Physical model-based permafrost hydrothermal state simulation method and system

Abstract

The invention discloses a permafrost hydrothermal state simulation method and system based on a physical model, and relates to the technical field of permafrost region engineering. The method comprises the following steps: acquiring meteorological driving data and land surface feature data; constructing a numerical model of a permafrost hydrothermal process by considering a permafrost freezing and thawing process and a soil ice effect by utilizing land surface feature data; preparing boundary constraint conditions of the numerical model according to the meteorological driving data and the existing temperature field data; and performing permafrost hydrothermal state simulation on the land surface characteristic data by using meteorological driving data, constraint condition data, key vegetation and soil parameters and a driving numerical model to obtain a sample set formed by key permafrost hydrothermal information. According to the method, the dependence of an existing model on data and structures is broken through, the multi-scale application requirement of the permafrost region is met, and finally a high-quality meteorological driving and boundary constraint condition data set suitable for permafrost simulation without an observation region can be generated.

Description

Technical Field

[0001] The present invention relates to the field of frozen soil regional engineering technology, and in particular to a permafrost hydrothermal state simulation method and system based on a physical model. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] At present, the research on the hydrothermal status of permafrost mainly relies on three technical paths: field observation, remote sensing technology and numerical simulation. However, in regional applications such as the Qinghai-Tibet Plateau, there are technical bottlenecks to varying degrees, which limit the accurate identification and simulation of the spatial distribution characteristics of permafrost, the hydrothermal change process and its response mechanism.

[0004] Field observations and remote sensing technologies provide fundamental data support for permafrost research, but they still face limitations in terms of regional coverage, penetration, and spatiotemporal continuity. Field observations rely on manually deployed monitoring points, primarily located in areas with convenient transportation. These observations lack spatial representation and cannot fully reflect the spatial heterogeneity of hydrothermal regimes across large permafrost areas. Remote sensing, while offering the advantages of monitoring over large areas and long time series, focuses primarily on surface information and struggles to penetrate the soil to obtain key parameters such as temperature, moisture, and ice content within deep permafrost. Furthermore, due to limitations in resolution, meteorological conditions, and data assimilation accuracy, these technologies suffer from unstable spatiotemporal accuracy and insufficient ability to resolve subsurface processes. Therefore, while field observations and remote sensing offer complementary data acquisition capabilities, neither can independently support high-precision simulations of hydrothermal regimes and forecast future trends in permafrost areas.

[0005] Numerical simulation methods are currently an important technical means of studying permafrost evolution and predicting future trends. Empirical models based on empirical relationships are simple and require minimal data. However, their statistical relationships are difficult to adapt to the complex and changing permafrost environment, lack a physical basis, and cannot accurately reflect the coupled energy and moisture processes within permafrost. Their regional applicability and scalability are poor. In contrast, physical models based on energy and water balance (such as Noah-MP, CLM, SHAW, and CryoGrid) can systematically characterize the exchange of thermal water within and outside the permafrost layer. They have a strong theoretical foundation and process traceability, making them an important tool for studying permafrost response mechanisms and conducting scenario forecasts.

[0006] However, existing numerical models based on physical processes still have the following key technical deficiencies: (1) The model's physical processes are incomplete and cannot accurately represent the complex energy and water exchange processes during the freeze-thaw process of permafrost. Key parameters in permafrost simulation, such as soil ice content, freeze-thaw curves, and ice-water phase transition mechanisms, are highly nonlinear and time-varying. Existing models often simplify these parameters and fail to fully consider their impact on simulation results.

[0007] (2) The model is highly dependent on meteorological driving data and boundary constraints, making it difficult to apply in data-scarce areas. The simulation of permafrost hydrothermal processes is highly dependent on high-resolution meteorological driving data, soil parameters, vegetation distribution, boundary hydrothermal fluxes, and other data. However, in areas such as the Qinghai-Tibet Plateau, observational data are sparse and of varying quality, making it difficult to accurately obtain boundary hydrothermal flux data. This results in large model simulation errors, poor stability, and unreliable output results.

[0008] (3) Model parameters are difficult to calibrate efficiently, and it is difficult to achieve both universality and accuracy. Current frozen soil simulation models contain a large number of physical process parameters (such as thermal conductivity, porosity, evaporation resistance, and freezing latent heat treatment parameters). However, due to the lack of sufficient observational data for calibration, regional average values ​​or empirical settings are often used, resulting in significant local simulation deviations.

