CN108319772A - A kind of analysis method again of wave long term data - Google Patents

Abstract

The invention relates to a method for reanalyzing long-term wave simulation data, comprising the following steps: constructing long-term observation data of wave elements (effective wave height and average period) based on a satellite altimeter and long-term simulation post-report data of wave elements based on a wave numerical model, and The time-space registration is carried out on the two; the time-space synchronous data is used to analyze and correct the wave post-report data to obtain the re-analysis data of the wave elements. The present invention proposes a new reanalysis method applied to long-term simulation post-report data of wave elements. Compared with commonly used data assimilation methods, the inventive method has high processing efficiency and flexible integrated observation data, and is especially suitable for long-term post-report of wave elements. The processing of simulated data can provide strong support for the research and analysis of climate variability, interannual variation of wave elements and marine engineering environment.

Description

A Reanalysis Method for Wave Long-Term Data

technical field

The invention relates to a data reanalysis method, in particular to a reanalysis method for long-term wave simulation data.

Background technique

Climate change has an important impact on the marine dynamic environment and the safety of human life and production. Waves are one of the main dynamic processes in the marine dynamic environment. Obtaining high-quality, high-resolution, and continuous wave elements (including effective wave height and average period) for a long-term Historical data is crucial to the study of long-term wave trends, interannual changes, and extreme value calculations of wave elements used in ocean engineering design. At present, the technical means of obtaining long-term wave data include fixed-point observation, satellite altimeter observation and wave numerical model. Fixed-point observations are mainly based on buoy observations, whose spatial distribution is extremely limited and not spatially representative; satellite altimeter observations have accumulated more than 30 years (1985-present) For observation at the next point, the temporal and spatial resolution of the data is very low, and for a certain spatial position, it is discontinuous in time. Numerical models of ocean waves can obtain long-term simulation data of wave elements with high resolution and continuous time and space, but the quality of the data is greatly affected by the reliability of the model's physical mechanism and the accuracy of wind and terrain fields.

The reanalysis method is to use the objective analysis method to analyze and correct the numerical simulation data by using the observation data; among them, the data assimilation method is the most commonly used method, which is based on the consideration of the temporal and spatial distribution of data and the errors of the observation field and the background field, and in the numerical model. A method for fusing new observational data during dynamic operation. Data assimilation is usually applied to the forecast of wave elements, which can provide the optimal initial conditions and model parameters for the forecast of the next period. However, it is expensive to process data and is not conducive to obtaining long-term posterior data of wave elements. At present, the reanalysis data of the wave elements of the European Weather Forecasting Center has been publicly released. It is mainly based on the wave numerical model to simulate the wave elements, and uses the four-dimensional variational assimilation method to correct the numerical model data with satellite altimeter data. It assimilates the 1991 6 altimeter data products out of 12 satellite altimeters launched since. Previous research reports have shown that the wave reanalysis data can well describe the climate variability of wave elements, but the effective wave heights given by it are generally low, especially in the case of high wind speed, which may be due to the limited satellite altimeter data used. Or the effectiveness of the assimilation method, but due to the technical limitations of the assimilation method, it is determined that it cannot be updated and optimized quickly using other altimeter observation data.

Therefore, it is necessary to develop a technical method for long-term wave simulation data, with higher analysis and processing efficiency and more flexible integration of observation data, so as to meet the requirements of climate variability, interannual change of wave elements, and ocean engineering environmental impact. data requirements.

Contents of the invention

Aiming at the deficiencies of the prior art, the purpose of the present invention is to provide a reanalysis method for long-term wave simulation data. Compared with the data assimilation method, which is mainly used for the simulation forecast of wave elements, this method is mainly used for the long-term simulation and forecast data of wave elements, which has the advantages of high analysis and processing efficiency and flexible integration of observation data.

In order to achieve the above-mentioned purpose of the invention, the present invention adopts the following technical solutions to achieve: a method for reanalyzing wave long-term simulation data, comprising the following steps:

1) Construct the observation data of wave elements based on satellite altimeter and the simulated data of wave elements based on wave numerical model, and perform temporal and spatial registration on the two to obtain the temporal and spatial synchronization data of wave elements;

2) The simulated data of the wave element is analyzed and corrected by using the space-time synchronous data to obtain the reanalysis data of the wave element.

Described step 1) comprises the following steps:

Obtain the wave element simulation data based on the wave numerical model;

Correct and grid the observation data of wave elements of each satellite altimeter;

Time-space registration and synchronization are performed on the wave element simulation data and the satellite altimeter wave element observation data to obtain time-space synchronization data.

Correcting and gridding the wave element observation data of each satellite altimeter includes the following steps:

Select the wave element observation data of a single satellite altimeter. For the wave element observation data of any buoy at a certain moment, the average value of the wave element observation data of the altimeter whose space interval is less than the set length and time interval is less than the set time is taken as the space-time Synchronous Data;

Linear fitting is performed on the time-space synchronization data of a single satellite; different satellite altimeters obtain different linear fitting relationships, that is, the fitting relationship between the wave element observation data observed by the buoy and the wave element observation data observed by the altimeter;

Apply the fitting relationship to correct multiple altimeters one by one to obtain the corrected altimeter wave element data;

Corresponding to the gridded simulation data, divide the altimeter data into different grids: for any observation position of the altimeter, select the grid point of the simulation data with the smallest spatial distance to it, and divide it into the grid point; At the same grid point, the altimeter wave element observation data whose time span does not exceed the set value is algebraically averaged to obtain the gridded data of the altimeter wave element at this moment.

The described wave element simulation data and satellite altimeter wave element observation data are carried out time-space registration and synchronization, and obtaining time-space synchronization data includes the following steps:

According to the gridded simulation data and the gridded data of the altimeter wave elements, the two are space-time registered: corresponding to the grid data of the altimeter wave elements at a certain moment, select the closest data point in the simulated data, Finally, the time-space synchronization data set of the wave elements of the two is obtained; each altimeter grid data point obtains a corresponding synchronization simulation data point; all data are combined to form time-space synchronization data.

Described step 2) comprises the following steps:

Based on the time-space synchronization data, the correction model is constructed by statistical analysis method: in the time-space synchronization data, for a single year and a subset of synchronization data B _jk under a certain grid point, j represents the year, k is the grid label, and the simulated data of the wave element is calculated Only when the correlation coefficient between the two is greater than the threshold value, the simulated data of the wave element is considered to be used for subsequent correction, otherwise, it is not used for correction and marked as bad point data; the simulated data used for correction is simulated Combined to obtain the correction relationship between the altimeter observed wave elements and wave element simulation data in B _jk ;

The simulated data is corrected according to the correction relationship to obtain the wave reanalysis data.

The present invention has the following beneficial effects and advantages:

1. The method of the present invention is applied to the long-term simulation post-report data of wave elements. The method is simple, fast and effective, and can provide necessary data support for research on climate variability, interannual variation and ocean engineering environmental impact of wave elements.

2. The method of the present invention integrates the observation data flexibly, and is convenient for continuously improving the reliability and precision of the wave element reanalysis data.

Description of drawings

Fig. 1 is the implementation flowchart of the reanalysis method that the present invention is applied to wave long-term post-report data;

Fig. 2a is the post-report data of wave significant wave height simulation based on multi-platform combined wind field (CCMP) forcing;

Figure 2b is the simulated aftercast data of wave significant wave height based on the wind field of the Climate Forecast Center (CFSR);

Figure 2c is based on the reanalysis data reported after the simulation of the significant wave height of CCMP waves;

Figure 2d is the reanalysis data based on the simulation of CFSR significant wave height;

Figure 2e is the reanalysis data of wave significant wave height based on data assimilation released by the European Forecasting Center.

Detailed ways

The technical solutions of the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

A re-analysis method for long-term wave simulation data, including the construction process of long-term observation data of wave elements (effective wave height and average period) based on satellite altimeter, the construction process of long-term simulation post-report data of wave elements based on wave numerical model, and through space-time registration Methods The process of obtaining the time-space synchronization data of the two wave elements, using the time-space synchronization data of the wave elements to analyze and construct the correction model process, using the correction model to correct and optimize the simulated post-report data of the wave elements, and obtaining the reanalysis data process of the wave elements. The invention is proposed aiming at the long-term simulation data of wave elements, which has high analysis and processing efficiency and flexible integration of wave element observation data, and is especially suitable for research on climate variability, interannual variation and marine engineering environment of wave elements.

As shown in Figure 1, a reanalysis method for long-term data of wave elements includes the following steps:

Construct the long-term observation data of wave elements based on the satellite altimeter and the long-term simulation post-report data of wave elements based on the wave numerical model, and perform temporal and spatial registration on the two to obtain the temporal and spatial synchronization data of the wave elements;

Using the space-time synchronous data to analyze and correct the aftercast data of the wave element, the reanalysis data of the wave element is obtained.

The long-term observation data of the wave element based on the satellite altimeter and the long-term follow-up data of the wave element based on the wave numerical model are constructed, and the time-space registration is performed on the two, and the time-space synchronization data of the wave element is obtained. The steps include the following steps:

Post-report of long-term data simulation of wave elements based on wave numerical model;

Multi-satellite altimeter wave element observation data correction and gridding;

Space-time registration and synchronous data acquisition of numerical simulation hindcast data of wave elements and satellite altimeter observation data.

Said utilizing the space-time synchronous data to analyze and correct the post-report data of the wave element, and obtain the reanalysis data of the wave element, comprising the following steps:

Based on the time-space synchronization data of wave elements, a correction model is constructed by analysis and fitting method;

The post-report data of the wave element is corrected by using the correction model to obtain the reanalysis data of the wave element.

The specific steps of this embodiment are as follows:

1. Long-term simulation and forecast simulation of wave elements (significant wave height and average period) based on the wave numerical model. The wave numerical mode can choose the more mature modes at present, such as Wavewatch III mode, SWAN mode, WAM mode and so on. This embodiment selects the Wavewatch III wave mode, which is released publicly and has a good performance in business applications. Driven by a wind field, the temporal and spatial resolution of the wind field is required to be relatively high. In this embodiment, the publicly released high-resolution wind field is selected: Multi-Platform Convergence Wind Field (CCMP) and the US Climate Forecast Center Wind Field (CFSR). The time resolution of the wind field is 6 hours, and the spatial resolution is 0.25°. In this example, these two wind fields are selected to drive the wave mode separately, and the simulation results of the latter report are obtained. The terrain background field of the model uses the publicly released bathymetric terrain data from the US Geophysical Center. In this embodiment, the backward forecast simulation of the wave element is carried out. The simulation time is from January 1, 1992 to December 31, 2010, and a total of 19 years of wave element backward forecast data are obtained. The simulated output time interval is 1 hour and the spatial resolution is 1.5°.

2. Multi-satellite altimeter wave element observation data correction and gridding.

(1) Collect observation data of wave elements (significant wave height and average period) of satellite altimeters that have been released publicly. Satellite sources include HY-2 satellite of my country Satellite Center, ERS-1, ERS-2 and Envisat satellites of ESA , TOPEX and Jason-1, Jason-2 satellites of the US Space Center. The altimeter adopts the polar orbit observation mode to observe the wave elements of sub-satellite points. The observation time interval between two adjacent sub-satellite points is about 1 second, and the space interval is about 6.8 kilometers.

(2) Although the wave element inversion accuracy of the satellite altimeter is high, due to the different internal settings of the instrument, there will be a certain systematic deviation in the data observation accuracy, which has a great impact on the study of the climate variability of the wave element, and it is necessary to combine the wave element The sea surface buoy observation data is corrected. The existing long-term buoy observation data mainly comes from the buoy observation data publicly released by the National Buoy Center of the United States. The selection standard of the buoy is to select an offshore distance greater than 50km, and there are 44 buoy stations that meet the conditions . A common correction method is used to correct multiple satellites one by one. Firstly, space-time registration: select the along-track observation data of wave elements from a single satellite altimeter, and perform space-time registration with the wave element observation data of sea surface buoys. The time-space registration window is selected as 50km and 30 minutes. The time-space registration method is: The wave element observation data of any buoy at a time, and the average value of the altimeter wave element observation data with a space interval of less than 50km and a time interval of less than 30 minutes are used as time-space synchronization data, so that the synchronization data at multiple times can be aggregated into synchronization data set. In order to distinguish it from the following, the synchronous data set is called A _i , where the subscript i represents different satellite altimeter numbers.

(3) Based on the time-space synchronization data set A _i of a single satellite, linear fitting is performed by the least square method: y=gx+h, where y is the effective wave height or average period data of the wave observed by the buoy, and x is the wave observed by the corresponding altimeter Significant wave height or average period data, g and h are undetermined fitting coefficients. In this way, different satellite altimeters can obtain different linear fitting relationships, and the fitting relationships between the significant wave height and the average period of the wave are also different. Multiple altimeters can be calibrated one by one by applying the fitting relationship: Substituting the effective wave height or average cycle data observed by the altimeter along the track into x, the obtained y is the corrected altimeter wave element data.

(4) Carry out grid processing on the corrected multi-satellite altimeter wave element observation data: corresponding to the grid information of wave numerical simulation, divide the altimeter data into different grids. Select the numerical model grid point with the smallest spatial distance to it, and divide it under this grid point. Algebraically average the along-track observation data of the altimeter wave elements with a time span of no more than 1 hour at the same grid point to obtain the gridded data of the altimeter wave elements at this moment. It should be noted that for the altimeter wave elements at the same grid point Feature observation data, which is discontinuous in time.

3. Time-space registration and synchronous data acquisition of the numerical simulation data of wave elements and satellite altimeter observation data.

According to the gridded data of wave elements based on the numerical model obtained in step 1 and the corrected satellite altimeter gridded data based on step 2, the space-time registration of the two is carried out: as described in step 2, the grid space of the two The information is the same, so only the time information needs to be registered here. The time information of the altimeter grid data is discontinuous, while the time information of the numerical post-report data is continuous (1 hour interval). Therefore, the altimeter corresponding to a certain moment Grid data, select the closest data point in the reported data after the numerical value, and finally get the space-time synchronization data group of the wave elements of the two; each altimeter grid data point will get the corresponding synchronized numerical simulation data point, all data Combined to form a data set, the synchronized data set is called B here.

4. Based on the space-time synchronization data set of the registered wave elements, the correction model is constructed through the analysis and fitting method;

According to the synchronous data set B obtained in step 3, analyze and build a correction model. In the study of climate change and engineering environment, the main focus is on the spatio-temporal change process of the annual average or annual extreme value, therefore, analysis and correction need to be carried out year by year and grid by grid. First of all, it is necessary to test and analyze the reliability of the model: in the synchronous data set B, for a single year and the synchronous data subset B _jk under a certain grid point (the subscript j indicates the year, and k is the grid label), the calculation is effective The correlation coefficient between the numerical simulation data of wave height or average period and the observation data of the altimeter, only when the correlation coefficient between the two is greater than 0.8, the numerical simulation data of the wave element is considered reliable and can be used for subsequent corrections, otherwise, it cannot be used for corrections, marked is the bad point data; through the numerical simulation data of the correlation test, the model relationship between the altimeter observation and the numerical simulation significant wave height or average period in B _jk is obtained by fitting: due to the error characteristics of the uncertain model data, a variety of Curve fitting methods, including polynomial fitting, exponential fitting, logarithmic fitting, etc., select the one with the smallest fitting error as the calibration model, here we take linear fitting as an example: y ₁ =px ₁ +q, where y ₁ is The significant wave height or average period observed by the altimeter, x ₁ is the significant wave height or average period of the numerical simulation, p and q are undetermined simulation coefficients, which can be obtained by the least square method.

5. Use the correction model to optimize the numerical simulation data of wave elements to obtain wave reanalysis data.

Through the correction model obtained in step 4, the wave forecast results are corrected year by year and grid by grid. Here, it is assumed that the optimal correction model is a linear relationship, that is, the example of step 4: y ₁ '=px ₁ +q, and the obtained Substituting the significant wave height or average period simulation data of the wave into x ₁ , the obtained y ₁ ' is the wave reanalysis data.

Next, the method of the present invention will be used to reanalyze and correct the post-report data of the long-term wave simulation, and check its actual effect in calculating the climate variability of significant wave height. Climate variability is defined as the trend in the annual mean significant wave height.

Figures 2a-2e show the climatic variability of the annual average significant wave height of China's coastal waters calculated from the long-term data of different wave significant wave heights, where "+" means passing the significance test. Figure 2a is the aftercast data of wave significant wave height simulation based on multi-platform combined wind field (CCMP); Figure 2d is the reanalysis data based on the CFSR wave significant wave height simulation forecast; Figure 2e is the wave significant wave height reanalysis data based on data assimilation released by the European Forecasting Center.

As shown in Fig. 2a-Fig. 2e, comparing Fig. 2a and Fig. 2b, it can be seen that there are obvious differences in the significant wave height climate variability of the wave simulation data driven by different wind fields. Using the reanalysis method proposed in this example to analyze After they are corrected separately, comparing Figure 2c and Figure 2d, it can be seen that their results are very similar, and they are also in good agreement with the calculation results of the wave reanalysis data based on the data assimilation method of the European Forecasting Center, thus confirming the Effectiveness of the present invention's reanalysis method for wave long-term simulation data.

In terms of computational efficiency, the processing time of single-machine single-core CPU in step 1 of this paper is about 10-15 days, and the processing time of other steps is about 1-2 days, while the data assimilation method takes several months to complete the reanalysis process , it can be seen that the method of the present invention greatly improves the efficiency of data reanalysis, and is especially suitable for the construction of long-term wave historical data required for studying climate impact.

In terms of data use, this example can flexibly use the altimeter observation data, for example: after obtaining the reanalysis data through steps 1-6 of this example, if you can obtain new altimeter wave element observation data, you can directly use step 1 on the basis of Only by repeating steps 2-5, the reanalysis process of wave element simulation data can be completed, new wave reanalysis data can be generated quickly and effectively, and the reliability of data can be improved.

Claims (5)

1. a reanalysis method of wave long-term simulation data, is characterized in that comprising the following steps: 1) Construct the observation data of wave elements based on satellite altimeter and the simulated data of wave elements based on wave numerical model, and perform temporal and spatial registration on the two to obtain the temporal and spatial synchronization data of wave elements; 2) The simulated data of the wave element is analyzed and corrected by using the space-time synchronous data to obtain the reanalysis data of the wave element. 2. the reanalysis method of a kind of wave long-term simulation data according to claim 1, is characterized in that described step 1) comprises the following steps: Obtain the wave element simulation data based on the wave numerical model; Correct and grid the observation data of wave elements of each satellite altimeter; Time-space registration and synchronization are performed on the wave element simulation data and the satellite altimeter wave element observation data to obtain time-space synchronization data. 3. the reanalysis method of a kind of wave long-term simulation data according to claim 2, it is characterized in that described correcting and gridding comprise the following steps to each satellite altimeter wave element observation data: Select the wave element observation data of a single satellite altimeter. For the wave element observation data of any buoy at a certain moment, the average value of the wave element observation data of the altimeter whose space interval is less than the set length and time interval is less than the set time is taken as the space-time Synchronous Data; Linear fitting is performed on the time-space synchronization data of a single satellite; different satellite altimeters obtain different linear fitting relationships, that is, the fitting relationship between the wave element observation data observed by the buoy and the wave element observation data observed by the altimeter; Apply the fitting relationship to correct multiple altimeters one by one to obtain the corrected altimeter wave element data; Corresponding to the gridded simulation data, divide the altimeter data into different grids: for any observation position of the altimeter, select the grid point of the simulation data with the smallest spatial distance to it, and divide it into the grid point; At the same grid point, the altimeter wave element observation data whose time span does not exceed the set value is algebraically averaged to obtain the gridded data of the altimeter wave element at this moment. 4. the reanalysis method of a kind of wave long-term simulation data according to claim 2, it is characterized in that described carry out space-time registration and synchronization to wave element simulation data and satellite altimeter wave element observation data, obtain time-space synchronous data comprising following step: According to the gridded simulation data and the gridded data of the altimeter wave elements, the two are space-time registered: corresponding to the grid data of the altimeter wave elements at a certain moment, select the closest data point in the simulated data, Finally, the time-space synchronization data set of the wave elements of the two is obtained; each altimeter grid data point obtains a corresponding synchronization simulation data point; all data are combined to form time-space synchronization data. 5. the reanalysis method of a kind of wave long-term simulation data according to claim 1, is characterized in that described step 2) comprises the following steps: Based on the time-space synchronization data, the correction model is constructed by statistical analysis method: in the time-space synchronization data, for a single year and a subset of synchronization data B _jk under a certain grid point, j represents the year, k is the grid label, and the simulated data of the wave element is calculated Only when the correlation coefficient between the two is greater than the threshold value, the simulated data of the wave element is considered to be used for subsequent correction, otherwise, it is not used for correction and marked as bad point data; the simulated data used for correction is simulated Combined to obtain the correction relationship between the altimeter observed wave elements and wave element simulation data in B _jk ; The simulated data is corrected according to the correction relationship to obtain the wave reanalysis data.