A method for targeted screening of dominant aquatic plants for the remediation of heavy metal pollution in karst water
By combining remote sensing imagery data and field surveys with deep learning models, highly adaptable aquatic plants were selected, which solved the problem of poor remediation effect of heavy metal pollution in karst water bodies in existing technologies and achieved efficient remediation of karst water bodies.
Patent Information
- Application Number
- CN202611276006.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-21
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies lack a targeted, systematic, and standardized screening system for aquatic plants used in the remediation of heavy metal pollution in karst waters in Southwest China. This makes it impossible to accurately, comprehensively, and safely screen dominant species in karst habitats, resulting in unsatisfactory remediation outcomes.
By acquiring remote sensing image data of the target karst area, determining community distribution data, conducting field surveys and tests, classifying water environment categories, continuously monitoring heavy metal concentrations, carrying out stress acclimatization and multi-dimensional weighted quantitative evaluation, using deep learning models to predict restoration and development trends, and selecting the best dominant species.
It achieves precise, systematic and standardized screening for different karst water environments, ensuring that the screened aquatic plants have long-term sustainable heavy metal remediation capacity and ecological adaptability in karst water bodies, thus improving the reliability and efficiency of the remediation effect.
Smart Images

Figure CN123117528A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the fields of karst water ecological environment restoration, water heavy metal pollution control, and aquatic plant resource screening and ecological adaptation technology. Specifically, it relates to a method for targeted screening of superior aquatic plants for the remediation of heavy metal pollution in karst water with high calcium and alkaline water environment in Southwest China and for scenarios of heavy metal complex pollution. Background Technology
[0002] The karst region in Southwest my country is the world's largest and most typical karst ecological distribution area, with karst water serving as the core water source for regional production, daily life, and ecology. Influenced by its geological formation, the karst water in Southwest China is characterized by high calcium content, alkalinity, complex hydrological structure, and weak self-purification capacity. Due to the continuous and intensifying human interference from industrial and mining activities, agricultural non-point source pollution, and rural sewage discharge, karst water bodies in the Southwest karst region generally suffer from complex pollution of multiple heavy metals such as cadmium, lead, zinc, and copper. This pollution is highly concealed, has a long retention period, and is difficult to treat, severely restricting the water ecological security of the karst region.
[0003] Aquatic plant ecological restoration technology has become the mainstream green technology for the treatment of heavy metal pollution in water bodies due to its advantages of low cost, no secondary pollution, good ecological compatibility, and strong sustainability. Aquatic plants such as emergent plants (reeds, cattails), submerged plants (Hydrilla verticillata, duckweed) and floating plants (water hyacinth, duckweed) have a certain ability to accumulate and tolerate single or complex heavy metals. However, the hydrological conditions of karst water bodies are complex, and the restoration performance of the same aquatic plant may vary greatly in different water bodies. Therefore, how to select aquatic plants that are suitable for various karst water environments directly affects the quality of restoration.
[0004] Most existing aquatic plant screening systems and methods are designed for non-karst water environments and do not simulate the high calcium and alkaline stress characteristics of karst water. Remediation plants that perform well in non-karst water are prone to calcium stress and acid-base stress damage in karst-specific water environments, resulting in problems such as photosynthetic decline, root rot, sharp drop in survival rate, significant reduction in enrichment efficiency, and even death. They have extremely poor engineering practicality and cannot meet the specific needs of karst scenarios.
[0005] Traditional screening methods for karst aquatic environments suffer from drawbacks such as crude and untargeted germplasm resource assessment, relying heavily on random field surveys and empirical sampling. The resulting aquatic plants exhibit poor adaptability and are insufficient to meet the actual remediation needs of various karst water bodies and dominant heavy metals. Furthermore, these methods lack reasonable and reliable simulation and verification procedures, leading to a severe disconnect between the screened habitats and the actual karst aquatic environment.
[0006] Current technologies for evaluating remediation effectiveness rely on a single, limited dimension, resulting in incomplete information. They often use the accumulation of a single heavy metal in plants as the sole screening indicator. However, the hydrological conditions in karst areas are complex, with various water body types exhibiting different adaptability to aquatic plants. A high heavy metal accumulation does not necessarily indicate that the aquatic plant possesses the ability for long-term, sustainable, and effective remediation in all types of water bodies. Furthermore, due to the diversified and dispersed nature of pollution sources in karst areas, the heavy metal pollution status of karst water bodies varies significantly, particularly in the dominant heavy metal species. Existing screening methods struggle to provide targeted adaptation to diverse hydrological conditions, and the selected species often fail to achieve ideal remediation results in the field, making it difficult to guarantee scenario suitability.
[0007] Therefore, the industry currently lacks a targeted, systematic, and standardized screening system for aquatic plants used in the remediation of heavy metal pollution in karst waters in Southwest China. This makes it impossible to accurately, comprehensively, and safely screen dominant species in karst habitats, which severely restricts the standardized and large-scale application of ecological remediation technologies for heavy metal pollution in karst waters in Southwest China. Summary of the Invention
[0008] The purpose of this invention is to provide a method for targeted screening of superior aquatic plants for the remediation of heavy metal pollution in karst water, so as to solve the technical problems mentioned in the background art.
[0009] This invention provides a method for targeted screening of dominant aquatic plants for the remediation of heavy metal pollution in karst water, comprising: Acquire remote sensing image data of the target karst area, determine the community distribution data of each karst water body based on the remote sensing image data; conduct field investigation and testing of each karst water body and identify the heavy metal with the highest concentration as the dominant heavy metal species; classify karst water bodies with the same dominant heavy metal species and water body type into Class I water environment. For each type of water environment, the concentration of dominant heavy metals in karst water bodies is continuously monitored. Based on the concentration change data of dominant heavy metals, a preliminary assessment of the remediation effectiveness is conducted to determine the representative karst water bodies of the current water environment. Based on the community distribution data of the representative karst water bodies, the target community is determined. Field surveys are conducted to record the species types covered by the target community, as well as the coverage area and biomass of each species. Based on the analysis of the recorded data, the range of dominant aquatic plant species is determined. The range of dominant aquatic plant species was screened to identify dominant candidate species. For each dominant candidate species, water samples were collected in the field according to the associated aquatic environment type. Based on the water samples, a karst-specific culture medium was artificially simulated, and the experimental plants of the dominant candidate species were subjected to stress acclimatization culture using the karst-specific culture medium. For the dominant candidate species that underwent stress acclimatization culture, a comprehensive score was obtained through multi-dimensional weighted quantitative evaluation based on heavy metal composite remediation performance, karst-specific environment adaptability, long-term growth stability, and overall ecological safety. For each type of water environment, field verification experiments are conducted on the dominant candidate species whose comprehensive scores are greater than the preset score threshold. The concentration change sequences of all heavy metal species in the field test water, the environmental data change sequences, the coverage area change sequences and growth change sequences of the dominant candidate species are input into a pre-trained prediction model based on a deep learning model to predict the remediation development trend. The dominant candidate species with the best prediction results is determined as the best dominant species for the current water environment.
[0010] As a preferred option, the community distribution data of each karst water body determined based on remote sensing image data includes: The community distribution data includes the distribution status of emergent, floating, and submerged plants; An object-oriented random forest algorithm combined with a deep learning CNN spectral classification model is used to extract the specific spectral features, texture features, and spatial distribution features of aquatic vegetation to distinguish emergent, floating, and submerged plant communities. The vegetation phenology and the wet-dry cycle of water bodies in the target karst area are used as prior knowledge inputs into the community classification model to optimize the model parameters.
[0011] As a preferred approach, a preliminary assessment of remediation effectiveness is conducted based on data on changes in the concentration of dominant heavy metals to determine representative karst water bodies in the current aquatic environment, including: The daily average reduction and residue ratio of the dominant heavy metal concentration are determined based on the concentration change data. The daily average reduction is expressed as: (initial concentration - final concentration) ÷ number of monitoring days; the residue ratio is expressed as: (final concentration ÷ initial concentration) × 100%. After normalizing the daily reduction, a weighted average score for the remediation effectiveness is calculated by combining the reduction with the residual ratio. The karst water body with the highest preliminary remediation effectiveness score is identified as the representative karst water body of the current water environment. Based on the community distribution data of representative karst water bodies, the target communities were identified as including: The communities with the highest area coverage ratio among all communities are identified as target communities.
[0012] As a preferred method, field surveys should record the species diversity of the target community, as well as the area coverage and biomass percentage of each species, including: For the target communities, a combination of transect and quadrat method was used for verification. Emergent plant communities were divided into 1m×1m quadrats, and floating and submerged plant communities were divided into 0.5m×0.5m quadrats. Each transect had 5 to 8 replicate quadrats evenly distributed. Each quadrat records all species and the percentage of coverage area and biomass of each species within the current quadrat. Based on the analysis of recorded data, the range of dominant aquatic plant species included: For each quadrat, the compatibility of each species with the current dominant heavy metal species is determined based on the coverage area ratio and biomass ratio; the species with the highest compatibility is selected for each quadrat to determine the range of dominant aquatic plant species.
[0013] As a preferred approach, screening the range of dominant aquatic plant species to identify dominant candidate species includes: Ecological security screening is conducted in advance to preserve native species in the karst landscape of Southwest China, while eliminating invasive species, artificially hybridized and improved species, and rare and protected species. Basic stress tolerance screening was conducted to eliminate vulnerable species that are sensitive to high-calcium and alkaline water environments, have a natural habitat survival rate of less than 85%, or suffer from obvious calcium stress and acid-base stress damage. Growth controllability screening was conducted to eliminate invasive species that reproduced explosively and spread too quickly, as well as unstable species with short growth cycles, severe seasonal wilting, and complete winter dormancy.
[0014] As a preferred method, when collecting water samples in the field according to the associated water environment type, it is necessary to test the water's pH value, calcium ion concentration, water hardness, dissolved oxygen, and water temperature on-site. After sampling, the types and concentrations of heavy metals in the water should be determined in the laboratory to finally determine the water samples that meet the current water environment and can be used for artificial simulation of karst-specific culture solutions.
[0015] As a preferred method, an artificially simulated karst-specific culture medium is used based on water samples. The stress acclimatization culture of the test plants of the dominant candidate species is then conducted using this karst-specific culture medium, including: Standardized culture medium preparation: Based on water quality test results, artificial simulation of karst-specific culture medium is constructed. Standardized pretreatment of plants: all experimental plants were uniformly cleaned of root mud, damaged branches and leaves were pruned, plant height and biomass were uniformly specified, and they were uniformly planted in artificial culture devices. The three-stage gradient acclimatization cycle, with a total acclimatization time of 45 days, is divided into three stages of gradient stress cultivation: the first stage is a 15-day low-stress adaptation period, during which heavy metal stress is weakened, and no or only very low concentrations of dominant heavy metals are added; the second stage is a 20-day moderate constant stress period, during which multiple heavy metal pollutants are added, with the amount of the dominant heavy metal added being greater than that of other heavy metals; maintaining a high-calcium, slightly alkaline, and multi-metal complex stress environment; the third stage is a 10-day stabilization acclimatization period, during which the plant's metabolic level and accumulation performance are stabilized, and the cultivation process simultaneously simulates the natural characteristics of karst waters, with a diurnal temperature range of 6-10℃ and intermittent hydrological wet-dry alternation.
[0016] As a preferred option, a comprehensive score was obtained through a multi-dimensional weighted quantitative evaluation of heavy metal composite remediation performance, adaptability to karst-specific environments, long-term growth stability, and overall ecological safety. The score included: A combined subjective and objective scoring model using the Analytic Hierarchy Process (AHP) and the entropy weight method was constructed, with a total score of 100 points. The weights were precisely allocated to meet the core needs of karst remediation: heavy metal remediation performance 40 points, karst environment adaptability 30 points, growth stability 20 points, and ecological safety 10 points. Overall score = Heavy metal remediation performance score × 40% + Karst environment adaptability score × 30% + Growth stability score × 20% + Ecological safety score × 10%; Subjective experience weights are constructed using the analytic hierarchy process, and objective weights are corrected based on measured data using the entropy weight method. Set three-level standardized grading thresholds: A comprehensive score of ≥85 indicates a Class I typical dominant species; A score of 70-84 indicates a candidate species for secondary restoration. A species with a comprehensive score of <70 is eliminated.
[0017] As a preferred option, the core indicators of the heavy metal composite remediation performance include: the total amount of dominant heavy metals accumulated in the aboveground and underground parts of the plant, the bioaccumulation coefficient, and the translocation coefficient. The core indicators of adaptability to karst-specific environments include: the detected root activity, relative chlorophyll content, soluble sugar content, and proline content of the plants. The core indicators of long-term growth stability include: the relative growth rate of plant biomass, plant height growth rate, and overall survival rate during the domestication period, as statistically obtained. The core indicators of overall ecological security include: plant decomposition rate, total nitrogen and total phosphorus release during the decomposition process, as obtained through closed-loop testing.
