Climate mode weighted set system and method based on three-factor method

Abstract

The invention discloses a climate mode weighted set system and method based on a three-factor method, and belongs to the technical field of climate change information analysis. The system comprises a data preprocessing module, a weight calculation module and a weighted set module, wherein the data preprocessing module integrates meteorological station observation data and climate mode simulation data, and provides a basis for subsequent steps. The weight calculation module evaluates each mode from the three dimensions of spatial correlation, station error and time variation rate by comparing the current observation value with the mode simulation value, calculates and homogenizes the weight, and sums the weight to obtain the total weight in each mode. And the weighted set module performs weighted processing on the meteorological elements according to the total weight of each mode to obtain a final result. According to the method, the simulation results of a plurality of climate modes can be integrated and optimized by integrating three factors, and more accurate and more reliable climate change prediction is provided.

Description

Technical Field

[0001] The present invention relates to the technical field of climate change information analysis, and more particularly to a climate model weighted set system and method based on a three-factor method. Background Art

[0002] Climate change has a profound impact on human society and the natural environment. Reliable climate prediction information is needed to formulate policies and adaptation measures to address climate change. Climate models can provide a wider range of climate predictions and provide a scientific basis for policy making.

[0003] As the study of climate change deepens, scientists are increasingly aware of the limitations of a single climate model. Thanks to the promotion of the international Coupled Model Intercomparison Project (CMIP), through international cooperation, climate models from different countries and regions can be compared and integrated under a unified framework, promoting data sharing and model improvement. However, different climate models still have different biases and uncertainties when simulating the Earth's climate system.

[0004] By weighting the results of climate model simulations, we can combine the results of multiple models in the hope of obtaining more accurate and reliable climate predictions. At the same time, due to the differences in emissions and technology development scenarios under different socioeconomic pathways, the errors in climate models, and the uncertainties caused by internal variability, the dispersion between models is still large. How to reduce the uncertainty of multi-model projections has become a hot topic in the field of climate change research. By considering multiple factors and reasonably allocating weights to different models, the uncertainty of climate projections can be reduced.

[0005] Therefore, how to provide a climate model weighted ensemble system and method based on the three-factor method is a problem that technical personnel in this field urgently need to solve. Summary of the invention

[0006] In view of this, the present invention provides a climate model weighted ensemble system and method based on the three-factor method. It sets reasonable weights for each model from three aspects, namely spatial correlation, error amount of each station and temporal variability, and constructs a climate model weighted ensemble system based on the three factors. This can improve the accuracy of climate forecasts, reduce estimation uncertainty, enhance the simulation capability of multi-scale interactions of the climate system, and respond to the needs of climate change policies.

[0007] In order to achieve the above object, the present invention provides the following technical solutions: A climate model weighted ensemble system based on a three-factor approach, including: Data preprocessing module, weight calculation module and weighted collection module; The data preprocessing module processes the contemporary observation data and climate scenario data of the meteorological station to obtain the observation and model results required for subsequent calculations, and the observation and model results include: contemporary observation values, contemporary climate simulation values, and future climate simulation values; The weight calculation module uses the contemporary observation values ​​and contemporary climate simulation values ​​obtained by the data preprocessing module to compare the climate states, calculates the weight values ​​of the three factors in each mode from the three factors of spatial correlation, station error and time variability, and homogenizes the weight values ​​of the three factors respectively, adds the homogenized weight values ​​of the three factors in each mode, and obtains the total weight in each mode respectively; The weighted set module calculates a weighted set for each meteorological element based on the contemporary climate simulation value and future climate simulation value obtained by the data preprocessing module and the total weight under each mode obtained by the weight calculation module, and outputs the result.

