Phytoplankton community stability evaluation method based on cascade reservoir influence

Abstract

The invention discloses a phytoplankton community stability evaluation method based on cascade reservoir influence. The method comprises the following steps: step 1, river reach division and sample collection; step 2, community structure and diversity analysis; step 3, analyzing a community construction mechanism; step 4, co-occurrence network construction and stability evaluation; step 5, environment driving factor modeling; and step 6, performing multi-group comparative analysis. According to the method, multi-dimensional analysis of a community structure, a construction mechanism, network stability and environment driving factors is integrated, and a complete evaluation chain is formed; a neutral community model and a co-occurrence network analysis method are introduced, and the influence mechanism of the cascade reservoir is revealed from the angles of the ecological process and interaction; the method is suitable for river and reservoir systems of different scales and different types, and is especially suitable for ecological assessment of areas without data or with less data; the method has important theoretical and practical values for scientifically guiding ecological scheduling and biodiversity protection of the drainage basin.

Description

Technical Field

[0001] This invention belongs to the field of water ecological monitoring and river ecological management technology, and in particular relates to a method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs. Background Technology

[0002] With the continuous growth of global energy demand and the rapid development of water conservancy engineering technology, the construction of cascade reservoirs has been widely implemented in major river basins around the world. The cascade development model plays a crucial role in ensuring energy supply, regulating flood resources, improving irrigation efficiency, and enhancing navigation conditions. However, the continuous construction of a series of dams has significantly altered the natural hydrological rhythms and geomorphological features of rivers, leading to a series of physical, chemical, and ecological effects such as slower flow, increased water depth, water temperature stratification, and nutrient retention, which in turn have a profound impact on the structure and function of river ecosystems.

[0003] Phytoplankton, as primary producers in aquatic ecosystems, play a central role in energy flow, material cycling, and maintaining system stability. Due to their short generation cycles, wide distribution, and extreme sensitivity to environmental changes, phytoplankton are often considered indicator organisms reflecting the health of aquatic ecosystems. Currently, most research focuses on analyzing the impact of environmental factors within a single reservoir on phytoplankton community structure, or on changes in species composition in local river sections. However, these studies have significant limitations: First, they lack a systematic perspective, making it difficult to reveal the gradient impact of cascade development on the ecological processes of the entire river continuum. Second, the analysis of community building mechanisms is insufficient; existing studies mostly focus on describing the statistical correlation between environmental factors and community structure, failing to deeply quantify the relative contributions of deterministic and stochastic processes in community assembly and their spatial differentiation patterns, thus limiting the understanding of the intrinsic mechanisms of the ecological impacts of cascade reservoirs. Third, the assessment methods for ecosystem stability and resilience are relatively limited; traditional studies rely heavily on static indicators such as species diversity indices, failing to dynamically assess the community's ability to cope with disturbances and the recovery potential of downstream ecosystems based on the structural attributes of interspecific interaction networks. Finally, there is a lack of causal modeling that integrates multiple driving factors; most analyses remain at the correlation level and fail to construct path models that include geographical, physicochemical, habitat characteristics and biological interactions, in order to systematically elucidate the causal relationship between the maintenance and changes in phytoplankton community stability under the influence of cascade reservoirs.

[0004] Therefore, there is an urgent need for a systematic, multi-dimensional, and mechanism-oriented assessment method to evaluate the stability of phytoplankton communities under the influence of cascade reservoirs, so as to provide theoretical and practical references for a deeper understanding of the ecological effects of cascade reservoirs and for scientifically guiding watershed ecological scheduling and biodiversity conservation. Summary of the Invention

[0005] The purpose of this invention is to provide a method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs, in order to solve the above-mentioned technical problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention discloses a method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs, the method comprising the following steps:

[0008] Step 1: River Section Division and Sample Collection: A complete river section containing a cascade reservoir group was selected as the study area, which was then divided into three sub-sections: the upstream natural river section, the cascade reservoir-controlled river section, and the downstream naturally restored river section. Sampling points were set up in each of the three sub-sections, and surface water samples and phytoplankton samples were collected at each sampling point. Environmental parameters were measured using the water samples, and geographic and meteorological data were collected to form environmental factor data. Species identification and cell counting were performed using the phytoplankton samples to obtain species abundance data for each sampling point.

[0009] Step 2, Community Structure and Diversity Analysis: Based on phytoplankton species abundance data, the species richness and Shannon-Wiener diversity index of each sampling point are calculated to characterize the community diversity level; the relative abundance of different phytoplankton phyla is statistically analyzed based on taxonomic information to describe the community composition structure; the sampling points are grouped by river segment, and the relative abundance of phytoplankton and Shannon-Wiener diversity index are spatially compared, and statistical tests are used to assess the significance of diversity differences between different river segments;

[0010] Step 3: Community Building Mechanism Analysis: A neutral community model was used to evaluate the relative contributions of stochastic and deterministic processes in phytoplankton community building. Specifically, phytoplankton species abundance data from the upstream natural river section, the cascade reservoir-controlled river section, and the downstream naturally restored river section were fitted using a neutral community model. The model used the average relative abundance of each species in the regional species pool as the independent variable and the frequency of occurrence of the species at each sampling point as the dependent variable. A prediction curve of species occurrence frequency was constructed, which is the fitting curve. The goodness of fit R² of the model was calculated through regression analysis, and the community building mechanism was analyzed based on the goodness of fit R². The goodness of fit R² was used as an indicator of the community building mechanism.

