Method and system for comprehensive comparison and selection of cmip6 climate models

Abstract

The disclosure discloses a CMIP6 climate model comprehensive comparison and selection method and system, relates to the technical field of data processing, calculates first index data of each climate model through meteorological measured data and CMIP6 climate model data, calculates first index scores of simulation capabilities of each climate model on one meteorological element according to the first index data and second index scores of simulation capabilities of each climate model on another meteorological element, averages the sum of the first index scores and the second index scores corresponding to each climate model to obtain a comprehensive index score corresponding to each climate model, and selects a preset number of climate models from the climate models in order from large to small according to the comprehensive index scores, wherein the selection process of the climate models comprehensively considers simulation effects when different climate elements simulate different meteorological elements, so that the selected climate models can achieve good simulation effects when simulating different meteorological elements.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of data processing, and particularly relates to a CMIP6 climate model comprehensive comparison and selection method and system. BACKGROUND

[0002] In recent years, global climate continues to warm, and climate change has become a focus problem of common concern in today's society. Climate warming will intensify global and regional water cycles, leading to frequent occurrence of extreme hydrological events such as floods, droughts, extreme high temperatures and heavy precipitation, which has caused extremely serious impact on human safety, social economy and ecological environment, so that the demand for accurate prediction of future climate change is also increasing. In addition, accurate simulation and prediction of climate change can provide scientific basis for policy making, risk assessment and resource management.

[0003] In order to realize the simulation and prediction of climate change, CMIP6 (the sixth phase of the international coupled model intercomparison project) collects various climate models developed by global research institutions, which contains various climate forcing scenarios, and has become one of the most effective tools for international climate simulation and prediction. However, due to the differences in data structure, parameter setting, calculation method and other aspects of different climate models in CMIP6, the simulation performance and applicable scenarios of each model are significantly different. Therefore, when studying the climate of a certain region, it is crucial to select the climate model with the best applicability to the certain region from CMIP6.

[0004] At present, the comparison and selection of climate models are usually realized by calculating the correlation coefficient (R), standard deviation (STD) and root mean square error (RMSE) and other indicators of the simulation values of each climate model in CMIP6 and the observation values, and the simulation capabilities of each climate model are judged based on the foregoing indicators to select the appropriate climate model. However, when selecting the appropriate climate model based on the foregoing method, only a single meteorological element of each climate model is usually evaluated, and while ensuring the optimal simulation of a certain meteorological element, the same climate model cannot make other meteorological elements also achieve optimal simulation, thereby leading to the problem of low climate prediction accuracy. SUMMARY

[0005] The present disclosure provides a CMIP6 climate model comprehensive comparison and selection method and system, and the main purpose is to solve the problem that related technologies only evaluate a single meteorological element, and while ensuring the optimal simulation of a certain meteorological element, the same climate model cannot make other meteorological elements also achieve optimal simulation.

[0006] According to a first aspect of the present disclosure, a CMIP6 climate model comprehensive comparison and selection method is provided, which comprises:

[0007] Obtaining CMIP6 climate model data and meteorological measured data for a target study area, wherein the spatial range of the CMIP6 climate model data includes the target study area;

[0008] Calculating first indicator data of each climate model in the CMIP6 climate model data based on the measured meteorological data to obtain an indicator factor when each climate model in the climate models simulates at least two meteorological elements respectively, wherein the types of the first indicator data are normalized standard deviation, spatial correlation coefficient, mean deviation, and interannual standard deviation;

[0009] Calculating, based on a preset basis function and according to the index factors, a first index score of each climate model's ability to simulate one meteorological element and a second index score of each climate model's ability to simulate another meteorological element, respectively, the at least two meteorological elements including the one meteorological element and the another meteorological element;

[0010] averaging the sum of the first index score and the second index score corresponding to each climate mode to obtain a comprehensive index score corresponding to each climate mode;

[0011] Selecting a preset number of climate models from the climate models in descending order of the comprehensive index scores;

[0012] Based on the Bayesian model weighted averaging method, a weighted ensemble average is performed on the preset number of climate models to obtain a coupled climate model, which is used to simulate the changing laws of future meteorological elements.

[0013] Optionally, the calculating, based on the measured meteorological data, first indicator data of each climate model in the CMIP6 climate model data to obtain an indicator factor when each climate model in the climate models simulates at least two meteorological elements respectively includes:

[0014] sequentially calculating a target normalized standard deviation, a target spatial correlation coefficient, a target mean deviation, and a target interannual standard deviation of each climate model in each climate model relative to the meteorological measured data;

[0015] Performing factor conversion on the target normalized standard deviation based on a normalized standard deviation conversion function to obtain a first conversion factor;

[0016] Performing factor conversion on the target spatial correlation coefficient based on a spatial correlation coefficient conversion function to obtain a second conversion factor;

[0017] Performing factor conversion on the target average deviation based on the average deviation conversion function to obtain a third conversion factor;

[0018] The target interannual standard deviation is factor-transformed based on an interannual standard deviation transformation function to obtain a fourth transformation factor, and the index factor is obtained by weighting the first transformation factor, the second transformation factor, the third transformation factor and the fourth transformation factor, so as to obtain the index factor when each climate mode simulates at least two meteorological elements respectively.

[0019] The formula of the normalization standard deviation transformation function is:

[0020] X1=-|x1-1|

[0021] wherein X1 is the first transformation factor, and x1 is the target normalization standard deviation.

[0022] The formula of the spatial correlation coefficient transformation function is:

[0023] X2=x2-1

[0024] wherein X2 is the second transformation factor, and x2 is the target spatial correlation coefficient.

[0025] The formula of the average deviation transformation function is:

[0026] X3=-|x3|

[0027] wherein X3 is the third transformation factor, and x3 is the target average deviation.

[0028] The formula of the interannual standard deviation transformation function is:

[0029] X4=-x4

[0030] wherein X4 is the fourth transformation factor, and x4 is the target interannual standard deviation.

[0031] Optionally, the formula of the preset base function is:

[0032] S(X)=β σ(X+δ)

[0033] wherein δ represents a tolerance, δ=|maxE(X)|, E(X) is a set composed of the index factors corresponding to the simulation of the meteorological element by the climate modes or a set composed of the index factors corresponding to the simulation of the other meteorological element by the climate modes, X is the index factor corresponding to the simulation of the meteorological element by the climate modes or the index factor corresponding to the simulation of the other meteorological element by the climate modes; σ is a correction factor; β is a base number, and S(X) is the first exponential score or the second exponential score.

[0034] Optionally, the method comprises:

[0035] According to the CMIP6 climate model data and the meteorological measured data, spatial distribution diagrams corresponding to the at least two meteorological elements in each climate model are drawn in sequence, to obtain a first spatial distribution diagram corresponding to one meteorological element in each climate model, and a second spatial distribution diagram corresponding to another meteorological element in each climate model;

[0036] The spatial distribution feature differences between the first spatial distribution diagrams in each climate model are compared to obtain a first spatial distribution feature difference, and the spatial distribution feature differences between the second spatial distribution diagrams in each climate model are compared to obtain a second spatial distribution feature difference;

[0037] A first spatial correlation coefficient between first simulation data of the one meteorological element simulated by each climate model and first target measured data corresponding to the first simulation data in the meteorological measured data is calculated respectively;

[0038] A second spatial correlation coefficient between second simulation data of the another meteorological element simulated by each climate model and second target measured data corresponding to the second simulation data in the meteorological measured data is calculated respectively, and the CMIP6 climate model data includes the first simulation data and the second simulation data;

[0039] According to the first spatial distribution feature difference and the first spatial correlation coefficient, a first target climate model with the best simulation effect on the spatial feature of the one meteorological element is selected from the climate models;

[0040] According to the second spatial distribution feature difference and the second spatial correlation coefficient, a second target climate model with the best simulation effect on the spatial feature of the another meteorological element is selected from the climate models.

[0041] Optionally, the method comprises:

[0042] According to a first interannual standard deviation, a third target climate model with the best simulation effect on the time feature of the one meteorological element is selected from the climate models, and the first interannual standard deviation is used to describe the difference in interannual fluctuation amplitude between the first simulation data and the first target measured data, so as to judge the simulation effect of each climate model on the time feature of the one meteorological element;

[0043] selecting a fourth target climate model having the best simulation effect on the time characteristic of the other meteorological element from the climate models according to the second interannual standard deviation, the second interannual standard deviation being used to describe the difference in interannual fluctuation amplitude between the second simulation data and the second target measured data so as to judge the simulation effect of the climate models on the time characteristic of the other meteorological element, the first interannual standard deviation and the second interannual standard deviation being calculated by an interannual standard deviation function;

[0044] The formula of the interannual standard deviation function is as follows:

[0045]

[0046] wherein STD m represents the interannual standard deviation of the climate models, STD o represents the interannual standard deviation of the meteorological measured data, and IVS represents the interannual standard deviation.

[0047] Optionally, the method comprises:

[0048] calculating the spatial correlation coefficient, the standard deviation and the root mean square error of the climate models and the meteorological measured data;

[0049] calculating a first performance score of the climate models in simulating the meteorological element according to the spatial correlation coefficient, the standard deviation and the root mean square error, and selecting a fifth target climate model having the best simulation performance on the meteorological element from the climate models based on the first performance score;

[0050] calculating a second performance score of the climate models in simulating the other meteorological element according to the spatial correlation coefficient, the standard deviation and the root mean square error, and selecting a sixth target climate model having the best simulation performance on the other meteorological element from the climate models based on the second performance score.

[0051] According to a second aspect of the present disclosure, a CMIP6 climate model comprehensive comparison and selection system is provided, comprising:

[0052] an acquisition unit configured to acquire CMIP6 climate model data and meteorological measured data of a target research region, a spatial range of the CMIP6 climate model data including the target research region;

[0053] The first computing unit is configured to calculate first index data of each climate model in the CMIP6 climate model data based on the meteorological observation data, so as to obtain an index factor of each climate model when simulating at least two meteorological elements respectively, and the first index data includes a normalized standard deviation, a spatial correlation coefficient, an average deviation and an interannual standard deviation.

[0054] The second computing unit is configured to calculate a first index score of the simulation capability of each climate model for one meteorological element and a second index score of the simulation capability of each climate model for another meteorological element based on the index factor and a preset basis function.

[0055] The third computing unit is configured to average the sum of the first index score and the second index score corresponding to each climate model, so as to obtain a comprehensive index score corresponding to each climate model.

[0056] The selecting unit is configured to select a preset number of climate models from the climate models in a descending order of the comprehensive index scores.

[0057] The processing unit is configured to perform weighted set averaging on the preset number of climate models based on a Bayesian model weighted average method, so as to obtain a coupled climate model, and the coupled climate model is used to simulate a change rule of a future meteorological element.

[0058] According to a third aspect of the present disclosure, an electronic device is provided, comprising:

[0059] at least one processor; and

[0060] a memory connected with the at least one processor; wherein

[0061] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of the first aspect.

[0062] According to a fourth aspect of the present disclosure, a non-transitory computer readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to perform the method of the first aspect.

[0063] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the method of the first aspect.

[0064] The CMIP6 climate model comprehensive comparison and selection method and system provided by the present disclosure, obtain CMIP6 climate model data and meteorological measured data of a target research area, the spatial range of the CMIP6 climate model data includes the target research area; calculate first index data of each climate model in the CMIP6 climate model data based on the meteorological measured data, obtain index factors of each climate model when simulating at least two meteorological elements respectively, the types of the first index data are normalized standard deviation, spatial correlation coefficient, average deviation and interannual standard deviation; calculate a first index score of the simulation capability of each climate model for one meteorological element and a second index score of the simulation capability of each climate model for another meteorological element based on a preset basis function according to the index factors, the at least two meteorological elements include the one meteorological element and the another meteorological element; average the sum of the first index score and the second index score corresponding to each climate model to obtain a comprehensive index score corresponding to each climate model; select a preset number of climate models from the climate models in order from large to small according to the comprehensive index scores; perform weighted set average on the preset number of climate models based on a Bayesian model weighted average method to obtain a coupled climate model, the coupled climate model is used to simulate the change rule of future meteorological elements. Compared with related technologies, the first index data of each climate model is calculated through the meteorological measured data and the CMIP6 climate model data, and the first index score of the simulation capability of each climate model for one meteorological element and the second index score of the simulation capability of each climate model for another meteorological element are calculated according to the first index data, the sum of the first index score and the second index score corresponding to each climate model is averaged to obtain a comprehensive index score corresponding to each climate model; a preset number of climate models are selected from the climate models in order from large to small according to the comprehensive index scores, and the selection process of the foregoing climate models comprehensively considers the simulation effects of the climate models when simulating different meteorological elements, so that the selected climate models can achieve good simulation effects when simulating different meteorological elements, and further, a coupled climate model is obtained by performing weighted set average on the selected climate models, and the simulation effect is improved when the coupled climate model is used to simulate different meteorological elements. BRIEF DESCRIPTION OF DRAWINGS

[0065] The accompanying drawings, which form a part of the present disclosure, are intended to provide further understanding of the present disclosure, and the illustrative embodiments of the present disclosure and their descriptions serve the purpose of explaining the present disclosure, and do not constitute improper limitations on the present disclosure. In the drawings:

[0066] Figure 1A CMIP6 climate model comprehensive comparison and selection method flowchart provided by an embodiment of the present disclosure;

[0067] Figure 2 A spatial distribution diagram of multi-year average precipitation and air temperature under climate models and measured models provided by an embodiment of the present disclosure;

[0068] Figure 3 A comparison diagram of simulation performance of interannual variability of meteorological elements by each climate model provided by an embodiment of the present disclosure;

[0069] Figure 4 A multi-model evaluation Taylor diagram provided by an embodiment of the present disclosure;

[0070] Figure 5 A performance ranking heatmap of each climate model for three indicators of mean, root mean square error and correlation coefficient provided by an embodiment of the present disclosure;

[0071] Figure 6 A multi-element simulation effect comprehensive evaluation diagram provided by an embodiment of the present disclosure;

[0072] Figure 7 A schematic diagram of future air temperature and rainfall relative to historical period changes under two scenarios provided by an embodiment of the present disclosure;

[0073] Figure 8 A structure schematic diagram of a CMIP6 climate model comprehensive comparison and selection system provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0074] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0075] The following detailed description is exemplary description, which is intended to provide further detailed description of the present application. Unless otherwise specified, all technical terms used in the present application have the same meaning as generally understood by those skilled in the art to which the present application belongs. The terms used in the present application are only for describing the specific embodiments, and are not intended to limit the exemplary embodiments according to the present application.

[0076] In addition, the terms "first", "second", and the like in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein.

[0077] The CMIP6 climate model comprehensive comparison and selection method and system of the embodiments of the present disclosure are described below with reference to the accompanying drawings.

[0078] To at least solve the problem that the related art only evaluates a single meteorological element, resulting in that when a same climate model ensures optimal simulation of a certain meteorological element, other meteorological elements cannot be simulated optimally. The embodiments provide a CMIP6 climate model comprehensive comparison and selection method.

[0079] Figure 1 A CMIP6 climate model comprehensive comparison and selection method flowchart provided by the embodiments of the present disclosure is shown in FIG. 1. As shown in FIG. 1, the method comprises the following steps: Figure 1

[0080] Step 101, obtaining CMIP6 climate model data and meteorological measured data of a target research region, wherein a spatial range of the CMIP6 climate model data comprises the target research region.

[0081] Step 102, calculating first index data of each climate model in the CMIP6 climate model data based on the meteorological measured data, to obtain an index factor of each climate model when simulating at least two meteorological elements, wherein types of the first index data are normalized standard deviation, spatial correlation coefficient, average deviation, and interannual standard deviation.

[0082] Step 103, calculating a first exponential score of a simulation capability of each climate model for one meteorological element and a second exponential score of a simulation capability of each climate model for another meteorological element based on a preset base function according to the index factor, wherein the at least two meteorological elements comprise the one meteorological element and the another meteorological element.

[0083] Step 104, averaging a sum of the first exponential score and the second exponential score corresponding to each climate model, to obtain a comprehensive exponential score corresponding to each climate model.

[0084] Step 105, selecting a preset number of climate models from the climate models in a descending order of the comprehensive exponential scores.

[0085] Step 106, performing weighted set averaging on the preset number of climate models based on a Bayesian model weighted averaging method, to obtain a coupled climate model, wherein the coupled climate model is used to simulate a change rule of a future meteorological element.

[0086] ​The CMIP6 climate model comprehensive comparison and selection method provided by the present disclosure obtains CMIP6 climate model data and meteorological measured data of a target research region, the spatial range of the CMIP6 climate model data includes the target research region; based on the meteorological measured data, first index data of each climate model in the CMIP6 climate model data is calculated, to obtain an index factor of each climate model when simulating at least two meteorological elements, respectively, the types of the first index data are normalized standard deviation, spatial correlation coefficient, average deviation and interannual standard deviation; based on a preset basis function, a first exponential score of the simulation capability of each climate model for one meteorological element and a second exponential score of the simulation capability of each climate model for another meteorological element are calculated according to the index factor, the at least two meteorological elements include the one meteorological element and the another meteorological element; the sum of the first exponential score and the second exponential score corresponding to each climate model is averaged to obtain a comprehensive exponential score corresponding to each climate model; a preset number of climate models are selected from the climate models in order from large to small according to the comprehensive exponential score; the preset number of climate models are weighted and set-averaged based on a Bayesian model weighted average method to obtain a coupled climate model, the coupled climate model is used to simulate the change rule of future meteorological elements. Compared with related technologies, the first index data of each climate model is calculated through the meteorological measured data and the CMIP6 climate model data, the first exponential score of the simulation capability of each climate model for one meteorological element and the second exponential score of the simulation capability of each climate model for another meteorological element are calculated according to the first index data, the sum of the first exponential score and the second exponential score corresponding to each climate model is averaged to obtain a comprehensive exponential score corresponding to each climate model; a preset number of climate models are selected from the climate models in order from large to small according to the comprehensive exponential score, the selection process of the foregoing climate model comprehensively considers the simulation effect of each climate element when simulating different meteorological elements, so that the selected climate model can achieve good simulation effect when simulating different meteorological elements, further, the coupled climate model is obtained by weighted set-averaging the selected climate model, and the simulation effect is improved when the coupled climate model is used to simulate different meteorological elements.

[0087] As a refinement of the disclosed embodiment, when executing step 101 to obtain CMIP6 climate model data and measured meteorological data for the target study area, where the spatial range of the CMIP6 climate model data includes the target study area, the following implementation methods may also be used, but are not limited to: downloading and collecting the NEX-GDDP-CMIP6 climate model dataset, where the NEX-GDDP-CMIP6 climate model dataset is the CMIP6 climate model data. The purpose of this step is to directly obtain 23 high-resolution downscaled CMIP6 climate models. All models in the NEX-GDDP-CMIP6 climate model dataset have been uniformly downscaled to 0.25° based on bias correction and spatial decomposition methods. Downloading the National Ground Meteorological Station Basic Meteorological Element Daily Value Dataset (V3.0), from which historical long-term series daily meteorological data from each station in the target study area are collected to obtain the measured meteorological data, which are used to evaluate and verify the simulation effects of each climate model. If the target study area has relatively sparse meteorological stations, the fifth-generation global reanalysis dataset (ERA5) released by the European Centre for Medium-Range Weather Forecasts (ECWMF) can be collected as ground-based data, which is an effective supplement to the meteorological measured data. The meteorological models mentioned include but are not limited to 23 high-resolution downscaled CMIP6 climate models. Strict quality control is performed on the collected meteorological measured data, and a small number of missing values ​​and outliers in the data are processed in a unified and centralized manner to ensure the continuity, completeness and accuracy of the data. The length of the time series of the collected CMIP6 climate model historical data must be consistent with the historical data of the meteorological stations, and the spatial range of the data must ensure that it covers the entire target study area.

[0088] As a refinement of the above embodiment, in the step of calculating the first index data of each climate model in the CMIP6 climate model data based on the meteorological observation data, to obtain the index factor of each climate model when simulating at least two meteorological elements, the following implementation manners can also be used, but are not limited to, for example: sequentially calculating the target normalized standard deviation, target spatial correlation coefficient, target mean deviation and target interannual standard deviation of each climate model with respect to the meteorological observation data; factor transforming the target normalized standard deviation based on the normalized standard deviation transformation function to obtain a first transformation factor; factor transforming the target spatial correlation coefficient based on the spatial correlation coefficient transformation function to obtain a second transformation factor; factor transforming the target mean deviation based on the mean deviation transformation function to obtain a third transformation factor; factor transforming the target interannual standard deviation based on the interannual standard deviation transformation function to obtain a fourth transformation factor, and the index factor is obtained by weighting the first transformation factor, the second transformation factor, the third transformation factor and the fourth transformation factor, so as to obtain the index factor of each climate model when simulating at least two meteorological elements; the formula of the normalized standard deviation transformation function is:

[0089] X1=-|x1-1|

[0090] wherein X1 is the first transformation factor, and x1 is the target normalized standard deviation;

[0091] The formula of the spatial correlation coefficient transformation function is:

[0092] X2=x2-1

[0093] wherein X2 is the second transformation factor, and x2 is the target spatial correlation coefficient;

[0094] The formula of the mean deviation transformation function is:

[0095] X3=-|x3|

[0096] wherein X3 is the third transformation factor, and x3 is the target mean deviation;

[0097] The formula of the interannual standard deviation transformation function is:

[0098] X4=-x4

[0099] wherein X4 is the fourth transformation factor, and x4 is the target interannual standard deviation.

[0100] As a refinement of the above embodiment, the formula of the preset base function is:

[0101] S(X)=βσ(X+δ)

[0102] wherein, δ represents a tolerance, δ is the absolute value of the maximum value in the set E(X), i.e. δ = |maxE(X)|, ensuring that the best performing model S(X) = 1; E(X) is a set composed of the index factors corresponding to the simulation of the one meteorological element by the climate models or a set composed of the index factors corresponding to the simulation of the other meteorological element by the climate models, X is the index factor corresponding to the simulation of the one meteorological element by the climate models or the index factor corresponding to the simulation of the other meteorological element by the climate models; σ is a correction factor aiming to correct the difference in order of magnitude between the correction coefficients, ensuring that S(X) ≤ 1, σ = aveE(x+δ) is selected; β is the base, the value of the parameter β needs to be iteratively calibrated, and the value of β that satisfies the maximum standard deviation of S(X) should be selected, S(X) is the first exponential score or the second exponential score.

[0103] In order to facilitate the understanding of the above-mentioned embodiments, the embodiments are further explained in combination with the detailed description of step 101, including: calculating the normalized standard deviation (NSTD), the spatial correlation coefficient (R), the average deviation from the measured value (MB), and the interannual standard deviation (IVS) of different climate models, and converting the aforementioned four indicators into index factors corresponding to the climate models according to the normalized standard deviation conversion function, the spatial correlation coefficient conversion function, the average deviation conversion function, and the interannual standard deviation conversion function, arranging the index factors X of the 23 climate models into a set E(X ∈ E), and further scoring each climate model by using an exponential function as a base function in order to amplify the final score of the model with better simulation effect, and the climate model with the highest final score meets the most evaluation criteria.

[0104] In order to facilitate the understanding of the processes involved in steps 104, 105, and 106, the embodiments are exemplarily explained by defining the at least two meteorological elements as rainfall and temperature, i.e. the rainfall is the one meteorological element and the temperature is the other meteorological element, and the specific explanation includes: averaging the sum of the simulation capability index scores of rainfall and temperature of each climate model to obtain the comprehensive score of different climate elements coupled with temperature and rainfall, ranking comprehensively according to the comprehensive score, and finally selecting the climate model with good simulation performance for both rainfall and temperature. The Bayesian model weighted average method is used to perform weighted set average on the output results of the top four climate models selected in the comprehensive ranking, and a coupled climate model is obtained. A future emission scenario is set, and the meteorological simulation values under the future scenario are output based on the coupled climate model, and then the change trend of each meteorological element under different future scenarios is obtained. The simulation capability index score of each climate model rainfall and temperature includes the first exponential score and the second exponential score, and the comprehensive score is the comprehensive exponential score.

[0105] As a refinement of the above-mentioned embodiment, the method further comprises: sequentially drawing spatial distribution diagrams corresponding to the at least two meteorological elements under each climate model according to the CMIP6 climate model data and the meteorological observation data, to obtain a first spatial distribution diagram corresponding to one meteorological element under each climate model, and a second spatial distribution diagram corresponding to another meteorological element under each climate model; comparing spatial distribution feature differences between the first spatial distribution diagrams under each climate model to obtain a first spatial distribution feature difference, and comparing spatial distribution feature differences between the second spatial distribution diagrams under each climate model to obtain a second spatial distribution feature difference; calculating a first spatial correlation coefficient between first simulation data and first target observation data for simulating the one meteorological element under each climate model, the first target observation data being data corresponding to the first simulation data in the meteorological observation data; calculating a second spatial correlation coefficient between second simulation data and second target observation data for simulating the another meteorological element under each climate model, the second target observation data being data corresponding to the second simulation data in the meteorological observation data, and the CMIP6 climate model data including the first simulation data and the second simulation data; selecting a first target climate model having the best simulation effect on the spatial feature of the one meteorological element from the climate models according to the first spatial distribution feature difference and the first spatial correlation coefficient; and selecting a second target climate model having the best simulation effect on the spatial feature of the another meteorological element from the climate models according to the second spatial distribution feature difference and the second spatial correlation coefficient.

[0106] As a refinement of the above-mentioned embodiment, the method further comprises: selecting a third target climate model having the best simulation effect on the time feature of the one meteorological element from the climate models according to a first interannual standard deviation, the first interannual standard deviation being used to describe a difference in interannual fluctuation amplitude between the first simulation data and the first target observation data to determine the simulation effect of each climate model on the time feature of the one meteorological element; and selecting a fourth target climate model having the best simulation effect on the time feature of the another meteorological element from the climate models according to a second interannual standard deviation, the second interannual standard deviation being used to describe a difference in interannual fluctuation amplitude between the second simulation data and the second target observation data to determine the simulation effect of each climate model on the time feature of the another meteorological element, and the first interannual standard deviation and the second interannual standard deviation being calculated by an interannual standard deviation function; and the formula of the interannual standard deviation function is as follows:

[0107]

[0108] Among them, STD m Represents the interannual standard deviation of each climate model, STD o represents the interannual standard deviation of the meteorological measured data, and IVS represents the interannual standard deviation.

[0109] In order to facilitate the understanding of the scheme involved in the above embodiment, this embodiment provides an exemplary explanation, including: drawing a schematic diagram of the spatial distribution of the multi-year average rainfall and temperature under 23 climate models and measured models, that is, under meteorological measured data, comparing and observing the differences in the spatial distribution characteristics of rainfall and temperature under each climate model, and statistically calculating the spatial correlation coefficient between each climate model and the measured value, that is, the spatial correlation coefficient between the simulated data of each climate model and its corresponding measured data, and preliminarily judging the climate model with better overall spatial simulation effect from the perspective of similar distribution patterns and higher spatial correlation coefficients. The performance of meteorological interannual variability simulation of each climate model is evaluated from the perspective of time change. Specifically, the difference in the interannual fluctuation amplitude of meteorological elements simulated by each climate model and the measured value is described by calculating the interannual standard deviation IVS index. The smaller the IVS, the better the climate model can simulate the interannual variability characteristics of meteorological elements.

[0110] As a refinement of the above embodiment, the method also includes: calculating the spatial correlation coefficient, standard deviation and root mean square error between the climate models and the measured meteorological data; calculating a first performance score of the climate models when simulating the one meteorological element based on the spatial correlation coefficient, the standard deviation and the root mean square error, and selecting the fifth target climate model with the best simulation performance of the one meteorological element from the climate models based on the first performance score; calculating a second performance score of the climate models when simulating the other meteorological element based on the spatial correlation coefficient, the standard deviation and the root mean square error, and selecting the sixth target climate model with the best simulation performance of the other meteorological element from the climate models based on the second performance score.

[0111] To facilitate understanding of the solutions involved in the above-mentioned embodiments, this embodiment provides an exemplary explanation, including: calculating three statistical indicators based on measured meteorological data and climate model data: spatial correlation coefficient (R), standard deviation (STD), and root mean square error (RMSE). A Taylor diagram is then plotted using these three indicators to visually demonstrate the performance of 23 climate models across the three simulation performance indicators. Based on the calculated STD, RMSE, and R results for each climate model, the performance of each indicator is scored and ranked, and the three indicators are combined to comprehensively evaluate the simulation performance of the 23 climate models for rainfall and temperature.

[0112] To more intuitively show the process of the implementation of the present disclosure, the following takes a certain water source area as an example, uses the CMIP6 climate model comprehensive comparison method to determine a climate model with higher applicability and higher accuracy for the rainfall and air temperature of the area, and then uses the climate model to simulate and predict the future air temperature and precipitation changes in the area as follows:

[0113] The daily precipitation and average air temperature measured data of 8 meteorological stations (Qingshui River, Shiqu, Ganzi, Xinlong, Daofu, Banma, Seda and Markang) from 1960 to 2020 in the water source area were collected, and the data were from the national ground meteorological station basic meteorological element daily value data set (V3.0). The data were strictly quality controlled, and a small amount of missing values and abnormal values were uniformly processed to ensure the continuity, integrity and accuracy of the data. At the same time, considering that the 8 meteorological stations are relatively dispersed and cannot cover the entire study area, the fifth generation global reanalysis data set (ERA5) released by the European Center for Medium-Range Weather Forecasts (ECWMF) was selected as an effective supplement to the ground measured data. The 23 high-resolution downscaling climate model data of NEX-GDDP-CMIP6 in the collection area were downloaded, and the two key meteorological element data of precipitation and air temperature in the study area from 1960 to 2020 (historical stage) were simulated and output by each climate model.

[0114] The simulation effects of the 23 climate models on the spatial distribution characteristics of rainfall and air temperature were evaluated. Based on the measured and model meteorological data collected in the study area, the spatial distribution maps of the multi-year average precipitation and the multi-year average air temperature from 1960 to 2020 under each climate model were drawn, as shown in Figure 2 Figure 2 (a) is the spatial distribution map of the multi-year average precipitation of each climate model. As can be seen from the figure, the simulation results of each climate model on the distribution characteristics of precipitation are generally similar, and basically can reflect the overall spatial variation trend of the multi-year precipitation. At the same time, the simulation effect is better for the northwest region with less precipitation, and the simulation value of the southeast region of the study area is slightly more. Further calculation of the spatial correlation coefficient of the rainfall simulation value of each climate model and the historical measured value shows that the spatial correlation coefficient of each climate model and the measured value reaches more than 0.86, among which CanESM5 has the best spatial simulation effect on rainfall, with a correlation coefficient of 0.88, and CNRM-CM6-1 and INM-CM4-8 and other climate models also have good simulation capabilities for the spatial distribution characteristics of precipitation. Figure 2 ​(b) is the spatial distribution map of the multi-year average temperature of each climate model. Overall, most of the climate models simulate lower temperatures in the study area compared to the measured values, but they can generally simulate the spatial distribution characteristics of temperature, i.e., the spatial characteristics of increasing from north to south. Among them, EC-Earth3-Veg-LR has the best simulation effect on the temperature of the study area, with a spatial correlation coefficient of 0.87, followed by Noresm2-MM and other climate models that also show good simulation ability of spatial distribution characteristics. The IVS index is used to further measure the simulation performance of each climate model on the interannual variability of meteorological elements. The results show that the best climate model in terms of simulation effect on the temporal characteristics of precipitation is MPI-ESM1-2-HR, and the worst is CanESM5. The better simulation models also include EC-Earth3, MRI-ESM2-0, CMCC-ESM2, etc. For temperature, the best climate model is Noresm2-MM, and the worst is CanESM5. Similarly, the climate models that simulate the temporal variation characteristics of temperature in the historical period well also include EC-Earth3-Veg-L, Taiesm1, IPSL-CM6A-LR, and CNRM-ESM2-1. Among them, the simulation performance of each climate model on the interannual variability of meteorological elements is shown in Figure 3 , Figure 3 Figure (a) is the IVS performance score comparison chart of precipitation, and figure (b) is the IVS performance score comparison chart of temperature.