[0018] Preferably, the concentration change sequences of all heavy metal species in the field test water, the environmental data change sequences, the coverage area change sequences and growth rate change sequences of dominant candidate species are input into a pre-trained prediction model based on a deep learning model to predict the remediation development trend, including: Training samples are constructed using historical case concentration change sequences, environmental data change sequences, dominant candidate species coverage area change sequences, growth change sequences, and real labels of restoration and development trends. A training set is constructed based on the training samples, and the LSTM model is trained using the training set until the loss function converges to obtain the prediction model. The expression for the prediction model is: The prediction model is built on an LSTM network, which performs well in processing time series data and is good at predicting performance state evolution. The LSTM unit includes an input gate, a forget gate, an output gate, and a unit state. The forget gate is represented as: in, Indicates a time step. This represents the input at the current time step, which is a feature vector constructed from the numerical values of the concentration change sequence, environmental data change sequence, dominant candidate species coverage area change sequence, and growth change sequence at time t. This represents the hidden state of the previous time step. The weight matrix represents the forget gate. This represents the Sigmoid activation function. Indicates the bias term. Indicates the output of the forget gate; The input gate is represented as: in, , These are the weight matrices for the input gate and the candidate neuron states, respectively. and These represent the bias terms for the input gate and the candidate neuron states, respectively. This represents the activation function. Indicates use for control The proportion parameter of the proportion of information stored. Indicates the status of the candidate cell; The cell state is represented as follows: in, It is the cell state of the previous time step. It represents the cell state at the current time step; The output gate is represented as: in, It is the weight matrix of the output gate. It is the bias term of the output gate. It is the output of the output gate. It is the output of the LSTM unit.
[0019] Compared to existing technologies, this invention offers the following advantages: The present invention provides a method for targeted screening of dominant aquatic plants for the remediation of heavy metal pollution in karst water. First, remote sensing image data of the target karst area is acquired. Based on this data, the community distribution data of each karst water body is determined. Field investigations and tests are conducted on each karst water body to identify the heavy metal with the highest concentration as the dominant heavy metal species. Karst water bodies with the same dominant heavy metal species and water body type are classified into one type of water environment. This invention uses dominant heavy metal species and karst water body type as the classification criteria for water environments. Since the dominant heavy metal species is the primary remediation target, it is necessary to identify aquatic remediation plants that are well-suited to the dominant heavy metal species. Water body types include pipe-cave karst water, cave-fissure karst water, fissure-pore karst water, etc. Different types of karst water will affect the growth status of aquatic plants. Therefore, it is necessary to identify aquatic plants suitable for each type of karst water, thereby making the targeted screening method of this invention more effective. This invention continuously monitors the concentration of dominant heavy metals in karst water bodies for each type of aquatic environment. Based on the concentration change data of dominant heavy metals, a preliminary assessment of remediation effectiveness is conducted to determine representative karst water bodies in the current aquatic environment. Target communities are identified based on the community distribution data of these representative karst water bodies. Field surveys are conducted to record the species included in the target communities, as well as the coverage area and biomass percentage of each species. The range of dominant aquatic plant species is determined based on the analysis of the recorded data. In the preliminary remediation effectiveness assessment stage, this invention first uses dominant heavy metals as the assessment object, thereby selectively screening out the range of dominant aquatic plant species for subsequent specific assessments and field tests. This invention further screens the range of dominant aquatic plant species to identify dominant candidate species. For each dominant candidate species, water samples are collected in the field according to the associated aquatic environment type. Based on the water samples, a karst-specific culture medium is artificially simulated, and the experimental plants of the dominant candidate species are subjected to stress acclimatization culture using the karst-specific culture medium. For the dominant candidate species that have undergone stress acclimatization culture, a comprehensive score is obtained through multi-dimensional weighted quantitative evaluation based on heavy metal composite remediation performance, karst-specific environment adaptability, long-term growth stability, and overall ecological safety. This invention establishes a multi-dimensional systematic evaluation system to systematically evaluate the dominant candidate species suitable for each type of aquatic environment.
[0020] For each type of water environment, field verification experiments are conducted on dominant candidate species whose comprehensive scores exceed a preset threshold. The concentration change sequences of all heavy metal species in the field test water, environmental data change sequences, and the coverage area and growth rate change sequences of dominant candidate species are input into a pre-trained prediction model based on deep learning to predict the remediation trend. The dominant candidate species with the best prediction results is determined as the optimal dominant species for the current water environment. This invention considers all heavy metals during the field testing phase, combining environmental data change sequences with the coverage area and growth rate change sequences of dominant candidate species to comprehensively assess the remediation effect. Thus, through standardized experimental methods, the optimal dominant species suitable for the current water environment is ultimately determined, achieving targeted, systematic, and standardized screening of the optimal dominant species adapted to various karst water environments. Attached Figure Description
[0021] Figure 1 A flowchart illustrating a method for targeted screening of superior aquatic plants for the remediation of heavy metal pollution in karst water, provided in an embodiment of the present invention. Detailed Implementation
[0022] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.
[0023] Combination Figure 1 This embodiment provides a method for targeted screening of advantageous aquatic plants for the remediation of heavy metal pollution in karst water, which includes: Step S1: Obtain remote sensing image data of the target karst area, and determine the community distribution data of each karst water body based on the remote sensing image data; conduct field investigation and testing on each karst water body and determine the heavy metal with the highest concentration as the dominant heavy metal species; classify karst water bodies with the same dominant heavy metal species and water body type into Class I water environment.
[0024] By integrating satellite multispectral remote sensing data and UAV hyperspectral remote sensing data, the invention provides full coverage of all karst water bodies in the target karst area. These karst water bodies are classified according to their water-bearing medium into pipe-cave karst water, cave-fissure karst water, and fissure-pore karst water. In actual operation, this invention generates an initial aquatic plant germplasm resource bank based on the community distribution data within each karst water body.
[0025] The community distribution data includes the distribution status of emergent, floating, and submerged plants; the community distribution data of each karst water body determined based on remote sensing image data includes: using a community classification model to extract the specific spectral features, texture features, and spatial distribution features of aquatic vegetation to distinguish emergent, floating, and submerged plant communities.
[0026] Emergent, floating, and submerged plant communities exhibit distinctly different physical morphologies in remote sensing images. Emergent plants (such as reeds and cattails) contain dual spectral signals from both vegetation and water. They have high reflectance in the near-infrared band but lower than that of purely terrestrial vegetation, with a coarser texture, obvious shadows and a three-dimensional effect. They exhibit high heterogeneity within patches and are mainly distributed in shallow coastal waters, appearing as stripes or rings with clear patch boundaries. Floating plants (such as water lilies and duckweed) show a "red-edge" effect in visible light reflectance, but the background water exhibits absorption valleys in the infrared band. The texture is relatively smooth, with characteristic circular or sheet-like textures. The patches are uniform inside and distributed on the open water surface in an aggregated sheet-like distribution. The patch shapes are irregular. Submerged plants (such as goldfish algae and Vallisneria natans) are severely affected by water absorption and scattering, and their spectral signals are relatively weak. In this embodiment, the weak texture generated by water flow is indirectly identified through water color remote sensing inversion. As an exemplary embodiment, this embodiment uses an object-oriented random forest algorithm combined with a deep learning CNN spectral classification model to extract the specific spectral features, texture features, and spatial distribution features of aquatic vegetation based on the different characteristics of emergent, floating, and submerged plant communities. This accurately distinguishes emergent, floating, and submerged plant communities and eliminates interference from terrestrial vegetation, artificial vegetation, and crops.
[0027] This invention uses the vegetation phenology patterns and the wet-dry cycle of water bodies in the target karst area as prior knowledge to input into the community classification model in order to optimize the model parameters. The community classification model can understand what plants should appear in certain areas at a specific time and at a specific water level, thereby automatically correcting and reducing misjudgments.
[0028] When conducting field investigations and tests on various karst water bodies and identifying the heavy metal with the highest concentration as the dominant heavy metal species, it is necessary to set up multiple sampling points at different depth gradients, and the concentration of the heavy metal detected is the average of the concentrations at each sampling point.
[0029] This invention uses dominant heavy metal species and karst water body type as the criteria for classifying aquatic environments. Dominant heavy metal species are the primary remediation target of this invention, so it is necessary to identify aquatic remediation plants that are best suited to dominant heavy metal species. Karst water bodies are classified according to their water-bearing medium into conduit-cave karst water, cave-fissure karst water, and fissure-pore karst water. Different water body types affect the growth status of aquatic plants. For example, cave-fissure karst water is usually found in underground seepage or narrow fissures, with extremely weak or even no light, making it virtually impossible for submerged and floating plants to survive; emergent plants only exist at the fissure outlets. Conduit-cave karst water is located in caves or underground rivers, also lacking light, resulting in very few aquatic plants; emergent plant communities only exist near the cave entrance or skylights. Given the characteristics of different karst water bodies, it is necessary to identify aquatic plants suitable for each type of karst water body, thereby making the targeted screening method of this invention more effective.
[0030] Step S2: For each type of water environment, continuously monitor the concentration of dominant heavy metals in karst water bodies, conduct preliminary assessment of remediation effectiveness based on the concentration change data of dominant heavy metals to determine the representative karst water bodies of the current water environment; determine the target community based on the community distribution data of the representative karst water bodies, conduct field surveys and record the species types covered by the target community as well as the coverage area and biomass proportion of each species, and analyze the recorded data to determine the range of dominant aquatic plant species.
[0031] This invention classifies water environments into multiple categories using remote sensing-based comprehensive surveys and heavy metal detection as prerequisites. Then, for each category, representative karst water bodies are identified based on continuously monitored data on the concentration changes of dominant heavy metals as typical samples with significant remediation effects. Subsequently, target communities are selected from these representative karst water bodies, and on-site investigations and records are conducted to achieve comprehensive screening of aquatic plant resources under complex karst landforms. This allows for the precise discovery of native high-potential remediation species, ensuring the comprehensiveness and typicality of germplasm selection.
[0032] When continuously monitoring the concentration of dominant heavy metals, this embodiment will deploy multiple sampling units and take the average value of all sampling units. Each sampling unit needs to cover multiple depth gradients.
[0033] The preliminary assessment of remediation effectiveness based on the concentration change data of dominant heavy metals to determine representative karst water bodies in the current aquatic environment includes: determining the daily average reduction and residual ratio of dominant heavy metal concentrations based on the concentration change data; the daily average reduction is expressed as (initial concentration - final concentration) ÷ number of monitoring days; the residual ratio is expressed as (final concentration ÷ initial concentration) × 100%; after normalizing the daily average reduction, a weighted average score is calculated using the residual ratio to obtain a preliminary remediation effectiveness assessment score; the karst water body with the highest preliminary remediation effectiveness assessment score is determined as the representative karst water body in the current aquatic environment. A higher degradation rate indicates a higher efficiency of heavy metal removal by plants in the karst water body, and a better remediation effect. A lower residual ratio indicates a more thorough removal of heavy metals in the karst water body, and a better remediation effect. Determining target communities based on the area coverage ratio of each community in a representative karst water body includes: identifying the communities with the highest pre-defined area coverage ratio among all communities as target communities; this invention mainly studies the dominant communities in each karst water body, which is more representative and can effectively avoid ineffective work.
[0034] Field surveys recorded the species diversity of the target community, as well as the area coverage and biomass percentage of each species, including: For the target communities, a combination of transect and quadrat method was used for verification. Emergent plant communities were laid out with 1m×1m quadrats, and floating and submerged plant communities were laid out with 0.5m×0.5m quadrats. 5 to 8 replicate quadrats were evenly distributed along each transect to ensure the representativeness of the survey. Each quadrat records all species and the percentage of coverage area and biomass of each species within the current quadrat. The range of dominant aquatic plant species was determined by analyzing the recorded data, including: for each quadrat, the compatibility of each species with the current dominant heavy metal species was determined based on the coverage area ratio and biomass ratio; in this embodiment, the compatibility was determined by a weighted calculation method of coverage area ratio and biomass ratio, and the species with the highest compatibility was selected for each quadrat to determine the range of dominant aquatic plant species.
[0035] During field surveys, species identification and verification are performed on-site for communities interpreted by remote sensing, correcting model errors and misjudgments. The percentage of species coverage and biomass in each community is recorded simultaneously. Furthermore, this invention conducts manual supplementary surveys in remote sensing blind spots such as karst fissures, hidden depressions, and pipe outlets to screen for less common and concealed aquatic plant resources. Due to the small sample size, determining the suitability of these less common and concealed aquatic plants for the aquatic environment is challenging; therefore, they can be used as supplementary dominant candidate species in subsequent field experiments.
[0036] Step S3: Screen the range of dominant aquatic plant species to determine dominant candidate species. For each dominant candidate species, collect water samples in the field according to the associated water environment type. Based on the water samples, artificially simulate karst-specific culture medium and use the karst-specific culture medium to conduct stress acclimatization culture on the test plants of the dominant candidate species. For the dominant candidate species that have undergone stress acclimatization culture, conduct a multi-dimensional weighted quantitative evaluation from the perspectives of heavy metal composite remediation performance, karst-specific environment adaptability, long-term growth stability and overall ecological safety to obtain a comprehensive score.
[0037] This invention establishes a three-tiered progressive initial screening criterion: ecological safety pre-screening, basic stress tolerance screening, and growth controllability screening. This criterion eliminates unqualified germplasm in batches, identifies candidate germplasm suitable for karst basic habitats, and provides a standard basis for subsequent precise targeted sampling.
[0038] The screening of dominant aquatic plant species to identify candidate dominant species includes pre-ecological safety screening, basic stress tolerance screening, and growth controllability screening. The pre-ecological safety screening retains native species of the southwestern karst landscape while eliminating invasive species, artificially hybridized species, and rare and protected species, ensuring that candidate species can integrate into the native karst ecological community without the risk of biological invasion. This invention screens typical native karst species, strictly controlling the risks of biological invasion and secondary pollution from plant decomposition. The remediation process maintains the structure of the native karst ecological community, achieving a synergistic unity between pollution remediation and ecological protection. In another specific embodiment, other species can be tested in certain experimental areas to obtain dominant species with the best remediation effect, strong reproductive capacity, and low maintenance cost. However, the experimental areas need to be conducted in independent environments to avoid uncontrolled species growth. The basic stress tolerance screening eliminates vulnerable species that are sensitive to high-calcium and alkaline water environments, have a natural habitat survival rate of less than 85%, or exhibit significant calcium stress and acid-base stress damage; only germplasm that can naturally adapt to the basic water quality of karst and have stable physiological states is retained. Growth controllability screening eliminates invasive species that reproduce explosively, spread too quickly, and are difficult to manage. It also eliminates unstable species that have short growth cycles, severe seasonal wilting, complete winter dormancy, and cannot achieve continuous restoration throughout the year.
[0039] After screening is completed, the present invention can form a standardized candidate germplasm list that is ecologically safe, trait-stable, habitat-suitable, and highly controllable. This list can serve as the target species for subsequent targeted field sampling, effectively avoiding the problems of invalid sampling and sampling of vulnerable species, and significantly improving screening efficiency and accuracy.
[0040] When collecting water samples in the field according to the associated water environment type, it is necessary to test the water pH, calcium ion concentration, water hardness, dissolved oxygen, and water temperature on-site to ensure that the sampled germplasm is native, ecologically safe, and suitable for the basic karst habitat. After sampling, the types and concentrations of heavy metals in the water are measured in the laboratory to finally determine the water samples that meet the current water environment and can be used for artificial simulation of karst-specific culture solutions, providing a basis for real environmental parameters for subsequent simulation domestication and index screening.
[0041] Based on water samples, a karst-specific culture medium was artificially simulated. This medium was used to conduct stress acclimatization training on experimental plants of dominant candidate species. The process included standardized culture medium preparation, standardized plant pretreatment, and a three-stage gradient acclimatization cycle. The standardized culture medium preparation was based on water quality testing results from field-collected water samples, artificially simulating the karst-specific culture medium. As an exemplary embodiment, this invention controls the pH value to 7.8–8.3 and the calcium ion concentration to 120–180 mg / L, precisely proportioning the concentrations of cadmium, lead, and zinc, with the dominant heavy metal concentration being prominent, at least 30%–50% higher than other heavy metal concentrations, completely replicating the core water quality characteristics of karst-polluted water bodies in Southwest China. Standardized pretreatment of plants involved uniformly cleaning the roots of all experimental plants to remove silt, pruning damaged branches and leaves, and standardizing plant height and biomass specifications to eliminate the interference of individual differences on the experimental results. All plants were then uniformly planted in artificial cultivation devices. The three-stage gradient acclimatization cycle lasted a total of 45 days, divided into three phases of gradient stress cultivation: The first phase was a 15-day low-stress adaptation period, where heavy metal stress was weakened, and no or only very low concentrations of dominant heavy metals were added, allowing the plants to gradually adapt to a high-calcium, slightly alkaline basic water environment; the second phase was a 20-day moderate constant stress period, during which multiple heavy metal pollutants were added, with the amount of the dominant heavy metal added being greater than that of other heavy metals; maintaining a high-calcium, slightly alkaline + multiple metal composite stress environment activated the plant's stress resistance mechanism and heavy metal accumulation potential; the third phase was a 10-day stabilization acclimatization period, where the stress intensity was no longer increased, maintaining the current environmental conditions, stabilizing the plant's metabolic level and accumulation performance. The cultivation process simultaneously simulated the natural characteristics of karst waters, with diurnal temperature variations of 6-10℃ and intermittent hydrological alternations of wet and dry periods.
[0042] In addition, during the acclimatization period, the root system status, leaf morphology, and photosynthetic status of the plants were continuously monitored. After the acclimatization was completed, the results of physiological index testing were combined to remove the germplasm with root rot, yellowing leaves, stunted growth, wilting and decay, and weak stress resistance, and retain the high-quality candidate plants with stable growth, strong stress tolerance, and no obvious stress damage.
[0043] A four-dimensional quantitative evaluation system was constructed to assess the performance of heavy metal composite remediation, adaptability to karst environment, long-term growth stability, and local ecological safety in karst habitats in Southwest China. All indicators were quantitatively tested using standardized experimental methods.
[0044] A comprehensive score was obtained through a multi-dimensional weighted quantitative evaluation of heavy metal composite remediation performance, adaptability to karst-specific environments, long-term growth stability, and overall ecological safety. A combined subjective and objective scoring model using the Analytic Hierarchy Process (AHP) and the entropy weight method was constructed, with a total score of 100 points. The weights were precisely allocated to meet the core needs of karst remediation: heavy metal remediation performance 40 points, karst environment adaptability 30 points, growth stability 20 points, and ecological safety 10 points. Overall score = Heavy metal remediation performance score × 40% + Karst environment adaptability score × 30% + Growth stability score × 20% + Ecological safety score × 10%; By constructing subjective experience weights using the analytic hierarchy process (AHP) and correcting objective weights based on measured data using the entropy weight method, the evaluation results are made accurate, fair, and traceable.
[0045] Set three-level standardized grading thresholds: A comprehensive score of ≥85 indicates a first-class typical dominant species, which combines local typicality, high environmental adaptability, high efficiency in restoration and ecological safety, and can be directly applied on a large scale in engineering. A comprehensive score of 70-84 indicates a secondary candidate species for restoration, which can be used as auxiliary restoration germplasm. Species with a comprehensive score of <70 are eliminated because their comprehensive performance does not meet the requirements for karst remediation. Finally, a set of typical and advantageous species for karst water remediation is selected.
[0046] This invention employs a four-dimensional quantitative index combined with a subjective and objective scoring model, which completely avoids the one-sidedness of single-index screening and the subjectivity of manual evaluation, taking into account restoration function, environmental adaptability, growth stability and ecological safety, and the screening results are in line with the actual application needs of engineering.
[0047] The core indicators of heavy metal composite remediation performance include: the total enrichment of dominant heavy metals in the aboveground and underground parts of the plant, the bioaccumulation coefficient, and the translocation coefficient. A weighted sum of these core indicators yields the heavy metal composite remediation performance score, quantitatively assessing the plant's ability to absorb, accumulate, and migrate karst metal pollution, and screening functional species with high enrichment, high translocation, and low residual risk. Specifically, the aboveground enrichment total = aboveground heavy metal concentration (mg / kg) × aboveground biomass (kg); the underground enrichment total = underground heavy metal concentration (mg / kg) × underground biomass (kg); the aboveground bioaccumulation coefficient = aboveground heavy metal concentration (mg / kg) ÷ heavy metal concentration in water / sediment (mg / kg); the underground bioaccumulation coefficient = underground heavy metal concentration (mg / kg) ÷ heavy metal concentration in water / sediment (mg / kg); and the translocation coefficient = aboveground heavy metal concentration (mg / kg) ÷ underground heavy metal concentration (mg / kg).
[0048] The core indicators of karst-specific environmental adaptability include: root activity, relative chlorophyll content, soluble sugar content, and proline content. A weighted sum of these core indicators yields the karst environment adaptability score. This invention quantifies the plant's ability to regulate osmotic pressure, photosynthetic stability, and long-term stress tolerance to high calcium stress and alkaline water environments through four major stress-resistance physiological indicators. Among these, root activity (mg TTF·g) is a key indicator. - ¹·h -¹) = (TTC reduction amount) / (root fresh weight × reaction time); relative chlorophyll content was directly read by a portable chlorophyll meter (SPAD meter); soluble sugar content (mg·g) - ¹) = (C × V) S ) / (V1×W × 1000); C represents the glucose concentration (μg / mL) obtained from the standard curve; V S V1 represents the total volume of the sample extract (mL); W represents the volume of sample liquid taken during color development (mL); W represents the fresh weight of the sample (g); and proline content (μg·g). - ¹) = (C × Vp) / (V1 × W); C represents the proline concentration (μg / mL) obtained from the standard curve; Vp represents the total volume of the extract (mL); V1 represents the volume of the sample liquid used in the determination (mL); W represents the fresh weight of the sample (g).
[0049] The core indicators of long-term growth stability include the relative growth rate of plant biomass, plant height growth rate, and overall survival rate during the domestication period. A weighted sum of these core indicators yields a growth stability score, assessing the plant's growth persistence and trait stability under long-term combined stress of high calcium and alkalinity and heavy metals, thus avoiding pseudo-dominant species that exhibit short-term superior growth followed by rapid decline. The relative growth rate of plant biomass (RGR) is calculated as: RGR = (ln W2 - ln W1) / (t2 - t1), where W1 is the initial biomass (dry weight), W2 is the biomass after domestication (dry weight), and t1 and t2 are the corresponding time points (days), representing the net increase in biomass per unit time, in g·g. - ¹·d - ¹; Plant height growth rate = (H2- H1) / (t2- t1); where H1 is the initial plant height, H2 is the final plant height, and t1 and t2 are the corresponding times; Overall survival rate = (number of surviving individuals / initial total number of individuals) × 100%, at the end of the experiment, the number of surviving plants is counted, divided by the initial total number of plants, and multiplied by 100%.
[0050] The core indicators of overall ecological security include the plant decomposition rate, total nitrogen and total phosphorus release during decomposition, obtained through closed-system testing. The ecological security score is obtained by weighted summation of multiple core indicators to quantitatively assess the risk of secondary pollution from eutrophication. Combined with the species' local community succession characteristics, the ecological adaptability of the species is verified to ensure ecological security, no invasion risk, and no secondary pollution during the restoration process. Among them, the plant decomposition rate (% / d) = (W0 - Wt) / (W0 × t) × 100%; where W0 is the initial dry weight (g), Wt is the remaining dry weight (g) after t days of decomposition, and t is the number of days of decomposition; the release amount (mg) = initial plant total nitrogen / total phosphorus content (mg) - residual plant total nitrogen / total phosphorus content (mg).
[0051] Step S4: For each type of water environment, conduct field verification experiments on the dominant candidate species whose comprehensive score is greater than the preset score threshold. Input the recorded concentration change sequence of all heavy metal species in the field test water, the environmental data change sequence, the coverage area change sequence and growth change sequence of the dominant candidate species into a pre-trained prediction model based on a deep learning model to predict the remediation and development trend. The dominant candidate species with the best prediction result is determined as the best dominant species in the current water environment. The environmental data includes pH value (water acidity and alkalinity, reflecting the alkaline characteristics of karst water), calcium ion concentration (reflecting the core indicator of high calcium stress in karst), water hardness (referring to total hardness, comprehensively reflecting the calcium and magnesium ion content), dissolved oxygen (reflecting the water's self-purification capacity and the influence of plant metabolism), and water temperature.
[0052] The concentration change sequences of all heavy metal species in the field test water, the environmental data change sequences, the coverage area change sequences of dominant candidate species, and the growth rate change sequences are input into a pre-trained prediction model based on deep learning to predict the remediation development trend, including: Training samples are constructed using historical case concentration change sequences, environmental data change sequences, dominant candidate species coverage area change sequences, growth change sequences, and ground truth labels of remediation development trends. A training set is then built based on these training samples, and the LSTM model is trained using this training set until the loss function converges to obtain the prediction model. The historical cases are taken from field verification data recorded in several previous karst heavy metal pollution remediation projects. As an exemplary embodiment, the dataset contains, for example, more than 50 completed project cases covering different karst hydrogeological types and different dominant heavy metal (such as Cd, Pb, Zn) pollution scenarios in Southwest China. The data included in the training set all include continuous monitoring data from the start to the end of the project (usually 3-6 months) and have a clear final remediation development trend as the ground truth label.
[0053] The true labels for the repair development trend are categorized into high-efficiency repair, effective repair, and ineffective repair based on the overall reduction rate; the overall reduction rate is calculated as follows: Label=∑(W i ×R i ), W i and R i R represents the weighting coefficient and reduction rate of the i-th heavy metal, respectively; i =(C 初始 C 末期 ) / C 初始×100%; the final period can be set to 30 days or 60 days. An overall reduction rate >60% is marked as a highly effective repair (the actual label is recorded as label 2); an overall reduction rate between 20% and 60% is marked as an effective repair (the actual label is recorded as label 1); an overall reduction rate <20% is marked as an ineffective repair (the actual label is recorded as label 0).
[0054] The prediction model is built on an LSTM network, which performs well in processing time-series data and is good at predicting state evolution. The LSTM unit includes an input gate, a forget gate, an output gate, and a unit state.
[0055] The forget gate is represented as: in, Indicates a time step. This represents the input at the current time step, which is a feature vector constructed from the numerical values of the concentration change sequence, environmental data change sequence, dominant candidate species coverage area change sequence, and growth change sequence at time t. This represents the hidden state of the previous time step. The weight matrix represents the forget gate. This represents the Sigmoid activation function. Indicates the bias term. Indicates the output of the forget gate; The input gate is represented as: in, , These are the weight matrices for the input gate and the candidate neuron states, respectively. and These represent the bias terms for the input gate and the candidate neuron states, respectively. This represents the activation function. Indicates use for control The proportion parameter of the proportion of information stored. Indicates the status of the candidate cell; The cell state is represented as follows: in, It is the cell state of the previous time step. It represents the cell state at the current time step; The output gate is represented as: in, It is the weight matrix of the output gate. It is the bias term of the output gate. It is the output of the output gate. It is the output of the LSTM unit.
[0056] This invention specifically simulates the high-calcium, alkaline aquatic environment of karst waters in Southwest China. Through gradient stress acclimatization, it achieves targeted adaptation of germplasm. The selected dominant species effectively resist karst calcium stress, acid-base stress, and heavy metal stress, exhibiting high in-situ survival rates and stable remediation efficacy. This effectively solves the problems of poor adaptability and rapid efficacy decay associated with traditional screening methods, improving remediation efficiency by more than 30% compared to traditional techniques. The entire process of this invention is standardized, with quantified parameters and good repeatability, making it widely adaptable to germplasm screening in karst waters of different regions and pollution levels in Southwest China. It provides a standardized technical system for large-scale ecological remediation of heavy metal pollution in karst waters, demonstrating extremely high engineering application and promotion value. The selected native dominant species have broad adaptability, strong reproductive capacity, and low maintenance costs. Compared to chemical remediation and constructed wetlands, this significantly reduces the cost of treating heavy metal pollution in karst waters, making it suitable for large-scale, routine application.
[0057] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0058] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0059] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0060] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0061] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.
Claims
1. A method for directional screening of superior aquatic plants for remediation of heavy metal pollution in karst water, characterized in that, include: Acquire remote sensing image data of the target karst area, and determine the community distribution data of each karst water body based on the remote sensing image data; Field investigations and tests were conducted on each karst water body to identify the heavy metal with the highest concentration as the dominant heavy metal species; karst water bodies with the same dominant heavy metal species and water body type were classified into Class I water environment. For each type of water environment, the concentration of dominant heavy metals in karst water bodies is continuously monitored. Based on the concentration change data of dominant heavy metals, a preliminary assessment of the remediation effectiveness is conducted to determine the representative karst water bodies of the current water environment. Based on the community distribution data of the representative karst water bodies, the target community is determined. Field surveys are conducted to record the species types covered by the target community, as well as the coverage area and biomass of each species. Based on the analysis of the recorded data, the range of dominant aquatic plant species is determined. The range of dominant aquatic plant species was screened to identify dominant candidate species. For each dominant candidate species, water samples were collected in the field according to the associated aquatic environment type. Based on the water samples, a karst-specific culture medium was artificially simulated, and the experimental plants of the dominant candidate species were subjected to stress acclimatization culture using the karst-specific culture medium. For the dominant candidate species that underwent stress acclimatization culture, a comprehensive score was obtained through multi-dimensional weighted quantitative evaluation based on heavy metal composite remediation performance, karst-specific environment adaptability, long-term growth stability, and overall ecological safety. For each type of water environment, field verification experiments are conducted on the dominant candidate species whose comprehensive scores are greater than the preset score threshold. The concentration change sequences of all heavy metal species in the field test water, the environmental data change sequences, the coverage area change sequences and growth change sequences of the dominant candidate species are input into a pre-trained prediction model based on a deep learning model to predict the remediation development trend. The dominant candidate species with the best prediction results is determined as the best dominant species for the current water environment.
2. The method for karst water heavy metal pollution remediation of the dominant aquatic plant directional screening according to claim 1, characterized in that, The community distribution data of each karst water body determined based on remote sensing image data includes: The community distribution data includes the distribution status of emergent, floating, and submerged plants; An object-oriented random forest algorithm combined with a deep learning CNN spectral classification model is used to extract the specific spectral features, texture features, and spatial distribution features of aquatic vegetation to distinguish emergent, floating, and submerged plant communities. The vegetation phenology and the wet-dry cycle of water bodies in the target karst area are used as prior knowledge inputs into the community classification model to optimize the model parameters.
3. The method for karst water heavy metal pollution remediation of the dominant aquatic plant directional screening according to claim 1, characterized in that, A preliminary assessment of remediation effectiveness was conducted based on data on changes in the concentration of dominant heavy metals to identify representative karst water bodies in the current aquatic environment, including: The daily average reduction and residue ratio of the dominant heavy metal concentration are determined based on the concentration change data. The daily average reduction is expressed as: (initial concentration - final concentration) ÷ number of monitoring days; the residue ratio is expressed as: (final concentration ÷ initial concentration) × 100%. After normalizing the daily reduction, a weighted average score for the remediation effectiveness is calculated by combining the reduction with the residual ratio. The karst water body with the highest preliminary remediation effectiveness score is identified as the representative karst water body of the current water environment. Based on the community distribution data of representative karst water bodies, the target communities were identified as including: The communities with the highest area coverage ratio among all communities are identified as target communities.
4. The method for karst water heavy metal pollution remediation of the dominant aquatic plant directional screening according to claim 1, characterized in that, Field surveys recorded the species diversity of the target community, as well as the area coverage and biomass percentage of each species, including: For the target communities, the transect-quad plot method was used for verification. Emergent plant communities were laid out with 1m×1m quadrats, and floating and submerged plant communities were laid out with 0.5m×0.5m quadrats. 5 to 8 replicate quadrats were evenly laid out along each transect. Each quadrat records all species and the percentage of coverage area and biomass of each species within the current quadrat. Based on the analysis of recorded data, the range of dominant aquatic plant species included: For each quadrat, the compatibility of each species with the current dominant heavy metal species is determined based on the coverage area ratio and biomass ratio; the species with the highest compatibility is selected for each quadrat to determine the range of dominant aquatic plant species.
5. The method for karst water heavy metal pollution remediation of the dominant aquatic plant directional screening according to claim 1, characterized in that, Screening the range of dominant aquatic plant species to identify dominant candidate species includes: Ecological security screening is conducted in advance to preserve native species in the karst landscape of Southwest China, while eliminating invasive species, artificially hybridized and improved species, and rare and protected species. Basic stress tolerance screening was conducted to eliminate vulnerable species that are sensitive to high-calcium and alkaline water environments, have a natural habitat survival rate of less than 85%, or suffer from obvious calcium stress and acid-base stress damage. Growth controllability screening is conducted to eliminate invasive species that reproduce explosively and spread too quickly, as well as unstable species with short growth cycles, severe seasonal wilting, and complete winter dormancy.
6. The method for karst water heavy metal pollution remediation of the dominant aquatic plant directional screening according to claim 1, characterized in that, When collecting water samples in the field according to the associated water environment type, it is necessary to test the water's pH value, calcium ion concentration, water hardness, dissolved oxygen, and water temperature on-site. After sampling, the types and concentrations of heavy metals in the water are determined in the laboratory to finally determine the water samples that meet the current water environment and can be used for artificial simulation of karst-specific culture solutions.
7. The method for karst water heavy metal pollution remediation of the dominant aquatic plant directional screening according to claim 1, characterized in that, Based on water samples, artificially simulated karst-specific culture solutions were used to conduct stress acclimatization culture on test plants of dominant candidate species using these solutions, including: Standardized culture medium preparation: Based on water quality test results, artificial simulation of karst-specific culture medium is constructed. Standardized pretreatment of plants: all experimental plants were uniformly cleaned of root mud, damaged branches and leaves were pruned, plant height and biomass were uniformly specified, and they were uniformly planted in artificial culture devices. The three-stage gradient acclimatization cycle, with a total acclimatization time of 45 days, is divided into three stages of gradient stress cultivation: the first stage is a 15-day low-stress adaptation period, during which heavy metal stress is weakened, and no or only very low concentrations of dominant heavy metals are added; the second stage is a 20-day moderate constant stress period, during which multiple heavy metal pollutants are added, with the amount of the dominant heavy metal added being greater than that of other heavy metals; maintaining a high-calcium, slightly alkaline, and multi-metal complex stress environment; the third stage is a 10-day stabilization acclimatization period, during which the plant's metabolic level and accumulation performance are stabilized, and the cultivation process simultaneously simulates the natural characteristics of karst waters, with a diurnal temperature range of 6-10℃ and intermittent hydrological wet-dry alternation.
8. The method for karst water heavy metal pollution remediation of the dominant aquatic plant directional screening according to claim 1, characterized in that, A comprehensive score was obtained through a multi-dimensional weighted quantitative evaluation of heavy metal composite remediation performance, adaptability to karst-specific environments, long-term growth stability, and overall ecological safety. A combined subjective and objective scoring model using the Analytic Hierarchy Process (AHP) and the entropy weight method was constructed, with a total score of 100 points. The weights were precisely allocated to meet the core needs of karst remediation: heavy metal remediation performance 40 points, karst environment adaptability 30 points, growth stability 20 points, and ecological safety 10 points. Overall score = Heavy metal remediation performance score × 40% + Karst environment adaptability score × 30% + Growth stability score × 20% + Ecological safety score × 10%; Subjective experience weights are constructed using the analytic hierarchy process, and objective weights are corrected based on measured data using the entropy weight method. Set three-level standardized grading thresholds: A comprehensive score of ≥85 indicates a Class I typical dominant species; A score of 70-84 indicates a candidate species for secondary restoration. A species with a comprehensive score of less than 70 is considered to be eliminated.
9. The method for karst water heavy metal pollution remediation of the superior aquatic plant directional screening according to claim 1, characterized in that, The core indicators of heavy metal composite remediation performance include: the total amount of dominant heavy metals accumulated in the aboveground and underground parts of the plant, the bioaccumulation coefficient, and the translocation coefficient. The core indicators of adaptability to karst-specific environments include: the detected root activity, relative chlorophyll content, soluble sugar content, and proline content of the plants. The core indicators of long-term growth stability include: the relative growth rate of plant biomass, plant height growth rate, and overall survival rate during the domestication period, as statistically obtained. The core indicators of overall ecological security include: plant decomposition rate, total nitrogen and total phosphorus release during the decomposition process, as obtained through closed-loop testing.
10. The method for karst water heavy metal pollution remediation of the dominant aquatic plant directional screening according to claim 1, characterized in that, The concentration change sequences of all heavy metal species in the field test water, the environmental data change sequences, the coverage area change sequences of dominant candidate species, and the growth rate change sequences are input into a pre-trained prediction model based on deep learning to predict the remediation development trend, including: Training samples are constructed using historical case concentration change sequences, environmental data change sequences, dominant candidate species coverage area change sequences, growth change sequences, and real labels of restoration and development trends. A training set is constructed based on the training samples, and the LSTM model is trained using the training set until the loss function converges to obtain the prediction model. The expression for the prediction model is: The prediction model is built on an LSTM network, which performs well in processing time series data and is good at predicting performance state evolution. The LSTM unit includes an input gate, a forget gate, an output gate, and a unit state. The forget gate is represented as: in, Indicates a time step. This represents the input at the current time step, which is a feature vector constructed from the numerical values of the concentration change sequence, environmental data change sequence, dominant candidate species coverage area change sequence, and growth change sequence at time t. This represents the hidden state of the previous time step. The weight matrix represents the forget gate. This represents the Sigmoid activation function. Indicates the bias term. Indicates the output of the forget gate; The input gate is represented as: in, , These are the weight matrices for the input gate and the candidate neuron states, respectively. and These represent the bias terms for the input gate and the candidate neuron states, respectively. This represents the activation function. Indicates use for control The proportion parameter of the proportion of information stored. Indicates the status of the candidate cell; The cell state is represented as follows: in, It is the cell state of the previous time step. It represents the cell state at the current time step; The output gate is represented as: in, It is the weight matrix of the output gate. It is the bias term of the output gate. It is the output of the output gate. It is the output of the LSTM unit.