[0009] In summary, the current numerical simulation methods based on physical processes have exposed a series of technical shortcomings in their application in plateau permafrost areas, such as poor adaptability, strong data dependence, and large uncertainty, which seriously restrict the high-precision simulation and trend prediction of the hydrothermal state of permafrost. Summary of the Invention

[0010] In response to the shortcomings of the existing technology, the purpose of the present invention is to provide a method and system for simulating the hydrothermal state of permafrost based on a physical model, fully considering the hydrothermal exchange process of permafrost, thereby developing a new parameterization scheme, and constructing a physical model that can accurately simulate the internal energy and moisture changes of permafrost. Based on observation data and remote sensing products, artificial intelligence methods are used to break through the existing model's dependence on data and structure, meet the multi-scale application needs of permafrost areas, and ultimately generate a high-quality meteorological driving and boundary constraint condition dataset suitable for permafrost simulation in unobserved areas.

[0011] In order to achieve the above object, the present invention is implemented through the following technical solutions: A first aspect of the present invention provides a method for simulating the hydrothermal state of permafrost based on a physical model, comprising the following steps: Obtain meteorological driving data and land surface characteristic data, wherein the land surface characteristic data include key vegetation parameters and key soil parameters; A numerical model of permafrost hydrothermal processes is constructed using land surface characteristic data, taking into account permafrost freeze-thaw processes and soil ice effects. Prepare boundary constraints of the numerical model based on meteorological driving data and existing temperature field data; The permafrost hydrothermal state is simulated by using meteorological data to drive the numerical model based on land surface characteristic data, and a sample set of key permafrost hydrothermal information is obtained.

[0012] Furthermore, the specific steps for constructing a numerical model of permafrost hydrothermal processes by using land surface characteristic data and considering the permafrost freeze-thaw process and soil ice action are as follows: Construct an energy balance equation based on the energy changes in the soil layer; Construct a water balance equation based on the changes in soil water content and the water absorption capacity of vegetation; The calculation equation of soil freeze-thaw process is constructed based on the physical mechanism of frozen soil freeze-thaw.

[0013] Furthermore, the specific steps for constructing the soil freeze-thaw process calculation equation based on the physical mechanism of frozen soil freeze-thaw are as follows: Calculate unfrozen water and ice content based on the frozen soil water potential according to soil pore ice action; The specific heat capacity and thermal conductivity are calculated by weighted calculation considering the composite effects of multiple soil components; The hydraulic conductivity affected by soil ice is calculated based on the effect of soil ice on hydraulic conductivity.

[0014] Furthermore, the specific steps for preparing the boundary constraints of the numerical model based on the meteorological driving data and the existing temperature field data are as follows: Use machine learning to simulate the data set composed of temperature field data to obtain the lower boundary temperature field constraint data, thereby constructing the lower boundary constraint condition; The model is preheated to obtain cyclic preheating data, and the temperature and moisture initial field constraints are constructed based on the cyclic preheating data.

[0015] Furthermore, the specific steps for simulating the data set composed of temperature field data using machine learning are as follows: Get lower boundary temperature constraint data; Obtain actual temperature field data to form a data set; Add random perturbation preprocessing operations to the data set; Verify the accuracy of the preprocessed dataset, filter and update the data in the dataset to obtain an updated dataset; Use the updated dataset to train the random forest model; The updated random forest model is used to correct the lower boundary temperature constraint data.

[0016] A second aspect of the present invention provides a permafrost hydrothermal state simulation system based on a physical model, comprising: a data acquisition module configured to acquire meteorological driving data and land surface characteristic data, wherein the land surface characteristic data includes key vegetation parameters and key soil parameters; A model building module is configured to construct a numerical model of permafrost hydrothermal processes by using land surface characteristic data and considering permafrost freeze-thaw processes and soil ice effects; A constraint preparation module is configured to prepare boundary constraint conditions of the numerical model based on meteorological driving data and existing temperature field data; The state simulation module is configured to use meteorological driving data to drive the numerical model to simulate the permafrost hydrothermal state of the land surface characteristic data, and obtain a sample set consisting of key permafrost hydrothermal information.

[0017] Furthermore, the model building module includes: an energy balance calculation module configured to construct an energy balance equation according to energy changes in the soil layer; a water balance calculation module configured to construct a water balance equation based on changes in soil water content, ice-water phase transition, and vegetation water absorption capacity; The soil freeze-thaw cycle calculation module is configured to construct a soil freeze-thaw process calculation equation based on the physical mechanism of frozen soil freeze-thaw.

[0018] Furthermore, the soil freeze-thaw cycle calculation module is also configured to: Calculate unfrozen water and ice content based on the frozen soil water potential according to soil pore ice action; The specific heat capacity and thermal conductivity are calculated by weighted calculation considering the composite effects of multiple soil components; The hydraulic conductivity affected by soil ice is calculated based on the effect of soil ice on hydraulic conductivity.

[0019] Furthermore, the constraint preparation module includes: a lower boundary constraint module, configured to simulate a data set consisting of temperature field data using machine learning to obtain lower boundary temperature field constraint data, thereby constructing a lower boundary constraint condition; The initial field constraint module is configured to perform preheating processing on the lower boundary temperature field constraint data to obtain cyclic preheating data, and construct the initial field constraint conditions according to the cyclic preheating data.

[0020] Furthermore, the lower boundary constraint module is also configured as follows: Get lower boundary temperature constraint data; Obtain actual temperature field data to form a data set; Add random perturbation preprocessing operations to the data set; Verify the accuracy of the preprocessed dataset, filter and update the data in the dataset to obtain an updated dataset; Use the updated dataset to train the random forest model; The updated random forest model is used to correct the lower boundary temperature constraint data.

[0021] One or more of the above technical solutions have the following beneficial effects: The present invention discloses a method and system for simulating the hydrothermal state of permafrost based on a physical model. The method overcomes the shortcomings of existing numerical simulation methods, fully considers the hydrothermal exchange process of permafrost, and thus develops a new parameterization scheme to construct a physical model that can accurately simulate the internal energy and moisture changes of permafrost. Based on observation data and remote sensing products, an artificial intelligence method is used to generate a high-quality meteorological driving and boundary constraint condition data set suitable for permafrost simulation in unobserved areas, breaking through the existing model's dependence on data and structure and meeting the multi-scale application needs of permafrost areas.

[0022] This method improves the permafrost parameterization scheme by taking into account the hydrothermal interactions of frozen soils. This improves the quality of meteorological driving data, vegetation and soil parameters, and boundary constraints, significantly enhancing the model's accuracy in simulating the coupled heat and moisture mechanisms during freeze-thaw processes. This method is adaptable to environments where observations are scarce and offers the advantages of high accuracy, strong generalization, and regional scalability.

[0023] Traditional permafrost models often use a fixed temperature as the lower boundary constraint, ignoring seasonal fluctuations in the lower boundary temperature field. This can easily lead to significant errors in permafrost temperature simulations. This paper simulates lower boundary data using the Noah model and performs error correction on the simulated data through machine learning, thereby obtaining high-quality lower boundary data and providing data support for efficient numerical model simulations.

[0024] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0026] Figure 1 This is a flow chart of the permafrost hydrothermal state simulation method based on a physical model in Example 1 of the present invention. DETAILED DESCRIPTION

[0027] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs.

[0028] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or their combinations; The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0029] Example 1: The first embodiment of the present invention provides a method for simulating the hydrothermal state of permafrost based on a physical model, such as Figure 1 As shown, taking the Qinghai-Tibet Plateau as an example, the specific steps include: Step 1: Obtain meteorological driving data and land surface characteristic data.

[0030] Step 1.1: Obtain meteorological driving data.

[0031] In a specific embodiment, meteorological conditions are the key control factors for controlling the hydrothermal process of permafrost, and are also necessary driving data for the operation of the hydrothermal coupling model SHAW (Simultaneous Heat and Water Model). There are few meteorological observation stations on the Qinghai-Tibet Plateau, and it is difficult to provide suitable meteorological driving data for the entire plateau. In order to drive the improved SHAW model to simulate permafrost in areas where there is no observational meteorological data, this embodiment collects key meteorological data from the China Meteorological Forcing Data. This dataset is a grid meteorological data product with a spatial resolution of 10 km. The data time range is from 1979 to 2018, with a time step of 3 hours, including 2 m temperature, precipitation, wind speed, wind direction, relative humidity, air pressure, shortwave radiation, longwave radiation and other data of a 10 km grid. Independently build R language code to extract the required meteorological data. The specific steps are: Step 1.1.1: Open the meteorological data saved in the netcdf file format.

[0032] Step 1.1.2: Use the Qinghai-Tibet Plateau boundary shapefile to clip the meteorological data and extract the part inside the Qinghai-Tibet Plateau.

[0033] Step 1.1.3: Filter the required meteorological data elements, including temperature, precipitation, wind speed, relative humidity, shortwave radiation, etc., with a time range from 1979 to 2018.

[0034] Step 1.1.4: Store all meteorological data for each 10km cell into a separate text file named with the coordinates of the cell center point.

[0035] Step 1.2: Obtain land surface characteristic data, which includes key vegetation parameters and key soil parameters.

[0036] In a specific embodiment, for areas without measured vegetation and soil data, this embodiment uses the MODIS MCD12Q1 land use classification product of NASA remote sensing satellite images with a spatial resolution of 500 m as the vegetation type. The classification standard adopts the IGBP standard (Loveland and Belward, 1997). The soil type data comes from the HWSD (Harmonized World Soil Database) dataset with a spatial resolution of 1 km. Each key vegetation parameter and key soil parameter is provided by the NLDAS (NASA Land Data Assimilation System) vegetation parameter lookup table. The specific steps are as follows: Step 1.2.1: Download the MODIS MCD12Q1 land use type product and the HWSD soil type raster product from the data website.

[0037] Step 1.2.2: Using ArcGIS, a comprehensive geospatial platform, crop key vegetation and soil parameters based on the shapefile data of the Qinghai-Tibet Plateau boundary.

[0038] In this embodiment, key vegetation parameters include vegetation coverage, leaf area index, canopy height, root depth, etc., and key soil parameters include soil texture, saturated hydraulic conductivity, porosity, etc.

[0039] Step 1.2.3: Resample the cropped vegetation key parameters and soil key parameters to a 10 km resolution to match the spatial resolution of the meteorological data.

[0040] Step 1.2.4: Obtain the key vegetation and soil parameter values ​​for each grid according to the parameter lookup table, and create a parameter configuration file that can be read by the model in the format required by the model.

[0041] Step 2: Use land surface characteristic data to consider the permafrost freeze-thaw process and soil ice effect to construct a numerical model of permafrost hydrothermal process.

[0042] Based on the widely used hydrothermal coupling model SHAW, this paper proposes a new simulation framework and parameterization scheme for frozen soil temperature and moisture changes that considers the permafrost freeze-thaw process and the effects of soil ice. A high-precision permafrost hydrothermal process numerical model is constructed to simulate the hydrothermal changes of permafrost on the Qinghai-Tibet Plateau over the past 40 years, producing high-precision data on soil temperature, unfrozen water content, and ground ice content.

[0043] Step 2.1: Construct an energy balance equation based on the energy changes in the soil layer.

[0044] .

[0045] In the above formula, The energy change of the soil layer caused by temperature change; is the latent heat flux of soil water freezing and thawing; is the heat transfer flux of the soil layer; is the convective heat flux of liquid water in the soil layer; is the latent heat flux of evapotranspiration in the soil layer.

[0046] in, is the volumetric specific heat capacity of soil, 、 and are the densities of ice, liquid water, and steam, respectively. is the soil ice content, is the thermal conductivity of soil, is the specific heat of liquid water, is the liquid water flux, is the steam flux, is the temperature, and are the latent heat consumed during freezing and vaporization, is the time step, is the depth of the soil layer.

[0047] Step 2.2: Construct a water balance equation based on the changes in soil water content and the water absorption capacity of vegetation.

[0048] .

[0049] in, is the change in the amount of liquid water in the soil; is the change in ice content; is the net liquid water flux; is the net steam flux; The amount of water absorbed by vegetation roots. is the hydraulic conductivity, is the soil potential energy.

[0050] Step 2.3: Construct the calculation equation of soil freeze-thaw process based on the physical mechanism of frozen soil freeze-thaw.

[0051] Specifically, this embodiment combines the hydrothermal interaction mechanism in permafrost to calculate the parameterization scheme of ice-water phase transition during soil freeze-thaw cycles.

[0052] Step 2.3.1: Modify the traditional Campbell frozen soil water potential scheme based on the action of soil pore ice to calculate the unfrozen water and ice content.

[0053] .

[0054] in, is the osmotic potential caused by the solute, and the left side of the equation is the improved one considering the pore ice The soil water potential equation of frozen soil is the Clapeyron equation for estimating soil water potential based on temperature. The unfrozen water content can be calculated by solving the two equations simultaneously. After that, the ice content is: .

[0055] in is the total water equivalent, is the soil ice content. This example fully considers the multiple effects of pore ice in frozen soil, such as the reduction in effective porosity and ice suction. This overcomes the problem that traditional soil water potential equations such as the Campbell equation, the Brooks-Corey equation, and the van Genuchten equation are not suitable for calculating the soil water potential of frozen soil. Integrating this equation into the SHAW model greatly improves the model's ability to simulate the unfrozen soil water content, ice content, and water movement in permafrost.

[0056] Step 2.3.2: Calculate the specific heat capacity and thermal conductivity by considering the combined effects of multiple soil components.

[0057] This example uses a novel weighted method that considers the combined effects of multiple components such as soil matrix, liquid water, ice, and air for calculation. The specific heat capacity is: .

[0058] in, 、 and They are elements The density, specific heat capacity and volume fraction of the soil are mainly composed of the three phases of soil texture, rock, organic matter, liquid water, soil ice and air. The thermal conductivity is: .

[0059] in, and They are elements mass fraction and thermal conductivity.

[0060] Step 2.3.3: Calculate the soil ice-affected hydraulic conductivity based on the effect of soil ice on hydraulic conductivity.

[0061] This embodiment uses a new formula that considers the effect of soil ice on hydraulic conductivity. The hydraulic conductivity affected by soil ice is: , .

[0062] in, is the hydraulic conductivity at saturation, is the soil porosity distribution coefficient, is the ice suction coefficient, is the soil potential energy under saturated soil conditions, is the soil porosity, soil potential energy Calculations were performed according to the improved Campbell matrix potential parameterization scheme proposed in this example.

[0063] Step 3: Prepare the boundary constraints of the numerical model based on the meteorological driving data and the existing temperature field data.

[0064] Among them, the constraints include lower boundary constraints and initial field constraints.

[0065] Step 3.1: Use machine learning to simulate the data set consisting of temperature field data to obtain the lower boundary temperature field constraint data, thereby constructing the lower boundary constraint condition.

[0066] In one specific embodiment, the lower boundary is the boundary of the SHAW model that defines the energy and water flux conditions at the bottom of the soil profile. The SHAW model uses a Newtonian iteration-based differential equation solving method to calculate the energy and water balance equations (equations in steps 2.1 and 2.2) for each layer of the model (vegetation, snow cover, surface, and soil) at each moment. This method requires not only upper boundary meteorological data but also a set of high-quality lower boundary constraint data at the bottom of the lowest soil layer to control the heat and water exchange between the bottom of the model soil layer and the underlying underground environment. In this embodiment, the lower boundary temperature field controls the transmission path of heat and ground temperature changes at the bottom of the permafrost layer, significantly affecting the simulation accuracy of permafrost internal temperature and ground ice changes. It is a key constraint condition that supports accurate simulation of deep permafrost. In traditional permafrost models, the lower boundary constraint condition often uses a fixed temperature as the lower boundary temperature field. This setting ignores the seasonal fluctuations of the lower boundary temperature field, which can easily introduce significant errors in permafrost temperature simulations.

[0067] Since observational data is only available at certain times and in certain areas, it cannot be extended to areas without observations. Some models can provide data from areas without observations through simulation, but the simulated data has certain errors, so error correction is required. This embodiment combines site observation data with existing numerical model simulation data and uses a random forest model to construct a lower boundary temperature construction method. This method constructs accurate and reliable lower boundary temperature field data in areas without observations, supporting the construction of permafrost hydrothermal process simulation models.

[0068] Step 3.1.1: Obtain the lower boundary temperature constraint data.

[0069] The soil temperature data at a depth of 2.7 m simulated by the Noah Land Surface Model (Noah LSM) were obtained from public data.

[0070] Step 3.1.2: Obtain actual temperature field data and construct a data set.

[0071] Long-term soil temperature observation data are obtained from multiple existing typical permafrost observation stations on the Qinghai-Tibet Plateau.

[0072] Step 3.1.3: Add preprocessing operations such as random perturbations to the dataset.

[0073] To address the problem of errors in the 2.7 m temperature field provided by Noah LSM, the 2.7 m soil temperature series provided by Noah LSM at each station was randomly perturbed from -10°C to 10°C to obtain the reduced virtual lower boundary temperature series for each station.

[0074] Step 3.1.4: Verify the accuracy of the preprocessed dataset, filter and update the data in the dataset to obtain the updated dataset.

[0075] In this example, a virtual temperature sequence is fed into the SHAW model as the lower boundary data. Combined with the meteorological driving data and key vegetation and soil parameter data obtained in step 1, this sequence drives the model to simulate permafrost hydrothermal processes. Simulated soil temperature and moisture content are then compared with the observed data for the corresponding site obtained in step 3.1.2 to verify simulation accuracy. If the accuracy meets the set threshold, this set of virtual lower boundary temperatures is used as the target ground truth to train the AI ​​model. Otherwise, a new virtual lower boundary temperature sequence is generated until a lower boundary temperature sequence that meets the threshold is found for each site.

[0076] Step 3.1.5: Train the random forest model using the updated dataset.

[0077] In this example, the target ground-truth lower boundary temperature series of all stations were integrated and used as the training set for the random forest model. The shallow layer (0.02 m), middle layer (0.39 m), and lower boundary layer (2.7 m) soil temperature series provided by Noah LSM during the same period were used as input variables.

[0078] Specifically, the target true lower boundary temperature sequence generated in step 3.1.4 is used as the training target to construct a random forest model to generate an accurate lower boundary temperature field, where 70% of the sample data is used for training and 30% is used for accuracy verification. In the random forest model hyperparameters, the number of trees is set to 500, and the number of branches of each tree model is set to 3. 100 random forest models are repeatedly generated, and the model with the best accuracy verification is selected.

[0079] Step 3.1.6: Use the updated random forest model to correct the lower boundary temperature constraint data.

[0080] Specifically, the trained optimal random forest model was used to correct the Noah LSM 2.7 m soil temperature data for the entire Qinghai-Tibet Plateau from 1979 to 2018. This corrected data served as the lower boundary temperature field constraint data, driving the constructed permafrost hydrothermal process model to simulate permafrost conditions across the entire Qinghai-Tibet Plateau. This method allows this embodiment to obtain lower boundary temperature constraint data for unobserved areas.

[0081] Step 3.2: Preheat the model to obtain cyclic preheat data. Use this data to construct initial temperature and moisture field constraints. Model-driven data, such as meteorological data, lower boundary temperature field constraints, and vegetation and soil parameters, are used to preheat the model.

[0082] In one specific embodiment, the initial field is the initial values ​​assigned to soil temperature and moisture content by the SHAW model at the start of the simulation. This field is used to initiate the time integration process and determine the evolution path of subsequent simulation results. Using inappropriate initial temperature and moisture content data for simulation can significantly impact subsequent simulation accuracy. This embodiment utilizes a preheating method driven by multiple years of real-world meteorological data to generate high-precision initial temperature and moisture fields, improving simulation accuracy.

[0083] Step 3.2.1: Obtain meteorological driving data and lower boundary temperature field constraint data for preheating.

[0084] Specifically, the meteorological data of each simulation grid and the data of the first four years (1979–1983) of the lower boundary constraint conditions were extracted.

[0085] Step 3.2.2: Perform preheating on the extracted meteorological driving data and lower boundary temperature field constraint data of the first four years to obtain cyclic preheating data.

[0086] Specifically, the meteorological driving data of the first four years and the lower boundary temperature field constraint data were replicated 10 times to generate a total of 40 years of cyclic preheating data.

[0087] Step 3.2.3: Perform error correction on the cycle warm-up data.

[0088] Specifically, the cyclic preheating data is input into the improved SHAW model in step 2 for preheating simulation, and the overall differences in simulated temperature and water content in the last 4-year cycle and the penultimate 4-year cycle are compared. If the simulation results of the two 4-year cycles are consistent within the set error range (set to 0.01), it means that the error caused by the initial field has been corrected and the cyclic preheating process is ended. If there is still inconsistency, the number of cyclic preheating times is increased until consistency is achieved.

[0089] Step 3.2.4: Extract the temperature and moisture data of each soil depth at the end of the cycle preheating as the initial field data for the formal simulation, providing the initial starting conditions for solving the energy and water balance equations in steps 2.1 and 2.2.

[0090] Step 4: Use meteorological data to drive the numerical model to simulate the permafrost hydrothermal state of the land surface characteristic data to obtain a sample set of key permafrost hydrothermal information.

[0091] In a specific implementation method, an automated parallel script is written, which integrates meteorological driving data, land surface characteristic data, lower boundary constraints and initial field constraints, and inputs the constructed numerical model of permafrost hydrothermal process. The meteorological driving data is used to drive the model to simulate the entire Qinghai-Tibet Plateau, and a model simulation result extraction and analysis script is constructed to extract key permafrost hydrothermal information such as soil temperature, unfrozen water content, soil ice content, and active layer thickness to create a sample set.

[0092] Step 4.1: For a single 10 km resolution simulation unit, build an interface and operating framework for driving data input into the model, and input meteorological driving data, vegetation and soil parameters, and boundary constraint data into the model.

[0093] Step 4.2: Use multiple CPUs on the computer to build a model runtime environment that supports parallel computing across multiple simulation units. Simulate 40 years of permafrost changes on the Qinghai-Tibet Plateau (1979-2018) and store the results on the hard disk.

[0094] Step 4.3: Write a data processing script to extract key permafrost hydrothermal characteristics data at a spatial resolution of 10 km and a temporal resolution of 1 day for the entire Qinghai-Tibet Plateau, including soil temperature, unfrozen water content, soil ice content, and active layer thickness.

[0095] Step 4.4: Write a data processing script to calculate monthly and annual data based on the 1-day time resolution data, and automatically calculate the seasonal and interannual change trends to form a sample set.

[0096] Compared with existing permafrost numerical models, this embodiment offers the following significant advantages: By improving the permafrost parameterization scheme to account for permafrost hydrothermal interactions, and by enhancing the quality of meteorological driving data, vegetation and soil parameters, and boundary constraints, the model significantly improves the accuracy of the heat and moisture coupling mechanism during freeze-thaw processes. Verification results show that compared with the traditional numerical model Noah LSM, the mean NSE of simulated soil temperature (0.35) is 0.82 higher than that of Noah LSM (-0.47), and the mean NSE of simulated soil moisture content (-0.37) is 1.02 higher than that of Noah LSM (-1.39), representing significant improvements. Furthermore, the integrity and spatial resolution of the model input data in unobserved areas are effectively improved, enhancing regional generalization capabilities. Overall, this invention improves the spatial adaptability and physical rationality of permafrost simulation, making it suitable for large-scale, multi-scale simulation and prediction of permafrost hydrothermal states in high-altitude cold regions such as the Qinghai-Tibet Plateau.

[0097] Example 2: A second embodiment of the present invention provides a permafrost hydrothermal state simulation system based on a physical model, comprising: a data acquisition module configured to acquire meteorological driving data and land surface characteristic data, wherein the land surface characteristic data includes key vegetation parameters and key soil parameters; A model building module is configured to construct a numerical model of permafrost hydrothermal processes by using land surface characteristic data and considering permafrost freeze-thaw processes and soil ice effects; In this embodiment, the model building module includes: an energy balance calculation module configured to construct an energy balance equation according to energy changes in the soil layer; A water balance calculation module is configured to construct a water balance equation based on the change of water content in the soil and the water absorption capacity of vegetation; The soil freeze-thaw cycle calculation module is configured to construct a soil freeze-thaw process calculation equation based on the physical mechanism of frozen soil freeze-thaw.

[0098] In this embodiment, the soil freeze-thaw cycle calculation module is further configured to: Calculate unfrozen water and ice content based on the frozen soil water potential according to soil pore ice action; The specific heat capacity and thermal conductivity are calculated by weighted calculation considering the composite effects of multiple soil components; The hydraulic conductivity affected by soil ice is calculated based on the effect of soil ice on hydraulic conductivity.

[0099] A constraint preparation module is configured to prepare boundary constraint conditions of the numerical model based on meteorological driving data and existing temperature field data; In this embodiment, the constraint preparation module includes: The lower boundary constraint module is configured to use machine learning to simulate a data set consisting of temperature field data to obtain lower boundary temperature field constraint data, thereby constructing a lower boundary constraint condition.

[0100] The lower boundary constraint module is also configured as follows: Get lower boundary temperature constraint data; Obtain actual temperature field data to form a data set; Add random perturbation preprocessing operations to the data set; Verify the accuracy of the preprocessed dataset, filter and update the data in the dataset to obtain an updated dataset; Use the updated dataset to train the random forest model; The updated random forest model is used to correct the lower boundary temperature constraint data.

[0101] The initial field constraint module is configured to use meteorological driving data and lower boundary temperature field constraint data for preheating processing to obtain cyclic preheating data, drive the model for preheating according to the cyclic preheating data, and construct initial field constraint conditions.

[0102] The state simulation module is configured to use meteorological driving data, constraints and vegetation soil parameters to drive the numerical model to simulate the permafrost hydrothermal state of land surface characteristic data, and obtain a sample set of key permafrost hydrothermal information.

[0103] The steps involved in the above embodiment 2 correspond to those in the method embodiment 1. For the specific implementation method, please refer to the relevant description part of the embodiment 1.

[0104] Those skilled in the art will appreciate that the units and algorithmic steps of each example described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technical personnel may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application. In the above embodiments, all or part of the embodiments can be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments can be implemented in the form of a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted via a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data processing device such as a server or data center that integrates one or more available media. Available media can include magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)). The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any technical object of a person skilled in the art that can be easily conceived of within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for simulating the hydrothermal state of permafrost based on a physical model, characterized in that: The following steps are involved: Obtain meteorological driving data and land surface characteristic data, wherein the land surface characteristic data include key vegetation parameters and key soil parameters; A numerical model of permafrost hydrothermal processes is constructed using land surface characteristic data, taking into account permafrost freeze-thaw processes and soil ice effects. Prepare boundary constraints of the numerical model based on meteorological driving data and existing temperature field data; The permafrost hydrothermal state is simulated by using meteorological data to drive the numerical model based on land surface characteristic data, and a sample set of key permafrost hydrothermal information is obtained.

2. The method for simulating the hydrothermal state of permafrost based on a physical model according to claim 1, characterized in that: The specific steps for constructing a numerical model of permafrost hydrothermal processes using land surface characteristic data and considering permafrost freeze-thaw processes and soil ice effects are as follows: Construct an energy balance equation based on the energy changes in the soil layer; Construct a water balance equation based on the changes in soil water content and the water absorption capacity of vegetation; The calculation equation of soil freeze-thaw process is constructed based on the physical mechanism of frozen soil freeze-thaw.

3. The method for simulating the hydrothermal state of permafrost based on a physical model according to claim 2, characterized in that: The specific steps for constructing the soil freeze-thaw process calculation equation based on the physical mechanism of frozen soil freeze-thaw are: Calculate unfrozen water and ice content based on the frozen soil water potential according to soil pore ice action; The specific heat capacity and thermal conductivity are calculated by weighted calculation considering the composite effects of multiple soil components; The hydraulic conductivity affected by soil ice is calculated based on the effect of soil ice on hydraulic conductivity.

4. The method for simulating the hydrothermal state of permafrost based on a physical model according to claim 1, wherein: The specific steps for preparing the boundary constraints of the numerical model based on meteorological driving data and existing temperature field data are as follows: Use machine learning to simulate the data set composed of temperature field data to obtain the lower boundary temperature field constraint data, thereby constructing the lower boundary constraint condition; The model is preheated to obtain cyclic preheating data, and the temperature and moisture initial field constraints are constructed based on the cyclic preheating data.

5. The method for simulating the hydrothermal state of permafrost based on a physical model according to claim 4, characterized in that: The specific steps for simulating a data set consisting of temperature field data using machine learning are as follows: Get lower boundary temperature constraint data; Obtain actual temperature field data to form a data set; Add random perturbation preprocessing operations to the data set; Verify the accuracy of the preprocessed dataset, filter and update the data in the dataset to obtain an updated dataset; Use the updated dataset to train the random forest model; The updated random forest model is used to correct the lower boundary temperature constraint data.

6. A permafrost hydrothermal state simulation system based on a physical model, characterized in that: include: a data acquisition module configured to acquire meteorological driving data and land surface characteristic data, wherein the land surface characteristic data includes key vegetation parameters and key soil parameters; A model building module is configured to construct a numerical model of permafrost hydrothermal processes by using land surface characteristic data and considering permafrost freeze-thaw processes and soil ice effects; A constraint preparation module is configured to prepare constraint conditions of a numerical model based on meteorological driving data and existing temperature field data; The state simulation module is configured to use meteorological driving data to drive the numerical model to simulate the permafrost hydrothermal state of the land surface characteristic data, and obtain a sample set consisting of key permafrost hydrothermal information.

7. The permafrost hydrothermal state simulation system based on a physical model according to claim 6, characterized in that: Model building modules include: an energy balance calculation module configured to construct an energy balance equation according to energy changes in the soil layer; A water balance calculation module is configured to construct a water balance equation based on the change of water content in the soil and the water absorption capacity of vegetation; The soil freeze-thaw cycle calculation module is configured to construct a soil freeze-thaw process calculation equation based on the physical mechanism of frozen soil freeze-thaw.

8. The permafrost hydrothermal state simulation system based on a physical model according to claim 7, characterized in that: The soil freeze-thaw cycle calculation module is also configured to: Calculate unfrozen water and ice content based on the frozen soil water potential according to soil pore ice action; The specific heat capacity and thermal conductivity are calculated by weighted calculation considering the composite effects of multiple soil components; The hydraulic conductivity affected by soil ice is calculated based on the effect of soil ice on hydraulic conductivity.

9. The permafrost hydrothermal state simulation system based on a physical model according to claim 6, characterized in that: The constraint preparation module includes: a lower boundary constraint module, configured to simulate a data set consisting of temperature field data using machine learning to obtain lower boundary temperature field constraint data, thereby constructing a lower boundary constraint condition; The initial field constraint module is configured to perform preheating processing on the model, obtain cyclic preheating data, and construct temperature and moisture initial field constraint conditions based on the cyclic preheating data.

10. The permafrost hydrothermal state simulation system based on a physical model according to claim 9, characterized in that: The lower boundary constraint module is also configured as follows: Get lower boundary temperature constraint data; Obtain actual temperature field data to form a data set; Add random perturbation preprocessing operations to the data set; Verify the accuracy of the preprocessed dataset, filter and update the data in the dataset to obtain an updated dataset; Use the updated dataset to train the random forest model; The updated random forest model is used to correct the lower boundary temperature constraint data.

Images