[0011] The formula for calculating the goodness-of-fit R² is:

[0012] (2)

[0013] In the formula: For the first The actual frequency of occurrence of each species at each sampling point; The predicted species frequency from the neutral community model; is the average actual frequency of occurrence for all species; S is the number of species involved in the fitting.

[0014] The parameters of the Beta distribution function are determined by the relative abundance of the regional species pool and the model influx rate.

[0015] (3)

[0016] In the formula: For the first The relative abundance of each species in the regional species pool ; For the first The mean abundance of a species across all sampling sites; N: the average total number of individuals in each sampling site; m: the influx rate parameter of the neutral community model; d= The lower limit of detection; BetaCDF is the cumulative distribution function of the Beta distribution;

[0017] Step 4: Co-occurrence Network Construction and Stability Assessment: Construct a co-occurrence network of phytoplankton species, calculate network topology parameters, and assess network stability; the specific process includes the following steps:

[0018] (1) Co-occurrence network construction: Based on phytoplankton species abundance data, the correlation between any two phytoplankton species was calculated and characterized by the Spearman rank correlation coefficient; species pairs with a correlation coefficient |r|>0.6 and p<0.05 were selected as effective connections to construct an undirected weighted co-occurrence network; the formula for calculating the Spearman rank correlation coefficient between two species is:

[0019] (4)

[0020] In the formula: is the Spearman rank correlation coefficient; is the number of sampling points involved in the calculation; This represents the difference in abundance rank between two species at the j-th sampling point;

[0021] (2) Calculation of topology parameters: Calculate the overall and node topology parameters of the co-occurring network, including the number of network nodes and connections, average degree, average path length, network diameter, clustering coefficient, and modularity;

[0022] (3) Network stability assessment: The network robustness value is used as the quantitative indicator of community stability, namely the network stability index. The network stability index and the community building mechanism index constitute the community characteristic index. The environment disturbance is simulated by random node removal attack. In each simulation, a preset proportion of nodes in the network are randomly removed, and the proportion of the number of nodes that remain connected in the remaining network after the disturbance is calculated to the original total number of nodes. This proportion is defined as the network robustness value under one simulation. The random removal simulation is repeated multiple times, and the robustness value obtained from each simulation is averaged as the final robustness index of the co-occurring network.

[0023] Step 5: Environmental Driving Factor Modeling: Based on the environmental factor data and community characteristic indicators obtained in the preceding steps, a partial least squares path model is constructed. Specifically, the environmental factors and community characteristics are divided into several latent variables. The environmental factors are exogenous latent variables, including geographical factor latent variables, physicochemical property factor latent variables, and habitat characteristic factor latent variables. The community characteristics are endogenous latent variables, including community building mechanism latent variables and network stability latent variables. Secondly, the corresponding observation variables are selected according to the ecological meaning of each latent variable. Then, based on the impact mechanism of cascade reservoirs, the path relationships between latent variables are set, a partial least squares path model is constructed, the parameters of the model are estimated, the path coefficients of each path are obtained, and the intensity of different environmental factors is compared and analyzed through the model fitting results.

[0024] Step 6: Multi-group comparative analysis: Based on the completion of the partial least squares path model construction, the multi-group path comparative analysis method is used to identify the specific impacts of cascade reservoirs on phytoplankton communities. The specific process is as follows:

[0025] While maintaining consistency in latent variable settings, observed variable composition, and path structure, the sample was divided into two groups: the reservoir-affected section and the natural river section. Partial least squares path models were constructed for each group, and path coefficients between latent variables were calculated. These path coefficients are standardized regression coefficients used to quantify the direct influence of one latent variable on another, and their relationship can be expressed as follows:

[0026] (11)

[0027] In the formula: Indicates endogenous latent variables; The number of exogenous latent variables that have a direct path relationship with the endogenous latent variable Y; Let p be the p-th exogenous latent variable; The path coefficients for the corresponding paths; For residual terms;

[0028] After obtaining the path coefficients of the corresponding paths in different groups, the differences in the magnitude and direction of the path coefficients of the same paths in the reservoir-affected section and the natural river section are compared to calculate the inter-group differences in the path coefficients. This is used to determine whether the construction of cascade reservoirs has changed the intensity and direction of the effects of environmental factors on the phytoplankton community building mechanism and community stability.

[0029] Furthermore, the upstream natural river section mentioned in step 1 is located upstream of the starting point of the cascade reservoir construction, is not affected by reservoir regulation, and maintains its natural flow state; the cascade reservoir controlled river section includes the reservoir areas and backwater variation areas of multiple existing and under-construction cascade reservoirs, and is affected by reservoir scheduling; the downstream naturally restored river section is located downstream of the dam of the last-level reservoir, and the water flow has returned to its natural state, which is used to assess the downstream ecological restoration status.

[0030] Furthermore, the sampling point deployment described in step 1 specifically involves: within each river segment, the sampling points are deployed to cover the upper, middle, and lower reaches of that river segment, while also taking into account the main stream and tributaries, as well as areas with different water depths and flow velocities, ensuring that the number of sampling points in each river segment is no less than 5.

[0031] Furthermore, the environmental parameters mentioned in step 1 include chemical indicators, physical indicators, and habitat indicators; the chemical indicators include total phosphorus, total nitrogen, total silicon, ammonia nitrogen, nitrate nitrogen, and nitrite nitrogen; the physical indicators include water temperature, dissolved oxygen, and pH; the habitat indicators include substrate type, river width, and flow velocity; and the geographic meteorological data include the altitude of the sampling point and light intensity.

[0032] Furthermore, the formula for calculating the Shannon-Wiener diversity index mentioned in step 2 is as follows:

[0033] (1)

[0034] In the formula: The Shannon-Wiener diversity index; For species richness; For the first The relative abundance of a species is the proportion of individuals of that species to the total number of individuals in the community.

[0035] Furthermore, the analysis of the community building mechanism based on the goodness-of-fit R² described in step 3 is as follows: the goodness-of-fit R² represents the proportion of species frequency variation explained by the neutral community model, and its value is between 0 and 1; the higher the R² value, the stronger the model's explanatory power for community structure, and the greater the contribution of stochastic processes in community building; the lower the R² value, the more dominant the deterministic process is in community assembly; by comparing the differences in R² values ​​in different river sections, the spatial differentiation characteristics of the phytoplankton community building mechanism under the influence of cascade reservoirs can be revealed.

[0036] Furthermore, the calculation process for each topology parameter in step 4 is as follows:

[0037] Number of nodes M and number of connections E: The number of nodes M represents the number of species participating in the co-occurrence analysis in the network; the number of connections E represents the number of species pairs that meet the correlation threshold condition.

[0038] Average degree k: node degree Defined as a node The formula for calculating the average degree k of a network is:

[0039] (5)

[0040] Average path length L: The average path length L represents the average of the shortest path lengths between any two nodes in the network. Its calculation formula is:

[0041] (6)

[0042] In the formula: For nodes With nodes The shortest path length between them;

[0043] Network diameter D: The calculation formula is as follows:

[0044] (7)

[0045] Clustering coefficient C: Node Local clustering coefficient for:

[0046] (8)

[0047] In the formula: For nodes The actual number of connections between adjacent nodes;

[0048] The average clustering coefficient C of the network is:

[0049] (9)

[0050] Modularity Q: Modularity Q measures the degree to which nodes in a network are divided into different modules. Its calculation formula is as follows:

[0051] (10)

[0052] In the formula: These are the adjacency matrix elements of the network; They are nodes and nodes The degree; Indicates the module number to which the node belongs; For the Kronecker function, when The value is 1 if it is true, and 0 otherwise.

[0053] Furthermore, in step 5, the corresponding observation variables are selected based on the ecological meaning of each latent variable. Specifically, the geographical factors latent variable is represented by altitude and light intensity; the physicochemical indicators latent variable is represented by total nitrogen, total phosphorus, dissolved oxygen, and pH; the habitat elements latent variable is represented by flow velocity and river width; the community building mechanism latent variable is represented by the goodness of fit R²; and the community stability latent variable is represented by the robustness value of the co-occurrence network.

[0054] The beneficial effects of the present invention are as follows:

[0055] (1) Systematic: It integrates multi-dimensional analysis of community structure, construction mechanism, network stability and environmental driving factors to form a complete evaluation chain;

[0056] (2) Scientific nature: The neutral community model and co-occurrence network analysis method are introduced to reveal the impact mechanism of cascade reservoirs from the perspective of ecological processes and interactions;

[0057] (3) Practicality: Applicable to river and reservoir systems of different scales and types, especially suitable for ecological assessment in areas with no or little data;

[0058] (4) Guiding: The assessment results can be directly used for the ecological scheduling, water quality management and biodiversity protection of cascade reservoirs, supporting differentiated and precise ecological management, and have important theoretical and practical value for scientifically guiding the ecological scheduling and biodiversity protection of the basin.

[0059] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of the method flow described in this invention;

[0061] Figure 2 The graph shows the percentage of species, relative abundance of species, and diversity index of Example 1.

[0062] Figure 3 The figure shows the fitting curves of the neutral community model for the three river sections in Example 1.

[0063] Figure 4 This is a co-occurrence network diagram of Example 1;

[0064] Figure 5 This is a path model diagram of the "reservoir impact group" and "natural river section group" in Example 1. Detailed Implementation

[0065] This invention discloses a method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs, such as... Figure 1 As shown, the method includes the following steps:

[0066] Step 1: River Segment Division and Sample Collection: A complete river segment containing a cascade reservoir group was selected as the study area, which was then divided into three sub-segments: the upstream natural river segment, the cascade reservoir-controlled river segment, and the downstream naturally restored river segment. The upstream natural river segment is located upstream of the starting point of the cascade reservoir construction and is unaffected by reservoir regulation, maintaining its natural flow pattern. The cascade reservoir-controlled river segment includes the reservoir areas and backwater fluctuation zones of multiple existing and under-construction cascade reservoirs, and is significantly affected by reservoir scheduling. The downstream naturally restored river segment is located downstream of the dam of the last-stage reservoir, where the water flow has returned to its natural state, and is used to assess the downstream ecological restoration status.

[0067] Sampling points were established in three separate river sections, and surface water and phytoplankton samples were collected at each point. Within each section, sampling points should cover the upper, middle, and lower reaches, taking into account tributaries and areas with different water depths and flow velocities, ensuring at least five sampling points per section to guarantee spatial coverage and ecological representativeness. Water samples were used to determine environmental parameters, including chemical indicators (total phosphorus, total nitrogen, total silicon, ammonia nitrogen (NH3-N), nitrate nitrogen (NO3⁻-N), nitrite nitrogen (NO2⁻-N), etc.), physical indicators (water temperature, dissolved oxygen, pH, etc.), and habitat indicators (substrate type, river width, flow velocity, etc.). Geographic and meteorological data such as altitude and light intensity were also collected at the sampling points to form environmental factor data. Phytoplankton samples were fixed, precipitated, and concentrated, and then species identification and cell counting were performed under a microscope to obtain species abundance data for each sampling point. Phytoplankton species identification should be conducted with reference to "Chinese Freshwater Algae - Systematics, Classification and Ecology", and identification should be at least down to the family level.

