CN110717611A

CN110717611A - Assimilation method for inverting sea fog humidity by meteorological satellite

Info

Publication number: CN110717611A
Application number: CN201910017023.6A
Authority: CN
Inventors: 高山红; 高小雨; 王永明
Original assignee: Ocean University of China
Current assignee: Ocean University of China
Priority date: 2019-01-08
Filing date: 2019-01-08
Publication date: 2020-01-21
Anticipated expiration: 2039-01-08
Also published as: CN110717611B

Abstract

The invention discloses a method for assimilating meteorological satellite inversion sea fog humidity, which comprises an all-weather sea fog monitoring module, a sea fog humidity construction module, a sea fog temperature construction module and a temperature and humidity data assimilation module; the all-weather sea fog monitoring module is used for monitoring the yellow sea fog in all weather based on MTSAT satellite data to obtain a three-dimensional distribution state of the sea fog; the sea fog humidity construction module and the sea fog temperature construction module are used for constructing humidity and temperature observation in the fog area based on sea fog information; the temperature and humidity data assimilation module utilizes cycling-3DVAR for assimilation. The method has the advantages that the forecasting initial field can be improved according to the inversion fog area and the change characteristics of the temperature and humidity structure in the sea fog development process, so that the short-time near forecasting result of the sea fog is further improved.

Description

Assimilation method for inverting sea fog humidity by meteorological satellite

Technical Field

The invention belongs to the technical field of short-time approach numerical prediction and data assimilation, and relates to an assimilation method for inverting sea fog humidity by using meteorological satellites MTSAT (Multi functional Transport satellites).

Background

Sea fog is a weather phenomenon that sea surface atmospheric level visibility is lower than 1km, which is generated in an offshore atmospheric boundary layer and is influenced by ocean action, and the sea fog has serious influence on offshore and coastal traffic and operation. The yellow sea is one of offshore sea fog high-incidence areas in China. In recent years, with the rapid increase of computing power, numerical simulation has become an advantageous tool for studying and forecasting sea fog. Sea fog numerical prediction, especially short-term approach prediction, is urgent in demand. It has been shown that the results of numerical simulations are very sensitive to errors in the initial field of the model. In order to improve the initial field quality, various researchers have made a lot of research work on data assimilation methods. The yellow sea is the sea area where sea fog occurs most frequently in offshore areas of China, and the important mechanism for generating the yellow sea is that warm and wet air advects to a cold sea surface so as to be cooled to saturation, and the process is influenced by an offshore atmospheric boundary layer. Therefore, the wind field, temperature and humidity structure of the offshore atmosphere in the initial field are of great importance to the sea fog simulation result. In order to improve the quality of an initial field of sea fog simulation, a cyclic three-dimensional variation (cyclic-3 DVAR) assimilation scheme is utilized to assimilate non-conventional marine observation data, particularly satellite radiation and inversion data to make up for the shortage of marine observation data, and the method is a feasible approach. The research results of assimilating satellite radiation data, satellite inversion sea surface wind data and radar radial wind data show that the temperature and the wind field structure in the sea fog simulation initial field can be improved to different degrees. However, the moisture conditions in the initial field did not improve significantly, which is often a significant cause of failure in sea fog forecasting. How to reasonably assimilate sea fog humidity information inverted by a meteorological satellite and improve the short-time near forecasting level of sea fog needs to develop an assimilation method.

Disclosure of Invention

The invention aims to provide a assimilation method for inverting sea fog humidity by a meteorological satellite. The method has the advantages that the forecasting initial field can be improved according to the inversion fog region and the change characteristics of the temperature and humidity structure in the sea fog development process, so that the forecasting result is improved. In the numerical forecast of 10 sea fog cases, the invention improves the just warning score (ETS) of the fog area by 16.9 percent compared with the ETS before improvement.

The technical scheme adopted by the invention comprises an all-weather sea fog monitoring module, a sea fog humidity construction module, a sea fog temperature construction module and a temperature and humidity data assimilation module; the all-weather sea fog monitoring module is used for monitoring the yellow sea fog in all weather based on MTSAT satellite data to obtain a three-dimensional distribution state of the sea fog; the sea fog humidity construction module and the sea fog temperature construction module are used for constructing humidity and temperature observation in the fog area based on sea fog information; the temperature and humidity data assimilation module utilizes cycling-3DVAR for assimilation.

Firstly, the all-weather sea fog monitoring module respectively inverts the three-dimensional structures of the sea fog at night and in the day based on MTSAT satellite data: at night, calculating the bright temperature difference (SLTD) between the short-wave channel IR4 wavelength of MTSAT and the long-wave channel IR1 wavelength of 3.5-4.0 μm, wherein the region satisfying-5.5K and SLTD and-2.5K is the fog region, and calculating the fog top height H by SLTD_n，H_n-212+191| SLTD × 0.5 |; calculating the difference ISTD, ISTD between IR4 bright temperature and SST of MTSAT in daytime<The 10K area is a high cloud and clear sky area, SLTD of a double-channel method is calculated, a fog area is preliminarily judged according to offshore visibility fitted by a dynamic SLTD threshold and a visible light albedo, the cloud area with rough texture is removed according to the uniformity of the visible light albedo to obtain a final fog area, the optical thickness delta is calculated according to the albedo and the solar zenith angle, and the fog top height H is calculated_dIn which H is_d＝45δ^2/3。

And then, diagnosing the three-dimensional sea fog body information by using a sea fog humidity construction module. Since the air in the sea fog is nearly saturated, the module assumes that the relative humidity in the inverted fog regions is 100%, and places 100% relative humidity on the pattern lattice points in each inverted fog region, thereby constructing a humidity profile family that can be identified by cycling-3DVAR, which is denoted as MTSAT-RH.

Further, temperature information in the three-dimensional sea fog body is constructed by utilizing a sea fog temperature construction module. The mixture ratio q of cloud water existing in the mode layer with the height less than or equal to 400m_cTaking an area with the fog top height of more than or equal to 0.016g/kg and less than or equal to 400M as a mode forecast fog area, comparing the forecast and inversion fog areas to obtain a forward report area, inverting the area with fog forecast in the fog area, marking as an H area, inverting the area without fog forecast in the fog area in a missed report area, marking as an M area and an area with no fog forecast in the false report area, marking as an F area, correcting the mode forecast temperature to construct a temperature profile, and marking as MTSAT-T: the forecast temperature of the M area is higher, and no neutral layer junction near the sea surface appears, so that the temperature profile of the adjacent H area is utilized to carry out the measurementCorrecting the air temperature at the pattern lattice point

Set as sea temperature

Plus the nearest H-zone gas-sea temperature difference SAT^H-SST^HAnd (3) vertical uniformity:

the forecast temperature of the F area is lower, so the temperature profile of the nearby forecast clear sky area is used for correcting the forecast temperature, and the air temperature on the pattern lattice point is corrected

Set as sea temperature

Plus the difference between the same-layer air temperature and the sea temperature in the nearest clear sky area

Finally, the temperature and humidity data assimilation module sets a period of time before a forecast starting point as an assimilation window through a circling-3 DVAR, wherein the assimilation window comprises a plurality of assimilation updating moments, a background field generates an initial field through multiple integration and assimilation, and the process is as follows: at the 1 st updating moment, interpolating global forecast or analysis data to lattice points of a WRF mode to serve as a background field; assimilating observation data by using 3DVar to obtain an analysis field; integrating the analysis field input mode to the next synchronization moment; assimilation and integration were repeated until the start of prediction.

Drawings

FIG. 1 is an operational flow of all-weather sea fog monitoring;

FIG. 2 is a looping-3 DVAR operational flow;

FIG. 3 is a flowchart of an operation of assimilating MTSAT-RH at the time of assimilation update;

FIG. 4 is a principal method of constructing a temperature profile within the sea fog region;

FIG. 5 is a flowchart of operations for assimilating MTSAT-RH and MTSAT-T at the time of assimilation update;

FIG. 6 is a pattern area setting with dots giving along-shore weather station locations;

FIG. 7 is a comparison of predicted fog regions, where behavior 1 observed fog region changes over time, and

rows

2, 3, and 4 are predicted fog regions for Exp-A, Exp-B, and Exp-C, respectively;

FIG. 8 is the 5 th step assimilation analysis increment for model bottom layers Exp-B (a) and Exp-C (b), with thick solid lines representing observed fog regions, dashed lines representing predicted fog regions, contour lines representing water vapor increment (in g/kg), and gray fill representing temperature increment (in K);

FIG. 9 is a comparison of the predicted 12h temperature (a) and steam mix ratio (b) against sounding observations for 10 cases, with the solid line representing the root mean square error RMSE and the dashed line representing the average deviation bias;

FIG. 10 is a graph of the predicted moisture versus observed bias at 850hPa to the surface layer for 10 cases of the initial field.

Detailed Description

The present invention will be described in detail with reference to the following embodiments.

Fig. 1 shows the operation flow of the all-weather sea fog monitoring module. The module respectively inverts the sea fog three-dimensional structures at night and in the day based on MTSAT satellite data: at night, calculating the bright temperature difference SLTD between MTSAT short-wave channel IR4 (with the wavelength of 3.5-4.0 μm) and long-wave channel IR1 (with the wavelength of 10.3-11.3 μm), wherein the region satisfying-5.5K and SLTD and-2.5K is the fog region, and calculating the height H of the fog top by using SLTD_n(H_n-212+191| SLTD × 0.5 |); during the daytime, the difference (ISTD) between the IR4 bright temperature and SST of MTSAT is calculated<The 10K area is a high cloud and clear sky area, the SLTD of a double-channel method is calculated, the fog area is preliminarily judged according to the offshore visibility fitted by a dynamic SLTD threshold value (which changes along with the solar altitude angle) and the visible light albedo, the cloud area with rough texture is eliminated according to the uniformity of the visible light albedo to obtain a final fog area, the optical thickness delta is calculated according to the albedo and the solar zenith angle, and the fog area is calculatedHeight of roof H_d(H_d＝45δ^2/3)。

The main idea of the sea fog humidity construction module is as follows: since the air in the sea fog is nearly saturated, the module assumes that the relative humidity in the inverted fog regions is 100%, and places 100% relative humidity on the pattern lattice points in each inverted fog region, thereby constructing a set of humidity profiles (denoted as MTSAT-RH).

Fig. 4 shows the main idea of the sea fog temperature building block. The mixture ratio q of cloud water existing in the mode layer with the height less than or equal to 400m_cAnd the area with the fog top height of not less than 400m and not less than 0.016g/kg is used as a mode forecast fog area. Comparing the forecast with the inversion fog area to obtain a positive report area (an area forecasting fog in the inversion fog area and recorded as an H area), a negative report area (an area forecasting fog-free in the inversion fog area and recorded as an M area) and a false report area (an area forecasting inversion fog-free in the fog area and recorded as an F area), and correcting the mode forecast temperature to construct a temperature profile (recorded as MTSAT-T): the forecast temperature of the M area is higher, and no neutral layer junction near the sea surface appears, so the temperature profile line of the adjacent H area is utilized to correct the forecast temperature, and the air temperature on the pattern lattice point is corrected

Set as sea temperature

Set as sea temperaturePlus the difference between the same-layer air temperature and the sea temperature in the nearest clear sky area

The operation flow of the temperature and humidity data assimilation module is shown in fig. 2 to 5. The ring-3 DVAR sets a period of time before the forecast starting point as an assimilation window, wherein the assimilation window comprises a plurality of assimilation updating moments, and a background field is subjected to multiple integration and assimilation to generate an initial field. The main process is as follows: at the 1 st updating moment, interpolating global forecast or analysis data to lattice points of a WRF mode to serve as a background field; assimilating observation data by using 3DVar to obtain an analysis field; integrating the analysis field input mode to the next synchronization moment; assimilation and integration were repeated until the start of prediction. The temperature and humidity profile assimilation module utilizes cycling-3DVAR to assimilate, each assimilation update is divided into two steps, the 1 st step assimilates other observation data (marked as obs), then MTSAT-RH and MTSAT-T are constructed by utilizing the first three modules, the height and air pressure data of an assimilation analysis field are combined, the analysis field is used as a background field, and the assimilation is carried out in the 2 nd step of updating.

The method is applied to the short-time approach prediction of the yellow sea fog, and the influence of the assimilation method on an initial field and a prediction result is analyzed. In 28 nightfall of 4 months in 2008, the upper layer of the south of the yellow sea is stable under the control of a marine high-pressure system, and small-range sea fog is generated. In the morning from 28 days at night to 29 days at night, the high-pressure system moves to the east, the sea fog in the air of the yellow sea rapidly develops to the northeast, and the whole yellow sea area is covered in the daytime of 29 days. For the sea fog example, an assimilation and numerical prediction test is carried out, simulation is carried out by adopting an ARW (Advanced Research core of WRF; V3.5.1) version of a mesoscale model WRF (weather Research and forecasting), and data assimilation is carried out by using an assimilation module WRFDA. The FNL data and SST of NEAR-GOOS provide initial field and boundary conditions for the mode. The mode uses the dual fixed nesting area of the lambertian projection (fig. 6), with the main parameter settings as in table 1. The prediction starting point is 12UTC on 28 days of 4 months, the assimilation window is 12h before the prediction starting point, and the assimilation updating is carried out every 3 h. Three experiments were set up, Exp-A, Exp-B and Exp-C: Exp-A utilizes circling-3 DVAR to only assimilate obs data (including conventional ground and sounding observation, aircraft newspaper, AIRS satellite inversion temperature and humidity profile and SSM/I inversion atmospheric water degradable amount); Exp-B utilizes an assimilation method of MTSAT inversion sea fog humidity to assimilate obs and MTSAT-RH; Exp-C utilizes MTSAT to invert an improved assimilation method of sea fog humidity, and assimilates obs, MTSAT-RH and MTSAT-T simultaneously (Table 2).

TABLE 1 WRF mode settings

*η＝1.0000，0.9975，0.9925，0.9850，0.9775，0.9700，0.9540，0.9340，0.9090，0.8800，0.8506，0.8212，0.7918，0.7625，0.7084，0.6573，0.6090，0.5634，0.5204，0.4798，0.4415，0.4055，0.3716，0.3397，0.3097，0.2815，0.2551，0.2303，0.2071，0.1854，0.1651，0.1461，0.1284，0.1118，0.0965，0.0822，0.0689，0.0566，0.0452，0.0346，0.0249，0.0159，0.0076，0.0000。

TABLE 2 assimilation test setup

Fig. 7 is a simulation result of 3 experiments. The predicted fog area of Exp-A is small, the speed of the Exp-A developing to the north is slow, and the difference from the real situation is large. The Exp-B fog region is larger, expanding to the north faster, but forecast in the southeast. The prediction result of Exp-C is further improved, the southeast part of the predicted fog region is quite consistent with satellite observation, and the development of a false fog region is almost avoided; on the other hand, the northern haze region of Exp-C is further expanded, closer to the observation fact.

Comparing the predicted fog region and the observed fog region within 24h time by time, calculating the fog region score (table 3), and finding that: Exp-A has a poor prediction effect on the cases, the POD, FAR and bias scores are 0.302, 0.746 and 0.492 respectively, and the ETS score is only 0.157; the ETS score of the fog region of Exp-B is 0.369, which is improved by 135 percent compared with that of Exp-A, POD and bias are respectively improved by 101 percent and 70.1 percent due to the enlargement of positive fog region, FAR is reduced to 0.511, which is improved by 92.5 percent; the Exp-C is improved to some extent compared with Exp-B, the ETS score reaches 0.499, and the FAR of Exp-C is improved by 23.5 percent compared with Exp-B due to the reduction of false fog areas.

TABLE 3 fog score for the test (number in brackets is percent improvement from the last test)

All three experiments are performed with 5 steps of assimilation, taking the 5 th step as an example, as shown in FIG. 8, Exp-B assimilates 100% of relative humidity in the observation fog region, and obvious water vapor change is generated. The water-vapor mixing ratio in the M area (near 31 degrees N and 123 degrees E) is increased by 1.0-2.5 g/kg. Exp-C makes a large correction to the temperature and humidity distribution of the background field. . Since the temperature profile is inverted based on the accuracy of the simulated fog region, the temperature changes in the M region (around 31 ℃ N, 123 ℃ E) and the F region (around 32 ℃ N, 126 ℃ E) are greatly different. At the bottom of the mode, the temperature in zone M is lowered and the temperature in zone F is raised. The change in moisture of Exp-C is not as drastic as that of Exp-B, and the mixing ratio is increased by about 1.7g/kg at most. This is because the decrease in the Exp-C temperature in the M zone decreases the saturation mixture ratio.

In order to further verify the effect and the precision of the invention, the data assimilation and simulation were performed on another 9 times of sea fog cases between 2007 and 2012 by using the same mode and assimilation setting. 3 groups of assimilation tests were performed for each example, and the data for assimilation were identified as Group-A, Group-B, and Group-C, respectively, in the same manner as for Exp-A, Exp-B, and Exp-C.

The Group-A has small predicted fog areas for some cases, and the assimilation of MTSAT-RH can greatly improve the sea fog prediction; in some examples, although Group-A has a larger bias, POD is smaller, which indicates that the area of the simulated fog region is larger, but the accuracy is not enough, in this case, Group-B can only reduce the M region by increasing the area of the fog region, and improve the ETS score, which sometimes causes the bias to become too large. In conclusion, Group-B improved the average ETS by 37.3%, but its predicted mist zone only increased and did not decrease, resulting in a systematic enlargement of the false mist zone with a bias of 1.315.

Group-C further improves the problem that the fog area of Group-B is larger, the correction of the temperature profile reduces the false fog area to a certain extent, bias is reduced to 1.264, ETS is improved by 16.9% compared with Group-B, and is improved by 60.6% compared with Group-A.

FIG. 9 shows a comparison of the 12h simulation results of three tests with the offshore meteorological station sounding data (located as in FIG. 6), with the Group-A and Group-B inshore RMSE around 1.4K, the Group-C inshore RMSE reduced to 1.1K, and the temperature bias reduced; compared with the water vapor of the Group-A, the Group-B has obvious improvement, the bias near the sea surface is increased from-0.4 g/kg to-0.2 g/kg, the RMSE is reduced by about 0.1g/kg, the Group-C has certain improvement compared with the Group-B, the RMSE below 950hPa is reduced by 0.18g/kg, and the bias is slightly increased compared with the Group-B. FIG. 10 shows the water vapor bias below 850hPa for each case, and it is clear that the water vapor bias decreases sequentially for Group-A, Group-B and Group-C, that Group-B assimilation MTSAT-RH primarily increases water vapor, and that Group-C assimilation with MTSAT-T improves the water vapor bias for Group-B (case-5 and case-8).

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not intended to limit the present invention in any way, and all simple modifications, equivalent variations and modifications made to the above embodiments according to the technical spirit of the present invention are within the scope of the present invention.

Claims

1. A assimilation method for inverting sea fog humidity by a meteorological satellite is characterized by comprising the following steps: the system comprises an all-weather sea fog monitoring module, a sea fog humidity construction module, a sea fog temperature construction module and a temperature and humidity data assimilation module; the all-weather sea fog monitoring module is used for monitoring the yellow sea fog in all weather based on MTSAT satellite data to obtain a three-dimensional distribution state of the sea fog; the sea fog humidity construction module and the sea fog temperature construction module are used for constructing humidity and temperature observation in the fog area based on sea fog information; the temperature and humidity data assimilation module utilizes cycling-3DVAR for assimilation.

2. The assimilation method for inverting the sea fog humidity by the meteorological satellite according to claim 1, is characterized in that: the all-weather sea fog monitoring module respectively inverts the sea fog three-dimensional structures at night and in the day based on MTSAT satellite data: at night, calculating the bright temperature difference SLTD of the MTSAT short wave channel IR4 wavelength of 3.5-4.0 μm and the long wave channel IR1 wavelength of 10.3-11.3 μm, wherein the region satisfying-5.5K and SLTD of-2.5K is the fog region, and calculating the fog top height H by using the SLTD_n，H_n-212+191| SLTD × 0.5 |; calculating the difference ISTD, ISTD between IR4 bright temperature and SST of MTSAT in daytime<The 10K area is a high cloud and clear sky area, SLTD of a double-channel method is calculated, a fog area is preliminarily judged according to offshore visibility fitted by a dynamic SLTD threshold and a visible light albedo, the cloud area with rough texture is removed according to the uniformity of the visible light albedo to obtain a final fog area, the optical thickness delta is calculated according to the albedo and the solar zenith angle, and the fog top height H is calculated_d，H_d＝45δ^2/3。

3. The assimilation method for inverting the sea fog humidity by the meteorological satellite according to claim 1, is characterized in that: the sea fog humidity building module is characterized in that the air in sea fog is close to saturation, the relative humidity in the inversion fog area is assumed to be 100%, and 100% of the relative humidity is placed on the pattern lattice point in each inversion fog area, so that a group of humidity profiles are built and are recorded as MTSAT-RH.

4. The assimilation method for inverting the sea fog humidity by the meteorological satellite according to claim 1, is characterized in that: the sea fog temperature construction module is used for constructing a mode layer with the height less than or equal to 400m and the cloud water mixing ratio q_cTaking an area with the fog top height of more than or equal to 0.016g/kg and less than or equal to 400M as a mode forecast fog area, comparing the forecast and inversion fog areas to obtain a forward report area, inverting the area with fog forecast in the fog area, marking as an H area, inverting the area without fog forecast in the fog area in a missed report area, marking as an M area and an area with no fog forecast in the false report area, marking as an F area, correcting the mode forecast temperature to construct a temperature profile, and marking as MTSAT-T: the forecast temperature of the M region is higher, and no neutrality near the sea surface appearsThe layer junction is corrected by using the nearby H-zone temperature profile to correct the air temperature at the lattice point of the pattern

Set as sea temperature

Set as sea temperature

5. The assimilation method for inverting the sea fog humidity by the meteorological satellite according to claim 1, is characterized in that: the temperature and humidity data assimilation module sets a period of time before a forecast starting point as an assimilation window through a circling-3 DVAR, wherein the assimilation window comprises a plurality of assimilation updating moments, a background field generates an initial field through multiple integration and assimilation, and the process is as follows: at the 1 st updating moment, interpolating global forecast or analysis data to lattice points of a WRF mode to serve as a background field; assimilating observation data by using 3DVar to obtain an analysis field; integrating the analysis field input mode to the next synchronization moment; assimilation and integration were repeated until the start of prediction.