A mountainous area-based flow water ecological landscape water demand calculation method

Abstract

The application belongs to the field of data processing and water resources, and particularly relates to a water demand calculation method for a flowing water ecological landscape based on a mountainous area. The method comprises the following steps: identifying features such as karst landforms and ancient terraces through remote sensing data; constructing a nonlinear regulation and storage model of underground reservoirs to simulate the base flow lag release effect; calculating the landscape fragmentation and connectivity index, and establishing the mapping relationship between the spatial structure and the hydrological process; constructing an ecological water demand quota correction matrix based on nonlinear regression; and combining the base flow supply sequence and the dynamic landscape parameter to output the final water demand calculation result. The application realizes the deep coupling and dynamic correction of ecological water demand and spatial structure by introducing the nonlinear regulation and storage theory and the landscape pattern index, and improves the accuracy and scientificity of water demand calculation in a complex environment.

Description

[0001] This invention belongs to the fields of data processing and water resources, specifically relating to a method for calculating water demand in mountainous flowing ecological landscapes. Background Technology

[0002] With the deepening of ecohydrological research in mountainous areas, the calculation of water demand for flowing water ecological landscapes in mountainous regions has become a key technical support for ensuring regional ecological security and sustainable water resource utilization. The complex geographical environment and diverse ecosystem structures of mountainous areas determine the spatiotemporal variability of their water cycle processes. Traditional ecological water allocation schemes typically rely on macroscopic hydraulic indicators to maintain basic ecological functions. In current ecological restoration practices, accurately identifying and calculating the water demand characteristics of flowing water landscapes under different seasons and geological backgrounds has significant theoretical and practical value for optimizing mountain landscape connectivity and protecting biodiversity.

[0003] Research on groundwater retreat mechanisms and landscape pattern evolution under complex geological backgrounds such as karst landforms and ancient terraces is a core approach to achieving accurate calculation of water demand in mountainous areas. This research simulates the dynamic regulation and storage process of water sources by constructing underground reservoir models and combines them with landscape ecological indices to characterize the connectivity of surface ecological spaces. Its fundamental principle lies in analyzing the nonlinear lag effect of geological structures on runoff and, by combining land use data obtained from remote sensing inversion, establishing a correlation matrix between landscape fragmentation and ecological water demand quotas, providing a quantitative basis for the scientific scheduling of ecological flow in mountainous areas.

[0004] Traditional computational models based on linear reservoir theory struggle to accurately depict the nonlinear regression characteristics of fissure karst water or special geological bodies under gravity, limiting the accuracy of simulations of groundwater delayed release effects. Existing methods generally lack quantitative assessments of the impact of surface landscape patterns on ecological water demand and fail to integrate the moderating effect of landscape connectivity on hydrological processes, easily leading to a systematic underestimation of ecological water demand during dry seasons. Single computational logic cannot adapt to the dynamic fusion requirements of multi-source heterogeneous data in mountainous areas, resulting in discrepancies between the calculated results and actual eco-hydrological evolution patterns, failing to meet the practical needs of high-precision landscape ecological protection. Summary of the Invention

[0005] The purpose of this invention is to provide a method for calculating the water demand of a flowing ecological landscape in mountainous areas, which can solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for calculating water demand in a mountainous, flowing ecological landscape includes the following specific steps:

[0008] Step 1: Acquire basic geographic information data and multi-source remote sensing image data of mountainous areas. By classifying the multi-source remote sensing image data, identify the land use types in the study area and extract the distribution characteristics of karst landforms, ancient terraced fields and conventional mountain landscapes.

[0009] Step 2: Construct a nonlinear regulation and storage model for underground reservoirs in mountainous areas. Use the principles of groundwater dynamics to describe the nonlinear relationship between underground water storage and discharge flow, and simulate the nonlinear retreat process of karst water in mountainous fissures or special geological bodies under gravity and the delayed release effect on the base flow.

[0010] Step 3: Calculate the water ecological landscape pattern index. Using geographic information system analysis tools, quantitatively characterize the landscape fragmentation index and landscape connectivity index within the study area, and establish the mapping relationship between the surface landscape spatial structure and eco-hydrological processes.

[0011] Step 4: Establish an ecological water demand quota correction matrix, using the landscape fragmentation index and the landscape connectivity index as constraint variables, and construct a functional relationship between the ecological water demand quota and the landscape pattern index through nonlinear regression analysis to determine the water demand correction coefficient under different landscape patterns.

[0012] Step 5: Encapsulate and run the water demand correction module for the flowing water ecological landscape. Input the base flow supply sequence calculated by the nonlinear regulation and storage model of the underground reservoir in the mountainous area into the correction module, and perform correction processing in combination with the dynamic parameters of the landscape pattern retrieved by real-time remote sensing, and output the final calculation result of the water demand of the flowing water ecological landscape in the mountainous area.

[0013] Preferably, step 1, acquiring multi-source remote sensing image data, includes acquiring satellite remote sensing multispectral data and radar mapping data at a preset spatial resolution. When performing land cover classification, a combination of supervised classification and object-oriented classification is used to extract vegetation cover, water body boundaries, and the extent of artificial structures in mountainous areas. For ancient terraced fields, the spatial morphology of terraced field ridges and water storage surfaces is identified through a combination of elevation gradient and slope variability logic. For karst landforms, surface subsidence characteristics and exposed rock rates are used for identification. All remote sensing data undergoes radiometric calibration, atmospheric correction, and geometric fine correction based on a digital elevation model to eliminate the distortion effects of terrain undulations on remote sensing information.

[0014] Preferably, the core logic of the nonlinear regulation model for underground reservoirs in mountainous areas constructed in step 2 lies in analyzing the power function relationship between groundwater storage and outflow velocity. This model defines the discharge flow rate as the product of a specific regulation coefficient and a specific exponent of the effective underground water storage. This specific exponent is pre-set based on the pore structure characteristics and fracture development density of the mountainous geological body, reflecting the resistance changes between unsaturated and saturated zones during gravity discharge. When the water storage is high, the model simulates a high discharge rate to reflect the rapid response of gravity flow in mountainous areas; when the water storage decreases to a preset threshold, the decay of the discharge rate exhibits a nonlinear, slow trend, accurately characterizing the strong retention and delayed release mechanism of water by karst fractures or terraced soils.

[0015] Preferably, the process of calculating the landscape ecological index in step 3 includes the quantification of landscape fragmentation and landscape connectivity. Landscape fragmentation is defined as the ratio of the total number of patches of a specific type to the total area of ​​that type of landscape, reflecting the degree of ecological spatial segmentation in mountainous areas caused by human activities or natural evolution. The landscape connectivity index is based on graph theory, defining different aquatic habitat patches as nodes and ecological corridors or diffusion paths between patches as edges. The ease of water flow transmission between different patches is assessed by calculating the probabilistic connectivity index. The calculation of this index involves a comprehensive calculation of the geometric distance between patches, habitat resistance coefficients, and the area weight of each patch. All computational logic is converted into descriptive text logic and does not involve algebraic symbols.