[0068] Step 2, Community Structure and Diversity Analysis: Based on phytoplankton species abundance data, the species richness (i.e., total number of species) and Shannon-Wiener diversity index were calculated for each sampling point to characterize the community diversity level. The formula for calculating the Shannon-Wiener diversity index is as follows:

[0069] (1)

[0070] In the formula: The Shannon-Wiener diversity index; For species richness; For the first The relative abundance of a species (i.e., the proportion of individuals of that species to the total number of individuals in the community). The Shannon-Wiener diversity index is used to comprehensively characterize the species diversity level of a phytoplankton community.

[0071] Meanwhile, the relative abundance of different phytoplankton phyla was statistically analyzed based on taxonomic information to describe the community composition structure. Furthermore, the sampling points were grouped by river segment, and spatial comparisons of the relative abundance and diversity indices of phytoplankton were performed. Statistical tests were then used to assess the significance of diversity differences between different river segments.

[0072] Step 3: Community Building Mechanism Analysis: The Neutral Community Model (NCM) is used to assess the relative contributions of stochastic and deterministic processes in phytoplankton community building. Based on the ecological equivalence assumption, the NCM assumes that different species do not differ significantly in birth, death, and dispersal abilities, and that community structure is primarily driven by stochastic processes such as random dispersal and random drift. By comparing the predicted species frequencies with actual observations, the extent to which stochastic processes and deterministic processes such as environmental selection play a role in community assembly can be quantitatively determined.

[0073] Specifically, a neutral community model was applied to fit phytoplankton species abundance data from the upstream natural river section, the cascade reservoir-controlled river section, and the downstream naturally restored river section, respectively. The model used the average relative abundance of each species in the regional species pool as the independent variable and the frequency of occurrence of the species at each sampling point as the dependent variable, constructing a prediction curve for species occurrence frequency, i.e., the fitting curve. Regression analysis was used to calculate the goodness of fit R², which characterizes the degree of agreement between the model's predictions and actual observations.

[0074] (2)

[0075] In the formula: For the first The actual frequency of occurrence of each species at each sampling point; The predicted species frequency from the neutral community model; is the average actual frequency of occurrence for all species; S is the number of species involved in the fitting.

[0076] Species occurrence frequency predicted by the model The parameters of the Beta distribution function are determined by the relative abundance of the regional species pool and the model influx rate.

[0077] (3)

[0078] In the formula: For the first The relative abundance of each species in the regional species pool ; For the first The mean abundance of a species across all sampling sites; N: the average total number of individuals in each sampling site; m: the influx rate parameter of the neutral community model (estimated by nonlinear least squares); d= The lower limit of detection is given; BetaCDF is the cumulative distribution function of the Beta distribution.

[0079] The goodness-of-fit R² represents the proportion of species frequency variation explained by a neutral community model, with a value between 0 and 1. A higher R² value indicates a stronger explanatory power for community structure and a greater contribution of stochastic processes (including diffusion restriction and stochastic drift) to community assembly. Conversely, a lower R² value indicates that deterministic processes (including environmental factor selection and habitat differences) dominate community assembly. By comparing the differences in R² values ​​across different river sections, the spatial differentiation characteristics of phytoplankton community assembly mechanisms under the influence of cascade reservoirs can be revealed; therefore, the goodness-of-fit R² is used as an indicator of community assembly mechanisms.

[0080] Step 4: Co-occurrence Network Construction and Stability Assessment: Construct a co-occurrence network of phytoplankton species, calculate the network topology parameters, and assess the network stability. The specific process is as follows:

[0081] (1) Co-occurrence network construction: Based on phytoplankton species abundance data, the correlation between any two phytoplankton species was calculated and characterized by the Spearman rank correlation coefficient; species pairs with a correlation coefficient |r|>0.6 and p<0.05 were selected as effective connections to construct an undirected weighted co-occurrence network. The formula for calculating the Spearman rank correlation coefficient between two species is:

[0082] (4)

[0083] In the formula: is the Spearman rank correlation coefficient; is the number of sampling points involved in the calculation; This represents the difference in abundance rank between the two species at the j-th sampling point.

[0084] (2) Calculation of topology parameters: Calculate the overall and node topology parameters of the co-occurring network, including the number of network nodes and connections, average degree, average path length, network diameter, clustering coefficient, and modularity. The formulas for calculating each topology parameter are as follows:

[0085] Number of nodes M and number of connections E: The number of nodes M represents the number of species participating in the co-occurrence analysis in the network; the number of connections E represents the number of species pairs that meet the correlation threshold condition.

[0086] Average degree k: node degree Defined as a node The formula for calculating the average degree k of a network is:

[0087] (5)

[0088] Average path length L: The average path length L represents the average of the shortest path lengths between any two nodes in the network. Its calculation formula is:

[0089] (6)

[0090] In the formula: For nodes With nodes The shortest path length between them.

[0091] Network diameter D:

[0092] (7)

[0093] Clustering coefficient C: Node Local clustering coefficient for:

[0094] (8)

[0095] In the formula: For nodes The actual number of connections between adjacent nodes.

[0096] The average clustering coefficient C of the network is:

[0097] (9)

[0098] Modularity Q: Modularity Q measures the degree to which nodes in a network are divided into different modules. Its calculation formula is as follows:

[0099] (10)

[0100] In the formula: These are the adjacency matrix elements of the network; They are nodes and nodes The degree; Indicates the module number to which the node belongs; For the Kronecker function, when The value is 1 if it is true, and 0 otherwise.

[0101] (3) Network stability assessment:

[0102] In network stability assessment, network robustness is used as a quantitative indicator of community stability, i.e., the network stability index. The network stability index, together with the community building mechanism index, constitutes the community characteristic index. The robustness value is used to characterize the ability of a co-occurring network to maintain structural connectivity when subjected to random perturbations.

[0103] Specifically, the system simulates environmental disturbances through random node removal attacks. In each simulation, a predetermined proportion (e.g., 50%) of the nodes in the network are randomly removed, and the proportion of nodes remaining connected in the remaining network after the disturbance is calculated relative to the original total number of nodes. This proportion is defined as the network robustness value for one simulation. Multiple random removal simulations are repeated (generally no less than 1000 times), and the robustness values ​​obtained from each simulation are averaged to obtain the final robustness index for the co-occurring network.

[0104] Under the same node removal ratio and number of simulations, the larger the network robustness value, the more interconnected the network can be under random perturbations, the stronger its structural resistance to perturbations, and the higher the corresponding community stability; conversely, it indicates that the network is more sensitive to perturbations and the community stability is lower.

[0105] Step 5: Environmental driving factor modeling: Construct a partial least squares path model based on the environmental factor data and community characteristic indicators obtained in the previous steps.

[0106] First, environmental factors and community characteristics are divided into several latent variables. Environmental factors are exogenous latent variables, including geographical factors, physicochemical properties, and habitat characteristics. Community characteristics are endogenous latent variables, including community building mechanisms and network stability.

[0107] Secondly, corresponding observation variables are selected based on the ecological meaning of each latent variable. For example, altitude and light intensity are used to characterize geographical factors, total nitrogen, total phosphorus, dissolved oxygen and pH are used to characterize physicochemical indicators, and flow velocity and river width are used to characterize habitat factors. The goodness of fit R² of the neutral community model is used to characterize community building mechanism latent variables, and the robustness value of co-occurrence network is used to characterize community stability latent variables.

[0108] Based on this, and according to the impact mechanism of cascade reservoirs, path relationships between latent variables were defined, and a partial least squares path model was constructed to describe the influence of environmental factors on community building mechanisms and community stability through direct or indirect paths. Subsequently, the model parameters were estimated to obtain the path coefficients for each path, and the intensity of different environmental factors was compared and analyzed based on the model fitting results.

[0109] Step 6: Multi-group comparative analysis: Based on the completion of the partial least squares path model construction, the multi-group path comparative analysis method is used to identify the specific impact of cascade reservoirs on phytoplankton communities.

[0110] Specifically, while maintaining consistency in the latent variable settings, observed variable composition, and path structure, the sample was divided into two groups: the reservoir-affected section and the natural river section. Partial least squares path models were constructed for each group, and path coefficients between latent variables were calculated. These path coefficients are standardized regression coefficients used to quantify the direct influence of one latent variable on another, and their relationship can be expressed as follows:

[0111] (11)

[0112] In the formula: Indicates endogenous latent variables; The number of exogenous latent variables that have a direct path relationship with the endogenous latent variable Y; Let p be the p-th exogenous latent variable; The path coefficients for the corresponding paths; This represents the residual term. The path coefficients are obtained through regression analysis of the latent variable scores, and their solution process is based on parameter estimation by minimizing the sum of squared residuals.

[0113] After obtaining the path coefficients for corresponding paths in different groups, the inter-group differences in path coefficients were calculated by comparing the magnitude and direction of the path coefficients for the same paths in the reservoir-affected sections and natural river sections. This was used to determine whether the construction of cascade reservoirs altered the intensity and direction of environmental factors' influence on the phytoplankton community building mechanism and community stability. The greater the inter-group difference in path coefficients, the more significant the impact of the cascade reservoirs on that path. Based on the results of multi-group path comparison analysis, differentiated ecological protection and reservoir scheduling recommendations for different river sections were further proposed, including optimizing water release timing, maintaining ecological flow, and improving hydrodynamic conditions.

[0114] Example 1

[0115] This embodiment is an application example of the above method.

[0116] This embodiment discloses a method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs, including the following steps:

[0117] Step 1: River Section Division and Sample Collection: The upper reaches of the Yangtze River, from Zhimen to Zhutuo, were selected as the study area. Based on the spatial distribution of the cascade reservoir group, the study area was divided into three sub-sections:

[0118] River section S1 (upstream natural river section): Located upstream of the starting point of the cascade reservoir construction, it is not affected by reservoir regulation and maintains its natural flow.

[0119] River section S2 (cascade reservoir controlled section): includes the reservoir area and backwater variation area of ​​multiple existing and under-construction cascade reservoirs, and is significantly affected by reservoir scheduling.

[0120] River section S3 (downstream naturally restored section): Located downstream of the last reservoir dam, the water flow has returned to its natural state and is used to assess the downstream ecological restoration status.

[0121] Sampling point layout: In each river section, a total of 34 representative sampling points were laid out along the longitudinal, transverse (main stream and major tributaries) and vertical (considering different water depths) directions of the river (9 points for river section S1, 12 points for river section S2, and 13 points for river section S3) to ensure comprehensive spatial coverage and ecological representativeness.

[0122] Data Acquisition and Integration: Water and phytoplankton samples were collected quarterly from June 2021 to November 2023. After fixation and concentration using Luger's iodine solution, species identification and cell counting were performed under a microscope to obtain a species abundance matrix for each sampling point. Simultaneously, water samples were collected to measure total phosphorus (TP), total nitrogen (TN), dissolved oxygen (DO), and pH; flow velocity and river width were measured on-site; and grid data on elevation and light intensity at the sampling points were integrated. All data were standardized, outliers were removed, and a structured dataset was created for subsequent analysis.

[0123] Step 2, Community Structure and Diversity Analysis: Import the preprocessed phytoplankton species abundance data into analysis software (such as R language) and perform the following operations: Calculate the species richness and Shannon-Wiener diversity index for each sampling point, and plot the percentage of species abundance, relative species abundance, and diversity index, as shown below. Figure 2 As shown, the proportion of different phyla of phytoplankton (such as diatoms, green algae, cyanobacteria, etc.) at each sampling point is displayed intuitively, and the diversity differences of the three river sections S1, S2, and S3 are compared and significance tests are performed.

[0124] Example of output results: The analysis shows that from upstream to downstream, the species richness follows the trend of S3>S2>S1; while the Shannon diversity index shows the highest value at S2, indicating that the species composition in the reservoir's influence area is the most complex.

[0125] Step 3: Community Building Mechanism Analysis: Using the corresponding packages in R (such as Hmisc), a neutral community model was applied to fit the phytoplankton species abundance data of the three river segments S1, S2, and S3. The goodness-of-fit R² value of the model was calculated. A higher R² value indicates a greater contribution of stochastic processes to community assembly; conversely, a lower R² value indicates that deterministic processes dominate.

[0126] Plot the neutral model community fit diagrams for the three river segments, showing the fit curves and 95% confidence intervals, such as... Figure 3 As shown, from left to right, these are the fitted curves for river segments S1, S2, and S3.

[0127] Example output: The higher R² values ​​in sections S1 and S3 (0.564 and 0.535 respectively) indicate that stochastic processes dominate; the lower R² value in section S2 (0.213) indicates that deterministic processes dominate. This quantifies that the cascade reservoirs enhance environmental screening by creating a stable environment.

[0128] Step 4: Co-occurrence network construction and stability assessment:

[0129] Network Construction: Based on species abundance data from all sampling points, the Spearman rank correlation coefficient between any two species is calculated. A threshold is set (e.g., |r|>0.6 and p<0.05) to retain significantly correlated species pairs, thus constructing a phytoplankton species co-occurrence network.

[0130] Topology parameter calculation: Calculate the topology parameters of the network as a whole and its nodes, such as the number of nodes, number of edges, average degree, average path length, clustering coefficient, modularity, etc., as shown in Table 1, to quantify the complexity of the network structure.

[0131] Table 1

[0132]

[0133] Stability (robustness) assessment: Write a simulation script to randomly remove a certain percentage (e.g., 50%) of nodes from the network. Calculate the proportion of species remaining connected in the network after the removal. Repeat the simulation multiple times and take the average value as a quantitative indicator of network stability. Plot a stability curve to show the decline in network connectivity as nodes are randomly removed, such as... Figure 4 As shown.

[0134] Visualization: Use network visualization tools (such as ggplot2) to draw co-occurrence network diagrams, such as... Figure 4 As shown, node size can represent species abundance, and edge color can represent positive / negative correlation.

[0135] Example output results: S2 segment network has the most nodes and connections, but lower modularity; S1 network is simple but has high natural connectivity; S3 network shows a recovery trend in terms of complexity and organization. Stability simulation shows that the robustness of S3 segment network is about 15% higher than that of S2.

[0136] Step 5: Environmental Driving Factor Modeling: Partial Least Squares Path Model Construction: Using software such as SmartPLS, observed variables were divided into different latent variables based on the ecological attributes of environmental factors. Specifically: altitude and light intensity were used as observed variables to construct geographical factor latent variables; total phosphorus (TP), total nitrogen (TN), dissolved oxygen (DO), and pH were used as observed variables to construct physicochemical index latent variables; and flow velocity and river width were used as observed variables to construct habitat element latent variables. Subsequently, based on the impact mechanism of cascade reservoirs and the logic of ecological processes, the path relationships between latent variables were established. The above-mentioned environmental factor latent variables were used as exogenous latent variables, and community building mechanism indicators and community stability indicators were introduced as endogenous latent variables to form a complete structural model. The model uses phytoplankton community assembly mechanism (represented by the R² value of NCM) and network stability as endogenous latent variables to explore the direct and indirect influence paths of environmental factors.

[0137] Step 6: Multi-group comparative analysis: Divide the data into the "reservoir impact group" (S2) and the "natural river section group" (S1+S3), and run MGA. Compare the differences in path coefficients between the two groups to examine whether the reservoir construction has changed the impact pattern of environmental factors on the community.

[0138] Generate a path model diagram, such as Figure 5 As shown, the path relationships and coefficients between the variables are clearly displayed.

[0139] Example of output results: The model shows that in section S2, habitat characteristics such as flow velocity and river width have a significant positive direct impact on community stability; while in sections S1 and S3, the direct impact of these environmental factors is not significant. MGA confirms a significant difference between the two groups of path coefficients (P=0.023).

[0140] Based on the analysis results of the above steps, a systematic diagnosis is performed:

[0141] Current Status Assessment: Clarify the spatial differentiation characteristics of phytoplankton community structure, diversity, assembly mechanism, and network stability under the influence of cascade reservoirs.

[0142] Mechanism analysis: This study elucidates how reservoir construction enhances environmental screening (a deterministic process) by altering hydrological conditions (such as reducing flow velocity and increasing residence time), and thus affects the structure and stability of interspecific interaction networks.

[0143] Resilience assessment: Based on signs of recovery in the community characteristics of the S3 segment, assess the natural recovery potential of the downstream ecosystem.

[0144] Management Recommendations: Generate a comprehensive assessment report and propose targeted management recommendations. For example: Implement ecological scheduling in S2 (reservoir impact zone) to optimize downstream water flow patterns and maintain suitable habitats; strengthen non-point source pollution control in S3 (downstream recovery zone) to prevent excessive nutrient input leading to excessive algal proliferation.

[0145] The method and system described in this invention have been applied in actual river ecological assessments. The results show that they have good scientific validity, practicality and operability, and can provide technical support for the ecological protection, water environment management and sustainable operation of cascade reservoirs. They have broad prospects for promotion and application.

[0146] Finally, it should be noted that the above description is only used to illustrate the technical solution of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred arrangement, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs, characterized in that, The method includes the following steps: Step 1: River Section Division and Sample Collection: A complete river section containing a cascade reservoir group was selected as the study area, which was then divided into three sub-sections: the upstream natural river section, the cascade reservoir-controlled river section, and the downstream naturally restored river section. Sampling points were set up in each of the three sub-sections, and surface water samples and phytoplankton samples were collected at each sampling point. Environmental parameters were measured using the water samples, and geographic and meteorological data were collected to form environmental factor data. Species identification and cell counting were performed using the phytoplankton samples to obtain species abundance data for each sampling point. Step 2, Community Structure and Diversity Analysis: Based on phytoplankton species abundance data, the species richness and Shannon-Wiener diversity index of each sampling point are calculated to characterize the community diversity level; the relative abundance of different phytoplankton phyla is statistically analyzed based on taxonomic information to describe the community composition structure; the sampling points are grouped by river segment, and the relative abundance of phytoplankton and Shannon-Wiener diversity index are spatially compared, and statistical tests are used to assess the significance of diversity differences between different river segments; Step 3: Community Building Mechanism Analysis: A neutral community model was used to evaluate the relative contributions of stochastic and deterministic processes in phytoplankton community building. Specifically, phytoplankton species abundance data from the upstream natural river section, the cascade reservoir-controlled river section, and the downstream naturally restored river section were fitted using a neutral community model. The model used the average relative abundance of each species in the regional species pool as the independent variable and the frequency of occurrence of the species at each sampling point as the dependent variable. A prediction curve of the species occurrence frequency was constructed, which is the fitting curve. The goodness of fit R² of the model was calculated through regression analysis, and the community building mechanism was analyzed based on the goodness of fit R². The goodness of fit R² was used as an indicator of the community building mechanism. The formula for calculating the goodness-of-fit R² is: （2） In the formula: For the first The actual frequency of occurrence of each species at each sampling point; The predicted species frequency from the neutral community model; is the average actual frequency of occurrence for all species; S is the number of species involved in the fitting. The parameters of the Beta distribution function are determined by the relative abundance of the regional species pool and the model influx rate. （3） In the formula: For the first The relative abundance of each species in the regional species pool ; For the first The mean abundance of a species across all sampling sites; N: the average total number of individuals in each sampling site; m: the influx rate parameter of the neutral community model; d= The lower limit of detection; BetaCDF is the cumulative distribution function of the Beta distribution; Step 4: Co-occurrence Network Construction and Stability Assessment: Construct a co-occurrence network of phytoplankton species, calculate network topology parameters, and assess network stability; the specific process includes the following steps: (1) Co-occurrence network construction: Based on phytoplankton species abundance data, the correlation between any two phytoplankton species was calculated and characterized by the Spearman rank correlation coefficient; species pairs with a correlation coefficient |r|>0.6 and p<0.05 were selected as effective connections to construct an undirected weighted co-occurrence network; the formula for calculating the Spearman rank correlation coefficient between two species is: （4） In the formula: is the Spearman rank correlation coefficient; is the number of sampling points involved in the calculation; This represents the difference in abundance rank between two species at the j-th sampling point; (2) Calculation of topology parameters: Calculate the overall and node topology parameters of the co-occurring network, including the number of network nodes and connections, average degree, average path length, network diameter, clustering coefficient, and modularity; (3) Network stability assessment: The network robustness value is used as the quantitative indicator of community stability, namely the network stability index. The network stability index and the community building mechanism index constitute the community characteristic index. The environment disturbance is simulated by random node removal attack. In each simulation, a preset proportion of nodes in the network are randomly removed, and the proportion of the number of nodes that remain connected in the remaining network after the disturbance is calculated to the original total number of nodes. This proportion is defined as the network robustness value under one simulation. The random removal simulation is repeated multiple times, and the robustness value obtained from each simulation is averaged as the final robustness index of the co-occurring network. Step 5: Environmental Driving Factor Modeling: Based on the environmental factor data and community characteristic indicators obtained in the preceding steps, a partial least squares path model is constructed. Specifically, the environmental factors and community characteristics are divided into several latent variables. The environmental factors are exogenous latent variables, including geographical factor latent variables, physicochemical property factor latent variables, and habitat characteristic factor latent variables. The community characteristics are endogenous latent variables, including community building mechanism latent variables and network stability latent variables. Secondly, the corresponding observation variables are selected according to the ecological meaning of each latent variable. Then, based on the impact mechanism of cascade reservoirs, the path relationships between latent variables are set, a partial least squares path model is constructed, the parameters of the model are estimated, the path coefficients of each path are obtained, and the intensity of different environmental factors is compared and analyzed through the model fitting results. Step 6: Multi-group comparative analysis: Based on the completion of the partial least squares path model construction, the multi-group path comparative analysis method is used to identify the specific impacts of cascade reservoirs on phytoplankton communities. The specific process is as follows: While maintaining consistency in latent variable settings, observed variable composition, and path structure, the sample was divided into two groups: the reservoir-affected section and the natural river section. Partial least squares path models were constructed for each group, and path coefficients between latent variables were calculated. These path coefficients are standardized regression coefficients used to quantify the direct influence of one latent variable on another, and their relationship can be expressed as follows: （11） In the formula: Indicates endogenous latent variables; The number of exogenous latent variables that have a direct path relationship with the endogenous latent variable Y; Let p be the p-th exogenous latent variable; These are the path coefficients for the corresponding paths; For residual terms; After obtaining the path coefficients of the corresponding paths in different groups, the differences in the magnitude and direction of the path coefficients of the same paths in the reservoir-affected section and the natural river section are compared to calculate the inter-group differences in the path coefficients. This is used to determine whether the construction of cascade reservoirs has changed the intensity and direction of the effects of environmental factors on the phytoplankton community building mechanism and community stability.

2. The method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs according to claim 1, characterized in that, The upstream natural river section mentioned in step 1 is located upstream of the starting point of the cascade reservoir construction, and is not affected by reservoir regulation, maintaining its natural flow state; the cascade reservoir controlled river section includes the reservoir areas and backwater variation areas of multiple existing and under-construction cascade reservoirs, and is affected by reservoir scheduling; the downstream naturally restored river section is located downstream of the dam of the last-level reservoir, and the water flow has returned to its natural state, which is used to assess the downstream ecological restoration status.

3. The method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs according to claim 1, characterized in that, The specific steps for setting up sampling points in step 1 are as follows: within each river segment, sampling points are set up to cover the upper, middle, and lower reaches of that river segment, taking into account the main stream and tributaries, as well as areas with different water depths and flow velocities, to ensure that the number of sampling points in each river segment is no less than 5.

4. The method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs according to claim 1, characterized in that, The environmental parameters mentioned in step 1 include chemical indicators, physical indicators, and habitat indicators; the chemical indicators include total phosphorus, total nitrogen, total silicon, ammonia nitrogen, nitrate nitrogen, and nitrite nitrogen; the physical indicators include water temperature, dissolved oxygen, and pH; the habitat indicators include substrate type, river width, and flow velocity; and the geographic meteorological data include the altitude of the sampling point and light intensity.

5. The method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs according to claim 1, characterized in that, The formula for calculating the Shannon-Wiener diversity index in step 2 is as follows: （1） In the formula: The Shannon-Wiener diversity index; For species richness; For the first The relative abundance of a species is the proportion of individuals of that species to the total number of individuals in the community.

6. The method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs according to claim 1, characterized in that, The analysis of the community building mechanism based on the goodness-of-fit R² in step 3 is as follows: the goodness-of-fit R² represents the proportion of species frequency variation explained by the neutral community model, and its value is between 0 and 1; the higher the R² value, the stronger the model's explanatory power for community structure, and the greater the contribution of stochastic processes in community building; the lower the R² value, the more dominant the deterministic process is in community assembly; by comparing the differences in R² values ​​in different river sections, the spatial differentiation characteristics of the phytoplankton community building mechanism under the influence of cascade reservoirs can be revealed.

7. The method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs according to claim 1, characterized in that, The calculation process for each topology parameter in step 4 is as follows: Number of nodes M and number of connections E: The number of nodes M represents the number of species participating in the co-occurrence analysis in the network; the number of connections E represents the number of species pairs that meet the correlation threshold condition. Average degree k: node degree Defined as a node The formula for calculating the average degree k of a network is: （5） Average path length L: The average path length L represents the average of the shortest path lengths between any two nodes in the network. Its calculation formula is: （6） In the formula: For nodes With nodes The shortest path length between them; Network diameter D: The calculation formula is as follows: （7） Clustering coefficient C: Node Local clustering coefficient for: （8） In the formula: For nodes The actual number of connections between adjacent nodes; The average clustering coefficient C of the network is: （9） Modularity Q: Modularity Q measures the degree to which nodes in a network are divided into different modules. Its calculation formula is as follows: （10） In the formula: These are the elements of the adjacency matrix of the network; They are nodes and nodes The degree; Indicates the module number to which the node belongs; For the Kronecker function, when The value is 1 if it is true, and 0 otherwise.

8. The method for assessing the stability of phytoplankton communities based on the influence of cascade reservoirs according to claim 1, characterized in that, Step 5 describes selecting corresponding observation variables based on the ecological meaning of each latent variable. Specifically, altitude and light intensity are used to characterize geographical factor latent variables, total nitrogen, total phosphorus, dissolved oxygen and pH are used to characterize physicochemical index latent variables, and flow velocity and river width are used to characterize habitat element latent variables. The goodness-of-fit R² is used to characterize the latent variables of community building mechanism, and the robustness value of co-occurrence network is used to characterize the latent variables of community stability.

Images