Water network reconstruction method based on forest water function combination

Abstract

The invention provides a water network reconstruction method based on forest water function combination. The method comprises the following steps: identifying water system features and constructing a topological structure; carrying out multi-dimensional forest stand attribute fusion and evaluation; optimizing a water network model based on a circuit theory; the regulation and storage function of the water network is simulated through multiple paths; and constructing a water forest habitat. According to the method, a water system network topological structure is established by adopting a spectrogram convolutional network, a forest-water interaction model is constructed by fusing multi-dimensional forest stand attributes and a circuit theory, a hydrological process is simulated, a riverbed form is adjusted, a water flow dynamic state can be simulated in various situations, the regulation and storage efficiency of a water network under different paths can be evaluated, and a water system layout and a hydrological analysis result are combined to obtain a high-performance water system. The landform transformation is implemented, the problems of water-green space splitting, fragile ecological function, single greening landscape and the like are solved, the effective river network regulation and storage space of green forest land which is green at ordinary times and is water in disaster is increased, the water forest ecological landscape is built, and important tools and strategies are provided for realizing forest-water compounding and improving city toughness.

Description

Technical Field

[0001] The present invention relates to the technical fields of forest land ecological restoration and hydrological model design, and in particular to a water network reconstruction method based on the combination of forest and water functions. Background Art

[0002] With the continuous development and construction of cities, the contradiction in land use has become extremely prominent. Due to the control of national land space and land policy restrictions, control elements such as forest land green lines and river blue lines have fragmented the continuous water-green integration space of "water-bank-land". An important means to solve this contradiction is the combined utilization of forest and water, that is, within a certain area, through the superposition of the functions of forest and water, and water and greenery, the effective regulation space of the river network that is "green in normal times and water in times of disaster" is expanded, forming the overall ecological effect of the water-green space. At present, most studies mainly focus on single water systems or the network planning of green forest lands, and it is difficult to comprehensively consider the ecological elements of both water and forest (greenery). Therefore, combining relevant disciplinary theories such as landscape ecology, fluvial geomorphology, and environmental engineering, optimizing the utilization of land and space resources, improving the water-land connectivity and regional flood control capabilities, creating an ecological landscape of "forest on water", and realizing the self-regulation and maintenance of the ecosystem. Summary of the Invention

[0003] The main purpose of the present invention is to provide a water network reconstruction method based on the combination of forest and water functions, so as to solve the problems of fragmentation of water-green space, fragile ecological functions, and single greening landscape.

[0004] To solve the above technical problems, the technical solution adopted by the present invention is: a water network reconstruction method based on the combination of forest and water functions, including the following steps:

[0005] S1. Water system feature identification and topological structure construction: Using a spectral graph convolutional network to analyze time series water flow distribution data, identifying the spatial distribution characteristics of the water system in the forest belt and extracting the topological structure of the water system network

[0006] S2. Multi-dimensional forest stand attribute fusion and evaluation: On the identified water system framework, superimposing the key attributes of the forest stand, including the small patch area A patch , surface runoff generation and concentration capacity C runoff , water system storage and detention capacity S storage , as well as vegetation water tolerance W tolerance and water demand characteristics D demand , to form a composite evaluation system;

[0007] S3. Optimization of the water network model based on circuit theory: Using circuit theory to construct a refined forest-water interaction model, simulating the hydrological process in the water network, and ensuring the minimization of Err, that is, the model simulation is highly consistent with the actual hydrological phenomenon;

[0008] S4. Regulation function of multi-path simulated water network: Based on the optimized water network model, by adjusting the riverbed width W and riverbed depth D, it is connected to the outer river, and the multi-scenario simulation technology is used to evaluate the water flow rate Q under different paths, as well as the dynamic regulation ability of the water volume at high water level, low water level, and normal water level;

[0009] S5. Construction of water forest habitat: Combine the water system layout and the results of hydrological analysis to implement terrain transformation.

[0010] In the preferred solution, in step S1, the specific process includes the following steps:

[0011] S11. Collect time series water flow distribution data within the forest belt, including satellite remote sensing images and ground monitoring station data;

[0012] S12. Use the spectral graph convolution network technology to process the time series water flow data and abstract the forest belt water system into a graph structure

[0013] Among them, is the node set, representing water areas or key points, ε represents the edge set, representing the connection relationship between water areas;

[0014] S13. Learn the water flow feature vector through spectral graph convolution Φ(G) = f(A, X) to identify the spatial distribution pattern of the water system; Apply the Kruskal algorithm to extract the minimum spanning tree of the water system network from the feature map and construct the topological structure;

[0015] Among them, Φ is the feature map, A is the adjacency matrix, X is the node feature matrix, and f is the convolution operation.

[0016] In the preferred solution, in step S2, the specific process includes the following steps:

[0017] S21. Clean and preprocess the collected forest stand attribute data, and based on the correlation analysis, select the attributes that have an impact on the ecological value and hydrological regulation ability of the forest stand, including: small patch area A patch 、Surface runoff generation and confluence ability C runoff 、Water system retention ability S storage And vegetation water tolerance W tolerance And water demand characteristics D demand ;

[0018] S22. Adopt the gradient boosting machine algorithm to learn the complex non-linear relationship between forest stand attributes by integrating multiple decision tree models. The gradient boosting machine algorithm conducts iterative training, and in each step, a new decision tree is built on the basis of the previous step's residual, gradually reducing the prediction error, and automatically evaluating and ranking the importance of each forest stand attribute;

[0019] S23. Obtain a comprehensive ecological value score for each forest stand area based on the prediction results output by the gradient boosting machine algorithm, and overlay the comprehensive score layer on the GIS platform;

[0020] S24. Establish a monitoring and feedback mechanism to dynamically adjust the evaluation system.