[0016] Preferably, in step 4, when establishing the ecological water demand quota correction matrix, a multi-factor regression analysis method is used. Landscape fragmentation is used as a negative correction factor, and landscape connectivity is used as a positive correction factor. Specifically, when an increase in landscape fragmentation is detected within the study area, the minimum ecological flow required for the ecosystem to maintain its original functions will change due to the enhanced habitat edge effect. The correction module will automatically increase the ecological water demand guarantee coefficient to offset the risk of water dissipation caused by landscape fragmentation. Conversely, when landscape connectivity is at a high level, the efficiency of water circulation between patches increases, and the correction module will appropriately optimize the water demand quota based on a preset correlation matrix. The range of correction coefficient values ​​is determined by fitting historical ecological observation data using the least squares method.

[0017] Preferably, the process of encapsulating and running the water demand correction module for the flowing water ecological landscape in step 5 involves dynamic processing of time series data. This module calls the groundwater receding curve generated in step 2 to determine the baseflow contribution within a specific time step. The module also receives the landscape pattern change parameters extracted in step 3 in real time. If the landscape connectivity within the current time step is lower than a preset ecological safety threshold, the correction module will trigger an early warning compensation mechanism, increasing the calculation weight of ecological water allocation to enhance the connectivity between landscape patches. The final output water demand result is a dynamic value after dual correction based on geological regulation characteristics and landscape spatial structure, reflecting the real demand for water use in complex mountainous environments.

[0018] Preferably, the basic geographic information data for the mountainous area also includes a high-precision digital elevation model (DEM). The DEM is used to extract runoff accumulation, slope aspect, slope length, and topographic moisture index for the mountainous area. The topographic moisture index is obtained by calculating the natural logarithm of the ratio of the tangent of the runoff area to the tangent of the slope, and is used to characterize the potential distribution characteristics of surface moisture in the mountainous area. During the calculation process, the topographic moisture index and landscape pattern index are spatially overlaid to identify the most sensitive ecological key nodes in the mountainous area to water response.

[0019] Preferably, the simulation of fracture karst water in step 2 further includes characterizing the dual-porosity media effect. The geological body is divided into a bedrock block system and a fracture network system, with water exchange between the two driven by pressure gradients. When water in the fracture network rapidly drains, water in the bedrock block is slowly replenished to the fractures due to matrix suction. This exchange logic is integrated into a nonlinear regulation model, which simulates the outflow process under different geological structures by setting different permeability parameters.

[0020] Preferably, the calculation of the landscape fragmentation index also involves assessing the complexity of patch shape. Shape complexity is obtained by calculating the ratio of the patch perimeter to the circumference of a circle of equal area. The more complex the patch shape, the larger its contact interface with the external environment, and the higher the risk of water evaporation and loss; therefore, it will be assigned a high water demand weight in the correction matrix.

[0021] Preferably, the calculation of the landscape connectivity index also incorporates a habitat suitability evaluation factor. Based on the distribution patterns of key protected species within the study area, different migration resistance values ​​are assigned to different land use types. Resistance values ​​for forests and water bodies are set at a lower level, while resistance values ​​for built-up land and wasteland are set at a higher level. Ecological connectivity between patches is assessed by calculating the minimum cost path, improving the relevance of water demand calculations to biodiversity conservation.

[0022] Preferably, the nonlinear regression analysis in step 4 also includes sensitivity correction for rainfall pulse effects. By analyzing the hysteretic correlation between historical rainfall sequences and baseflow changes, the time constant for rainfall infiltration into underground reservoirs and its conversion into effective discharge flow is determined. This time constant, as an important parameter of the nonlinear regulation and storage model, ensures that the water demand calculation model maintains high stability and accuracy under extreme drought seasons or continuous rainfall conditions.

[0023] Preferably, the correction module is encapsulated using a layered architecture design. The bottom layer is the data access layer, responsible for reading remote sensing images and hydrological monitoring data; the middle layer is the logic operation layer, responsible for performing nonlinear water storage simulation and landscape index calculation; and the top layer is the result output layer, responsible for generating a visualized spatiotemporal distribution map of water demand and a decision recommendation report.

[0024] Preferably, the calculation method further includes a verification step for the calculation results. Automatic flow monitoring stations are set up at typical cross-sections of mountainous water landscapes to obtain measured flow data. The measured flow is compared with the ecological water demand calculated by the method of this invention, and the residuals and correlation coefficients are calculated. If the correlation coefficient is lower than a preset accuracy threshold, an adaptive parameter adjustment mechanism is activated, and the calculation model is iteratively optimized by changing the exponential power of the nonlinear water storage model or the weight ratio of the landscape correction matrix.

[0025] Preferably, when dealing with ancient terraced areas, the method specifically considers the artificial wetland effect within the terraces. Due to the unique stepped structure of terraces, their ability to intercept surface runoff is higher than that of ordinary slopes. By introducing the terrace storage volume parameter, the model simulates the gradual cascading and infiltration process of water between different levels of terraces, realistically reflecting the landscape water demand characteristics of this special artificial ecosystem.

[0026] Preferably, considering the seasonal differences in the mountainous water ecological landscape, the method sets three different calculation modes: dry season, wet season, and normal season. During the dry season, the model focuses on the supporting role of nonlinear regulation on the baseflow; during the wet season, it emphasizes assessing the contribution of landscape connectivity to flood peak reduction. By switching different weighting coefficients, all-weather, full-cycle mountainous eco-hydrological monitoring and water demand management can be achieved.

[0027] Preferably, the calculation of the landscape connectivity index also involves assessing the importance of ecological patches. By systematically removing specific patches from the landscape and observing the degree of decline in overall landscape connectivity, the core value of each patch in the ecological network is determined. For patches with high core value, the correction module will allocate more ecological water to ensure the integrity of their ecological functions.

[0028] Preferably, the method further includes predicting future landscape evolution trends in mountainous areas. By inputting a preset land use planning scheme, the model can simulate the impact of changes in landscape pattern on ecological water demand. This provides a scientific quantitative prediction tool for balancing development and ecological protection in mountainous areas, and can avoid the risk of ecological water shortage caused by over-development of the landscape.

[0029] Preferably, the accuracy of the land cover classification in step 1 is required to reach a preset percentage threshold. After classification, a confusion matrix is ​​constructed by randomly selecting verification points and combining them with field survey data, and the classification accuracy index is calculated. If the classification accuracy does not meet the requirements, the classification results are optimized by increasing the number of training samples or improving the classification algorithm logic.

[0030] Preferably, the calculation results output in step 5 also include a breakdown of water requirements for different ecosystem service functions. Specifically, this includes the minimum environmental water required to maintain the survival of aquatic organisms, the landscape water required to maintain the visual aesthetic value of the landscape, and the dilution and purification water required to maintain the self-purification capacity of the river. Based on the specific functional positioning of the mountain ecological landscape, each water quantity is weighted and summed in the correction module using different weighting functions to obtain the final total water requirement.

