32 results about "Climate forecast" patented technology

A method of generating short-, medium-range and seasonal-timescale weather or climate forecasts by running an ensemble of computer models on a distributed computing system or network. Individual model integrations are interrogated to select those that most closely ressemble observed conditions in the present and recent past and the forecast based on a weighted average of future predictions based on this subset of the ensemble. The selection criteria determining which models are deemed to fit the observations most closely may be adjusted to optimize the use of observations in forecasting specific climate variables or geographic regions in order to develop forcasts tailored to particular applications.

Methods and systems for generating a neural network (NN)-based climate forecasting model are disclosed. The methods and systems perform steps of generating a multi-model ensemble of global climate simulation data by combining simulation data from at least two global climate simulation models; pre-processing the multi-model ensemble of global climate simulation data, where the pre-processing comprises at least one action of spatial re-gridding, temporal homogenization, and data augmentation; training the NN-based climate forecasting model on the pre-processed multi-model ensemble of global climate simulation data; and validating the NN-based climate forecasting model using observational historical climate data. Embodiments of the present invention enable accurate climate forecasting without the need to run new dynamical global climate simulations on supercomputers. Also disclosed are benefits of the new methods, and alternative embodiments of implementation.

The invention discloses a short-term climate forecast method based on Kalman filtering and evolution modeling. The method comprises the following steps of: establishing a linear model about a forecast factor by the Kalman filtering at first; and simulating an error sequence approaching the Kalman filtering by using a non-linear ordinary differential equation math model on the basis of the linear model a, and performing error forecast. An evolution algorithm is an evolution process for simulating the nature by using a computer, in particular a calculation method for solving complicated problems by simulating biological evolution processes, and has the intelligent characteristics of self-adaptation, self-organization, self-learning, internal parallelism and the like. The two algorithms are combined with each other, so the natural characteristic of the climate can be simulated better than being simulated by a pure linear model, so the climate forecast precision is enhanced. By the method, short-term sunshine duration, temperature and rainfall can be forecast, so future knowledge of the short-term climate can be provided.

The invention discloses a terahertz and millimeter wave cloud radar data joint inversion method. The method comprises the following steps of: establishing a relationship among terahertz radar reflectivity factors, a millimeter wave radar reflectivity factors and cloud attenuation; establishing a relationship between the radar reflectivity factors of a terahertz frequency band and a millimeter wavefrequency band and cloud physical parameters, establishing a probability formula based on a Bayesian theory, iteratively operating the cloud physical parameters, and taking an error between a physical model value and an observed value as a criterion of iterative convergence to obtain an optimal solution; and iteratively calculating the backscattering cross section and attenuation cross section ofcloud particles according to the cloud physical parameters obtained by inversion, updating the function relationship between the radar reflectivity factors and the cloud attenuation to obtain a new inversion value, and taking the change of the cloud attenuation coefficient in iterative operation as a convergence condition to obtain a dual optimal solution of the cloud physical parameters. Data support is provided for climate prediction, weather forecast and atmospheric science research, and effective application of terahertz cloud radar and millimeter wave cloud radar in a meteorological observation system is promoted.

The invention discloses an intelligent climate prediction system. The prediction system comprises a data preprocessing subsystem, an intelligent similar year search prediction subsystem, a time series-based prediction subsystem, a climatic characteristic-based prediction subsystem, a circulation index-based prediction subsystem, an EOF reconstruction-based prediction subsystem, a re-analysis fieldand multi-mode prediction product-based prediction subsystem, a decision tree and any-time forest prediction subsystem, a machine learning-based extended period prediction subsystem, an MJO-based prediction subsystem and an intelligent recommendation subsystem. The intelligent climate prediction system is high in accuracy, stable in prediction and high in pertinence, and prediction objects, prediction areas and prediction methods can be expanded and increased.

The present invention relates to a South China area summer average temperature short term climate prediction method and system which comprises the steps of calculating the summer average temperature of a South China area, calculating the meteorological parameter data of an early winter tropical India ocean area corresponding to the above summer average temperature data, especially using a TWNIO index which expresses the tropical northwest India Ocean region thermal condition abnormity as the above meteorological parameter, using the above summer average temperature data and meteorological parameter data to construct a prediction model, and predicting the summer average temperature of South China through the prediction model. According to the method, the temperature can be accurately and effectively predicted through early meteorological parameter, especially South China area summer high temperature, the method has strong prediction ability, and effective warning information can be provided for a state climate disaster.

The embodiment of the invention provides a climate prediction method and device, a computer readable storage medium and a server, and relates to the field of weather calculation. The method comprisesthe steps: acquiring climate prediction data and weather observation data of each day in a past first time period, wherein the weather data of the (K + i) th day are predicted on the Kth day; obtaining climate prediction data of the ith day in the future; calculating a prediction error according to the climate prediction data and the weather observation data of each day in the past first time period; and correcting the climate prediction data of the ith day in the future according to the prediction error. The historical data contains a large number of error characteristics; the prediction error is calculated according to the weather prediction data and the weather observation data of each day in the past first time period, and the prediction error is used for correcting the future weatherprediction data, so that the corrected weather prediction data is more accurate, and the weather prediction capability is improved.

The invention discloses a whole-process numerical simulation and dangerous case forecasting method for mountain disasters. The method comprises the following steps of S1, forecasting high space-time rainfall in a mountain area; S2, carrying out hydrodynamic process and numerical simulation, namely establishing a hydrodynamic process model, and solving the hydrodynamic process model; S3, performing mountain torrent and debris flow disaster motion model and numerical simulation; and S4, analyzing the small watershed disaster risk and forecasting a dangerous case. According to the disaster whole-process scene simulation driven by a climate forecast result, the disaster harmfulness prediction and the risk loss dynamic quantitative evaluation are realized, the current disaster grade forecast is improved to dangerous case forecast, and the accurate disaster prevention and accurate rescue are served.

The invention discloses an energy distribution management system and method for an off-grid micro-grid. The system comprises: a data information comparison processing module, wherein a cloud computing data center for comparing and processing natural energy data and weather predication data; and a data receiving and storing module, wherein the input end of a far-end client receives external natural energy data and cloud climate prediction data and stores the data in a database. According to the invention, climate prediction data is received while electric quantity energy storage data of a generator is transmitted, and in cooperation with real-time monitoring and management of the micro-grid system,and effective control of energy is realized while the utilization rate of the energy is improved; and in addition, through a set system control scheme, and combining parameters such as fluctuation frequency of output power, power provided by a battery, power fluctuation frequency of load equipment and the like and utilizing the advantages of the micro-grid distributed power supply, the voltage of a direct-current bus can be maintained, and the problem of power fluctuation caused by renewable energy sources can be stabilized.

The invention provides an inter-year forecast initialization method and system, which relate to the technical field of inter-year climate prediction. The method comprises the steps that a forecast initial field is obtained based on historical simulation of a global coupling mode, and multiple sets of atmosphere forecast initial fields are obtained; based on ocean and sea ice historical simulation driven by atmosphere reanalysis data, the climate state of the ocean and sea ice historical simulation is corrected to be the climate state of ocean and sea ice of global coupling mode historical simulation, and an ocean and sea ice forecast initial field is obtained; the obtained multiple sets of atmosphere forecast initial fields are combined with land, ocean and sea ice initial fields to construct multiple sets of initial fields of inter-year ensemble forecast; a pre-constructed inter-year forecast system is initialized, and historical post-reporting or inter-year forecast of a future business type set is implemented. According to the method, the business type inter-year prediction can be ensured, the influence caused by impact after initialization is reduced, the accuracy of natural change rate prediction of the prediction system is improved, and thus the business prediction level of the inter-year prediction system is enhanced.

The invention discloses a power-statistics combined seasonal climate prediction method based on a predictable climate mode. A predictable climate mode for determining the annual tendency of the rainfall in the same period in the Chinese seasons is extracted by utilizing historical observation data, and a physical statistical model of the same period relation between the predictable climate mode and the annual tendency of target variables such as rainfall is established; predication of a climate mode by a dynamic mode is substituted into a physical statistical model, and prediction of interannual tendency of target variables such as rainfall is realized; and the predicted interannual tendency is superposed with the observation distance flatness of the last year to obtain the distance flatness prediction of the target variable. Compared with the defect of direct prediction of target variables such as Chinese rainfall in a power mode, the prediction capability of the power mode for main climate modes is fully utilized, and prediction of the predictable climate modes by the power mode is combined with a physical statistical model established according to historical data; and the optimal climate mode is selected through independent sample inspection, so that the prediction of the target variable is realized, and the accuracy of the prediction of the seasonal drought and flood in China can be effectively improved.

The invention relates to the technical field of irrigation prediction, and provides a method and a system for predicting irrigation flow according to climate change. The system comprises a historicaldatabase, a climate prediction database, an on-site acquisition sensor and an irrigation water volume prediction and correction module. The method comprises the following steps: obtaining weather prediction data from a network according to original historical irrigation water volume data; calculating predicted irrigation water volume, and correcting the predicted irrigation water amount by using the field environment data; finally, acquiring the actual irrigation water amount, adding climate change factors to the basis of the crop irrigation flow, predicting the irrigation flow of different crops, and designing intelligent irrigation flow prediction according to the climate factors, so the defects of existing consistent quantitative irrigation can be overcome, and water supply is more reasonable.

The invention discloses a sub-seasonal climate prediction method and system based on ten-day tendency and physical modal modeling. The method includes calculating the ten-day tendency flatness of the atmospheric element field and the predictive variable by using historical data; through singular value decomposition, extracting main physical modes of early-stage tropical atmosphere outward long-wave radiation field and mid-high latitude atmosphere 500 hPa potential height field ten-day tendency bathymetric variation for determining predictive variable ten-day tendency bathymetric variation and taking same as predictive factors; utilizing a multiple regression method to construct a prediction model for predicting the relationship between the variable ten-day tendency distance and the early-stage main physical mode, and determining the optimal physical mode through historical return; substituting the optimal physical mode observed in the early stage into the prediction model, so that the 10-day tendency bay prediction of the prediction variable is realized; and superposing the predicted 10-day tendency flat distance with the observation or prediction flat distance of the previous 10-day so as to obtain the flat distance prediction of the current 10-day. The modeling method and system based on the ten-day tendency flatness and the physical mode established by the invention can effectively improve the sub-seasonal climate prediction accuracy.

The invention discloses an offshore climate prediction method and device based on a fast matrix decomposition method. Using an existing Matern kernel function-based Gaussian random field model and collected offshore climate physical data to carry out physical value prediction on any spatial point location of the region; generating a task dependency graph by using OpenMP and scheduling decomposition tasks in the task dependency graph according to a dependency sequence; and after the MPI is used for controlling matrix partitioning, non-blocking communication among different processes is carried out. The invention provides a brand-new MPI + X normal form which can be operated in a distributed environment and can also be programmed by using a native OpenMP task model, and multi-thread execution communication effectively improves the operation efficiency of a large-scale decomposition algorithm. The invention provides a brand-new two-dimensional data layout in a distributed environment, and the load on each process in a distributed decomposition algorithm is balanced; and the tasks with relatively high calculation amount are allocated to different processes, so that the communication overhead of a decomposition algorithm is reduced, and the efficiency is improved.

The invention discloses a forecasting method of Chinese seasonal climate based on main SVD modal modeling. The forecasting method selects outward long radiation (OLR) of low latitudes and 500hPa height fields of middle and high latitudes as forecasting factor variables to conduct seasonal climate forecasting of Chinese rainfall and temperature. The method comprises steps of inter-annual increments of the forecasting factor variables, and extracting the main SVD modal time coefficient of the forecasting factor variables by singular value decomposition (SVD) as the actual forecasting factor; and constructing the climate statistical forecasting model by using multiple linear regression, forecasting the inter-annual increment of the forecasting object of the specified year, adding inter-annual increment to the observation anomaly of the previous year, and obtaining the seasonal anomaly of the climate forecasting variable. The method adopts the forecasting factor variables by taking into account the influence of tropical and extra-tropical atmospheric anomalous signals, extracts the most closely coupled coupling modes as the forecasting factors, and predicts the inter-annual increment to avoid the interference of the inter-decadal variation rate, to ensure that the seasonal climate forecasting better and more stable.

The invention discloses a sub-season-season-interannual scale integrated climate mode set prediction system. The system comprises an initialization module, a high-resolution climate mode module, a set prediction module and a prediction post-processing module. The initialization module is used for downloading, extracting and processing multi-source data including atmosphere, land surface, ocean and sea ice, and using external parameters for controlling and achieving preprocessing such as data selection, checking and horizontal and vertical interpolation of meteorological elements, and the initialization module is suitable for data preprocessing of operation in all climate modes. The high-resolution climate mode module is used for carrying out objective quantitative prediction on multi-circle coupling collaborative change of atmosphere, land surface, ocean and sea ice; the set prediction module is used for generating any number of set prediction sample members combined by physical parameter tendency random disturbance and an initial value time lag method. The invention provides the climate prediction with high resolution, high prediction accuracy and good controllability, and provides multi-time-scale and multi-space-scale climate prediction of atmosphere, land surface, ocean and sea ice.

The invention discloses a method and device for predicting environmental pollution, and relates to the technical field of computers. According to a specific embodiment, the method comprises the following steps of: constructing a new target climate prediction model based on a plurality of climate prediction models; utilizing the target climate prediction model to predict various target climate data in a future set time range; and predicting the environmental pollution condition in the future set time range according to the various target climate data, the climate and the correlation characteristics of environmental pollution. By constructing the target climate prediction model and the correlation characteristics of the environmental pollution, the problem that the accuracy of predicting the environmental pollution data within a long time range is low is solved, the automation degree of predicting the environmental pollution is improved, and the efficiency of predicting the environmental pollution is improved.

The invention discloses an abnormal relative tendency generation method and system for climate prediction. Performing time smoothing on the original data of the climate variables to remove the high-frequency variation rate; subtracting the climate state of the smooth data to obtain a distance plane; on the basis, according to the target frequency band, selecting the length of the early-stage average time period of each moment, and defining a corresponding recent abnormal background; the recent abnormal background is subtracted, the low frequency variation rate is removed, and the abnormal relative tendency is obtained; and modeling and predicting the abnormal relative tendency, and substituting and adding the low-frequency variation rate as a known recent abnormal background to realize prediction of a distance and an original field. According to the method, the problems of time boundary and multi-scale of climate prediction are solved, target frequency band information is extracted at the tail end of a time sequence only by using historical and current data, the relative tendency of high-frequency anomaly on a known recent anomaly background can be highlighted, only the relative tendency of anomaly needs to be predicted, errors caused by introduction of prediction low-frequency variation rate are avoided, and the prediction accuracy is improved. And the accuracy and the stability of climate prediction are effectively improved.