[0021] In the preferred solution, in step S3, using circuit theory, simulate infiltration, runoff, confluence, and storage, specifically including:

[0022] Runoff and confluence: Simplify the water system network into a circuit model, where the river section is the edge in the graph structure and R ij represents the river section resistance on the edge, corresponding to the resistance in the circuit. The volume of water flowing through a certain point or a certain river section per unit time is the flow rate Q ij , corresponding to the current in the circuit, and the head difference Δh ij represents the water level difference between any two points in the water system. According to Ohm's law, the flow rate of each river section is simplified as:

[0023]

[0024] where Q ij is the flow rate from node i to node j, Δh ij is the head difference between the two end nodes, and R ij is the river section resistance;

[0025] Infiltration process: Add infiltration simulation to each river section. The infiltration capacity R sink,i,j of a river section characterizes the soil permeability, corresponding to the adjustable resistance in the circuit. Connect the end of the river section to the soil layer to simulate the infiltration flow. The calculation formula for the infiltration flow rate is:

[0026]

[0027] where Δh i is the head difference between the river section from node i to the next node;

[0028] Storage process: Introduce the storage capacity C storage at the water storage area, corresponding to the node capacitance in the circuit. The water storage area is the node in the graph structure , reflecting its storage capacity. The calculation formula for the rate of change of the water volume in the water storage area is:

[0029]

[0030] where C starage,k represents the storage capacity of the kth node, and I in,k and I out,k are the inflow and outflow flows of the water storage area respectively.

[0031] The storage capacity C of a certain reservoir starage,k can help understand the dynamic changes of the nodes, such as C starage,k is large. Even with a large flow difference (I in,k -I out,k ), the rate of change of the water volume will be relatively small because the nodes have sufficient capacity to buffer this change, which is convenient for the staff to adjust and control.

[0032] In the preferred solution, in step S3, it specifically includes the following steps:

[0033] S31. Construct a comprehensive circuit equations set including all resistors, capacitors, and infiltration resistors;

[0034] S32. According to step S1 and step S2, set initial values for the river section resistors, infiltration resistors, and storage area capacitors, and construct a comprehensive error function:

[0035]

[0036] where obs represents the observed value, model represents the model predicted value. Specifically, Δh ij,model and Δh ij,obs respectively represent the predicted head difference and the observed head difference between node i and node j, Q sink,i,model and Q sink,i,obs respectively represent the predicted infiltration flow and the actual infiltration flow of the river section between node i and the next node, and represent the predicted water volume change rate and the actual water volume change rate of the storage area at the i-th node;

[0037] S33. Adopt genetic algorithm or particle swarm optimization to adjust the parameters and minimize Err.

[0038] In the preferred solution, in step S31, it specifically includes the following steps:

[0039] Utilize the simulated infiltration, runoff, confluence, and storage analysis in step S1 to establish the equations set for each node respectively, and obtain the dynamic solution of the water level V of each node in the water network system changing with time. The specific equation is, for node k there is:

[0040]

[0041] where Δh iv represents the head difference between the i-th node and the k-th node, that is, for the k-th node, Δh vj represents the head difference between the k-th node and the j-th node.

[0042] In the preferred solution, in step S4, by using the optimized water network model obtained in step S3, the water storage capacity of the water network is enhanced by adjusting the riverbed parameters. The specific process is as follows: The variation of the river width W and depth D affects the hydraulic radius R. Monte Carlo or multi - scenario simulation is used to evaluate the water flow rate Q and the water surface regulation ability under different W and D.

[0043] In the preferred solution, in step S4, the specific formula for the hydraulic radius R is:

[0044]

[0045] Where A is the cross - sectional area of the water flow, P is the wetted perimeter, W is the river width, and D is the river depth.

[0046] In the preferred solution, in step S4, the calculation formula for the water flow rate Q is:

[0047]

[0048] Where n is the Manning roughness coefficient, R is the hydraulic radius, and S is the slope of the channel or river, that is, the ratio of the head difference to the horizontal distance.

[0049] In the preferred solution, in step S4, when the river width W and depth D are changed, the change amounts W - W 0 ≥1 m, D - D 0 ≤0.5 m, where W 0 and D 0 are the original river bottom width and river depth.

[0050] In the preferred solution, in step S5, the construction of the aquatic forest habitat includes terrain transformation, vegetation configuration, monitoring and feedback;

[0051] Terrain transformation: According to the results of hydrological simulation, design a terrain transformation plan, including building dikes, digging and filling ditches, and optimizing the surface runoff path;

[0052] Vegetation configuration: Under the guidance of the composite evaluation system in step S2, select suitable tree species and planting patterns;

[0053] Monitoring and feedback: Establish a long - term monitoring mechanism, evaluate the eco - hydrological effects, and continuously adjust and optimize management measures.

[0054] In the preferred solution, in step S5, the landscape creation specifically includes the following steps:

[0055] S51. Divide the forest - water composite plot into three areas according to the water level: the long - term flooded area, the short - term flooded area, and the bank slope area;

[0056] S52. Plant adaptable tree species in different areas. For long-term and short-term flooded areas, plant water-tolerant tree species, and select container seedlings with a diameter at breast height of 6.0 - 8.0 cm. Control the planting density at 2,000 - 2,500 plants / hm 2 ; For the bank slope area, plant tree species that like humidity and have well-developed root systems, and the planting density is 3,000 - 4,000 plants / hm 2 .

[0057] S53. Implement water level regulation measures, including gate control, to achieve dynamic management of flooded areas;

[0058] S54. Establish a connection mechanism with the external water network.

[0059] The present invention provides a water network reconstruction method based on the integration of forest and water functions, innovatively integrating modern data analysis and ecohydrology, aiming to optimize the collaborative functions of forests and water systems. First, analyze water flow data with a spectral graph convolutional network to accurately identify and construct the water system structure in the forest; incorporate multi-dimensional stand attributes, and use the gradient boosting machine algorithm to evaluate the ecological value and guide vegetation configuration; through circuit theory simulation and optimization algorithms, carefully simulate the hydrological cycle process to ensure that the model accurately reflects reality; adjust the riverbed size to simulate various water flow scenarios and analyze the regulation potential; finally, based on these analysis results, carry out terrain transformation, vegetation planning and long-term monitoring to form a water-tolerant and ecological forest habitat. This solution not only improves the scientificity and efficiency of water resource management, but also promotes the sustainable development of the ecological environment, and is an effective strategy to address environmental challenges.

