ensures that normal samples participate fully in the fitting of the logistic function parameters.

[0084] Based on the absolute value of the residuals, a second fitting weight is assigned to outlier samples; Wherein, the second fitting weight decreases continuously as the absolute value of the residual increases, and the second fitting weight satisfies:

[0085] In the formula, This is the weight decay coefficient.

[0086] Specifically, when the absolute value of the residual of the outlier increases, the corresponding second fitting weight decreases; when the absolute value of the residual of the outlier decreases, the corresponding second fitting weight increases, thereby achieving a continuous change between the degree of anomaly and the fitting weight.

[0087] Based on the first and second fitting weights, a sample weight sequence is constructed; The sample weight sequence is corresponding one-to-one with the APEI sample sequence and the soil relative humidity observation sample sequence according to the sample number.

[0088] Based on the obtained sample weight sequence, the contribution of different samples to the fitting of logistic function parameters is adjusted; In this model, normal samples are assigned larger fitting weights, while abnormal samples are assigned smaller fitting weights. This reduces the interference of abnormal samples on the optimization results of the logistic function parameters and improves the stability and robustness of the soil relative humidity meteorological model parameter fitting.

[0089] In one embodiment, step S5 is followed by: using a second-weighted Kappa coefficient to verify the consistency between the simulated soil relative humidity results and the measured soil relative humidity levels.

[0090] In this embodiment, the consistency between the simulated soil relative humidity results and the measured soil relative humidity levels is verified using a quadratic weighted Kappa coefficient, including: Based on the obtained simulated soil relative humidity results and measured soil relative humidity observation data, the soil moisture level is classified. Soil moisture is divided into multiple levels according to the range of relative soil humidity values; The soil moisture levels include: extreme drought, severe drought, moderate drought, mild drought, slightly dry, slightly wet, and excessively wet. in: Soil relative humidity is less than 30%, which is classified as extreme drought; Soil relative humidity is greater than or equal to 30% but less than 40%, and is classified as severe drought. Soil relative humidity is classified as moderately dry when it is greater than or equal to 40% but less than 50%. Soil relative humidity is greater than or equal to 50% but less than 60%, and is classified as mild drought. Soil with a relative humidity of 60% or higher but less than 75% is classified as slightly dry. Soil with a relative humidity of 75% or higher but less than 90% is classified as slightly moist. Soil relative humidity of 90% or higher is classified as excessively wet.

[0091] Based on the obtained soil moisture level, the simulated results of soil relative humidity and the measured soil relative humidity observation data are respectively coded into levels; Different soil moisture levels correspond to different level numbers 1 to 7, with the level numbers increasing sequentially from dry to wet soil moisture.

[0092] Based on the grade coding results, a confusion matrix between the simulated grade and the measured grade of soil relative humidity is constructed. In this system, the rows of the confusion matrix represent the measured soil moisture levels, the columns of the confusion matrix represent the simulated soil moisture levels, and the elements in the confusion matrix represent the number of samples that appear in the corresponding level combination.

[0093] Based on the constructed confusion matrix, calculate the quadratic weighting matrix between different levels; The secondary weighted matrix is ​​used to characterize the degree of deviation between different soil moisture levels. The greater the difference between levels, the greater the corresponding weight value, thus imposing a higher penalty on samples with larger level deviations. The specific calculation formula is as follows:

[0094] In the formula, The weights are the quadratic weights corresponding to the i-th and j-th levels. This is the measured grade number. For simulation level numbering, This represents the total number of soil moisture levels.

[0095] Based on the obtained quadratic weighted weight matrix and confusion matrix, the weighted observation consistency is calculated; The specific formula for calculating the weighted observation consistency is as follows:

[0096] In the formula, Weighted observation consistency refers to the actual degree of consistency between the simulated level and the measured level calculated based on the quadratic weighting matrix and confusion matrix. It is used to characterize the true matching between the model simulation results and the measured results. Let be the number of observed samples corresponding to the i-th row and j-th column in the confusion matrix. This represents the total number of samples.

