A method for synergistically improving soil fertility and yield capacity of corn straw full-amount field covering and returning

Abstract

This invention relates to the field of agricultural data analysis, and particularly to a method for synergistically improving soil fertility and productivity through full coverage of corn straw. The method includes: acquiring a synchronous multidimensional time-series data set of the full corn straw cover layer and the underlying soil; performing cross-interface hydrothermal inversion transport resistance analysis to obtain a retardation attenuation coefficient; performing pore evolution discrete phase space curvature analysis on the volumetric water content sequence of the straw cover layer to obtain a curvature compensation factor; obtaining the actual decomposition driving dependent variable by fusing and reconstructing the original soil hydrothermal comprehensive driving dependent variable, the retardation attenuation coefficient, and the curvature compensation factor; and obtaining fertilization decision instructions by calculating the cumulative decomposition process and assessing the fertilizer demand gap based on the actual decomposition driving dependent variable. This solves the problem that existing straw decomposition process prediction algorithms cannot accurately characterize the interfacial hydrothermal inversion and internal water-locking inertia between the cover layer and the underlying soil, leading to misalignment of fertilization decision nodes and dosages.

Description

Technical Field

[0001] This invention relates to the field of agricultural data analysis technology, and in particular to a method for synergistically improving soil fertility and productivity through full coverage of corn stalks and returning them to the field. Background Technology

[0002] Full coverage of corn stalks and their return to the field is a common practice in conservation tillage. It reduces water evaporation, improves soil structure, and increases soil organic matter levels, making it highly valuable for improving farmland fertility. However, in the early stages of stalk decomposition, microbial activity consumes available nitrogen in the soil, potentially creating competition for nitrogen with corn seedling growth. Therefore, existing farmland management systems typically combine field environmental monitoring data to dynamically predict the stalk decomposition process and estimate nitrogen release levels based on the predictions to guide compensatory nitrogen fertilizer application. Current stalk decomposition prediction methods often use soil temperature and humidity monitoring data at a single depth. At each time point, temperature and humidity values ​​are directly multiplied or linearly combined to generate a single-step hydrothermal driving force, which is then integrated and accumulated along the time axis as an assessment of the overall stalk decomposition process.

[0003] However, in scenarios where corn stalks are fully covered and returned to the field, the thick stalk cover layer blocks direct exchange between the atmosphere and the soil. After rainfall or irrigation, a hydrothermal inversion phenomenon easily occurs, where the surface of the stalk cover dries rapidly while the underlying soil remains moist. Simultaneously, the porous structure within the stalks creates a certain water-locking inertia during the initial shallow drying stage. Existing prediction methods that rely solely on single-step transient product integration of soil temperature and humidity data at a single depth are insufficient to accurately perceive the cross-interface physical barrier between the stalk cover and the underlying soil, and also fail to reflect the continued effect of the water-locking inertia within the stalks on the decomposition process. This easily leads to both overestimation and underestimation of the decomposition process, resulting in misalignment of compensatory nitrogen fertilizer application timing and dosage. Therefore, improving the accuracy of decomposition process assessment and fertilization decisions in scenarios where corn stalks are fully covered and returned to the field has become an urgent problem to be solved. Summary of the Invention

[0004] In view of this, the present invention aims to propose a method for synergistic improvement of soil fertility and productivity through full coverage of corn straw and returning it to the field, in order to solve the problem that existing straw decomposition process prediction algorithms cannot accurately characterize the interface hydrothermal inversion and internal water-locking inertia between the cover layer and the underlying soil, resulting in misalignment of fertilization decision nodes and dosages.

[0005] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0006] A method for synergistically improving soil fertility and productivity through full coverage of corn stalks and their return to the field, the method comprising:

[0007] Step S1: Collect and preprocess time-series monitoring data of the full coverage layer of corn stalks and the underlying soil in farmland to obtain a synchronous multidimensional time-series data set;

[0008] Step S2: Obtain the stagnation attenuation coefficient by performing cross-interface hydrothermal inversion transport resistance analysis on the synchronous multidimensional time series data set;

[0009] Step S3: Obtain the curvature compensation factor by performing discrete phase space curvature analysis on the volumetric moisture content sequence of the straw cover layer;

[0010] Step S4: Obtain the actual decomposition driving dependent variable by fusing and reconstructing the original soil hydrothermal integrated driving dependent variable, retardation attenuation coefficient and curvature compensation factor;

[0011] Step S5: Obtain fertilization decision instructions by calculating the cumulative decomposition process and assessing the fertilizer gap for the actual decomposition-driving dependent variables.

[0012] Furthermore, by collecting and preprocessing time-series monitoring data of the full coverage layer of corn stalks and the underlying soil in farmland, a synchronous multi-dimensional time-series data set is obtained, including:

[0013] In the target monitoring area of ​​farmland where corn stalks are fully covered and returned to the field, temperature and humidity sensor nodes are set up inside the corn stalks covering layer and in the shallow soil below the surface. Sampling time intervals are set and continuous monitoring is carried out according to the sampling time intervals to obtain time-series monitoring data of the corn stalks covering layer and the underlying soil.

[0014] The time-series monitoring data of the full corn stalk cover layer and the underlying soil in the farmland shall include at least: volumetric moisture content data of the stalk cover layer, volumetric moisture content data of the underlying soil layer, temperature data of the stalk cover layer, and temperature data of the underlying soil layer.

[0015] Data cleaning, outlier removal, missing value imputation, and timestamp alignment were performed on the volumetric moisture content data of the straw cover layer, the volumetric moisture content data of the underlying soil layer, the temperature data of the straw cover layer, and the temperature data of the underlying soil layer. The preprocessed data were then stored in association according to a unified sampling time order to obtain a synchronous multidimensional time-series data set.

[0016] Furthermore, the step of obtaining the stagnation attenuation coefficient by performing cross-interface hydrothermal inverted transport resistance analysis on a synchronous multidimensional time-series data set includes:

[0017] By performing cross-interface difference extraction processing on the cover layer moisture content data, soil layer moisture content data, cover layer temperature data, and soil layer temperature data in the synchronous multidimensional time series dataset, interface hydrothermal inversion characterization data is obtained.