[0060] The beneficial effects of the present invention are as follows

[0061] 1) Through high-precision water system feature identification and stand attribute evaluation, it provides a scientific basis for water resource management and ecological restoration, making decision-making more accurate and efficient.

[0062] 2) The application of the circuit theory model greatly improves the accuracy of hydrological process simulation, helps to deeply understand the operation mechanism of the water system, and provides a powerful tool for flood control and disaster reduction, and water resource allocation.

[0063] 3) Dynamically adjust the riverbed parameters, combined with multi-scenario simulation, effectively improves the adaptability and regulation efficiency of the water network to different hydrological conditions, and ensures the reasonable utilization of water resources and ecological security.

[0064] 4) The vegetation configuration strategy based on ecological value evaluation, and comprehensive terrain transformation, promote the natural restoration of the ecosystem and the protection of biodiversity, and construct a more stable and rich water forest habitat.

[0065] 5) The establishment of a comprehensive monitoring and feedback mechanism ensures the continuous optimization and adjustment of eco-hydro management measures, supports the sustainable management of water resources and the long-term health of ecosystems, and is of great significance for promoting the construction of ecological civilization and addressing climate change. Brief Description of the Drawings

[0066] The present invention will be further described below in conjunction with the drawings and embodiments:

[0067] Figure 1 is a flowchart of the water network reconstruction method of the present invention;

[0068] Figure 2 is the spatial distribution characteristics of the water system before adjustment in the embodiment of the present invention;

[0069] Figure 3 is the adjusted forest-water interaction model in the embodiment of the present invention;

[0070] Figure 4 is the adjusted low-water-level water network model in the embodiment of the present invention;

[0071] Figure 5 is the adjusted high-water-level water network model in the embodiment of the present invention;

[0072] In the figure: Taxodium ascendens forest 1; Koelreuteria paniculata forest 2; Salix babylonica forest 3; Zelkova serrata forest 4; Taxodium distichum forest 5; Pterocarya stenoptera forest 6. Detailed Embodiment

[0073] Embodiment 1

[0074] As Figures 1 to 5 shown, a water network reconstruction method based on the compound function of forest and water includes the following steps:

[0075] S1. Identification of water system characteristics and construction of topological structure: Using a spectral graph convolutional network to analyze time series water flow distribution data, identify the spatial distribution characteristics of the water system in the forest belt and extract the topological structure of the water system network

[0076] S2. Fusion and evaluation of multi-dimensional forest stand attributes: On the identified water system framework, overlay the key attributes of the forest stand, including the small patch area A patch , surface runoff generation and concentration capacity C runoff , water storage and retention capacity S storage as well as vegetation water tolerance W tolerance and water demand characteristics D demand , to form a composite evaluation system;

[0077] S3. Optimization of the water network model based on circuit theory: Using circuit theory, construct a fine forest-water interaction model to simulate the hydrological process in the water network and ensure the minimization of Err, that is, the model simulation is highly consistent with the actual hydrological phenomenon;

[0078] S4. Regulation function simulation of multi-path simulated water network: Based on the optimized water network model, by adjusting the riverbed width W and riverbed depth D, it is connected to the external river, and the multi-scenario simulation technology is used to evaluate the water flow rate Q under different paths, as well as the dynamic regulation ability of the water volume at high water level, low water level, and normal water level;

[0079] S5. Construction of water forest habitat: Combine the water system layout and the results of hydrological analysis to implement terrain transformation.

[0080] This optimal solution systematically realizes the precise management and optimization of the forest water ecosystem through five steps, aiming to construct a healthy and fully functional water forest habitat. The advanced spectral graph convolutional network technology is used to accurately identify the water system distribution characteristics and construct its topological structure, laying a solid foundation for subsequent analysis; through the integration and evaluation of multi-dimensional stand attributes, not only the ecological value is considered, but also the hydrological regulation ability is deeply analyzed to form a comprehensive ecological evaluation system; the circuit theory is introduced to optimize the water network model, making the simulation closer to the real hydrological process and ensuring the accuracy of the model; on this basis, by adjusting the riverbed parameters, the regulation function of the water network is simulated to enhance its ability to respond to different water levels and ensure the stability of the water environment; combined with terrain transformation and vegetation configuration, a water forest habitat that can both resist floods and conserve water sources is created, and at the same time, a monitoring and feedback mechanism is established to continuously optimize the management strategy to ensure the long-term stable operation of the ecosystem. This solution not only improves the self-recovery ability of the forest ecosystem, but also enhances its adaptability to climate change, which is of great significance for maintaining biodiversity and promoting sustainable development.

[0081] In the optimal solution, in step S1, the specific process includes the following steps:

[0082] S11. Collect time-series water flow distribution data within the forest belt, including satellite remote sensing images and ground monitoring station data;

[0083] S12. Use the spectral graph convolutional network technology to process the time-series water flow data and abstract the forest belt water system into a graph structure

[0084] Among them, is the node set, representing water areas or key points, and ε represents the edge set, representing the connection relationship between water areas;

[0085] S13. Learn the water flow feature vector through spectral graph convolution Φ(G) = f(A, X) to identify the water system spatial distribution pattern; apply the Kruskal algorithm to extract the minimum spanning tree of the water system network from the feature map and construct the topological structure;

[0086] Among them, Φ is the feature map, A is the adjacency matrix, X is the node feature matrix, and f is the convolution operation.

[0087] The above solution can efficiently and accurately identify and construct the water system network structure within the forest belt. By collecting and analyzing the time-series water flow distribution data and adopting the advanced spectral graph convolutional network technology, the complex water system information can be transformed into an intuitive graph structure, where the nodes and edges represent water areas and their connections respectively. The key advantage of this step is the ability to learn the feature vectors of water flow and reveal the spatial distribution law of the water system. With the help of the Kruskal algorithm to extract the minimum spanning tree, the topological structure of the water system network is further refined, providing solid data support for subsequent water resource management and ecological balance analysis, and ensuring the accuracy and effectiveness of the subsequent steps. This data-driven method greatly improves the accuracy and efficiency of water system identification and lays a foundation for the scientific management of forest belt water resources.

[0088] In the preferred solution, in step S2, the specific process includes the following steps:

[0089] S21. Clean and preprocess the collected forest stand attribute data. Based on the correlation analysis, select the attributes that have an impact on the ecological value and hydrological regulation ability of the forest stand, including: small patch area A patch , surface runoff generation and concentration ability C runoff , water system retention ability S storage and vegetation water tolerance W tolerance and water demand characteristics D demand ;

[0090] S22. Adopt the gradient boosting machine algorithm to learn the complex non-linear relationships among forest stand attributes by integrating multiple decision tree models. The gradient boosting machine algorithm conducts iterative training, and in each step, a new decision tree is established based on the residuals of the previous step, gradually reducing the prediction error, and automatically evaluating and ranking the importance of each forest stand attribute;

[0091] S23. According to the prediction results output by the gradient boosting machine algorithm, obtain a comprehensive ecological value score for each forest stand area, and overlay the comprehensive score layer on the GIS platform;

[0092] S24. Establish a monitoring and feedback mechanism to dynamically adjust the evaluation system.

[0093] The goal is to refine the assessment of the ecological value and hydrological regulation function of forest stands. Through in-depth cleaning and preprocessing of forest stand attribute data, combined with correlation analysis, indicators crucial for ecology and hydrology are selected, such as small patch area, surface runoff and confluence capacity, etc. Then, using the gradient boosting machine algorithm, which can capture the intricate non-linear relationships between attributes, by integrating multiple decision trees, the prediction error is gradually reduced layer by layer, and at the same time, the influence weights of each attribute on the forest stand value are quantified. Finally, based on the algorithm prediction, a comprehensive ecological value score is assigned to each forest stand area, providing a comprehensive score for the subsequent selection of suitable tree species and planting patterns. This not only provides a scientific basis for guiding the optimal allocation of forestry resources, but also establishes a dynamic monitoring and feedback mechanism to ensure the continuous improvement of the evaluation system with environmental changes, effectively promoting the health of the forest ecosystem and the sustainable management of water resources.

[0094] In the preferred solution, in step S3, using circuit theory, infiltration, runoff, confluence, and storage are simulated, specifically including:

[0095] Runoff and confluence: Simplify the water system network into a circuit model, with river reaches as the edges in the graph structure R ij represents the resistance of the river reach on the edge, corresponding to the resistance in the circuit. The volume of water flowing through a certain point or a certain river reach per unit time is the flow rate Q ij , corresponding to the current in the circuit, and the head difference Δh ij represents the water level difference between any two points in the water system. According to Ohm's law, the flow rate of each river reach is simplified as:

[0096]

[0097] where Q ij is the flow rate from node i to node j, Δh ij is the head difference between the two end nodes, and R ij is the river reach resistance;

[0098] Infiltration process: Add infiltration simulation to each river reach. The infiltration capacity R sink,i,j of a river reach characterizes the soil permeability, corresponding to the adjustable resistance in the circuit. Connect the end of the river reach to the soil layer to simulate the infiltration flow. The calculation formula for the infiltration flow rate is:

[0099]

[0100] where Δh i is the head difference between the river reach from node i to the next node;

[0101] Storage process: Introduce the storage capacity C storage at the storage area, corresponding to the node capacitance in the circuit. The storage area is the graph structure The nodes in the [specific context] reflect their storage capacity. The calculation formula for the rate of change of water volume in the storage area is:

[0102]

[0103] Among them, C starage,k represents the storage capacity of the k-th node, and I in,k and I out,k are the inflow and outflow rates of the storage area respectively.

[0104] The storage capacity C of a certain reservoir starage,k can help understand the dynamic changes of the nodes. For example, if C starage,k is large, even with a large flow difference (I in,k -I out,k ), the rate of change of water volume will be relatively small because the nodes have enough capacity to buffer this change, which is convenient for the staff to adjust and control.

[0105] By mapping the hydrological process to the framework of circuit theory, the infiltration, runoff, confluence, and storage phenomena in the forest belt water system are creatively simulated. The complex hydrological problems are transformed into an easy-to-understand and calculable circuit model, where the river section is regarded as a resistor, the head difference is equivalent to voltage, and the water flow corresponds to current. Through Ohm's law, the flow rate of the river section can be accurately calculated, and thus how the water body flows and converges in the forest belt can be understood. The infiltration process is ingeniously modeled as an adjustable resistor, reflecting the permeability of the soil, while the storage area is presented in the form of a capacitor, quantifying its water storage capacity. The application of this circuit theory not only greatly simplifies the complexity of hydrological simulation, improves the calculation efficiency, but also makes the adjustment of model parameters more intuitive, facilitating researchers to deeply explore the internal mechanism of the hydrological process, thereby providing a scientific basis for water resource management and ensuring the health and stability of the forest ecosystem.

[0106] In the preferred solution, in step S3, it specifically includes the following steps:

[0107] S31. Construct a comprehensive circuit equation system including all resistors, capacitors, and infiltration resistors;

[0108] S32. According to step S1 and step S2, set the initial values of the river section resistor, infiltration resistor, and storage area capacitor, and construct a comprehensive error function:

[0109]

[0110] Among them, obs represents the observed value, model represents the model prediction value. Specifically, Δh ij,model and Δh ij,obs respectively represent the predicted head difference and the observed head difference between the i-th node and the j-th node, and Q sink,i,model and Q sink,i,obsrespectively represent the predicted infiltration flow and the actual infiltration flow of the river section between the i-th node and the next node, and represent the predicted change rate of the water volume in the storage area and the actual change rate of the water volume in the storage area of the i-th node;

[0111] S33. Adopt genetic algorithm or particle swarm optimization to adjust the parameters and minimize Err.