[0097] Calculate the random consistency probability matrix based on the confusion matrix; The random consistency probability matrix is ​​obtained by calculating the proportion of the number of samples corresponding to each measured level in the confusion matrix to the total number of samples as the marginal probability of the measured level, and calculating the proportion of the number of samples corresponding to each simulated level to the total number of samples as the marginal probability of the simulated level. The random consistency probability corresponding to each level combination is obtained by multiplying the marginal probability of the measured level and the marginal probability of the simulated level. This matrix is ​​used to characterize the expected occurrence of different level combinations under random conditions.

[0098] Based on the obtained quadratic weighted weight matrix and random consistency probability matrix, the weighted random consistency is calculated; The specific calculation formula for the weighted random consistency is as follows:

[0099] In the formula, Weighted random consistency is used to characterize the impact of random consistency on model test results. Let be the expected number of samples corresponding to the i-th row and j-th column of the random consistency probability matrix.

[0100] The quadratic weighted Kappa coefficient is calculated based on weighted observation consistency and weighted random consistency. The specific formula for calculating the weighted Kappa coefficient is as follows:

[0101] In the formula, The Kappa coefficient is a second-weighted coefficient used to quantitatively characterize the degree of consistency between simulated and measured soil relative humidity levels.

[0102] The consistency between the simulated soil relative humidity results and the measured soil relative humidity level is evaluated based on the quadratic weighted Kappa coefficient.

[0103] Furthermore, based on the quadratic weighted Kappa coefficient, the consistency between the simulated soil relative humidity results and the measured soil relative humidity levels is evaluated, including: Based on soil moisture levels, a historical soil moisture level transition sequence for the target area was constructed; The grade transition sequence is used to characterize the process of soil moisture grade change between adjacent dates. The grade transition process includes grade retention and transition between different grades.

[0104] Based on the aforementioned level transition sequence, the number of level transitions between each level is counted. Based on the number of grade transitions, a soil moisture grade transition matrix for the target area is constructed; In the rank transfer matrix, the elements represent the number of times the corresponding rank transfer process occurs, the matrix rows represent the ranks before the transfer, and the matrix columns represent the ranks after the transfer.

[0105] Based on the rank transition matrix, calculate the rank transition probability between each rank; The specific formula for calculating the level transition probability is as follows:

[0106] In the formula, Let i be the level transition probability corresponding to the transition from level i to level j. This represents the number of level transitions corresponding to the transition from level i to level j.

[0107] Based on the probability of grade transition, the stability index of soil moisture grade in the target area is calculated; The grade stability index is used to characterize the frequency of changes in soil moisture grade in the target area. The specific calculation formula for the grade stability index is as follows:

[0108] In the formula, This is a stability index, with a value range of 0 to 1. The probability of level transition for keeping the i-th level unchanged.

[0109] Specifically, a stability index close to 1 indicates stable soil moisture levels in the target area, while a stability index close to 0 indicates frequent changes in soil moisture levels in the target area. Based on the grade stability index, the consistency evaluation intervals of each Kappa are dynamically adjusted; Specifically, a basic Kappa consistency evaluation interval is pre-defined, which is used to characterize the consistency level corresponding to different Kappa value ranges; and the basic Kappa consistency evaluation interval is dynamically corrected based on the level stability index; the basic Kappa consistency evaluation interval can be set using a domain-general Kappa consistency division standard.

[0110] Specifically, when the stability index S decreases, the Kappa consistency evaluation threshold is lowered accordingly; when the stability index S increases, the Kappa consistency evaluation threshold is raised accordingly.

[0111] The dynamically adjusted Kappa consistency evaluation threshold satisfies:

[0112] In the formula, The dynamically adjusted Kappa conformity assessment threshold. Basic Kappa evaluation threshold, For adjustment coefficients, .

