CN114862249A - River basin non-point source pollution prevention and control method and system based on key landscape indexes - Google Patents
River basin non-point source pollution prevention and control method and system based on key landscape indexes Download PDFInfo
- Publication number
- CN114862249A CN114862249A CN202210577988.2A CN202210577988A CN114862249A CN 114862249 A CN114862249 A CN 114862249A CN 202210577988 A CN202210577988 A CN 202210577988A CN 114862249 A CN114862249 A CN 114862249A
- Authority
- CN
- China
- Prior art keywords
- landscape
- index
- source pollution
- point source
- indexes
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A20/00—Water conservation; Efficient water supply; Efficient water use
- Y02A20/152—Water filtration
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Theoretical Computer Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a river basin non-point source pollution prevention and control method and system based on key landscape indexes, which comprises the following steps: collecting various non-point source pollution indexes of each sampling point in a target flow field; acquiring various landscape indexes at each sampling point in a target flow domain; inputting the same non-point source pollution indexes of all sampling points and corresponding landscape indexes into a random forest model, screening out key landscape indexes of the non-point source pollution indexes, and applying a non-parameter deviation reduction algorithm and a bootstrap method to obtain mutation threshold data of the key landscape indexes; and optimizing the landscape pattern according to the response relation between the key landscape index and the non-point source pollution index and the mutation threshold value of the key landscape index, and realizing the prevention and control of the non-point source pollution. By screening key landscape indexes affecting the non-point source pollution in the drainage basin and determining the mutation threshold of the key landscape indexes, the landscape pattern of the target drainage basin is optimized more accurately, and the non-point source pollution in the drainage basin is prevented and treated accurately and efficiently.
Description
Technical Field
The invention relates to the field of river basin non-point source pollution prevention and control, in particular to a river basin non-point source pollution prevention and control method and system based on key landscape indexes.
Background
The data of the second national pollution source census official gazette shows that the non-point source pollution becomes the main pollution source of the water body in China and is an important factor restricting the economic development. The unreasonable landscape configuration accelerates the loss of nutrient substances such as nitrogen, phosphorus and the like to be the main cause of non-point source pollution. As the natural landscape configuration in the river basin is greatly changed by human activities, the problem of non-point source pollution is increasingly prominent, and the regulation and control of the spatial configuration of different types of plaques from the perspective of landscape planning is the key for improving the non-point source pollution treatment effect and maintaining the long-term effective treatment result. The reasonable landscape pattern planning has extremely high potential for preventing and controlling non-point source pollution, protecting water quality health and guaranteeing economic development.
However, researchers mainly refer to relevant documents to select landscape indexes and quantify landscape patterns to qualitatively analyze the relation between landscape patterns and non-point source pollution based on self knowledge of the landscape. Researchers know that the landscape pattern has angle and scale difference, and have long-term debate on the selection of key landscape indexes. And due to the nonlinear response of landscape indexes to the surface source pollution, the regulation and control threshold values of different landscape indexes of the drainage basin are different, but the screening of the landscape indexes and the setting of the threshold values in the prior art are low in accuracy and large in limitation, so that an effective and reliable standard system cannot be established for the surface source pollution prevention and control of the drainage basin. Therefore, the invention provides a drainage basin non-point source pollution prevention and control method and system based on key landscape indexes.
Disclosure of Invention
The invention aims to provide a watershed non-point source pollution prevention and control method and system based on key landscape indexes, which can accurately screen out the key landscape indexes and quantify mutation thresholds causing water quality deterioration, and provide specific and quantifiable design basis for future landscape sustainability planning, so that the watershed non-point source pollution is effectively prevented and controlled.
In order to achieve the purpose, the invention provides the following scheme:
a watershed non-point source pollution prevention and control method based on key landscape indexes comprises the following steps:
collecting various non-point source pollution indexes of each sampling point in a target drainage basin;
acquiring various landscape indexes at each sampling point in the target flow domain, and recording the various landscape indexes as a landscape index set; each non-point source pollution index of all the sampling points corresponds to the landscape index set;
for the non-point source pollution indexes with index values exceeding a preset value, respectively inputting the same non-point source pollution indexes of all the sampling points and the corresponding landscape index sets into a random forest model, and screening out key landscape indexes corresponding to each non-point source pollution index by combining a significance test method;
applying a nonparametric deviation reduction algorithm and a bootstrap method to each non-point source pollution index and the corresponding key landscape index to obtain a mutation threshold of each key landscape index;
and for each non-point source pollution index, designing a landscape planning scheme by taking the mutation threshold of the key landscape index as a standard based on the relationship between the key landscape index and the non-point source pollution index.