[0031] Preferably, the regulation coefficient in the nonlinear regulation model of the underground reservoir in the mountainous area is obtained by comprehensive interpolation of the lithology, soil thickness, and slope of the mountainous strata. In areas with high rock exposure, the regulation coefficient is set to a smaller value, reflecting its weaker water-holding capacity; in forest and ancient terraced areas, the regulation coefficient is set to a larger value, reflecting its stronger water storage and buffering function.

[0032] Preferably, the method involves the storage and concurrent processing of massive amounts of spatiotemporal data during operation. A distributed database system is used to store remote sensing image slices and hydrological time series, and a parallel computing framework is used to accelerate the extraction of landscape indices and the numerical solution of nonlinear equations, ensuring the real-time performance and efficiency of water demand calculation in large-scale mountainous areas.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1. Improved the accuracy of groundwater receding simulation in mountainous areas. By introducing nonlinear regulation theory, the limitations of traditional linear reservoir models are overcome, enabling a realistic depiction of the delayed release effect of water under complex geological backgrounds such as karst and ancient terraces, thus solving the problem of inaccurate baseflow estimation during the dry season.

[0035] 2. This invention achieves deep coupling between ecological water demand calculation and landscape pattern. By quantitatively introducing landscape fragmentation and connectivity indices, it can identify the feedback mechanism of surface landscape heterogeneity on hydrological processes, making the water demand calculation results dependent on climatic conditions and the spatial structure of mountain landscapes.

[0036] 3. A dynamic ecological water demand correction mechanism is provided. The correction module encapsulated in this invention can automatically adjust the water demand quota and correction coefficient based on real-time landscape evolution data retrieved by remote sensing, thus addressing the uncertain impact of land use changes in mountainous areas on ecological water allocation.

[0037] 4. Enhanced the scientific rigor and predictability of ecological protection in mountainous areas. By simulating water demand characteristics under different geological and landscape conditions, this invention provides highly accurate quantitative data for the ecological restoration of mountain stream landscapes, the protection of terraced cultural heritage, and water resource allocation in karst regions, demonstrating its engineering application value.

[0038] 5. Excellent adaptability and scalability. The method logic of this invention is not only applicable to specific mountainous terrains, but can also be adaptively adjusted according to different geographical parameters, providing a standardized computational framework for eco-hydrological research in multi-source heterogeneous data environments. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the overall technical solution architecture according to the present invention;

[0040] Figure 2 This is a schematic diagram of the core principle framework of the present invention based on the coupling correction of nonlinear regulation and landscape pattern;

[0041] Figure 3 A logical flow diagram illustrating the nonlinear regulation and storage model of underground reservoirs in Zhongshan District based on the present invention;

[0042] Figure 4 This is a flowchart illustrating the logical process of calculating and mapping the water ecological landscape pattern index according to the present invention.

[0043] Figure 5 This is a schematic diagram of the multi-level interaction relationship and data flow of the water demand correction module according to the present invention. Detailed Implementation

[0044] Example 1: Please refer to the appendix Figure 1 To be continued Figure 5 In the specific implementation of a method for calculating water demand in a mountainous ecological landscape, step 1 involves acquiring basic geographic information data and multi-source remote sensing image data of the mountainous area. The multi-source remote sensing image data includes satellite remote sensing multispectral data and radar mapping data with a preset spatial resolution. In this embodiment, the satellite remote sensing multispectral data uses panchromatic and multispectral band fusion data with a spatial resolution better than 10 meters to provide rich spectral features of ground features. The radar mapping data uses synthetic aperture radar data to penetrate cloud and fog interference and obtain microscopic features of terrain undulations.

[0045] After acquiring the raw data, land use types within the study area are identified by performing land cover classification processing on the multi-source remote sensing image data. A combination of supervised classification and object-oriented classification is used to divide the study area into forest land, grassland, water area, construction land, cultivated land, and bare land. In object-oriented classification, the system aggregates pixels with similar spectral and textural features into objects by setting scale, shape, and compactness parameters.

[0046] During the identification process, this invention focuses on extracting the distribution characteristics of karst landforms, ancient terraced fields, and conventional mountain landscapes. For ancient terraced fields, the spatial morphology of terraced field ridges and water storage surfaces is identified through a combination of elevation gradient and slope variability logic. Specifically, a digital elevation model is used to calculate the second derivative of the slope, identifying ridges with abrupt slope changes as candidate ridge lines. This is then combined with water or vegetation indices from spectral features to determine the water storage or vegetation cover status within the ridge enclosure. For karst landforms, surface subsidence characteristics and exposed rock rates are used for identification.

[0047] The system determines the degree of karst development by identifying closed depressions in the digital elevation model (DEM) and combining this with the proportion of exposed rock area in multispectral data. All remote sensing data undergoes radiometric calibration, atmospheric correction, and geometric fine correction based on the DEM before classification to eliminate the influence of topographic relief on the geometric distortion of remote sensing information. The accuracy of land cover classification is required to reach a preset percentage threshold, such as above 85%. After initial classification, a confusion matrix is ​​constructed by randomly selecting validation points and combining them with field survey data, and classification accuracy indices such as the Kappa coefficient are calculated. If the classification accuracy does not meet the requirements, the classification results are optimized by increasing the representativeness of the training samples or improving the feature vector space of the classification algorithm.

[0048] The basic geographic information data for the mountainous area also includes a high-precision digital elevation model (DEM), with a resolution matching that of the remote sensing imagery. The DEM is used to extract runoff accumulation, slope aspect, slope length, and topographic moisture index for the mountainous area. The topographic moisture index is obtained by calculating the natural logarithm of the ratio of the tangent of the runoff area to the tangent of the slope. This index characterizes the potential distribution characteristics of surface moisture in the mountainous area, reflecting the gravitational accumulation trend of moisture under topographic influence. During the calculation process, the topographic moisture index is spatially overlaid with the landscape pattern index to identify the most water-sensitive ecological key nodes in the mountainous area, such as the core patches at valley confluences or terraced catchment areas.

[0049] Step 2: Construct a nonlinear regulation model for underground reservoirs in mountainous areas. This model utilizes groundwater dynamics principles to describe the nonlinear relationship between groundwater storage and discharge flow. The core logic of the model lies in analyzing the power function relationship between groundwater storage and outflow velocity. The model defines the discharge flow as the product of a specific regulation coefficient and a specific exponent of the effective groundwater storage. This specific exponent is not a fixed constant but a parameter pre-set based on the pore structure characteristics and fracture development density of the mountainous geological body. In karst regions with extremely high fracture development, this exponent is typically set between 1.5 and 2.5 to reflect the nonlinear evolution of resistance between unsaturated and saturated zones during gravity discharge. When the storage volume is high, the groundwater level is high, the hydraulic gradient is large, and the model simulates a high discharge rate to reflect the rapid response of gravity flow in mountainous areas. When the storage volume decreases to a preset threshold, groundwater is mainly released through micro-fractures or bedrock pores, and the decay of the discharge rate exhibits a nonlinear and slow trend. This mechanism precisely describes the strong retention and delayed release of water by karst fissures or terraced soil.

[0050] In one specific embodiment, the construction and operation of the nonlinear regulation and storage model of the underground reservoir in the mountainous area adopts the discrete time step method. Specifically, this model is a first-order nonlinear dynamic system, the core of which lies in defining the state variables of the underground reservoir. And its equations of evolution over time.

[0051] Input data includes: the initial effective groundwater storage. Time step (For example, set to 1 day), and the set of model parameters determined based on geological exploration and experiments in the study area, including the regulation coefficient. and nonlinear exponent .

[0052] The internal flow logic of the model is as follows: at time... The system determines the current water storage level. Calculate the discharge flow rate during this period. This flow rate, on the one hand, is output to step 5 as part of the baseflow supply sequence, and on the other hand, it leads to a reduction in the reservoir water volume. Subsequently, the system superimposes the recharge from rainfall infiltration. (This can be estimated from the remote sensing and hydrological data in step 1), and the water storage capacity at the next moment can be calculated. This allows for cyclic simulation. The process accurately characterizes the nonlinear decline and dynamic replenishment of water in the reservoir. Discharge flow rate. The calculation formula is:

[0053] ;

[0054] in, Indicates time Discharge flow of underground reservoirs (unit: ); This represents a specific regulation and storage coefficient, whose value ranges from 1 to 2. The relationship between these parameters is positively correlated with the water-conducting capacity of the geological body. Indicates time Effective underground water storage (unit: ); The nonlinear exponent is a dimensionless parameter greater than 1. The reservoir state update equation is: ;

[0055] in, Indicates the effective underground water storage capacity at the next moment; Indicates the time step of the simulation; Indicates in The amount of rainfall infiltrating into the reservoir during the specified period.

[0056] In step 2, the simulation of fissure karst water also includes characterizing the dual-porosity media effect. The geological body is logically divided into a bedrock block system and a fissure network system. A pressure gradient-driven water exchange exists between the two systems, proportional to the hydraulic conductivity between the bedrock blocks and the fissures. When water in the fissure network is rapidly drained by gravity after rainfall, water in the bedrock blocks is constrained by matrix suction, slowly replenishing horizontally to the fissures at a lower rate. This exchange logic is integrated into a nonlinear storage model, simulating the outflow process under different geological structures (such as anticlines, synclines, or fault zones) by setting different permeability parameters and exchange coefficients. The storage coefficient itself is obtained through comprehensive interpolation of the lithology, soil thickness, and slope of the mountainous strata. In areas with high rock exposure and thin soil layers, the water retention coefficient is set to a smaller value to reflect its weak water holding capacity; in forest areas and areas covered by ancient terraces, the water retention coefficient is set to a larger value due to the water retention effect of vegetation roots and artificial terrace structures.

[0057] Step 3: Calculate the watercourse ecological landscape pattern index. This step uses Geographic Information System (GIS) analysis tools to quantitatively characterize the landscape fragmentation index and landscape connectivity index within the study area. Landscape fragmentation is defined as the quotient obtained by dividing the total number of patches of a specific type in the landscape by the total area of ​​that type of landscape. A higher value reflects a more severe degree of ecological spatial fragmentation in mountainous areas caused by human activities or natural evolution. The calculation of landscape fragmentation also involves assessing the complexity of patch shapes. Shape complexity is obtained by calculating the ratio of the patch perimeter to the circumference of a circle of equal area. The more complex the patch shape, the larger its contact interface with the external environment, and the higher the risk of water loss through transpiration and surface runoff.

[0058] The landscape connectivity index is constructed based on graph theory principles. In this logic, different aquatic habitat patches (such as stream sections, ponds, and terraced wetlands) are defined as nodes, and ecological corridors, seasonal overflow paths, or diffusion paths between patches are defined as edges. The ease of water flow and its carried ecological elements transport between different patches is assessed by calculating the probabilistic connectivity index. The calculation of the probabilistic connectivity index involves a comprehensive calculation of the negative exponential function of the geometric distance between patches, the habitat resistance coefficient, and the area weight of the patch itself. A habitat suitability evaluation factor is also introduced in this process. Based on the distribution patterns of key protected species in the study area, different migration resistance values ​​are assigned to different land use types. Specifically, the calculation of the probabilistic connectivity index adopts a minimum cost path model based on landscape resistance surfaces. First, the study area is rasterized, with each raster unit... A habitat migration resistance value is assigned based on its land use type. For example, natural forest and water area grids. Set to 1, for the construction land grid. Set to 10 to form the resistance surface of the study area. The contribution of individual landscape patch pairs to connectivity. The calculation formula is:

[0059] ;

[0060] Overall landscape probability connectivity index The sum of contributions from all plaques:

[0061] ;

[0062] in, Indicates plaque With plaque The cumulative resistance distance of the least cost path between them; This represents the distance attenuation coefficient, which is positive and is used to control the rate at which connectivity decreases with increasing resistance distance. , They represent plaques. and plaque The area; This represents the total landscape area of ​​the study region; This indicates the total number of landscape patches.

[0063] For any two landscape patches and The probability of effective connections between them The following comprehensive calculations were performed: First, the plaques on the resistance surface were calculated. From centroid to plaque The path of minimum cumulative resistance at the center of mass, whose resistance distance is denoted as . Then, the patch area weighting is combined. and The connectivity contribution between individual patch pairs is calculated. Finally, the summation is performed over all patch pairs to obtain the overall landscape connectivity index. This index quantifies the overall connectivity level of landscape ecological networks under specific resistance conditions.

[0064] For example, resistance values ​​for natural forests and permanent water bodies were set at a lower level, while resistance values ​​for built-up land, intensively developed farmland, and wasteland were set at a higher level. The connectivity efficiency of water and biological transport was assessed by calculating the minimum cost paths between patches. The assessment also included evaluating the importance of ecological patches by systematically removing specific patches from the landscape and observing the decrease in overall landscape connectivity values ​​to determine the core value of each patch within the entire mountainous water system ecological network, providing a priority reference for subsequent water allocation.

[0065] Step 4: Establishing the Ecological Water Demand Correction Matrix. This step uses the landscape fragmentation index and landscape connectivity index as constraint variables, and constructs a functional relationship between the ecological water demand quota and the landscape pattern index through multi-factor nonlinear regression analysis. Landscape fragmentation is used as a negative correction factor, while landscape connectivity is used as a positive correction factor. The core logic is: when increased landscape fragmentation is detected in the study area, the minimum ecological flow required for the ecosystem to maintain its original ecological functions (such as biodiversity maintenance and microclimate regulation) will change due to enhanced edge effects. The correction module will automatically increase the ecological water demand guarantee coefficient according to the magnitude of the increase in fragmentation, that is, multiply the basic water demand quota by a correction ratio greater than 1 to offset the risk of water dissipation caused by landscape fragmentation. Conversely, when landscape connectivity is at a high level, the efficiency of water circulation between patches is improved, and the system function is more robust. The correction module will then appropriately adjust and optimize the water demand quota according to the preset correlation matrix, that is, multiply by a correction ratio less than 1 but greater than the preset safety threshold.