[0113] It should be noted that since the value range of the stability index S is 0 to 1, "0.5" is introduced as a stability neutral benchmark when dynamically adjusting the Kappa evaluation threshold. When S is greater than 0.5, it indicates that the soil moisture level in the target area is relatively stable, and the Kappa evaluation threshold is increased accordingly to enhance the rigor of the evaluation. When S is less than 0.5, it indicates that the soil moisture level changes more frequently, and the Kappa evaluation threshold is decreased accordingly to improve the adaptability of the evaluation. When S equals 0.5, the dynamic adjustment term is 0, so that the adjusted Kappa evaluation threshold remains unchanged with respect to the basic evaluation standard, thereby achieving symmetrical dynamic adjustment of the Kappa evaluation threshold around the basic threshold.

[0114] The simulation performance of the soil relative humidity meteorological model was tested based on the quadratic weighted Kappa coefficient and the dynamically adjusted Kappa consistency evaluation interval. Specifically, when the second-weighted Kappa coefficient meets the requirements of the consistency evaluation interval after dynamic adjustment, the soil relative humidity meteorological model is determined to meet the soil moisture simulation requirements of the target area; otherwise, the parameters of the soil relative humidity meteorological model are re-optimized and fitted.

[0115] It should be noted that by constructing a grade stability index using the soil moisture grade transition probability, and dynamically adjusting the Kappa consistency evaluation interval based on the grade stability index, the evaluation standard of the secondary weighted Kappa coefficient can adapt to the dynamic change characteristics of soil moisture in different regions. This effectively solves the problem of insufficient adaptability of the traditional fixed Kappa evaluation interval between highly fluctuating and stable regions. At the same time, by introducing grade transition process and grade stability analysis, the model's ability to respond to the temporal change characteristics of soil moisture is enhanced, and the objectivity, accuracy, and regional applicability of the consistency evaluation results of the soil relative humidity meteorological model are improved.

[0116] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0117] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0118] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. 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.

[0119] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0120] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for simulating soil relative humidity based on meteorological indices, characterized in that, include: Obtain daily meteorological data for the target area and calculate the evapotranspiration index of previous precipitation, including: Acquire daily meteorological data for the target area and perform data preprocessing; Calculate the reference evapotranspiration based on the preprocessed daily meteorological data; Calculate the potential evapotranspiration of the crop based on the crop coefficient and the reference evapotranspiration. The specific formula for calculating the potential evapotranspiration of the crop is as follows: In the formula, This represents the daily potential evapotranspiration of crops. For crop coefficients, For daily reference evapotranspiration; Calculate the evapotranspiration index based on daily precipitation and potential crop evapotranspiration. The specific formula for calculating the evapotranspiration index of the preceding precipitation is as follows: In the formula, This represents the evapotranspiration index of previous precipitation. Here, is the attenuation coefficient, and is the preceding date sequence number. Let be the daily precipitation corresponding to the i-th leading day. Let m be the daily potential crop evapotranspiration corresponding to the i-th leading day, and m be the number of leading days. Based on the S-shaped variation characteristics of the soil moisture absorption process, a logistic function relationship between the evapotranspiration index of the preceding precipitation and the relative humidity of the soil is constructed. By fitting the parameters of the logistic function using soil relative humidity observation samples, a soil relative humidity meteorological simulation model is obtained for outputting soil relative humidity simulation results.

2. The method according to claim 1, characterized in that, Based on the S-shaped variation characteristics of the soil moisture absorption process, the logistic function relationship between the anterior precipitation evapotranspiration index and soil relative humidity is constructed, including: Obtain the evapotranspiration index of previous precipitation and the soil relative humidity observation data of the target area during the same period; Based on the S-shaped variation characteristics of the soil moisture absorption process, the logistic function is selected as the nonlinear mapping model between APEI and soil relative humidity. The logistic function relationship between the evapotranspiration index of previous precipitation and the simulated value of soil relative humidity is constructed as follows: In the formula, This is a simulated value for soil relative humidity. This represents the maximum relative humidity of the soil. These are parameters for adjusting the shape of the logistic function. The growth rate parameter of the logistic function. This represents the evapotranspiration index of previous precipitation.