A watershed non-point source pollution prevention and control system based on key landscape indexes, the system comprises:
the non-point source pollution index acquisition module is used for acquiring various non-point source pollution indexes of each sampling point in the target drainage basin;
the landscape index acquisition module is used for acquiring various landscape indexes at each sampling point in the target flow domain and recording the various landscape indexes as a landscape index set; each non-point source pollution index of all the sampling points corresponds to the landscape index set;
the key landscape index screening module is used for inputting the same surface source pollution indexes of all the sampling points and the corresponding landscape index sets into a random forest model respectively for surface source pollution indexes with index values exceeding a preset value and screening out the key landscape indexes corresponding to each surface source pollution index by combining a significance test method;
the key landscape index threshold value calculation module is used for applying a nonparametric deviation reduction algorithm and a bootstrap method to each non-point source pollution index and the corresponding key landscape index to obtain a mutation threshold value of the key landscape index;
and the landscape pattern optimization module adjusts the landscape pattern in the target flow domain according to the mutation threshold data of each key landscape index corresponding to each non-point source pollution index and the response relation between each non-point source pollution index and each corresponding key landscape index, so as to realize the prevention and control of non-point source pollution.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention relates to a river basin non-point source pollution prevention and control method and system based on key landscape indexes, which comprises the following steps: collecting various non-point source pollution indexes of each sampling point in a target drainage basin; acquiring various landscape indexes at each sampling point in a target flow domain, and recording the various landscape indexes as a landscape index set; for each non-point source pollution index, inputting the same non-point source pollution index of all sampling points and a corresponding landscape index set into a random forest model, and screening out a key landscape index corresponding to each non-point source pollution index by combining a significance test method; applying a nonparametric deviation reduction algorithm and a bootstrap method to each non-point source pollution index and each corresponding key landscape index to obtain mutation threshold data of the key landscape index corresponding to each non-point source pollution index; and adjusting the landscape pattern in the target watershed according to the response relation between the non-point source pollution index and the corresponding key landscape index and the mutation threshold data of the corresponding key landscape index, so as to realize the prevention and control of the non-point source pollution of the watershed. The method comprises the steps of screening key landscape indexes affecting the non-point source pollution in the drainage basin by using a random forest algorithm, and determining mutation thresholds of the key landscape indexes by using a nonparametric deviation reduction algorithm and a bootstrap method, so that the landscape pattern of a target drainage basin is optimized according to the mutation thresholds of the key landscape indexes, and the non-point source pollution in the drainage basin is accurately and efficiently prevented.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a flow chart of a drainage basin non-point source pollution prevention and control method based on key landscape indexes provided in embodiment 1 of the present invention;
fig. 2 is a block diagram of a drainage basin non-point source pollution prevention and control system based on key landscape indexes, provided in embodiment 2 of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Over 200 landscape indexes are available to describe landscape patterns in terms of fragmentation, shape, connectivity, diversity, etc. However, from the perspective of practical planning, a few indexes can comprehensively evaluate the watershed landscape pattern characteristics. Therefore, a widely applicable index screening method is developed, key landscape indexes with a function of regulating the non-point source pollutants are selected in a target watershed (a riparian zone scale and/or a sub-watershed scale), the mutation threshold value of water quality change caused by the key landscape indexes is quantified, the process of quantifying the mutation threshold value of the landscape indexes is popularized into a general research model, scientific basis is provided for watershed landscape planning, and the method becomes a key technology for preventing and treating non-point source pollution of the watershed.
The invention aims to provide a method and a system for preventing and controlling the river basin non-point source pollution based on key landscape indexes, which can accurately screen out the key landscape indexes which obviously influence the river basin non-point source pollution and quantify mutation thresholds causing water quality deterioration, and provide a specific and quantifiable design scheme for future landscape sustainability planning, thereby effectively preventing and controlling the river basin non-point source pollution.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Example 1
As shown in fig. 1, this embodiment 1 provides a method for preventing and controlling drainage basin non-point source pollution based on key landscape indicators, where the method includes:
s1: collecting various non-point source pollution indexes of each sampling point in a target drainage basin;
setting a plurality of sampling points in the target basin, collecting surface water bodies of the basin at different periods, and screening indexes with surface source pollution by referring to a surface water environment quality standard (GB 3838-2002); in order to guarantee the accuracy of subsequent calculation, the accumulated sampling is required to be more than 50.