[0115] The spatial correlation coefficient (R), standard deviation (STD), and root mean square error (RMSE) of each climate model and the measured value are calculated and standardized. At the same time, the Taylor diagram is drawn to evaluate the simulation performance of the selected 23 climate models in NEX-GDDP-CMIP6 on the temperature and precipitation of the study area in the historical period. According to the calculation results of the STD, RMSE and R of each model, the comprehensive index ranking of each model is obtained. The best simulation effect of precipitation is CanESM5, and the simulation effects of NroESM2-LM, INM-CM4-8, CNRM-CM6-1 and FGOALS-g3 climate models are also good. The climate models with good temperature simulation performance are NorESM2-LM, GISS-E2-1-G, CNRM-ESM2-1, CanESM5 and INM-CM5-0. The Taylor diagram is shown in Figure 4 , Figure 4 Figure (a) is the precipitation Taylor diagram of each climate model, and figure (b) is the temperature Taylor diagram of each climate model. The performance ranking of each climate model on the three indicators of mean, root mean square error and correlation coefficient is shown in Figure 5 , Figure 5 Figure (a) is the ranking chart of precipitation simulation performance, and figure (b) is the ranking chart of temperature simulation performance.

[0116] Since the same climate model cannot achieve the optimal simulation of precipitation and temperature in space and time, the exponential function is further used as the basis function to score each climate model, and the sum of the exponential scores of the precipitation and temperature simulation capabilities of each climate model is averaged to obtain the comprehensive score of the coupling of temperature and precipitation of different climate elements. According to the comprehensive score, a comprehensive ranking is made, and finally the climate model with good simulation performance of both precipitation and temperature is selected. Through the exponential scoring method, the simulation performance of temperature and precipitation of each climate model is comprehensively evaluated, as shown in Figure 6 , Figure 6 Fig. (a) is the precipitation simulation ranking, (b) is the temperature simulation ranking, and (c) is the comprehensive simulation ranking of precipitation and temperature. The comparison result shows that EC-Earth3, MRI-ESM2-0, Noresm2-MM and MPI-ESM1-2-HR have the best simulation effect in the study area, and the comprehensive scores of EC-Earth3 and MRI-ESM2-0 are both more than 0.5, indicating that they perform well in both temperature and precipitation.

[0117] Based on the comprehensive evaluation result, the Bayesian model weighted average method is used to couple the four climate models of EC-Earth3, MRI-ESM2-0, Noresm2-MM and MPI-ESM1-2-HR obtained by the foregoing comparison to form a climate coupling model.

[0118] According to the climate coupling model, the future change trend of temperature and precipitation in the study area is predicted under two emission scenarios SSP1-2.6 and SSP5-8.5. Considering the large difference between precipitation, the change is measured by anomaly percentage (%), and the difference of temperature is measured by anomaly value (℃). The trend prediction result is shown in Figure 7 , Figure 7 Fig. (a) is a schematic diagram of the change of future precipitation relative to the historical period, and Fig. (b) is a schematic diagram of the change of future temperature relative to the historical period. It can be seen that the future temperature and precipitation in the study area show a fluctuating upward trend under the two scenarios. The average growth of future precipitation under the two scenarios is 7.9% (SSP1-2.6) and 10.21% (SSP5-8.5). For temperature, by the end of the 21st century, the future temperature under the two scenarios increases by 1.64℃

[0119] (SSP1-2.6) and 3.72℃ (SSP5-8.5) respectively.

[0120] In summary, the present disclosure selects the best climate model for regional climate prediction by using a multi-dimensional, multi-model and multi-element climate model evaluation and comprehensive method, and obtains the optimal model by using the Bayesian model weighted average coupling, realizes the optimization of regional climate model selection, and more accurately predicts the change rule of future meteorological elements.

[0121] In summary, the embodiments of the present disclosure can achieve the following effects:

[0122] 1. The first index data of each climate model is calculated by the meteorological observation data and the CMIP6 climate model data, and the first index score of the simulation capability of each climate model for one meteorological element and the second index score of the simulation capability of each climate model for another meteorological element are calculated according to the first index data, respectively, the sum of the first index score and the second index score corresponding to each climate model is averaged to obtain the comprehensive index score corresponding to each climate model; the climate models of a preset number are selected from the climate models in order from large to small according to the comprehensive index score, and the selection process of the climate models comprehensively considers the simulation effects of the climate elements when simulating different meteorological elements, so that the selected climate models can achieve good simulation effects when simulating different meteorological elements, and further, the coupled climate model is obtained by weighted set average of the selected climate models, which realizes the improvement of the simulation effect when the coupled climate model is used to simulate different meteorological elements.

[0123] 2. The present disclosure introduces Taylor diagram, interannual standard deviation (IVS) and other evaluation indexes, which not only evaluates the consistency of the distribution characteristics of temperature and precipitation from the spatial dimension, but also evaluates the interannual variation from the time dimension, comprehensively reflects the simulation performance of different models on meteorological elements in space and time, and improves the comprehensiveness and reliability of climate model evaluation.

[0124] 3. In view of the current situation that the same climate model is difficult to achieve optimal simulation of temperature and precipitation in space and time, the present disclosure applies a comprehensive ranking method based on multiple statistical indexes, which forms a multi-element comprehensive ranking method by weighted summation of the normalized standard deviation (NSTD), spatial correlation coefficient (R), mean bias (MB) and other indexes of the climate model, effectively realizes the unified evaluation of temperature and precipitation multi-meteorological elements, and ensures the scientificity and objectivity of the ranking.

[0125] 4. The present disclosure uses Bayesian model averaging (BMA) to weightedly output the excellent climate models. Compared with the traditional model averaging method, BMA further reduces the prediction error and uncertainty, significantly improves the accuracy of future meteorological change trend prediction, and is especially suitable for water source area climate simulation and future scenario analysis, thereby providing more guiding meteorological data support for water resource management and regulation.

[0126] Corresponding to the above-mentioned CMIP6 climate model comprehensive comparison and selection method, the present application also provides a CMIP6 climate model comprehensive comparison and selection system. Since the system embodiments of the present application correspond to the above-mentioned method embodiments, for the details not disclosed in the system embodiments, reference can be made to the above-mentioned method embodiments, which will not be described in detail in the present application.

[0127] Figure 8 The structure diagram of a CMIP6 climate model comprehensive comparison and selection system provided by the embodiments of the present disclosure is shown in Figure 8 , which comprises:

[0128] The acquisition unit 21 is configured to acquire CMIP6 climate model data and meteorological measured data of a target research area, and the spatial range of the CMIP6 climate model data includes the target research area.

[0129] The first calculation unit 22 is configured to calculate first index data of each climate model in the CMIP6 climate model data based on the meteorological measured data, so as to obtain an index factor of each climate model when simulating at least two meteorological elements, respectively, and the types of the first index data are normalized standard deviation, spatial correlation coefficient, average deviation and interannual standard deviation.

[0130] The second calculation unit 23 is configured to calculate a first exponential score of the simulation capability of each climate model for one meteorological element and a second exponential score of the simulation capability of each climate model for another meteorological element based on the index factor according to a preset base function, and the at least two meteorological elements include the one meteorological element and the another meteorological element.

[0131] The third calculation unit 24 is configured to average the sum of the first exponential score and the second exponential score corresponding to each climate model, so as to obtain a comprehensive exponential score corresponding to each climate model.

[0132] The selection unit 25 is configured to select a preset number of climate models from the climate models in order from large to small according to the comprehensive exponential score.

[0133] The processing unit 26 is configured to perform a weighted ensemble average on the preset number of climate models based on a Bayesian model weighted average method to obtain a coupled climate model, and the coupled climate model is used to simulate a variation law of a future meteorological element.

[0134] The CMIP6 climate model comprehensive selection system provided by the present disclosure obtains CMIP6 climate model data and meteorological measured data of a target research region, the spatial range of the CMIP6 climate model data includes the target research region; calculates first index data of each climate model in the CMIP6 climate model data based on the meteorological measured data to obtain an index factor of each climate model when simulating at least two meteorological elements respectively, the types of the first index data are normalized standard deviation, spatial correlation coefficient, average deviation and interannual standard deviation; calculates a first exponential score of the simulation capability of each climate model on one meteorological element and a second exponential score of the simulation capability of each climate model on another meteorological element based on a preset base function according to the index factor, the at least two meteorological elements include the one meteorological element and the another meteorological element; averages the sum of the first exponential score and the second exponential score corresponding to each climate model to obtain a comprehensive exponential score corresponding to each climate model; selects a preset number of climate models from the climate models in a descending order of the comprehensive exponential scores; and performs a weighted ensemble average on the preset number of climate models based on a Bayesian model weighted average method to obtain a coupled climate model, and the coupled climate model is used to simulate a variation law of a future meteorological element. Compared with related technologies, the first index data of each climate model is calculated based on the meteorological measured data and the CMIP6 climate model data, the first exponential score of the simulation capability of each climate model on one meteorological element and the second exponential score of the simulation capability of each climate model on another meteorological element are calculated based on the first index data, the sum of the first exponential score and the second exponential score corresponding to each climate model is averaged to obtain a comprehensive exponential score corresponding to each climate model, and a preset number of climate models are selected from the climate models in a descending order of the comprehensive exponential scores. The selection process of the foregoing climate models comprehensively considers the simulation effects of the climate models when simulating different meteorological elements, so that the selected climate models can achieve good simulation effects when simulating different meteorological elements. Further, a coupled climate model is obtained by performing a weighted ensemble average on the selected climate models, and the simulation effect is improved when the coupled climate model is used to simulate different meteorological elements.

[0135] It should be noted that the foregoing description of the method embodiment is also applicable to the system of the present embodiment, and the principle is the same, and the present embodiment is not limited.

[0136] According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0137] The present disclosure provides an electronic device, comprising: at least one processor; and a memory connected in communication with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the CMIP6 climate model comprehensive comparison and selection method described in the above embodiments.

[0138] The present disclosure provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the CMIP6 climate model comprehensive comparison and selection method described in the above embodiments.

[0139] The present disclosure provides a computer program product comprising a computer program which, when executed by a processor, implements the CMIP6 climate model comprehensive comparison and selection method described in the above embodiments.

[0140] Those skilled in the art will appreciate that embodiments of the present disclosure can be provided as methods, systems, or computer program products. Therefore, the present disclosure can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present disclosure can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0141] The present disclosure is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one flow or multiple flows and / or blocks Figure 1 an apparatus that performs the functions specified in one or more blocks or flows.

[0142] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the flow Figure 1 The functions of a flow or multiple flows and / or a block or multiple blocks in conjunction with the disclosed methods can be implemented on a computer or processor. Figure 1

[0143] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flow Figure 1 The functions of a flow or multiple flows and / or a block or multiple blocks in conjunction with the disclosed methods can be implemented on a computer or processor. Figure 1

[0144] Finally, it should be noted that the above-mentioned embodiments are merely used to illustrate the technical solutions of the present application, rather than limiting the same. Even though the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced equivalently, and any modification or replacement without departing from the spirit and scope of the present application should be covered within the protection scope of the claims of the present application.​​

Claims

1. A CMIP6 climate model comprehensive comparison and selection method, characterized in that, The method comprises the following steps: obtaining CMIP6 climate model data and meteorological observation data of a target research region, wherein the spatial range of the CMIP6 climate model data comprises the target research region; calculating first index data of each climate model in the CMIP6 climate model data based on the meteorological observation data, to obtain index factors of each climate model in the each climate model when simulating at least two meteorological elements respectively, wherein the types of the first index data are normalized standard deviation, spatial correlation coefficient, average deviation and interannual standard deviation; calculating a first index score of the simulation capability of each climate model for one meteorological element and a second index score of the simulation capability of each climate model for another meteorological element based on a preset base function according to the index factors, wherein the at least two meteorological elements comprise the one meteorological element and the another meteorological element; averaging the sum of the first index score and the second index score corresponding to each climate model to obtain a comprehensive index score corresponding to each climate model; selecting a preset number of climate models from the each climate model in order from large to small according to the comprehensive index score; comprising: calculating the spatial correlation coefficient, the standard deviation and the root mean square error of the each climate model and the meteorological observation data; calculating a first performance score of the each climate model when simulating the one meteorological element according to the spatial correlation coefficient, the standard deviation and the root mean square error, and selecting a fifth target climate model with the best simulation performance for the one meteorological element from the each climate model based on the first performance score; calculating a second performance score of the each climate model when simulating the another meteorological element according to the spatial correlation coefficient, the standard deviation and the root mean square error, and selecting a sixth target climate model with the best simulation performance for the another meteorological element from the each climate model based on the second performance score; performing weighted ensemble averaging on the preset number of climate models based on a Bayesian model weighted average method to obtain a coupled climate model, wherein the coupled climate model is used to simulate the change rule of future meteorological elements.

2. The method of claim 1, wherein, The method for calculating first index data of each climate model in the CMIP6 climate model data based on the meteorological observation data to obtain index factors of each climate model in the each climate model when simulating at least two meteorological elements respectively comprises the following steps: calculating target normalized standard deviation, target spatial correlation coefficient, target average deviation and target interannual standard deviation of each climate model relative to the meteorological observation data in the each climate model in turn; performing factor conversion on the target normalized standard deviation based on a normalized standard deviation conversion function to obtain a first conversion factor; performing factor conversion on the target spatial correlation coefficient based on a spatial correlation coefficient conversion function to obtain a second conversion factor; performing factor conversion on the target average deviation based on an average deviation conversion function to obtain a third conversion factor; performing factor conversion on the target interannual standard deviation based on an interannual standard deviation conversion function to obtain a fourth conversion factor. The target interannual standard deviation is factor-transformed based on an interannual standard deviation transformation function to obtain a fourth transformation factor, and the index factor is obtained by weighting the first transformation factor, the second transformation factor, the third transformation factor and the fourth transformation factor, so as to obtain the index factor when each climate mode simulates at least two meteorological elements respectively. The formula of the normalization standard deviation transformation function is: X1=-|x1-1| wherein X1 is the first transformation factor, and x1 is the target normalization standard deviation. The formula of the spatial correlation coefficient transformation function is: X2=x2-1 wherein X2 is the second transformation factor, and x2 is the target spatial correlation coefficient. The formula of the average deviation transformation function is: X3=-|x3| wherein X3 is the third transformation factor, and x3 is the target average deviation. The formula of the interannual standard deviation transformation function is: X4=-x4 wherein X4 is the fourth transformation factor, and x4 is the target interannual standard deviation.

3. The method of claim 2, wherein, The formula of the preset base function is: wherein, represents a tolerance, = |maxE(X)|, E(X) is the set of the index factors corresponding to the climate model simulating the one meteorological element or the set of the index factors corresponding to the climate model simulating the other meteorological element, X is the index factor corresponding to the climate model simulating the one meteorological element or the index factor corresponding to the climate model simulating the other meteorological element; is a correction factor; is a base number, S(X) is the first index score or the second index score.

4. The method of claim 1, wherein, The method comprises: The spatial distribution diagrams corresponding to the at least two meteorological elements respectively under each climate mode are sequentially drawn according to the CMIP6 climate mode data and the meteorological measured data, to obtain a first spatial distribution diagram corresponding to one meteorological element under each climate mode and a second spatial distribution diagram corresponding to another meteorological element under each climate mode. The spatial distribution feature differences between the first spatial distribution diagrams under the respective climate modes are compared to obtain a first spatial distribution feature difference, and the spatial distribution feature differences between the second spatial distribution diagrams under the respective climate modes are compared to obtain a second spatial distribution feature difference. A first spatial correlation coefficient between first simulation data simulated by each climate mode for the one meteorological element and first target measured data corresponding to the first simulation data in the meteorological measured data is calculated respectively. A second spatial correlation coefficient between second simulation data simulated by each climate mode for the another meteorological element and second target measured data corresponding to the second simulation data in the meteorological measured data is calculated respectively, and the CMIP6 climate mode data comprises the first simulation data and the second simulation data. A first target climate mode having the best simulation effect on the spatial feature of the one meteorological element is selected from the respective climate modes according to the first spatial distribution feature difference and the first spatial correlation coefficient. A second target climate mode having the best simulation effect on the spatial feature of the another meteorological element is selected from the respective climate modes according to the second spatial distribution feature difference and the second spatial correlation coefficient.

5. The method of claim 4, wherein, The method comprises: selecting a third target climate model having the best simulation effect on the time characteristic of the one meteorological element from the climate models according to a first interannual standard deviation, the first interannual standard deviation being used to describe the difference in interannual fluctuation amplitude between the first simulation data and the first target measured data so as to determine the simulation effect of the climate models on the time characteristic of the one meteorological element; selecting a fourth target climate model having the best simulation effect on the time characteristic of the other meteorological element from the climate models according to a second interannual standard deviation, the second interannual standard deviation being used to describe the difference in interannual fluctuation amplitude between the second simulation data and the second target measured data so as to determine the simulation effect of the climate models on the time characteristic of the other meteorological element, the first interannual standard deviation and the second interannual standard deviation being calculated by an interannual standard deviation function; the interannual standard deviation function is expressed as: wherein, represents the interannual standard deviation of the respective climate pattern, represents the interannual standard deviation of the meteorological observation data, IVS represents the interannual standard deviation.

6. A CMIP6 climate model comprehensive comparison and selection system, which executes the method of any one of claims 1-5, characterized in that, comprising: an acquisition unit configured to acquire CMIP6 climate model data and meteorological measured data of a target research region, a spatial range of the CMIP6 climate model data including the target research region; a first calculation unit configured to calculate first index data of each climate model in the CMIP6 climate model data based on the meteorological measured data, to obtain an index factor of each climate model in the climate models when the each climate model respectively simulates at least two meteorological elements, the types of the first index data being normalized standard deviation, spatial correlation coefficient, average deviation and interannual standard deviation; a second calculation unit configured to calculate a first exponential score of simulation capability of each climate model on one meteorological element and a second exponential score of simulation capability of each climate model on another meteorological element based on a preset base function according to the index factor, the at least two meteorological elements including the one meteorological element and the another meteorological element; a third calculation unit configured to average the sum of the first exponential score and the second exponential score corresponding to each climate model, to obtain a comprehensive exponential score corresponding to each climate model; a selection unit configured to select a preset number of climate models from the climate models in a descending order of the comprehensive exponential score; comprising: calculating the spatial correlation coefficient, the standard deviation and the root mean square error of each climate model and the meteorological measured data; calculating a first performance score of each climate model when simulating the one meteorological element according to the spatial correlation coefficient, the standard deviation and the root mean square error, and selecting a fifth target climate model having the best simulation performance on the one meteorological element from the climate models based on the first performance score; calculating a second performance score of each climate model when simulating the other meteorological element according to the spatial correlation coefficient, the standard deviation and the root mean square error, and selecting a sixth target climate model having the best simulation performance on the other meteorological element from the climate models based on the second performance score; The processing unit is configured to perform a weighted ensemble average on the preset number of climate models based on a Bayesian model weighted average method to obtain a coupled climate model, and the coupled climate model is used to simulate a variation law of a future meteorological element.

7. An electronic device, comprising: The computer program product comprises: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-5.

8. A non-transitory computer-readable storage medium having stored thereon computer instructions, wherein, The computer instructions are used to enable the computer to perform the method of any one of claims 1-5.

9. A computer program product, characterised in that, The computer program, when executed by a processor, implements the method of any one of claims 1-5.

Images