[0018] The retardation decay coefficient was obtained by performing exponential decay mapping on the interfacial hydrothermal inversion characterization data.

[0019] Furthermore, the step of extracting cross-interface differences from the overburden moisture content data, soil moisture content data, overburden temperature data, and soil temperature data in the synchronous multidimensional time-series dataset to obtain interface hydrothermal inversion characterization data includes:

[0020] For any target acquisition time in the synchronous multidimensional time series data set, extract the straw cover layer volume moisture content data, the underlying soil layer volume moisture content data, the straw cover layer temperature data, and the underlying soil layer temperature data corresponding to the target acquisition time from the synchronous multidimensional time series data set.

[0021] The difference between the volumetric moisture content data of the underlying soil layer and the volumetric moisture content data of the straw cover layer is used as the interface moisture difference assessment corresponding to the target collection time. When the interface moisture difference assessment is less than the constant 0, the interface moisture difference assessment corresponding to the target collection time is set to the constant 0, and the interface moisture difference assessment after non-negative screening is used as the interface moisture inversion difference data corresponding to the target collection time.

[0022] The absolute value of the difference between the temperature data of the straw cover layer and the temperature data of the underlying soil layer is used as the interface temperature gradient data at the target acquisition time.

[0023] The interface moisture inversion difference data and interface temperature gradient data are used as the interface hydrothermal inversion characterization data corresponding to the target acquisition time.

[0024] Furthermore, the step of obtaining the retardation attenuation coefficient by performing exponential decay mapping processing on the interfacial hydrothermal inversion characterization data includes:

[0025] For any target collection time, acquire the interface moisture inversion difference data and interface temperature gradient data corresponding to the target collection time, and acquire the straw cover layer volume moisture content data and the underlying soil layer volume moisture content data corresponding to the target collection time.

[0026] The square of the interface moisture inversion difference data is used as the interface moisture inversion intensity assessment at the target acquisition time; the product of the volumetric moisture content data of the underlying soil layer and the volumetric moisture content data of the straw cover layer is used as the interface moisture coupling base at the target acquisition time, and a preset small positive constant is added to the interface moisture coupling base to obtain the interface moisture normalized base at the target acquisition time; the interface moisture inversion intensity assessment divided by the calculation result of the interface moisture normalized base is used as the interface moisture inversion penalty assessment at the target acquisition time.

[0027] The interface temperature gradient data is added to a constant 1 and the result is processed by natural logarithmic mapping to obtain the temperature difference excitation term corresponding to the target acquisition time.

[0028] The product of the interface moisture inversion penalty assessment and the temperature difference excitation term is used as the hydrothermal coupling penalty term corresponding to the target acquisition time. The negative of the hydrothermal coupling penalty term is subjected to exponential mapping with the natural constant as the base to obtain the stagnation attenuation coefficient corresponding to the target acquisition time.

[0029] Furthermore, the step of obtaining a curvature compensation factor by performing discrete phase space curvature analysis on the volumetric moisture content sequence of the straw cover layer includes:

[0030] By performing local temporal difference processing on the volumetric moisture content sequence of straw cover layer, data representing the moisture evolution curvature were obtained.

[0031] By statistically processing the historical water holding status of the straw cover layer volume moisture content sequence, historical water holding background characterization data is obtained.

[0032] Curvature compensation factor is obtained by coupling and compensating water evolution curvature characterization data with historical water holding background characterization data.

[0033] Furthermore, the step of obtaining moisture evolution curvature characterization data by performing local temporal difference processing on the volumetric moisture content sequence of the straw cover layer includes:

[0034] For any target sampling time in the straw cover layer volume moisture content sequence, obtain the straw cover layer volume moisture content data corresponding to the target sampling time, the straw cover layer volume moisture content data corresponding to the sampling time before the target sampling time, and the straw cover layer volume moisture content data corresponding to the two sampling times before the target sampling time.

[0035] The difference between the volumetric moisture content data of the straw cover layer at the target collection time and the volumetric moisture content data of the straw cover layer at the previous collection time is used as the first-order temporal difference data at the target collection time.

[0036] The second-order temporal difference data corresponding to the target acquisition time is calculated by subtracting twice the volumetric moisture content data of the straw cover layer corresponding to the previous acquisition time from the volumetric moisture content data of the straw cover layer corresponding to the target acquisition time, and then adding the volumetric moisture content data of the straw cover layer corresponding to the two acquisition times before the target acquisition time. When the second-order temporal difference data is less than the constant 0, the second-order temporal difference data corresponding to the target acquisition time is set to the constant 0, and the second-order temporal difference data after non-negative screening is used as the curvature resistance component data corresponding to the target acquisition time.

[0037] First-order temporal difference data and curvature resistance component data are used as the water evolution curvature characterization data corresponding to the target acquisition time.

[0038] Furthermore, the step of obtaining historical water-holding background characterization data by statistically processing the historical water-holding state of the straw cover layer volume moisture content sequence includes:

[0039] For any target sampling time in the straw cover layer volume moisture content sequence, acquire all straw cover layer volume moisture content data from the start sampling time to the target sampling time.

[0040] The calculation result of summing all the volumetric moisture content data of the straw cover layer from the start time to the target time is used as the historical water holding cumulative value corresponding to the target time.

[0041] Obtain the total number of collection times from the start collection time to the target collection time, and divide the historical water holding cumulative value by the total number of collection times as the historical average water holding level corresponding to the target collection time;

[0042] The historical average water holding level and the volumetric moisture content of the straw cover layer corresponding to the target collection time are used as the historical water holding background characterization data corresponding to the target collection time.

[0043] Furthermore, the process of coupling and compensating water evolution curvature characterization data with historical water-holding background characterization data to obtain a curvature compensation factor includes:

[0044] For any target sampling time in the straw cover layer volumetric moisture content sequence, obtain the first-order temporal difference data and curvature resistance component data corresponding to the target sampling time, and obtain the historical average water holding level and straw cover layer volumetric moisture content data corresponding to the target sampling time.