[0066] The range of correction coefficients is determined by fitting historical ecological observation data using the least squares method, ensuring that the corrected water demand remains highly correlated with the actual ecosystem health. In step 4, the nonlinear regression analysis also includes sensitivity correction for rainfall pulse effects. The system determines the time constant for rainfall infiltration into underground reservoirs and its eventual conversion into effective discharge flow by analyzing the lag correlation between historical long-series rainfall data and baseflow changes. This time constant, as a crucial parameter of the nonlinear regulation model, ensures the water demand calculation model maintains high stability under extreme drought conditions or continuous heavy rainfall, avoiding unreasonable jumps in water demand calculation results due to short-term rainfall fluctuations. As a preferred implementation, the nonlinear regression analysis employs a multivariate nonlinear regression model to establish the water demand correction coefficients. The functional relationship between the landscape pattern index and the model. The input variable of the model is the landscape fragmentation index calculated in step 3. Landscape connectivity index The output variable is the water demand correction coefficient. Water demand correction factor The formula for the nonlinear regression model with the landscape pattern index is as follows:

[0067] ;

[0068] in, Indicates the landscape fragmentation index; Indicates the landscape connectivity index; , , , These are undetermined parameters, determined by fitting historical data using the nonlinear least squares method. A positive value indicates that increased fragmentation leads to... Increase; parameter A positive value indicates that increased connectivity leads to... Decrease; As a base correction value, ensure The basic rationality.

[0069] Specifically, an exponential function was chosen as the basic form of the regression model to reflect the nonlinear response relationship between the correction coefficient and the exponent. The model was determined by fitting long-term historical observation data. The historical dataset includes landscape pattern indices from different years. , And the corresponding actual water demand correction coefficient determined through ecological experiments or historical hydrological and ecological surveys. The nonlinear least squares method is used as the optimization algorithm. By iteratively adjusting the model parameters, the sum of squared residuals between the model's predicted and observed values ​​is minimized, thereby determining the final functional relationship and parameters. This function clearly defines how an increase in fragmentation and a decrease in connectivity will quantitatively increase the water demand guarantee coefficient.

[0070] Step 5: Encapsulate and run the water demand correction module for the flowing water ecological landscape. This module adopts a layered architecture design to ensure the decoupling of data flow processing and functions. The bottom layer is the data access layer, responsible for reading remote sensing image interpretation results, real-time monitoring data from hydrological stations, and digital terrain parameters. The middle layer is the logic operation layer, responsible for sequentially executing nonlinear regulation simulation and real-time calculation of landscape pattern indices. The top layer is the result output layer, responsible for generating a visualized spatiotemporal distribution map of water demand and a decision recommendation report. During the module's operation, dynamic processing of time-series data is involved. The module calls the groundwater receding curve generated in Step 2 to determine the baseflow contribution within the current specific time step. The module also receives the latest extracted landscape pattern change parameters from Step 3 in real time.

[0071] If the landscape connectivity within the current time step falls below a preset ecological safety threshold, the correction module will automatically trigger an early warning compensation mechanism. This mechanism prioritizes meeting the water needs of ecological patches with higher core value by increasing the calculation weight of ecological water allocation, thereby forcibly enhancing the connectivity between landscape patches. The final output calculation results also include a breakdown of water demand for different ecosystem services. Specifically, this includes: the minimum environmental water volume required to maintain the survival of aquatic organisms, determined based on the flow rate and water depth requirements of suitable habitats; the landscape water volume required to maintain the visual aesthetic value of the landscape, mainly targeting the maintenance of the mirror effect of waterfalls, cascades, or terraces; and the dilution and purification water volume required to maintain the self-purification capacity of river channels. Based on the specific functional positioning of the mountain ecological landscape (such as scenic spots, ecological protection red line areas, etc.), each water volume is weighted and summed in the correction module using different weighting functions to obtain the final dynamic value of total water demand.

[0072] In practical operation, this method also includes a verification step for the calculation results. Automatic flow monitoring stations are set up at typical cross-sections of mountainous water landscapes to obtain measured flow data. The measured flow is compared with the ecological water demand calculated by the method of this invention, and the residuals, root mean square error, and correlation coefficients are calculated. If the correlation coefficient is lower than 0.8 or the residual exceeds a preset allowable range, a parameter adaptive adjustment mechanism is activated. This mechanism iteratively optimizes the calculation model by changing the exponential parameters in the nonlinear water storage model or adjusting the weight ratio of fragmentation and connectivity in the landscape correction matrix until the goodness of fit between the calculated results and the measured data meets the standard.

[0073] When dealing with specific ancient terraced field areas, the method specifically considers the artificial wetland effect within the terraces. Due to their unique stepped artificial structure, terraces have a higher capacity to intercept surface runoff and promote water infiltration than ordinary natural slopes. The model simulates the gradual cascading, storage, and infiltration of water between different terrace levels by introducing terrace storage volume parameters. Each terrace level is considered a micro-reservoir, and its overflow conditions depend on the height of the ridge. This refined simulation allows the calculation results to realistically reflect the landscape water demand characteristics of this unique artificial ecosystem of ancient terraced fields.

[0074] To address the seasonal variations in mountain streamline ecological landscapes, the proposed method employs three distinct calculation modes: dry season, wet season, and normal season. During the dry season, when rainfall is scarce, the model automatically switches to a focus on the supporting role of nonlinear regulation in baseflow, increasing the weight of baseflow contribution to ensure the riverbed does not dry up. During the wet season, with abundant rainfall, the model emphasizes assessing the contribution of landscape connectivity to flood peak reduction and the self-purification needs of the water body. By switching different weighting coefficients, all-weather, full-cycle mountain ecological hydrological monitoring and water demand management are achieved.

[0075] The method in this embodiment also includes the prediction of future landscape evolution trends in mountainous areas. By inputting a preset land use planning scheme (such as the area of ​​new afforestation or the scope of urbanization expansion), the model can predict the long-term impact of changes in landscape pattern on ecological water demand. This provides a scientific quantitative prediction tool for balancing development and ecological protection in mountainous areas. During operation, it involves the storage and concurrent processing of massive amounts of spatiotemporal data. A distributed database system is used to store multi-period remote sensing image slices and hydrological time series data. A parallel computing framework is used to accelerate the extraction of landscape indices and the numerical solution of nonlinear equations, ensuring good real-time performance of water demand calculations even in large-scale mountainous areas.

[0076] Example 2: Building upon Example 1, Example 2 further elaborates on the application of a method for calculating the water demand of a mountainous water ecosystem in a complex karst cone region. In this scenario, the surface landscape pattern is fragmented, and the groundwater hydrological processes exhibit spatial heterogeneity.

[0077] In step 1, for karst funnel areas, the feature classification process particularly emphasized the identification of funnel centers, sinkholes, and exposed rock clusters. Utilizing high-precision micro-topographic features provided by radar mapping data, the system accurately locates the spatial coordinates of sinkholes by calculating the contrast in local topographic closure. These sinkholes are considered channels through which surface runoff flows into underground rivers. In multi-source remote sensing image processing, a combination of infrared and visible light bands is used to calculate the thermal infrared characteristics of exposed rocks, aiding in the identification of karst areas with extremely high exposure rates.

[0078] When constructing the nonlinear regulation and storage model in step 2, a dual-channel retreat logic for both fast and slow flows was introduced. The discharge flow rate depends not only on the power function relationship of the overall water storage volume but also on the pulsed discharge term controlled by the sinkhole. For the slow flow component, the regulation and storage coefficient was set to a dynamic value related to the lithological karstification degree. For the fast flow component, it was activated by identifying a threshold for rainfall intensity. When the amount of a single rainfall event exceeds the preset threshold, the nonlinear exponent of the underground reservoir increases rapidly within a short period, simulating the high-speed response of the karst conduit system to rainfall.

[0079] When calculating the landscape pattern index in step 3, the landscape fragmentation calculation is given higher weight due to the unique characteristics of karst areas. Because the funnel-shaped topography causes surface vegetation patches to be fragmented by rocks, the system calculates the edge length density of the rock-soil complex as a supplementary indicator of fragmentation. The landscape connectivity calculation focuses on the potential connecting effect of underground river networks on surface patches. If two surface patches are located beneath the same underground river tributary, an additional positive correction coefficient is assigned when calculating the probabilistic connectivity index, reflecting the characteristic that although the surface landscape is discontinuous, it possesses hydrological connectivity.

[0080] When establishing the correction matrix in step 4, the system employed a geographically weighted regression-based analysis method. Since the response of ecological water demand to landscape changes in karst areas is spatially uneven, the regression coefficients were set as functions of coordinates. This means that in the core area of ​​the funnel, the increase in water demand due to increased landscape fragmentation is greater than in the peak-cluster ridge area. The correction matrix also includes a drought vulnerability variable, obtained by calculating the decay rate of vegetation cover during historical dry seasons, used to enhance ecological water demand protection in drought-sensitive areas.

[0081] During step 5, the correction module accessed real-time groundwater level monitoring well data. The module's output calculations now include a component for calculating the maintenance capacity of underground river ecosystems. This is because in karst mountain areas, the biodiversity of underground rivers is also an important component of the flowing water ecological landscape. The final output water demand not only covers surface flowing water landscapes but also simultaneously provides the minimum replenishment required to maintain underground river water levels above a safe threshold.

[0082] In the verification phase, this embodiment incorporates tracer test data. By adding tracers at known inflow points, the time distribution curves of their arrival at typical cross-sections are measured. The actual storage parameters obtained by fitting these curves are used to perform secondary calibration of the nonlinear storage model's parameters. This verification mechanism based on physical experiments improves the reliability of calculation results under extremely complex terrain conditions.

[0083] Example 3: This example describes the application of the method in large-scale terraced cultural heritage protection areas (such as large-scale terraced fields spanning multiple elevation gradients), focusing on the vertical differentiation characteristics of ecological water demand caused by elevation differences.

[0084] In step 1, when acquiring multi-source remote sensing imagery, an orthophoto system mounted on a drone was specifically used to obtain ground feature data with centimeter-level resolution. This enabled the system to accurately extract the width of the ridges, water depth, and internal vegetation type of each terrace. When classifying ground features, the system not only identifies land use types but also the maintenance status of the terraces, such as cultivated terraces, seasonally abandoned terraces, and permanently abandoned terraces. Terraces in different maintenance states exhibit differences in water retention capacity.

[0085] In step 2, the constructed nonlinear water storage model is refined into a series structure of multiple altitude levels. Since water flows from high-altitude terraces to lower-altitude terraces via infiltration and overflow, the effective water storage capacity of each level is affected not only by its own infiltration but also by lateral inflow from the upper level. A potential energy correction factor reflecting the altitude difference is added to the power function model of the discharge flow. In low-altitude areas, due to the accumulated delayed release flow from the upper terraces, the storage coefficient is set as a function that increases with decreasing altitude.

[0086] Step 3, when calculating the landscape pattern index, incorporates the vertical connectivity index. This index is calculated by assessing the slope continuity between patches at different elevations and the integrity of artificial irrigation and drainage systems. Landscape connectivity decreases if the irrigation and drainage system is disrupted. The calculation of landscape fragmentation includes an analysis of field fragmentation, assessing the extent to which a single continuous cultivated surface is broken into smaller fragments due to terrace collapse.

[0087] Step 4, in establishing the correction matrix, incorporates a meteorological gradient correction for temperature increases with decreasing altitude. Due to the high evaporation rate in low-altitude areas, the baseline value for ecological water demand is set higher than that in high-altitude areas. The correction matrix couples landscape connectivity with irrigation efficiency indicators. For areas with well-preserved irrigation canal systems and high vertical connectivity, the system assigns a higher water demand weight to maintain their unique cultural landscape value; for fragmented abandoned farmland, the focus is on calculating the minimum ecological restoration water volume required for natural vegetation succession.

[0088] During step 5, the correction module dynamically adjusts the flow distribution instructions between terraces by accessing real-time data from multi-level distributed water level gauges. The calculation results output layer includes a terrace landscape stability index, which is obtained by comparing the current ecological water supply with the minimum moisture content required to maintain the stability of the ridge structure. If the calculated water supply is insufficient to maintain the moisture content of the ridge soil, the module issues an alarm, indicating a potential risk of terrace landslides.

[0089] In the verification step of this embodiment, vegetation canopy transpiration rate data monitored by thermal infrared remote sensing were utilized. The actual water consumption was calculated using an energy balance equation, and this calculation served as an upper limit constraint on the water demand calculated by this method. In this way, it is ensured that in large and complex terraced systems, water allocation meets both landscape aesthetic requirements and does not exceed the carrying capacity of the watershed's water resources.

[0090] Example 4: This example explores the application of this method in the dynamic water demand monitoring process of urbanization in mountainous areas, focusing on the impact of abrupt changes in landscape structure caused by the expansion of construction land on eco-hydrology.

[0091] In step 1, the system periodically (e.g., monthly) acquires the latest high-resolution satellite data and uses differential image analysis technology to quickly identify newly added construction land patches, road corridors, and interrupted original runoff paths. In the land feature classification process, special emphasis is placed on extracting the area of ​​artificial impermeable surfaces, as impermeable surfaces reduce the recharge efficiency of underground reservoirs.

[0092] In step 2, during model construction, the effective water storage parameter of the nonlinear regulation and storage model was adjusted in real time to address the structural damage to the unsaturated zone caused by urbanization. For areas covered by hardened surfaces, the regulation and storage coefficient in the discharge flow model was corrected to approach 0, while the lateral drainage volume was converted into urban stormwater runoff logic for calculation.

[0093] In step 3, the calculation of the landscape fragmentation index became the core monitoring indicator. As roads and buildings fragmented existing woodlands and waterways, the fragmentation index showed a rapid upward trend. The calculation of landscape connectivity incorporated the resistance coefficient of artificial barriers; for example, the obstructive effects of highways and retaining walls on surface runoff and biological migration were quantified as resistance values.

[0094] In step 4, a heat island effect correction term is introduced into the correction matrix. Due to higher temperatures in urban areas, evaporation losses in the flowing water ecological landscape increase. The correction module automatically increases the water allocation weight of surrounding ecological water bodies to cool the area and maintain the ecological microclimate. The correction coefficient is also linked to the landscape diversity index, ensuring sufficient diversity levels of the flowing water landscape in the remaining green spaces surrounded by the city by increasing water demand.

[0095] The output of step 5 is fed into the city's smart water management platform. The calculation results not only include water demand figures but also generate an ecological landscape damage assessment map. This map uses color depth to reflect the degree of ecological water shortage caused by the deterioration of the landscape pattern. When the expansion of construction land causes the landscape connectivity of a certain area to fall below a threshold, the system will automatically generate ecological corridor restoration suggestions, including specific water volume requirements for adding artificial wetlands or restoring connecting canals at specific locations.

[0096] Example 5: This Example 5 illustrates how this method can ensure the basic safety of mountain stream ecological landscapes by adjusting the core logic under extreme climate events (such as severe droughts lasting for months).

[0097] In step 1, the accuracy of traditional spectral classification decreases as surface vegetation withers due to water shortage. The system automatically switches to an auxiliary classification logic based on multi-temporal texture features and radar-derived soil moisture, maintaining accurate identification of landform patches by recognizing micro-topographic features under the withered vegetation.

[0098] In step 2, the nonlinear regulation model of the underground reservoir enters the extreme recession mode. In this mode, the focus is on simulating the slow release of stagnant water in deep bedrock fissures under extreme pressure gradients. Specific exponential parameters are adjusted to reflect the high resistance characteristics under low water cut conditions to accurately predict the maximum number of days the baseflow can support the flowing water.

[0099] In step 3, a dynamic drying risk weight is introduced into the calculation of the landscape connectivity index. As small tributaries dry up, the number of landscape nodes decreases, and connectivity declines sharply. The system identifies refuge patches that still have water resources during drought by calculating the aggregation degree of the remaining core habitat patches.

[0100] Step 4's correction matrix executes emergency preparedness logic. The negative correction effect of landscape fragmentation is temporarily suppressed, while the positive support effect of landscape connectivity is given the highest weight. This means the system will concentrate limited water resources to prioritize key corridors that connect multiple refuge patches, preventing structural disruptions to the ecosystem.

[0101] The output of step 5 is transformed into a drought management scheme. The scheme details the order of water distribution for pressure reduction at each level of landscape nodes. For example, priority is given to sacrificing water volume for aesthetic purposes to ensure sufficient water for the survival of key aquatic organisms. During the verification phase, the system incorporated data from pressure gauges deployed at the bottom of the riverbed. By monitoring flow changes at extremely low water levels, the accuracy of the nonlinear regression model at extreme points was verified.

[0102] Example 6: This example describes the application of the method in a large-scale forest fire prevention and ecological restoration project in mountainous areas. After a forest fire, the landscape pattern and geological water storage capacity of mountainous areas will undergo drastic changes. This method is used to calculate the ecological water allocation required during the restoration period.

[0103] In step 1, the normalized fire index is calculated using multispectral images before and after the fire to quickly extract the extent and intensity of the burned area. In the land cover classification process, the burned area is divided into lightly, moderately, and severely burned zones. Vegetation cover and litter layer (which has water-holding capacity) in severely burned zones are considered to be zero.

[0104] In step 2, the infiltration parameters of the storage model were adjusted to address the changes in soil physicochemical properties after fire (such as the formation of a hydrophobic layer). With reduced rainfall infiltration and increased surface runoff, the storage coefficient describing delayed release in the nonlinear storage model was lowered, simulating a rapid inflow and outflow hydrological response pattern.

[0105] Step 3, when calculating the landscape pattern index, focuses on the hollowing out of landscape patches caused by fire. The fragmentation index rises sharply at the edges of burned areas. The connectivity calculation incorporates a soil erosion sensitivity factor, meaning that although water transport is connected between damaged patches with steep slopes, the lack of vegetation to impede water flow increases the risk of erosion.

[0106] Step 4's correction matrix links water requirement quotas to the ecological restoration stage. Initially, the correction module allocates higher water volumes to moisten the topsoil and promote seed germination of pioneer vegetation. The correction coefficient dynamically decreases as landscape fragmentation improves (vegetation patch merging).

[0107] Step 5 outputs a dynamic water demand blueprint for ecological restoration. This blueprint guides the precise deployment of artificial rain enhancement operations or water diversion projects. The validation phase involves monitoring the recovery rate of vegetation cover within the burned area and comparing it with the ideal water demand path predicted by this method to evaluate the efficiency of ecological restoration and adjust the parameter weights in the correction matrix accordingly.

[0108] This embodiment illustrates and describes the basic principles and main features of the present invention, as well as its advantages. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for calculating water demand in a mountainous flowing ecological landscape, characterized in that, Includes the following steps: Step 1: Acquire basic geographic information data and multi-source remote sensing image data of mountainous areas, perform land cover classification processing on the multi-source remote sensing image data, identify land use types in the study area, and extract the distribution characteristics of karst landforms, ancient terraced fields and conventional mountain landscapes. Step 2: Construct a nonlinear regulation and storage model for underground reservoirs in mountainous areas. Use the principles of groundwater dynamics to describe the nonlinear relationship between underground water storage and discharge flow, and simulate the nonlinear retreat process of karst water in mountainous fissures or special geological bodies under gravity and the delayed release effect on the base flow. The nonlinear regulation and storage model for underground reservoirs in mountainous areas is constructed, and its discharge flow calculation logic is as follows: the discharge flow is equal to the product of a specific regulation and storage coefficient and a specific exponent power of the effective underground water storage volume; The specific exponent power is preset based on the pore structure characteristics and fracture development density of the geological body in the mountainous area; Step 3: Calculate the water ecological landscape pattern index. Using geographic information system analysis tools, quantitatively characterize the landscape fragmentation index and landscape connectivity index within the study area, and establish the mapping relationship between the surface landscape spatial structure and eco-hydrological processes. The process of calculating the water ecological landscape pattern index includes: defining the landscape fragmentation as the quotient obtained by dividing the total number of patches of a specific type in the landscape by the total area of ​​that type of landscape. The calculation logic of the landscape connectivity index is as follows: based on the principle of graph theory, aquatic habitat patches are defined as nodes, and ecological corridors and seasonal overflow paths between patches are defined as edges. The ease of water flow transmission between different patches is evaluated by calculating the probability connectivity index. The probability connectivity index is obtained by performing a comprehensive calculation using a negative exponential function of the geometric distance between patches, the habitat resistance coefficient, and the patch's own area weight. Step 4: Establish an ecological water demand quota correction matrix, using the landscape fragmentation index and the landscape connectivity index as constraint variables, and construct a functional relationship between the ecological water demand quota and the landscape pattern index through nonlinear regression analysis to determine the water demand correction coefficient under different landscape patterns. The specific logic for establishing the ecological water demand quota correction matrix is ​​as follows: landscape fragmentation is used as a negative correction factor, and landscape connectivity is used as a positive correction factor. When landscape fragmentation increases, the correction module increases the ecological water demand guarantee coefficient, and offsets the water dissipation caused by habitat edge effect by multiplying the basic water demand quota by a correction ratio greater than 1. When landscape connectivity is improved, the correction module performs a downward optimization on the water demand quota based on the correlation matrix, that is, multiplies it by a correction ratio that is less than 1 and greater than the safety threshold. Step 5: Encapsulate and run the water demand correction module for the flowing water ecological landscape. Input the base flow supply sequence calculated by the nonlinear regulation and storage model of the underground reservoir in the mountainous area into the correction module, and perform correction processing in combination with the dynamic parameters of the landscape pattern retrieved by real-time remote sensing, and output the final calculation result of the water demand of the flowing water ecological landscape in the mountainous area.

2. The method for calculating water demand in a mountainous flowing ecological landscape according to claim 1, characterized in that: The specific process of land feature classification in step 1 is as follows: using a combination of supervised classification and object-oriented classification, the study area is divided into forest land, grassland, water area, construction land, cultivated land and bare land; In object-oriented classification, pixels with similar spectral and texture features are aggregated into objects by setting scale parameters, shape parameters, and compactness parameters; For the ancient terraced fields, the second derivative of the slope is calculated using a digital elevation model to identify candidate ridge lines where the slope changes and exhibits ridge features. The water storage status inside the ridges is determined by combining water index and vegetation index. For karst landforms, the degree of karst development is determined by identifying closed depressions in the digital elevation model and combining them with the proportion of exposed rock area in multispectral data. All remote sensing data undergo radiometric calibration, atmospheric correction, and geometric fine correction based on the digital elevation model before classification.

3. The method for calculating water demand in a mountainous flowing ecological landscape according to claim 2, characterized in that: Step 1 also includes using a digital elevation model to extract the runoff accumulation, slope aspect, slope length, and topographic humidity index of the mountainous area. The topographic humidity index is obtained by taking the ratio between the tangent of the catchment area and the tangent of the slope, and then performing a natural logarithmic operation on the ratio. During the calculation process, the topographic humidity index and landscape pattern index are spatially overlaid to identify key ecological nodes in mountainous areas that are sensitive to water response, and assign them corresponding geographical weights.

4. The method for calculating water demand in a mountainous flowing ecological landscape according to claim 3, characterized in that: The construction of the nonlinear regulation and storage model for underground reservoirs in mountainous areas in step 2 also includes: in karst regions, the specific exponent power is set to a value greater than 1.5 and less than 2.

5. When the water storage volume is higher than the preset high threshold, the nonlinear adjustment model of the reservoir simulates the rapid response discharge of gravity flow in mountainous areas through a large hydraulic gradient. When the water storage volume decreases to a preset low threshold, the decay of the discharge flow rate exhibits a non-linear, slow trend to simulate the retention and delayed release mechanism of water in cracks and soil.

5. The method for calculating water demand in a mountainous flowing ecological landscape according to claim 4, characterized in that: The simulation of fractured karst water in step 2 also includes characterizing the dual-porosity medium effect, that is, dividing the geological body into a bedrock block system and a fracture network system. There is a pressure gradient-driven water exchange between the bedrock block system and the fracture network system, and the water exchange is proportional to the hydraulic conductivity between the bedrock block and the fracture. When water in the fracture network is drained by gravity, water in the bedrock block is horizontally replenished to the fractures by matrix suction. The specific water storage coefficient is obtained by spatial interpolation of the lithology, soil thickness and slope of the mountainous strata. The water storage coefficient is set to a small value in areas with high rock exposure and a large value in forest land and ancient terraced fields.

6. The method for calculating water demand in a mountainous flowing ecological landscape according to claim 5, characterized in that: The process of calculating the water ecological landscape pattern index in step 3 also includes: while calculating the landscape fragmentation, assessing the shape complexity of the patches, which is obtained by calculating the ratio of the patch perimeter to the circumference of a circle of equal area. As the complexity of the patch shape increases, it is determined that the contact interface between the patch and the external environment expands, and the weights of water evaporation and loss risk are increased in the subsequent correction matrix.

7. The method for calculating water demand in a mountainous flowing ecological landscape according to claim 6, characterized in that: The calculation logic of the landscape connectivity index in step 3 also includes: assigning migration resistance values ​​according to land use type, setting the resistance values ​​of natural forests and permanent water bodies to low level, and setting the resistance values ​​of construction land and high-intensity developed farmland to high level. The connectivity efficiency of water and biological circulation is assessed by calculating the minimum cost path between patches; at the same time, the core value of patches in the aquatic ecological network is determined by removing patches one by one and calculating the decrease in landscape connectivity values.

8. The method for calculating water demand in a mountainous flowing ecological landscape according to claim 7, characterized in that: The nonlinear regression analysis also includes sensitivity correction for rainfall pulse effects. By analyzing the hysteretic correlation between historical long-sequence rainfall data and baseflow changes, the time constant of rainfall infiltration into underground reservoirs and conversion into effective discharge flow is determined, and this time constant is introduced into the nonlinear regulation and storage model.

9. The method for calculating water demand in a mountainous flowing ecological landscape according to claim 8, characterized in that: The water demand correction module for the flowing water ecological landscape in step 5 adopts a layered architecture design, including a data access layer, a logic operation layer and a result output layer. The final calculation results generated by the output layer include a breakdown of water requirements for different ecosystem services: the minimum environmental water required to maintain the survival of aquatic organisms, the landscape water required to maintain the visual aesthetic value of the landscape, and the dilution and purification water required to maintain the self-purification capacity of the river. The environmental water volume is determined based on the flow rate and water depth requirements of suitable biological habitats; the landscape water volume is determined to maintain the mirror-like effect of waterfalls and terraced fields. Based on the functional positioning of the mountain ecological landscape, the various water quantities are weighted and summed using a weighting function to obtain the final dynamic value of total water demand. If the landscape connectivity within a time step is lower than the preset ecological security threshold, an early warning compensation mechanism is triggered, increasing the calculation weight of ecological water allocation.

10. The method for calculating water demand in a mountainous flowing ecological landscape according to claim 9, characterized in that: The method also includes a verification and adjustment step for the calculation results, the process of which is as follows: by setting up automatic flow monitoring stations at typical cross sections in mountainous areas to obtain measured flow data, the residual, root mean square error and correlation coefficient between the measured flow and the calculated ecological water demand are calculated. If the correlation coefficient is lower than the preset accuracy threshold, the parameter adaptive adjustment mechanism is activated. The calculation model is iteratively optimized by changing the exponential power parameter in the nonlinear storage model or adjusting the weight ratio of fragmentation and connectivity in the landscape correction matrix. To address the seasonal differences in the ecological landscape of mountain streams, three calculation modes were set up: dry season, wet season, and normal season. In the dry season mode, the weight of the baseflow contribution was increased, while in the wet season mode, the focus was on evaluating the contribution of landscape connectivity to flood peak reduction and the self-purification requirements of the water body.

Images