Sand eco-hydrological process critical threshold identification system and applications thereof

Abstract

The present application provides a sand ecological-hydrological process critical threshold identification system and application thereof, comprising a multi-parameter sample collection device, a soil physicochemical index determination module, a vegetation water stress response monitoring module, a critical threshold calculation unit and a steady state transition early warning module; by determining soil aggregate stability index data and soil water characteristic curve data, monitoring vegetation transpiration conductance data and leaf water potential data, using multi-level lag response analysis and dynamic response surface construction method, calculating critical water content threshold value and steady state transition critical point data, accurate identification and dynamic early warning of sand ecological system steady state transition are realized.

Description

Technical Field

[0001] This invention relates to the fields of environmental science and ecohydrology, specifically to a critical threshold identification system for sandy land eco-hydrological processes and its application. Background Technology

[0002] Sandy ecosystems are a typical fragile ecosystem type in arid and semi-arid regions, and their stability is profoundly affected by hydrological processes. Driven by both climate change and human activities, sandy ecosystems face the risk of transitioning from one steady state to another. Accurately identifying the critical point of this transition is of great significance for ecosystem protection and management.

[0003] In existing technologies, numerous studies have been conducted on methods for determining ecological water level thresholds. Taking Chinese invention CN118275651A as an example, it discloses a method for determining groundwater ecological hydrological thresholds. This method measures the maximum water absorption depth of tracer plants through field plant water absorption depth tracing experiments, measures the maximum evaporation depth of soil moisture in the study area and the maximum depth at which soil moisture is replenished to plant roots through vertical transport through field soil moisture transport experiments, and then uses soil column experiments to find the groundwater level depth that can transport groundwater to a specific height, thereby determining the range of groundwater ecological water levels in the study area. This method, based on hydrogen and oxygen isotope tracing and soil column capillary experiments, can determine a reasonable range of groundwater levels, which is of guiding significance for preventing soil salinization and maintaining vegetation survival.

[0004] However, the existing technologies have the following shortcomings. First, existing methods mainly focus on the static threshold range of groundwater levels, failing to reveal the dynamic critical characteristics of the transition from a stable to a degraded state in sandy ecosystems, making it difficult to predict and warn of steady-state transitions. Second, existing methods analyze soil moisture transport and vegetation water absorption as relatively independent processes, failing to establish a synergistic correlation mechanism between soil structural stability and vegetation water stress response, resulting in a lack of systematic judgment of critical thresholds. Third, existing methods rely on discrete measurement methods such as isotope tracing and soil column experiments, which are time-consuming and costly, making it difficult to achieve real-time monitoring and dynamic early warning of critical states. Fourth, existing methods lack consideration of soil aggregate stability, a key soil structural indicator, which directly affects the soil's water-holding capacity and erosion resistance, and is an important foundation for the stability of sandy ecosystems.

[0005] Therefore, it is necessary to develop a system that can comprehensively analyze soil aggregate stability, soil moisture characteristics, and vegetation water stress response to achieve accurate identification and dynamic early warning of the critical point of steady-state transition in sandy ecosystems. Summary of the Invention

[0006] The purpose of this invention is to provide a critical threshold identification system for sandy land eco-hydrological processes and its application, so as to solve the above-mentioned technical problems existing in the prior art.

[0007] To achieve the above objectives, the present invention provides a critical threshold identification system for sandy land eco-hydrological processes, including a multi-parameter sample collection device, a soil physicochemical index measurement module, a vegetation water stress response monitoring module, a critical threshold calculation unit, and a steady-state transition early warning module.

[0008] A multi-parameter sample collection device is configured to collect soil and vegetation samples at different depths within the sandy study area and monitor in-situ environmental parameters in the collected area. A soil physicochemical index measurement module is connected to the multi-parameter sample collection device and configured to receive soil samples and measure soil aggregate stability index data and soil moisture characteristic curve data. A vegetation water stress response monitoring module is configured to monitor the water physiological state of target vegetation within the study area and output vegetation transpiration conductance data and leaf water potential data. A critical threshold calculation unit is connected to both the soil physicochemical index measurement module and the vegetation water stress response monitoring module, and is configured to calculate the critical water content threshold value and steady-state transition critical point data based on the synergistic analysis of the response characteristics of soil aggregate stability with water changes and the vegetation water stress response characteristics. A steady-state transition early warning module is connected to the critical threshold calculation unit and is configured to compare real-time monitored environmental parameters with the critical threshold and output an early warning signal.

[0009] This invention also provides applications of the above-mentioned system, including sample collection steps, soil index measurement steps, vegetation response monitoring steps, critical threshold calculation steps, and early warning judgment steps. Through multi-level hysteresis response analysis and dynamic response surface construction, accurate identification of the critical point of steady-state transition in sandy ecosystems is achieved.

[0010] This invention offers the following advantages. First, it innovatively combines soil aggregate stability analysis with vegetation water stress response monitoring, establishing a critical threshold identification system for soil-vegetation synergistic response, overcoming the limitations of single-indicator judgment in existing technologies. Second, the multi-level hysteresis response analysis method proposed in this invention can reveal the coupling relationship between soil structure changes and vegetation responses at different time scales, providing a temporal basis for predicting steady-state transitions. Third, the dynamic response surface constructed in this invention can comprehensively reflect the multi-dimensional relationship between soil moisture content, aggregate stability, and vegetation stress level, achieving precise location of critical points. Fourth, the steady-state transition early warning module of this invention can provide dynamic early warnings based on real-time monitoring data, exhibiting higher timeliness and practicality compared to existing discrete measurement methods. Attached Figure Description

[0011] Figure 1This is an architecture diagram of the critical threshold identification system for sandy land ecological-hydrological processes provided in an embodiment of the present invention.