[0045] The square of the first-order temporal difference data is added to the square of the curvature resistance component data, and a preset small positive constant is added to the sum to obtain the curvature normalization basis corresponding to the target acquisition time; the curvature resistance component data is divided by the square root of the curvature normalization basis as the curvature angle characterization value corresponding to the target acquisition time.

[0046] The curvature angle characterization value is processed by arcsine mapping, and the result of multiplying the arcsine mapping result by a constant two and dividing by pi is used as the local curvature compensation evaluation corresponding to the target acquisition time.

[0047] The historical average water holding level is divided by the volumetric moisture content of the straw cover layer at the target collection time to obtain the historical water holding amplification assessment at the target collection time.

[0048] The product of the local curvature compensation assessment and the historical water holding capacity amplification assessment is used as the curvature compensation factor corresponding to the target acquisition time.

[0049] Furthermore, the actual decomposition driving variables are obtained by fusing and reconstructing the original soil hydrothermal integrated driving dependent variable, the retardation attenuation coefficient, and the curvature compensation factor, including:

[0050] For any target acquisition time, obtain the original soil hydrothermal integrated driving dependent variable, retardation attenuation coefficient, and curvature compensation factor corresponding to the target acquisition time; perform exponential mapping processing on the negative of the retardation attenuation coefficient with the natural constant as the base to obtain the gated activation term corresponding to the target acquisition time; take the product of the curvature compensation factor and the gated activation term as the compensation driving component corresponding to the target acquisition time; take the calculation result of adding the retardation attenuation coefficient and the compensation driving component as the reconstruction weight factor corresponding to the target acquisition time; take the product of the original soil hydrothermal integrated driving dependent variable and the reconstruction weight factor as the actual decomposition driving dependent variable corresponding to the target acquisition time.

[0051] Compared with the prior art, the present invention has the following advantages:

[0052] This invention presents a method for synergistically improving soil fertility and productivity through full-coverage corn stalk return to the field. Addressing the significant separation between the surface straw and underlying soil microenvironments in scenarios involving full-coverage corn stalk return, this method simultaneously acquires temporal data on moisture and temperature in both the straw cover layer and the underlying soil. By performing constraint analysis on the cross-interface hydrothermal inversion state, it effectively identifies false high-humidity drivers resulting from "surface drying and deep moisturization" after rainfall or irrigation. By using a retardation attenuation coefficient to dynamically suppress driving forces that do not conform to actual decomposition conditions, it avoids the erroneous assumption by existing algorithms that continuous high humidity in the underlying soil equates to active decomposition of straw surface microorganisms. This significantly reduces the risk of overestimating decomposition and consequently leading to premature fertilization and excessive fertilizer dosage, making the assessment of the decomposition process under straw return conditions more closely reflect the real field environment. Meanwhile, this invention further combines the local temporal difference characteristics of the volumetric moisture content sequence of straw mulch with historical water-holding background to compensate for the inertial effect of the straw's internal pore structure maintaining a certain water-locking capacity in the early stages of shallow air drying. This more realistically reflects the continuous decomposition kinetic energy when the activity of microorganisms inside the straw has not completely ceased. By fusing and reconstructing this compensation mechanism with the original soil hydrothermal comprehensive driving dependent variable, it not only avoids missing the effective period of straw decomposition but also better matches the available nitrogen release level calculated based on the decomposition process with the actual nitrogen requirement rhythm of maize seedlings. This improves the accuracy of the compensatory nitrogen fertilizer application nodes and dosages, taking into account both the soil fertility release effect after straw return to the field and the maize seedling yield guarantee effect. Attached Figure Description

[0053] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0054] Figure 1 This is a flowchart illustrating a method for synergistically improving soil fertility and productivity through full coverage of corn stalks in an embodiment of the present invention. Detailed Implementation

[0055] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0056] See Figure 1 This is a flowchart of a method for synergistically improving soil fertility and productivity through full coverage of corn stalks in the field, as provided in Embodiment 1 of the present invention. Figure 1 As shown, a method for synergistically improving soil fertility and productivity through full coverage and return of corn stalks to the field may include:

[0057] Step S1 involves collecting and preprocessing time-series monitoring data of the full coverage layer of corn stalks and the underlying soil in farmland to obtain a synchronous multidimensional time-series data set.

[0058] First, within the target monitoring area of ​​farmland where full coverage of corn stalks is implemented, temperature and humidity sensor nodes are deployed both inside the full corn stalk coverage layer and in the shallow soil below the surface. Sampling time intervals are set, and continuous monitoring is performed according to these intervals to obtain time-series monitoring data of the full corn stalk coverage layer and the underlying soil. In this embodiment, the sampling time interval is set to once every 5 minutes, and the sensor locations are chosen at 5 cm below the surface. The time-series monitoring data of the full corn stalk coverage layer and the underlying soil includes at least: volumetric moisture content data of the stalk coverage layer, volumetric moisture content data of the underlying soil layer, temperature data of the stalk coverage layer, and temperature data of the underlying soil layer.

[0059] Data cleaning, outlier removal, missing value imputation, and timestamp alignment were performed on the volumetric moisture content data of the straw cover layer, the volumetric moisture content data of the underlying soil layer, the temperature data of the straw cover layer, and the temperature data of the underlying soil layer. The preprocessed data were then stored in association according to a unified sampling time order to obtain a synchronous multidimensional time-series data set.

[0060] It should be noted that, in the embodiments of the present invention, when cleaning data, duplicate sampling records, records with incorrect timestamps, and invalid records exceeding the sensor's measurement range are preferentially removed; when removing outliers, the median absolute deviation discrimination method under a sliding time window can be used, combined with the outlier detection method based on local density, to identify and remove outliers; during the missing value imputation process, for data segments with a small number of consecutive missing points, linear interpolation or interpolation based on the average of adjacent valid times is preferred, while for data segments with a long consecutive missing time, a segmented imputation method based on the changing trend of adjacent historical time periods in the same layer is preferred; when performing timestamp alignment processing, a standard time axis is preferably constructed with a uniform sampling time interval, and the data collected by each sensor is matched to the standard time axis according to the nearest timestamp.

[0061] Thus, the collection and preprocessing of time-series monitoring data on the full coverage layer of corn stalks and the underlying soil in farmland has been completed, resulting in a synchronous multidimensional time-series data set.