S2: acquiring various landscape indexes at each sampling point in the target flow domain, and recording the various landscape indexes as a landscape index set; the non-point source pollution indexes of all sampling points correspond to the landscape index set;
considering that there is a scale effect on the influence of landscape patterns on the area source pollution in the target watershed, in order to realize landscape pattern optimization of the target watershed, the target watershed may be scaled, for example, into sub-watersheds and a bank zone. Setting the sampling points as water outlets to divide sub-watersheds in Arcgis software, selecting a proper buffer distance to divide a riparian zone according to the size of the watersheds, selecting a buffer area of 100m as the riparian zone, and adjusting the value according to actual requirements.
Inputting the scale division result of the target watershed into 109 commonly used landscape indexes such as shape, connectivity, edge density and the like under a calculation type scale in Fragstatss 4.2, and obtaining a riparian zone landscape index set and a sub-watershed landscape index set which respectively represent the riparian zone scale and the landscape pattern of the sub-watershed scale.
S3: inputting the same surface source pollution indexes of all sampling points and corresponding landscape index sets into a random forest model for surface source pollution indexes (namely surface source pollution indexes with pollution) with index values exceeding a preset value, and screening out key landscape indexes corresponding to the surface source pollution indexes by combining a significance inspection method; the preset value can be determined according to the national standard of the non-point source pollution index.
The area source indexes (the embodiment takes total nitrogen TN as an example for explanation, and has no limitation effect) screened in the step S1 and the corresponding landscape indexes calculated in the step S2 are input into a random forest model, and the importance of different landscape indexes on the change of each area source pollution index (total nitrogen TN) in the current domain is evaluated. And (3) taking the mean square error (% IncMSE) of the average increase of the model parameters as a measurement index, quantifying the influence degree of landscape indexes with different scales on the change of the river basin TN, and performing significance test.
If the target drainage basin is divided into a sub drainage basin and a riparian zone, respectively inputting each non-point source pollution index and a corresponding landscape index in the sub drainage basin or the riparian zone into a random forest model to obtain each non-point source pollution index and a key landscape index in the sub drainage basin and each non-point source pollution index and the key landscape index in the riparian zone.
Step S3 specifically includes:
step S31: combining the indexes of all sampling points and the corresponding landscape index sets to obtain an original sample data set of the non-point source pollution indexes for the non-point source indexes with pollution; one non-point source pollution index and the corresponding landscape index set of each sampling point are taken as a sample; the landscape index is used as independent variable, and the TN index is used as dependent variable.
Step S32: and extracting the same number of samples from the original sample data set of each non-point source pollution index based on a mode of replacing and resampling to form a plurality of new data sets, and training a decision tree in the random forest model by using the new data sets. Wherein a new data set corresponds to a decision tree.
Step S33: for each decision tree, calculating the mean square error of each decision tree by using the out-of-bag data of the decision tree, and recording the mean square error as a first Mean Square Error (MSE) 1 。
The data outside the bag refers to data which does not participate in the establishment of the decision tree when the decision tree is trained through the replaceable resampling. Assuming that D is a data set containing M samples, we re-sample it M times, which results in a new data set D' of size M, it can be assured that the new data set must contain a re-sampling of some sample in the original data set, and thus can be estimated. In each round of sampling, the probability that the sample x is decimated is 1/M, and the probability that the sample x is not decimated is 1-1/M, so after M rounds of sampling, the probability calculation formula that the sample is not decimated is as follows:
thus, approximately 36.8% of the samples in the original data set are not drawn, and these samples can be used to validate the decision tree trained by the new data set D'.
MSE is calculated by the following equation:
in the formula, E is the sample number of the data outside the bag, f (x) e ) One non-point source contamination index (TN) prediction value, y, for the e-th sample in the out-of-bag data e And representing the actual value of a one-side source pollution index (TN) corresponding to the e-th sample in the original sample data set.