[0012] Figure 2 This is a flowchart of the identification method provided in the embodiments of the present invention. Detailed Implementation

[0013] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0014] Reference Figure 1 This embodiment provides a critical threshold identification system for sandy land eco-hydrological processes. The system includes a multi-parameter sample collection device, a soil physicochemical index measurement module, a vegetation water stress response monitoring module, a critical threshold calculation unit, and a steady-state transition early warning module.

[0015] The multi-parameter sample acquisition device is the data acquisition front end of this system. It is configured to collect soil and vegetation samples for analysis within the sandy study area, while simultaneously monitoring in-situ environmental parameters in real time. The device consists of two main parts: a stratified sampler assembly and an in-situ sensor array.

[0016] The stratified sampler assembly is used to collect soil samples vertically at preset depth intervals. In sandy ecosystems, soil properties exhibit significant vertical differentiation with depth. Topsoil is significantly affected by atmospheric evaporation and vegetation root activity, while deeper soil is primarily influenced by groundwater level fluctuations. To comprehensively characterize this vertical differentiation, the stratified sampler assembly performs stratified sampling at 20-centimeter depth intervals, extending from the surface to 50 centimeters below the groundwater level, ensuring complete coverage of the soil profile. At least 500 grams of soil sample are collected at each sampling depth to meet the needs of subsequent aggregate stability analysis and moisture characteristic curve determination. The sampling tool consists of a stainless steel ring cutter and a soil drill. The ring cutter has an inner diameter of 50 millimeters and a height of 51 millimeters, used for collecting undisturbed soil samples for aggregate analysis; the soil drill has a diameter of 38 millimeters, used for collecting disturbed soil samples for moisture characteristic curve determination.

[0017] The in-situ sensor array is configured for real-time monitoring of soil temperature, soil conductivity, and soil moisture content. The sensor array is deployed at the same depth as the stratified sampling depth, with one set of sensor probes at each monitoring depth. The soil temperature sensor is a thermistor type, with a measurement accuracy of ±0.1 degrees Celsius and a measurement range of -20 degrees Celsius to 60 degrees Celsius. The soil conductivity sensor uses the four-electrode method, with a measurement accuracy of ±2% and a measurement range of 0 to 20 millisiemens per centimeter. The soil moisture content sensor uses the time-domain reflectometry principle, with a measurement accuracy of ±0.02 cubic meters per cubic meter and a measurement range of 0 to 0.60 cubic meters per cubic meter. The sensor array's data acquisition frequency is set to once every 15 minutes, and the data is uploaded to the data processing center in real time via a wireless transmission module.

[0018] The soil physicochemical index determination module is connected to the multi-parameter sample collection device, receives the collected soil samples, performs aggregate stability analysis and moisture characteristic curve determination, and outputs soil aggregate stability index data and soil moisture characteristic curve data.

[0019] The soil aggregate stability index was determined using the wet sieving method. The specific procedure is as follows: First, the collected soil samples were naturally air-dried indoors until the soil moisture content was approximately 60% of field capacity. At this state, the soil structure would not crack due to excessive dryness, nor deform due to excessive wetness. Then, the air-dried soil sample was sieved through an 8 mm sieve, and 50 grams of soil with a particle size of 2 to 8 mm were taken as the test sample. The test sample was placed on the sieve assembly of the aggregate analyzer, which included stainless steel sieves with apertures of 5 mm, 2 mm, 1 mm, 0.5 mm, and 0.25 mm. The sieve assembly was immersed in distilled water and shaken for 30 minutes at an amplitude of 30 mm and a frequency of 30 times per minute. After shaking, the aggregates on each sieve were rinsed into an aluminum box and dried at 105°C to constant weight. The mass of aggregates of each particle size was then weighed. Based on the test results, the mass percentage of water-stable aggregates larger than 0.25 mm was calculated. This percentage is a key indicator for evaluating soil structural stability.

[0020] Based on the results of wet sieving, the average weight diameter and geometric mean diameter were calculated as quantitative indicators of soil aggregate stability. The average weight diameter was calculated as the sum of the products of the average diameter for each particle size interval and the mass fraction of the aggregates in that interval. The geometric mean diameter was calculated by taking the weighted sum of the logarithmic values ​​of the average diameter for each particle size interval and the mass fraction, followed by taking the exponent. Larger values ​​for these two indicators indicate stronger soil aggregate stability and better soil structure. In sandy ecosystems, when the average weight diameter is below 0.5 mm, the soil structure tends to be unstable, and its resistance to wind and water erosion decreases; when the average weight diameter is above 1.0 mm, the soil structure is stable.

[0021] Soil moisture characteristic curve data were determined using the pressure membrane apparatus method. The pressure membrane apparatus gradually removes water from the soil by applying different air pressures, measuring the equilibrium water content of the soil under different pressures to obtain the correlation between soil water potential and water content. During the measurement, soil samples were placed in a 100 cubic centimeter ring sampler, saturated for 24 hours, and then placed in a pressure chamber. Air pressures of 0.01 MPa, 0.03 MPa, 0.05 MPa, 0.1 MPa, 0.3 MPa, 0.5 MPa, 1.0 MPa, and 1.5 MPa were applied sequentially. After equilibration at each pressure level for 48 hours, the samples were weighed, and the corresponding soil volumetric water content was calculated. The measurement results were fitted using the van Genuchten model to obtain the model parameters describing the soil moisture characteristic curve. The van Genuchten model expresses soil volumetric water content as a function of soil water potential, and the functional form includes residual water content, saturated water content, and parameters reflecting the characteristics of soil pore distribution. By fitting the model, the predicted value of soil moisture content under any water potential condition can be obtained, providing continuous data support for subsequent critical threshold calculation.

[0022] The vegetation water stress response monitoring module is configured to monitor the water physiological status of target vegetation within the study area and output vegetation transpiration conductance data and leaf water potential data. The selection criteria for target vegetation are dominant or constructive species that are widely distributed within the study area and sensitive to water changes.

[0023] Vegetation transpiration conductance data were acquired using a combination of stomatalometer and thermal diffusion probe. The stomatalometer measures leaf stomatal conductance by measuring the diffusion rate of air through a leaf area. During measurement, fully expanded mature leaves from the top layer of the target vegetation were selected, with 5 to 10 leaves measured per plant. The average value was taken as the stomatal conductance for that plant. The thermal diffusion probe measures trunk sap flow rate. The probe is installed on the sunny side of the trunk at a height of 1.3 meters above the ground. The probe is 20 mm long, with a 40 mm gap between the upper and lower probes. The upper probe is the heating probe, and the lower probe is the reference probe. The trunk sap flow density is calculated by measuring the temperature difference between the two probes, which is then converted into the transpiration rate of the entire plant. The combined result of stomatal conductance and transpiration rate is the vegetation transpiration conductance data, which reflects the vegetation's transpiration regulation capacity under current moisture conditions. Under conditions of sufficient water, vegetation maintains a high transpiration conductance; when soil moisture deficit intensifies, vegetation reduces transpiration conductance by closing stomata to reduce water loss. This response is a direct manifestation of vegetation water stress.

[0024] Leaf water potential data were monitored using the pressure chamber method. Measurements were taken twice, once before sunrise and again at noon. The sunrise measurement represents the baseline water potential when the water potential between the leaf and soil is close to equilibrium, while the noon measurement represents the minimum water potential when transpiration is at its peak. For each measurement, 3 to 5 mature leaves were selected, cut from the plant, and immediately placed in the pressure chamber. Pressure was slowly increased until water droplets appeared at the cut surface; the pressure inside the chamber at this point is the absolute value of the leaf water potential. Leaf water potential was monitored twice daily, covering the entire growing season. The measurement range for leaf water potential data was -0.1 MPa to -3.0 MPa. When the noon leaf water potential was below -1.5 MPa, it indicated that the vegetation was under moderate water stress; when it was below -2.5 MPa, the vegetation was under severe water stress, which could lead to irreversible damage.

[0025] The vegetation water stress response monitoring module is also equipped with data quality control functions. For stomatal conductance measurements, outlier rejection rules are set up, marking data as suspicious when a single measurement deviates from the mean by more than twice the standard deviation. For leaf water potential measurements, a physical upper limit test is set, stipulating that the water potential in the early morning should not be lower than the water potential at noon; otherwise, the data is invalid. Data after quality control enters the subsequent processing flow.

[0026] The critical threshold calculation unit is the core data processing module of this system. It is connected to both the soil physicochemical index measurement module and the vegetation water stress response monitoring module. It receives soil aggregate stability index data, soil moisture characteristic curve data, vegetation transpiration conductance data, and leaf water potential data, and outputs critical water content threshold values ​​and steady-state transition critical point data. This unit includes a data preprocessing subunit, a multi-level hysteresis response analysis subunit, and a dynamic response surface construction subunit.

[0027] The data preprocessing subunit standardizes and aligns the input data over time. Due to differences in the collection frequency and time base of different data types, time alignment is necessary to ensure the reliability of subsequent analyses. Soil aggregate stability index data are discretely sampled, typically with a sampling period of one week to one month; soil moisture content data are continuously monitored, collected every 15 minutes; and vegetation water stress data are discrete data collected twice daily. The data preprocessing subunit uses interpolation algorithms to unify data from different frequencies to a daily scale. Cubic spline interpolation is used to maintain data continuity and smoothness. Standardization employs a min-max normalization method, mapping all data types to the 0-1 range to eliminate the impact of dimensional differences on subsequent calculations.

[0028] The multi-level hysteresis response analysis subunit is one of the core innovations of this invention. It is configured to analyze the time lag relationship between changes in soil moisture and changes in aggregate stability, as well as the time lag relationship between changes in soil moisture and changes in vegetation transpiration conductance. In sandy ecosystems, the responses of soil structure and vegetation physiology to changes in moisture do not occur instantly, but rather exhibit varying degrees of time lag. Changes in soil aggregate stability are a relatively slow process, typically laging behind changes in soil moisture by several days to several weeks; changes in vegetation transpiration conductance are relatively rapid, with a lag time typically ranging from several hours to several days. This hysteresis response characteristic is an important basis for determining whether a system is approaching a critical state.

[0029] Multi-level hysteresis response analysis employs the cross-correlation function method. Let the soil moisture content time series be... The time series of aggregate stability index is The time series of vegetation transpiration conductance is Cross-correlation function between soil moisture content and aggregate stability Defined as:

[0030] ,

[0031] in: The time lag is in days; The length of the time series; The mean of the soil moisture content sequence; This represents the mean of the aggregate stability index sequence. The cross-correlation function ranges from -1 to +1, with a larger absolute value indicating a stronger correlation. When the cross-correlation function reaches a certain lag time... When the maximum value is obtained at this point, This refers to the characteristic lag time of aggregate stability in response to changes in soil moisture.

[0032] Similarly, the cross-correlation function between soil moisture content and vegetation transpiration conductance Using the same calculation method, the characteristic lag time of vegetation transpiration conductance to soil moisture changes was obtained.

[0033] Under steady-state conditions, the hysteresis response exhibits regular periodic characteristics. As the system approaches a critical state, the regularity of the hysteresis response weakens, the hysteresis time lengthens, and the peak value of the cross-correlation function decreases. The multi-level hysteresis response analysis subunit provides early warning signals for critical state identification by continuously tracking changes in the hysteresis response characteristics.

[0034] The dynamic response surface construction subunit is based on the results of multi-level hysteresis response analysis, constructing a multi-dimensional response surface describing the state of the soil-vegetation system. The response surface uses soil moisture content as a starting point. and lag time As the independent variable, the aggregate stability index is used. and vegetation stress index The dependent variable is denoted by leaf water potential and transpiration conductance. The vegetation stress index is calculated by combining leaf water potential and transpiration conductance.

[0035] ,

[0036] in: The current leaf water potential is expressed in megapascals (MPA). For reference, the leaf water potential was taken at dawn under sufficient water conditions at the beginning of the growing season, and the unit is megapascals. This represents the current porosity, expressed in moles per square meter per second. For reference, the maximum stomatal conductance under sufficient water conditions is used, and the unit is moles per square meter per second. The vegetation stress index ranges from 0 to 1, with a higher value indicating a more severe stress.

[0037] The response surface was constructed using tensor decomposition. The four-dimensional dataset, consisting of soil moisture content, hysteresis time, aggregate stability index, and vegetation stress index, was represented as a third-order tensor. Perform CP decomposition on the tensor:

[0038] ,

[0039] in: To decompose the rank; For the first The weighting coefficients of each component; , , These are the factor vectors corresponding to soil moisture content, lag time, and system state, respectively. This represents the vector outer product operation. Tensor decomposition can reveal the underlying structural relationships between multidimensional data and extract the key patterns that dominate system state changes.

[0040] The steady-state transition critical point data is defined as the set of coordinates of gradient abrupt change points on the dynamic response surface. Gradient abrupt change points are identified using partial derivative analysis of the response surface. The gradient vector at each point on the response surface is calculated; when the magnitude of the gradient vector undergoes a significant jump, that point is considered a candidate critical point. Confirmation of a critical point also requires the following conditions to be met: the aggregate stability index decreases to below 70% of its initial value, and the vegetation stress index increases to above 0.5. Gradient abrupt change points that meet both conditions are confirmed as steady-state transition critical points.

[0041] The critical moisture content threshold was calculated using a dual-constraint method. The soil moisture content corresponding to a decrease in the soil aggregate stability index to 70% of its initial value was taken as the critical moisture content of the aggregates. The soil moisture content corresponding to a decrease in vegetation transpiration conductance data to 50% of the potential transpiration conductance is taken as the critical moisture content for vegetation. Critical moisture content threshold value Take the larger of the two values:

[0042] ,

[0043] The rationale for choosing a larger value is that when the soil moisture content falls below this threshold, at least one aspect of soil structure and vegetation physiology is approaching a critical state, threatening the overall stability of the system. This dual-constraint method embodies the concept of synergistic response in the soil-vegetation system and is more robust than judging by a single indicator.

[0044] The steady-state transition early warning module is connected to the critical threshold calculation unit. It receives the critical moisture content threshold value and the steady-state transition critical point data, and compares them with the environmental parameters monitored in real time by the multi-parameter sample acquisition device. When the system approaches the critical state, it outputs an early warning signal. This module includes a real-time comparison subunit and an early warning signal generation subunit.

[0045] The real-time comparison sub-unit is configured to compare the current soil moisture content with the critical moisture content threshold value and calculate the early warning margin value. :

[0046] ,

[0047] in: This is the current real-time monitoring value of soil moisture content, in cubic meters per cubic meter. This represents the critical moisture content threshold value, expressed in cubic meters per cubic meter. The warning margin value indicates the distance between the current state and the critical state; the smaller the value, the closer to the critical state.

[0048] The real-time comparison sub-units also monitor changes in hysteresis response characteristics. When the hysteresis time is significantly prolonged or the peak value of the cross-correlation function is significantly reduced, it indicates that the system is evolving towards an unstable state, even if the current soil moisture content has not yet reached the critical threshold. The criteria for judging abnormal hysteresis response are: hysteresis time exceeding 1.5 times the historical average, or the peak value of the cross-correlation function being lower than 0.7 times the historical average.

[0049] The early warning signal generation subunit is configured to generate three levels of early warning signals based on real-time comparison results. Level 1 Early Warning (Yellow): The warning margin value is less than the preset safety threshold (default value is 0.05 cubic meters per cubic meter) but greater than zero, or an abnormal delayed response occurs. A Level 1 early warning indicates that the system is approaching a critical state, and it is recommended to increase the monitoring frequency. Level 2 Early Warning (Orange): The warning margin value is less than or equal to zero, meaning the current soil moisture content has reached or fallen below the critical threshold. A Level 2 early warning indicates that the system has entered a critical region, and it is recommended to take manual intervention measures. Level 3 Early Warning (Red): The current state point falls within the data range of the steady-state transition critical point confirmed on the dynamic response surface. A Level 3 early warning indicates that the system is undergoing a steady-state transition, and immediate emergency response measures are required.

[0050] Warning signals are delivered through various means, including SMS, email, and audible and visual alarms, to ensure that relevant management personnel can obtain warning information in a timely manner and take appropriate measures.

[0051] Reference Figure 2 This embodiment provides an application of the above-mentioned critical threshold identification system for sandy land eco-hydrological processes, including the following steps.

[0052] Step S1, Sample Collection. Using a multi-parameter sample collection device, soil and target vegetation samples were collected at different depths in the sandy study area. Simultaneously, an in-situ sensor array was activated for environmental parameter monitoring. Sampling deployment followed the principle of representativeness, selecting 3 to 5 typical plots within the study area. Each plot included one soil profile sampling point and three individual target vegetation specimens for monitoring. Soil sampling depth extended from the surface to 50 cm below the groundwater level, with each 20 cm layer constituting a sampling layer. Sampling was conducted once at the beginning, middle, and end of the growing season, with more frequent sampling every two weeks during the growing season. The in-situ sensor array began continuous monitoring immediately after deployment, with data acquisition occurring every 15 minutes.

[0053] Step S2, Soil Index Measurement Step. The soil physicochemical index measurement module is used to perform aggregate stability analysis and moisture characteristic curve determination on the collected soil samples. Aggregate stability analysis employs the wet sieving method described in the system embodiment to obtain the average weight diameter and geometric mean diameter data for each sampling layer. Moisture characteristic curve determination employs the pressure membrane method described in the system embodiment to obtain the van Genuchten model parameters for each sampling layer. After the measurements are completed, the data are entered into the database of the critical threshold calculation unit.

[0054] Step S3, Vegetation Response Monitoring. The target vegetation is continuously monitored using a vegetation water stress response monitoring module. Stomatal conductance is measured daily from 9:00 AM to 11:00 AM, with 5 to 10 leaves measured per plant; leaf water potential is measured daily before sunrise and at 12:00 PM, with 3 to 5 leaves measured per plant. Monitoring data is uploaded in real-time to the critical threshold calculation unit via a wireless transmission module. The monitoring period covers the entire growing season, typically from May to October.

[0055] Step S4, Critical Threshold Calculation Step. The critical threshold calculation unit processes and analyzes the data obtained in steps S2 and S3. First, the data preprocessing subunit performs time alignment and standardization. Then, the multi-level hysteresis response analysis subunit calculates the cross-correlation functions between soil moisture and aggregate stability, and between soil moisture and vegetation transpiration conductance, to determine the characteristic lag times. Next, the dynamic response surface construction subunit uses tensor decomposition to construct a multi-dimensional response surface and identify the steady-state transition critical point. Finally, the critical moisture content threshold value is calculated based on the dual-constraint method. The calculation results are output to the steady-state transition early warning module.

[0056] Step S5, Early Warning Judgment Step. The steady-state transition early warning module compares the real-time soil moisture content data monitored by the in-situ sensor array with the critical moisture content threshold value to calculate the early warning margin. Simultaneously, changes in the hysteresis response characteristics are monitored. When an early warning condition is triggered, the early warning signal generation subunit outputs an early warning signal of the corresponding level. The early warning signal includes information such as the early warning level, early warning time, current state parameters, and recommended measures.

[0057] The effective operation of the system of this invention depends on reasonable parameter configuration. The system parameters need to be adaptively adjusted for different types of sandy ecosystems.

[0058] Regarding the configuration of sampling depth parameters, the setting of sampling depth intervals should take into account the vertical differentiation characteristics of the soil profile. For sandy areas with shallow groundwater levels (less than 3 meters), a sampling interval of 20 centimeters is recommended; for sandy areas with deeper groundwater levels (greater than 5 meters), this interval can be appropriately increased to 30 centimeters. The total sampling depth should exceed the groundwater level by at least 50 centimeters to ensure coverage of the entire capillary rise zone. In strata with significant changes in soil texture, sampling should be intensified to capture interface effects.

[0059] Regarding the configuration of monitoring frequency parameters, the data acquisition frequency setting of the in-situ sensor array needs to strike a balance between data accuracy and storage capacity. Before and after precipitation events and during periods of high evaporation, it is recommended to increase the acquisition frequency to once every 5 minutes to capture rapid changes; during stable periods, it can be reduced to once every 30 minutes to save storage space. The frequency of vegetation water stress monitoring should match the physiological rhythm of the vegetation, and the measurement times for water levels at dawn and noon should be fixed to ensure data comparability.

[0060] Regarding parameter configuration for lag response analysis, the calculation window length of the cross-correlation function directly affects the stability of the analysis results. A window that is too short will lead to large fluctuations in the results, while a window that is too long will smooth out meaningful changes. Based on the response timescale characteristics of sandy ecosystems, it is recommended that the window length for soil moisture and aggregate stability analysis be set to 30 days, and the window length for soil moisture and vegetation transpiration conductance analysis be set to 14 days. The search range for lag time should cover the expected maximum lag time, and it is recommended to set it to half the window length.

[0061] Regarding the configuration of critical threshold parameters, the 70% threshold in the critical criterion for aggregate stability and the 50% threshold in the critical criterion for vegetation transpiration conductance are default values ​​and can be adjusted according to specific ecosystem characteristics. For sandy areas with high vulnerability, these thresholds can be appropriately increased (e.g., 80% and 60%) to obtain a more conservative critical judgment; for sandy areas with strong recovery capacity, the thresholds can be appropriately decreased (e.g., 60% and 40%) to avoid over-warning.

[0062] Regarding the configuration of early warning parameters, the setting of the early warning margin threshold should consider the accuracy of the monitoring data and the system response time. When the soil moisture sensor accuracy is ±0.02 cubic meters per cubic meter, the early warning margin threshold should not be less than 0.03 cubic meters per cubic meter to avoid false alarms caused by measurement errors. The gradient setting of the thresholds at each level in the three-level early warning system should match the management response capability to ensure that each level of early warning can trigger corresponding management actions.

[0063] The hardware architecture of the system of this invention adopts a distributed design, including three layers: the front-end acquisition layer, the edge computing layer, and the central processing layer.

[0064] The front-end acquisition layer consists of an in-situ sensor array within a multi-parameter sample acquisition device. The sensors employ an industrial-grade protective design with an IP67 protection rating, enabling long-term stable operation in harsh sandy environments. Power is supplied by a combination of solar panels and batteries, allowing for over 30 days of continuous operation on a single full charge. Communication between the sensors and edge computing nodes utilizes the LoRa wireless communication protocol, achieving a communication distance of up to 5 kilometers, suitable for long-distance transmission in open sandy terrain.

[0065] The edge computing layer consists of edge computing nodes installed within the study area. These nodes possess preliminary data processing capabilities, enabling preprocessing operations such as outlier removal, missing value imputation, and data compression on raw sensor data, thus reducing the computational burden and communication bandwidth requirements of the central processing layer. The edge computing nodes utilize low-power ARM architecture processors, have 4GB of memory, and 64GB of storage, capable of caching two weeks' worth of raw monitoring data. Communication between the edge computing nodes and the central processing layer is via a 4G cellular network, supporting breakpoint resume functionality to ensure data integrity.

[0066] The central processing layer is the core computing platform of the system, housing the software system for critical threshold calculation units and steady-state transition early warning modules. The central processing layer adopts a server cluster architecture, configured with an 8-core processor, 32GB of memory, and a 2TB solid-state drive, running a Linux operating system. The software system is developed using Python, with the core calculation modules using the NumPy and SciPy libraries for numerical computation, and the tensor decomposition module using the TensorLy library. The system uses a MySQL database to store historical monitoring data and calculation results, and Redis to cache real-time monitoring data to improve access efficiency.

[0067] In terms of communication protocols, a custom lightweight binary protocol is used between the front-end acquisition layer and the edge computing layer. The data frame format includes seven fields: frame header, device identifier, timestamp, data type, data length, data content, and checksum. The frame header is a fixed two-byte identifier, the device identifier is a four-byte unsigned integer, the timestamp is an eight-byte Unix timestamp, the data type is a one-byte enumerated value, the data length is a two-byte unsigned integer, the data content is a variable-length byte array, and the checksum is a two-byte CRC16 checksum. The edge computing layer communicates with the central processing layer using the MQTT protocol, and the message format uses JSON encoding for easy parsing and expansion.

[0068] Regarding data security, this invention employs multiple measures to ensure the integrity and confidentiality of monitoring data. Data transmission is encrypted using the TLS 1.3 protocol to prevent interception or tampering during transmission. Data storage uses the AES-256 encryption algorithm to encrypt sensitive fields, with encryption keys centrally managed by a key management service. System access utilizes a role-based access control mechanism, with different users holding different data access and operation permissions. System operation logs comprehensively record all user login, query, and modification operations, supporting operation auditing and problem tracing.

[0069] Regarding system reliability, this invention employs a redundancy design to improve the availability of key components. The in-situ sensor array deploys two sensors at each monitoring depth; if one fails, the other can continue to provide monitoring data. The edge computing nodes utilize a dual-machine hot standby architecture, with the standby node automatically taking over when the primary node fails. The central processing layer employs load balancing and failover mechanisms, ensuring that the failure of a single server does not affect the overall system operation. The database uses a master-slave replication architecture, with data written to the master database synchronized to the slave database in real time, ensuring no data loss.

[0070] The system possesses self-diagnostic and self-recovery capabilities. The sensor self-diagnostic function runs hourly, checking if sensor readings are within physically reasonable ranges. Upon detecting anomalies, it automatically flags them and sends alarms to maintenance personnel. The edge computing node watchdog program continuously monitors the application's running status and automatically restarts it if it detects a program crash. The central processing layer's health check service performs a health check on each component every minute, triggering an automatic recovery process or sending alarms to maintenance personnel upon detecting anomalies.

[0071] The system of this invention was applied to a typical study area in the Mu Us Desert for verification. The study area is located at 38°45′N, 109°12′E, with an average annual precipitation of 320 mm and an average annual evaporation of 2100 mm. The soil type is aeolian sandy soil, and the dominant vegetation is *Salix psammophila*. The verification period was from May to October 2023. Five monitoring plots were set up in the study area, and a complete monitoring device was installed in each plot.

[0072] The results of soil aggregate stability measurements showed that the average weight diameter of the topsoil (0–20 cm) in the study area was 0.68 mm, the average weight diameter of the subsoil (20–60 cm) was 0.52 mm, and the average weight diameter of the deep soil (below 60 cm) was 0.41 mm. The contents of water-stable aggregates larger than 0.25 mm were 42%, 35%, and 28%, respectively. This indicates that the soil structural stability in the study area decreases with increasing depth, consistent with the general pattern of sandy soils.

[0073] The results of soil moisture characteristic curve measurements show that the van Genuchten model parameters for the soil in the study area are: residual moisture content 0.021 m³ / m³, saturated moisture content 0.385 m³ / m³, parameter α 0.068 m / cm, and parameter n 1.65. These parameter values ​​reflect the typical moisture characteristics of sandy soils, namely, low water holding capacity and high hydraulic conductivity.

[0074] Monitoring results of vegetation water stress response showed that during the validation period, the stomatal conductance of *Salix psammophila* varied between 0.05 and 0.35 mol / m² / s, the leaf water potential at dawn varied between -0.3 and -1.8 MPa, and the leaf water potential at noon varied between -0.8 and -2.6 MPa. During the sustained drought from late July to mid-August, the stomatal conductance dropped below 0.08 mol / m² / s, and the leaf water potential at noon fell below -2.2 MPa, indicating that the vegetation was under severe water stress.

[0075] The results of multi-level lag response analysis showed that the cross-correlation function between soil moisture content and aggregate stability index peaked at a lag of 8 days, with a peak correlation coefficient of 0.72. The cross-correlation function between soil moisture content and vegetation transpiration conductance peaked at a lag of 2 days, with a peak correlation coefficient of 0.81. During the drought, these two lag times were extended to 12 days and 4 days, respectively, and the peak values ​​of the cross-correlation functions decreased to 0.58 and 0.65, respectively, indicating that the regularity of the system response weakened and it was evolving towards an unstable state.

[0076] Verification results show that the critical moisture content threshold identified by the system of this invention is 0.078 cubic meters per cubic meter, with the critical moisture content of aggregates at 0.065 cubic meters per cubic meter and the critical moisture content of vegetation at 0.078 cubic meters per cubic meter. The system takes the larger of the two values. This threshold is highly consistent with the soil moisture content of 0.082 cubic meters per cubic meter corresponding to severe vegetation wilting observed in actual studies, with a relative error of only 4.9%. The dynamic response surface successfully identified three steady-state transition critical points, corresponding to the states on July 28, August 5, and August 12, respectively. The critical points on August 5 and August 12 were verified to show vegetation wilting in subsequent observations.

[0077] The steady-state transition early warning module issued a Level 1 warning 5 days before actual vegetation withering and a Level 2 warning 2 days before actual withering, demonstrating good timeliness. During the entire verification period, the system issued 12 Level 1 warnings, 3 Level 2 warnings, and 0 Level 3 warnings. The false alarm rate for Level 1 warnings was 16.7% (2 false alarms), and the false alarm rate for Level 2 warnings was 0%.

[0078] Compared with existing technologies, the system of this invention has the following advantages. First, the accuracy of critical threshold identification is improved, with the relative error reduced from over 15% in existing methods to less than 5%. Second, the timeliness of early warning is enhanced, with the early warning lead time increased from 1 to 2 days in existing methods to 3 to 5 days. Third, the system is more systematic, achieving synergistic analysis of soil structure and vegetation response, overcoming the limitations of single-indicator judgment. Fourth, practicality is enhanced, with dynamic early warning based on real-time monitoring replacing periodic discrete measurements, reducing operation and maintenance costs by approximately 40%. Fifth, reliability is enhanced, with redundant design and self-diagnostic functions ensuring long-term stable operation of the system in harsh sandy environments.

[0079] The embodiments of the present invention are not limited to the specific embodiments described above. Those skilled in the art can make various equivalent changes or substitutions based on the technical solutions of the present invention, and all such changes or substitutions should be included within the protection scope of the present invention.

Claims

1. A critical threshold identification system for sandy land eco-hydrological processes, characterized in that, include: The multi-parameter sample collection device is configured to collect soil and vegetation samples at different depths within the sandy research area and to monitor in-situ environmental parameters in the collection area. The soil physicochemical index determination module is connected to the multi-parameter sample acquisition device and is configured to receive the soil sample and determine the soil aggregate stability index data and soil moisture characteristic curve data. The soil aggregate stability index data includes the mass distribution ratio of aggregates of different particle sizes and the content of water-stable aggregates. The soil moisture characteristic curve data includes the correspondence between soil volumetric water content and soil water potential. The vegetation water stress response monitoring module is configured to monitor the water physiological state of the target vegetation in the study area and output vegetation transpiration conductance data and leaf water potential data, wherein the vegetation transpiration conductance data characterizes the degree of vegetation stomatal response to water deficit, and the leaf water potential data characterizes the degree of vegetation water stress. The critical threshold calculation unit is connected to the soil physicochemical index measurement module and the vegetation water stress response monitoring module, respectively. It is configured to receive the soil aggregate stability index data, the soil moisture characteristic curve data, the vegetation transpiration conductance data and the leaf water potential data. Based on the synergistic analysis of the response characteristics of soil aggregate stability with water change and the response characteristics of vegetation water stress, it calculates the critical water content threshold value and the steady-state transition critical point data. The steady-state transition early warning module is connected to the critical threshold calculation unit and is configured to compare the environmental parameters monitored in real time by the multi-parameter sample acquisition device with the critical moisture content threshold value and the steady-state transition critical point data. When the difference between the monitored parameter and the critical moisture content threshold value is less than a preset safety threshold or the monitored parameter is lower than the critical moisture content threshold value, an early warning signal is output. The critical threshold calculation unit includes a multi-level hysteresis response analysis subunit, which is configured to analyze the time lag relationship between soil moisture changes and aggregate stability changes, as well as the time lag relationship between soil moisture changes and vegetation transpiration conductance changes, and output hysteresis response characteristic parameters at different levels. The critical threshold calculation unit further includes a dynamic response surface construction subunit. The dynamic response surface construction subunit is configured to construct a multidimensional response surface based on the hysteresis response characteristic parameters, with soil moisture content and time as independent variables and aggregate stability index and vegetation stress index as dependent variables. The steady-state transition critical point data is the set of coordinates of gradient abrupt change points on the multidimensional response surface. When calculating the critical moisture content threshold value, the critical threshold calculation unit uses the soil moisture content corresponding to the decrease of the soil aggregate stability index data to 70% of the initial value as the critical moisture content of the aggregate, and the soil moisture content corresponding to the decrease of the vegetation transpiration conductance data to 50% of the potential transpiration conductance as the critical moisture content of the vegetation. The larger value between the critical moisture content of the aggregate and the critical moisture content of the vegetation is taken as the critical moisture content threshold value.

2. The critical threshold identification system for sandy land eco-hydrological processes according to claim 1, characterized in that, When the soil physicochemical index determination module measures the soil aggregate stability index data, it uses a wet sieving method to obtain the mass percentage of water-stable aggregates larger than 0.25 mm. The measurement range of the mass percentage is multiple stratification depths between the soil surface and the groundwater level.

3. The critical threshold identification system for sandy land eco-hydrological processes according to claim 1, characterized in that, The soil moisture characteristic curve data were determined using the pressure membrane method, with the soil water potential ranging from -0.01 MPa to -1.5 MPa, corresponding to a soil volumetric water content ranging from 0.02 cubic meters per cubic meter to 0.45 cubic meters per cubic meter.

4. The critical threshold identification system for sandy land eco-hydrological processes according to claim 1, characterized in that, The vegetation water stress response monitoring module monitors the leaf water potential data with a time resolution of once per hour, and the monitoring period covers the entire growing season. The measurement range of the leaf water potential data is from -0.1 MPa to -3.0 MPa.

5. The critical threshold identification system for sandy land eco-hydrological processes according to claim 1, characterized in that, The multi-parameter sample collection device includes a stratified sampler assembly and an in-situ sensor array. The stratified sampler assembly is configured to collect soil samples at preset depth intervals in the vertical direction. The in-situ sensor array is configured to monitor soil temperature, soil electrical conductivity, and soil moisture content in real time.

6. The critical threshold identification system for sandy land eco-hydrological processes according to claim 1, characterized in that, The steady-state transition early warning module includes a real-time comparison subunit and an early warning signal generation subunit. The real-time comparison subunit is configured to calculate the difference between the current soil moisture content and the critical moisture content threshold value as an early warning margin value. The early warning signal generation subunit is configured to generate an early warning signal when the early warning margin value is less than a preset safety threshold.

7. The application of the critical threshold identification system for sandy land eco-hydrological processes according to any one of claims 1 to 6, characterized in that, Includes the following steps: The sample collection step involves using the multi-parameter sample collection device to collect soil samples and target vegetation samples at different depths in the sandy research area, while simultaneously initiating in-situ environmental parameter monitoring. The soil index determination step involves using the soil physicochemical index determination module to perform aggregate stability analysis and moisture characteristic curve determination on the collected soil samples, thereby obtaining the soil aggregate stability index data and the soil moisture characteristic curve data. The vegetation response monitoring step involves continuously monitoring the target vegetation using the vegetation water stress response monitoring module to obtain the vegetation transpiration conductance data and the leaf water potential data. The critical threshold calculation step involves using the critical threshold calculation unit to perform multi-level hysteresis response analysis and dynamic response surface construction on the soil aggregate stability index data, the soil moisture characteristic curve data, the vegetation transpiration conductance data, and the leaf water potential data, and outputting the critical moisture content threshold value and the steady-state transition critical point data. The early warning judgment step involves comparing the real-time monitoring data with the critical moisture content threshold value and the critical point data of the steady-state transition using the steady-state transition early warning module, and outputting an early warning signal.

Images