3. The method according to claim 2, characterized in that, The S-shaped change characteristics based on the soil moisture absorption process also include: Obtain the previous precipitation evapotranspiration index sequence and the corresponding soil relative humidity observation sequence for the target area; The sequences are sorted in ascending order of the evapotranspiration index of previous precipitation to construct a response relationship sequence; Calculate the rate of change of soil relative humidity corresponding to adjacent intervals in the response relationship sequence, and perform smoothing processing; Based on the continuous trend of the change rate of relative humidity in smoothed soil, the soil moisture absorption process is divided into an initial slow water absorption stage, a rapid humidification stage, and a saturated stable stage. The average rate of change for each stage was calculated. Based on the fact that the average rate of change in the initial slow water absorption stage was less than that in the rapid humidification stage and that the average rate of change in the rapid humidification stage was greater than that in the saturated stable stage, it was determined that the soil moisture change process exhibited an S-shaped change characteristic.

4. The method according to claim 1, characterized in that, The process of fitting the parameters of the logistic function using soil relative humidity observation samples to obtain a soil relative humidity meteorological simulation model for outputting soil relative humidity simulation results includes: Input the evapotranspiration index samples from previous precipitation into the logistic function to obtain the simulated values ​​of soil relative humidity; The fitting residuals are calculated based on the simulated and observed values ​​of soil relative humidity, and abnormal and normal samples are identified based on the statistical characteristics of the residuals. Different fitting weights are assigned to abnormal samples and normal samples respectively, and a weighted least squares objective function is constructed. The logistic function parameters that minimize the weighted least squares objective function are solved by iterative optimization to obtain the meteorological simulation model for soil relative humidity. Real-time or forecasted meteorological data are input into the soil relative humidity meteorological simulation model to obtain the soil relative humidity simulation results for the target area.

5. The method according to claim 4, characterized in that, The method of assigning different fitting weights to abnormal samples and normal samples includes: Based on the identified abnormal and normal samples, calculate the absolute value of the residual for each sample; Assign a constant first fitting weight to normal samples; A second fitting weight is assigned to outlier samples, which decreases continuously as the absolute value of the residual increases. Based on the first fitting weight and the second fitting weight, a sample weight sequence corresponding one-to-one with the sample is constructed, which is used to adjust the contribution of different samples in the fitting of the logistic function parameters.

6. The method according to claim 4, characterized in that, It also includes using a quadratic weighted Kappa coefficient to verify the consistency between the simulated soil relative humidity results and the measured soil relative humidity levels, including: Based on the simulation results of soil relative humidity and the measured observation data, soil moisture is divided into multiple levels; The simulation results and measured data are respectively graded, and a confusion matrix between the simulation grade and the measured grade is constructed. Based on the deviation between each level, a quadratic weighted weight matrix is ​​calculated, where the weights are proportional to the square of the level deviation; Calculate the weighted observation consistency based on the confusion matrix and the quadratic weighted weight matrix; Calculate weighted random consistency based on the marginal probabilities of the measured and simulated levels; The quadratic weighted Kappa coefficient is calculated based on the weighted observation consistency and the weighted random consistency.

7. The method according to claim 6, characterized in that, The test also includes: Based on soil moisture levels, a grade transition sequence between adjacent dates in the target area is constructed; The probability of transition between each level is calculated based on the level transition sequence, and a level stability index is calculated to characterize the frequency of changes in soil moisture level. Based on the aforementioned level stability index, the preset basic Kappa consistency evaluation interval is dynamically adjusted to obtain the adjusted Kappa consistency evaluation threshold. The adjusted Kappa consistency evaluation threshold is used to evaluate the consistency between the simulation results reflected by the quadratic weighted Kappa coefficient and the measured level.

Images