Data assimilation method and system for inferring seawater quality changes using tide gauge records

Abstract

The present application provides a data assimilation method and system for predicting seawater quality change by using tide station records, and relates to the technical field of marine observation. The method comprises the following steps: selecting a tide station; calculating the global land mass migration change to obtain the sea level fingerprint effect; extracting the dynamic sea level change of the climate model and interpolating the dynamic sea level to the tide station record data; establishing an observation equation according to the selected tide station and using a sparse matrix to associate the state quantity and the observation quantity; introducing a random variable at each tide station to make up for the deficiency of the climate model in simulating the local dynamic sea level change; determining the initial time state of the tide station, optimizing and adjusting the constraint parameters of the state quantity and the correlation coefficient between adjacent tide stations; and based on Kalman filtering and smoothing, estimating the optimal state quantity and predicting the global seawater quality increase. The present application improves the accuracy of seawater quality estimation and realizes the coincidence and compatibility of model prediction and tide station observation data.

Description

Technical Field

[0001] This invention belongs to the field of marine observation technology, and in particular relates to a data assimilation method and system for inferring seawater quality changes using tide gauge records. Background Technology

[0002] The main causes of long-term sea-level changes include two parts: changes in seawater density caused by temperature and salinity variations, and increases in seawater mass caused by land mass migration. Tide gauges record long-term sea-level changes along the coast, providing valuable data. Climate models can simulate dynamic sea-level changes, primarily those caused by temperature and salinity variations and ocean circulation. Glacier models, hydrological models, and observational data can provide the contributions of different components of land mass change to sea-level rise, and the sea-level fingerprint effect can describe the spatial variation characteristics of these contributions in the marine region.

[0003] To investigate the contribution of seawater quality changes to sea-level rise in the 20th century, Hay et al. (2015) proposed a data assimilation method. The core idea is to assume that the main contributors to seawater quality increase are Greenland, West Antarctica, and global mountain glaciers. The melting of Greenland and West Antarctica is spatially homogenized. Then, the sea-level fingerprint theory is used to predict the spatial characteristics of their contributions to sea-level rise. Furthermore, it is assumed that the spatial characteristics of the sea-level fingerprint caused by global mountain glacier melting are homogenized. Using dynamic sea-level data from climate models, and with selected global tide gauges as observational constraints, the contributions of the three sources to seawater quality increase are estimated through Kalman filtering and smoothing. In addition to estimating the contribution of seawater quality, this method can also reconstruct the sea-level rise trend at tide gauge locations, supplementing missing records from these stations.

[0004] The inventors discovered that directly using land mass migration to study seawater mass increase faces several technical challenges, such as significant differences in error characteristics between different data sources and incompatibility between model outputs and observational data. Therefore, it is necessary to develop technical methods that ensure compatibility between the various components of land mass migration and tide gauge records. While existing sea level data assimilation techniques can reconstruct long-term sea level changes at tide gauge stations, estimating multiple mass contribution sources simultaneously faces significant uncertainties or requires strong prior constraints to achieve accurate quantitative estimation. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this invention provides a data assimilation method and system for inferring seawater quality changes using tide gauge records. This avoids the strong prior constraints in the prior art, reduces the number of parameters to be estimated, and introduces random variables to compensate for the insufficient simulation capabilities of climate models in local areas. Through these technical improvements, the accuracy of seawater quality estimation can be enhanced, achieving good agreement and compatibility between model predictions and tide gauge observation data.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0007] The first aspect of the present invention provides a data assimilation method for inferring changes in seawater quality using tide gauge records.

[0008] A data assimilation method for inferring changes in seawater quality using tide gauge records includes the following steps:

[0009] The data preprocessing includes: selecting tide gauge stations and acquiring their recorded data; calculating global land mass migration and change to obtain the sea level fingerprint effect; extracting dynamic sea level changes from climate models and interpolating the dynamic sea level data to the tide gauge station recorded data.

[0010] An observation equation is established based on the selected tide gauge stations, and a sparse matrix is ​​used to correlate state variables and observations. Random variables are introduced at each tide gauge station to compensate for the shortcomings of climate models in simulating local dynamic sea level changes. The initial state of the tide gauge stations is determined, and the constraint parameters of the state variables and the correlation coefficients between adjacent tide gauge stations are optimized and adjusted. Based on Kalman filtering and smoothing, the optimal state variables are estimated, and the increase in global seawater mass is inferred.

[0011] A second aspect of the present invention provides a data assimilation system for inferring changes in seawater quality by using tide gauge records.

[0012] A data assimilation system for inferring changes in seawater quality using tide gauge records includes:

[0013] The data preprocessing module is configured to: select tide gauge stations and acquire the data recorded by the tide gauge stations; calculate the global land mass migration and change to obtain the sea level fingerprint effect; extract the dynamic sea level change from the climate model and interpolate the dynamic sea level to the data recorded by the tide gauge stations;

[0014] The assimilation modeling module is configured to: establish observation equations based on the selected tide gauge stations, and use sparse matrices to correlate state variables and observations; introduce random variables at each tide gauge station to compensate for the shortcomings of climate models in simulating local dynamic sea-level changes; determine the initial state of the tide gauge stations, optimize and adjust the constraint parameters of the state variables and the correlation coefficients between adjacent tide gauge stations; and estimate the optimal state variables based on Kalman filtering and smoothing to infer the increase in global seawater mass.

[0015] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of the data assimilation method for inferring seawater quality changes using tide gauge records as described in the first aspect of the present invention.

[0016] The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the data assimilation method for inferring changes in seawater quality using tide gauge records as described in the first aspect of the present invention.

[0017] The above one or more technical solutions have the following beneficial effects:

[0018] This invention provides a data assimilation method and system for inferring seawater quality changes using tide gauge records. It constructs a new framework for sea level data assimilation theory. Based on tide gauge records, climate models, and sea level fingerprint effects, it introduces random variables to enhance the simulation capability of climate models in local areas, compensate for the shortcomings of climate models in local simulation capabilities, improve the estimation of global seawater quality increases, avoid strong prior constraints, and reduce the number of parameters to be estimated. Through these technical improvements, the accuracy of seawater quality estimation is enhanced, and the consistency and compatibility between model predictions and tide gauge observation data are achieved.

[0019] This invention assumes that land mass migration based on different data sources has the same error characteristics, and integrates the sea level fingerprint effect caused by different land mass migrations into a whole for estimation, thereby improving the estimation of land mass migration, enhancing the compatibility between different mass components, and ultimately achieving an accurate estimation of seawater mass increase.

[0020] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0022] Figure 1 This is a flowchart of the method in the first embodiment.

[0023] Figure 2(a) is a comparison of the sea level rise curves obtained by the method of the present invention and the existing technology.

[0024] Figure 2(b) shows the seawater quality and dynamic sea level curves obtained using the method of the present invention.

[0025] Figure 3 This is a comparison chart of seawater quality obtained using the method of the present invention and the total amount of sea level fingerprints input.

[0026] Figure 4 This is a long-term sea level trend map obtained using the method of the present invention. Detailed Implementation

[0027] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0028] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0029] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0030] Terminology Explanation:

[0031] GIA: Glacial isostatic adjustment

[0032] CMIP6: Coupled Model Intercomparison Project Phase 6 - The Sixth International Comparison of Coupled Models

[0033] SLF: Sea Level Fingerprint

[0034] Sterody: sterodynamic sea level

[0035] GMSL: Global Mean Sea Level

[0036] Example 1

[0037] This embodiment discloses a data assimilation method for inferring changes in seawater quality using tide gauge records.

[0038] like Figure 1 As shown, the data assimilation method for inferring seawater quality changes using tide gauge records includes the following steps:

[0039] The data preprocessing includes: selecting tide gauge stations and acquiring their recorded data; calculating global land mass migration and change to obtain the sea level fingerprint effect; extracting dynamic sea level changes from climate models and interpolating the dynamic sea level data to the tide gauge station recorded data.

[0040] An observation equation is established based on the selected tide gauge stations, and a sparse matrix is ​​used to correlate state variables and observations. Random variables are introduced at each tide gauge station to compensate for the shortcomings of climate models in simulating local dynamic sea level changes. The initial state of the tide gauge stations is determined, and the constraint parameters of the state variables and the correlation coefficients between adjacent tide gauge stations are optimized and adjusted. Based on Kalman filtering and smoothing, the optimal state variables are estimated, and the increase in global seawater mass is inferred.

[0041] This embodiment mainly includes three steps:

[0042] Step 1: Data preprocessing, mainly including tide gauge station selection, sea level fingerprint calculation, and dynamic sea level extraction. The specific implementation process is as follows:

[0043] (1) Select a suitable tide gauge station.

[0044] Although tide gauge stations are widely distributed along the coast of the world, there are still significant differences in daily observation and maintenance in different regions, resulting in obvious data gaps. In addition, tide gauge stations in some regions are also affected by local geological, geophysical, climatic and meteorological conditions, resulting in some skipped records. These skipped records are not signals of sea level change, or have little correlation with large-scale sea level change signals.

[0045] Therefore, tide gauge stations need to be carefully selected based on the research objectives. First, select tide gauge stations with more than 20 years of valid records since 1950. Then, conduct a detailed examination of the initially selected stations and delete any abnormal jumps in records. Furthermore, remove tide gauge stations with linear velocities exceeding 10 mm / yr, as these stations are highly likely to reflect local geophysical signals.

[0046] (2) Calculate the sea level fingerprint effect.

[0047] The study primarily collects components related to changes in land mass migration, including Greenland glacier melting, Antarctic ice sheet melting, mountain glacier melting, and changes in land water storage.

[0048] Based on image data, the spatiotemporal characteristics of Greenland's 20th-century melting were reconstructed, and the melting of Greenland's glaciers and ice sheets was quantitatively estimated.

[0049] Based on meteorological observations and ice sheet models, the melting of the Antarctic ice sheet is estimated.

[0050] Based on glacier models, simulate the ablation changes of global mountain glaciers in the 20th century;

[0051] By combining observational data and model-driven approaches, we can reconstruct changes in terrestrial water storage during the 20th century.

[0052] Using the above four components of change, we obtain the global land mass migration change. Based on the sea level fingerprint theory, under the elastic assumption of mass load, we consider the effects of the Earth's gravitational field, surface deformation and Earth's rotation to calculate the seawater mass redistribution characteristics formed by land mass change in the ocean region, i.e., the sea level fingerprint effect.

[0053] (3) Extract dynamic sea level changes from climate models.

[0054] The CMIP6 model provides dynamic sea-level changes on a global scale, but these outputs are based on regular grids or grids defined by climate models, not on dynamic sea-level changes at tide gauge stations. Therefore, two-dimensional interpolation techniques are needed to interpolate the dynamic sea-level changes output by the model to the tide gauge locations. Furthermore, the contribution of post-glacial rebound to the current relative sea level needs to be considered, and corrections should be made using the output of the GIA model.

[0055] Step 2: Construct a data assimilation framework, which includes five processes: establishing observation equations, introducing random variables, estimating initial states, optimizing parameters, and filtering and smoothing.

[0056] The mathematical structure and computational process of the data assimilation method will be described in detail below:

[0057] An observation equation was established based on tide gauge records at time node t:

[0058] Z t =H t X t +ε(S1)

[0059] In the formula, Z t It is the observation vector, which at time t contains m valid tide gauge records (m varies with time), ε is the noise of the tide gauge records, and matrix H t It is a transformation matrix that transforms the quantity to be estimated (i.e., the state variable) into the observation vector. It consists of two parts:

[0060]

[0061] In the formula, Sparse matrix: For each row, when the i-th tide gauge station has a record, then... The i-th element is 1, otherwise it is 0; The dimension is m×n, where n is the total number of tide gauge stations.

[0062] In data assimilation, the state vector X t Defined as:

[0063]

[0064] In the formula, This represents the sea level at time t for the i-th tide gauge station; similarly, α represents the random velocity of sea level at the tide gauge station. t The amplitude t represents the fingerprint at sea level.

[0065] During the filtering process, the state transition matrix Φ transforms the state variables to the next time step:

[0066]

[0067] In the formula, w represents process noise, subscript f represents prediction, and subscript a represents the analytical solution, i.e. the filtered result; Indicates the instantaneous rate of dynamic sea level; This indicates the relative sea-level change caused by the isostatic adjustment of glaciers; it is important to emphasize that... and In data assimilation, it is a driving factor, meaning that each climate model output can obtain a corresponding state quantity estimate (or complete a data assimilation calculation).

[0068] The structure of the transition matrix Φ is as follows:

[0069]

[0070] In the formula, y SLF It is a vector containing the rate change of the sea level fingerprint at the tide gauge station, i.e. It is important to note that This indicates a linear trend in sea-level fingerprints from 1950 to 2020, while Let represent the rate of the sea surface fingerprint at time t, and the relationship between the two is:

[0071]

[0072] In the formula, α represents the amplitude of the linear velocity of the sea-level fingerprint at time t.

[0073] Formulas S3-S6 can be used to represent the process of sea level change between two adjacent moments:

[0074]

[0075] Formulas S1 and S4 constitute the basic formulas for data assimilation:

[0076]

[0077] In the formula, R represents the observation noise, and matrix Q represents the covariance matrix of the state variables, which has the following form:

[0078]

[0079] In the formula, I (n+1)×(n+1) It is an identity matrix, σ 2 As a constraint parameter, it affects Given the range of α(t) as a function of time, this paper sets σ to 1 mm yr. -1 V TG The correlation between tide gauge stations is defined, which mainly depends on the distance between them:

[0080]

[0081] In the formula, The variance of the tide gauge stations after removing the linear trend; τ is taken as 500km; D ​​represents the distance between tide gauge stations. When D≤300km, the correlation is calculated using formula S10. If the distance D>300km, the correlation between tide gauge stations is no longer considered.

[0082] The specific formula for data assimilation is as follows:

[0083]

[0084] In the formula, It is the Kalman gain matrix, v t It is an update of the observations, with a variance of F. t .

[0085] The smoothing calculation process for data assimilation is as follows:

[0086]

[0087] In the formula, L t =Φ-K t H t ,and This represents the smoothed state quantity.

[0088] After making preliminary estimates of the state of seawater quality changes, random variables, and tide gauge reconstruction, the initial state of the tide gauge is determined to reduce the uncertainty of other state variables. The constraint parameters of the state variables and the correlation coefficients between adjacent tide gauges are further optimized and adjusted. Finally, Kalman filtering and smoothing are implemented to estimate the optimal state variables.

[0089] Step 3: Output and evaluate the causes of sea level rise, which mainly includes comparing and quantifying the contribution of seawater quality, analyzing the contribution of seawater specific volume, analyzing and reconstructing the sea level time series, explaining the causes of sea level rise, and uncertainty analysis.

[0090] The sea-level data assimilation framework can not only infer the increase in global sea-level mass but also reconstruct the long-term trend of sea-level change at tide gauge stations and estimate the contribution of dynamic sea-level change. The spatial characteristics of global sea-level mass change follow the spatial characteristics of the input sea-level fingerprint effect, with only a difference in amplitude between the two. The inferred sea-level mass change is compatible with tide gauge observations. Based on the dynamic sea-level change output by climate models, combined with the estimation of random variables, new dynamic sea-level changes at tide gauge stations will be obtained. If the inferred sea-level mass change and the newly estimated dynamic sea-level change can explain the reconstructed sea-level change, then the cause of sea-level rise is considered to have been successfully revealed. Since data assimilation considers the outputs of multiple climate models, time series driven by different models can be obtained, and the degree of dispersion between them can be estimated to reflect the uncertainty of the inferred results.

[0091] The experimental results obtained by this invention are described as follows:

[0092] As shown in Figure 2(a), a comparison is made between the global mean sea level rise curve reconstructed by the present invention (solid line) and the sea level rise curve reconstructed by the prior art (dashed line); as shown in Figure 2(b), the long-term trends of seawater mass (ocean mass, short dashed line), dynamic sea level (sterodynamic, long dashed line), and the sum of the two (Sum, long and short dashed lines) are the same as those in Figure 2(a).

[0093] like Figure 3 The figure shown is a comparison chart of the seawater mass predicted by this invention and the input seawater mass. The solid line represents the global seawater mass increase predicted by this invention; the dashed line represents the total sea level fingerprint (SLF total) input.

[0094] Figure 4 The following is a map showing the long-term trend of sea level (mm / yr):

[0095] Figure 4 (a) in the figure is the sea level rise trend at the tide gauge station reconstructed by the present invention; Figure 4 (b) in the figure represents the trend of the sum of seawater mass and seawater specific volume estimated by the present invention; Figure 4 In the equation (c), the difference in trends (ab) is represented.

[0096] Figures 2(a) and 2(b) present the main outputs of the data assimilation. For the reconstructed global mean sea level, the results of this invention are largely consistent with those in the references, and the inferred changes in seawater mass are also provided. The results show that the increase in seawater mass is the dominant factor in sea level rise, especially before 1980. The inferred changes in seawater mass and the newly estimated dynamic sea level well explain the reconstructed global mean sea level, providing an interpretive perspective for the study of the causes of sea level rise.

[0097] Compare the input seawater quality with the estimated seawater quality. Figure 3 The study found a significant difference between the two around 1980. Although the inferred seawater quality was constrained by tide gauge records within the framework of data assimilation theory, it was also driven by climate model outputs. Therefore, the accuracy of model outputs would significantly affect the inferred seawater quality changes.

[0098] In addition to the global average, the data assimilation results can also provide information on seawater quality changes at tide gauge stations and their contribution to sea-level rise. See [link to relevant documentation]. Figure 4 Among all the selected tide gauge stations, the sum of the inferred seawater mass and the newly estimated dynamic sea level matches the reconstructed sea-level rise trend, which is reflected in the difference in trends ( Figure 4 c). This result provides a new perspective for understanding the causes of local sea-level rise.

[0099] Example 2

[0100] This embodiment discloses a data assimilation system for inferring changes in seawater quality using tide gauge records.

[0101] A data assimilation system for inferring changes in seawater quality using tide gauge records includes:

[0102] The data preprocessing module is configured to: select tide gauge stations and acquire the data recorded by the tide gauge stations; calculate the global land mass migration and change to obtain the sea level fingerprint effect; extract the dynamic sea level change from the climate model and interpolate the dynamic sea level to the data recorded by the tide gauge stations;

[0103] The assimilation modeling module is configured to: establish observation equations based on the selected tide gauge stations, and use sparse matrices to correlate state variables and observations; introduce random variables at each tide gauge station to compensate for the shortcomings of climate models in simulating local dynamic sea-level changes; determine the initial state of the tide gauge stations, optimize and adjust the constraint parameters of the state variables and the correlation coefficients between adjacent tide gauge stations; and estimate the optimal state variables based on Kalman filtering and smoothing to infer the increase in global seawater mass.

[0104] In this embodiment, we used MATLAB to independently compile a data processing software system that can preprocess sea level observation data and model data, automatically connect to the data assimilation module, assimilate and absorb these data, run Kalman filtering and smoothing, and save the output variables.

[0105] Example 3

[0106] The purpose of this embodiment is to provide a computer-readable storage medium.

[0107] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the data assimilation method for inferring seawater quality changes using tide gauge records as described in Embodiment 1 of this disclosure.

[0108] Example 4

[0109] The purpose of this embodiment is to provide an electronic device.

[0110] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the data assimilation method for inferring seawater quality changes using tide gauge records as described in Embodiment 1 of this disclosure.

[0111] The steps and methods involved in the apparatuses of Embodiments 2, 3, and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0112] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computer device. The present invention is not limited to any particular combination of hardware and software.

[0113] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A data assimilation method for inferring seawater quality changes using tide gauge records, characterized in that, Includes the following steps: The data preprocessing includes: selecting tide gauge stations and acquiring their recorded data; calculating global land mass migration and change to obtain the sea level fingerprint effect; extracting dynamic sea level changes from climate models and interpolating the dynamic sea level data to the tide gauge station recorded data. An observation equation is established based on the selected tide gauge stations, and a sparse matrix is ​​used to correlate state variables and observations. Random variables are introduced at each tide gauge station to compensate for the shortcomings of climate models in simulating local dynamic sea level changes. The initial state of the tide gauge stations is determined, and the constraint parameters of the state variables and the correlation coefficients between adjacent tide gauge stations are optimized and adjusted. Based on Kalman filtering and smoothing, the optimal state variables are estimated, and the increase in global seawater mass is inferred. An observation equation was established based on tide gauge records, at the time nodes. : In the formula, It is the observation vector, in Time, which includes Records from one valid tide gauge station. It is noise recorded by the tide gauge station, matrix It is a transformation matrix; State vector Defined as: In the formula, Indicates the first Each tide gauge station at any moment sea ​​level; This represents the random rate of sea level at the tide gauge station; Indicates the amplitude of sea level fingerprints ; During the filtering process, the state transition matrix Transition the state variables to the next time step: In the formula, Indicates process noise, subscript Indicates a prediction, subscript This represents the analytical solution, i.e., the filtered result; Represents the instantaneous rate of dynamic sea level; This indicates the relative sea level change caused by the isostatic adjustment of glaciers; The process of sea level change between two adjacent moments is as follows: In the formula, The current sea level at the tide gauge station; The sea level at the tide gauge station at the next moment; The random velocity of the sea level at the tide gauge station; Indicates sea level fingerprints Rate of time: In the formula, The rate of change of the sea level fingerprint at the tide gauge station; express The amplitude of the linear velocity of the sea-level fingerprint at any given time.

2. The data assimilation method for inferring seawater quality changes using tide gauge records as described in claim 1, characterized in that, Select tide gauge stations with more than 20 years of valid records, delete abnormal jump records, and delete tide gauge stations with linear rates exceeding the set value.

3. The data assimilation method for inferring seawater quality changes using tide gauge records as described in claim 1, characterized in that, Assuming that land mass migration based on different data sources has the same error characteristics, the sea level fingerprint effect caused by different land mass migrations is fused into a whole for estimation.

4. The data assimilation method for inferring seawater quality changes using tide gauge records as described in claim 3, characterized in that, The process of obtaining the sea level fingerprint effect is as follows: Considering the melting of Greenland glaciers, Antarctic ice sheet, mountain glaciers, and changes in land water storage, the changes in global land mass migration are obtained; Under the elastic assumption of mass load, considering the effects of Earth's gravitational field, surface deformation, and Earth's rotation, the redistribution characteristics of seawater mass caused by land mass changes in the ocean region are calculated, and the sea level fingerprint effect is obtained.

5. The data assimilation method for inferring seawater quality changes using tide gauge records as described in claim 1, characterized in that: Process noise ,matrix Covariance matrix representing state variables: In the formula, It is a unit array. As a constraint parameter, it affects and The range that varies over time; The correlation between tide gauge stations was defined: In the formula, Variance of tide gauge stations after removing linear trends; The value is 500 km; Indicates the distance between tide gauge stations, when If the value is greater than the set value, the correlation between tide gauge stations will no longer be considered.

6. A data assimilation system for inferring seawater quality changes using tide gauge records, characterized in that: include: The data preprocessing module is configured to: select a tide gauge station and acquire the data recorded at the tide gauge station; Calculate global land mass migration changes to obtain the sea level fingerprint effect; Extract the dynamic sea level changes from the climate model and interpolate the dynamic sea level data to the tide gauge station's recorded data. The assimilation modeling module is configured to: establish observation equations based on the selected tide gauge stations and use sparse matrices to correlate state variables and observations; Random variables are introduced at each tide gauge station to compensate for the shortcomings of climate models in simulating local dynamic sea-level changes; Determine the initial state of the tide gauge station, and optimize and adjust the constraint parameters of the state variables and the correlation coefficients between adjacent tide gauge stations. Based on Kalman filtering and smoothing, the optimal state variables are estimated to predict the increase in global seawater mass. An observation equation was established based on tide gauge records, at the time nodes. : In the formula, It is the observation vector, in Time, which includes Records from one valid tide gauge station. It is noise recorded by the tide gauge station, matrix It is a transformation matrix; State vector Defined as: In the formula, Indicates the first Each tide gauge station at any moment sea ​​level; This represents the random rate of sea level at the tide gauge station; Indicates the amplitude of sea level fingerprints ; During the filtering process, the state transition matrix Transition the state variables to the next time step: In the formula, Indicates process noise, subscript Indicates a prediction, subscript This represents the analytical solution, i.e., the filtered result; Represents the instantaneous rate of dynamic sea level; This indicates the relative sea level change caused by the isostatic adjustment of glaciers; The process of sea level change between two adjacent moments is as follows: In the formula, The current sea level at the tide gauge station; The sea level at the tide gauge station at the next moment; The random velocity of the sea level at the tide gauge station; Indicates sea level fingerprints Rate of time: In the formula, The rate of change of the sea level fingerprint at the tide gauge station; express The amplitude of the linear velocity of the sea-level fingerprint at any given time.

7. A computer-readable storage medium having a program stored thereon, characterized in that, When executed by the processor, the program implements the steps in the data assimilation method for inferring changes in seawater quality using tide gauge records as described in any one of claims 1-5.

8. An electronic device, comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the data assimilation method for inferring seawater quality changes using tide gauge records as described in any one of claims 1-5.

Images