Step S34: respectively carrying out disorder arrangement on each landscape index in the out-of-bag data of each decision tree to obtain disorder out-of-bag data, respectively calculating the mean square error of each disorder out-of-bag data according to the predicted value and the actual value of the decision tree of the non-point source pollution index to obtain a second Mean Square Error (MSE) 2 ；
When the sequence of the characteristic values of a certain column of landscape indexes is disordered, other characteristic variables of the data outside the bag need to be ensured to be unchanged, the decision tree is used for predicting samples after the disordered sequence, and the mean square error of the predicted value and the real value is MSE 2 。
Step S34 specifically includes:
step S341: randomly selecting one landscape index from each out-of-bag data, recording the landscape index as a selected landscape index, and performing disorder arrangement on the sequence of the current selected landscape indexes at all the sampling points to obtain disorder out-of-bag data;
step S342: for each out-of-bag data after disorder, according to the data after disorderCalculating the mean square error of the predicted value and the actual value of the decision tree of the non-point source pollution index of the data outside the bag to obtain a second mean square error MSE corresponding to the current selected landscape index 2 ；
Step S343: obtaining the unselected landscape indexes and returning to the step of randomly selecting one landscape index to be recorded as a selected landscape index, carrying out disorder arrangement on the sequence of the current selected landscape indexes at all the sampling points to obtain disorder data outside the bag until all the landscape indexes in the bag data are traversed to obtain MSE corresponding to each landscape index 2 。
For the convenience of understanding, the non-point source pollution index is assumed to be TN; the landscape index set corresponding to TN comprises landscape index X 1 ，X 2 ，X 3 ，....，X m (ii) a Therefore, the original data set can be regarded as a matrix which takes each landscape index as a column vector and takes the sampling point as a row vector. When the sampling points are disordered, the sequence of TN values of all sampling points of a landscape index is disordered, for example, a certain landscape index has values of 5 sampling points, the data is 1, 2, 3, 4, 5, and the data in the column may be 5, 2, 3, 1, 4 after random arrangement. Keeping other column data unchanged after disorder, namely establishing the original variable as a 'bad' variable which does not accord with the actual variable, and calculating the MSE of the landscape index at the moment 2 . MSE of all landscape indexes can be obtained by each decision tree 2 。
Step S35: determining a first mean score for each of the landscape indicators according to the first mean square error and the second mean square error;
step S35 specifically includes:
comparing the first mean square error with the second mean square error of each landscape index respectively, and determining a first score of each landscape index;
and averaging the first scores of the same landscape index of all the data outside the bag to obtain a first average score of each landscape index.
MSE due to index of each landscape 2 Is created using completely random and realistic inconformity 'bad' variablesThus we predict MSE 2 >MSE 1 (the higher the MSE the worse the model result). IncMSE ═ MSE 2 –MSE 1 Therefore, the higher the IncMSE is, the larger the error of the current corresponding disordered landscape index on the prediction result is, the more important the unordered landscape index is;
traversing the whole random forest model to obtain a landscape index X m Influence on all decision trees (each decision tree includes incMSE) m1 ，incMSE m2 ) Then calculate the landscape index X m The overall effect on the accuracy of the random forest model is an average increased mean square error (% incMSE), i.e. the incMSE of all decision trees is accumulated m Then taking the average value to obtain the mean square error% incMSE m Mixing% incMSE Xm As a landscape index X m A score of importance.
Step S36: randomly replacing the first mean square error and the second mean square error corresponding to each landscape index in all the decision trees, and determining a second mean score of each landscape index according to the replaced mean square errors;
step S36 specifically includes:
step S361: the first Mean Square Error (MSE) corresponding to each landscape index in each decision tree 1 And said second mean square error MSE 2 Carrying out random replacement;
the random permutation is explained in detail: MSE of each decision tree of each landscape index in random forest 1 、MSE 2 The sequence is disturbed. For example, the original and the out-of-order mean square errors of the ith decision tree are MSE i1 、MSE x2 The original and the out-of-order mean square errors of the xth decision tree are MSE x1 、MSE i2 。
Step S362: determining a second score of each landscape index according to the first mean square error and the second mean square error after the replacement corresponding to each landscape index;
after the mean square error of the ith decision tree and the xth decision tree is replaced, IncMSE i ＝MSE x2 -MSE i1 ，IncMSEx＝MSE i2 -MSE x1 。
Step S363: and averaging the second scores of the same landscape index corresponding to all the decision trees to obtain a second average score of each landscape index.
Step S37: and carrying out significance test on the first average score and the second average score of each landscape index, and screening out the key landscape index corresponding to each non-point source pollution index.
And (3) evaluating whether the difference between the replaced% incMSE (second mean score) and the original% incMSE (first mean score) reaches a significant degree (p <0.05), and judging the index to be a key landscape index influencing TN through significance detection.
S4: applying a nonparametric deviation reduction algorithm (nCPA) and a Bootstrap method (Bootstrap) to each non-point source pollution index and the corresponding key landscape index to obtain a mutation threshold of each key landscape index;
step S4 specifically includes:
step S41: sorting the non-point source pollution indexes in the original sample data set of each non-point source pollution index according to the numerical value to obtain each sorted non-point source pollution index;
step S42: calculating a total deviation value D of the non-point source pollution indexes for each sorted non-point source pollution index;
in which D is the whole number set Y 1 ，Y 2 ，Y 3 ，……Y n K is the number of samples in the original sample data set, Y k Is the value of a non-point source pollution index (TN variable value) in the kth sample, and mu is K Y k Average value of (a).
Step S43: randomly selecting a plurality of index numerical value division points for each sequenced non-point source pollution index, and dividing the sequenced non-point source pollution index into a first group of indexes and a second group of indexes according to each index numerical value division point;
by selecting the division point, the sample is divided into two different random samples (a first group of indexes and a second group of indexes) without significant trend change, for example, TN indexes are divided into two groups (Y) 1 ，Y 2 ，……，Y g And Y g+1 …Y n G is more than or equal to 1 and less than or equal to n), so that the respective characteristics of two random samples and the difference between the two random samples are analyzed in a comparative way.
Step S44: calculating the whole set of deviation values of the non-point source pollution indexes for the first set of indexes and the second set of indexes respectively for each index numerical value division point, and recording the whole set of deviation values as a first set of deviation values and a second set of deviation values;
calculating the sum of the first group of deviation values and the second group of deviation values, and recording the sum as a group deviation sum;
calculating the difference value between the total deviation value and the group deviation sum, and recording the difference value as a total group difference value;
step S45: selecting the index value division point corresponding to the maximum value in the total difference value as a mutation point of the non-point source pollution index, and determining the value of each key landscape index corresponding to the mutation point of each non-point source pollution index as a mutation threshold of the key landscape index.
Two groups of data divided by each division point g have two deviations, and the division point with the largest difference value between the sum and the total deviation is a catastrophe point;
Δg＝D-(D ≤g +D ＞g )
in which D is the whole number set Y 1 ，Y 2 ，Y 3 ，……，Y n D ≦ g and D > g are the deviations of the two columns of samples, respectively, g is 1, 2, 3, … …, n, Δ g is the difference between the divided deviations and the total deviation. At this time, the landscape index corresponding to the value of g of the divided TN is the mutation threshold of the landscape index, namely, the corresponding non-point source pollution index is mutated when the landscape index changes beyond the mutation threshold.
In order to ensure that the selection of the key landscape index mutation threshold is more accurate, a plurality of samples can be newly established based on a bootstrap method for calculation, so that the variation range of the same key landscape index mutation threshold is determined, and the landscape pattern can be subsequently optimized based on the variation range. In step S4, calculating a mutation threshold of each of the key landscape indicators, specifically including:
extracting a plurality of subsample sets with the same number of samples from the original sample data set of each non-point source pollution index;
for each subsample set, combining a nonparametric deviation reduction algorithm according to the area source pollution indexes and each corresponding key landscape index to obtain the mutation threshold value of each key landscape index corresponding to each area source pollution index;
and sequencing the mutation threshold values of the same key landscape index of all the subsample sets for each key landscape index corresponding to each non-point source pollution index to obtain the mutation threshold value range of the key landscape index corresponding to each non-point source pollution index. Specifically, the obtained mutation threshold values are arranged according to the size, and mutation point values with confidence intervals of [ 2.5%, 97.5% ] are reserved according to the data normality, so that the mutation threshold value range of the key landscape index is obtained.
S5: and for each non-point source pollution index, designing a landscape planning scheme by taking the mutation threshold of the key landscape index as a standard based on the relationship (positive correlation or negative correlation) between the key landscape index and the non-point source pollution index.
By taking TN as a research object, finding that key landscape indexes of TN are influenced by riparian zone dimensions and sub-watershed dimensions, and correspondingly adopted landscape management measures are changed, so that the landscape planning scheme for preventing and controlling total nitrogen is obtained as follows;
(1) the maximum patch index (LPI) of the grassland in the riparian zone is maintained to be more than 4.19 percent, the grassland edge density index (ED) is more than 27.99m/ha, and the farmland shape index (LSI) is less than 3.2.
(2) Maintaining an edge density index (ED) <1.69m/ha, a shape index (LSI) <2.46, and a farm land similarity percentage index (PLADJ) < 89.0% for a construction land.
And establishing a random forest model to definitely influence the key landscape indexes of the non-point source pollution by investigating and counting the main non-point source pollution types and the landscape indexes of the drainage basin. The potential of non-point source pollution prevention and control is developed from qualitative landscape pattern analysis, and a scientific planning standard is established for watershed landscape planning in a quantitative mode by calculating a mutation threshold value of a key landscape index influencing non-point source pollutants. The method realizes pollution control by development from the perspective of landscape ecology, achieves the purposes of preventing and controlling non-point source pollution, protecting water quality health and guaranteeing economic development, provides a theoretical basis for improving environmental control measures from the perspective of landscape planning in the follow-up process, and further provides possibility for realizing farmland intensification, economic maximization, environmental optimization and reduction of regional non-point source pollution in urban expansion.
And (3) screening key landscape indexes at different scales (sub-basin scale and riparian zone scale) by considering the type of the main non-point source pollutants of the basin to quantify the landscape pattern. The influence of landscape space configuration on the area source pollution is simulated and optimized, a landscape pattern optimization scheme is established, and the cooperative regulation and control among economic benefit, agricultural production, urban planning and area source pollution are realized. The method converts qualitative regulation into a quantitative planning scheme, combines ecological monitoring with an intelligent algorithm to form a clear, simple and unified research paradigm, is favorable for macroscopically preventing and controlling the non-point source pollution of the watershed, and solves the problem of water eutrophication.
Example 2
As shown in fig. 2, the present embodiment provides a drainage basin non-point source pollution prevention and control system based on key landscape indexes, the system includes:
the non-point source pollution index acquisition module T1 is used for acquiring various non-point source pollution indexes of each sampling point in the target drainage basin;
a landscape index acquisition module T2, configured to acquire multiple landscape indexes at each sampling point in the target flow domain, and record the multiple landscape indexes as a landscape index set; each non-point source pollution index of all the sampling points corresponds to the landscape index set;
the key landscape index screening module T3 is used for inputting the same surface source pollution indexes of all the sampling points and the corresponding landscape index sets into a random forest model respectively for the surface source pollution indexes with index values exceeding a preset value, and screening out the key landscape indexes corresponding to each surface source pollution index by combining a significance test method;
a key landscape index threshold calculation module T4, configured to apply a nonparametric deviation reduction algorithm and a bootstrap method to each non-point source pollution index and the corresponding key landscape index, so as to obtain a mutation threshold of each key landscape index;
and the landscape pattern optimization module T5 is used for designing a landscape planning scheme for each non-point source pollution index by taking the mutation threshold of the key landscape index as a standard based on the relationship between the key landscape index and the non-point source pollution index.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principle and the embodiment of the present invention are explained by applying specific examples, and the above description of the embodiments is only used to help understanding the method and the core idea of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (8)
1. A watershed non-point source pollution prevention and control method based on key landscape indexes is characterized by comprising the following steps:
collecting various non-point source pollution indexes of each sampling point in a target drainage basin;
acquiring various landscape indexes at each sampling point in the target flow domain, and recording the various landscape indexes as a landscape index set; each non-point source pollution index of all the sampling points corresponds to the landscape index set;
inputting the same non-point source pollution indexes of all the sampling points and the corresponding landscape index sets into a random forest model for the non-point source pollution indexes with index values exceeding a preset value, and screening out key landscape indexes corresponding to each non-point source pollution index by combining a significance test method;
applying a nonparametric deviation reduction algorithm and a bootstrap method to each non-point source pollution index and the corresponding key landscape index to obtain mutation threshold data of each key landscape index;
and for each non-point source pollution index, designing a landscape planning scheme by taking the mutation threshold of the key landscape index as a standard based on the relationship between the key landscape index and the non-point source pollution index.
2. The method according to claim 1, wherein the non-point source pollution indexes with index values exceeding a preset value, the same non-point source pollution indexes and the corresponding landscape index sets of all the sampling points are respectively input into a random forest model, and a significance test method is combined to screen out a key landscape index corresponding to each non-point source pollution index, specifically comprising:
combining the indexes of all sampling points and the corresponding landscape index sets to obtain an original sample data set of the non-point source pollution indexes for the non-point source indexes with pollution; one non-point source pollution index and the corresponding landscape index set of each sampling point are taken as a sample;
extracting the same number of samples from the original sample data set of each non-point source pollution index based on a mode of replacement and resampling to form a plurality of new data sets, and training a decision tree in a random forest model by using the new data sets;
for each decision tree, calculating the mean square error of each decision tree by using the data outside the bag of the decision tree, and recording the mean square error as a first mean square error;
respectively carrying out disorder arrangement on each landscape index in the out-of-bag data of each decision tree to obtain disorder out-of-bag data, and respectively calculating a mean square error of each disorder out-of-bag data according to a predicted value and an actual value of the decision tree of the non-point source pollution index to obtain a second mean square error;
determining a first mean score for each of the landscape indicators according to the first mean square error and the second mean square error;
randomly replacing the first mean square error and the second mean square error corresponding to each landscape index in all the decision trees, and determining a second mean score of each landscape index according to the replaced mean square errors;
and carrying out significance test on the first average score and the second average score of each landscape index, and screening out the key landscape index corresponding to each non-point source pollution index.
3. The method according to claim 2, wherein the step of performing disorder arrangement on each landscape index in the out-of-bag data of each decision tree to obtain disorder out-of-bag data, and the step of calculating a mean square error for each disorder out-of-bag data according to the predicted value and the actual value of the decision tree of the non-point source pollution index to obtain a second mean square error comprises:
randomly selecting one landscape index from each out-of-bag data, recording the landscape index as a selected landscape index, and performing disorder arrangement on the sequence of the current selected landscape indexes at all the sampling points to obtain disorder out-of-bag data;
calculating a mean square error of each out-of-bag data after disorder according to a decision tree predicted value and an actual value of the non-point source pollution index of the out-of-bag data after disorder, and obtaining a second mean square error corresponding to the current selected landscape index;
and acquiring the unselected landscape indexes and returning to the step of randomly selecting one landscape index to be recorded as a selected landscape index, and performing disorder arrangement on the sequence of the current selected landscape indexes at all the sampling points to obtain disorder data outside the bag until all the landscape indexes in the out-of-bag data are traversed to obtain the second mean square error corresponding to each landscape index.
4. The method of claim 3, wherein determining the first mean score for each of the landscape indicators according to the first mean square error and the second mean square error comprises:
comparing the first mean square error with the second mean square error of each landscape index respectively, and determining a first score of each landscape index;
and averaging the first scores of the same landscape index of all the data outside the bag to obtain a first average score of each landscape index.
5. The method of claim 2, wherein the randomly permuting the first and second mean-square errors for each of the landscape indicators in all of the decision trees, and determining a second mean score for each of the landscape indicators according to the permuted mean-square errors comprises:
randomly permuting the first and second mean square errors corresponding to each of the landscape indicators in each of the decision trees;
determining a second score of each landscape index according to the first mean square error and the second mean square error after the replacement corresponding to each landscape index;
and averaging the second scores of the same landscape index corresponding to all the decision trees to obtain a second average score of each landscape index.
6. The method of claim 2, wherein the applying a nonparametric deviation reduction algorithm and a bootstrap method to each of the area-source pollution indicators and the corresponding key landscape indicators to obtain a mutation threshold for each of the key landscape indicators comprises:
sorting the non-point source pollution indexes in the original sample data set of each non-point source pollution index according to the numerical value to obtain each sorted non-point source pollution index;
calculating a total deviation value of the non-point source pollution indexes for each sorted non-point source pollution index;
randomly selecting a plurality of index numerical value division points for each sequenced non-point source pollution index, and dividing the sequenced non-point source pollution index into a first group of indexes and a second group of indexes according to each index numerical value division point;
calculating a whole set of deviation values of the non-point source pollution indexes for the first set of indexes and the second set of indexes respectively for each index value division point, and recording the deviation values as a first set of deviation values and a second set of deviation values;
calculating the sum of the first group of deviation values and the second group of deviation values, and recording the sum as a group deviation sum;
calculating the difference value between the total deviation value and the group deviation sum, and recording the difference value as a total group difference value;
selecting the index numerical value division point corresponding to the maximum value in the total difference value as a mutation point of the non-point source pollution index, and determining the numerical value of the key landscape index corresponding to the mutation point of each non-point source pollution index as a mutation threshold value of each key landscape index.
7. The method of claim 6, wherein determining the value of each key landscape indicator corresponding to the mutation point of each non-point source pollution indicator is after the mutation threshold of each key landscape indicator further comprises:
extracting a plurality of subsample sets with the same number of samples from the original sample data set of each non-point source pollution index;
for each subsample set, obtaining the mutation threshold value of each key landscape index corresponding to each non-source pollution index according to the non-source pollution index and each corresponding key landscape index by combining a non-parameter deviation reduction algorithm and a bootstrap method;
and sequencing the mutation threshold values of the same key landscape index of all the subsample sets for each key landscape index corresponding to each non-point source pollution index to obtain the mutation threshold value range of each key landscape index corresponding to each non-point source pollution index.