[0062] Step S2: Obtain the stagnation attenuation coefficient by performing cross-interface hydrothermal inverted transport resistance analysis on the synchronous multidimensional time series data set.

[0063] In actual farmland scenarios where corn stalks are fully covered and returned to the field, the farmland enters a long evapotranspiration period after natural rainfall or artificial irrigation. During this time, the surface stalks directly exposed to the air rapidly dehydrate under wind and sunlight, while the underlying soil, due to the shading effect of the stalk cover, has its moisture loss to the atmosphere blocked, thus maintaining a high moisture content for an extended period. This physical structure leads to a situation where the surface layer is dry while the deeper soil layer is moist. At this point, the microorganisms attached to the surface stalks have essentially entered a dormant state due to dehydration, slowing down or even halting the stalk decomposition process. However, existing prediction algorithms typically rely solely on single temperature and humidity sensor data from the deep soil layer when extracting temporal features. Faced with this moisture inversion phenomenon, the algorithm continuously reads high humidity signals from the deep soil layer and uses this to determine that the current moisture driving force is strong, thus continuously accumulating an artificially inflated decomposition progress in the time integral calculation. This fundamental algorithmic flaw, which equates deep water retention errors with surface decomposition moisture, directly leads to severe overestimation of decomposition volume and supercomputing errors. To address this algorithmic disconnect caused by physical interface barriers, this step abandons static mapping relying on a single monitoring sequence and instead introduces a dynamic state game metric across interfaces. By simultaneously extracting data states from both the straw cover layer and the underlying soil layer, the moisture inversion dispersion between the two is quantified. Combined with the temperature gradient characterizing energy exchange intensity, an exponential decay suppression factor is constructed to counter spurious high-humidity signals. This factor is used to dynamically suppress ineffective driving forces detached from the real physical environment during single-step hydrothermal product calculations.

[0064] In summary, this invention first extracts interface hydrothermal inversion characterization data by performing cross-interface difference extraction processing on the cover layer moisture content data, soil layer moisture content data, cover layer temperature data, and soil layer temperature data in a synchronous multidimensional time-series dataset. Specifically, for any target acquisition time in the synchronous multidimensional time-series dataset, the volumetric moisture content data of the straw cover layer, the volumetric moisture content data of the underlying soil layer, the straw cover layer temperature data, and the underlying soil layer temperature data corresponding to the target acquisition time are extracted from the synchronous multidimensional time-series dataset. The difference between the volumetric moisture content data of the underlying soil layer and the volumetric moisture content data of the straw cover layer is used as the interface moisture difference assessment corresponding to the target acquisition time. When the interface moisture difference assessment is less than a constant 0, the interface moisture difference assessment corresponding to the target acquisition time is set to a constant 0, and the interface moisture difference assessment after non-negative screening is used as the interface moisture inversion difference data corresponding to the target acquisition time. The absolute value of the difference between the straw cover layer temperature data and the underlying soil layer temperature data is used as the interface temperature gradient data corresponding to the target acquisition time. The interface moisture inversion difference data and interface temperature gradient data are used as the interface hydrothermal inversion characterization data corresponding to the target acquisition time.

[0065] After obtaining the interface hydrothermal inversion characterization data, the data is further processed by exponential decay mapping to obtain the retardation decay coefficient. Specifically, for any target acquisition time, the interface moisture inversion difference data and interface temperature gradient data corresponding to the target acquisition time are obtained, as well as the volumetric moisture content data of the straw cover layer and the underlying soil layer. The square of the interface moisture inversion difference data is used as the interface moisture inversion intensity assessment corresponding to the target acquisition time; the product of the volumetric moisture content data of the underlying soil layer and the straw cover layer is used as the interface moisture coupling basis corresponding to the target acquisition time, and a preset small positive constant is added to the interface moisture coupling basis to obtain the interface moisture normalized basis corresponding to the target acquisition time. The result of dividing the interface moisture inversion intensity assessment by the interface moisture normalized basis is used as the interface moisture inversion penalty assessment corresponding to the target acquisition time. The result of adding the interface temperature gradient data to the constant 1 is processed by natural logarithmic mapping to obtain the temperature difference excitation term corresponding to the target acquisition time. The product of the interface moisture inversion penalty assessment and the temperature difference excitation term is used as the hydrothermal coupling penalty term corresponding to the target acquisition time. The negative of the hydrothermal coupling penalty term is subjected to exponential mapping with the natural constant as the base to obtain the stagnation attenuation coefficient corresponding to the target acquisition time.

[0066] In one implementation, assume the first The volumetric moisture content data of the straw cover layer at each sampling time were: ;No. The volumetric water content data of the underlying soil layer at each sampling time were: ;No. The straw cover temperature data at each sampling time were: ;No. The underlying soil temperature data at each sampling time were: The preset small positive constant is: Then the first The formula for calculating the attenuation coefficient at each acquisition moment is:

[0067]

[0068] in, Indicates the first The attenuation coefficient at each acquisition moment; This represents an exponential function with the natural constant e as the base. Indicates the first Data on the volumetric moisture content of the straw cover layer at each collection point; Indicates the first Volumetric moisture content data of the underlying soil layer at each sampling time; Indicates the first Temperature data of the straw cover layer at each collection time; Indicates the first Temperature data of the underlying soil layer at each collection time; This represents a tiny positive constant, set in the embodiments of the present invention. ; Represents the absolute value function; This represents the logarithmic function with the natural constant e as the base. This represents the maximum value function.

[0069] It should be noted that, since existing algorithms lack the ability to detect the blocking effect of interface isolation, this invention constructs the attenuation coefficient... The primary challenge is accurately identifying and quantifying the true degree of moisture inversion to filter out false high-humidity driving signals. To this end, the core of the formula is designed with... This is a dimensionless relative divergence penalty basis. When farmland has just experienced thorough rain, and the straw moisture content is greater than or equal to the soil moisture content, the physical environment is in a period of smooth positive conduction and full decomposition. The function's output is zero, causing the entire exponential term to become zero. equal This ensures that the computational load of the original algorithm is not affected at this stage. However, if the soil is wet and the straw is dry, The function captures this difference and amplifies the error nonlinearly through a squared term. This squared structure effectively filters out minor differences caused by sensor noise and imposes a high numerical penalty on severe inversions that truly cause decomposition stagnation. Simultaneously, to eliminate the influence of different absolute moisture content baselines across different plots, the formula introduces the cross-product of the two sequences as the denominator, allowing the penalty baseline to adapt to different soil textures. Furthermore, relying solely on the moisture inversion difference is insufficient to fully characterize the dormancy rate of surface microorganisms. At night or on rainy days, even with moisture inversion, the extremely low temperature difference results in weak surface evaporation, and the deterioration of the decomposition environment is relatively gradual. During normal daytime conditions, however, the surface temperature rises, and heat transfer leads to increased dehydration rates. Based on this physical causal relationship, the formula independently couples a natural logarithm excitation term driven by the temperature gradient. When the temperature difference is significant, this logarithmic term increases significantly, acting as an amplification factor directly on the aforementioned moisture-punishing substrate, thus affecting the external... The decay exhibits a rapid decline, forcing It rapidly approaches zero. It precisely transforms the misjudgment kinetic energy under a single deep high humidity illusion into an exponentially decaying penalty weight, cutting off the accumulation path of supercomputing errors in existing algorithms from the bottom up.

[0070] Thus, the analysis of cross-interface hydrothermal inverted transport resistance by synchronous multidimensional time-series data sets was completed, and the attenuation coefficient was obtained.

[0071] Step S3: Obtain the curvature compensation factor by performing pore evolution discrete phase space curvature analysis on the volumetric moisture content sequence of the straw cover layer.

[0072] After effectively suppressing the illusion of deep high humidity by introducing a retardation attenuation coefficient, the algorithm faces new challenges in the microscopic transition stage. The full-coverage stack of corn stalks exhibits a significant porous structure. Even after sufficient rainfall, when the surface begins to dry out, triggering the initial penalty of moisture inversion, the capillary resistance within the stalks still maintains a certain amount of liquid water. This internal water-holding phase allows for a certain degree of continuation of microbial activity and represents the actual effective timeframe for decomposition. If suppression relies solely on transient moisture difference comparisons, the algorithm will readily detect surface dehydration and prematurely impose penalties, misjudging the active decomposition kinetic energy within the stalks as stagnation, leading to underestimation of nitrogen release. To compensate for this hidden inertial kinetic energy, it is necessary to move beyond simply measuring absolute values ​​and further explore the evolutionary resistance of the moisture sequence itself. In this step, the first-order descent velocity and second-order resistance acceleration of moisture evolution are extracted in the local state space. By calculating the phase space curvature of the time-series curve during the descent process, the resistance of the straw's internal physical water-locking mechanism to the drying process is quantified, thereby incrementally compensating for the initial real driving force.

[0073] In summary, this invention first obtains moisture evolution curvature characterization data by performing local temporal difference processing on the straw cover layer volumetric moisture content sequence. Specifically, for any target sampling time in the straw cover layer volumetric moisture content sequence, the straw cover layer volumetric moisture content data corresponding to the target sampling time, the straw cover layer volumetric moisture content data corresponding to the sampling time before the target sampling time, and the straw cover layer volumetric moisture content data corresponding to the two sampling times before the target sampling time are obtained. The difference between the straw cover layer volumetric moisture content data corresponding to the target sampling time and the straw cover layer volumetric moisture content data corresponding to the sampling time before the target sampling time is used as the first-order temporal difference data corresponding to the target sampling time. The second-order temporal difference data for the target sampling time is calculated by subtracting twice the volumetric moisture content of the straw cover layer at the sampling time preceding the target sampling time from the volumetric moisture content of the straw cover layer at the sampling time preceding the target sampling time, and then adding the volumetric moisture content of the straw cover layer at the sampling times preceding the target sampling time. When the second-order temporal difference data is less than a constant 0, the second-order temporal difference data for the target sampling time is set to a constant 0, and the second-order temporal difference data after non-negative screening is used as the curvature resistance component data for the target sampling time. The first-order temporal difference data and the curvature resistance component data are used as the moisture evolution curvature characterization data for the target sampling time.

[0074] After obtaining the moisture evolution curvature characterization data, the historical water holding status statistical processing of the straw cover layer volumetric moisture content sequence is further performed to obtain historical water holding background characterization data. Specifically, for any target sampling time in the straw cover layer volumetric moisture content sequence, all straw cover layer volumetric moisture content data from the initial sampling time to the target sampling time are obtained. The sum of all straw cover layer volumetric moisture content data from the initial sampling time to the target sampling time is used as the historical water holding cumulative value corresponding to the target sampling time. The total number of sampling times from the initial sampling time to the target sampling time is obtained, and the historical water holding cumulative value divided by the total number of sampling times is used as the historical average water holding level corresponding to the target sampling time. The historical average water holding level and the straw cover layer volumetric moisture content data corresponding to the target sampling time are used as the historical water holding background characterization data corresponding to the target sampling time.

[0075] After obtaining historical water-holding background characterization data, a curvature compensation factor is obtained by coupling and compensating the water evolution curvature characterization data with the historical water-holding background characterization data. Specifically, for any target sampling time in the straw cover layer volumetric moisture content sequence, the first-order temporal difference data and curvature resistance component data corresponding to the target sampling time are obtained, along with the historical average water-holding level and straw cover layer volumetric moisture content data corresponding to the target sampling time. The square of the first-order temporal difference data and the square of the curvature resistance component data are added together, and a preset small positive constant is added to the sum to obtain the curvature normalization basis corresponding to the target sampling time. The result of dividing the curvature resistance component data by the square root of the curvature normalization basis is used as the curvature angle characterization value corresponding to the target sampling time. The curvature angle characterization value is processed by arcsine mapping, and the result of multiplying the arcsine mapping result by a constant two and dividing by pi is used as the local curvature compensation evaluation corresponding to the target sampling time. The historical average water holding capacity is divided by the volumetric moisture content of the straw cover layer at the target sampling time to obtain the historical water holding capacity amplification assessment at the target sampling time. The product of the local curvature compensation assessment and the historical water holding capacity amplification assessment is used as the curvature compensation factor at the target sampling time.

[0076] In one implementation, it is assumed that in the first... The volumetric moisture content data of the straw cover layer at each sampling time were: ;No. The volumetric moisture content data of the straw cover layer at each sampling time were: Then the first The formula for calculating the curvature compensation factor corresponding to each acquisition time is:

[0077]

[0078] in, Indicates the first Curvature compensation factor corresponding to each acquisition moment; The sequence of volumetric moisture content of straw mulch is represented in the th order. Second-order temporal difference data at each acquisition time; The sequence of volumetric moisture content of straw mulch is represented in the th order. First-order temporal difference data at each acquisition time; Indicates the first The historical water holding cumulative value corresponding to each collection moment; Represents the arcsine function; Represents pi; Represents the maximum value function; This represents a tiny positive constant, set in the embodiments of the present invention. .

[0079] It should be noted that, since the water-holding effect of the pores inside straw is difficult to observe directly using single-point numerical values, this application utilizes the calculus-topological properties of discrete sequences to evaluate water-locking resistance. The formula front end utilizes... Extracting the second difference reveals that when the straw begins to dehydrate but the internal capillaries resist moisture loss, the rate of moisture loss gradually slows down. At this point, the time series curve shows a concave shape, and the second difference is positive. The function extracts this component representing the resistance to water retention. To eliminate errors caused by the absolute dimensions of the sensor, the formula constructs a rectangular vector system for the first-order velocity and second-order acceleration, and extracts the angle of deceleration in phase space of this evolution trajectory through arcsine mapping. The more pronounced the concavity of the curve, the greater the internal resistance to water retention, and the closer this angular component is to the... The output coefficient at the front end increases accordingly. Furthermore, the effectiveness of local second-order kinetic energy must be judged in conjunction with the macroscopic historical water supply background; if it is merely a curve fluctuation caused by brief dew, no compensation should be generated. Therefore, the tail end of the formula... The structure calculates the ratio of historical average water holding capacity to current residual moisture. If there was a prolonged period of immersion followed by drying, this ratio will amplify the resistance curvature extracted at the front end, forming an effective... The product term compensates for the incremental change; conversely, if the historical context is insufficient, the product term will limit the compensation output of the pseudo-fluctuation. This structure achieves the coupling of micro-curvature changes and macro-hydrological context, objectively restoring the true water-locking inertial kinetic energy inside the straw.

[0080] Thus, the curvature compensation factor was obtained by performing discrete phase space curvature analysis on the pore evolution of the straw cover layer volume moisture content sequence.

[0081] Step S4: Obtain the actual decomposition driving variables by fusing and reconstructing the original soil hydrothermal integrated driving dependent variables, retardation attenuation coefficients and curvature compensation factors.

[0082] After obtaining the attenuation coefficient characterizing physical blockage and the curvature compensation factor characterizing internal water retention, it is necessary to integrate these two factors into the existing straw decomposition prediction algorithm. Existing decomposition algorithms typically generate single-step driving variables by calculating the combined response function of soil temperature and moisture at a single time point. To ensure a stable transition between different physical stages in the model, this step uses the attenuation coefficient as the basis for the penalty term and utilizes the negative exponent of the attenuation coefficient as a release gating. While suppressing deep-seated artifacts, it autonomously and smoothly releases the inertial compensation increment, thereby reconstructing a comprehensive driving variable that closely reflects the actual physical cover environment.

[0083] Specifically, for any target acquisition time, the original soil hydrothermal integrated driving dependent variable, retardation attenuation coefficient, and curvature compensation factor corresponding to the target acquisition time are obtained. The inverse of the retardation attenuation coefficient is subjected to exponential mapping with the natural constant as the base to obtain the gated activation term corresponding to the target acquisition time. The product of the curvature compensation factor and the gated activation term is taken as the compensation driving component corresponding to the target acquisition time. The sum of the retardation attenuation coefficient and the compensation driving component is taken as the reconstruction weight factor corresponding to the target acquisition time. The product of the original soil hydrothermal integrated driving dependent variable and the reconstruction weight factor is taken as the actual decomposition driving dependent variable corresponding to the target acquisition time.

[0084] It should be noted that when water conduction in farmland is smooth, the retardation attenuation coefficient is close to 1. In this case, the model primarily calculates the existing original driving dependent variables, while the exponential function generates a small constant decay, limiting the excessive intervention of the curvature compensation factor and preventing unnecessary incremental interference during normal decomposition. However, when farmland exhibits an inverted phenomenon of surface drought and deep moisture, the retardation attenuation coefficient rapidly decreases, approaching 0, to cut off the false accumulation caused by deep high moisture. At this point, as the retardation attenuation coefficient decreases, the value of the gating activation term increases in the opposite direction, approaching 1. The activation of this exponential gating smoothly switches the structure to a compensation-dominated mode, accurately and seamlessly superimposing the internal water-locking and drought-resistant kinetic energy represented by the curvature compensation factor onto the original variables. This design, without altering the main architecture of the existing algorithm, completely solves the overcomputing and undercomputation problems caused by interface retardation in fully covered micro-environments.

[0085] Furthermore, in existing straw decomposition prediction algorithms based on Cumulative Hydrothermal Time (CHT), the single-step driving dependent variables over the entire observation period (from time node 1 to T) are typically linearly accumulated, i.e., using the formula... The cumulative total driving force characterizing the decomposition stage is calculated. In the optimization process of this invention, the system obtains the actual decomposition driving dependent variable after reconstruction. Then, it directly replaces the static input item in the original algorithm's integration and accumulation module. Perform the corrected time-series integration operation to calculate the true cumulative hydrothermal time. By replacing variables at the single-step level, this application can fundamentally eliminate the invalid physical differences accumulated on the time axis by the original algorithm without changing the integral framework of the existing macroscopic algorithm, using the optimized multidimensional feature constraints.

[0086] Thus, the actual decomposition driving variables were obtained by fusing and reconstructing the original soil hydrothermal integrated driving dependent variables, retardation attenuation coefficients, and curvature compensation factors.

[0087] Step S5: Obtain fertilization decision instructions by calculating the cumulative decomposition process and assessing the fertilizer gap for the actual decomposition-driving dependent variables.

[0088] After reconstructing the driving variables at each time point, the system integrates and accumulates the reconstructed actual interface decomposition driving dependent variables along the time axis to obtain the true cumulative total decomposition progress of the straw cover layer at the current moment. Based on this cumulative decomposition progress, the total amount of available nitrogen actually released in the soil substrate due to straw degradation is calculated using an agronomic empirical mapping function, which serves as a dynamic evaluation indicator of the current farmland fertility recovery level.

[0089] Simultaneously, the system acquires the theoretical nitrogen requirement of maize seedlings at their current growth and development stage. By comparing the actual released available nitrogen with the theoretical nitrogen requirement of the maize seedlings over time, the system calculates the dynamic nutrient gap caused by competition for nitrogen from straw decomposition microorganisms. Finally, based on this nutrient gap, the system automatically generates and outputs precise compensatory nitrogen fertilizer application nodes and dosage decisions. This precise decision-making bridges the time lag in nutrient distribution under the mulch-and-return model, effectively avoiding nutrient loss or crop nitrogen starvation caused by indiscriminate fertilization, and achieving true synergy between straw return to the field for soil fertility release and maize seedling productivity assurance.

[0090] This completes the calculation of the cumulative decomposition process and the assessment of fertilizer demand gap by the actual decomposition-driving dependent variable, thus obtaining fertilization decision instructions.

[0091] 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 synergistically improving soil fertility and productivity through full coverage of corn stalks and returning them to the field, characterized in that, The method includes: Step S1: Collect and preprocess time-series monitoring data of the full coverage layer of corn stalks and the underlying soil in farmland to obtain a synchronous multidimensional time-series data set; Step S2: Obtain the stagnation attenuation coefficient by performing cross-interface hydrothermal inversion transport resistance analysis on the synchronous multidimensional time series data set; Step S3: Obtain the curvature compensation factor by performing discrete phase space curvature analysis on the volumetric moisture content sequence of the straw cover layer; Step S4: Obtain the actual decomposition driving dependent variable by fusing and reconstructing the original soil hydrothermal integrated driving dependent variable, retardation attenuation coefficient and curvature compensation factor; Step S5: Obtain fertilization decision instructions by calculating the cumulative decomposition process and assessing the fertilizer gap for the actual decomposition-driving dependent variables.

2. The method for synergistically improving soil fertility and productivity through full coverage and return of corn stalks to the field according to claim 1, characterized in that, The process involves collecting and preprocessing time-series monitoring data of the full corn stalk cover layer and the underlying soil in farmland to obtain a synchronous multidimensional time-series data set, including: In the target monitoring area of ​​farmland where corn stalks are fully covered and returned to the field, temperature and humidity sensor nodes are set up inside the corn stalks covering layer and in the shallow soil below the surface. Sampling time intervals are set and continuous monitoring is carried out according to the sampling time intervals to obtain time-series monitoring data of the corn stalks covering layer and the underlying soil. The time-series monitoring data of the full corn stalk cover layer and the underlying soil in the farmland shall include at least: volumetric moisture content data of the stalk cover layer, volumetric moisture content data of the underlying soil layer, temperature data of the stalk cover layer, and temperature data of the underlying soil layer. Data cleaning, outlier removal, missing value imputation, and timestamp alignment were performed on the volumetric moisture content data of the straw cover layer, the volumetric moisture content data of the underlying soil layer, the temperature data of the straw cover layer, and the temperature data of the underlying soil layer. The preprocessed data were then stored in association according to a unified sampling time order to obtain a synchronous multidimensional time-series data set.

3. The method for synergistically improving soil fertility and productivity through full coverage and return of corn stalks to the field according to claim 1, characterized in that, The method involves performing cross-interface hydrothermal inverted transport resistance analysis on a synchronous multidimensional time-series data set to obtain the stagnation and attenuation coefficient, including: By performing cross-interface difference extraction processing on the cover layer moisture content data, soil layer moisture content data, cover layer temperature data, and soil layer temperature data in the synchronous multidimensional time series dataset, interface hydrothermal inversion characterization data is obtained. The retardation decay coefficient was obtained by performing exponential decay mapping on the interfacial hydrothermal inversion characterization data.

4. The method for synergistically improving soil fertility and productivity through full coverage and return of corn stalks to the field according to claim 3, characterized in that, The process involves extracting cross-interface differences from overburden moisture content data, soil moisture content data, overburden temperature data, and soil temperature data in a synchronous multidimensional time-series dataset to obtain interfacial hydrothermal inversion characterization data, including: For any target acquisition time in the synchronous multidimensional time series data set, extract the straw cover layer volume moisture content data, the underlying soil layer volume moisture content data, the straw cover layer temperature data, and the underlying soil layer temperature data corresponding to the target acquisition time from the synchronous multidimensional time series data set. The difference between the volumetric moisture content data of the underlying soil layer and the volumetric moisture content data of the straw cover layer is used as the interface moisture difference assessment corresponding to the target collection time. When the interface moisture difference assessment is less than the constant 0, the interface moisture difference assessment corresponding to the target collection time is set to the constant 0, and the interface moisture difference assessment after non-negative screening is used as the interface moisture inversion difference data corresponding to the target collection time. The absolute value of the difference between the temperature data of the straw cover layer and the temperature data of the underlying soil layer is used as the interface temperature gradient data at the target acquisition time; the interface moisture inversion difference data and the interface temperature gradient data are used as the interface hydrothermal inversion characterization data at the target acquisition time.

5. The method for synergistically improving soil fertility and productivity through full coverage of corn stalks and returning them to the field, as described in claim 3, is characterized in that... The step of obtaining the retardation decay coefficient by performing exponential decay mapping processing on the interface hydrothermal inversion characterization data includes: For any target collection time, acquire the interface moisture inversion difference data and interface temperature gradient data corresponding to the target collection time, and acquire the straw cover layer volume moisture content data and the underlying soil layer volume moisture content data corresponding to the target collection time. The square of the interface moisture inversion difference data is used as the interface moisture inversion intensity assessment at the target acquisition time; the product of the volumetric moisture content data of the underlying soil layer and the volumetric moisture content data of the straw cover layer is used as the interface moisture coupling base at the target acquisition time, and a preset small positive constant is added to the interface moisture coupling base to obtain the interface moisture normalized base at the target acquisition time; the interface moisture inversion intensity assessment divided by the calculation result of the interface moisture normalized base is used as the interface moisture inversion penalty assessment at the target acquisition time. The interface temperature gradient data is added to the constant 1 and the result is processed by natural logarithmic mapping to obtain the temperature difference excitation term corresponding to the target acquisition time. The product of the interface moisture inversion penalty assessment and the temperature difference excitation term is used as the hydrothermal coupling penalty term corresponding to the target acquisition time. The negative of the hydrothermal coupling penalty term is processed by exponential mapping with the natural constant as the base to obtain the stagnation attenuation coefficient corresponding to the target acquisition time.

6. The method for synergistically improving soil fertility and productivity through full coverage and return of corn stalks to the field according to claim 1, characterized in that, The method involves performing discrete phase space curvature analysis on the pore evolution of the straw cover layer's volumetric moisture content sequence to obtain a curvature compensation factor, including: By performing local temporal difference processing on the volumetric moisture content sequence of straw cover layer, data representing the moisture evolution curvature were obtained. By statistically processing the historical water holding status of the straw cover layer volume moisture content sequence, historical water holding background characterization data is obtained. Curvature compensation factor is obtained by coupling and compensating water evolution curvature characterization data with historical water holding background characterization data.

7. The method for synergistically improving soil fertility and productivity through full coverage of corn stalks and returning them to the field, as described in claim 6, is characterized in that... The process of obtaining moisture evolution curvature characterization data by performing local temporal difference processing on the volumetric moisture content sequence of the straw cover layer includes: For any target sampling time in the straw cover layer volume moisture content sequence, obtain the straw cover layer volume moisture content data corresponding to the target sampling time, the straw cover layer volume moisture content data corresponding to the sampling time before the target sampling time, and the straw cover layer volume moisture content data corresponding to the two sampling times before the target sampling time. The difference between the volumetric moisture content of the straw cover layer at the target sampling time and the volumetric moisture content of the straw cover layer at the sampling time preceding the target sampling time is used as the first-order temporal difference data for the target sampling time. The second-order temporal difference data for the target sampling time is calculated by subtracting twice the volumetric moisture content of the straw cover layer at the sampling time preceding the target sampling time from the volumetric moisture content of the straw cover layer at the sampling time preceding the target sampling time, and then adding the volumetric moisture content of the straw cover layer at the sampling times preceding the target sampling time. When the second-order temporal difference data is less than a constant 0, the second-order temporal difference data for the target sampling time is set to a constant 0, and the second-order temporal difference data after non-negative screening is used as the curvature resistance component data for the target sampling time. The first-order temporal difference data and the curvature resistance component data are used as the moisture evolution curvature characterization data for the target sampling time.

8. The method for synergistically improving soil fertility and productivity through full coverage of corn stalks and returning them to the field, as described in claim 6, is characterized in that... The process involves statistically processing the historical water-holding status of the straw cover layer volumetric moisture content sequence to obtain historical water-holding background characterization data, including: For any target sampling time in the straw cover layer volumetric moisture content sequence, acquire all straw cover layer volumetric moisture content data from the start sampling time to the target sampling time; sum the total volumetric moisture content data from the start sampling time to the target sampling time as the historical water holding cumulative value corresponding to the target sampling time; acquire the total number of sampling times from the start sampling time to the target sampling time, and divide the historical water holding cumulative value by the total number of sampling times as the historical average water holding level corresponding to the target sampling time; use the historical average water holding level and the straw cover layer volumetric moisture content data corresponding to the target sampling time as the historical water holding background characterization data corresponding to the target sampling time.

9. A method for synergistically improving soil fertility and productivity through full coverage of corn stalks and returning them to the field, as described in claim 6, is characterized in that... The process of coupling and compensating water evolution curvature characterization data with historical water-holding background characterization data to obtain curvature compensation factors includes: For any target sampling time in the straw cover layer volumetric moisture content sequence, obtain the first-order temporal difference data and curvature resistance component data corresponding to the target sampling time, and obtain the historical average water holding level and straw cover layer volumetric moisture content data corresponding to the target sampling time. The squares of the first-order temporal difference data and the squares of the curvature resistance component data are added together, and a preset small positive constant is added to the sum to obtain the curvature normalization basis corresponding to the target acquisition time. The curvature resistance component data is divided by the square root of the curvature normalization basis as the curvature angle characterization value corresponding to the target acquisition time. The curvature angle characterization value is processed by arcsine mapping, and the result of multiplying the arcsine mapping result by a constant two and dividing by pi is used as the local curvature compensation evaluation corresponding to the target acquisition time. The historical average water holding level is divided by the straw cover layer volume moisture content data corresponding to the target acquisition time as the historical water holding amplification evaluation corresponding to the target acquisition time. The product of the local curvature compensation evaluation and the historical water holding amplification evaluation is used as the curvature compensation factor corresponding to the target acquisition time.

10. A method for synergistically improving soil fertility and productivity through full coverage and return of corn stalks to the field according to claim 1, characterized in that, The actual decomposition driving variables are obtained by fusing and reconstructing the original soil hydrothermal integrated driving dependent variables, retardation attenuation coefficients, and curvature compensation factors, including: For any target acquisition time, obtain the original soil hydrothermal integrated driving dependent variable, retardation attenuation coefficient, and curvature compensation factor corresponding to the target acquisition time; perform exponential mapping processing on the negative of the retardation attenuation coefficient with the natural constant as the base to obtain the gated activation term corresponding to the target acquisition time; take the product of the curvature compensation factor and the gated activation term as the compensation driving component corresponding to the target acquisition time; take the calculation result of adding the retardation attenuation coefficient and the compensation driving component as the reconstruction weight factor corresponding to the target acquisition time; take the product of the original soil hydrothermal integrated driving dependent variable and the reconstruction weight factor as the actual decomposition driving dependent variable corresponding to the target acquisition time.

Images