[0112] Under the framework of circuit theory, a hydrological model is constructed and optimized to achieve accurate simulation of the forest belt water system process. By integrating the circuit equations of resistors (representing river sections and infiltration) and capacitors (reflecting storage areas), a comprehensive model is formed, which can describe the flow, infiltration and storage of water in detail. Based on the data in steps S1 and S2, initial values are set for the model parameters, an error function is constructed, and the differences between the observed values and the model predicted values are compared. Using intelligent optimization techniques such as genetic algorithm or particle swarm optimization, these parameters are dynamically adjusted to minimize the error and ensure the high accuracy of the model. The advantage of this method is that it can efficiently calibrate the model parameters, improve the reliability of hydrological simulation, provide strong support for the rational planning of forest water resources and ecological management, and at the same time provide a scientific basis for formulating strategies to cope with climate change and extreme hydrological events.

[0113] In the preferred solution, in step S31, it specifically includes the following steps:

[0114] Using the simulated infiltration, runoff, confluence and storage analysis in step S1, establish the equations of each node respectively, and obtain the dynamic solution of the water level V of each node in the water network system changing with time by solving the equations. The specific equation is, for node k There is: There is:

[0115]

[0116] Among them, Δh iv represents the head difference between the i-th node and the k-th node, that is, the, Δh vj represents the head difference between the k-th node and the j-th node.

[0117] By simulating the infiltration, runoff, confluence, and storage processes, it is transformed into a mathematical expression under circuit theory, and a set of equations describing the water level change over time is established for each node. The concepts of components such as resistors and capacitors in circuit theory are used to characterize the physical properties of water flow in the water system network, and then a complete hydrological dynamic model is formed. By solving these equations, the dynamic solutions of the water levels at each node in the water network system over time can be obtained. Doing so not only simplifies the complex hydrological process into an easily solvable computational problem but also can more accurately predict the change trend of the water level, providing an important scientific basis for flood warning, water resource management, and water conservancy project design.

[0118] In the preferred solution, in step S4, using the optimized water network model obtained in step S3, by adjusting the riverbed parameters to enhance the water storage and regulation capacity of the water network, the specific process is as follows: Changing the river width W and depth D affects the hydraulic radius R, and Monte Carlo or multi-scheme simulation is used to evaluate the water flow rate Q and the water surface regulation ability under different W and D.

[0119] By adjusting the width and depth of the riverbed, optimizing the hydraulic radius, the water storage and regulation capacity of the water network and the control of the water flow rate can be significantly improved. Using the optimized water network model and combining Monte Carlo or multi-scheme simulation technology, the hydrodynamic performance under different riverbed parameters can be systematically evaluated to ensure effective regulation of the water body at high, low, and normal water surfaces, enhancing the resistance of the forest belt to floods and droughts and promoting the sustainable management of water resources. This strategy helps to maintain ecological balance, ensure the safety of surrounding communities, and at the same time optimize the natural landscape and biodiversity protection.

[0120] In the preferred solution, in step S4, the specific formula for the hydraulic radius R is:

[0121]

[0122] Where A is the cross-sectional area of flow, P is the wetted perimeter, W is the river width, and D is the river depth.

[0123] The hydraulic radius is a key parameter for evaluating the water flow efficiency. Its calculation formula considers the proportional relationship between the cross-sectional area of flow and the wetted perimeter, intuitively reflecting the influence of the river channel geometry on the water flow velocity. Optimizing the hydraulic radius by adjusting the river width and depth in the above formula can effectively improve the water conveyance capacity and water storage and regulation function of the water network. This not only helps to accelerate flood discharge and reduce the risk of water disasters but also can maintain the necessary water level during the dry period to ensure the ecological water demand, thus realizing the balanced distribution of water resources and enhancing the ecological stability and resistance of the water system.

[0124] In the preferred solution, in step S4, the calculation formula for the water flow rate Q is:

[0125]

[0126] Among them, n is the Manning roughness coefficient, R is the hydraulic radius, and S is the slope of the channel or river, that is, the ratio of the head difference to the horizontal distance.

[0127] Calculating the flow rate through the Manning formula can accurately quantify the influence of the hydraulic radius, slope, and roughness coefficient on the flow velocity. It helps to optimize the river channel design, ensure smooth water flow, reduce hydraulic losses, and also plays a key role in flood management, such as accelerating flood discharge and reducing flood disasters. At the same time, during the dry season, a reasonable flow rate helps to maintain the necessary water level, ensure the water supply of the ecosystem, and thus achieve the effective management of water resources and the balanced protection of the ecological environment.

[0128] In the preferred solution, in step S4, when changing the channel width W and depth D, the change amount W - W 0 ≥1 m, D - D 0 ≤0.5 m, where W 0 and D 0 are the original river bottom width and channel depth.

[0129] Adjusting the channel width and depth affects the flow velocity and capacity of the river, improves the self-purification ability of the water body, enhances the flood control efficiency, and promotes biodiversity, achieving better water resource management and ecological balance.

[0130] In the preferred solution, in step S5, the construction of the aquatic forest habitat includes terrain transformation, vegetation configuration, monitoring, and feedback;

[0131] Terrain transformation: According to the results of hydrological simulation, design a terrain transformation plan, including building dams, excavating and filling ditches, and optimizing the surface runoff path;

[0132] Vegetation configuration: Under the guidance of the composite evaluation system in step S2, select suitable tree species and planting patterns;

[0133] Monitoring and feedback: Establish a long-term monitoring mechanism, evaluate the eco-hydrological effects, and continuously adjust and optimize the management measures.

[0134] The construction of a water forest habitat aims to create an environment that is beneficial for both water resource management and the promotion of biodiversity. Through terrain transformation, such as building dikes and excavating ditches, surface runoff can be effectively controlled, flood risks reduced, and water supply ensured during dry periods. Vegetation configuration strategies are based on the evaluation of the stand attributes in the early stage, accurately positioning tree species and planting patterns, which not only enhance the self-recovery ability of the ecosystem but also improve the hydrological regulation function of the forest land. In addition, the long-term monitoring and feedback mechanism ensure the continuous optimization of the ecological project, promptly responding to changes in climate and ecosystem needs, guaranteeing the healthy and stable development of the water forest complex system, and thus achieving the sustainable management of water resources and the harmonious coexistence of the ecological environment. This series of measures jointly promote soil and water conservation, water quality purification, and the improvement of biological habitats, providing multiple benefits for both humans and nature.

[0135] In the preferred solution, in step S5, the landscape creation specifically includes the following steps:

[0136] S51. Divide the forest-water complex plot into three regions according to the water level: the long-term flooded area, the short-term flooded area, and the bank slope area;

[0137] S52. Plant adaptable tree species in different regions. In the long-term flooded area and the short-term flooded area, plant water-tolerant tree species, select container seedlings with a diameter at breast height of 6.0 - 8.0 cm, and control the planting density at 2000 - 2500 plants / hm 2 ; in the bank slope area, plant tree species that like moisture and have well-developed roots, and the planting density is 3000 - 4000 plants / hm 2 .

[0138] S53. Implement water level regulation measures, including gate control, to achieve dynamic management of the flooded area;

[0139] S54. Establish a connection mechanism with the external water network.

[0140] The vegetation configuration plans the ecological pattern of the forest-water complex plot. By subdividing the plot into the long-term flooded area, the short-term flooded area, and the dry bank slope area, and specifically planting adaptable tree species, it not only strengthens the hydrological regulation function of the ecosystem but also significantly improves biodiversity. The reasonable layout of water-tolerant tree species in the flooded area, combined with the dynamic management of the water level controlled by the gate, effectively alleviates the flood threat and promotes the natural circulation of water resources. The vegetation configuration in the bank slope area stabilizes the soil, reduces erosion, and beautifies the landscape. Establishing a connection mechanism with the external water network further optimizes water body flow, enhances the resistance and self-repair ability of the entire system, provides a rich and diverse habitat environment for wild animals and plants, and achieves a win-win situation for ecological, economic, and social benefits. This scientific ecological engineering design demonstrates the beautiful vision of the harmonious coexistence of humans and nature.

[0141] Example 2

[0142] As further illustrated in Embodiment 1, as Figures 1 - 5 shown,

[0143] Step 1: Identification of water system characteristics and construction of topological structure: Collect time series water flow distribution data within the forest belt, including satellite remote sensing images and ground monitoring station data, specifically including the topography, vegetation coverage rate, and soil type in the forest belt area. Based on the above information, divide the forest belt area into several nodes, and connect the nodes with edges to represent the geographical adjacency relationship.

[0144] As Figure 2 shown; Use the spectral graph convolutional network to analyze the time series water flow distribution data, identify the spatial distribution characteristics of the water system in the forest belt, and extract the topological structure of the water system network where, is the node set, representing water areas or key points, ε represents the edge set, representing the connection relationship between water areas;

[0145] Learn the water flow feature vectors through the spectral graph convolution Φ(G) = f(A, X), including the average flow velocity, flow rate change trend, and vegetation coverage rate, and identify the spatial distribution pattern of the water system.

[0146] Apply the Kruskal algorithm to extract the minimum spanning tree of the water system network from the feature map and construct the topological structure, where Φ is the feature map, A is the adjacency matrix, X is the node feature matrix, and f is the convolution operation.

[0147] Step 2: Multi-dimensional forest stand attribute fusion and evaluation: Clean and preprocess the collected forest stand attribute data. Based on the correlation analysis, select the attributes that have an impact on the ecological value and hydrological regulation ability of the forest stand as shown in Table 1, including: small patch area A patch , surface runoff generation and concentration ability C runoff , soil water storage and retention ability S storage , and vegetation water tolerance W tolerance and water demand characteristics D demand ; Use the gradient boosting machine algorithm to learn the complex non-linear relationships between forest stand attributes by integrating multiple decision tree models. The gradient boosting machine algorithm performs iterative training, and in each step, a new decision tree is built based on the residuals of the previous step, gradually reducing the prediction error, and automatically evaluating and ranking the importance of each forest stand attribute; According to the prediction results output by the gradient boosting machine algorithm, obtain a comprehensive ecological value score for each forest stand area as shown in Table 2, and overlay the comprehensive score layer on the GIS platform; Establish a monitoring and feedback mechanism to dynamically adjust the evaluation system.

[0148] Table 1 Classification of evaluation indicators

[0149]

[0150] Table 2 Comprehensive ecological values of different forest stands

[0151]

[0152]

[0153] S3. Optimization of the water network model based on circuit theory: Using circuit theory, a refined forest-water interaction model is constructed, as Figure 3 shown, to simulate the hydrological processes in the water network and ensure the minimization of Err, that is, the model simulation highly coincides with the actual hydrological phenomena;

[0154] Using circuit theory, simulate infiltration, runoff, confluence, and storage, specifically including:

[0155] Runoff and confluence: Simplify the water system network into a circuit model, where the river section is the edge in the graph structure and represents the resistance R ij , and the volume of water flowing through a certain point or a certain river section per unit time is the flow rate Q ij , corresponding to the current in the circuit, and the water head difference Δh ij represents the water level difference between any two points in the water system. According to Ohm's law, the flow rate of each river section is expressed as:

[0156]

[0157] where Q ij is the flow rate from node i to node j, Δh ij is the water head difference between the two end nodes, and R ij is the resistance of the river section;

[0158] Infiltration process: Add infiltration simulation to each river section, and use an adjustable resistance R sink,i to characterize the soil permeability, connect the end of the river section to the ground, and simulate the infiltration flow. The calculation formula for the infiltration flow rate is:

[0159]

[0160] where Δh i is the water head of this river section;

[0161] Storage process: Introduce a capacitance C storage,k at the water storage area. The water storage area is the point in the graph structure , which reflects its storage capacity. The calculation formula for the rate of change of the water volume in the water storage area is:

[0162] where I in,k and I out,k are the flow rates flowing into and out of the water storage area respectively.

[0163] Construct a comprehensive error function:

[0164]

[0165] Among them, obs represents the observed value, and model represents the model predicted value; use genetic algorithm or particle swarm optimization to adjust the parameters to minimize Err.

[0166] S4. Regulation function of multi-path simulated water network: Based on the optimized water network model, by adjusting the riverbed width W channel and the riverbed depth D channel , realize the connection with the outer river, and use multi-scenario simulation technology to evaluate the water flow rate Q under different paths, as well as the dynamic regulation ability of the water volume at high water level, low water level, and normal water level. As shown in Table 3, obtain examples of the influence of different riverbed widths, riverbed depths, precipitation, and water surface heights on the water flow rate;

[0167] The specific formula for the hydraulic radius R is:

[0168]

[0169] Among them, A is the cross-sectional area of the water flow, P is the wetted perimeter, W is the river channel width, D is the river channel depth, K is a constant, and n is the Manning coefficient.

[0170] The calculation formula for the water flow rate Q is:

[0171]

[0172] Among them, n is the Manning roughness coefficient, R is the hydraulic radius, and S is the slope of the channel or river, that is, the ratio of the water head difference to the horizontal distance.

[0173] Vary the river channel width W and depth D, then the change amount W - W 0 ≥1 m, D - D 0 ≤0.5 m, where W 0 and D 0 are the original river bottom width and river channel depth.

[0174] Table 3 Influence of different riverbed widths, riverbed depths, precipitation, and water surface heights on the water flow rate

[0175]

[0176]

[0177] Statistically according to the normal water level range, the adjusted water area is about 1.93 hectares, the forest area is about 6.70 hectares, and the forest-water composite area is about 0.88 hectares, as Figure 4 , Figure 5 shown.

[0178] S5. Construction of aquatic forest habitat: Combining the water system layout and the results of hydrological analysis, implement topographical transformation, including building dikes, excavating and filling ditches, and optimizing the surface runoff path.

[0179] Divide the forest-water composite plot into three areas: long-term flooded area, short-term flooded area, and bank slope area. In the long-term flooded area (water depth not exceeding 50 cm), plant water-tolerant tree species such as Taxodium distichum and Taxodium ascendens. In the short-term flooded area, plant Salix babylonica and Pterocarya stenoptera, and select container seedlings with a breast diameter of 6.0 - 8.0 cm for planting in April - May, with the planting density controlled at 2300 plants / hm 2 . In the bank slope area, plant Koelreuteria paniculata and Zelkova schneideriana, which are moisture-loving and have well-developed roots, with a planting density of 3500 plants / hm 2 . After August, control the water level through the gate for two intermittent floods; after September, connect with the external water network. Through long-term monitoring and management, the "aquatic forest" created will form a forest and a scenic view after three years.

[0180] Based on the above analysis results, implement measures such as topographical transformation and plant planting to scientifically configure the aquatic forest habitat. The entire plan integrates modern data analysis technology and the principles of ecological hydrology, realizes the precise management of the aquatic forest ecosystem and the sustainable development strategy, and improves the efficiency of ecological restoration and the sustainable utilization level of water resources.

[0181] The above embodiments are only the preferred technical solutions of the present invention and should not be regarded as limitations on the present invention. The protection scope of the present invention should be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A water network reconstruction method based on forest-water functional integration, characterized by: The following steps are involved: S1. Identification of water system characteristics and construction of topological structure: Spectral convolutional network is used to analyze time series water flow distribution data, identify the spatial distribution characteristics of water systems in forest belts and extract the topological structure of water system networks. S2. Multi-dimensional forest stand attribute fusion and evaluation: On the identified water system framework, key forest stand attributes are superimposed, including small patch area A patch , surface runoff generation and convergence capacity C runoff , water system storage capacity S storage and vegetation water tolerance tolerance and water demand characteristics D demand , forming a composite evaluation system; S3. Optimization of water network model based on circuit theory: Using circuit theory, a sophisticated forest-water interaction model is constructed to simulate the hydrological process in the water network to ensure that Err is minimized, that is, the model simulation is highly consistent with the actual hydrological phenomenon; S4. Multi-path simulation of water network regulation and storage function: Based on the optimized water network model, the riverbed width W and riverbed depth D are adjusted to achieve connectivity with the external river. Multi-scenario simulation technology is used to evaluate the flow rate Q under different paths, as well as the dynamic regulation capacity of water volume at high water level, low water level and normal water level. S5. Construction of water forest habitat: Implement terrain transformation based on the water system layout and hydrological analysis results.

2. The water network reconstruction method based on forest-water functional integration according to claim 1 is characterized by: In step S1, the specific process includes the following steps: S11. Collect time series water flow distribution data within the forest belt, including satellite remote sensing images and ground monitoring station data; S12. Use spectral graph convolutional network technology to process time series water flow data and abstract the forest belt water system into a graph structure G(v,ε); in, is a node set, representing water areas or key points, and ε is an edge set, representing the connection relationship between water areas; S13. Learn the water flow feature vector through spectral graph convolution Φ(G)=f(A,X) to identify the spatial distribution pattern of the water system; apply the Kruskal algorithm to extract the minimum spanning tree of the water system network from the feature graph and construct the topological structure; Among them, Φ is the feature graph, A is the adjacency matrix, X is the node feature matrix, and f is the convolution operation.

3. The water network reconstruction method based on forest-water functional integration according to claim 1 is characterized in that: step S2 The specific process includes the following steps: S21. Clean and preprocess the collected stand attribute data, and select the attributes that have an impact on the stand ecological value and hydrological regulation capacity based on correlation analysis, including: small patch area A patch , surface runoff generation and convergence capacity C runoff , water system storage capacity S storage and vegetation water tolerance tolerance and water demand characteristics D demand ; S22, using the gradient boosting algorithm to learn the complex nonlinear relationship between forest stand attributes by integrating multiple decision tree models. The gradient boosting algorithm uses iterative training to build a new decision tree based on the residual of the previous step at each step, gradually reducing the prediction error, and automatically evaluating and ranking the importance of each forest stand attribute; S23, obtaining a comprehensive ecological value score for each forest area according to the prediction results output by the gradient boosting algorithm, and overlaying the comprehensive score layer on the GIS platform; S24. Establish a monitoring and feedback mechanism and dynamically adjust the evaluation system.

4. The water network reconstruction method based on forest-water functional integration according to claim 1 is characterized by: In step S3, circuit theory is used to simulate infiltration, runoff, confluence, and accumulation, including: Runoff and confluence: Simplify the water network into a circuit model, and the river section is a graph structure The edge in R ij It represents the resistance of the river section on the side, corresponding to the resistance in the circuit. The volume of water flowing through a certain point or a certain river section per unit time is the flow rate Q ij , corresponding to the current in the circuit, the head difference Δh ij Represents the water level difference between any two points in the water system. According to Ohm's law, the flow of each river section is simplified as: Among them, Q ij is the flow from node i to node j, Δh ij is the water head difference between the two nodes, R ij is the river section resistance; Infiltration process: Add infiltration simulation for each river section, and use the infiltration capacity R of a river section to calculate the infiltration capacity R of a river section. sink,i,j Characterize the soil permeability, corresponding to the adjustable resistance in the circuit, connect the end of the river section to the soil layer, simulate the seepage flow, and the calculation formula of the seepage flow is: Where Δh i is the water head difference of the river section from the i node to the next node; Storage process: Introducing water storage capacity C in the water storage area storage , corresponding to the node capacitance in the circuit, the water storage area is the graph structure The nodes in the middle reflect its storage capacity. The water volume change rate calculation formula of the water storage area is: Among them, C starage,k represents the water storage capacity of the kth node, I in,k and I out,k are the flows into and out of the storage area, respectively.

5. The water network reconstruction method based on forest-water functional integration according to claim 4 is characterized by: Step S3 specifically includes the following steps: S31. Construct a comprehensive circuit equation system including all resistors, capacitors and seepage resistances; S32, according to step S1 and step S2, the river section resistance, infiltration resistance and water storage area capacitance are set to initial values, and a comprehensive error function is constructed: Among them, obs represents the observed value, model represents the model predicted value, and specifically, Δh ij,model and Δh ij,obs They represent the predicted head difference and observed head difference between node i and node j, respectively. sink,i,model and Q sink,i,obs They represent the predicted infiltration flow and the actual infiltration flow of the river section from the i node to the next node, and It represents the predicted water volume change rate of the storage area and the actual water volume change rate of the storage area at the i-th node; S33. Use genetic algorithm or particle swarm optimization to adjust parameters and minimize Err.

6. The water network reconstruction method based on forest-water functional integration according to claim 5 is characterized by: Step S31 specifically includes the following steps: Using the simulated infiltration, runoff, confluence, and storage analysis in step S1, the equation group for each node is established respectively, and the water level V of each node in the water network system is obtained by solving the equation group. k The dynamic solution that changes with time is as follows: have: Where Δh iv It represents the water head difference from the i-th node to the k-th node, that is, the k-th node, Δh vj Represents the water head difference from the kth node to the jth node.

7. The water network reconstruction method based on forest-water functional integration according to claim 1 is characterized by: In step S4, the optimized water network model obtained in step S3 is used to enhance the water network regulation and storage capacity by adjusting the riverbed parameters. The specific process is: changing the river width W and depth D to affect the hydraulic radius R, using Monte Carlo or multi-scheme simulation to evaluate the water flow rate Q and water surface regulation capacity under different W and D.

8. The water network reconstruction method based on forest-water functional integration according to claim 7 is characterized by: In step S4, the specific formula of hydraulic radius R is: Among them, A is the cross-sectional area of ​​water flow, P is the wetted perimeter, W is the width of the river, and D is the depth of the river.

9. The water network reconstruction method based on forest-water functional integration according to claim 7 is characterized by: In step S4, the calculation formula of the water flow rate Q is: Where n is the Manning roughness coefficient, R is the hydraulic radius, and S is the slope of the channel or river, which is the ratio of the head difference to the horizontal distance.

10. The water network reconstruction method based on forest-water functional integration according to claim 7 is characterized by: In step S4, the river channel width W and depth D are changed, and the change amount W-W0≥1 meter, D-D0≤0.5 meter, where W0 and D0 are the original river bottom width and river channel depth.

11. The water network reconstruction method based on forest-water functional integration according to claim 1 is characterized by: In step S5, water forest habitat construction includes terrain transformation, vegetation configuration, monitoring and feedback; Terrain transformation: Based on the results of hydrological simulation, design terrain transformation plans, including building dams, digging and filling ditches, and optimizing surface runoff paths; Vegetation configuration: Select appropriate tree species and planting patterns according to the guidance of the composite evaluation system in step S2; Monitoring and feedback: Establish a long-term monitoring mechanism, evaluate eco-hydrological effects, and continuously adjust and optimize management measures.

12. The water network reconstruction method based on forest-water functional integration according to claim 10 is characterized by: In step S5, landscape creation specifically includes the following steps: S51. Divide the forest-water complex into three areas according to the height of the water level: long-term flooding area, short-term flooding area and bank slope area; S52. Plant adaptive tree species in different areas. Plant water-resistant tree species in long-term and short-term flooded areas. Select container seedlings with a breast diameter of 6.0-8.0 cm. Control the planting density at 2000-2500 plants / hm2. 2 ; Planting of moisture-loving and well-developed root-system tree species on the bank slope area, with a planting density of 3000-4000 trees / hm2 2 ; S53. Implement water level control measures, including gate control, to achieve dynamic management of flooded areas; S54. Establish a connection mechanism with the external water network.

Images