129 results about "Downscaling" patented technology

The invention discloses a physical information neural network-driven ocean island typhoon scene generation method and system. The method comprises the steps of collecting multi-source heterogeneous meteorological data and performing space-time alignment preprocessing; constructing a coarse-scale space-time probability prediction model, capturing space correlation of meteorological elements by using a graph topology learning network, efficiently processing long-time-sequence dependence of typhoon evolution by integrating a state space model with linear complexity, and generating a probabilistic typhoon scene with coarse resolution through a multivariable joint distribution probability model; further constructing a physical downscaling model, taking a coarse-scale prediction result as condition input, and performing physical consistency downscaling on a coarse-scale scene by embedding an atmospheric fluid mechanics equation in a loss function as a physical hard constraint; and finally, outputting a high-resolution typhoon scene with probability reliability and physical authenticity.

The invention discloses a summer rainfall forecast correction downscaling method and system based on deep learning, and belongs to the technical field of meteorological prediction and climate simulation. In order to solve the technical problems of systematic deviation and spatial detail missing in forecasting, a deep learning downscaling model is constructed for summer multi-mode climate forecasting data published six months ahead of time in combination with high-resolution observation data and topographic data, deviation correction and spatial super-resolution reconstruction are performed on a summer rainfall forecasting result, and a deep learning downscaling model is constructed. And outputting the kilometer-level summer precipitation field of the target area. The model improves the spatial feature capture capability of multi-scale climate elements through an attention mechanism, and realizes the adaptive fusion of meteorological elements and topographic factors through a topographic perception module, thereby enhancing the medium and long term prediction precision of summer rainfall in a complex topographic region. According to the method, global climate information and local topographic features can be effectively integrated, and high-precision technical support is provided for medium-and-long-term prediction of regional summer rainfall, disaster prevention and reduction in flood seasons and water resource scheduling.

The invention discloses a multi-drive process multi-factor agricultural non-point source pollution prediction method based on coupling meteorological numerical forecasting. The method comprises the following steps: firstly, introducing numerical weather forecast data, and constructing a high-precision weather driving field with kilometer-level space and hour-level time resolution by combining WRF dynamic downscaling, DEM terrain correction and conservation resampling; secondly, establishing a multi-drive process coupling model system which comprises a meteorological drive layer, a hydrological response layer, a pollutant migration layer and a crop feedback layer and is used for simulating runoff production, sediment erosion, nitrogen and phosphorus migration and transformation and crop transpiration and root nutrient absorption processes; thirdly, performing precision verification on a simulation result by using observation data, and identifying key meteorological and hydrological factors through an error transfer matrix and a sensitivity analysis method; and finally, realizing parameter adaptive correction by adopting a long short-term memory network, finishing parameter optimization in combination with a multi-target genetic algorithm, packaging the model chain through a containerization technology, and realizing cross-platform deployment and visual output of a pollution load result. The method can be used for agricultural non-point source pollution prediction and management.

The invention relates to the technical field of satellite hydrological data processing, and particularly discloses a GRACE data super-resolution network space downscaling method fusing geographic information and environmental variables, which comprises the following steps: S1, acquiring original resolution GRACE data and original GLDAS data of a research area, and preprocessing the data; s2, dividing the data obtained by preprocessing in the step S1 into a training set and a test set, and training the GRACE data space downscaling model by using the training set to obtain a trained discriminator and a trained generator; and S3, inputting the GRACE low-resolution data in the test set and the high-resolution environment variable at the moment corresponding to the data into the generator trained in the step S2, and finally obtaining a downscaled high-resolution GRACE image. The method not only can be used for dynamic monitoring of regional scale underground water reserves and flood risk assessment, but also can be expanded and applied to scenes such as agricultural drought monitoring and ecological hydrological process simulation.

The invention relates to the technical field of remote sensing image intelligent processing, in particular to a multi-scale progressive surface temperature fusion downscaling method and device, and the method comprises the steps: preprocessing a surface temperature image in a multi-scale image library; constructing a multi-scale training data set containing the surface temperature and at least one auxiliary parameter; constructing a surface temperature downscaling network based on multi-parameter fusion, and constructing a domain transformation network based on heterogeneous high-frequency information guidance; training a surface temperature downscaling network and a domain transformation network based on a progressive dual-network joint training strategy; finely adjusting the downscaling model meeting a preset low-medium resolution condition; and performing low-medium-high progressive downscaling on the low-resolution surface temperature image to be processed to generate a multi-scale progressive surface temperature meeting a target high-resolution condition. The method fully considers the problem of descending scale difference of different resolutions, achieves the maintenance of the physical characteristics of the surface thermal field, and improves the processing efficiency of large-area remote sensing data.

The RIME-RF spatial downscaling-based photovoltaic output prediction method comprises the steps of collecting photovoltaic power data and local meteorological observation LMD data of a photovoltaic power station in a target area, extracting common data of numerical weather forecast NWP data and the local meteorological observation LMD data, and constructing an input feature set; the method comprises the following steps: optimizing hyper-parameters of a random forest (RF) algorithm based on a frost ice optimization (RIME) algorithm, constructing an RIME-RF model, and performing spatial downscaling on numerical weather forecast NWP data; a VMD-CNN-GRU-SE attention mechanism photovoltaic power prediction model optimized based on BKA is adopted, original numerical weather forecast NWP data is combined with photovoltaic power data to train the prediction model, and numerical weather forecast NWP data after spatial downscaling is combined with the photovoltaic power data to train the prediction model. According to the prediction method, changes of fine meteorological factors influencing the photovoltaic power can be more accurately captured, downscaling errors are remarkably reduced, and short-term power prediction precision is improved.

The invention relates to the technical field of meteorological data processing, in particular to a progressive wind speed field depth downscaling method and device fusing multi-source data and space-time perception, electronic equipment and a storage medium. Multi-scale topographic features are introduced and self-adaptive fusion is performed, so that the wind field reconstruction precision in a complex topographic region is improved; a progressive multi-stage downscaling network architecture is adopted, large-multiple stable downscaling from the kilometer level to the hectometer level is effectively achieved, and error accumulation is controlled; reconstruction is carried out by utilizing a network trained through physical constraint loss, it is ensured that the divergence and the vorticity of an output wind field conform to the aerodynamics law, the physical rationality and the scientific credibility of a result are enhanced, and therefore an efficient solution is provided for obtaining a near-ground wind field with high resolution and high physical credibility.

The invention discloses an SMAP soil hydrodynamic downscaling method based on a TOPMODEL theory, and the method comprises the steps: enabling original SMAP data to be matched with DEM data in spatial attributes through standardization processing, and obtaining standardized SMAP data; on the basis of DEM data, a confluence area and a gradient are calculated, and then a terrain index is calculated according to a TOPMODEL model; constructing a time sequence through the standardized SMAP data, and executing space-time stabilization processing to obtain a basic soil moisture field; based on the terrain index and the basic soil moisture field, preliminary soil moisture data is obtained through calculation by means of a downscaling model, and target soil water grating data with high resolution are obtained by optimizing the quality of the preliminary soil moisture data; according to the downscaling method, the structure is simple and clear, batch operation is facilitated, and low-resolution SMAP soil water data can be upgraded into high-resolution soil water data with the same resolution as DEM data in complex terrain areas such as mountainous areas, hills and river valleys.

The invention discloses a wind field correction method based on multi-stage hybrid physics-data driving, and relates to the technical field of wind field data processing and climate mode application. Executing dynamic downscaling simulation by using a regional climate simulation tool to obtain a multi-dimensional candidate variable field; screening target candidate climate variables according to ERA5 wind speed data, and forming a multi-dimensional key variable field; constructing a to-be-trained model comprising a multi-scale spatial feature extraction module, a time dynamic feature extraction module, a multi-scale spatio-temporal feature fusion module and an output module, and training by using the multi-dimensional key variable field to obtain a hierarchical multi-scale spatio-temporal fusion network for correcting wind field data; and by taking the wind field correction field as a boundary condition, generating a high-resolution wind field correction field by using a complex terrain micro-scale flow field refinement solver. The method not only improves the accuracy of wind speed prediction, but also ensures the physical authenticity and consistency of a high-resolution wind field under the influence of complex terrains through physical constraints.

The invention discloses an earth surface downlink short-wave radiation space downscaling method based on a super-resolution reconstruction technology, belongs to the technical field of meteorological data analysis, and solves the problems of insufficient solar radiation data space resolution and weak earth surface heterogeneity characterization capability in the prior art. Carrying out super-resolution reconstruction on the radiation data set by adopting a super-resolution combination model, and carrying out weighted fusion on a reconstructed data set output by the super-resolution combination model; according to the method, the output results of the multiple models are subjected to weighted fusion, weight optimization of the super-resolution sub-models is performed in combination with the high-resolution terrain factors, spatial downscaling from 5km to 1km resolution is finally realized, the spatial precision of the solar radiation data is remarkably improved on the basis of keeping the hour-level time resolution of the original data, and the accuracy of the solar radiation data is improved. And more earth surface heterogeneity information can be captured, and meanwhile, the high-time-resolution application requirement is met.

The present invention belongs to the field of multi-energy complementary coordinated dispatching. Disclosed is a meteorological-downscaling-coupled dispatching method for the power generation and consumption of a hydro-wind-solar system. The method comprises: using a support vector machine regression algorithm to identify different hydrometeorological variable data, and establishing a statistical relationship between observation data and a meteorological factor to implement high-resolution spatial downscaling; using a wind-solar power-generation empirical formula to calculate a wind-solar output process; and introducing a series of time-series peak regulation modes to determine a regulated peak and a hydro-wind-solar power consumption linkage equation, thereby avoiding overestimation of energy consumption caused by neglecting climate change impacts and short-term power generation rules. By means of the analysis of engineering examples consisting of Yunnan Lancang River Basin and surrounding wind-solar power stations thereof, the result shows that the present invention can effectively reduce hydrometeorological downscaling errors, and more accurately describe hydro-wind-solar energy power generation rules by means of a hydro-wind-solar power consumption linkage equation, thereby making the dispatching result show better accuracy and reliability.

The invention discloses a unified rainfall downscaling method based on a multi-stream analysis diffusion model, and the method comprises the steps: collecting the meteorological data of a target region, including low-resolution rainfall data and preset auxiliary variable data, generating a low-resolution rainfall field, and calculating the rainfall deviation between a high-resolution real rainfall field and the low-resolution rainfall field; based on preset auxiliary variable data, combining the low-resolution rainfall data to form multi-modal condition input; constructing a precipitation downscaling model which takes the analysis diffusion model as a framework and is combined with a mixed attention U-shaped network, and predicting precipitation deviation; and training the rainfall downscaling model, performing reasoning by adopting the rainfall downscaling model, inputting meteorological data of the target area, outputting predicted rainfall deviation, and superposing the predicted rainfall deviation to the low-resolution rainfall field to obtain super-resolution rainfall. According to the method, multi-source meteorological and geographic information is effectively fused, high-fidelity and high-resolution rainfall data is generated, the generalization ability is high, and the calculation efficiency is high.

The invention relates to a multi-modal meteorological data downscaling generation method and device and computer equipment. The method comprises the following steps: acquiring topographic data and meteorological data in a to-be-measured area; inputting the topographic data and the meteorological data into a pre-trained deep learning model, carrying out feature coding on the topographic data and the meteorological data through the deep learning model to obtain a topographic feature vector and a meteorological feature vector, carrying out feature interaction on the topographic feature vector and the meteorological feature vector to obtain interaction features, and carrying out feature extraction on the interaction features; and carrying out downscaling decoding on the interaction features to obtain downscaling meteorological data. By adopting the method, the downscaling generation efficiency and the physical rationality of the meteorological data can be considered.

The invention discloses a coastal wind turbine group generation power multi-scale space-time prediction method, which mainly comprises the following steps: carrying out space interpolation and error correction on a target wind turbine position by adopting a numerical weather forecast statistical downscaling technology, generating a high-resolution wind speed prediction sequence covering a short term and a long term, aligning and splicing the predicted wind speed, the field actually-measured wind speed, the environment and the unit operation variables into node dynamic input characteristics; the method comprises the following steps: encoding longitude and latitude and time sequence monitoring data of N fans of a coastal fan group into graph nodes, determining an edge weight according to geographic distance and wake flow coupling, and forming a fan graph network containing static and dynamic characteristics; and inputting the static and dynamic feature sequences into a graph neural network comprising a space attention layer, a time recursion layer and a physical constraint regular term, completing model training, and outputting the generated power of each fan in a plurality of time steps in the future and the total power predicted value of the fan group. The method can provide powerful support for wind power plant operation scheduling, power grid-connected management and new energy consumption.

The invention discloses a power transaction supply and demand prediction method and system based on a regional meteorological large model, and the method comprises the steps: collecting and fusing multi-source historical meteorological data and geographic data of a target region, and forming a special data set; training the large base meteorological model by using the special data set, and outputting a preliminary meteorological forecast according to the large base meteorological model and the initial meteorological data of the target area; inputting the preliminary meteorological forecast into the super-resolution downscaling model, and outputting super-resolution meteorological element forecast by taking the super-resolution geographic data of the target area as a condition; and inputting the super-resolution meteorological element forecast, the historical power data of the target area and the market supply and demand data into the supply and demand prediction model, and outputting a supply and demand prediction curve according to a prediction task. According to the invention, super-resolution meteorological element forecast is generated through the large base meteorological model and the super-resolution downscaling model, so that an accurate supply and demand prediction curve can be obtained by using the supply and demand prediction model.

The invention discloses a rainfall downscaling method and system based on a deep learning network model fusing rainfall priori knowledge, and the method comprises the steps: firstly collecting the topographic data and low-resolution day-by-day rainfall data of a target region, and taking the data as input data; a short-term high-resolution precipitation field generated in a mesoscale weather forecast WRF mode is used as training truth value data; according to the method, the function of accurately downscaling the rainfall data in combination with the convolutional neural network and the long and short term memory network is realized, the spatial-temporal correlation of rainfall is fully considered in the downscaling process, and meanwhile, a likelihood function combined with coupled censored data, Box-Cox conversion and time variation variance Gaussian distribution is adopted as a rainfall loss function; the method not only can represent zero expansibility, skewness and heterovariance characteristics of rainfall, but also can improve the rainfall downscaling precision and quantify the uncertainty of rainfall downscaling, and is suitable for wide popularization and use.

The invention provides a hectometer-level short temporary rainfall forecasting method based on generative adversarial network downscaling and physical constraint, which belongs to the technical field of weather forecast, and is characterized in that S-band radar jigsaw and high-resolution X-band radar observation data are fused through an adaptive rule to generate a fused radar echo combined reflectivity, and the combined reflectivity is used for forecasting the rainfall of the hectometer-level short temporary rainfall. A super-resolution downscaling model is constructed by using a generative adversarial network, a long-sequence high-resolution radar echo training data set is reconstructed, a deep learning forecasting model is constructed, a high-resolution radar echo sequence is used as input, space-time modeling training is performed by using a composite loss function, and a future radar echo forecasting field is output. The Z-R relation is converted into quantitative rainfall forecast, and finally the quantitative rainfall forecast with the spatial resolution reaching the hectometer level is generated. The interpretability of the model is enhanced through physical constraint, the bottleneck of insufficient hectometer-level resolution training data is effectively solved, and the spatial refinement degree of short temporary rainfall forecasting and the forecasting capacity for a severe convection system are remarkably improved.

The invention discloses a meteorological data spatial downscaling training method and system based on self-supervised learning. An inverted meteorological data pyramid is constructed, pre-training is performed based on meteorological data internal repeatability, and self-supervised fine tuning is performed based on a spatial position relation. According to the invention, the spatial downscaling of the meteorological data can be realized without using high-resolution data, so that the dependence on the high-resolution data in the meteorological downscaling process is reduced; the method can be flexibly applied to various deep learning models to realize self-supervised downscaling of meteorological data; in actual tasks such as typhoon positioning, the method can achieve the same downscaling effect as a supervised learning method. Experimental results show that the method is small in size under the condition that a high-resolution training set is not needed, high-quality meteorological variable downscaling is achieved, and the problem that actual meteorological forecast lacks a high-resolution data set is solved.

The invention discloses a feature fusion and region correction-based weather forecast method and a medium. The method comprises the following steps of: acquiring an initial weather forecast result generated by a weather forecast basic large model; inputting the initial weather forecast result and the multi-source environment auxiliary data of the target area into an area fine correction module; and performing deviation correction and spatial downscaling processing on the initial weather forecast result based on the multi-source environment auxiliary data through a region fine correction module to generate a high-resolution weather element forecast result of the target region. According to the invention, the initial forecast is generated by the basic large model, and the multi-source data is fused by the area correction module to carry out deviation correction and downscaling, so that the high-resolution weather forecast of the target area is generated. Through two-stage cooperative processing, the precision and resolution of weather forecast in a specific area are remarkably improved, end-to-end conversion from a forecast result to an industry decision index is realized, and the practicability is high.

The invention discloses an extension period temperature downscaling method and system based on deep learning in a complex terrain, and belongs to the technical field of temperature prediction. The method comprises the following steps: aiming at the problem of defects in an existing temperature downscaling task, acquiring a back-calculation data set and a reanalysis data set, preprocessing the back-calculation data set and the reanalysis data set to obtain a multivariable training data set, acquiring terrain data, and extracting multi-scale features of a terrain by using a convolutional neural network; constructing an extended-period temperature downscaling model based on deep learning, training the model, fusing topographic data, determining an objective function, and obtaining a trained model; and inputting the low-resolution air temperature forecast data into the trained model to obtain a high-resolution downscaling result. According to the method, complex topographic data and extended period forecast data are adopted, a deep learning method is utilized to extract nonlinear features of air temperature data, respective advantages of multi-scale features are mined and integrated, the forecast air temperature is downscaled, and a high-resolution downscaling result is output.

The invention relates to a complex terrain air temperature refined downscaling method based on a TopoMW model, and belongs to the technical field of meteorological element downscaling, and the method comprises the steps: obtaining multi-source data of a target region, carrying out the preprocessing of the multi-source data, and fusing the multi-source data into a feature set; constructing a complex terrain multi-weighting matrix based on the preprocessed multi-source data; a TopoMW model improved based on UNet is established, a weighted MSE loss function is adopted to train the TopoMW model on the basis of a feature set in combination with a complex terrain multiple weighting matrix, an optimal downscaling model is obtained, and the contradiction between model statistic multiplexing and performance retention is solved; inputting a feature set to be predicted into the optimal downscaling model to obtain an initial prediction result; the initial prediction result is corrected based on the complex terrain multiple weighting matrix to obtain a final result, extreme deviation is eliminated, and accurate generation of the 30-meter high-resolution air temperature under the complex terrain is achieved.

The application belongs to the field of sub-seasonal climate prediction, and provides a sub-seasonal climate prediction method and system based on the combination of dynamic model downscaling and machine learning downscaling, which comprises the following steps: S1, generating initial field and model boundary field information required for dynamic downscaling based on global climate model output data; S2, driving regional climate model to perform dynamic downscaling by using the initial field and model boundary field information; S3, generating input field required for machine learning downscaling based on the output circulation field information of the regional climate model; S4, performing machine learning downscaling correction optimization on the output circulation field of dynamic downscaling based on a convolution model; S5, performing machine learning super-resolution based on the output data of machine learning downscaling correction; and S6, generating sub-seasonal prediction information of double downscaling of dynamic model and machine learning. The application utilizes the complementary advantages of dynamic downscaling and deep learning downscaling, improves the sub-seasonal prediction skill, and copes with new challenges brought by climate change.

The invention relates to the technical field of rainfall analysis, and discloses a rainfall data high-precision downscaling method based on terrain fusion and residual optimization, which comprises the following steps: firstly obtaining original rainfall data of a target area and aggregating the original rainfall data into downscaling reference data, and marking an effective rainfall area of a digital elevation model; extracting terrain factors of the region, dynamically normalizing the terrain factors, constructing a random forest regression model by taking the reference data as response variables and standard terrain factors as explanatory variables, and predicting to obtain a preliminary rainfall result; and then calculating a residual error correction preliminary result to obtain an intermediate result, and comparing the reference data with the local mean value of the intermediate result to zoom the intermediate result to obtain a final rainfall prediction result. According to the method, the refinement degree of downscaling prediction of rainfall data is improved in combination with terrain factors, errors are controlled through residual correction and mean scaling, dynamic normalization adapts to different terrains, and the efficiency is improved through block processing.

The application discloses a kind of spatiotemporal modeling ground surface temperature downscaling method and system considering energy constraint, comprising: 1) acquisition and preprocessing of multi-source remote sensing and meteorological data;2) feature grouping and spatiotemporal feature tensor construction;3) time feature and spatial feature extraction;4) spatiotemporal feature modeling and high-resolution ground surface temperature estimation;5) loss construction and cross-scale consistency constraint;6) model iterative training and result output.The application breaks through the modeling limitation of time variation law and spatial detail description in the prior art, realizes the collaborative promotion of time continuity and spatial fine expression of ground surface temperature.On this basis, the overall temperature deviation and the problem of insufficient physical rationality commonly existing in the prior art are also effectively avoided, and the application is significantly superior to the prior art in terms of timing stability, spatial accuracy and physical reliability.

The invention discloses a physical information guided wind speed field data downscaling method fusing terrain and time perception. The method comprises the steps of 1, multi-source physical constraint modeling and data construction; step 2, constructing a generative downscaling framework based on a conditional diffusion probability model; step 3, designing a noise prediction network structure fusing terrain and time prior; and 4, model training and rapid sampling reasoning. The method solves the problems that the texture of a wind speed field generated by an existing deep learning downscaling method is excessively smooth, high-frequency turbulence details are lost, and the terrain forcing effect and the wind speed seasonal / daily periodic physical law are ignored, and provides a framework combining explicit physical feature engineering and a conditional diffusion model. Through dual deep fusion of terrain and time information, high fidelity, physical consistency and reasoning efficiency of a generated result are considered.

The invention discloses a high-temporal-spatial-resolution surface temperature integration method and system based on optimal interpolation, and relates to the field of surface temperature integration, and the method comprises the steps: obtaining a surface temperature image, and carrying out the preprocessing of the image; constructing an auxiliary data set and performing preprocessing; performing spatial interpolation by adopting a thin-plate spline function to obtain a high-temporal-spatial-resolution surface temperature interpolation result; establishing a mapping relation between the low-spatial-temporal-resolution surface temperature and the auxiliary variables, and obtaining a high-spatial-resolution surface temperature spatial downscaling result; performing space-time fusion on the surface temperature by adopting a space-time adaptive reflectivity fusion model to obtain a high-space-time resolution surface temperature space-time fusion result; and carrying out weighted integration on the surface temperature space downscaling result and the space-time fusion result by adopting an optimal interpolation method. According to the method, the problem that an existing method still has certain limitations in surface temperature downscaling and space-time fusion technologies is solved.

The invention discloses a correction method and device for a multi-classification rainfall phase state, and relates to the technical field of weather forecast, and the method comprises the steps: employing a step-by-step condition dichotomy rainfall phase state probability correction algorithm to determine the probability of multi-phase rainfall after preliminary correction for a full space-time sample; and carrying out spatial downscaling processing on the probability of the multi-phase rainfall after preliminary correction by adopting a rainfall phase probability change rate spatial downscaling algorithm to obtain a high-resolution rainfall phase probability forecast correction product. And a poor correction effect caused by imbalance between event sample sizes can be eliminated to the greatest extent. And in the face of kilometer-level resolution downscaling, the calculation efficiency is improved, and meanwhile, the spatial continuity of the precipitation phase state probability can be ensured.

The application discloses a kind of physical information neural network driven ocean island typhoon scene generation method and system, comprising: collecting multi-source heterogeneous meteorological data and carrying out space-time alignment preprocessing;Build coarse scale space-time probability prediction model, utilize graph topology learning network to capture the spatial correlation of meteorological elements, integrate the state space model of linear complexity to efficiently process the long-time sequence dependence of typhoon evolution, and generate coarse resolution probabilistic typhoon scene through multivariate joint distribution probability model;Further build physical downscaling model, input coarse scale prediction result as condition, embed atmospheric fluid mechanics equation as physical hard constraint in loss function, and physically consistent downscaling is carried out to coarse scale scene;Finally output high-resolution typhoon scene with probability reliability and physical authenticity.

The invention discloses a regional high-resolution atmosphere reanalysis data construction method based on dynamic downscaling, and aims to improve the depiction capability of medium and small-scale weather processes. According to the method, global reanalysis data is used as a background field, multi-source observation data and high-resolution land utilization information are fused, multi-time and short-period simulation is executed by constructing a multi-layer nested WRF region mode and combining an assimilation strategy combining analysis Nudging and observation Nudging, and a reanalysis data set with hour-level and kilometer-level resolution is constructed step by step. Through comparison and verification of multi-source observation data, the constructed data set is obviously superior to ERA5 in the aspects of ground meteorological elements and vertical profile structure reduction, and has higher regional adaptability and simulation precision. The method is suitable for a plurality of high-resolution meteorological application scenes such as regional artificial intelligence meteorological large models, urban fine meteorological services, high-influence weather disaster early warning, regional climate change evaluation, energy meteorological resource evaluation and the like.

The application discloses a kind of based on multi-stage mixed physics-data driven wind field revision method, it is related to wind field data processing and climate model application technical field. Utilize regional climate simulation tool to execute dynamic downscaling simulation, obtain multidimensional candidate variable field;According to ERA5 wind speed data, target candidate climate variable is screened, and multidimensional key variable field is formed;Hierarchical multi-scale spatio-temporal fusion network for revising wind field data is obtained by constructing the to-be-trained model including multi-scale spatial feature extraction module, time dynamic feature extraction module, multi-scale spatio-temporal feature fusion, output module and training using multidimensional key variable field;With wind field revision field as boundary condition, high-resolution wind field revision field is generated by using complex terrain microscale flow field refining solver.The application not only improves the accuracy of wind speed prediction, but also ensures the physical authenticity and consistency of high-resolution wind field under the influence of complex terrain through physical constraint.