[0008] Furthermore, the data preprocessing module includes: Based on contemporary observation data from meteorological stations, extract observation data site information and quality control information; Count the number of effective value years and the number of effective value stations in each year for each meteorological element, and select the inspection period and inspection station for each meteorological element based on the preset standards; Based on the contemporary climate simulation values ​​and future climate simulation values ​​of the climate model, multiple sets of global meteorological element data under different scenarios of the climate model are input, the target area data are intercepted and format converted, the binary file is converted into an ASCII text file and output, and the observation and model results are obtained. The observation and model results include: contemporary observation values, contemporary climate simulation values, and future climate simulation values.

[0009] Furthermore, the weight calculation module includes: The spatial correlation coefficient of the multi-year average of contemporary observations and contemporary climate simulations is used as an indicator, and the result is proportional to the weight to obtain the first weight value of each model; The root mean square error between the multi-year average of contemporary observations and contemporary climate simulations is used as an indicator, and the result is inversely proportional to the weight, to obtain the second weight value of each model; The distance between the multi-year variance ratio of contemporary observations and contemporary climate simulations and 1 (i.e. the absolute value of the variance ratio of contemporary model values ​​to contemporary observations minus 1) is used as an indicator. The result is inversely proportional to the weight, and the third weight value of each model is obtained. Since the order of magnitude ranges of the above three weights are quite different, in order to ensure that each weight value has a relatively uniform importance in the mode collection process, it is necessary to use a homogenization method to convert each weight value into a decimal between (0, 1), thereby eliminating the value range differences between the three weight values ​​and making different indicators comparable. Therefore, the first weight value, the second weight value, and the third weight value under each mode are homogenized respectively to obtain the homogenized first weight value, the homogenized second weight value, and the homogenized third weight value of each mode. The homogenized first weight value, the homogenized second weight value, and the homogenized third weight value of each mode are added together to obtain the total weight under each mode.

[0010] A climate model weighted ensemble method based on a three-factor method comprises the following steps: Processing contemporary observation data and climate scenario data of meteorological stations to obtain observation and model results required for subsequent calculations, wherein the observation and model results include: contemporary observation values, contemporary climate simulation values, and future climate simulation values; The obtained contemporary observations and contemporary climate simulation values ​​are used to compare the climate states, and the weights of the three factors in each mode are calculated from the three factors of spatial correlation, station error and time variability. The weights of the three factors are homogenized respectively, and the homogenized weights of the three factors in each mode are added together to obtain the total weights in each mode. Based on the obtained contemporary climate simulation values ​​and future climate simulation values ​​and the total weights obtained under each model, a weighted set is calculated for each meteorological element and the results are output.

[0011] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a climate model weighted set system and method based on the three-factor method, which integrates and optimizes the simulation results of multiple climate models, provides more accurate and reliable climate change predictions, and supports scientific research, policy formulation, and the implementation of climate adaptation strategies. The specific beneficial effects are as follows: Improve the accuracy of climate prediction: By integrating and optimizing the simulation results of multiple climate models and comprehensively considering the advantages and characteristics of different models, more accurate climate change predictions can be provided.

[0012] Reducing forecast uncertainty: Taking into account spatial correlation, station error, and temporal variability, the weights of different models are reasonably allocated, which effectively reduces the uncertainty of multi-model forecasts and improves the reliability of climate forecasts.

[0013] Improving simulation capabilities: The simulation capabilities of the multi-scale interactions of the climate system have been enhanced, providing stronger support for scientific research and policy making.

[0014] Responding to policy needs: It meets the needs of climate change policy-making for reliable climate prediction information, and helps to formulate more effective policies and adaptation measures to address climate change. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0016] Figure 1 It is a schematic diagram of the system structure of the present invention; Figure 2 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] See also Figure 1 The embodiment of the present invention discloses a climate model weighted ensemble system based on a three-factor method, which is applicable to multiple sets of global climate model simulation results with corrected and unified horizontal resolutions, and the system comprises: Data preprocessing module, weight calculation module and weighted collection module; The data preprocessing module processes the contemporary observation data and climate scenario data of the meteorological station to obtain the observation and model results required for subsequent calculations, and the observation and model results include: contemporary observation values, contemporary climate simulation values, and future climate simulation values; The weight calculation module uses the contemporary observation values ​​and contemporary climate simulation values ​​obtained by the data preprocessing module to compare the climate states, calculates the weight values ​​of the three factors in each mode from the three factors of spatial correlation, station error and time variability, and homogenizes the weight values ​​of the three factors respectively, adds the homogenized weight values ​​of the three factors in each mode, and obtains the total weight in each mode respectively; The weighted set module calculates a weighted set for each meteorological element based on the contemporary climate simulation value and future climate simulation value obtained by the data preprocessing module and the total weight under each mode obtained by the weight calculation module, and outputs the result.

[0019] Furthermore, the data preprocessing module includes: Based on contemporary observation data from meteorological stations, extract observation data site information and quality control information; Count the number of effective value years and the number of effective value stations in each year for each meteorological element, and select the inspection period and inspection station for each meteorological element based on the preset standards; Based on the contemporary climate simulation values ​​and future climate simulation values ​​of the climate model, multiple sets of global meteorological element data under different scenarios of the climate model are input, the target area data are intercepted and format converted, the binary file is converted into an ASCII text file and output, and the observation and model results are obtained. The observation and model results include: contemporary observation values, contemporary climate simulation values, and future climate simulation values.

[0020] Furthermore, the weight calculation module includes: The spatial correlation coefficient of the multi-year average of contemporary observations and contemporary climate simulations is used as an indicator, and the result is proportional to the weight to obtain the first weight value of each model; The root mean square error between the multi-year average of contemporary observations and contemporary climate simulations is used as an indicator, and the result is inversely proportional to the weight, to obtain the second weight value of each model; The distance between the multi-year variance ratio of contemporary observations and contemporary climate simulations and 1 (i.e. the absolute value of the variance ratio of contemporary model values ​​to contemporary observations minus 1) is used as an indicator. The result is inversely proportional to the weight, and the third weight value of each model is obtained. The first weight value, the second weight value, and the third weight value under each mode are respectively normalized to obtain the normalized first weight value, the normalized second weight value, and the normalized third weight value of each mode. The normalized first weight value, the normalized second weight value, and the normalized third weight value of each mode are added together to obtain the total weight under each mode.

[0021] Specifically, the total weight refers to the total weight of each mode. The total weight of each mode is used in the weighted collection module, that is, each mode has a total weight. The "total" weight is relative to the weight of the "three factors", which is the sum of the weights of the three factors, not the sum of the weights of all modes.

[0022] More specifically, since the magnitude ranges of the three weights mentioned above vary greatly, in order to ensure that each weight value has a relatively uniform importance in the pattern collection process, it is necessary to use a normalization method to convert each weight value into a decimal between (0, 1), thereby eliminating the range differences between the three weight values ​​and making different indicators comparable. For each pattern, add its normalized first, second, and third weight values ​​to obtain the total weight of the pattern.

[0023] Module 1: Data preprocessing module: This module is responsible for processing meteorological station observation data and climate model simulation results. Specifically, it includes: Extract the site information and observation quality control information of meteorological site observation data.

[0024] Statistics are made on the number of effective value years for each station and the number of effective value stations in each year for each meteorological element.

[0025] The longest consecutive years in which the number of effective value stations in each year reaches 90% or more of the total number of sample stations of the factor are selected as the "inspection period".

[0026] According to the standard of "the number of effective years of each station reaches 90% of the total number of years in the investigation period", the "inspection stations" of each meteorological element are selected to form a text file.

[0027] Input multiple global meteorological element values ​​under various scenarios from multiple global climate models (GCMs), extract the Chinese land area values ​​from the global data and convert the format.

[0028] Convert binary files to ASCII files and output them as txt documents.

[0029] Module 2, weight calculation module: This module uses the contemporary multi-year simulation values ​​extracted and converted by module 1 and the contemporary observation values ​​to compare the climate state, considering the three aspects of spatial correlation, station error, and time variability. Each aspect accounts for 1 / 3 of the weight, and each model is assigned a different weight through calculation. The specific calculation method is as follows: Factor 1: The spatial correlation coefficient between the multi-year average values ​​of contemporary observations and contemporary simulations is used as an indicator, and the result is proportional to the weight.

[0030] Factor 2: The root mean square error between the multi-year average of contemporary observations and contemporary simulations is used as an indicator, and the result is inversely proportional to the weight.

[0031] Factor three: The distance between the variance ratio of contemporary observations and contemporary simulations over the years and 1 (i.e. the absolute value of the variance ratio of simulations and observations minus 1) is used as an indicator, and the result is inversely proportional to the weight.

[0032] The homogenization method is used to obtain the homogenized weight values ​​of the three factors, and the total weight of the model is obtained by adding the three together.

[0033] Module 3, weighted collection module: This module inputs multiple sets of climate scenario data for China's land areas obtained in module 1 and the weights of each meteorological element of each model obtained in module 2. For each meteorological element, the output results of each model are weighted and aggregated according to different weight coefficients to obtain a set of aggregate data and provide the output function of the aggregate results.

[0034] In a specific embodiment, the system of the present invention can be implemented based on the following software and hardware environment: Development software: Compaq Visual Fortran Version 6.6, GrADS, Microsoft Excel; Software environment: Microsoft Windows 95 or above; Hardware environment: desktop computer or laptop computer and above, 16G memory or above, 5T hard disk space.

[0035] On the other hand, see Figure 2 The embodiment of the present invention discloses a climate model weighted set method based on the three-factor method, which runs the relevant programs in the data preprocessing module to complete data extraction, statistics, selection and format conversion.

[0036] Run the program in the weight calculation module to calculate the weight of each mode based on the preprocessed data.

[0037] Run the program in the weighted aggregation module, perform weighted aggregation on the output results of each mode according to the calculated weights, and obtain the final aggregate data.

[0038] A climate model weighted ensemble method based on a three-factor method specifically includes the following steps: Processing contemporary observation data and climate scenario data of meteorological stations to obtain observation and model results required for subsequent calculations, wherein the observation and model results include: contemporary observation values, contemporary climate simulation values, and future climate simulation values; The obtained contemporary observations and contemporary climate simulation values ​​are used to compare the climate states, and the weights of the three factors in each mode are calculated from the three factors of spatial correlation, station error and time variability. The weights of the three factors are homogenized respectively, and the homogenized weights of the three factors in each mode are added together to obtain the total weights in each mode. Based on the obtained contemporary climate simulation values ​​and future climate simulation values ​​and the total weights obtained under each model, a weighted set is calculated for each meteorological element and the results are output.

[0039] Specifically, based on contemporary observation data from meteorological stations, the observation data site information and quality control information are extracted; Count the number of effective value years for each station and the number of effective value stations in each year for each element; Filter out the inspection time periods and inspection sites for each meteorological element based on preset standards; Input contemporary and future global meteorological element data under different scenarios of multiple climate models, intercept the data of the target area and convert the format to obtain the model results required for subsequent steps; Based on the contemporary observation values ​​and contemporary model values ​​extracted and converted in the previous steps, the total weight of each model of each factor is calculated; Based on the model results and the total weight of each model, a weighted set is calculated for each meteorological element.

[0040] Specifically, the following is an example of the ensemble calculation of the simulated values ​​of eight meteorological elements under four greenhouse gas emission scenarios using five global climate models to list the specific implementation steps: Extraction of observation data site information and quality control information; Count the number of effective value years for each station and the number of effective value stations in each year for each element; Select the "inspection period" and "inspection site" for each element; Regional and format conversion of global climate model simulation results; Calculate the total weight of each mode of each factor; Perform a weighted aggregation of pattern results.

[0041] Specifically, step 1. Extraction of observation data site information and quality control information: Eight common meteorological elements closely related to climate change impact assessment were selected: average temperature, maximum temperature, minimum temperature, precipitation, total solar radiation, average relative humidity, surface air pressure, and near-surface average wind speed. Compaq VisualFortran Version 6.6 was used to extract the station information and quality control information of the observation data of meteorological stations in the land area of ​​China. The station information and quality control information of the observation data of the daily value dataset of China's surface climate data V3.0 (SURF_CLI_CHN_MUL_DAY_V3.0) from 1961 to 2012 and the observation data of the international exchange station of the Meteorological Administration of Radiation Data (RADI_MUL_CHN_DAY_CES) were read. If the number of days with a data quality control code of 0 for a certain element at a certain station in a certain year ≥ the threshold standard for the number of days (number of days per year), the station information and quality control information of the observation data were read. ), then the data of the element in that year at that station is a valid value, and the station is assigned a value of 1 for that year, otherwise it is assigned a value of 0, forming a two-dimensional matrix output to the document, with the horizontal axis being "year" and the vertical axis being "station".

[0042] Step 2. Count the number of effective value years for each station and the number of effective value stations in each year for each element: According to the output results of step 1, Compaq Visual Fortran Version 6.6 and Excel tables were used to count the number of effective value years and the number of effective value stations in each year for the eight meteorological elements.

[0043] Step 3. Select the “inspection period” and “inspection site” for each element: From the results of step 2, select the longest consecutive years in which the number of effective value stations in each year reaches 90% or more of the total number of sample stations of the element as the "inspection period", and select the "inspection sites" of each element according to the standard of "the number of effective value years of each station reaches 90% of the total number of years in the inspection period" to form a text file.

[0044] Step 4. Convert the global climate model simulation results to different regions and formats: Input the simulation results of 5 global climate models (GCM) from 1951 to 2050 under 4 RCP scenarios. The simulation results are obtained from ISI-MIP (The Inter-Sectoral Impact Model Intercomparison Project). These 5 GCMs are all from CMIP5 (Coupled Model IntercomparisonProject Phase 5), and the model names are GFDL-ESM2M, HadGEM2-ES, IPSL-CM5A-LR, MIROC-ESM-CHEM, and NorESM1-M. The horizontal resolutions of their original output results are different. ISI-MIP selects climate variables commonly used in climate change impact assessment work, such as temperature, precipitation, wind speed, radiation, relative humidity, and air pressure, and uses the bilinear interpolation method to interpolate them to a unified horizontal resolution of 0.5°. 0.5°, and the interpolation results were corrected one by one using the statistical deviation correction method based on probability distribution.

[0045] The RCP scenario is a greenhouse gas stabilization scenario proposed by the IPCC (Intergovernmental Panel on Climate Change) in 2007. The emission scenario is represented by representative concentration pathways (RCPs). In this example, the climate simulation results under four scenarios, namely RCP2.6 / 4.5 / 6.0 / 8.5, are calculated.

[0046] Compaq Visual Fortran Version 6.6 was used to read the global eight element values ​​of five GCMs under four RCP scenarios. The Chinese land area was intercepted from the global data and converted from nc format into ASCII code files and output as txt documents.

[0047] Step 5. Calculate the weight of each mode for each factor: Compaq Visual Fortran Version 6.6 was used to read the meteorological station observation data (SURF_CLI_CHN_MUL_DAY_V3.0, RADI_MUL_CHN_DAY_CES), the station information extracted in step 1, the survey station documents selected in step 3, and the model results output in step 4. The model grid whose difference between the center point of the model grid and the longitude and latitude of the survey station is within 0.25° is the model grid point that matches the survey station. The observation values ​​and contemporary simulation values ​​of the overlapping period of the "survey period" selected in step 3 (1961-2000 in this example) were used to compare the climate state. From the three aspects of spatial correlation, station error, and time variability, each aspect accounts for 1 / 3 of the weight, and different weights are assigned to the five models: Factor 1 uses the spatial correlation coefficient of the 40-year average value of the model value and the observed value as an indicator, and its result is proportional to the weight; Factor 2 uses the root mean square error of the 40-year average value of the model value and the observed value as an indicator, and its result is inversely proportional to the weight; Factor 3 uses the distance between the 40-year variance ratio of the model value and the observed value and 1 (that is, the absolute value of the variance ratio minus 1) as an indicator, and its result is inversely proportional to the weight. The homogenized weight values ​​of the three factors are obtained by processing with the homogenization method, and the total weight of the model is obtained by adding the three. In other words, compared with the observed values, the better the spatial correlation of the simulated values, the smaller the errors of each station, and the closer the time variability, the higher the weight of the model.

[0048] For each meteorological element, a set of weight coefficients corresponding to the kth mode is obtained by the following calculation method: Factor 1: Calculate the 40-year average values ​​of the observed values ​​and the simulated values ​​respectively, and use these two sets of sequences (the number of samples in each set of sequences is the number of sites nn) to calculate their spatial correlation coefficient COR. The formula is: COR=A / (B C) A=Σ[(VVo(i)-Avo) (VVm(i)-AVm)],i=1,nn B=SQRT(So) C=SQRT(Sm) So = Σ[VVo(i)-AVo] 2 Sm=Σ[VVm(i)-AVm] 2 AVo=ΣVVo(i),i=1,nn AVm=ΣVVm(i),i=1,nn VVo(i)= ΣVo(i,j),i=1,nn; j=1,nt VVm(i)= ΣVm(i,j),i=1,nn; j=1,nt Where Vo(i,j) is the observed value at the i-th station at the j-th time point, Vm(i,j) is the simulated value at the i-th station at the j-th time point, and nt is the number of days in 40 years.

[0049] The first weight value WT1(k)=COR.

[0050] Factor 2: Using the 40-year average of the observed and simulated values, the root mean square error RMSE of the two sets of series is calculated using the formula: RMSE = SQRT (RR / nn) RR=Σ[(VVo(i)-VVm(i)) 2],i=1,nn The meaning and calculation formula of VVo(i) and VVm(i) are the same as those described in factor one.

[0051] The second weight value WT2(k)=1.0 / RMSE.

[0052] Factor 3: Calculate the daily spatial average of the observed and simulated values, and calculate the variances of the two series, VRo (observed values) and VRm (simulated values), respectively. The formula is: VRo=Σ[(VSo(j)-ASo) 2] / nt,j=1,nt VRm=Σ[(VSm(j)-ASm) 2] / nt,j=1,nt ASo=ΣVSo(j),j=1,nt ASm=ΣVSm(j),j=1,nt VSo(j)= ΣVo(i,j),j=1,nt; i=1,nn VSm(j)= ΣVm(i,j),j=1,nt; i=1,nn Where Vo(i,j) is the observed value at the i-th station at the j-th time point, Vm(i,j) is the simulated value at the i-th station at the j-th time point, and nt is the number of days in 40 years.

[0053] The third weight value WT3(k)=VRo / abs(VRm-VRo).

[0054] Calculate the corresponding values ​​of 5 sets of patterns for the above 3 weights respectively, then each weight value has 5 samples, use these 5 samples to perform homogenization calculation, and obtain the homogenized weight of each factor. The homogenized weights of factors one, two, and three are FWT1, FWT2, and FWT3 respectively, and the calculation formula is: FWT1(k)=WT1(k) / SWT1(k) SWT1(k)=ΣWT1(k),k=1,nm FWT2(k)=WT2(k) / SWT2(k) SWT2(k)=ΣWT2(k),k=1,nm FWT3(k)=WT3(k) / SWT3(k) SWT3(k)=ΣWT3(k),k=1,nm Among them, k is the mode number and nm=5 is the number of modes.

[0055] Add the averaged weights of factors one, two, and three to get the total weight TWT(k) of the kth mode: TWT(k)=(FWT1(k)+FWT2(k)+FWT3(k)) / nm, k=1,nm Among them, k is the mode number, and nm=5 is the total number of modes.

[0056] Step 6. Weighted aggregation of pattern results: Compaq Visual Fortran Version 6.6 was used to read the model results output in step 4 and the total weights of each model obtained in step 5, and the weighted set was calculated for each meteorological element at each grid point. The formula is: Vout=ΣVV(k) TWT(k), k=1,nm; Where Vout is the aggregate result of the meteorological element, VV(k) is the value of the meteorological element in the kth mode, TWT(k) is the total weight of the meteorological element in the kth mode, and nm=5 is the total number of modes.

[0057] For each meteorological element, if all five mode values ​​of the grid at that time point are missing values, the weighted set result is set as the missing value.

[0058] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0059] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A climate model weighted ensemble system based on a three-factor method, characterized in that: include: Data preprocessing module, weight calculation module and weighted collection module; The data preprocessing module processes the contemporary observation data and climate scenario data of the meteorological station to obtain the observation and model results required for subsequent calculations, and the observation and model results include: contemporary observation values, contemporary climate simulation values, and future climate simulation values; The weight calculation module uses the contemporary observation values ​​and contemporary climate simulation values ​​obtained by the data preprocessing module to compare the climate states, calculates the weight values ​​of the three factors in each mode from the three factors of spatial correlation, station error and time variability, and homogenizes the weight values ​​of the three factors respectively, adds the homogenized weight values ​​of the three factors in each mode, and obtains the total weight in each mode respectively; The weighted set module calculates a weighted set for each meteorological element based on the contemporary climate simulation value and future climate simulation value obtained by the data preprocessing module and the total weight under each mode obtained by the weight calculation module, and outputs the result.

2. A climate model weighted ensemble system based on the three-factor method according to claim 1, characterized in that: The data preprocessing module comprises: Based on contemporary observation data from meteorological stations, extract observation data site information and quality control information; Count the number of effective value years and the number of effective value stations in each year for each meteorological element, and select the inspection period and inspection station for each meteorological element based on the preset standards; Based on the contemporary climate simulation values ​​and future climate simulation values ​​of the climate model, multiple sets of global meteorological element data under different scenarios of the climate model are input, the target area data are intercepted and format converted, the binary file is converted into an ASCII text file and output, and the observation and model results are obtained. The observation and model results include: contemporary observation values, contemporary climate simulation values, and future climate simulation values.

3. The climate model weighted ensemble system based on the three-factor method according to claim 1, characterized in that: The weight calculation module includes: The spatial correlation coefficient of the multi-year average of contemporary observations and contemporary climate simulations is used as an indicator, and the result is proportional to the weight to obtain the first weight value of each model; The root mean square error between the multi-year average of contemporary observations and contemporary climate simulations is used as an indicator, and the result is inversely proportional to the weight, to obtain the second weight value of each model; The distance between the multi-year variance ratio of contemporary observations and contemporary climate simulations and 1 is used as an indicator, and the result is inversely proportional to the weight, to obtain the third weight value of each model; The first weight value, the second weight value, and the third weight value under each mode are respectively normalized to obtain the normalized first weight value, the normalized second weight value, and the normalized third weight value of each mode. The normalized first weight value, the normalized second weight value, and the normalized third weight value of each mode are added together to obtain the total weight under each mode.

4. A climate model weighted ensemble method based on a three-factor method, characterized in that: The following steps are involved: Processing contemporary observation data and climate scenario data of meteorological stations to obtain observation and model results required for subsequent calculations, wherein the observation and model results include: contemporary observation values, contemporary climate simulation values, and future climate simulation values; The obtained contemporary observations and contemporary climate simulation values ​​are used to compare the climate states, and the weights of the three factors in each mode are calculated from the three factors of spatial correlation, station error and time variability. The weights of the three factors are homogenized respectively, and the homogenized weights of the three factors in each mode are added together to obtain the total weights in each mode. Based on the obtained contemporary climate simulation values ​​and future climate simulation values ​​and the total weights obtained under each model, a weighted set is calculated for each meteorological element and the results are output.

Images