A method and system for improving air quality forecast effectiveness

Abstract

The application belongs to the technical field of air quality numerical prediction, and relates to a method and system for improving air quality prediction effect, the method comprising the following steps: determining weather types of each time period of each sub-region and each buffer region; determining an optimal parameterization scheme combination; determining future weather types; generating a series of initial value set prediction members; predicting future weather to obtain a future meteorological field and performing ensemble averaging on the future meteorological field to obtain an ensemble average prediction field; performing east-west fitting and north-south fitting to obtain a final future meteorological field of an inner layer nested region of each sub-region; splicing to obtain a future meteorological field of a large region; and obtaining a meteorological field of each nested region of a target region under an air quality mode. The method uses weather typing technology, ensemble prediction technology and multi-region simulation result coupling technology, and further improves the prediction effect of the air quality mode.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of air quality numerical prediction, and relates to a method and system for improving air quality prediction effect. BACKGROUND

[0002] With the development of numerical models, numerical model method has become an important method in pollutant concentration prediction. In actual business prediction, three generations of air quality models are mainly used to construct air quality prediction system. The mainstream three generations of air quality models at present mainly include multiscale air quality model MODEL-3 / CAMQ (Community Multiscale Air Quality Modeling System), atmospheric chemical transport model CAMx (Community Atmosphere Model), WRF-CHEM model and nested grid air quality prediction system (NAQPMS). The three generations of air quality models contain complex and perfect gas phase chemistry and photochemistry mechanism, and have good simulation and prediction ability for the temporal and spatial distribution of pollutants.

[0003] However, no matter which three generations of air quality models are used to construct the air quality prediction system, a complete air quality prediction system mainly includes three parts of emission source processing system (providing emission source input), meteorological model (providing meteorological field such as temperature, pressure, humidity and wind) and air quality model (simulating the temporal and spatial distribution of pollutants). As can be seen, the meteorological field is an important input item of the air quality prediction system, and has great influence on the prediction accuracy of the air quality prediction system. Therefore, the prediction accuracy of the meteorological model has great influence on the prediction accuracy of the air quality.

[0004] The meteorological model contains a large number of physical parameterization schemes. These parameterization schemes are mainly based on mathematical modeling to describe various physical processes in the atmosphere, for example, the WRF meteorological model contains more than 10 types of parameterization schemes such as cumulus convection parameterization scheme, boundary layer parameterization scheme, microphysical parameterization scheme and land surface process parameterization scheme. There are multiple schemes for each type of parameterization scheme, for example, there are more than 10 boundary layer schemes such as MYJ boundary layer parameterization scheme, MRF boundary layer parameterization scheme, ACM2 boundary layer parameterization scheme, QNSE boundary layer parameterization scheme and MYNN boundary layer parameterization scheme.

[0005] Meteorological model parameter adjustment is to select the most suitable parameterization scheme for the local area from various types of physical parameterization schemes according to the actual situation of each place, for example, to select a scheme that is best for local simulation and prediction from various boundary layer schemes, and to select a scheme that is best for local simulation and prediction from other types of schemes. However, long-term test results show that for a certain region, there is no scheme that is always better in terms of prediction effect.

[0006] At the same time, the meteorological field required by the meteorological department can be applied to the actual meteorological prediction system of the local area as long as the meteorological model is adjusted and has a good prediction effect on the local area. However, the transport of pollutants has a greater impact, for example, for a medium-sized city, the contribution of particulate matter transport is about 60% to 70%, and the contribution of ozone transport is about 70% to 80%. Therefore, the air quality model needs to consider the contribution of transport, which requires the meteorological model not only to have a good prediction effect on the local area, but also to have a good prediction effect on the surrounding area. If the prediction effect on the surrounding area is poor, the prediction error of the air quality model for the surrounding area will be large, the transport contribution error of the surrounding area to the local area will be large, and ultimately the prediction effect of the air quality model for the local area will be affected. However, test results show that there is no parameterization scheme that is good for all regions.

[0007] This brings challenges to the prediction of the meteorological field required by the air quality model, and the selected parameterization scheme needs to have good prediction effects for all regions at all times. This is difficult to achieve under current technical conditions and is an important factor affecting the prediction accuracy of current air quality models.

[0008] As can be seen from the foregoing, a complete air quality prediction system is composed of a meteorological model, an emission source processing system, and an air quality model. Therefore, the current improvement of air quality prediction effect mainly starts from the above three parts. Improving the emission source processing system mainly improves the accuracy of the time, space, and species allocation of the atmospheric pollution source emission inventory, and then improves the prediction effect of the air quality model. Improving the air quality model mainly improves the prediction effect of the air quality model through improving the parameterization scheme of the air quality model, the assimilation of the initial field, etc. Improving the meteorological model mainly improves the prediction effect by improving the meteorological background field required by the air quality model. It can be seen that improving the prediction effect of the meteorological model is an important direction for improving the prediction effect of the air quality model.

[0009] The existing technology mainly uses data assimilation, ensemble prediction, and improved parameterization scheme to improve the prediction accuracy of the meteorological field of the meteorological model.

[0010] Among them, data assimilation mainly assimilates monitoring data to the initial field to improve the accuracy of the initial field and thus improve the accuracy of meteorological prediction.

[0011] The ensemble prediction mainly considers that the meteorological model, the initial field and each parameterization scheme all have errors. When simulating and predicting, a plurality of meteorological models are selected to construct an ensemble prediction system, which is called a multi-model ensemble prediction system. Or, a plurality of initial value ensemble prediction members are constructed based on initial field perturbation when simulating and predicting, which is called an initial value ensemble prediction system. Or, a plurality of physical process ensemble members are constructed based on different parameterization schemes, which is called a physical process ensemble prediction system. Or, the uncertainties of the meteorological model, the initial value and the physical process are simultaneously considered to construct an ensemble prediction system that comprehensively considers the uncertainties of the meteorological model, the initial value and the physical process.

[0012] The ensemble prediction technology can improve the prediction effect, and there are related researches in air quality prediction. For example, the Chinese invention patent with the patent number ZL202110873280.7. The patent first obtains a meteorological field with relatively high accuracy through meteorological model ensemble prediction, and then uses the meteorological field to drive the air quality model to improve the prediction effect of the air quality model. The patent considers the influence of the prediction accuracy of the meteorological model on the air quality model, and solves the influence of the meteorological prediction with large error on the air quality prediction, which is of great significance to improve the prediction effect of the air quality model. However, the patent does not consider that the optimal parameterization scheme combination in different time periods has great difference. Although the meteorological prediction field obtained by the ensemble prediction method is better than the prediction effect of the single deterministic prediction, the prediction effect in each period is lower than that of the optimal ensemble prediction member. If the optimal parameterization scheme combination in different time periods can be obtained, the physical process ensemble member does not need to be constructed.

[0013] The improved parameterization scheme mainly improves the physical parameterization scheme in the model, so that it can be closer to the physical process in the actual atmosphere, and then improve the prediction accuracy.

[0014] These technologies improve the prediction accuracy of the meteorological field to a certain extent, and then improve the accuracy of the air quality prediction. However, they still have certain determinations, and the accuracy of the air quality prediction still needs to be improved.

[0015] However, the existing technology is difficult to simultaneously have good simulation and prediction effects in each period and each sub-region in the region, so it is difficult to meet the needs of air quality model prediction. SUMMARY

[0016] In view of the defects of the prior art, the application provides a method for improving air quality prediction effect, which uses weather typing technology to solve the problem of large simulation effect difference of the same parameterization scheme combination in each period, uses initial value set prediction technology to solve the initial field error problem, uses multi-region simulation result coupling technology to solve the problem of large simulation effect difference of the same parameterization scheme combination in each sub-region, and improves the prediction effect of the air quality model by improving the accuracy of the meteorological model prediction.

[0017] In order to achieve the above-mentioned purpose, the application provides the following technical scheme.

[0018] A method for improving air quality prediction effect, characterized in that it comprises the following steps:

[0019] 1) dividing a large region into a plurality of sub-regions and setting a buffer region between two adjacent sub-regions, identifying the weather type of each sub-region and each buffer region in each period based on weather typing technology and combined with historical meteorological data of many years;

[0020] 2) using a meteorological model to simulate the weather of many years, identifying the optimal parameterization scheme combination of each sub-region and each buffer region under different weather types;

[0021] 3) identifying the future weather type of each sub-region and each buffer region based on the driving field meteorological data of the meteorological model combined with weather typing technology;

[0022] 4) perturbing the initial field of each sub-region and each buffer region to generate a series of initial value set prediction members;

[0023] 5) using the optimal parameterization scheme combination of each sub-region and each buffer region under the weather type and combined with the initial value set prediction member to predict the future weather according to the future weather type of each sub-region and each buffer region, obtaining the future meteorological field of each sub-region and each buffer region and performing ensemble averaging to obtain an ensemble average prediction field;

[0024] 6) fitting the ensemble average prediction field of the inner nested region of each sub-region and each buffer region in the east-west direction and the north-south direction to obtain the final future meteorological field of the inner nested region of each sub-region;

[0025] 7) splicing the final future meteorological field of the inner nested region of each sub-region to obtain the future meteorological field of the large region;

[0026] 8) intercepting the large region according to the geographical range of each nested region of the target region under the air quality model, and obtaining the meteorological field of each nested region of the target region under the air quality model based on the future meteorological field of the intercepted large region.

[0027] Preferably, in the step 4), singular vector method or growing mode method is used to perturb the initial field of each sub-region and each buffer region.

[0028] Preferably, the east-west fitting in the step 6) specifically comprises:

[0029] 61) dividing the buffer region between each sub-region of the east-west division into a center zone and two edge zones, wherein the ensemble mean prediction field of the inner nested region of the buffer region is used as the new future weather field of the inner nested region of the center zone;

[0030] 62) using the fitting result of the ensemble mean prediction field of the inner nested region of the buffer region and the ensemble mean prediction field of the inner nested region of the sub-region having overlapping part with the edge zone as the new future weather field of the inner nested region of the edge zone;

[0031] 63) using the ensemble mean prediction field of the inner nested region of each sub-region as the new future weather field of the inner nested region of the part of each sub-region not overlapping with the buffer region;

[0032] 64) integrating the new future weather field of the inner nested region of the center zone, the new future weather field of the inner nested region of the edge zone and the new future weather field of the inner nested region of the part of each sub-region not overlapping with the buffer region to obtain the new future weather field of the inner nested region of each sub-region.

[0033] Preferably, in the step 62), when fitting the ensemble mean prediction field of the inner nested region of the buffer region and the ensemble mean prediction field of the inner nested region of the sub-region having overlapping part with the edge zone, cosine type probability density function is used for fitting.

[0034] Preferably, the cosine type probability density function is specifically:

[0035]

[0036] wherein, is the ensemble mean prediction field of the i-th grid from the center zone in a certain row grid of the inner nested region of the edge zone in the east-west direction, and n is the total number of grids of the row grid of the inner nested region of the edge zone in the east-west direction, is the ensemble mean prediction field of the corresponding grid of the inner nested region of the buffer region, is the ensemble mean prediction field of the corresponding grid of the inner nested region of the sub-region having overlapping part with the edge zone.

[0037] Preferably, the step 8) specifically comprises:

[0038] 81) intercepting the large region according to the geographical range of the inner layer nested region of the target region under the air quality mode, and taking the future meteorological field of the intercepted large region as the meteorological field of the inner layer nested region of the target region under the air quality mode;

[0039] 82) intercepting the large region based on the geographical range of the other layer nested region of the target region under the air quality mode, and processing the future meteorological field of the intercepted large region to obtain the meteorological field of the other layer nested region of the target region under the air quality mode.

[0040] Preferably, the processing of the future meteorological field of the intercepted large region to obtain the meteorological field of the other layer nested region of the target region under the air quality mode in step 82) specifically comprises:

[0041] 821) the zonal wind in the meteorological field of each grid of the other layer nested region is obtained by weighted averaging the zonal wind in the future meteorological field of the westmost column of grids in each grid of the large region contained by each grid of the other layer nested region;

[0042] 822) the meridional wind in the meteorological field of each grid of the other layer nested region is obtained by weighted averaging the meridional wind in the future meteorological field of the northmost column of grids in each grid of the large region contained by each grid of the other layer nested region;

[0043] 823) the prediction result of the remaining meteorological elements in the meteorological field of each grid of the other layer nested region is obtained by weighted averaging the prediction results of the corresponding meteorological elements in the future meteorological field of all grids in the large region contained by each grid of the other layer nested region.

[0044] In addition, the present application also provides a system for improving the effect of air quality prediction, characterized in that it comprises:

[0045] a weather type determination module for dividing the large region into a plurality of sub-regions and setting a buffer region between two adjacent sub-regions, identifying the weather type of each sub-region and each buffer region at each time period based on weather typing technology and combined with historical meteorological data of many years;

[0046] an optimal parameterization scheme combination determination module for simulating the weather of many years using a meteorological model, identifying the optimal parameterization scheme combination of each sub-region and each buffer region under different weather types;

[0047] a future weather type determination module for identifying the future weather type of each sub-region and each buffer region based on the driving field meteorological data of the meteorological model combined with weather typing technology;

[0048] a set of initial value ensemble members determining module, configured to disturb initial fields of each sub-region and each buffer region to generate a set of initial value ensemble members;

[0049] a set average prediction field determining module, configured to predict future weather according to future weather types of each sub-region and each buffer region, using optimal parameterization schemes of each sub-region and each buffer region under the weather types, combining the initial value ensemble members to obtain future meteorological fields of each sub-region and each buffer region and performing set average to obtain a set average prediction field;

[0050] a final future meteorological field determining module, configured to perform east-west fitting and north-south fitting on the set average prediction field of the inner nested region of each sub-region and each buffer region to obtain a final future meteorological field of the inner nested region of each sub-region;

[0051] a large region future meteorological field determining module, configured to splice the final future meteorological field of the inner nested region of each sub-region to obtain a future meteorological field of the large region;

[0052] a nested region meteorological field determining module, configured to intercept the large region according to geographical ranges of each nested region of a target region in the air quality model, and obtain meteorological fields of each nested region of the target region in the air quality model based on the future meteorological field of the intercepted large region.

[0053] Furthermore, the present application also provides an apparatus for improving air quality prediction effect, characterized by comprising:

[0054] one or more processors;

[0055] a memory for storing one or more programs;

[0056] when the one or more programs are executed by the one or more processors, the one or more processors implement the method for improving air quality prediction effect as described above.

[0057] Finally, the present application also provides a computer readable storage medium having a computer program stored thereon, characterized in that the program is executed by a processor to implement the steps of the method for improving air quality prediction effect as described above.

[0058] Compared with the prior art, the method and system for improving air quality prediction effect have one or more of the following beneficial technical effects:

[0059] 1. The application uses the initial value set prediction technology to reduce the influence of initial field error on the prediction result, uses the weather typing technology and the multi-region simulation result coupling technology, and uses different parameterization scheme combinations at different time periods and different regions for simulation, which can not only improve the meteorological field prediction accuracy of each period and each region, thereby improving the air quality mode prediction effect, but also save computing resources by reducing the number of meteorological model initial value set members.

[0060] 2. The application reduces the influence of the abnormal discontinuity problem of the adjacent sub-regional meteorological field caused by the use of different parameterization scheme combinations on the air quality mode simulation by setting the buffer area and fitting in the east-west and north-south directions.

[0061] 3. The application obtains the prediction result of the meteorological elements of the outer nested region through the processing of the corresponding grid of the inner nested region, and solves the meteorological field mismatching problem of the inner and outer nested regions.

[0062] 4. The application not only provides another idea for improving the air quality mode prediction effect, and can further improve the accuracy of air quality prediction, but also provides an idea for improving the meteorological model ensemble prediction technology. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 is a flow chart of the method for improving the air quality prediction effect of the application.

[0064] Figure 2 The schematic diagram of the large area division of the application is shown.

[0065] Figure 3 The nested schematic diagram of the sub-region of the application is shown.

[0066] Figure 4 The schematic diagram of the buffer area setting of the application is shown.

[0067] Figure 5 The schematic diagram of the east-west fitting of the application is shown.

[0068] Figure 6 The schematic diagram of determining the meridional wind of each grid of the outer nested region of the application is shown.

[0069] Figure 7 The schematic diagram of determining the zonal wind of each grid of the outer nested region of the application is shown.

[0070] Figure 8 The schematic diagram of determining the prediction result of other meteorological elements of each grid of the outer nested region of the application is shown.

[0071] Figure 9is a schematic diagram of the system for improving the air quality prediction effect of the present application.

[0072] wherein, Figures 6-8 In the figure, the left side is a schematic diagram of the grid in the large area, and the right side is a schematic diagram of the grid of the outer nesting area of the air quality model, one grid of the outer nesting area includes nine grids in the large area. DETAILED DESCRIPTION

[0073] Before any embodiments of the application are explained in detail, it is to be understood that the application is not limited in its application to the details of construction and the arrangement of components set forth in the following description or illustrated in the following drawings. The application is capable of other embodiments and of being practiced or being carried out in various ways. Also, it is to be understood that the phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting. The use of "including," "comprising," or "having" and variations thereof herein is meant to encompass the items listed thereafter and equivalents thereof as well as additional items. Unless specified or limited otherwise, the terms "mounted," "connected," "supported," and "coupled" and variations thereof are used broadly and encompass both direct and indirect mountings, connections, supports, and couplings. Further, "connected" and "coupled" are not restricted to physical or mechanical connections or couplings.

[0074] Also, in the disclosure of the present application, the terms "longitudinal", "transverse", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, which is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, therefore the above terms cannot be understood as limiting the present application; secondly, the term "one" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of one element can be one, and in another embodiment, the number of the element can be multiple, the term "one" cannot be understood as limiting the number.

[0075] In view of the defects of the prior art, the present application provides a method and system for improving the air quality prediction effect, which uses weather typing technology, ensemble prediction technology and multi-region simulation result coupling technology on the basis of fully considering the meteorological field characteristics required by the air quality model, solves the prediction error of the meteorological model in time and space, improves the simulation prediction effect of the meteorological model in a large area, and further improves the accuracy of the air quality model prediction.

[0076] Figure 1 A flowchart of the method for improving the air quality prediction effect of the present application is shown. As Figure 1As shown, the method for improving the effect of air quality prediction of the present application comprises the following steps:

[0077] I. Determine the weather type.

[0078] 1. Divide a large area into multiple sub-areas, and identify the weather type of each sub-area at each time period based on the weather typing technology combined with historical meteorological data, and arrange the time period of each sub-area according to the weather type.

[0079] In the present application, the specific implementation mode is introduced taking the case of dividing a large area (for example, a certain large area) into 4 sub-areas, and implementing two-fold nesting in each sub-area (the outer nesting area is grid-divided and its grid is 9 km, i.e., a square grid with a side length of 9 km, also referred to as a 9 km resolution grid; the inner nesting area is also grid-divided and its grid is 3 km, i.e., a square grid with a side length of 3 km, also referred to as a 3 km resolution grid; thus, one grid of the outer nesting area includes 9 grids of the inner nesting area). In actual application, multiple sub-areas (8 or 16) and multiple nesting (for example, three-fold nesting, the grid of the outer nesting area is 27 km, the grid of the intermediate nesting area is 9 km, and the grid of the inner nesting area is 3 km) can be set according to actual needs, and the specific processing mode is the same as that of 4 sub-areas and two-fold nesting.

[0080] For ease of description, as shown in Figure 2 The present application divides a large area into 4 sub-areas (northeast, northwest, southeast and southwest, respectively denoted as NE, NW, SE and SW). At the same time, as shown in Figure 3 Each sub-area includes an inner nesting area and an outer nesting area. And when setting, the innermost nesting area of each sub-area is tightly connected and does not overlap (as shown in Figure 2 ), and the outer nesting areas can overlap. Since the outer nesting area of the present application only provides initial values and boundary values for the inner nesting area, the data of the outer nesting area is not processed subsequently, for ease of display, Figure 2 the outer nesting area is hidden, and only the inner nesting area is displayed.

[0081] In the present application, the weather type of each time period of the 4 sub-areas is identified based on the weather typing technology combined with historical 3-year or multi-year meteorological data, and the historical time period of each sub-area is arranged according to the weather type. For example, which weather type does the northeast sub-area belong to from xx month xx day to xx month xx day, etc.

[0082] The weather typing technology mainly takes the temperature, humidity, isentropic potential height line (air weather chart) or isobar (ground weather chart), and wind direction and wind speed as the basis to identify the air weather system (trough line, ridge line, cutting edge line, cold vortex, low vortex, and other weather systems) and the ground weather system (front, back of the front, back of the high pressure, uniform field, terrain trough, inverted trough, North China small low pressure, and other weather systems), and identify the local weather situation. Certain weather situation will appear a certain type of weather, and the local weather can be predicted by analyzing the local weather situation, which is a common technology in artificial weather forecasting. The weather typing technology is based on AI technology to train a weather typing large model by analyzing a large number of historical weather systems, which belongs to the prior art.

[0083] 2. A buffer area is arranged between two adjacent sub-areas, and the weather type of each buffer area at each time period is identified based on the weather typing technology and historical meteorological data for many years, and the time period of each buffer area is sorted according to the weather type.

[0084] In the example of the present application, the weather type of each buffer area at each time period specifically includes:

[0085] 1) As shown in Figure 4 , a buffer area 1 (the buffer area 1 is also arranged as a double nested area) is arranged between the northeast and northwest sub-areas, which is denoted as HC1. The grid of the inner nested area of the buffer area 1 needs to be completely overlapped with the grid of the inner nested area of the northeast and northwest sub-areas (as shown in Figure 4 , the outer nested area of the present application only provides the initial value and the boundary value for the inner nested area, and the data of the outer nested area is not processed subsequently, and for the convenience of display, Figure 4 the outer nested area is hidden, and only the inner nested area is displayed). The weather of the buffer area 1 is typed based on the weather typing technology and historical meteorological data for 3 years or more, to obtain the weather type of the buffer area 1 at each time period, and the time period of the buffer area 1 is sorted according to the weather type.

[0086] 2) Similarly, a buffer area 2 is set between the southeast and southwest sub-areas, a buffer area 3 is set between the northeast and southeast sub-areas, and a buffer area 4 is set between the southwest and northwest sub-areas (the buffer area 2, the buffer area 3, and the buffer area 4 are also two-fold nested). It can be seen that the buffer area 1 and the buffer area 2 are buffer areas between sub-areas divided in the east-west direction, and the buffer area 3 and the buffer area 4 are buffer areas between sub-areas divided in the south-north direction. Based on the weather typing technology and historical meteorological data of three years or more, the buffer area 2, the buffer area 3, and the buffer area 4 are weather typed, and the time periods of the buffer area 2, the buffer area 3, and the buffer area 4 are sorted according to the weather type, according to the above step 1.

[0087] In the present application, the purpose of setting the buffer area is to establish a transition zone between two adjacent sub-areas, so that the meteorological elements evolve slowly and avoid abnormal discontinuity of the meteorological field at the junction of two adjacent sub-areas. For example, if the northeast sub-area and the northwest sub-area are directly spliced, and different weather types are used in the two sub-areas, the meteorological field at the junction area will have abnormal discontinuity, which may form a false convergence zone at the junction. Using this meteorological field to drive the air quality model will increase the simulation error.

[0088] II. Determine the optimal parameterization scheme combination.

[0089] 1. Simulate each time period of each sub-area using a meteorological model to identify the optimal parameterization scheme combination of each sub-area under different weather types.

[0090] In the present application, since the corresponding weather types of each sub-area in each time period are obtained by step 1, and there is generally a certain rule for what type of weather appears in a particular weather situation, the optimal parameterization scheme combination under a particular weather type is relatively fixed. Therefore, by simulating each time period of each sub-area using a meteorological model, the optimal parameterization scheme combination of each sub-area under different weather types can be identified.

[0091] When using the WRF meteorological model, the optimal parameterization scheme combination includes more than 10 types of parameterization schemes such as cumulus convection parameterization scheme, boundary layer parameterization scheme, microphysical parameterization scheme, and land surface process parameterization scheme. Each parameterization scheme includes the optimal parameterization scheme used.

[0092] 2. Simulate each time period of each buffer area using a meteorological model to identify the optimal parameterization scheme combination of each buffer area under different weather types.

[0093] Similarly, in the present application, since the weather type corresponding to each time period of each buffer area is obtained by step one, and generally there is a certain rule for what type of weather will appear under a specific weather situation, the optimal parameterization scheme combination under a specific weather type is relatively fixed, therefore, by using the meteorological model to simulate each time period of each buffer area, the optimal parameterization scheme combination of each buffer area under different weather types can be identified.

[0094] III. Determining the future weather type.

[0095] In the process of forecasting the future air quality, the future weather type of each sub-region and buffer area is first identified based on the driving field meteorological data of the meteorological model and the weather typing technology.

[0096] There are currently various existing meteorological model driving field data, and all current meteorological model driving field data meet the needs of weather typing, so the future weather type of each sub-region and buffer area can be identified by using the driving field meteorological data of the meteorological model and the weather typing technology.

[0097] It should be noted that in the present application, the air quality forecast is a short-term forecast (1-7 day forecast), and for short-term forecasting, the dominant weather type in the future is identified before each forecast, and then the optimal parameterization scheme combination is selected to construct the initial value set of forecast members for air quality forecasting in the future. Because it is a short-term forecast, the prediction time is relatively short, and by default, the future is dominated by one weather type. Therefore, the future weather type determined in this step is a weather type that plays a dominant role.

[0098] IV. Determining the initial value set of forecast members.

[0099] The initial field of each sub-region and each buffer area is disturbed to generate a series of initial value set of forecast members.

[0100] In the present application, various existing methods can be used, for example, singular vector method or growth module breeding method can be used to disturb the initial field of each sub-region and each buffer area to generate a series of initial value set of forecast members. This belongs to the prior art, and in order to simplify, it will not be described in detail here.

[0101] V. Determining the ensemble average forecast field.

[0102] 1. According to the future weather type of each sub-region and each buffer area determined in step III, the optimal parameterization scheme combination of each sub-region and each buffer area under the weather type determined in step II is used to predict the future weather in combination with each initial value set of forecast members determined in step IV, to obtain multiple future meteorological fields of each sub-region and each buffer area.

[0103] The present application adopts the method of ensemble prediction when determining the future meteorological field. The basic idea of ensemble prediction is to fully consider the uncertainty of the model and input field, but not to increase the false uncertainty to affect the prediction effect. Therefore, when performing ensemble prediction, various uncertainties should be fully considered, but the false uncertainty factors should also be controlled. The present application only performs initial value ensemble and does not perform physical process ensemble. Because the present application has screened the optimal parameterization scheme combination under the weather type based on weather typing, if the physical process ensemble prediction members are constructed by considering the uncertainty of the physical process, the false uncertainty will be increased due to the increase of the physical process ensemble members, which will affect the final ensemble prediction effect. Meanwhile, constructing the physical process ensemble prediction members will also increase the calculation amount and consume the calculation resources. For example, if there are 10 initial value ensemble prediction members and 10 physical process ensemble prediction members, 100 sets of ensemble prediction members will be obtained by randomly combining the physical process ensemble prediction members and the initial value ensemble prediction members, and the calculation amount will be greatly increased. However, the present application determines the weather type and selects the optimal parameterization scheme combination under the weather type through the weather typing technology, and then combines the 10 initial value ensemble prediction members, so that only 10 ensemble prediction members are needed, which can greatly reduce the number of ensemble prediction members, reduce the calculation amount and reduce the false uncertainty.

[0104] Since each sub-region and each buffer region are multi-nested (doubly nested in the example of the present application), by predicting the future weather, multiple future meteorological fields of the inner nested regions of each sub-region and each buffer region can be obtained, and multiple future meteorological fields of the outer nested regions of each sub-region and each buffer region can also be obtained. However, the present application only uses the future meteorological fields of the inner nested regions of each sub-region and each buffer region in subsequent processing.

[0105] 2. The multiple future meteorological fields of the inner nested regions of each sub-region and each buffer region are respectively ensemble averaged to obtain the ensemble average prediction fields of the inner nested regions of each sub-region and each buffer region.

[0106] The ensemble average is a numerical processing method of performing arithmetic average on multiple prediction results in ensemble prediction, which is used to reduce the uncertainty of single prediction and improve the prediction accuracy. This belongs to the prior art, and will not be described in detail here in order to simplify.

[0107] Six, determine the final future meteorological field.

[0108] 1. The ensemble average prediction fields of the inner nested regions of the buffer regions between the east-west divided sub-regions and the ensemble average prediction fields of the inner nested regions of the sub-regions are combined to perform east-west fitting to obtain new future meteorological fields of the inner nested regions of the sub-regions, and the specific implementation steps are as follows:

[0109] 1) The buffer region between the east-west divided sub-regions is divided into a center region and two edge regions. The ensemble average prediction field of the inner nesting region of the buffer region is used as the new future weather field of the inner nesting region of the center region.

[0110] For example, as shown in Figure 5 The buffer region 1 is divided into a center region C and two edge regions D and E. The center region C and the edge regions D and E are also double nested. The width of the center region C and the two edge regions D and E can each account for one third of the width of the buffer region 1. The ensemble average prediction field of the inner nesting region of the buffer region 1 is used as the new future weather field of the inner nesting region of the C region in the middle of the buffer region 1.

[0111] 2) The fitting result of the ensemble average prediction field of the inner nesting region of the buffer region and the ensemble average prediction field of the inner nesting region of the sub-region overlapping with the edge region is used as the new future weather field of the inner nesting region of the edge region.

[0112] For example, as shown in Figure 5 The regions outside the C region in the middle of the buffer region 1, that is, the D region overlapping with the northeast sub-region and the E region overlapping with the northwest sub-region, need to be fitted to solve the problem of abnormal discontinuity of the weather field in the east-west direction of the adjacent regions and improve the accuracy of the weather field.

[0113] There are two ensemble average prediction fields for each grid of the inner nesting region of the D region, that is, the prediction results of the meteorological elements, that is, the prediction results of the meteorological elements of the inner nesting region of the northeast sub-region and the prediction results of the meteorological elements of the inner nesting region of the buffer region 1. When fitting the inner nesting region of the D region, in order to ensure the continuity of the weather field, the prediction results of the meteorological elements (temperature, pressure, humidity, radial wind, latitudinal wind, etc.) of the inner nesting region of the buffer region 1 should be given greater weight for the grids of the inner nesting region of the D region close to the C region, and the prediction results of the meteorological elements of the inner nesting region of the northeast sub-region should be given greater weight for the grids of the inner nesting region close to the B region. From the C region to the B region, the weight of the prediction results of the meteorological elements of the inner nesting region of the buffer region 1 gradually decreases, and the weight of the prediction results of the meteorological elements of the inner nesting region of the northeast sub-region gradually increases. At the same time, in the fitting, not only the continuity of the weather field should be considered, but also the non-uniformity of the specific characteristic quantity, so the present application uses the cosine type probability density function which can represent the non-uniformity of the characteristic quantity for fitting.

[0114] The fitting formula is as follows:

[0115]

[0116] wherein, is the ensemble average forecast field of the i-th grid from the center region in a certain row of grids in the east-west direction of the inner nesting region of the edge region, n is the total number of grids in the row of grids in the east-west direction of the inner nesting region of the edge region, is the ensemble average forecast field of the corresponding grid of the inner nesting region of the buffer region, is the ensemble average forecast field of the corresponding grid of the inner nesting region of the sub-region overlapping with the edge region.

[0117] The fitting formula is used to fit all the grids in a certain row of grids in the east-west direction of the inner nesting region of the edge region, and the ensemble average forecast field of all the grids in the row of grids is obtained.

[0118] As can be seen from the fitting formula, the first grid of the inner nesting region of the left side of the D region close to the C region uses the prediction result of the meteorological element of the inner nesting region of the buffer region 1, and the first grid of the inner nesting region of the right side close to the B region uses the prediction result of the meteorological element of the inner nesting region of the northeast sub-region. From the C region to the B region, the weight of the prediction result of the meteorological element of the inner nesting region of the buffer region gradually decreases, the weight of the prediction result of the meteorological element of the inner nesting region of the northeast sub-region gradually increases, and the change of the weight is non-uniform.

[0119] The fitting formula is used to fit all the grids in each row of grids in the east-west direction of the inner nesting region of the edge region, and the new future meteorological field of the inner nesting region of the edge region after fitting is obtained.

[0120] In the same way, the prediction result of the meteorological element of the E region (i.e., the meteorological field) can be obtained.

[0121] 3) The ensemble average forecast field of the inner nesting region of each sub-region is used as the new future meteorological field of the inner nesting region of the part of each sub-region not overlapping with the buffer region.

[0122] For example, outside the buffer region 1, the B region of the northeast sub-region and the A region of the northwest sub-region use the prediction result of the meteorological element of the inner nesting region of the respective sub-region.

[0123] 4) The new future meteorological field of the inner nesting region of each sub-region is obtained by integrating the new future meteorological field of the inner nesting region of the center region, the new future meteorological field of the inner nesting region of the edge region, and the new future meteorological field of the inner nesting region of the part of each sub-region not overlapping with the buffer region.

[0124] That is, the prediction results of the meteorological elements of the inner nesting region of the C region, the prediction results of the meteorological elements of the inner nesting region of the D region, the prediction results of the meteorological elements of the inner nesting region of the E region, and the prediction results of the meteorological elements of the inner nesting region of the A region and the prediction results of the meteorological elements of the inner nesting region of the B region are integrated to obtain the prediction results of the meteorological elements of the inner nesting region of the new northeast and northwest sub-regions, i.e., the new future meteorological field of the inner nesting region of the northeast and northwest sub-regions.

[0125] Similarly, the inner nesting region of the buffer region 2 between the southwest and southeast sub-regions is fitted in the manner of steps 1) to 4) above to obtain the prediction results of the meteorological elements of the inner nesting region of the new southwest and southeast sub-regions, i.e., the new future meteorological field of the inner nesting region of the southwest and southeast sub-regions.

[0126] However, the new future meteorological field is only the east-west fitting of the inner nesting regions of the buffer regions of the northeast and northwest sub-regions and the buffer regions of the southeast and southwest sub-regions, and the prediction results of the meteorological elements in the north-south direction of the inner nesting regions of the northeast and southeast sub-regions and the northwest and southwest sub-regions can still differ greatly, causing discontinuity of the meteorological field, which needs to be fitted in the north-south direction.

[0127] The inner nesting regions of the buffer regions between the sub-regions divided in the north-south direction are combined with the set average prediction field, and the new future meteorological field of the inner nesting region of each sub-region is fitted in the north-south direction to obtain the final future meteorological field of the inner nesting region of each sub-region, i.e., the prediction results of the meteorological elements of each grid in the inner nesting region of each sub-region.

[0128] In the present application, the same as the east-west fitting method, the new future meteorological field of the inner nesting region of the buffer region 4 and the northwest and southwest sub-regions is fitted in the north-south direction, and the new future meteorological field of the inner nesting region of the buffer region 3 and the northeast and southeast sub-regions is fitted in the north-south direction to obtain the final future meteorological field of the inner nesting region of each sub-region, i.e., the prediction results of the meteorological elements of each grid in each sub-region.

[0129] Seven, determine the future meteorological field of the large region.

[0130] After obtaining the new future meteorological field of the inner nesting region of each sub-region, the new future meteorological field of the inner nesting region of each sub-region is spliced to obtain the future meteorological field of the large region.

[0131] In the present application, since the above processing is based on the inner nesting area, it can be seen that the future meteorological field of the large area is only the future meteorological field of the inner nesting area, i.e. the future meteorological field of the 3 km resolution grid.

[0132] VIII. Determining the meteorological field of each nesting area.

[0133] When performing air quality model simulation, it is necessary to first determine the area of air quality model simulation, i.e. the target area under the air quality model. For example, if air quality model simulation is performed for a certain place, the target area is the geographical range of the certain place and the surrounding area within a certain distance.

[0134] At the same time, air quality model simulation is generally multi-nested, for example, double nested. Among them, the outer nesting area (for example, a certain place) is larger and the grid of the outer nesting area is larger, for example, the grid of the outer nesting area is 9 km, but the resolution is lower, and the inner nesting area (for example, a certain area in a certain place) is smaller than the outer nesting area and the grid of the inner nesting area is smaller than the grid of the outer nesting area, for example, the grid of the inner nesting area is 3 km, but the resolution is higher. Of course, it can also be triple nested or other nested areas. Among them, in triple nested, the grid of the outer nesting area is 27 km, the grid of the middle nesting area is 9 km, and the grid of the inner nesting area is 3 km.

[0135] Through the above steps one to seven, the present application can obtain the meteorological field of the 3 km resolution grid of the large area (for example, a certain large area). In this way, when performing air quality model simulation, the target area under the air quality model (for example, a certain place or a certain city in the certain large area, etc.) can directly obtain the meteorological field of the 3 km resolution grid within its own range based on the meteorological field of the large area according to its own geographical range, thereby saving a lot of cost without the need for each target area to obtain the meteorological field of the 3 km resolution grid within its own range through the above steps one to seven.

[0136] In the present application, when performing air quality model simulation, the large area can be intercepted according to the geographical range of each nesting area of the target area under the air quality model, and the meteorological field of each nesting area of the target area under the air quality model can be obtained based on the future meteorological field of the intercepted large area, which specifically includes:

[0137] 1. Directly intercepting the large area based on the geographical range of the inner nesting area of the target area under the air quality model, and taking the future meteorological field of the intercepted large area as the meteorological field of the inner nesting area of the target area under the air quality model.

[0138] For example, the future meteorological field of each grid in the intercepted large region is directly taken from the large region according to the size of the inner layer nested region of the target region under the air quality mode, and the future meteorological field of each grid in the intercepted large region is taken as the meteorological field of each grid of the inner layer nested region of the target region under the air quality mode.

[0139] 2) The geographical range of the other nested region under the air quality mode, for example, the outer layer nested region, intercepts the large region, and the meteorological field of each grid of the other layer nested region under the air quality mode is obtained by processing the future meteorological field of each grid in the large region contained by each grid of the other nested region.

[0140] For example, in the present application, the grid of the outer layer nested region of the target region under the air quality mode is a 9 km resolution grid, and the grid in the large region is a 3 km resolution grid, so that one grid of the outer layer nested region of the target region under the air quality mode includes 9 grids in the large region, and the meteorological field of the grid of the outer layer nested region of the target region under the air quality mode is obtained by processing the future meteorological field of the 9 grids in the large region.

[0141] In the present application, the meteorological field of each grid of the other layer nested region of the target region under the air quality mode is obtained by processing the future meteorological field of each grid in the large region contained by each grid of the other layer nested region under the air quality mode, which specifically includes:

[0142] 1) The meridional wind in the meteorological field of each grid of the other layer nested region is obtained by weighted averaging the meridional wind in the future meteorological field of the most west column of grids in the large region contained by each grid of the other layer nested region.

[0143] For example, as shown in Figure 6 The meridional wind u in one grid 1 of the outer layer nested region is obtained by weighted averaging the meridional wind of the grid 1 in the large region, the grid 2 in the large region and the grid 3 in the large region in the most west column of grids contained by the grid 1 of the outer layer nested region. Figure 6

[0144] 2) The zonal wind in the meteorological field of each grid of the other layer nested region is obtained by weighted averaging the zonal wind in the future meteorological field of the most north column of grids in the large region contained by each grid of the gas layer nested region.

[0145] For example, as shown in Figure 7 ​As shown, the zonal wind v in grid 1 of the outer nested region is determined by the westernmost column of the nine grids within the large area encompassed by grid 1 of the outer nested region. Figure 7 The weighted average of the zonal winds of grid 1, grid 2, and grid 3 within the large region is obtained.

[0146] 3) The prediction results of the remaining meteorological elements in the meteorological field of each grid in the other nested regions are obtained by weighted averaging the prediction results of the corresponding meteorological elements in the future meteorological field of all grids in the large region contained by each grid in the outer nested region.

[0147] For example, such as Figure 8 As shown, the prediction results of meteorological elements other than meridional wind u and zonal wind v in grid 1 of the outer nested region are derived from the nine grids within the large area contained in grid 1 of the outer nested region, that is, Figure 8 The weighted average of the prediction results of the corresponding meteorological elements in grid 1, grid 2, grid 3, ..., grid 9 within the large area is obtained.

[0148] In general meteorological model simulations, the outer nested region (using a 9km resolution grid) is larger than the inner nested region (using a 3km resolution grid), providing initial and boundary values ​​to the inner nested region. Simultaneously, the inner nested region provides feedback to the outer nested region, ensuring that meteorological elements in the overlapping areas of the two nested regions match. However, this invention, when obtaining the prediction results of meteorological elements from the inner nested region's grid, reduces the impact of discontinuities in the meteorological field between adjacent sub-regions by setting a buffer and performing secondary fitting. This results in a deviation between the predicted meteorological elements of the overlapping area's grid and the original meteorological element prediction results. If a similar method is used to process and stitch together the outer nested region to obtain the prediction results of its meteorological elements, then the predicted meteorological elements of each grid in the overlapping or adjacent areas of the two sub-regions in the outer nested region will also deviate from the original meteorological elements. This leads to a mismatch between the predicted meteorological elements of the 9km resolution grid and the 3km resolution grid in the overlapping areas of the inner and outer nested regions, thus preventing the direct processing of the outer nested region. Therefore, the prediction results of meteorological elements in the inner and outer nested regions of this invention are all based on the prediction results of meteorological elements in the 3km resolution grid (i.e., the 3km resolution grid in the large area of ​​the meteorological model obtained in step seven), and can be arbitrarily extracted according to the needs of air quality model simulation.

[0149] Figure 9 A schematic diagram of the system for improving air quality forecasting according to the present invention is shown.Figure 9 The system for improving the effect of air quality prediction of the present application comprises:

[0150] 1. A weather type determination module.

[0151] The weather type determination module is used to divide a large area into a plurality of sub-areas and set buffer areas between two adjacent sub-areas, and identify the weather type of each sub-area and each buffer area at each time period based on weather typing technology and in combination with historical meteorological data.

[0152] 2. An optimal parameterization scheme combination determination module.

[0153] The optimal parameterization scheme combination determination module is used to simulate the weather of each year using a meteorological model, and identify the optimal parameterization scheme combination of each sub-area and each buffer area under different weather types.

[0154] 3. A future weather type determination module.

[0155] The future weather type determination module is used to identify the future weather type of each sub-area and each buffer area based on the driving field meteorological data of the meteorological model and in combination with weather typing technology.

[0156] 4. An initial value ensemble prediction member determination module.

[0157] The initial value ensemble prediction member determination module is used to perturb the initial field of each sub-area and each buffer area to generate a series of initial value ensemble prediction members.

[0158] 5. An ensemble average prediction field determination module.

[0159] The ensemble average prediction field determination module is used to predict the future weather of each sub-area and each buffer area using the optimal parameterization scheme combination of each sub-area and each buffer area under the weather type and in combination with the initial value ensemble prediction members according to the future weather type of each sub-area and each buffer area, obtain the future meteorological field of each sub-area and each buffer area, and perform ensemble averaging to obtain the ensemble average prediction field.

[0160] 6. A final future meteorological field determination module.

[0161] The final future meteorological field determination module is used to perform east-west fitting and north-south fitting on the ensemble average prediction field of the inner nested area of each sub-area and each buffer area to obtain the final future meteorological field of the inner nested area of each sub-area.

[0162] 7. A large-area future meteorological field determination module.

[0163] The large-area future meteorological field determination module is used to splice the final future meteorological field of the inner nested area of each sub-area to obtain the future meteorological field of the large area.

[0164] 8. the each nested area meteorological field determining module.

[0165] The each nested area meteorological field determining module is configured to intercept the large area according to the geographical range of each nested area of the target area in the air quality model, and obtain the meteorological field of each nested area of the target area in the air quality model based on the future meteorological field of the intercepted large area.

[0166] In addition, the present application also provides an apparatus for improving air quality prediction effect, comprising: one or more processors; a memory for storing one or more programs; and when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the method for improving air quality prediction effect as described above.

[0167] Finally, the present application also provides a computer readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for improving air quality prediction effect as described above.

[0168] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the protection scope of the present application. Based on the technical solutions of the present application, those skilled in the art can modify or equivalently replace the technical solutions of the present application without departing from the essence and scope of the technical solutions of the present application.

Claims

1. A method of improving the effect of air quality forecasts, characterized in that, The method comprises the following steps: 1) dividing a large area into multiple sub-areas and setting buffer areas between two adjacent sub-areas, identifying weather types of each sub-area and each buffer area in each time period based on a weather classification technology and in combination with historical meteorological data of multiple years; 2) simulating weather of multiple years using a meteorological model to identify optimal parameterization scheme combinations of each sub-area and each buffer area under different weather types; 3) identifying future weather types of each sub-area and each buffer area based on driving field meteorological data of the meteorological model in combination with the weather classification technology; 4) perturbing initial fields of each sub-area and each buffer area to generate a series of initial value set prediction members; 5) predicting future weather according to the future weather types of each sub-area and each buffer area, using the optimal parameterization scheme combinations of each sub-area and each buffer area under the weather types and in combination with the initial value set prediction members, obtaining future meteorological fields of each sub-area and each buffer area and performing ensemble averaging to obtain an ensemble average prediction field; 6) performing east-west fitting and south-north fitting on the ensemble average prediction field of the inner nested area of each sub-area and each buffer area to obtain a final future meteorological field of the inner nested area of each sub-area; 7) splicing the final future meteorological field of the inner nested area of each sub-area to obtain a future meteorological field of the large area; 8) cutting the large area according to the geographical ranges of each nested area of the target area under the air quality model, and obtaining meteorological fields of each nested area of the target area under the air quality model based on the future meteorological field of the large area.

2. The method of improving air quality forecast effectiveness of claim 1, wherein, In the step 4), the initial fields of each sub-area and each buffer area are perturbed using a singular vector method or a growing mode breeding method.

3. The method of improving air quality forecast effectiveness of claim 1, wherein, The east-west fitting in the step 6) specifically comprises: 61) dividing the buffer area between the east-west divided sub-areas into a center zone and two edge zones, wherein the ensemble average prediction field of the inner nested area of the buffer area is used as a new future meteorological field of the inner nested area of the center zone; 62) using a fitting result of the ensemble average prediction field of the inner nested area of the buffer area and the ensemble average prediction field of the inner nested area of the sub-area having an overlapping part with the edge zone as a new future meteorological field of the inner nested area of the edge zone; 63) using the ensemble average prediction field of the inner nested area of each sub-area as a new future meteorological field of the inner nested area of the part of each sub-area not overlapping with the buffer area; 64) integrating the new future meteorological field of the inner nested area of the center zone, the new future meteorological field of the inner nested area of the edge zone, and the new future meteorological field of the inner nested area of the part of each sub-area not overlapping with the buffer area to obtain a new future meteorological field of the inner nested area of each sub-area.

4. The method of improving air quality forecast effectiveness of claim 3, wherein, In the step 62), when fitting the ensemble average prediction field of the inner nested area of the buffer area and the ensemble average prediction field of the inner nested area of the sub-area having an overlapping part with the edge zone, a cosine-type probability density function is used for fitting.

5. The method of improving air quality forecast effectiveness of claim 4, wherein, The cosine-type probability density function is specifically: wherein, is the ensemble mean forecast field of the i-th grid from the center region in a certain row of grids in the east-west direction of the inner nesting domain of the edge region, and n is the total number of grids in the row of grids in the east-west direction of the inner nesting domain of the edge region, is the ensemble mean forecast field of the corresponding grid of the inner nesting domain of the buffer region, is the ensemble mean forecast field of the corresponding grid of the inner nesting domain of the sub-region that overlaps with the edge region.

6. The method of improving air quality forecast effectiveness of claim 1, wherein, The step 8) specifically comprises: 81) intercepting the large region according to the geographical range of the inner-layer nested region of the target region under the air quality mode, taking the future meteorological field of the intercepted large region as the meteorological field of the inner-layer nested region of the target region under the air quality mode; 82) intercepting the large region based on the geographical range of the other-layer nested region of the target region under the air quality mode, and processing the future meteorological field of the intercepted large region to obtain the meteorological field of the other-layer nested region of the target region under the air quality mode.

7. The method of improving air quality forecast effectiveness of claim 6, wherein, The step 82) of processing the future meteorological field of the intercepted large region to obtain the meteorological field of the other-layer nested region of the target region under the air quality mode specifically comprises: 821) the zonal wind in the meteorological field of each grid of the other-layer nested region is obtained by weighted average of the zonal wind in the future meteorological field of the westmost column of grids in each grid of the large region contained by each grid of the other-layer nested region; 822) the meridional wind in the meteorological field of each grid of the other-layer nested region is obtained by weighted average of the meridional wind in the future meteorological field of the northmost column of grids in each grid of the large region contained by each grid of the other-layer nested region; 823) the prediction result of the remaining meteorological elements in the meteorological field of each grid of the other-layer nested region is obtained by weighted average of the prediction results of the corresponding meteorological elements in the future meteorological field of all grids in the large region contained by each grid of the other-layer nested region.

8. A system for improving air quality forecast effectiveness, characterized by, Comprise: a weather type determination module for dividing a large region into a plurality of sub-regions and setting a buffer region between two adjacent sub-regions, identifying the weather type of each sub-region and each buffer region at each time period based on a weather typing technology and in combination with historical meteorological data of multiple years; an optimal parameterization scheme combination determination module for simulating the weather of historical years using a meteorological model, identifying the optimal parameterization scheme combination of each sub-region and each buffer region under different weather types; a future weather type determination module for identifying the future weather type of each sub-region and each buffer region based on the driving field meteorological data of the meteorological model and in combination with the weather typing technology; an initial value ensemble prediction member determination module for perturbing the initial field of each sub-region and each buffer region to generate a series of initial value ensemble prediction members; an ensemble average prediction field determination module for predicting the future weather using the optimal parameterization scheme combination of each sub-region and each buffer region under the weather type and in combination with the initial value ensemble prediction member according to the future weather type of each sub-region and each buffer region, obtaining the future meteorological field of each sub-region and each buffer region and performing ensemble average to obtain an ensemble average prediction field; a final future meteorological field determination module for performing east-west fitting and south-north fitting on the ensemble average prediction field of the inner-layer nested region of each sub-region and each buffer region to obtain the final future meteorological field of the inner-layer nested region of each sub-region; a large region future meteorological field determination module for splicing the final future meteorological field of the inner-layer nested region of each sub-region obtained to obtain the future meteorological field in the large region; Each nested region meteorological field determination module for cropping the large region according to a geographical extent of each nested region of the target region under the air quality model and obtaining a meteorological field of each nested region of the target region under the air quality model based on a future meteorological field of the cropped large region.

9. An apparatus for improving the effect of air quality forecasts, characterized by comprising: one or more processors; memory to store one or more programs; when the one or more programs are executed by the one or more processors, cause the one or more processors to implement the method for improving air quality forecasting effect according to any one of claims 1-7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, the program is executed by the processor to implement the steps of the method for improving air quality forecasting effect according to any one of claims 1-7.

Images