CN114626577B - Method for forecasting winter precipitation phase state by utilizing artificial intelligence - Google Patents
Method for forecasting winter precipitation phase state by utilizing artificial intelligence Download PDFInfo
- Publication number
- CN114626577B CN114626577B CN202210158802.XA CN202210158802A CN114626577B CN 114626577 B CN114626577 B CN 114626577B CN 202210158802 A CN202210158802 A CN 202210158802A CN 114626577 B CN114626577 B CN 114626577B
- Authority
- CN
- China
- Prior art keywords
- data
- precipitation
- forecasting
- phase state
- utilizing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001556 precipitation Methods 0.000 title claims abstract description 82
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 12
- 238000012795 verification Methods 0.000 claims abstract description 10
- 238000012876 topography Methods 0.000 claims abstract description 4
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 239000007788 liquid Substances 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 3
- 239000007787 solid Substances 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 2
- 238000013277 forecasting method Methods 0.000 abstract description 5
- 230000007246 mechanism Effects 0.000 abstract description 5
- 230000008859 change Effects 0.000 abstract description 3
- 238000002372 labelling Methods 0.000 abstract 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 13
- 238000010801 machine learning Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000033001 locomotion Effects 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 230000008014 freezing Effects 0.000 description 3
- 238000007710 freezing Methods 0.000 description 3
- 239000002245 particle Substances 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 238000009833 condensation Methods 0.000 description 2
- 230000005494 condensation Effects 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000000630 rising effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 241000581364 Clinitrachus argentatus Species 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 239000013078 crystal Substances 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000053 physical method Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 229920006395 saturated elastomer Polymers 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000026676 system process Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Primary Health Care (AREA)
- Educational Administration (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method for forecasting winter precipitation phase state by utilizing artificial intelligence, which comprises the steps of selecting profile data of 16 air pressure layers with atmospheric temperature, humidity, wind speed and vertical speed between 500-100hpa, and ground surface 2m air temperature, dew point temperature, ground surface air pressure and topography information as forecast characteristic variables based on a physical mechanism of precipitation phase state change; labeling 0-99 kinds of weather phenomenon observation data to form three kinds of rainfall phase state labels; training a forecasting model by utilizing a lightGBM framework and a training data set, and checking and optimizing parameters of the forecasting model by utilizing a verification data set and corresponding characteristic data to obtain a rainfall phase forecasting model which is stored as an MLPT; and obtaining the rainfall phase forecast of the site by using an MLPT algorithm and inputting forecast characteristic variable data. According to the invention, the characteristic variable at the current moment is utilized to intelligently forecast the phase state of the precipitation after 6 hours, and compared with a numerical forecasting mode and a precipitation phase state objective forecasting method of the same type of artificial intelligent algorithm, the forecasting accuracy and the forecasting timeliness are improved.
Description
Technical Field
The invention relates to the technical field of meteorological service, in particular to a method for forecasting winter precipitation phase state by utilizing artificial intelligence.
Background
Precipitation is one of the key links of earth water circulation and energy circulation (Zhang T J, 2005), and the precipitation phase state includes various forms such as rain, snow, sleet, freezing rain, ice particles, hail, and the like. The rain, the snow and the ice are in different phases of water, so the rainfall prediction and the rainfall prediction are collectively called as rainfall prediction in meteorology. Weather factors such as large-scale circulation background, water vapor condition, power condition and the like which are required to be paid attention to in rainfall prediction are also important to the rainfall prediction, while rainfall in winter is often accompanied with a rain and snow phase state conversion process, and accurate phase state prediction is the basis of the rainfall and snow depth prediction. However, phase prediction has been a difficulty in winter precipitation prediction. In addition, rain, snow, freezing rain, ice particles, rime and the like are also various in winter precipitation forms, and the transition between the phases has a very complex mechanism, and the core is the vertical change of the temperature and the water vapor in the atmosphere. In general, there are four requirements for snowfall: the temperature is low, the water vapor is saturated, the rising motion and the condensation nucleus exists in the air, and on the basis of achieving the four conditions, the rising motion is required to be extended to a temperature layer which is favorable for the development of ice crystals, and the temperature is usually at least to-10 ℃. In fact, the vertical distribution of the temperature and humidity in the atmosphere is very complex, the height of the layer from-10 ℃ to 0 ℃, the thickness difference between the related temperature layers, the temperature from the ground to the vapor lifting condensation height, the cloud bottom height and the like, and the cold air path, the strength, the moving speed, the topography and other underlying conditions are all key factors influencing the transformation of the precipitation phase, and the elements are mutually restricted and have a non-fixed change relation with the formation of the precipitation phase. That is, precipitation of different phases has a great influence on the surface material and energy circulation of the land (Wu B Y et al 2009), and the influence of the same precipitation amount on different precipitation phases is significantly different (Sun Yan et al, 2014; wangchun et al, 2005). The rainfall phase state forecasting quality is a core influencing factor of the rainfall forecasting accuracy, particularly in winter, the time and place of whether to rain or snow is a technological problem to be pre-judged in winter, three-dimensional weather comprehensive observation data, a corresponding numerical forecasting mode and the support of an objective forecasting method are needed, and the rainfall phase state forecasting quality is one of the biggest problems faced by the forecaster in winter rainfall forecasting.
The prediction of the winter precipitation phase state starts from seventies of the twentieth century, and the main methods of development include an empirical prediction method based on observation data, a numerical mode prediction based on a physical mechanism, a statistical prediction combining observation and a numerical mode, and the like, and can be mainly classified into three categories. The first type of method is to establish indexes and regression equations based on observation or numerical weather forecast, and is called an index criterion method for short; the second type of method is a micro-physical method and an aggregate forecasting method based on a numerical weather forecasting mode; the third type of method is an artificial intelligence forecasting method based on observation data and numerical weather forecast products, which applies decision trees, artificial neural networks, deep learning and the like.
At present, the ground precipitation phase mode forecast is mainly obtained by combining vertical distribution of the water content in the cloud with vertical profile diagnosis of the ambient temperature, the subjective phase forecast is also mainly judged based on temperature thresholds of key layers, however, the actual precipitation phase is related to a complex cloud micro-physical process, the quantitative simulation precision of the numerical mode on the cloud is not high, and the phase distribution described based on a temperature profile model or a threshold value has larger deviation from the phase characteristics in the actual cloud.
Statistical methods based on linear regression models are also applied to numerical model precipitation phase diagnostic predictions. Bocchieri (1979) proposed Model Output Statistics (MOS) techniques to predict the conditional probability of precipitation types. They used the pressure layer thickness, boundary layer bit temperature, specific temperature, dew point temperature and wet bulb temperature as predictors to establish a regression function with site precipitation type records. Keeter and Cline (1991) analyzed the relationship between 1000-700 hPa, 850-700 hPa, and 1000-850 hPa thickness and precipitation type using stepwise linear regression and provided additional objective precipitation type guidance for MOS. Bourgouin (2000) developed a so-called area method to diagnose the type of surface precipitation. The area of the process is defined as the area between the temperature profile of the melted (refrozen) layer and the isotherm at 0 c on the aerograph and is considered to be proportional to the residence time of the precipitation particles. The main disadvantage of these methods is the linear assumption, which in many cases is not applicable to precipitation, which forms a typical nonlinear complex system process.
In the current weather forecast service, the forecast of the winter precipitation phase is based on a numerical mode, and the forecasting personnel is used for carrying out artificial correction, so that the requirements of accuracy and timeliness cannot be met. The method is more suitable for machine learning methods for solving the high-dimensional nonlinearity problem, and is introduced into the forecasting work of the precipitation phase state in recent years, and artificial intelligence methods such as decision trees, logistic regression, support vector machines, deep neural networks and the like are used for exploring the forecasting of the precipitation phase state by domestic and foreign meteorological departments and scholars. Dong Quandeng (2013) develops a China area rain and snow phase objective forecasting model and a product based on an artificial neural network method on the basis of a numerical forecasting product, so that a northern rain and snow boundary line can be accurately forecasted; yang Lu and the like (2021) respectively establish a high-resolution objective classification model of a precipitation phase based on XGBoost, SVM, DNN three machine learning methods, and compare and test the forecasting effects of 3 Jinjin Ji main precipitation phases of rain, snow and snow under the same conditions by the three machine learning methods, so that the objective classification forecasting skill of the rain, snow and complex precipitation phases is further improved, but most methods are characterized in that the near-earth meteorological elements are the main, and the vertical structures of atmospheric heat and water vapor playing a key role in the precipitation phases cannot be well represented; in addition, some algorithms feature historical climate information that prevents applicability of their application due to data availability issues.
However, the current rainfall phase prediction method based on machine learning mainly uses the weather element prediction product (air temperature, dew point temperature and the like) of a numerical prediction mode and simultaneous secondary observation data to carry out training modeling, and then uses the prediction time secondary weather element product output by the mode to carry out rainfall phase classification prediction based on a classification model (Yang Lu and the like, 2021; dong Quandeng, 2013). In practical business application, because of a certain delay in the acquisition of numerical forecast data, the method has the problem of insufficient forecast timeliness, or only a forecast product with a longer report time can be used as a model input, but the forecast accuracy is also influenced.
Compared with the existing method, the prediction characteristic adopted by the method increases the vertical section of the variables such as the atmospheric temperature, the humidity, the wind speed, the vertical movement and the like on the basis of the common prediction variable of the ground surface, builds a physical mechanism of an implicit prediction model on the formation of the precipitation phase to develop high-dimensional approximation, uses the basic meteorological variable with high observation and simulation precision instead of the diagnostic index with complex error structure, and reduces the prediction error.
In addition, the accurate prediction of the precipitation phase state has great application value in many aspects of people's production and life, and snowfall, freezing rain and even rainy and snowy weather in winter bring great harm to urban traffic, power communication and people's production and life. The different precipitation phases cause a large difference in influence, for example: the snow, water and ice on the road surface are different in coping means. Furthermore, the precipitation phase is also closely related to the winter outdoor exercises. In conclusion, the continuous improvement of the forecasting capability of the phase state of the precipitation in winter is significant for improving the service quality of modern weather forecasting, guaranteeing the important activities of the country, preventing and reducing disasters, improving the urban management level and the like.
Disclosure of Invention
Aiming at the technical problems in the related art, the invention provides a method for forecasting the winter precipitation phase state by utilizing artificial intelligence, which can overcome the defects of the prior art method.
In order to achieve the technical purpose, the technical scheme of the invention is realized as follows:
a method for forecasting winter precipitation phase state by utilizing artificial intelligence, comprising the following steps:
s1, firstly, reading and preprocessing rainfall type observation data and site information, and carrying out rainfall phase state conversion according to coding rules recorded by the rainfall type;
s2, randomly dividing the preprocessed rainfall type observation data into two parts according to the proportion of 70 percent to 30 percent, wherein 70 percent is used as a training data set and 30 percent is used as a verification data set;
s3, according to the observation time in the observation data, selecting each air pressure layer data and near ground data corresponding to 24 hours before 6 hours, matching grid data in a certain range around each site, extracting variable profile data in a certain range around each site, and taking 2m air temperature, dew point temperature and surface air pressure as characteristic quantities, adding the topographic parameters of site information, calculating final characteristic quantities, and forming the final characteristic quantities into a characteristic data set;
s4, taking the training data set as a label, after acquiring characteristic data, training a forecast model by utilizing a lightGBM algorithm frame, and simultaneously, utilizing the verification data set and the corresponding characteristic data to carry out verification and parameter optimization on the forecast model, and finally obtaining and storing a rainfall phase forecast model;
and S5, finally, directly forecasting by utilizing a LightBGM algorithm frame according to the forecast model parameters and the input air pressure layer and near-ground characteristic data to obtain a rainfall phase forecast data result after 6 hours at the station.
Further, in step S1, the phase state conversion of precipitation is specifically three types of precipitation including liquid precipitation, solid precipitation and solid-liquid mixed precipitation.
Further, in step S3, the profile data of the temperature, humidity and wind speed around the site in a certain range is extracted, specifically, the profile data of each barometric layer with the atmospheric temperature, humidity, horizontal wind speed u, v components and vertical wind speed between 500-1000hpa is extracted.
Further, the grid data is obtained by calculating nearest grid points of the analysis data according to longitude and latitude information of the site, and then grid point data in a certain range around the nearest grid points are selected.
Further, the topographical parameters of the site information include longitude, latitude, and altitude parameters.
The invention has the beneficial effects that: by utilizing the numerical forecasting data at the current moment to intelligently forecast the phase state of the precipitation after 6 hours and generating a forecasting model, compared with the international optimal weather forecasting numerical mode (business products of the European middle weather forecasting center) and the same type of artificial intelligent algorithm-based objective forecasting method for the phase state of the precipitation, the forecasting accuracy and the calculating efficiency are improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic diagram of an example of precipitation in a rainy phase in a certain area according to a method for forecasting precipitation in winter by using artificial intelligence according to an embodiment of the invention.
Fig. 2 is a schematic diagram of an example of precipitation in a rainy and snowy phase in a certain area according to a method for forecasting precipitation in winter by using artificial intelligence according to an embodiment of the invention.
Fig. 3 is a schematic diagram of an example of precipitation of a snow phase in a certain area according to a method for predicting precipitation phase in winter by using artificial intelligence according to an embodiment of the invention.
Fig. 4 is a schematic diagram showing characteristic data selection of a method for forecasting winter precipitation phases by using artificial intelligence according to an embodiment of the invention.
Fig. 5 is a flowchart of a precipitation phase prediction method based on a lightGBM intelligent algorithm framework according to the method for predicting winter precipitation phase by using artificial intelligence according to an embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of the present invention, and the above technical solutions of the present invention will be described in detail below by way of specific usage modes for convenience of understanding the present invention.
As shown in fig. 5, the method for forecasting winter precipitation phase by using artificial intelligence according to the embodiment of the invention, namely, a machine learning precipitation phase forecasting algorithm, abbreviated as MLPT, includes firstly, reading and preprocessing precipitation type observation data and site information, and performing precipitation phase conversion according to coding rules recorded by the precipitation types; the phase state conversion of the precipitation is specifically converted into three types of liquid precipitation, mixed phase precipitation and solid precipitation. The pretreated precipitation type observation data are randomly divided into two parts according to the proportion of 70 percent to 30 percent, wherein 70 percent is used as a training data set and 30 percent is used as a verification data set.
As shown in fig. 4, since the hour-level observation data of the vertical profile of the temperature, humidity and wind speed of the site and the periphery are not available in general, the present invention adopts european center analysis ERA5 data, which is the same as the observation data, as the pseudo-true value for the feature engineering construction. And then according to the observation time in the observation data, selecting the air pressure layer data and the near ground data (for example, when the observation time is 2016, 01, 11 and 06, the analysis data from 2016, 01, 10 and 00 to 2016, 01, 11 and 00) corresponding to 24 hours before 6 hours, and performing geographic position matching, namely, calculating according to longitude and latitude information of the site to obtain the nearest neighbor grid point of the analysis data and selecting the grid point data of a surrounding 3X 3 range. And acquiring profile data of 16 layers of temperature, humidity, horizontal wind speed u and v components below 500hPa and vertical wind speed, near-ground 2m temperature, dew point temperature and surface air pressure data in the matched analysis extraction data as characteristic values, adding topographic parameters (geographic longitude and latitude and altitude) of site information, calculating out final 17931 characteristics, and forming the characteristic values into a characteristic data set.
And taking the observation data in the training data set as a label, carrying out rainfall phase prediction model training by utilizing a lightGBM algorithm frame together with the characteristic data set, and simultaneously carrying out inspection and parameter optimization on the prediction model by utilizing the verification data set and the characteristic data corresponding to the verification data set, so as to finally obtain and store the rainfall phase prediction model.
And (3) reading a randomly selected 5% test data set in the observed data, acquiring re-analyzed 3×3 grid data 24 hours before 6 hours corresponding to the time according to the observed time of the data, performing dimension reduction processing, adding site information (longitude, latitude and altitude), and simultaneously reading stored forecast model parameters. And forecasting the precipitation phase state of 32 sites in the Beijing area according to the forecasting model parameters and the input air pressure layer and near-ground data by utilizing the lightBGM.
Finally, as shown in fig. 1-3, the accuracy and the forecast score are calculated for the rainfall phase forecast result and the ERA5 rainfall phase simulation result respectively, and the result shows that the forecast accuracy of the invention reaches acc=0.87, the forecast score hss=0.77, which are superior to acc=0.80 and hss=0.62 of the ECMWF, and the acc=0.75 and hss=0.48 of the ERA 5. That is, the method is used for forecasting the winter precipitation phases of a plurality of sites in Beijing area, and compared with the forecasting result of the European mid-term weather forecast center (ECMWF, 0.8), the forecasting accuracy is improved by 7 percent (0.87).
In summary, by means of the above technical solution of the present invention, by generating the prediction model driven by data, the prediction model is more suitable for simulating the complex nonlinear problem of precipitation phase state than the conventional physical model, dynamic model and linear model; the vertical section of the variables such as atmospheric temperature, humidity, wind speed, vertical movement and the like is increased on the basis of the common forecasting variable of the ground surface through the adopted forecasting characteristics, a physical mechanism formed by an implicit forecasting model on a precipitation phase is constructed to develop high-dimensional approximation, and the forecasting error is reduced by using the basic meteorological variable with high observation and simulation precision instead of the diagnostic index with complex error structure; compared with the product of the middle-term weather forecast center in Europe, which is most widely applied in the world, the forecast result provided by the technical method is more accurate; by utilizing the numerical forecasting data at the current moment to intelligently forecast the phase state of the precipitation after 6 hours, compared with the same type of optimization method for the phase state forecasting result of the precipitation based on an artificial intelligence algorithm, the forecasting aging is improved.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (2)
1. A method for forecasting winter precipitation phase state by utilizing artificial intelligence, which is characterized by comprising the following steps:
s1, firstly, reading and preprocessing precipitation type observation data and site information, and carrying out precipitation phase state conversion according to coding rules recorded by the precipitation types;
s2, randomly dividing the preprocessed rainfall type observation data into two parts according to the proportion of 70 percent to 30 percent, wherein 70 percent is used as a training data set and 30 percent is used as a verification data set;
s3, according to the observation time in the observation data, selecting each air pressure layer data and near ground data corresponding to 24h before 6h, matching grid data in a certain range around each site, extracting variable profile data in a certain range around each site, taking 2m air temperature, dew point temperature and surface air pressure as characteristic quantities, adding topography parameters of site information, calculating final characteristic quantities, forming a characteristic data set from the final characteristic quantities, wherein the grid data are nearest grid points of analysis data obtained by calculating according to the longitude and latitude information of the site, selecting surrounding grid point data in a certain range, and the topography parameters of the site information comprise longitude, latitude and altitude parameters, namely calculating nearest grid points of the re-analysis data according to the longitude and latitude information of the site and selecting surrounding grid point data in a 3X 3 range;
in the step S3, the variable profile data of the temperature, the humidity and the wind speed of the periphery of the extraction station in a certain range is specifically the profile data of each air pressure layer with the atmospheric temperature, the humidity, the horizontal wind speed u and v components and the vertical wind speed between 500 and 1000 hpa;
s4, taking the training data set as a label, after acquiring characteristic data, training a forecast model by utilizing a lightGBM algorithm frame, and simultaneously, utilizing the verification data set and the corresponding characteristic data to carry out verification and parameter optimization on the forecast model, and finally obtaining and storing a rainfall phase forecast model;
and S5, finally, directly forecasting by utilizing a lightGBM algorithm frame according to the forecast model parameters and the input air pressure layer and near-ground characteristic data to obtain a rainfall phase forecast data result after 6 hours at the station.
2. The method for forecasting winter precipitation phases by utilizing artificial intelligence according to claim 1, wherein the precipitation phases in the step S1 are converted into three categories of liquid precipitation, solid precipitation and solid-liquid mixed precipitation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210158802.XA CN114626577B (en) | 2022-02-21 | 2022-02-21 | Method for forecasting winter precipitation phase state by utilizing artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210158802.XA CN114626577B (en) | 2022-02-21 | 2022-02-21 | Method for forecasting winter precipitation phase state by utilizing artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114626577A CN114626577A (en) | 2022-06-14 |
| CN114626577B true CN114626577B (en) | 2023-05-12 |
Family
ID=81900135
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210158802.XA Active CN114626577B (en) | 2022-02-21 | 2022-02-21 | Method for forecasting winter precipitation phase state by utilizing artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114626577B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113267834A (en) * | 2020-11-30 | 2021-08-17 | 武汉超碟科技有限公司 | Fusion rainfall forecasting method based on multi-model integration |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107991722A (en) * | 2017-12-25 | 2018-05-04 | 北京墨迹风云科技股份有限公司 | Method for building up, Forecasting Methodology and the prediction meanss of weather prediction model |
| CN112684520B (en) * | 2020-12-30 | 2022-12-13 | 北京墨迹风云科技股份有限公司 | Weather forecast correction method and device, computer equipment and storage medium |
| CN114022335B (en) * | 2022-01-10 | 2022-03-18 | 中航信移动科技有限公司 | Electronic identity authentication method and device based on dynamic timeliness and electronic equipment |
-
2022
- 2022-02-21 CN CN202210158802.XA patent/CN114626577B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113267834A (en) * | 2020-11-30 | 2021-08-17 | 武汉超碟科技有限公司 | Fusion rainfall forecasting method based on multi-model integration |
Non-Patent Citations (1)
| Title |
|---|
| 安俊岭 ; 陈勇 ; 屈玉 ; 陈琦 ; 庄炳亮 ; 张平文 ; 吴其重 ; 徐勤武 ; 曹乐 ; 姜海梅 ; 陈学舜 ; 郑捷 ; .全耦合空气质量预报模式***.地球科学进展.2018,(05),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114626577A (en) | 2022-06-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102076426B1 (en) | System for managing detailed weather prediction information in real time and method to use for impact forecasting of heat-wave and tropical night using the system | |
| CN108227041B (en) | Horizontal visibility forecasting method based on site measured data and mode result | |
| CN109958588B (en) | Icing prediction method, icing prediction device, storage medium, model generation method and model generation device | |
| CN109165693B (en) | Automatic identification method suitable for dew, frost and icing weather phenomena | |
| CN111665575B (en) | Medium-and-long-term rainfall grading coupling forecasting method and system based on statistical power | |
| KR101383617B1 (en) | Method and apparatus for predicting daily solar radiation level | |
| CN105760970A (en) | Method for predicting AQI | |
| CN105809287A (en) | High-voltage transmission line icing process integrated prediction method | |
| CN106803129A (en) | A kind of wind power ensemble prediction method based on multi-source numerical weather forecast | |
| CN106845080B (en) | Based on the modified Scene Tourist meteorological disaster intelligent Forecasting of difference | |
| CN102520464A (en) | Regional waterlogging forecasting system and forecasting method thereof | |
| CN113435630A (en) | Basin hydrological forecasting method and system with self-adaptive runoff yield mode | |
| CN114882373A (en) | Multi-feature fusion sandstorm prediction method based on deep neural network | |
| CN114386654A (en) | Multi-scale numerical weather forecasting mode fusion weather forecasting method and device | |
| CN114626577B (en) | Method for forecasting winter precipitation phase state by utilizing artificial intelligence | |
| CN111967661B (en) | Method, system, storage medium and equipment for deciding running direction of airport runway | |
| CN113075751A (en) | Method and system for fusing observation data in short-term forecasting | |
| CN111986464A (en) | System for forecasting road icing based on dynamic method | |
| CN115047544B (en) | Icing early warning method and system | |
| CN116663393A (en) | Random forest-based power distribution network continuous high-temperature fault risk level prediction method | |
| CN116679355A (en) | Precipitation prediction correction method based on cascade Attention-U-Net | |
| CN115902812A (en) | Method, system, equipment and terminal for automatically distinguishing short-term rainstorm weather background | |
| CN114966898A (en) | Early warning method and system for rainy and snowy weather in highway section | |
| CN114626578B (en) | Method for forecasting freezing rain by using artificial intelligence | |
| CN115511140A (en) | Intelligent lamp pole electric leakage risk prediction method, device and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |