CN117113808B - Global climate mode coupling hydrologic model simulation processing method and system - Google Patents
Global climate mode coupling hydrologic model simulation processing method and system Download PDFInfo
- Publication number
- CN117113808B CN117113808B CN202310896277.6A CN202310896277A CN117113808B CN 117113808 B CN117113808 B CN 117113808B CN 202310896277 A CN202310896277 A CN 202310896277A CN 117113808 B CN117113808 B CN 117113808B
- Authority
- CN
- China
- Prior art keywords
- factors
- data
- model
- global
- adopting
- 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.)
- Active
Links
- 238000004088 simulation Methods 0.000 title claims abstract description 49
- 230000008878 coupling Effects 0.000 title claims abstract description 28
- 238000010168 coupling process Methods 0.000 title claims abstract description 28
- 238000005859 coupling reaction Methods 0.000 title claims abstract description 28
- 238000003672 processing method Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 72
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 41
- 238000012216 screening Methods 0.000 claims abstract description 29
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 24
- 238000012545 processing Methods 0.000 claims abstract description 23
- 230000001133 acceleration Effects 0.000 claims description 46
- 230000006870 function Effects 0.000 claims description 19
- 238000012360 testing method Methods 0.000 claims description 19
- 239000011159 matrix material Substances 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 14
- 238000005259 measurement Methods 0.000 claims description 13
- 238000005457 optimization Methods 0.000 claims description 13
- 230000015654 memory Effects 0.000 claims description 12
- 230000003416 augmentation Effects 0.000 claims description 11
- 238000012546 transfer Methods 0.000 claims description 11
- 230000001419 dependent effect Effects 0.000 claims description 7
- 210000002569 neuron Anatomy 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 6
- 238000011161 development Methods 0.000 claims description 6
- 230000033001 locomotion Effects 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 4
- 238000007689 inspection Methods 0.000 claims description 4
- 230000007787 long-term memory Effects 0.000 claims description 4
- 239000000463 material Substances 0.000 claims description 4
- 239000002689 soil Substances 0.000 claims description 4
- 230000003321 amplification Effects 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 claims description 3
- 230000008595 infiltration Effects 0.000 claims description 3
- 238000001764 infiltration Methods 0.000 claims description 3
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 3
- 238000013136 deep learning model Methods 0.000 abstract description 5
- 230000007246 mechanism Effects 0.000 abstract description 4
- 238000012805 post-processing Methods 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 10
- 238000004891 communication Methods 0.000 description 6
- 238000011160 research Methods 0.000 description 5
- 238000012795 verification Methods 0.000 description 5
- 238000001556 precipitation Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 239000012530 fluid Substances 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 230000010355 oscillation Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000001808 coupling effect Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000013480 data collection Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000005713 exacerbation Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010792 warming Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a processing method and a processing system for global climate mode coupling hydrologic model simulation, which belong to the technical field of hydrologic model simulation and post-processing, and comprise the following steps: obtaining simulated daily runoff according to a VIC distributed watershed hydrological model, and decomposing the simulated daily runoff into a plurality of natural modal components with different frequencies by adopting variational modal decomposition; the global key sea area sea temperature factors and atmospheric circulation factors are screened by adopting a correlation coefficient method and a stepwise multiple regression method, and screened factors are obtained; and constructing an input factor set by using a plurality of natural modal components with different frequencies and the factors after screening, constructing an LSTM model, and optimizing and calibrating the super parameters of the LSTM model by adopting an Archimedes algorithm to obtain an improved simulated runoff series. The invention considers the advantages of the deep learning model and the scientificity of the influence factors based on the physical mechanism, corrects the daily runoff simulated by the hydrologic model, and provides a reference basis for the coupling simulation of the global climate mode and the watershed hydrologic model.
Description
Technical Field
The invention relates to the technical field of hydrologic model simulation and post-treatment, in particular to a processing method and a processing system for global climate mode coupling hydrologic model simulation.
Background
Climate change due to human activity has been identified as a cause of global warming, which has profound effects on the land-atmosphere-ocean system, extreme weather and climate event exacerbation, and presents a great challenge to the natural ecosystem and the human socioeconomic system. The accuracy of the hydrologic model on runoff simulation is improved, and the method has important significance on flood risk management and water resource planning and utilization.
In recent years, research based on a global climate mode coupling watershed hydrologic model is rapidly developed, and regional application of the global climate mode effectively deepens understanding of a global climate system and hydrologic cycle. However, when the global climate mode is coupled with the watershed hydrologic model, the coupling effect is often poor because the resolution of the global climate mode is rough. At present, partial researches are carried out on improving the effect of the global climate mode coupling hydrologic model simulation based on a deep learning model, but only weather and hydrologic factors are often considered as input variables when the deep learning model is applied, and large-scale atmospheric circulation factors and sea Wen Yinzi are not considered, so that the accuracy and scientificity of the global climate mode coupling hydrologic model simulation result are restricted.
Disclosure of Invention
The invention provides a processing method and a processing system for a global climate mode coupling hydrological model simulation, which are used for solving the defects that in the prior art, influence factors considered when the global climate mode is coupled with a river basin hydrological model are limited, and the coupling effect is poor due to rough simulation process.
In a first aspect, the present invention provides a processing method for global climate mode coupled hydrological model simulation, including:
Establishing a VIC distributed drainage basin hydrological model based on drainage basin actual measurement meteorological hydrological data, global climate mode data and underlying surface data of a region to be measured;
Obtaining simulated daily runoff according to the VIC distributed watershed hydrological model, and decomposing the simulated daily runoff into a plurality of different-frequency natural modal components by adopting variational modal decomposition;
collecting global key sea area sea temperature factors and atmospheric circulation factors, and screening the global key sea area sea temperature factors and the atmospheric circulation factors by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors;
And constructing an input factor set by the natural modal components with different frequencies and the screened factors, constructing an LSTM model, and optimizing and calibrating the super parameters of the LSTM model by adopting an Archimedes algorithm to obtain an improved simulated runoff series.
According to the processing method for simulating the global climate mode coupling hydrologic model, which is provided by the invention, the VIC distributed drainage basin hydrologic model is established based on the drainage basin actually measured meteorological hydrologic data, the global climate mode data and the underlying surface data of the area to be detected, and the processing method comprises the following steps:
Collecting daily runoff process data of a drainage basin outlet station as actual measurement hydrological data, taking a CN05.1 grid point data set as actual measurement meteorological data, taking a sixth international coupling mode comparison plan CMIP6 data as global climate mode data, and collecting underlying surface data comprising soil data, vegetation type data, land utilization type data, topographic data and elevation data;
Interpolating the measured hydrological data, the measured meteorological data, the global climate mode data and the underlying surface data to a preset resolution grid through bilinear interpolation technology, and carrying out aggregate average, uniform time length and time resolution on the global climate mode data to obtain an initial VIC distributed drainage basin hydrological model;
And carrying out parameter calibration on preset key hydrologic parameters in the initial VIC distributed watershed hydrologic model by using a calibration objective function by adopting a complex evolution optimization algorithm SCE-UA algorithm, determining a Nash efficiency coefficient and a water yield error objective function, and training the initial VIC distributed watershed hydrologic model to obtain the VIC distributed watershed hydrologic model.
According to the processing method for simulating the global climate mode coupling hydrologic model, provided by the invention, simulated daily runoff is obtained according to the VIC distributed watershed hydrologic model, and the simulated daily runoff is decomposed into a plurality of different-frequency natural mode components by adopting variation mode decomposition, and the processing method comprises the following steps:
wherein f (t) is original runoff data, t is time component, For partial derivative calculation, delta (t) is an impulse function, j is an imaginary unit, e is a natural index, K is the total number of the natural modal components, K is any natural modal component number, { u k}＝{u1(t),u2(t),…,uk (t) } is K natural modal components after decomposition, { omega k}＝{ω1(t),ω2(t),…,ωk (t) } is the center frequency corresponding to each modal component, and x is a convolution operator;
And solving by adopting a secondary penalty factor and a Lagrange operator, and carrying out iterative search by adopting an alternate direction multiplier method to obtain the natural modal components with different frequencies.
According to the processing method for the global climate mode coupling hydrologic model simulation, which is provided by the invention, the global key sea area sea temperature factors and the atmospheric circulation factors are collected, and the global key sea area sea temperature factors and the atmospheric circulation factors are screened by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors, wherein the method comprises the following steps:
acquiring daily sea temperature data of a global preset key sea area within a preset time range, and acquiring a global key remote related atmospheric sphere index;
Based on annual time lag, calculating the pearson correlation coefficient of the daily sea temperature data and the remote correlation atmospheric sphere index, and screening a plurality of factors with preset ranking pearson correlation coefficients;
Screening a plurality of factors of the preset ranking pearson correlation coefficient by using a preset confidence inspection factor by adopting the stepwise multiple regression method to obtain a prediction candidate factor;
and rejecting and checking by using a regression coefficient augmentation matrix to obtain the screened factors.
According to the processing method for the global climate mode coupling hydrologic model simulation, provided by the invention, the regression coefficient augmentation matrix is utilized for removing and checking, and the screened factors are obtained, and the processing method comprises the following steps:
calculating variance contribution values, introducing factors which affect the dependent variables most significantly one by one into the regression coefficient amplification matrix, and performing F test on the introduced independent variables one by one;
and screening out factors with the smallest variance contribution from regression equations corresponding to the regression coefficient augmentation matrix which is introduced with factors with the most significant influence on dependent variables, performing rejection test, and screening out factors with the largest variance contribution from factors which are not introduced with the regression equations, performing introduction test, thus obtaining the screened factors.
According to the processing method for simulating the global climate mode coupling hydrologic model provided by the invention, an input factor set is constructed by the natural modal components with different frequencies and the screened factors, a long-term and short-term memory LSTM model is constructed, and the Archimedes algorithm is adopted to optimize and rate the super parameters of the LSTM model, so that an improved simulated runoff series is obtained, and the processing method comprises the following steps:
determining that the LSTM model comprises two full-connection layers, connecting a Dropout layer with preset parameters after each full-connection layer, optimizing the LSTM model by adopting an Adam algorithm, and converging by adopting a preset loss function;
Optimizing and calibrating the number of hidden layer neurons, the initial learning rate, the minimum batch size and the maximum training times of the LSTM model by adopting the Archimedes algorithm to obtain an optimized simulated runoff series;
And calibrating the optimized simulated runoff series by utilizing a calibration objective function to obtain the improved simulated runoff series.
According to the processing method for simulating the global climate mode coupling hydrologic model, provided by the invention, the Archimedes algorithm is adopted to carry out optimization calibration on the number of hidden layer neurons, the initial learning rate, the minimum batch size and the maximum training times of the LSTM model, and an optimized simulated runoff series is obtained, and the processing method comprises the following steps:
Step 1: randomly generating a plurality of individuals, initializing the volume, density and acceleration of each individual, and determining the optimal volume, optimal density and optimal acceleration of the individual with optimal fitness value;
step 2: updating the densities and volumes of the remaining individuals based on the individual optimal volumes, the individual optimal densities, and the individual optimal accelerations;
Step3: determining a transfer operator and a density reduction factor according to the current iteration times and the maximum iteration times;
Step 4: if the transfer operator is determined to be smaller than or equal to a preset threshold value, global exploration is carried out, a global individual acceleration updating formula is obtained through the acceleration, density and volume of random materials and the acceleration, density and volume of an individual in the next iteration, the global individual acceleration updating formula is normalized, and a global individual position updating formula is obtained through combining the current any one of the body positions and the current random individual position;
Step 5: if the transfer operator is determined to be greater than a preset threshold value, carrying out local development, obtaining a local individual acceleration updating formula by the individual optimal volume, the individual optimal density, the individual optimal acceleration and the acceleration, density and volume of an individual in the next iteration, normalizing the local individual acceleration updating formula, and obtaining a local individual position updating formula by combining the current optimal position of any one of the individual, the current optimal individual position and the movement direction change mark;
Step 6: and (5) repeatedly executing the steps 2 to 5 until the maximum iteration times are reached, and obtaining the optimized simulated runoff series.
In a second aspect, the present invention also provides a processing system for global climate mode coupled hydrological model simulation, comprising:
The establishing module is used for establishing a VIC distributed drainage basin hydrological model based on drainage basin actual measurement meteorological hydrological data, global climate mode data and underlying surface data of the to-be-detected region;
The decomposition module is used for obtaining simulated daily runoff according to the VIC distributed watershed hydrological model, and decomposing the simulated daily runoff into a plurality of natural modal components with different frequencies by adopting variation modal decomposition;
The screening module is used for collecting global key sea area sea temperature factors and atmospheric circulation factors, and screening the global key sea area sea temperature factors and the atmospheric circulation factors by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors;
and the optimization module is used for constructing an input factor set by the natural modal components with the different frequencies and the screened factors, constructing an LSTM model, and optimizing and calibrating the super parameters of the LSTM model by adopting an Archimedes algorithm to obtain an improved simulated runoff series.
In a third aspect, the invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing a processing method of global climate mode coupled hydrologic model simulation as described in any of the above when executing the program.
In a fourth aspect, the invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method of processing a global climate mode coupled hydrological model simulation as described in any of the above.
According to the processing method and the processing system for the global climate mode coupling hydrologic model simulation, provided by the invention, through considering the advantages of the deep learning model and the scientificity of the influence factors based on the physical mechanism, the daily runoff of the hydrologic model simulation is corrected, and a reference basis is provided for the global climate mode and the watershed hydrologic model coupling simulation.
Drawings
In order to more clearly illustrate the invention or the technical solutions of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow diagram of a method for processing a global climate model coupled hydrologic model simulation provided by the present invention;
FIG. 2 is a second flow chart of a method for processing a global climate model-coupled hydrological model simulation according to the present invention;
FIG. 3 is a schematic view of a variation modal decomposition provided by the present invention;
FIG. 4 is a schematic diagram of a long-short neural network model memory unit provided by the invention;
FIG. 5 is a flowchart of an Archimedes algorithm provided by the invention;
FIG. 6 is a schematic diagram of an initial runoff simulation sequence and a modified runoff simulation sequence provided by the present invention;
FIG. 7 is a schematic diagram of a processing system for global climate mode coupled hydrologic model simulation provided by the present invention;
Fig. 8 is a schematic structural diagram of an electronic device provided by the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
FIG. 1 is a schematic flow chart of a processing method for global climate mode coupling hydrological model simulation according to an embodiment of the present invention, as shown in FIG. 1, including:
step 100: establishing a VIC distributed drainage basin hydrological model based on drainage basin actual measurement meteorological hydrological data, global climate mode data and underlying surface data of a region to be measured;
Step 200: obtaining simulated daily runoff according to the VIC distributed watershed hydrological model, and decomposing the simulated daily runoff into a plurality of different-frequency natural modal components by adopting variational modal decomposition;
Step 300: collecting global key sea area sea temperature factors and atmospheric circulation factors, and screening the global key sea area sea temperature factors and the atmospheric circulation factors by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors;
Step 400: and constructing an input factor set by the natural modal components with different frequencies and the screened factors, constructing an LSTM model, and optimizing and calibrating the super parameters of the LSTM model by adopting an Archimedes algorithm to obtain an improved simulated runoff series.
Firstly, collecting actually measured meteorological hydrological data, global climate mode data and underlying surface data of a river basin, establishing a hydrological model of the river basin, and calibrating parameters to obtain a preliminary daily runoff simulation sequence; decomposing the simulated daily runoff into K different frequency modes by adopting a variational mode decomposition, collecting and extracting sea temperature factors and important atmospheric circulation factors of a global key sea area, screening the factors passing the saliency test by adopting a correlation coefficient method and a stepwise regression method, and forming an input factor set together with the runoff modes; and taking the constructed factor set as input, constructing a long-term memory neural network model, optimizing the hyper-parameters of the model by adopting an Archimedes algorithm, and obtaining an improved simulated runoff series after training and verification of the model. The specific implementation flow is shown in fig. 2:
Firstly, collecting hydrologic, meteorological and underlying data of a research area, extracting output data of a global climate mode, and constructing a variable infiltration capacity (Variable Infiltration Capacity, VIC) distributed watershed hydrologic model;
Secondly, adopting a self-adaptive and completely non-recursive variational modal decomposition (Variational Mode Decomposition, VMD) method to decompose the simulated daily runoff into K natural modal components (INTRINSIC MODE FUNCTION, IMF) with different frequencies;
Collecting N sea temperature factors and important atmospheric circulation factors of a global key sea area again, and considering the earlier 1-365 days of dead time, for example, predicting runoff of 1 month and 1 day in 2001, wherein the sea temperature factors and the atmospheric circulation factors of 1 month and 31 months in 2000 can be used, namely, the alternative factors have 365 XN items; performing primary screening by adopting a correlation coefficient method, selecting factors passing through significance test by using a stepwise multiple regression method, and forming an input factor set together with the runoff mode obtained by the decomposition;
Finally, the obtained factor set is used as the input of a Long Short-Term Memory (LSTM) model, the LSTM model is constructed, the Archimedes algorithm (ARCHIMEDES OPTIMIZATION ALGORITHM, AOA) is adopted to carry out optimization calibration on the super parameters of the LSTM model, and the model is trained and verified to obtain an improved simulated runoff series.
According to the invention, through considering the advantages of the deep learning model and the scientificity of the influence factors based on the physical mechanism, the daily runoff simulated by the hydrologic model is corrected, and a reference basis is provided for the coupling simulation of the global climate mode and the watershed hydrologic model.
Based on the above embodiment, step 100 includes:
Collecting daily runoff process data of a drainage basin outlet station as actual measurement hydrological data, taking a CN05.1 grid point data set as actual measurement meteorological data, taking a sixth international coupling mode comparison plan CMIP6 data as global climate mode data, and collecting underlying surface data comprising soil data, vegetation type data, land utilization type data, topographic data and elevation data;
Interpolating the measured hydrological data, the measured meteorological data, the global climate mode data and the underlying surface data to a preset resolution grid through bilinear interpolation technology, and carrying out aggregate average, uniform time length and time resolution on the global climate mode data to obtain an initial VIC distributed drainage basin hydrological model;
And carrying out parameter calibration on preset key hydrologic parameters in the initial VIC distributed watershed hydrologic model by using a calibration objective function by adopting a complex evolution optimization algorithm SCE-UA algorithm, determining a Nash efficiency coefficient and a water yield error objective function, and training the initial VIC distributed watershed hydrologic model to obtain the VIC distributed watershed hydrologic model.
Specifically, when the data collection is carried out, the hydrologic data are daily runoff process data of the drainage basin outlet station; the actually measured meteorological data uses CN05.1, which is a set of grid point data sets obtained by interpolation of the observed data of 2400 sites in China by using a distance-level approximation method, and the grid point data sets comprise variables such as daily precipitation, wind speed, highest air temperature, lowest air temperature and the like; global climate pattern the climate pattern comparison plan CMIP6 from the world climate study plan selects the r1i1p1f1 test data of the historical simulation "historical" representing the real world simulation of the historical period using the same data variables as the measured meteorological data; the underlying data includes soil data, vegetation type and land use type data, terrain and elevation data, and the like. The data are interpolated to a grid of the same resolution of 0.25 DEG x 0.25 DEG by bilinear interpolation technique, wherein global climate patterns are aggregate averaged and sorted to the same time length, and the time resolution is on the daily scale.
After the VIC distributed watershed hydrological model is established, carrying out parameter calibration on 6 key hydrological parameters (INFILT, ds, dsmax, ws, depth _2 and depth_3) of the model, adopting a complex evolutionary optimization algorithm (Shuffled Complex Evolution-University of Arizona, SCE-UA), and training to obtain an optimal hydrological model by using odd years as a periodic rate and even years as a verification period in the parameter calibration process. Selecting an objective function which simultaneously considers the Nash efficiency coefficient and the water quantity error, wherein the calculation formula is as follows:
Wherein Y i represents the measured runoff sequence value, Representing simulated runoff sequence values,/>Represents the average value of the simulated runoff sequence, and n represents the runoff sequence length. In particular, when obj=0, the representative analog sequence is identical to the measured sequence.
Based on the above embodiment, step 200 includes:
obtaining simulated daily runoff according to the VIC distributed watershed hydrological model, decomposing the simulated daily runoff into a plurality of different-frequency natural mode components by adopting variation mode decomposition, and comprising the following steps:
wherein f (t) is original runoff data, t is time component, For partial derivative calculation, delta (t) is an impulse function, j is an imaginary unit, e is a natural index, K is the total number of the natural modal components, K is any natural modal component number, { u k}＝{u1(t),u2(t),…,uk (t) } is K natural modal components after decomposition, { omega k}＝{ω1(t),ω2(t),…,ωk (t) } is the center frequency corresponding to each modal component, and x is a convolution operator;
Solving by adopting a secondary penalty factor and a Lagrange operator, and performing iterative search by adopting an alternate direction multiplier method (ALTERNATING DIRECTION METHOD OF MULTIPLIERS, ADMM) to obtain the natural modal components with different frequencies, wherein a comparison diagram of VMD decomposition and the frequency spectrums of the IMF components is shown in FIG. 3 in detail.
Based on the above embodiment, step 300 includes:
collecting global key sea area sea temperature factors and atmospheric circulation factors, and screening the global key sea area sea temperature factors and the atmospheric circulation factors by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors, wherein the method comprises the following steps of:
acquiring daily sea temperature data of a global preset key sea area within a preset time range, and acquiring a global key remote related atmospheric sphere index;
Based on annual time lag, calculating the pearson correlation coefficient of the daily sea temperature data and the remote correlation atmospheric sphere index, and screening a plurality of factors with preset ranking pearson correlation coefficients;
Screening a plurality of factors of the preset ranking pearson correlation coefficient by using a preset confidence inspection factor by adopting the stepwise multiple regression method to obtain a prediction candidate factor;
and rejecting and checking by using a regression coefficient augmentation matrix to obtain the screened factors.
And removing and checking by using a regression coefficient augmentation matrix to obtain the screened factors, wherein the method comprises the following steps of:
calculating variance contribution values, introducing factors which affect the dependent variables most significantly one by one into the regression coefficient amplification matrix, and performing F test on the introduced independent variables one by one;
and screening out factors with the smallest variance contribution from regression equations corresponding to the regression coefficient augmentation matrix which is introduced with factors with the most significant influence on dependent variables, performing rejection test, and screening out factors with the largest variance contribution from factors which are not introduced with the regression equations, performing introduction test, thus obtaining the screened factors.
Specifically, the embodiment of the invention downloads global daily sea temperature data of OI SST V2 High Resolution from NOAA, wherein the spatial Resolution is 0.25 degrees multiplied by 0.25 degrees, and the time range is 9 months in 1981 to date. Sea temperature data of sea areas with obvious influence on a research area are extracted, and sea Wen Qiu average value in each area represents sea Wen Daxiao of the whole area, so that sea Wen Yinzi daily for many years is obtained. Taking pacific and indian areas as an example, 10 key sea areas are extracted from the whole research scope, and the latitude and longitude ranges of the sea areas are A1: Zone (10 DEG S-0 DEG, 90 DEG W-80 DEG W); a2: /(I) Zone (5 DEG S-5 DEG N,150 DEG W-90 DEG W); a3: /(I)Zone (5 DEG S-5 DEG N,160 DEG E-150 DEG W); a4: /(I)Zone (5 DEG S-5 DEG N,170 DEG W-120 DEG W); a5: north Pacific (30-45, 160-145); a6: e Huoci g sea (50 DEG N-60 DEG N,140 DEG E-160 DEG E); a7: japanese sea (30 DEG N-45 DEG N,130 DEG E-145 DEG E); a8: south China sea (8-22 degree N, 110-120 degree E); a9: the Bengala bay (6 DEG N-20 DEG N,82 DEG E-98 DEG E); a10: arabian (10 DEG S-10 DEG N,50 DEG E-70 DEG E).
The global critical tele-related atmospheric flow index is then downloaded from Climate Prediction Center, mainly including the arctic Tao index (Arctic Oscillation, AO), the antarctic Tao index (ANTARCTIC OSCILLATION, AAO), the North Atlantic Tao index (North Atlantic Oscillation, NAO), and the North America tele-related index (Pacific/North America, PNA). Wherein AO, NAO, PNA index data began in 1950 and AAO index data began in 1979. The arctic wave motion refers to the phenomenon of potential height difference between high and low latitudes in the arctic region, and mainly affects the arctic region and surrounding regions, including climate elements such as air temperature, precipitation and the like in the high latitude region in the northern hemisphere; the phenomenon that the seesaw is changed reversely between the north direction and the south direction in the middle and high latitude areas of the south pole wave motion guide hemisphere always has good latitudinal symmetry, and the phenomenon of the good latitudinal symmetry mainly affects the south pole area and the surrounding areas thereof, such as the climate change in areas of south america, australia and the like; the north atlantic billows refer to the annual changes of the atmospheric circulation system in the north atlantic region, and mainly affect climate factors such as air temperature, precipitation and the like in the middle and high latitude regions of asia in europe and north america; the remote-related type of pacific north america mainly refers to the phenomenon of high-pressure system in north of pacific and low-pressure system in north america, the change of PNA index is closely related to temperature in northern hemisphere in winter, climate factors such as precipitation, etc., for example, the normal phase state of PNA index is usually accompanied by warmth and drought in western north america and canada, while the eastern portions of alaska and north america are more susceptible to cold tide and storm snow.
For each atmospheric circulation factor and sea Wen Yinzi, considering the dead time of 1-365 days, for example, predicting runoff of 1 month 1 day in 2001, sea temperature factors and atmospheric circulation factors of 31 days in 1 month 1 day in 2000 to 12 months in 2000 can be used, namely, the candidate factors have 365 XN items; and calculating the Pearson correlation coefficient of the factor and the runoff sequence, and primarily screening out the first 100 factors with the maximum correlation coefficient.
Wherein: r XY is the correlation coefficient between X and Y; n is the number of data samples; x i is the ith sample value of x; y i is the ith sample value of y; Sample mean value of x; /(I) Is the sample mean of y.
The factors that passed the F test at a 0.01 confidence level were further screened on the basis of the 100 factors obtained in step S3.3 using stepwise multiple regression analysis. The stepwise multiple regression method combines the advantages of the forward introduction method and the backward elimination method, assuming that the radial flow f (t) and the m predictors x i (t) satisfy the linear regression equation:
f(t)＝β1x1(t)+β2x2(t)+…βixi(t)+…+βmxm(t);t＝1,2,…,N;i＝1,2,…,m.
Where β i is the regression coefficient, the least squares estimation can be used.
In addition, the predictors are rejected and introduced for testing: calculating a regression coefficient augmentation matrix of the candidate factors, introducing factors with the most obvious influence on dependent variables one by one through calculating variance contribution, performing F test on the independent variables introduced into the model one by one, selecting factors with the smallest variance contribution from the variables introduced into a regression equation as rejection test, selecting factors with the largest variance contribution from factors not entering the equation, and performing introduction test to finally obtain the screened related factors. The specific calculation steps are as follows:
firstly, calculating regression coefficient augmentation matrix of factors as follows:
R(0)＝(rij),i,j＝1,2,…,m+1;xm+1＝y
The correlation coefficient is:
in the above (j＝1,2,…,m+1)
X it and x jt each represent a correlation value between any runoff data and any predictor, i and j being used to distinguish between different predictors;
then step-wise calculations are performed:
if the calculation of step i has been performed (l=0, 1, …), the calculation of step i+1 is:
First, the variance contribution U is calculated one by one for x i i
Where l represents the current calculation step number, l represents the next calculation step number,Cross-correlation coefficient representing the number of steps currently calculated,/>Representing the autocorrelation coefficient of the current number of calculation steps.
Next, selecting x i with the minimum U value from variables which are introduced into the regression equation as a rejection test
If it isRepresenting the minimum U value threshold, eliminating the factor and performing matrix transformation operation:
finally, the factor of the maximum U value is selected from the factors not entering the equation, and the introduction test is performed.
If it isRepresenting the maximum U value threshold, introducing the factor into the equation, performing matrix transformation operation, and if no introduction is performed, ending the step-by-step calculation.
Based on the above embodiment, step 400 includes:
determining that the LSTM model comprises two full-connection layers, connecting a Dropout layer with preset parameters after each full-connection layer, optimizing the LSTM model by adopting an Adam algorithm, and converging by adopting a preset loss function;
Optimizing and calibrating the number of hidden layer neurons, the initial learning rate, the minimum batch size and the maximum training times of the LSTM model by adopting the Archimedes algorithm to obtain an optimized simulated runoff series;
And calibrating the optimized simulated runoff series by utilizing a calibration objective function to obtain the improved simulated runoff series.
The method for optimizing and calibrating the hidden layer neuron number, the initial learning rate, the minimum batch size and the maximum training times of the LSTM model by adopting the Archimedes algorithm to obtain an optimized simulated runoff series comprises the following steps:
Step 1: randomly generating a plurality of individuals, initializing the volume, density and acceleration of each individual, and determining the optimal volume, optimal density and optimal acceleration of the individual with optimal fitness value;
step 2: updating the densities and volumes of the remaining individuals based on the individual optimal volumes, the individual optimal densities, and the individual optimal accelerations;
Step3: determining a transfer operator and a density reduction factor according to the current iteration times and the maximum iteration times;
Step 4: if the transfer operator is determined to be smaller than or equal to a preset threshold value, global exploration is carried out, a global individual acceleration updating formula is obtained through the acceleration, density and volume of random materials and the acceleration, density and volume of an individual in the next iteration, the global individual acceleration updating formula is normalized, and a global individual position updating formula is obtained through combining the current any one of the body positions and the current random individual position;
Step 5: if the transfer operator is determined to be greater than a preset threshold value, carrying out local development, obtaining a local individual acceleration updating formula by the individual optimal volume, the individual optimal density, the individual optimal acceleration and the acceleration, density and volume of an individual in the next iteration, normalizing the local individual acceleration updating formula, and obtaining a local individual position updating formula by combining the current optimal position of any one of the individual, the current optimal individual position and the movement direction change mark;
Step 6: and (5) repeatedly executing the steps 2 to 5 until the maximum iteration times are reached, and obtaining the optimized simulated runoff series.
Specifically, the embodiment of the invention takes the factor set obtained by combination in the previous embodiment as an LSTM model input, constructs the LSTM model, adopts AOA to carry out optimization calibration on the super parameters of the LSTM model, and obtains an improved simulated runoff series after training and verification of the model.
It should be noted that LSTM is a special recurrent neural network, and three gating mechanisms including an input gate, a forgetting gate and an output gate are introduced into LSTM, so that input, forgetting and output of information can be controlled, and thus the long-term dependence problem existing in the recurrent neural network can be effectively solved when sequence data is processed. In modeling, LSTM uses two full-connection layers, in order to prevent model overfitting, dropout layers with parameters of 0.1 are added after each full-connection layer, optimization training is carried out by adopting Adam algorithm, RMSE is selected as a loss function, and a schematic diagram of a memory unit of a common long-short-term memory neural network model is shown in FIG. 4 in detail.
Further, the super parameters of the LSTM model are optimized and calibrated by adopting an AOA algorithm, and the super parameters comprise parameters such as the number of neurons of an hidden layer, an initial learning rate, the size of the minimum batch, the maximum training times and the like. The archimedes algorithm is a new meta-heuristic algorithm proposed by Fatma a. Hashim et al under the heuristic of archimedes' law of buoyancy. The archimedes algorithm essentially mimics archimedes' law, i.e., when an object is fully or partially immersed in a fluid, the fluid exerts an upward buoyancy force on the object equal to the weight of the fluid displaced by the object. The archimedes algorithm is a population-based algorithm in which population individuals are immersed objects, each individual has three properties of density, volume and acceleration in addition to position, the acceleration of the individual is adjusted by changing the density and volume of the individual, and the updated density, volume and acceleration determine the new position of the individual. The archimedes algorithm flow chart is shown in fig. 5, and the main steps of the algorithm are as follows:
(1) Initializing a population: a number of individuals are randomly generated, and the volume (vol), density (den), acceleration (acc) of each individual are initialized. Evaluating the initial population, and selecting the volume (vol best), the density (den best) and the acceleration (acc best) of the individual with the optimal fitness value;
(2) Density and volume update: according to the volume (vol best), the density (den best) and the acceleration (acc best) of the optimal individual in the last step, the density and the volume of other individuals are updated, and the calculation formula is as follows:
Wherein: and/> Is the density of the ith and i+1 individuals in the t generation; /(I)And/>Volume for the ith and i+1 individuals in the t generation; rand is a random number of (0, 1);
(3) Transfer operator and density reduction factor update: the AOA algorithm classifies an object immersed in a liquid into a global exploration phase and a local development phase, depending on whether it collides. If collision does not occur, the AOA performs global exploration; and on the contrary, carrying out local development. The conversion of two phases is performed by a Transfer operator (TF), the calculation formula of which is as follows:
TF＝exp((t-tmax)/tmax)
wherein: t and t max represent the current iteration number and the maximum iteration number, respectively.
TF increases gradually over time until 1 is reached, and similarly the density reduction factor d also helps AOA to perform global to local search conversion:
(4) Global exploration phase: when TF is less than or equal to 0.5, the AOA performs global exploration, and the individual acceleration update formula is as follows:
Wherein the method comprises the steps of And/>The density, volume and acceleration of individual i are the t+1st iteration, respectively. Whereas acc mr、denmr and vol mr Respectively are provided with are the acceleration, density, and volume of the random material.
Normalization of individual acceleration:
Where u and l represent the left and right end points of the acceleration normalization map interval.
The location update formula for an individual is as follows:
/>
Wherein, And/>For the positions of the ith individual in the t+1 generation and the t generation, x rand is the position of the random individual in the t generation, C 1 is a constant 2, rand is a uniform random number with the value of [0,1], and d is a density reduction factor.
(5) Local development stage: when TF >0.5, AOA is developed locally, and the individual acceleration update formula is as follows:
acceleration is also normalized for individual position update, the position update formula is as follows:
Where C 2 is a constant of 6, t=c 3 ×tf, F is a sign for changing the direction of motion, Wherein p=2×rand-C 4;
(6) And (5) repeating the steps (2) to (5) until the maximum iteration number is reached, and obtaining the optimal scheme.
Finally, selecting an objective function consistent with the rate timing of the VIC model when the LSTM model is trained, using odd years as the rate period and even years as the verification period, and obtaining an improved simulated runoff series after training and verification. The schematic diagrams of the initial daily runoff simulation sequence and the improved daily runoff simulation sequence are shown in fig. 6.
The processing system for the global climate mode coupling hydrologic model simulation provided by the invention is described below, and the processing system for the global climate mode coupling hydrologic model simulation described below and the processing method for the global climate mode coupling hydrologic model simulation described above can be correspondingly referred to each other.
FIG. 7 is a schematic diagram of a processing system for global climate mode coupled hydrological model simulation according to an embodiment of the present invention, as shown in FIG. 7, including: the establishing module 71, the decomposing module 72, the screening module 73 and the optimizing module 74 include:
The establishing module 71 is configured to establish a VIC distributed drainage basin hydrological model based on drainage basin actually measured meteorological hydrological data, global climate mode data and underlying surface data of the area to be measured; the decomposition module 72 is configured to obtain a simulated daily runoff according to the VIC distributed watershed hydrological model, and decompose the simulated daily runoff into a plurality of natural modal components with different frequencies by adopting variation modal decomposition; the screening module 73 is configured to collect global critical sea area sea temperature factors and atmospheric circulation factors, and screen the global critical sea area sea temperature factors and the atmospheric circulation factors by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors; the optimization module 74 is configured to construct an input factor set from the natural modal components of the several different frequencies and the filtered factors, construct an LSTM model, and perform optimization calibration on super parameters of the LSTM model by using an archimedes algorithm to obtain an improved simulated runoff series.
Fig. 8 illustrates a physical structure diagram of an electronic device, as shown in fig. 8, which may include: processor 810, communication interface (Communications Interface) 820, memory 830, and communication bus 840, wherein processor 810, communication interface 820, memory 830 accomplish communication with each other through communication bus 840. The processor 810 may invoke logic instructions in the memory 830 to perform a method of processing a global climate mode-coupled hydrologic model simulation, the method comprising: establishing a VIC distributed drainage basin hydrological model based on drainage basin actual measurement meteorological hydrological data, global climate mode data and underlying surface data of a region to be measured; obtaining simulated daily runoff according to the VIC distributed watershed hydrological model, and decomposing the simulated daily runoff into a plurality of different-frequency natural modal components by adopting variational modal decomposition; collecting global key sea area sea temperature factors and atmospheric circulation factors, and screening the global key sea area sea temperature factors and the atmospheric circulation factors by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors; and constructing an input factor set by the natural modal components with different frequencies and the screened factors, constructing an LSTM model, and optimizing and calibrating the super parameters of the LSTM model by adopting an Archimedes algorithm to obtain an improved simulated runoff series.
Further, the logic instructions in the memory 830 described above may be implemented in the form of software functional units and may be stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method of processing a global climate mode coupled hydrological model simulation provided by the methods described above, the method comprising: establishing a VIC distributed drainage basin hydrological model based on drainage basin actual measurement meteorological hydrological data, global climate mode data and underlying surface data of a region to be measured; obtaining simulated daily runoff according to the VIC distributed watershed hydrological model, and decomposing the simulated daily runoff into a plurality of different-frequency natural modal components by adopting variational modal decomposition; collecting global key sea area sea temperature factors and atmospheric circulation factors, and screening the global key sea area sea temperature factors and the atmospheric circulation factors by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors; and constructing an input factor set by the natural modal components with different frequencies and the screened factors, constructing an LSTM model, and optimizing and calibrating the super parameters of the LSTM model by adopting an Archimedes algorithm to obtain an improved simulated runoff series.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for processing a global climate mode coupled hydrological model simulation, comprising:
establishing a variable infiltration capacity VIC distributed drainage basin hydrological model based on drainage basin actual measurement meteorological hydrological data, global climate mode data and underlying surface data of a region to be measured;
Obtaining simulated daily runoff according to the VIC distributed watershed hydrological model, and decomposing the simulated daily runoff into a plurality of different-frequency natural modal components by adopting variational modal decomposition;
collecting global key sea area sea temperature factors and atmospheric circulation factors, and screening the global key sea area sea temperature factors and the atmospheric circulation factors by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors;
and constructing an input factor set by the natural modal components with different frequencies and the screened factors, constructing a long-term memory LSTM model, and optimizing and calibrating the super parameters of the LSTM model by adopting an Archimedes algorithm to obtain an improved simulated runoff series.
2. The method for processing the global climate pattern coupled hydrologic model simulation according to claim 1, wherein establishing the VIC distributed drainage basin hydrologic model based on the drainage basin measured meteorological hydrologic data, the global climate pattern data and the underlying surface data of the area to be measured comprises:
Collecting daily runoff process data of a drainage basin outlet station as actual measurement hydrological data, taking a CN05.1 grid point data set as actual measurement meteorological data, taking a sixth international coupling mode comparison plan CMIP6 data as global climate mode data, and collecting underlying surface data comprising soil data, vegetation type data, land utilization type data, topographic data and elevation data;
Interpolating the measured hydrological data, the measured meteorological data, the global climate mode data and the underlying surface data to a preset resolution grid through bilinear interpolation technology, and carrying out aggregate average, uniform time length and time resolution on the global climate mode data to obtain an initial VIC distributed drainage basin hydrological model;
And carrying out parameter calibration on preset key hydrologic parameters in the initial VIC distributed watershed hydrologic model by using a calibration objective function by adopting a complex evolution optimization algorithm SCE-UA algorithm, determining a Nash efficiency coefficient and a water yield error objective function, and training the initial VIC distributed watershed hydrologic model to obtain the VIC distributed watershed hydrologic model.
3. The method for processing the global climate mode coupling hydrologic model simulation according to claim 1, wherein obtaining a simulated daily runoff according to the VIC distributed watershed hydrologic model, decomposing the simulated daily runoff into a plurality of different frequency natural mode components by adopting variation mode decomposition, comprises:
wherein f (t) is original runoff data, t is time component, For partial derivative calculation, delta (t) is an impulse function, j is an imaginary unit, e is a natural index, K is the total number of the natural modal components, K is any natural modal component number, { u k}＝{u1(t),u2(t),…,uk (t) } is K natural modal components after decomposition, { omega k}＝{ω1(t),ω2(t),…,ωk (t) } is the center frequency corresponding to each modal component, and x is a convolution operator;
And solving by adopting a secondary penalty factor and a Lagrange operator, and carrying out iterative search by adopting an alternate direction multiplier method to obtain the natural modal components with different frequencies.
4. The method for processing the global climate mode coupling hydrologic model simulation according to claim 1, wherein the method for acquiring the global key sea area sea temperature factor and the atmospheric circulation factor, and screening the global key sea area sea temperature factor and the atmospheric circulation factor by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors comprises:
acquiring daily sea temperature data of a global preset key sea area within a preset time range, and acquiring a global key remote related atmospheric sphere index;
Based on annual time lag, calculating the pearson correlation coefficient of the daily sea temperature data and the remote correlation atmospheric sphere index, and screening a plurality of factors with preset ranking pearson correlation coefficients;
Screening a plurality of factors of the preset ranking pearson correlation coefficient by using a preset confidence inspection factor by adopting the stepwise multiple regression method to obtain a prediction candidate factor;
And (5) rejecting and introducing inspection by using a regression coefficient augmentation matrix to obtain the screened factors.
5. The method for processing the global climate pattern coupled hydrologic model simulation according to claim 4, wherein the steps of eliminating and checking by using regression coefficient augmentation matrix to obtain the screened factors include:
calculating variance contribution values, introducing factors which affect the dependent variables most significantly one by one into the regression coefficient amplification matrix, and performing F test on the introduced independent variables one by one;
and screening out factors with the smallest variance contribution from regression equations corresponding to the regression coefficient augmentation matrix which is introduced with factors with the most significant influence on dependent variables, performing rejection test, and screening out factors with the largest variance contribution from factors which are not introduced with the regression equations, performing introduction test, thus obtaining the screened factors.
6. The method for processing the global climate mode coupling hydrologic model simulation according to claim 1, wherein the steps of constructing an input factor set from the natural modal components of the different frequencies and the screened factors, constructing a long-term memory LSTM model, optimizing and calibrating super parameters of the LSTM model by using archimedes algorithm, and obtaining an improved simulated runoff series include:
determining that the LSTM model comprises two full-connection layers, connecting a Dropout layer with preset parameters after each full-connection layer, optimizing the LSTM model by adopting an Adam algorithm, and converging by adopting a preset loss function;
Optimizing and calibrating the number of hidden layer neurons, the initial learning rate, the minimum batch size and the maximum training times of the LSTM model by adopting the Archimedes algorithm to obtain an optimized simulated runoff series;
And calibrating the optimized simulated runoff series by utilizing a calibration objective function to obtain the improved simulated runoff series.
7. The method for processing the global climate model coupled hydrologic model simulation according to claim 6, wherein optimizing the hidden layer neuron number, the initial learning rate, the minimum batch size and the maximum training number of the LSTM model by adopting the archimedes algorithm to obtain an optimized simulated runoff series comprises:
Step 1: randomly generating a plurality of individuals, initializing the volume, density and acceleration of each individual, and determining the optimal volume, optimal density and optimal acceleration of the individual with optimal fitness value;
step 2: updating the densities and volumes of the remaining individuals based on the individual optimal volumes, the individual optimal densities, and the individual optimal accelerations;
Step3: determining a transfer operator and a density reduction factor according to the current iteration times and the maximum iteration times;
Step 4: if the transfer operator is determined to be smaller than or equal to a preset threshold value, global exploration is carried out, a global individual acceleration updating formula is obtained through the acceleration, density and volume of random materials and the acceleration, density and volume of an individual in the next iteration, the global individual acceleration updating formula is normalized, and a global individual position updating formula is obtained through combining the current any one of the body positions and the current random individual position;
Step 5: if the transfer operator is determined to be greater than a preset threshold value, carrying out local development, obtaining a local individual acceleration updating formula by the individual optimal volume, the individual optimal density, the individual optimal acceleration and the acceleration, density and volume of an individual in the next iteration, normalizing the local individual acceleration updating formula, and obtaining a local individual position updating formula by combining the current optimal position of any one of the individual, the current optimal individual position and the movement direction change mark;
Step 6: and (5) repeatedly executing the steps 2 to 5 until the maximum iteration times are reached, and obtaining the optimized simulated runoff series.
8. A processing system for global climate mode coupled hydrological model simulation, comprising:
The establishing module is used for establishing a VIC distributed drainage basin hydrological model based on drainage basin actual measurement meteorological hydrological data, global climate mode data and underlying surface data of the to-be-detected region;
The decomposition module is used for obtaining simulated daily runoff according to the VIC distributed watershed hydrological model, and decomposing the simulated daily runoff into a plurality of natural modal components with different frequencies by adopting variation modal decomposition;
The screening module is used for collecting global key sea area sea temperature factors and atmospheric circulation factors, and screening the global key sea area sea temperature factors and the atmospheric circulation factors by adopting a correlation coefficient method and a stepwise multiple regression method to obtain screened factors;
and the optimization module is used for constructing an input factor set by the natural modal components with the different frequencies and the screened factors, constructing an LSTM model, and optimizing and calibrating the super parameters of the LSTM model by adopting an Archimedes algorithm to obtain an improved simulated runoff series.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements a processing method of the global climate mode-coupled hydrological model simulation as claimed in any of claims 1 to 7 when the program is executed.
10. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor, implements a method of processing a global climate mode-coupled hydrological model simulation as claimed in any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310896277.6A CN117113808B (en) | 2023-07-20 | 2023-07-20 | Global climate mode coupling hydrologic model simulation processing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310896277.6A CN117113808B (en) | 2023-07-20 | 2023-07-20 | Global climate mode coupling hydrologic model simulation processing method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117113808A CN117113808A (en) | 2023-11-24 |
| CN117113808B true CN117113808B (en) | 2024-05-10 |
Family
ID=88801080
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310896277.6A Active CN117113808B (en) | 2023-07-20 | 2023-07-20 | Global climate mode coupling hydrologic model simulation processing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117113808B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105893770A (en) * | 2016-04-15 | 2016-08-24 | 山东省水利科学研究院 | Method for quantifying influence on basin water resources by climate change and human activities |
| CN110490366A (en) * | 2019-07-15 | 2019-11-22 | 西安理工大学 | Runoff forestry method based on variation mode decomposition and iteration decision tree |
| CN111695290A (en) * | 2020-05-14 | 2020-09-22 | 天津大学 | Short-term runoff intelligent forecasting hybrid model method suitable for variable environment |
| CN112765912A (en) * | 2021-01-26 | 2021-05-07 | 武汉大学 | Evaluation method for social and economic exposure degree of flood disasters based on climate mode set |
| CN114386334A (en) * | 2022-01-19 | 2022-04-22 | 浙江大学 | Runoff rolling forecasting method based on distributed hydrological runoff simulation substitution model |
| CN115907062A (en) * | 2022-04-21 | 2023-04-04 | 河南大学 | Hydrological forecasting method based on uniform design and artificial neural network |
| CN116227785A (en) * | 2023-03-01 | 2023-06-06 | 河海大学 | Runoff change attribution method, device and system considering climate-vegetation-water taking influence |
| CN116451879A (en) * | 2023-06-16 | 2023-07-18 | 武汉大学 | Drought risk prediction method and system and electronic equipment |
-
2023
- 2023-07-20 CN CN202310896277.6A patent/CN117113808B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105893770A (en) * | 2016-04-15 | 2016-08-24 | 山东省水利科学研究院 | Method for quantifying influence on basin water resources by climate change and human activities |
| CN110490366A (en) * | 2019-07-15 | 2019-11-22 | 西安理工大学 | Runoff forestry method based on variation mode decomposition and iteration decision tree |
| CN111695290A (en) * | 2020-05-14 | 2020-09-22 | 天津大学 | Short-term runoff intelligent forecasting hybrid model method suitable for variable environment |
| CN112765912A (en) * | 2021-01-26 | 2021-05-07 | 武汉大学 | Evaluation method for social and economic exposure degree of flood disasters based on climate mode set |
| CN114386334A (en) * | 2022-01-19 | 2022-04-22 | 浙江大学 | Runoff rolling forecasting method based on distributed hydrological runoff simulation substitution model |
| CN115907062A (en) * | 2022-04-21 | 2023-04-04 | 河南大学 | Hydrological forecasting method based on uniform design and artificial neural network |
| CN116227785A (en) * | 2023-03-01 | 2023-06-06 | 河海大学 | Runoff change attribution method, device and system considering climate-vegetation-water taking influence |
| CN116451879A (en) * | 2023-06-16 | 2023-07-18 | 武汉大学 | Drought risk prediction method and system and electronic equipment |
Non-Patent Citations (3)
| Title |
|---|
| 基于VIC水文模型的滦河流域径流变化特征及其影响因素;陈宏;尉英华;王颖;张余庆;左涛;邢雯慧;;干旱气象;20171031(05);全文 * |
| 考虑次网格水文过程的区域气候-水文耦合方法;陈星;余钟波;崔广柏;;河海大学学报(自然科学版);20080525(03);全文 * |
| 陆面水文―气候耦合模拟研究进展;占车生;宁理科;邹靖;韩建;;地理学报;20180419(05);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117113808A (en) | 2023-11-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Samadianfard et al. | Wavelet neural networks and gene expression programming models to predict short-term soil temperature at different depths | |
| Najafi et al. | Statistical downscaling of precipitation using machine learning with optimal predictor selection | |
| Mendoza et al. | An intercomparison of approaches for improving operational seasonal streamflow forecasts | |
| Marofi et al. | Predicting spatial distribution of snow water equivalent using multivariate non-linear regression and computational intelligence methods | |
| Kar et al. | Rain gauge network design for flood forecasting using multi-criteria decision analysis and clustering techniques in lower Mahanadi river basin, India | |
| CN111915058B (en) | Flood prediction method and device based on long-time memory network and transfer learning | |
| CN115423163A (en) | Method and device for predicting short-term flood events of drainage basin and terminal equipment | |
| Abd-Elmaboud et al. | Estimation of groundwater recharge using simulation-optimization model and cascade forward ANN at East Nile Delta aquifer, Egypt | |
| Shiri et al. | Coupling wavelet transform with multivariate adaptive regression spline for simulating suspended sediment load: independent testing approach | |
| CN115859116A (en) | Marine environment field reconstruction method based on radial basis function regression interpolation method | |
| Heddam et al. | River water temperature prediction using hybrid machine learning coupled signal decomposition: EWT versus MODWT | |
| CN117408171B (en) | Hydrologic set forecasting method of Copula multi-model condition processor | |
| Song et al. | Application of a novel signal decomposition prediction model in minute sea level prediction | |
| Badyalina et al. | Streamflow estimation at ungauged basin using modified group method of data handling | |
| Gao et al. | A framework for automatic calibration of SWMM considering input uncertainty | |
| CN117113808B (en) | Global climate mode coupling hydrologic model simulation processing method and system | |
| Solaiman | Uncertainty estimation of extreme precipitations under climate change: A non-parametric approach | |
| Meng et al. | Variable infiltration capacity model with BGSA-based wavelet neural network | |
| Shin et al. | Development of non-parametric evolutionary algorithm for predicting soil moisture dynamics | |
| Nazemi et al. | Extracting a set of robust Pareto-optimal parameters for hydrologic models using NSGA-II and SCEM | |
| Dauji et al. | Improving numerical current prediction with Model Tree | |
| Achite et al. | Application of multiple spatial interpolation approaches to annual rainfall data in the Wadi Cheliff basin (north Algeria) | |
| Saha et al. | Application of machine learning and emerging remote sensing techniques in hydrology: A state-of-the-art review and current research trends | |
| Cui et al. | A secondary modal decomposition ensemble deep learning model for groundwater level prediction using multi-data | |
| Zamani et al. | Hybrid WT–CNN–GRU-based model for the estimation of reservoir water quality variables considering spatio-temporal features |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |