An alpine sandy soil moisture infiltration amount estimation method
By integrating multi-source data and improving the Hydrus-1D model, the accuracy problem of soil moisture infiltration estimation in high-altitude sandy areas has been solved, achieving high-precision spatiotemporal continuous estimation, which is applicable to water resource management and ecological restoration in high-altitude sandy areas.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- QINGHAI UNIVERSITY
- Filing Date
- 2025-06-26
- Publication Date
- 2026-07-21
AI Technical Summary
Existing technologies are insufficient to accurately estimate soil moisture infiltration in high-altitude sandy areas, especially under the influence of freeze-thaw cycles and biological crusting. Furthermore, traditional methods cannot achieve continuous spatiotemporal estimation of the 0-110cm soil layer, resulting in significant errors in the estimation results.
A multi-source data fusion method was adopted, combining cold-resistant soil hydrothermal sensor data and Sentinel-2 remote sensing data. The permeability coefficient and water-blocking factor were calculated by using an improved Hydrus-1D model and a multi-expression programming model. The spatiotemporal continuous estimation was performed using a random forest algorithm and a deep learning model, combined with the freeze-thaw-porosity dynamic relationship, and iterative correction was performed through error correction.
It achieves high-precision spatiotemporal continuous estimation of seepage in soil layers from 0 to 110 cm, with a root mean square error (RMSE) ≤ 0.048 m³/m³. It is suitable for extreme freeze-thaw environments and supports water resource management and ecological restoration engineering design.
Smart Images
Figure SMS_1 
Figure SMS_2 
Figure SMS_3
Abstract
Description
Technical Field
[0001] This invention relates to the field of soil moisture infiltration estimation technology in high-altitude sandy areas, specifically to a method for estimating soil moisture infiltration in high-altitude sandy areas. Background Technology
[0002] In alpine sandy ecosystems, soil moisture infiltration is a key parameter of the water resource cycle. Its accurate estimation is of great significance for regional water resource management, ecological restoration engineering design, and climate change response. The unique low-temperature environment and frequent freeze-thaw cycles of alpine sandy areas, coupled with the significant impact of biological crusts on soil hydrological processes, make the soil moisture infiltration mechanism in this region extremely complex.
[0003] Traditional methods for estimating soil moisture infiltration, such as the Hydrus-1D model, are largely based on steady-state flow assumptions and empirical parameters, failing to adequately consider the dynamic impact of freeze-thaw cycles on soil pore structure and permeability. Studies have shown that freeze-thaw cycles lead to soil particle redistribution, causing significant changes in porosity at soil depths of 30-80 cm, thus affecting moisture infiltration characteristics. However, existing methods struggle to accurately quantify these changes. Furthermore, the widespread presence of biocrusts in high-altitude sandy areas exhibits a significant water-blocking effect; when biocrust coverage exceeds 30%, soil infiltration rates can decrease by 23.41%-30.39%. Traditional models typically simplify this process, leading to substantial errors in the estimation results.
[0004] In terms of data acquisition and processing, existing technologies mostly rely on single-point in-situ monitoring or a single remote sensing data source, which cannot achieve continuous spatiotemporal estimation of seepage in the 0-110cm soil layer. Furthermore, traditional interpolation methods and statistical models struggle to effectively integrate multi-source heterogeneous data, failing to meet the high-precision estimation requirements of complex environments in high-altitude sandy areas. Therefore, there is an urgent need for a method that can comprehensively consider special factors such as freeze-thaw cycles and biological crusting, and combine multi-source data to achieve accurate spatiotemporal estimation, thus filling the gap in existing technologies for studying soil moisture seepage in high-altitude sandy areas. Summary of the Invention
[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for estimating soil moisture infiltration in high-altitude sandy areas, comprising the following steps:
[0006] S1. Multi-source data acquisition: Use cold-resistant soil hydrothermal sensor array to collect soil moisture content and temperature data in the 0-110cm layer, and simultaneously acquire Sentinel-2 remote sensing data to extract vegetation cover and surface temperature, and record the number of freeze-thaw cycles and biological crust cover types in the region.
[0007] S2. Key parameter inversion: Based on the number of freeze-thaw cycles N, the permeability coefficient correction factor k_FTC=α・N^(1 / 2) is calculated through a multi-expression programming model, where α is the soil texture coefficient. The water-blocking factor β is determined according to the biological crust coverage. When the coverage is >30%, β=0.72-0.85.
[0008] S3. Coupled model construction: The correction factor from step S2 is embedded into the improved Hydrus-1D model to establish a water balance equation that considers freeze-thaw and porosity dynamics.
[0009] S4. Leakage Calculation: The leakage rate of each soil layer is solved iteratively by combining the Koskoskov formula. The leakage of the 0-110cm soil layer is continuously estimated in time and space by fusing remote sensing and in-situ data through the random forest algorithm.
[0010] S5. Error correction: By comparing the measured porosity values from CT scans with the model predictions, when the deviation is >15%, k_FTC and β are iteratively corrected.
[0011] Preferably, the cold-resistant sensor array in step S1 includes TDR sensors buried at depths of 10, 30, 50, 70, 90, and 110 cm, with a data acquisition interval of ≤30 min, and the sensors have the ability to operate at temperatures from -30°C to 50°C.
[0012] Preferably, the method for determining the permeability coefficient correction factor in step S2 includes:
[0013] S21. Conduct freeze-thaw cycle tests on undisturbed soil samples;
[0014] S22. Use a laser particle size analyzer to test the particle distribution of soil samples and calculate the changes in sand, silt, and clay content Δm.
[0015] S23. Use the MEP algorithm to establish a permeability coefficient variation model.
[0016] Preferably, the freeze-thaw-porosity dynamic relationship in step S3 is determined by CT scanning. When the number of freeze-thaw cycles N≤10, the total porosity of the soil first decreases and then increases. There is a porosity-sensitive zone at a soil depth of 30-80cm. The relationship between the porosity change rate Δn and the number of freeze-thaw cycles is: Δn=0.017⋅sin(0.3N)+0.023.
[0017] Preferably, the multi-source data fusion algorithm in step S4 adopts the DBN deep learning model. The input variables include NDVI downscaled to 30m, surface temperature, and initial soil moisture content. The output is daily seepage data at a resolution of 30m, with a root mean square error RMSE ≤ 0.048m³ / m³.
[0018] A system for estimating soil moisture infiltration in high-altitude sandy areas includes:
[0019] The data acquisition module is used to acquire soil hydrothermal data, remote sensing data, and freeze-thaw cycle parameters;
[0020] The parameter inversion module is used to calculate the freeze-thaw correction factor and the biological crust water-blocking factor;
[0021] The model calculation module integrates the improved Hydrus-1D model with a multi-source data fusion algorithm and implements the specific method for spatiotemporal continuous estimation described in S4.
[0022] The output module generates spatiotemporally continuous leakage estimation results.
[0023] Preferably, the parameter inversion module has a built-in MEP algorithm engine, which can calculate the permeability coefficient correction value in real time based on the number of freeze-thaw cycles and changes in soil particle size.
[0024] Preferably, the model calculation module supports user-defined input of high-altitude sandy land type parameters, and when applied to mobile sand dunes, it automatically enables the 110cm depth infiltration threshold correction parameter.
[0025] It has the following beneficial effects:
[0026] This invention quantifies the dynamic effect of freeze-thaw cycles on the porosity of soil layers 30-80cm deep and the water-blocking effect of biocrust coverage exceeding 30% (23.41%-30.39%). It overcomes the limitations of traditional models based on steady-state flow and empirical parameters, accurately characterizing the impact of low-temperature freeze-thaw cycles and biocrust on seepage mechanisms in high-altitude sandy areas. Compared to traditional methods, it exhibits lower error. Furthermore, by fusing multi-source data from Sentinel-2 remote sensing data and a cold-resistant sensor array, combined with algorithms such as inverse distance weighted spatial interpolation and LSTM time series prediction, it achieves the ability to measure the permeability of soil layers 0-110cm deep. The system provides continuous spatiotemporal estimation of daily seepage at a resolution of 30m, with a root mean square error (RMSE) ≤ 0.048 m³ / m³, demonstrating high accuracy in predicting extreme freeze-thaw events. Furthermore, it is the first to embed innovative parameters such as the power function relationship between freeze-thaw cycles and permeability coefficients, and the biological crust impedance function into the Hydrus-1D model, forming a specialized estimation system applicable to low-temperature environments ranging from -6.0 to 2.7℃. This provides crucial technical support for water resource management, ecological restoration engineering design, and climate change response strategies in high-altitude sandy areas with an annual potential evaporation of 1500-2400 mm, such as the northeastern Qinghai-Tibet Plateau. Detailed Implementation
[0027] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0028] This invention provides a technical solution: a method for estimating soil moisture infiltration in high-altitude sandy areas, characterized by comprising the following steps:
[0029] S1. Multi-source data acquisition: The cold-resistant soil hydrothermal sensor array is used to collect soil moisture content and temperature data in the 0-110cm layer. Sentinel-2 remote sensing data is acquired simultaneously to extract vegetation cover and surface temperature. The number of freeze-thaw cycles and the type of biological crust cover in the region are recorded. The cold-resistant sensor array includes TDR sensors buried at depths of 10, 30, 50, 70, 90 and 110cm. The data acquisition interval is ≤30min, and the sensors have the ability to work from -30℃ to 50℃.
[0030] S2. Key Parameter Inversion: Based on the number of freeze-thaw cycles N, the permeability coefficient correction factor k_FTC=α・N^(1 / 2) is calculated using a multi-expression programming model, where α is the soil texture coefficient. The water-blocking factor β is determined based on the biological crust coverage. When the coverage > 30%, β = 0.72-0.85. The methods for determining the permeability coefficient correction factor include:
[0031] S21. Conduct freeze-thaw cycle tests on undisturbed soil samples. The specific steps are as follows:
[0032] S211. Sample collection and preparation: Select a representative area in the high-altitude sandy land, collect undisturbed soil samples from the 0-110cm soil layer, cut the soil samples into cylinders with a diameter of 5cm and a height of 10cm, and prepare 3 parallel samples for each soil layer.
[0033] S212. Initial parameter determination: The initial bulk density of the soil sample was determined by the ring cutter method, the initial moisture content was determined by the drying method, and the initial pore structure parameters of the soil sample were obtained by the mercury intrusion porosimetry method.
[0034] S213. Freeze-thaw cycle setting: Place the soil sample in a high and low temperature alternating test chamber, set the freezing temperature to -15℃, freezing time to 12 hours, and thawing temperature to 5℃, thawing time to 12 hours, to form a complete freeze-thaw cycle.
[0035] S214. Cyclic process monitoring: During each freeze-thaw cycle, the soil sample temperature is recorded every 2 hours, and the volume change of the soil sample during the freeze-thaw process is monitored using a displacement sensor.
[0036] S215. Post-cycle treatment: After completing the predetermined number of freeze-thaw cycles (1, 3, 5, 7, 10 times), take out the soil sample, air dry it naturally to constant weight, and measure the bulk density, moisture content and pore structure parameters of the soil sample again.
[0037] S22. Use a laser particle size analyzer to test the particle distribution of soil samples and calculate the changes in sand, silt, and clay content Δm.
[0038] S23. Establish a permeability coefficient variation model using the MEP algorithm:
[0039]
[0040] Where k0 is the initial permeability coefficient, and a1, a2, and a3 are the fitting coefficients;
[0041] S3. Coupled Model Construction: The correction factor from step S2 is embedded into the improved Hydrus-1D model to establish a water balance equation considering freeze-thaw and porosity dynamics.
[0042]
[0043] Where k(θ) is the temperature-corrected permeability coefficient, ψ is the soil water potential, and f(crust) is the biological crust resistance function;
[0044] The dynamic relationship between freeze-thaw cycles and porosity was determined by CT scan. When the number of freeze-thaw cycles N≤10, the total porosity of the soil first decreases and then increases. There is a porosity-sensitive zone in the soil layer with a depth of 30-80cm. The relationship between the porosity change rate Δn and the number of freeze-thaw cycles is: Δn=0.017⋅sin(0.3N)+0.023;
[0045] S4. Leakage Calculation: The leakage rate of each soil layer is iteratively solved using the Koskoskov formula. A random forest algorithm is then used to fuse remote sensing and in-situ data to continuously estimate the leakage rate of the 0-110cm soil layer in time and space. The specific method is as follows:
[0046] S41. The fused multi-source data is preprocessed, and abnormal data is removed using a quality control algorithm. Missing data is filled in using spatiotemporal interpolation, where spatial interpolation uses the inverse distance weighting method and temporal interpolation uses the cubic spline interpolation method.
[0047] S42. Divide the preprocessed data into a 30m×30m grid and establish a spatial grid database;
[0048] S43. Based on historical data, the leakage time series is decomposed using a seasonal decomposition algorithm to separate the trend term, seasonal term, and random term.
[0049] S44. For each spatial grid, a Long Short-Term Memory (LSTM) network model is used, with the decomposed trend term, seasonal term and environmental parameters of the current grid (vegetation coverage, surface temperature, initial soil moisture content, etc.) as inputs to predict the seepage of each soil layer in the future time period.
[0050] S45. Based on the prediction results of each spatial grid, generate spatiotemporal continuous data of daily seepage volume in the 0-110cm soil layer with a resolution of 30m.
[0051] The multi-source data fusion algorithm uses the DBN deep learning model. The input variables include NDVI downscaled to 30m, surface temperature, and initial soil moisture content. The output is daily seepage data at a resolution of 30m, with a root mean square error RMSE ≤ 0.048m³ / m³.
[0052] S5. Error Correction: By comparing the measured porosity values from CT scans with the model predictions, when the deviation exceeds 15%, k_FTC and β are iteratively corrected. The correction formula is as follows:
[0053]
[0054] Where γ is the correction coefficient, n_CT is the CT measured porosity, and n_model is the model calculated porosity.
[0055] Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art and related fields based on the embodiments of the present invention without inventive effort should fall within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described and explained in the present invention, unless otherwise specified or limited, shall be implemented according to conventional means in the art.
Claims
1. A method for estimating soil moisture infiltration in high-altitude sandy areas, characterized in that, Includes the following steps: S1. Multi-source data acquisition: Use cold-resistant soil hydrothermal sensor array to collect soil moisture content and temperature data in the 0-110cm layer, and simultaneously acquire Sentinel-2 remote sensing data to extract vegetation cover and surface temperature, and record the number of freeze-thaw cycles and biological crust cover types in the region. S2. Key parameter inversion: Based on the number of freeze-thaw cycles N, the permeability coefficient correction factor k_FTC=α•N^(1 / 2) is calculated through a multi-expression programming model, where α is the soil texture coefficient. The water-blocking factor β is determined according to the biological crust coverage. When the coverage is >30%, β=0.72-0.
85. S3. Coupled Model Construction: The correction factor from step S2 is embedded into the improved Hydrus-1D model to establish a water balance equation considering freeze-thaw and porosity dynamics. Where k(θ) is the temperature-corrected permeability coefficient, ψ is the soil water potential, and f(crust) is the biological crust resistance function; The dynamic relationship between freeze-thaw cycles and porosity was determined by CT scan. When the number of freeze-thaw cycles N≤10, the total porosity of the soil first decreases and then increases. There is a porosity-sensitive zone in the soil layer with a depth of 30-80cm. The relationship between the porosity change rate Δn and the number of freeze-thaw cycles is: Δn=0.017⋅sin(0.3N)+0.023; S4. Leakage Calculation: The leakage rate of each soil layer is solved iteratively by combining the Koskoskov formula. The leakage of the 0-110cm soil layer is continuously estimated in time and space by fusing remote sensing and in-situ data through the random forest algorithm. S5. Error correction: By comparing the measured porosity values from CT scans with the model predictions, when the deviation is >15%, k_FTC and β are iteratively corrected.
2. The method for estimating soil moisture infiltration in high-altitude sandy areas according to claim 1, characterized in that, The cold-resistant sensor array in step S1 includes TDR sensors buried at depths of 10, 30, 50, 70, 90, and 110 cm, with a data acquisition interval of ≤30 min, and the sensors have the ability to operate from -30℃ to 50℃.
3. The method for estimating soil moisture infiltration in high-altitude sandy areas according to claim 1, characterized in that, The method for determining the permeability correction factor in step S2 includes: S21. Conduct freeze-thaw cycle tests on undisturbed soil samples; S22. Use a laser particle size analyzer to test the particle distribution of soil samples and calculate the changes in sand, silt, and clay content Δm. S23. Use the MEP algorithm to establish a permeability coefficient variation model.
4. The method for estimating soil moisture infiltration in high-altitude sandy areas according to claim 1, characterized in that, In step S3, the dynamic relationship between freeze-thaw and porosity is determined by CT scanning. When the number of freeze-thaw cycles N ≤ 10, the total porosity of the soil first decreases and then increases. There is a porosity-sensitive zone at a soil depth of 30-80cm. The relationship between the porosity change rate Δn and the number of freeze-thaw cycles is: Δn = 0.017⋅sin(0.3N) + 0.
023.
5. The method for estimating soil moisture infiltration in high-altitude sandy areas according to claim 1, characterized in that, In step S4, the multi-source data fusion algorithm adopts the DBN deep learning model. The input variables include NDVI downscaled to 30m, surface temperature, and initial soil moisture content. The output is daily seepage data at a resolution of 30m, with a root mean square error RMSE ≤ 0.048m³ / m³.
6. A system for estimating soil moisture infiltration in high-altitude sandy areas, implementing the method for estimating soil moisture infiltration in high-altitude sandy areas as described in any one of claims 1-5, characterized in that, include: The data acquisition module is used to acquire soil hydrothermal data, remote sensing data, and freeze-thaw cycle parameters; The parameter inversion module is used to calculate the freeze-thaw correction factor and the biological crust water-blocking factor; The model computation module integrates the improved Hydrus-1D model with a multi-source data fusion algorithm; The output module generates spatiotemporally continuous leakage estimation results.
7. The soil moisture infiltration estimation system for high-altitude sandy areas according to claim 6, characterized in that, The parameter inversion module has a built-in MEP algorithm engine, which can calculate the permeability coefficient correction value in real time based on the number of freeze-thaw cycles and changes in soil particle size.
8. A soil moisture infiltration estimation system for high-altitude sandy areas according to claim 6, characterized in that, The model calculation module supports user-defined input parameters for high-altitude sandy terrain types. When applied to mobile sand dunes, it automatically enables the 110cm depth infiltration threshold correction parameter.