Medium and long term climate resource prediction method and system

Abstract

The invention discloses a medium and long term climate resource prediction method and system, and belongs to the technical field of resource prediction. The method comprises the following steps: establishing a multi-modal data set based on multi-source meteorological data, and preprocessing the multi-modal data set to generate a data time sequence; using an original numerical climate prediction model and a hybrid model to predict future medium and long term climate resources according to the data time sequence to obtain an original climate result predicted based on the original numerical climate prediction model and a deep learning predicted climate result predicted based on the hybrid model; fusing the original climate result and the deep learning predicted climate result, and correcting a fusion result to obtain a corrected predicted climate result; and carrying out probability prediction on the corrected predicted climate result through uncertainty quantification processing, and predicting future medium and long term climate resources. According to the method, future medium and long-term climate resources can be accurately predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of resource prediction, and more specifically, to a medium- and long-term climate resource prediction method and system. Background Art

[0002] Classic climate resource prediction techniques are mainly based on physical models and statistical methods. These techniques have been developed for decades and are widely used in medium- and long-term predictions of renewable energy resources such as wind speed, surface irradiance, and precipitation. Some of the main classic techniques are outlined below:

[0003] (1) Numerical Weather Prediction (NWP) models;

[0004] NWP models can be mainly divided into global climate models (GCM) and regional climate models (RCM).

[0005] GCM: GCM is one of the most widely used NWP models, designed to simulate the atmosphere and ocean system of the entire Earth. It takes into account a variety of factors such as atmospheric circulation, ocean circulation, land surface processes and cryosphere. GCM is often used for long-term climate prediction and climate change research.

[0006] RCM: Unlike GCM, RCM only simulates the climate of a specific region and uses data from GCM as boundary conditions. The advantage of RCM is that it can provide higher-resolution local climate predictions and is suitable for climate resource predictions in smaller areas.

[0007] (2) Statistical downscaling methods;

[0008] Statistical downscaling is a method of predicting local climate conditions based on the statistical relationship between historical observations and large-scale climate model outputs. It is often used to convert the output of GCMs or RCMs into smaller-scale local predictions.

[0009] Regression models: Use multiple linear regression or nonlinear regression models to predict local climate conditions based on the relationship between historical climate data and the output of large-scale climate models.

[0010] Principal Component Analysis (PCA): PCA is used to reduce dimensionality and identify the main patterns in the data. It can help extract key features from large-scale climate data and then perform downscaling predictions.

[0011] Random forests and decision trees: These machine learning methods are also used in statistical downscaling to handle large-scale data and identify complex nonlinear relationships.

[0012] (3) Time series analysis;

[0013] Time series analysis is used to forecast data that has time dependence, such as climate data. It uses past observations to predict future climate conditions.

[0014] Autoregressive Moving Average Model (ARMA): The ARMA model combines the autoregressive (AR) and moving average (MA) methods and is suitable for short-term climate prediction.

[0015] Autoregressive Integrated Moving Average Model (ARIMA): ARIMA is an extension of ARMA that can handle non-stationary time series data and is suitable for processing climate data with a long time span.

[0016] Seasonal Autoregressive Integrated Moving Average (SARIMA) model: The SARIMA model takes into account seasonal patterns in time series and is able to more accurately forecast climate resources with periodic fluctuations, such as monsoon rainfall or seasonal wind speed.

[0017] (4) Climate index method;

[0018] Climate indices such as the El Nino Southern Oscillation (ENSO) index and the North Atlantic Oscillation (NAO) index are often used for medium- and long-term climate forecasts. These indices are based on the main change patterns of the global or regional climate system and have a significant impact on climate resources such as precipitation, temperature, wind speed, etc.

[0019] ENSO: By analyzing the ENSO index, it is possible to predict the impact of El Niño or La Niña events on climate resources, such as affecting global rainfall patterns and wind speed distribution.

[0020] NAO: The NAO index is closely related to winter wind and precipitation patterns over the North Atlantic and Europe and is often used to predict seasonal climate in these regions.

[0021] (5) Empirical statistical methods;

[0022] Empirical statistical methods make predictions based on the statistical characteristics of historical climate data. These methods do not rely on physical models, but make predictions by analyzing the patterns of historical data.

[0023] Kriging interpolation: A geostatistical method used for interpolation prediction of spatial data, often used for spatial distribution prediction of meteorological data, such as the spatial distribution of wind speed and irradiance.

[0024] Binned regression: Climate data is binned according to specific classification standards, and regression analysis is performed within each bin to predict the distribution of climate resources.

[0025] In summary, the classic climate resource prediction technology is based on physical modeling and statistical analysis, using large-scale observation data and climate model simulation to predict climate resources. Although these methods are relatively mature, their prediction accuracy still cannot meet the prediction accuracy requirements of the rapid development of renewable energy in my country at this stage. Summary of the invention

[0026] In view of the above problems, the present invention proposes a medium- and long-term climate resource prediction method, comprising:

[0027] Acquire meteorological data from multiple sources, establish a multimodal data set based on the meteorological data from multiple sources, and preprocess the multimodal data set to generate a data time series;

[0028] Using the original numerical climate prediction model and the hybrid model, the mid- to long-term climate resources in the future are predicted respectively according to the data time series, and the original climate results predicted by the original numerical climate prediction model and the deep learning predicted climate results predicted by the hybrid model are obtained;

[0029] Fusing the original climate result and the deep learning predicted climate result, and correcting the fused result to obtain a corrected predicted climate result;

[0030] The corrected predicted climate results are subjected to probabilistic prediction through uncertainty quantification processing to predict the future medium- and long-term climate resources.

[0031] Optionally, the meteorological data includes at least one of the following: satellite observation data, weather station data, ocean observation data and geographic information data.

[0032] Optionally, a linear interpolation method is used to adjust the spatiotemporal resolution of meteorological data from multiple sources to be consistent, so as to establish a multimodal dataset.

[0033] Optionally, preprocessing the multimodal dataset includes:

[0034] The multimodal data set is sequentially subjected to data cleaning, interpolation and denoising processes.

[0035] Optionally, the hybrid model is a combination model based on CNN and RNN.

[0036] Optionally, an ensemble learning algorithm is used to fuse the climate results predicted by traditional predictions and the climate results predicted by deep learning, and the fusion results are corrected based on the Bayesian update technology.

[0037] Optionally, the uncertainty of the corrected predicted climate results is quantified based on the Monte Carlo simulation method, and a prediction confidence interval is generated according to the prediction results, and the reliability of future medium- and long-term climate resources is determined by the prediction confidence curve.

[0038] On the other hand, the present invention also proposes a medium- and long-term climate resource prediction system, comprising:

[0039] A data acquisition unit, used to acquire meteorological data from multiple sources, establish a multimodal data set based on the meteorological data from multiple sources, and preprocess the multimodal data set to generate a data time series;

[0040] A first prediction unit is used to use the original numerical climate prediction model and the hybrid model to predict the future medium- and long-term climate resources according to the data time series, and obtain the original climate results predicted by the original numerical climate prediction model and the deep learning predicted climate results predicted by the hybrid model;

[0041] A fusion unit, used to fuse the original climate result and the deep learning predicted climate result, and correct the fusion result to obtain a corrected predicted climate result;

[0042] The second prediction unit is used to perform probability prediction on the corrected predicted climate result through uncertainty quantification processing to predict the future medium- and long-term climate resources.

[0043] Optionally, the meteorological data includes at least one of the following: satellite observation data, weather station data, ocean observation data and geographic information data.

[0044] Optionally, the data acquisition unit uses a linear interpolation method to adjust the temporal and spatial resolutions of meteorological data from multiple sources to be consistent, so as to establish a multimodal data set.

[0045] Optionally, the data acquisition unit preprocesses the multimodal data set, including:

[0046] The multimodal data set is sequentially subjected to data cleaning, interpolation and denoising processes.

[0047] Optionally, the hybrid model is a combination model based on CNN and RNN.

[0048] Optionally, the fusion unit uses an integrated learning algorithm to fuse the climate results predicted by traditional predictions and the climate results predicted by deep learning, and corrects the fusion results based on the Bayesian update technology.

[0049] Optionally, the second prediction unit quantifies the uncertainty of the corrected predicted climate results based on the Monte Carlo simulation method, and generates a prediction confidence interval according to the prediction results, and determines the reliability of future medium- and long-term climate resources with the prediction confidence curve.

[0050] In yet another aspect, the present invention further provides a computing device, comprising: one or more processors;

[0051] a processor for executing one or more programs;

[0052] When the one or more programs are executed by the one or more processors, the above-described method is implemented.

[0053] In yet another aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed, the method described above is implemented.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] The present invention provides a medium- and long-term climate resource prediction method, comprising: obtaining meteorological data from multiple sources, establishing a multimodal data set based on the meteorological data from multiple sources, and preprocessing the multimodal data set to generate a data time series; using an original numerical climate forecast model and a hybrid model to respectively forecast future medium- and long-term climate resources according to the data time series, and obtaining an original climate result predicted by the original numerical climate forecast model and a deep learning predicted climate result predicted by the hybrid model; fusing the original climate result and the deep learning predicted climate result, and correcting the fusion result to obtain a corrected predicted climate result; performing probability prediction on the corrected predicted climate result through uncertainty quantification processing to predict future medium- and long-term climate resources. The present invention can accurately predict future medium- and long-term climate resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a flow chart of the method of the present invention;

[0057] Figure 2 A flowchart of an embodiment of the method of the present invention;

[0058] Figure 3 It is a structural diagram of the system of the present invention. DETAILED DESCRIPTION

[0059] Now, exemplary embodiments of the present invention are described with reference to the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. These embodiments are provided to disclose the present invention in detail and completely and to fully convey the scope of the present invention to those skilled in the art. The terms used in the exemplary embodiments shown in the accompanying drawings are not intended to limit the present invention. In the accompanying drawings, the same units / elements are marked with the same reference numerals.

[0060] Unless otherwise specified, the terms (including technical terms) used herein have the commonly understood meanings to those skilled in the art. In addition, it is understood that the terms defined in commonly used dictionaries should be understood to have the same meanings as those in the context of the relevant fields, and should not be understood as idealized or overly formal meanings.

[0061] Embodiment 1:

[0062] The present invention proposes a medium- and long-term climate resource prediction method, such as Figure 1 As shown, including:

[0063] Step 1: Acquire meteorological data from multiple sources, establish a multimodal data set based on the meteorological data from multiple sources, and preprocess the multimodal data set to generate a data time series;

[0064] Step 2: using the original numerical climate prediction model and the hybrid model, respectively predicting the future medium- and long-term climate resources according to the data time series, and obtaining the original climate results predicted by the original numerical climate prediction model and the deep learning predicted climate results predicted by the hybrid model;

[0065] Step 3: Fusing the original climate result and the deep learning predicted climate result, and correcting the fusion result to obtain a corrected predicted climate result;

[0066] Step 4: Probabilistically predict the corrected predicted climate results through uncertainty quantification to predict the future medium- and long-term climate resources.

[0067] Among them, meteorological data includes: satellite observation data, meteorological station data, ocean observation data and geographic information data.

[0068] Among them, the linear interpolation method is used to adjust the temporal and spatial resolutions of meteorological data from multiple sources to consistent to establish a multimodal dataset.

[0069] The multimodal data set is preprocessed, including:

[0070] The multimodal data set is sequentially subjected to data cleaning, interpolation and denoising processes.

[0071] Among them, the hybrid model is a combination model based on CNN and RNN.

[0072] Among them, an integrated learning algorithm is used to fuse the climate results of traditional predictions and the climate results of deep learning predictions, and the fusion results are corrected based on Bayesian updating technology.

[0073] Among them, the uncertainty of the corrected predicted climate results is quantified based on the Monte Carlo simulation method, and a prediction confidence interval is generated according to the prediction results. The reliability of future medium- and long-term climate resources is determined by the prediction confidence curve.

[0074] The following is further described in conjunction with the embodiments of the present invention:

[0075] Example steps, such as Figure 2 As shown, including:

[0076] (1) Multimodal data integration and preprocessing:

[0077] 1) Collect a variety of meteorological data sources, including satellite observation data, weather station data, ocean observation data, geographic information data, etc., and use linear interpolation methods to adjust the temporal and spatial resolutions of data from different sources to be consistent, and construct a multimodal data set:

[0078] D total ={D satellite ,D station ,D ocean ,D GIS}

[0079] Among them, D total Denotes the integrated multimodal dataset, D satellite is the satellite observation data, D station is the weather station data, D ocean is the ocean observation data, D GIS It is geographic information data.

[0080] 2) Data preprocessing includes cleaning, interpolation, denoising and other steps to ensure the integrity and accuracy of the data:

[0081] D preprocessed =Preprocess(D total )

[0082] =Clean(D)+Interpolate(D)+Denoise(D)

[0083] Among them, Preprocess (D total ) represents the cleaning, interpolation and denoising of the data set.

[0084] 3) The span of the data time series covers at least the past 20 years of historical data:

[0085] D time-series ={D preprocessed,t-20 ,D preprocessed,t-19 ,…,D preprocessed,t}

[0086] Here, Dtime-series Contains historical data from the past 20 years for building time series models.

[0087] (2) Multi-level climate modeling and prediction:

[0088] 1) Use traditional numerical climate prediction models (such as the global climate model CFS, etc.) to make basic predictions for the climate in the next few months:

[0089]

[0090] in, It is the short-term climate result predicted by the traditional climate model CFS.

[0091] 2) Use deep learning algorithms to perform pattern recognition and post-processing corrections on climate data to improve the prediction accuracy of future climate resources. In practical applications, a hybrid model (this patent uses a CNN-RNN combination) is used to handle long-span sequence prediction tasks:

[0092]

[0093] in, It is the climate prediction result after pattern recognition and post-processing correction through deep learning model.

[0094] (3) Multimodal fusion prediction:

[0095] 1) Use ensemble learning methods (this patent uses XGBoost) to fuse the prediction results from different models to improve the stability and accuracy of the final prediction:

[0096]

[0097] in, It is the predicted value after fusing the results of different models through ensemble learning methods (such as XGBoost).

[0098] 2) Introduce Bayesian update technology to make real-time corrections to the model's predictions based on actual observation data to further improve prediction accuracy: When introducing Bayesian update technology, it is recommended to clarify the prior information used and how to update the posterior distribution, especially in the management of computational complexity in high-dimensional data scenarios.

[0099]

[0100] here, It is the final prediction result after correction based on the actual observed data through Bayesian updating. The two variables on the right side of the formula are conditional probability distribution and prior probability distribution respectively.

[0101] (4) Uncertainty analysis and probability prediction:

[0102] 1) Using Monte Carlo simulation to quantify the uncertainty of the prediction results and provide a probability distribution of climate resources rather than just a single numerical prediction:

[0103]

[0104] Where P(Y) represents the probability distribution of climate resources, N is the number of Monte Carlo simulation runs, is the result of each simulation.

[0105] 2) Generate forecast confidence intervals so that users can better understand the reliability of the forecast:

[0106]

[0107] Where CI stands for confidence interval, μ is the predicted mean, σ is the standard deviation, and z α / 2 is the critical value in the normal distribution table.

[0108] The present invention is mainly used in the field of energy management, and the invention will significantly improve the prediction accuracy and management efficiency of renewable energy. The present invention provides more reliable data support for power companies, assists in optimizing medium- and long-term power production plans, rationally dispatches power supply, and reduces dependence on traditional energy. Specifically, accurate climate resource forecasts can identify in advance the medium- and long-term changing trends of wind, solar, and hydropower generation potential, allowing the power grid to better balance supply and demand and reduce power generation volatility and grid pressure caused by climate uncertainty. In addition, this invention will also support energy investors to make more informed decisions by predicting future climate resource conditions and optimizing the site selection and return on investment analysis of new energy projects. In general, the application of the present invention in energy management will effectively improve the utilization rate of renewable energy, reduce energy costs, enhance the sustainability and stability of the energy system, and provide strong technical support for achieving a low-carbon economy and addressing climate change goals.

[0109] Embodiment 2:

[0110] The present invention also proposes a medium- and long-term climate resource prediction system 200, such as Figure 2 As shown, including:

[0111] The data acquisition unit 201 is used to acquire meteorological data from multiple sources, establish a multimodal data set based on the meteorological data from multiple sources, and preprocess the multimodal data set to generate a data time series;

[0112] The first prediction unit 202 is used to use the original numerical climate prediction model and the hybrid model to predict the future medium- and long-term climate resources according to the data time series, and obtain the original climate results predicted by the original numerical climate prediction model and the deep learning predicted climate results predicted by the hybrid model;

[0113] A fusion unit 203 is used to fuse the original climate result and the deep learning predicted climate result, and correct the fusion result to obtain a corrected predicted climate result;

[0114] The second prediction unit 204 is used to perform probability prediction on the corrected predicted climate result through uncertainty quantification processing to predict the future medium- and long-term climate resources.

[0115] Among them, meteorological data includes: satellite observation data, meteorological station data, ocean observation data and geographic information data.

[0116] The data acquisition unit 201 uses a linear interpolation method to adjust the temporal and spatial resolutions of meteorological data from multiple sources to be consistent, so as to establish a multimodal data set.

[0117] The data acquisition unit 201 preprocesses the multimodal data set, including:

[0118] The multimodal data set is sequentially subjected to data cleaning, interpolation and denoising processes.

[0119] Among them, the hybrid model is a combination model based on CNN and RNN.

[0120] The fusion unit 203 uses an integrated learning algorithm to fuse the climate results predicted by traditional methods and the climate results predicted by deep learning methods, and corrects the fusion results based on the Bayesian update technology.

[0121] The second prediction unit 204 quantifies the uncertainty of the corrected predicted climate results based on the Monte Carlo simulation method, generates a prediction confidence interval according to the prediction results, and determines the reliability of future medium- and long-term climate resources with the prediction confidence curve.

[0122] The present invention can accurately predict the future medium- and long-term climate resources.

[0123] Embodiment 3:

[0124] Based on the same inventive concept, the present invention also provides a computer device, which includes a processor and a memory, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, and is specifically suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding functions, so as to implement the steps of the method in the above embodiment.

[0125] Embodiment 4:

[0126] Based on the same inventive concept, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both a built-in storage medium in a computer device and an extended storage medium supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the steps of the method in the above embodiment.

[0127] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may 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-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0128] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, 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 device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0131] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0132] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A medium- and long-term climate resource prediction method, characterized in that: include: Acquire meteorological data from multiple sources, establish a multimodal data set based on the meteorological data from multiple sources, and preprocess the multimodal data set to generate a data time series; Using the original numerical climate prediction model and the hybrid model, the mid- to long-term climate resources in the future are predicted respectively according to the data time series, and the original climate results predicted by the original numerical climate prediction model and the deep learning predicted climate results predicted by the hybrid model are obtained; Fusing the original climate result and the deep learning predicted climate result, and correcting the fused result to obtain a corrected predicted climate result; The corrected predicted climate results are subjected to probabilistic prediction through uncertainty quantification processing to predict the future medium- and long-term climate resources.

2. The method according to claim 1, characterized in that The meteorological data includes at least one of the following: satellite observation data, meteorological station data, ocean observation data and geographic information data.

3. The method according to claim 1, characterized in that The linear interpolation method is used to adjust the spatiotemporal resolution of meteorological data from multiple sources to create a multimodal dataset.

4. The method according to claim 1, characterized in that: The preprocessing of the multimodal data set includes: The multimodal data set is sequentially subjected to data cleaning, interpolation and denoising processes.

5. The method according to claim 1, characterized in that The hybrid model is a combination model based on convolutional neural network CNN and recurrent neural network RNN.

6. The method according to claim 1, characterized in that An integrated learning algorithm is used to fuse the climate results predicted by traditional predictions and those predicted by deep learning, and the fusion results are corrected based on the Bayesian updating technology.

7. The method according to claim 1, characterized in that The uncertainty of the corrected predicted climate results is quantified based on the Monte Carlo simulation method, and a prediction confidence interval is generated according to the prediction results. The reliability of future medium- and long-term climate resources is determined by the prediction confidence curve.

8. A medium- and long-term climate resource prediction system, characterized in that: include: A data acquisition unit, used to acquire meteorological data from multiple sources, establish a multimodal data set based on the meteorological data from multiple sources, and preprocess the multimodal data set to generate a data time series; A first prediction unit is used to use the original numerical climate prediction model and the hybrid model to predict the future medium- and long-term climate resources according to the data time series, and obtain the original climate results predicted by the original numerical climate prediction model and the deep learning predicted climate results predicted by the hybrid model; A fusion unit, used to fuse the original climate result and the deep learning predicted climate result, and correct the fusion result to obtain a corrected predicted climate result; The second prediction unit is used to perform probability prediction on the corrected predicted climate result through uncertainty quantification processing to predict the future medium- and long-term climate resources.

9. The system according to claim 8, characterized in that The meteorological data includes at least one of the following: satellite observation data, meteorological station data, ocean observation data and geographic information data.

10. The system according to claim 8, characterized in that The data acquisition unit uses a linear interpolation method to adjust the temporal and spatial resolutions of meteorological data from multiple sources to be consistent, so as to establish a multimodal data set.

11. The system according to claim 8, characterized in that The data acquisition unit preprocesses the multimodal data set, including: The multimodal data set is sequentially subjected to data cleaning, interpolation and denoising processes.

12. The system according to claim 8, characterized in that The hybrid model is a combination model based on CNN and RNN.

13. The system according to claim 8, characterized in that The fusion unit uses an integrated learning algorithm to fuse the climate results predicted by traditional methods and the climate results predicted by deep learning, and corrects the fusion results based on the Bayesian update technology.

14. The system according to claim 8, characterized in that The second prediction unit quantifies the uncertainty of the corrected predicted climate results based on the Monte Carlo simulation method, generates a prediction confidence interval according to the prediction results, and determines the reliability of future medium- and long-term climate resources with the prediction confidence curve.

15. A computer device, characterized in that: include: one or more processors; a processor for executing one or more programs; When the one or more programs are executed by the one or more processors, the method according to any one of claims 1 to 7 is implemented.

16. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed, the method according to any one of claims 1 to 7 is implemented.

Images