8. A river basin non-point source pollution prevention and control system based on key landscape indexes is characterized by comprising:
the non-point source pollution index acquisition module is used for acquiring various non-point source pollution indexes of each sampling point in the target drainage basin;
the landscape index acquisition module is used for acquiring various landscape indexes at each sampling point in the target flow domain and recording the various landscape indexes as a landscape index set; each non-point source pollution index of all the sampling points corresponds to the landscape index set;
the key landscape index screening module is used for inputting the same surface source pollution indexes of all the sampling points and the corresponding landscape index sets into a random forest model respectively for surface source pollution indexes with index values exceeding a preset value, and screening out the key landscape indexes corresponding to each surface source pollution index by combining a significance inspection method;
the key landscape index threshold value calculation module is used for applying a nonparametric deviation reduction algorithm and a bootstrap method to each non-point source pollution index and the corresponding key landscape index to obtain a mutation threshold value of the key landscape index;
and the landscape pattern optimization module adjusts the landscape pattern in the target flow domain according to the mutation threshold data of each key landscape index corresponding to each non-point source pollution index and the response relation between each non-point source pollution index and each corresponding key landscape index, so as to realize the prevention and control of non-point source pollution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210577988.2A CN114862249A (en) | 2022-05-25 | 2022-05-25 | River basin non-point source pollution prevention and control method and system based on key landscape indexes |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210577988.2A CN114862249A (en) | 2022-05-25 | 2022-05-25 | River basin non-point source pollution prevention and control method and system based on key landscape indexes |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114862249A true CN114862249A (en) | 2022-08-05 |
Family
ID=82639910
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210577988.2A Pending CN114862249A (en) | 2022-05-25 | 2022-05-25 | River basin non-point source pollution prevention and control method and system based on key landscape indexes |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114862249A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116340723A (en) * | 2023-05-22 | 2023-06-27 | 安徽中科大国祯信息科技有限责任公司 | Rural water pollution quick tracing method and system based on big data |
-
2022
- 2022-05-25 CN CN202210577988.2A patent/CN114862249A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116340723A (en) * | 2023-05-22 | 2023-06-27 | 安徽中科大国祯信息科技有限责任公司 | Rural water pollution quick tracing method and system based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108491970B (en) | Atmospheric pollutant concentration prediction method based on RBF neural network | |
| Kassouri | Monitoring the spatial spillover effects of urbanization on water, built-up land and ecological footprints in sub-Saharan Africa | |
| Filstrup et al. | Regional variability among nonlinear chlorophyll—phosphorus relationships in lakes | |
| CN109408848B (en) | Distributed attribution method considering runoff evolution space-time heterogeneity | |
| CN109858755B (en) | Method for evaluating water quality | |
| Ravallion | Assessing the poverty impact of an assigned program | |
| Jackman | Models for ordered outcomes | |
| CN111160776A (en) | Method for detecting abnormal working condition in sewage treatment process by utilizing block principal component analysis | |
| Milošević et al. | The potential of chironomid larvae-based metrics in the bioassessment of non-wadeable rivers | |
| CN110706213A (en) | Bridge cluster structure damage judgment method based on strain response cumulative distribution function difference | |
| Jafarian et al. | Which spatial distribution model best predicts the occurrence of dominant species in semi-arid rangeland of northern Iran? | |
| CN114862249A (en) | River basin non-point source pollution prevention and control method and system based on key landscape indexes | |
| Sabzipour et al. | Evaluation of the potential of using subsets of historical climatological data for ensemble streamflow prediction (ESP) forecasting | |
| Li et al. | Development of a Wilks feature importance method with improved variable rankings for supporting hydrological inference and modelling | |
| Adhikari et al. | Climate change-induced invasion risk of ecosystem disturbing alien plant species: An evaluation using species distribution modeling | |
| CN113378473A (en) | Underground water arsenic risk prediction method based on machine learning model | |
| Wlezien et al. | The “timeline” method of studying electoral dynamics | |
| Guo et al. | Impacts of GCM credibility on hydropower production robustness under climate change: CMIP5 vs CMIP6 | |
| Zheng et al. | Addressing the uncertainty in modeling watershed nonpoint source pollution | |
| CN116227692B (en) | Crop heavy metal enrichment risk quantification method, system and storable medium | |
| CN117235510A (en) | Joint roughness prediction method and training method of joint roughness prediction model | |
| CN111428861A (en) | Manufacturing enterprise overseas investment site selection decision method based on multi-scale multi-dimensional proximity | |
| CN113516387B (en) | Regional ecological security pattern construction method and system based on geographic space big data | |
| CN115618720A (en) | Soil salinization prediction method and system based on altitude | |
| Chen et al. | Effectively inferring overall spatial distribution pattern of species in a map when exact coordinate information is missing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |