A passive tracer diffusion simulation method and system based on the CROCO model

Abstract

The application discloses a passive tracer diffusion simulation method and system based on a CROCO mode, and the method comprises the following steps: firstly, obtaining observation data of passive tracer content in a sampling grid of a target sea area; the passive tracer is a stable gas compound type; and then, performing passive tracer diffusion simulation by using the CROCO mode to obtain the concentration distribution of the passive tracer; wherein, the gas exchange flux of the passive tracer at the sea-air interface is considered in the convection diffusion equation under the CROCO mode. The application fully considers the sea surface gas exchange process, and can significantly improve the simulation accuracy of the tracer concentration distribution.

Description

Technical Field

[0001] This invention relates to the field of marine numerical simulation technology, and in particular to a passive tracer diffusion simulation method and system based on the CROCO model. Background Technology

[0002] Ocean numerical models are important tools for studying ocean dynamic processes, material transport, and marine ecosystems. The CROCO (Coastal and Regional Ocean Community) model, one of the major existing regional ocean numerical models, evolved from the ROMS_AGRIF (Regional Ocean Modeling System, Adaptive Grid Refinement InFortran) model. It employs a three-dimensional set of ocean dynamic equations, combined with topographically following coordinates (s-coordinates) and free surfaces, to simulate physical processes in the ocean such as current fields, temperature fields, and salinity fields. The CROCO model can perform high-resolution simulations of coastal and regional oceans, capturing fine-scale ocean phenomena such as mesoscale eddies, fronts, sub-mesoscale processes, and internal ocean waves.

[0003] Marine passive tracers are substances introduced into or naturally occurring in seawater whose movement characteristics are highly consistent with the movement of the surrounding seawater. They do not participate in chemical reactions or biological activities, do not significantly alter the physical or chemical state of seawater, and simply move with the seawater. By monitoring parameters such as their diffusion and concentration changes, phenomena such as ocean circulation, mixing processes, and pollutant migration can be indirectly inferred. Passive tracers mainly include radioactive isotopes (such as tritium, carbon-14, and radium-226) and stable gaseous compounds (such as chlorofluorocarbons (CFCs) and sulfur hexafluoride (SF6).

[0004] The CROCO ocean model typically uses the PASSIVE_TRACER tool to simulate the transport and diffusion of passive tracers in the ocean. When simulating passive tracer diffusion, it assumes that tracers are only affected by physical transport processes in the ocean, neglecting surface material exchange. However, stable gaseous passive tracers may exhibit concentration changes at the air-sea interface due to gas exchange processes. Existing CROCO models ignore this process when simulating tracer diffusion, leading to a significant overestimation of tracer concentrations in the upper waters, especially with substantial error accumulation over long simulations, resulting in a large discrepancy between simulation results and actual observations. Summary of the Invention

[0005] The purpose of this invention is to provide a passive tracer diffusion simulation method and system based on the CROCO model, which takes into account the material exchange process at the air-sea interface and solves the problem of significantly overestimating the concentration of passive tracers in simulation results.

[0006] The passive tracer diffusion simulation method based on the CROCO mode comprises the following steps:

[0007] Obtaining observation data of passive tracer content in a sampling grid of a target sea area; the passive tracer is a stable gas compound type;

[0008] Simulating passive tracer diffusion by using the CROCO mode to obtain the concentration distribution of the passive tracer; wherein the gas exchange flux of the passive tracer at the sea-air interface is considered in the convection diffusion equation under the CROCO mode.

[0009] Further, the convection diffusion equation is:

[0010] ;

[0011] wherein, is a surface source and sink term, , is a gas exchange flux, is a Dirac function, is a sea surface height, is a passive tracer concentration, , and are flow velocities in , and directions, is a eddy diffusivity, is a molecular diffusivity, is a vertical diffusion term, is a source and sink term, is a dissipation term.

[0012] Further, the gas exchange flux ;

[0013] wherein, is a gas transfer velocity, is a surface ocean gas concentration, is an atmospheric gas concentration, is a dimensionless solubility coefficient; it is set according to experience and .

[0014] Further, the gas transfer velocity is set to 10 cm / hour, and the atmospheric gas concentration is set to 0.

[0015] Further, the method further comprises: processing the observation data into grid data, interpolating the concentration distribution of the passive tracer to the spatial positions of the grid data, calculating the average relative error between the observation values in the grid data and the simulation values in the concentration distribution;

[0016] If the average relative error exceeds its corresponding threshold value, the passive tracer diffusion simulation is re-performed after modifying the numerical value of and / or .

[0017] The passive tracer diffusion simulation system based on the CROCO model comprises:

[0018] A data acquisition unit is configured to acquire observation data of passive tracer content in a sampling grid of a target sea area; the passive tracer is a stable gas compound type.

[0019] An simulation unit is configured to perform passive tracer diffusion simulation by using the CROCO model to obtain a concentration distribution of the passive tracer; in the convection-diffusion equation under the CROCO model, the gas exchange flux of the passive tracer at the sea-air interface is considered.

[0020] The passive tracer diffusion simulation method based on the CROCO model comprises the following steps:

[0021] Acquiring observation data of passive tracer content in a sampling grid of a target sea area; the passive tracer is a stable gas compound type.

[0022] Performing passive tracer diffusion simulation by using the CROCO model to obtain a concentration distribution of the passive tracer.

[0023] The CROCO model comprises an analytical field calculation module, a boundary condition processing module and a three-dimensional advection term time integral solving module; performing passive tracer diffusion simulation by using the CROCO model comprises the following steps:

[0024] In the analytical field calculation module, the gas exchange flux is calculated for each sampling grid.

[0025] In the boundary condition processing module, the boundary condition of the gas exchange flux is calculated.

[0026] In the three-dimensional advection term time integral solving module, the boundary condition of the gas exchange flux is applied to the sampling grid on the surface of the target sea area, and the concentration distribution of the passive tracer is solved.

[0027] The boundary condition of the gas exchange flux is:

[0028] ;

[0029] wherein, is the gas exchange flux, is the eddy diffusivity, is the passive tracer concentration, is the ocean sea surface height in the direction.

[0030] Further comprising the following steps:

[0031] Setting a macro definition switch module, in which a file macro definition is configured, used to activate the macro of the conditional compilation branch of the gas exchange flux calculation;

[0032] After activation, the gas exchange flux is calculated in the parsed field calculation module by traversing each sampling grid; if not activated, the flux data is not read.

[0033] The computer readable storage medium of the present application stores a computer program, which is executed by a processor to implement the passive tracer diffusion simulation method based on the CROCO model.

[0034] The computer program product of the present application comprises a computer program, which is executed by a processor to implement the passive tracer diffusion simulation method based on the CROCO model.

[0035] Advantages: Compared with the prior art, the present application has the advantages that the present application considers the gas exchange of the gas passive tracer on the sea surface, embeds it into the CROCO ocean model, and improves the convection diffusion equation. At the same time, by adding a surface boundary condition to the vertical diffusion term, the gas exchange flux can run in the original CROCO ocean model. The present application significantly improves the simulation accuracy of the passive tracer concentration distribution, especially the consistency with the observation data in the surface water body. The present application has strong universality and portability, and can be extended to different types of gas tracers and other ocean numerical models. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is the passive tracer diffusion simulation method flowchart of the present application embodiment 1.

[0037] Figure 2 is the passive tracer diffusion simulation method flowchart of the present application embodiment 2.

[0038] Figure 3 is the running result chart of the CROCO model before and after improvement in the present application embodiment 3. DETAILED DESCRIPTION

[0039] The technical solutions of the present application will be further described below with reference to the accompanying drawings.

[0040] Example 1

[0041] like Figure 1 As shown in this embodiment, a passive tracer diffusion simulation method based on the CROCO model includes the following steps:

[0042] Step 1: Obtain observational data on the content of passive tracers in the sampling grid of the target sea area; the passive tracers are stable gaseous compounds.

[0043] Specifically, in step 1, samples are taken from several observation stations and their tracer content is analyzed. The observation data with unequal intervals are processed into grid data to facilitate subsequent comparison with model data.

[0044] Step 2: Use the CROCO model to simulate the diffusion of passive tracers and obtain the concentration distribution of passive tracers; wherein, the convection-diffusion equation in the CROCO model considers the gas exchange flux of passive tracers at the air-sea interface.

[0045] Specifically, in step 2, passive tracers of the stable gaseous compound type undergo gas exchange at the air-sea interface, that is, the passive tracers are released from the sea surface into the atmosphere, thus increasing the gas exchange flux of the passive tracers. (mol / m² / s) is expressed as:

[0046] (1)

[0047] in, (m / s) is the gas transport rate. (mol / m³) represents the concentration of gases in the surface ocean. (mol / m³) represents the atmospheric gas concentration. The solubility coefficient is dimensionless; it is set based on experience. and .

[0048] As a preferred embodiment, based on observational data, i.e., historical research conclusions, Set as a constant of 10 cm / hour (i.e., 2.78 × 10⁻⁶ cm / hour). -5 (m / s), and set It is zero.

[0049] Specifically, in step 2, the CROCO model performs numerical solutions based on the assumptions of hydrostatic equilibrium, Boussinesq, fluid incompressibility, the Earth's spherical approximation, the thin-shell approximation, and the Navier-Stokes equations under the turbulent closure approximation. The scalar concentration field in the CROCO model under Cartesian coordinates... The time evolution of a scalar concentration (e.g. temperature, salinity, or a user-defined passive tracer) is represented by the advection-diffusion equation:

[0050] (2)

[0051] where, is the scalar concentration, , and are the flow velocities in the , and directions, is the eddy diffusivity, is the molecular diffusivity, is the vertical diffusion term, is the source / sink term (representing external input / output of the scalar, e.g. sea surface heat flux, freshwater flux), is the dissipation term (representing momentum / scalar mixing or dissipation due to turbulence, etc.). This equation is the underlying control equation for the transport of all passive tracers (including temperature, salinity, and user-defined passive tracers) in CROCO.

[0052] In the original scalar concentration advection-diffusion equation (equation (2)) of CROCO model, the air-sea interface exchange process of user-defined passive tracers other than temperature and salinity is not included, leading to large deviations between simulation results and actual observations. To consider the gas exchange process of user-defined passive tracers other than temperature and salinity at the air-sea interface, equation (2) is rewritten as:

[0053] (3)

[0054] where the newly added surface source / sink term is defined as:

[0055] (4)

[0056] where, is the sea surface gas exchange flux, given by equation (1); is the Dirac function, representing that this flux only acts on the sea surface ( , is the sea surface height).

[0057] Specifically, the concentration distribution of the passive tracer is a three-dimensional spatial concentration distribution, with spatial dimensions in the east-west direction (X) and the north-south direction (Y) and the vertical layer (terrain-following coordinate layer S), which is used to analyze the distribution of the tracer, the location of the high-concentration area, and the comparison with the observations, and to reveal the horizontal and vertical transport and stratification structure of the tracer under the influence of large-scale ocean currents, mesoscale eddies, and frontal processes. All outputs are grid data based on the model grid, which can be directly used to draw spatial distribution maps, profile maps, and time series curves.

[0058] Step 3: Interpolate the concentration distribution of the passive tracer to the spatial position of the grid data, calculate the average relative error between the observed values and the simulated values in the concentration distribution; if the average relative error exceeds its corresponding threshold value, modify the numerical value of and / or and re-perform passive tracer diffusion simulation.

[0059] Specifically, after interpolating the concentration distribution of the passive tracer to the spatial position of the grid data, a "simulation-observation" paired data set is formed. Then, the regional average relative error (MRE) of the entire study area is calculated:

[0060] ; (5)

[0061] wherein, is the model simulation value, is the observed value, is the total number of data points.

[0062] In some optional embodiments, the MRE≤40% is set as the accuracy threshold for acceptable simulation results, which takes into account the uncertainty of ocean observations, the resolution limit of the model, and the representative error of the model parameterization. If the verification result shows that the MRE meets the threshold requirement, it indicates that the current model improvement scheme is effective and can be used for subsequent scientific research and operational application; if the MRE exceeds 40%, it indicates that the parameterization scheme or the key parameters (such as the gas transport rate k in step two) need to be further optimized. At this time, step 2 should be returned to, and the empirical coefficients should be re-calibrated and adjusted in combination with the observation data and relevant literature, and then steps 2 and 3 are repeated until the simulation accuracy meets the above verification standard, thereby forming a closed "simulation-verification-optimization" iterative process to ensure the stability and reliability of the model improvement.

[0063] Embodiment 2

[0064] Integrating the gas exchange flux into a regional ocean model such as CROCO with specific architecture, numerical kernel and code system is not a simple function migration or module coupling. Different models have significant differences in physical process parameterization, discretization method, boundary condition processing, variable transfer mechanism and code structure, etc. Direct transplantation often leads to numerical instability, inconsistent physical processes, reduced computational efficiency and even model collapse.

[0065] Specifically, the CROCO model adopts a terrain-following coordinate, a free surface treatment, and a non-hydrostatic balance option. The passive tracer module (PASSIVE_TRACER) in the original version does not open the sea surface flux interface. To introduce the gas exchange process, a comprehensive and in-depth understanding of its internal variable transfer, time integration process, boundary condition setting mechanism, etc. is required, and multiple model core files (such as analytical.F, get_vbc.F, step3d_t.F, etc.) need to be improved while maintaining consistency with the original dynamics and thermodynamics processes.

[0066] As shown in Figure 2 , the passive tracer diffusion simulation method based on the CROCO model according to the embodiment, by modularly embedding the improved formula in the CROCO model, and keeping code compatibility through conditional compilation, realizes the CROCO model running of the passive tracer diffusion simulation method according to embodiment 1, including the following steps.

[0067] Step S1, configure file macro definition, used to activate the passive tracer sea surface flux module.

[0068] In the model configuration file (i.e. preprocessor macro definition file) cppdefs.h, a physical process switch named "ANA_STFLUX_TPAS" is added to activate the passive tracer sea-air interface flux calculation module, which controls whether the PASSIVE_TRACER module has a sea surface flux process. The macro definition controls whether the related code is effective through conditional compilation. Through conditional compilation, code modularization can be achieved, avoiding invasive modification of the original model architecture. The function and flow are as follows:

[0069] a) Compilation stage: activate the conditional compilation branch through the macro definition ANA_STFLUX_TPAS.

[0070] b) Model initialization: when the model starts, the macro will trigger the following operations: compile the ana_stflux_tpas function in analytical.F; enable the passive tracer flux related branch in get_vbc.F and step3d_t.F.

[0071] Step S2, add a new analytical function to calculate the surface flux of each horizontal grid point.

[0072] In the analytical field calculation module croco / OCEAN / analytical.F, add the ana_stflux_tpas subroutine to calculate the passive tracer's surface flux. The function and flow are as follows:

[0073] a) Traverse all horizontal grid points, and calculate the passive tracer's surface flux according to the formula (4) in Embodiment 1, , with the unit of m / s (i.e. the form of flux divided by the reference density), which embodies the diffusion process of the tracer from the ocean surface to the atmosphere.

[0074] b) Periodic calculation: if the flux changes over time, it can be recalculated at each time step.

[0075] c) Periodic calculation: if the flux changes over time, it can be recalculated at each time step.

[0076] Step S3, boundary condition processing.

[0077] In the boundary condition processing module (croco / OCEAN / get_vbc.F), call ana_stflux_tpas to calculate the flux. The function and flow are as follows:

[0078] a) Time step loop: execute in the vertical boundary condition calculation stage of each model time step.

[0079] b) Add conditional compilation branch: if ANA_STFLUX_TPAS is enabled, call ana_stflux_tpas to calculate the flux.

[0080] c) Data transfer: write the calculated stflx into the model public variable pool for subsequent module calls.

[0081] Step S4, run the time integration module.

[0082] In the three-dimensional advection term time integration solution module (croco / OCEAN / step3d_t.F), add the passive tracer's surface flux in the vertical diffusion term calculation part, integrate the surface flux into the vertical diffusion process, and realize the interface exchange of the passive tracer. The integration method is to introduce the surface source-sink term in the vertical boundary condition , i.e. embed the surface source-sink term as the surface boundary condition in the vertical diffusion term. By discretizing the formula (3) in Embodiment 1, the boundary condition applied in the surface grid element is obtained:

[0083] ; (6)

[0084] where, is the gas exchange flux, is the eddy diffusivity, is the passive tracer concentration, is the ocean directional sea surface height.

[0085] Example 3

[0086] This example performs a CROCO model run to verify the passive tracer dispersion simulation method based on the CROCO model described in the present application through a specific experiment (tracer dispersion simulation in the coastal sea area along Peru). This example takes the coastal sea area along Peru in the tropical southeast Pacific (west longitude 70-100°, south latitude 5-35°) as an example, simulates the dispersion process of passive tracers released in this area in the ocean, and verifies the effectiveness of the method of the present application by comparing with the observation data of the POSTRE cruise.

[0087] Step 1: Observation data acquisition and preprocessing.

[0088] The observation data used in this experiment is derived from the “Peru Minimum Oxygen Zone System Tracer Release Experiment” (POSTRE). The experiment injected a total amount of 68.5 kg of artificial tracer sulfur pentafluoride trifluoromethyl (CF3SF5) at three different locations (10°42.9′S, 78°14.1′W; 12°21.87′S, 77°26.25′W; 14°02.16′S, 76°31.67′W) at the bottom boundary layer of the Peru continental slope in October 2015. Seventeen months after the release, a large-scale cruise survey was conducted, covering the Peru-Chile coastal sea area from south latitude 10.7° to 31° and west longitude 86° to the west, including a total of 132 CTD observation stations, each station collecting about 24 water samples at different depths for tracer concentration analysis.

[0089] For ease of comparison with the model output data, the non-equidistant observation data is processed into regular grid data by the bilinear interpolation method. The interpolated data format is a three-dimensional (longitude x latitude x depth) concentration field with a unit of fmol / kg.

[0090] Step 2: Determine the key parameters in the empirical formula of the passive tracer sea surface flux.

[0091] ​​For the passive tracer of the type of gas compound, there is a phenomenon of release to the atmosphere at the air-sea interface. According to the POSTRE observation data and related literature (Wanninkhof, R., Asher, W. E., Ho, D. T., Sweeney, C., & McGillis, W. R. (2009). Advances in quantifying air-sea gas exchange and environmental forcing. Ann Rev Mar Sci, 1, 213-244. http: / / doi.org / 10.1146 / annurev.marine.010908.163742), in this embodiment, the is set to be a constant of 10 cm / hour (i.e. 2.78*10 -5 m / s); since the background concentration can be ignored, it is set to be zero. The parameter setting embodies the process of net release of the tracer from the surface layer of the sea to the atmosphere. The background concentration can be ignored, so it is set to be zero. The parameter setting embodies the process of net release of the tracer from the surface layer of the sea to the atmosphere.

[0092] Step three: introduce the sea surface flux into the CROCO model control equation.

[0093] In order to introduce the passive tracer air-sea interface gas exchange, the surface boundary flux term is added in the convection diffusion equation, and in numerical discretization, the flux is embedded in the vertical diffusion term as a surface boundary condition, which is realized in the time integration module (step3d_t.F) in step four.

[0094] Step four: modular implementation of sea surface flux coupling in the CROCO model. Referring to Figure 2 , step four specifically includes the following steps.

[0095] Step 4.1, macro definition of configuration file: compilation switch of function module.

[0096] In the model preprocessing configuration file cppdefs.h, define the macro ANA_STFLUX_TPAS as a global switch to control the passive tracer sea surface flux module. This design realizes the modular management of functions. The functions and processes are as follows:

[0097] 1) Compilation control: in the model compilation stage, if ANA_STFLUX_TPAS is defined, the preprocessor will activate all the conditional compilation branches of the subsequent air-sea gas exchange flux. Otherwise, the relevant content will be excluded, and the model will return to the original state.

[0098] ​2) Modular integration: This macro serves as a unified "signal" that is invoked in multiple core modules such as analytical.F, get_vbc.F, and step3d_t.F, ensuring that related functionalities are enabled or disabled in coordination, avoiding scattered and intrusive code modifications.

[0099] Step 4.2, Add parsing function: Core calculator of sea surface flux.

[0100] In the analytical field calculation module analytical.F, a subroutine named ana_stflux_tpas is added. This function is the core implementation of the physical parameterization, responsible for calculating the tracer sea surface flux at each horizontal grid point. The function and process are as follows:

[0101] 1) Input and output: The function input is the grid index and tracer number itrc; the output is the kinematic surface flux stored at each horizontal grid point , with a unit of m / s (i.e. the form of flux divided by the reference density).

[0102] 2) Flux calculation: The function iterates through each horizontal grid point in the calculation domain, and calculates the instantaneous flux according to formula (1) in embodiment 1. The core calculation formula is embodied in the model run as:

[0103] ; (7)

[0104] Where, since the background concentration in the atmosphere can be ignored, it is set to zero, , , , take the tracer concentration at the surface of the previous time step of the model , , set as a constant (2.78×10 -5 m / s). Therefore, formula (7) is simplified as:

[0105] ; (8)

[0106] This negative flux explicitly represents the net release of tracers from the ocean surface to the atmosphere.

[0107] 3) Invocation timing: The function is called at the initialization of the model to set the initial flux field. If the flux needs to change over time (such as depending on the time-varying wind field), it can be re-invoked at the boundary condition update stage of each model time step.

[0108] Step 4.3, Boundary condition processing: Transmission channel of flux data.

[0109] In the vertical boundary condition processing module get_vbc.F, the surface flux assignment logic of the passive tracer is modified to enable the call of the newly added analytical function. The function and flow are as follows:

[0110] 1) Conditional branching: In the original flux assignment, a conditional compilation branch is added for the passive tracer. If the ANA_STFLUX_TPAS macro is enabled, the call ana_stflux_tpas_tile(...) statement is executed to call the above analytical function to calculate the flux. Otherwise, the flux data is not read.

[0111] 2) Data writing: the calculated flux value is assigned to the corresponding horizontal slice of the three-dimensional array , and stored in the public variable pool of the model for subsequent time integration module reading.

[0112] Step 4.4, time integration module running: the final coupling of the flux process and the physical equation.

[0113] In the time integration solution module step3d_t.F of the convection-diffusion equation, the surface flux is embedded as the upper boundary condition in the vertical diffusion term solving process of the passive tracer. This is the most critical step to realize the physical process of sea-air interface exchange. The function and flow are as follows:

[0114] 1) Positioning and insertion: a special branch is added for the passive tracer in the area where the vertical diffusion term is calculated and the surface boundary condition is processed. The itrc.eq.itpas condition is used to identify whether the current integration variable is a passive tracer.

[0115] 2) Boundary condition introduction: in this branch, the surface flux stored in is introduced as the upper boundary condition into the discretization form of the vertical diffusion term. In implementation, it is embodied that the flux is multiplied by the time step dt and added to the vertical diffusion flux array of the surface grid cell.

[0116] 3) Non-local mixing scheme compatibility: the compatibility with KPP and other non-local boundary layer mixing schemes is considered. When the KPP scheme and the module of the application are both enabled, the flux will also be added to the non-local transport term ghats, ensuring the physical consistency of the mixing process under convection conditions.

[0117] In summary, step four seamlessly couples the air-sea interface exchange processes to the dynamical framework of the model by four interlocking modular modifications without destroying the original architecture of CROCO. The macro definition provides flexible control switches, the analytical function encapsulates the core physical formula, the boundary condition module is responsible for data transfer and scheduling, and the time integration module finally completes the numerical implementation of the physical process. This implementation ensures improved physical consistency, numerical stability, and good code maintainability and portability.

[0118] Step five: model running and result output.

[0119] Step 5.1, model preparation.

[0120] Deploy the improved CROCO model on a high-performance computing server running the Linux operating system: upload the CROCO source code package and decompress it; upload the improved cppdefs.h, analytical.F, get_vbc.F, and step3d_t.F files to the server and replace the corresponding files in the source code package. Configure the necessary compilation environment, such as the Intel Fortran compiler, NetCDF library, and MPI parallel environment, etc.

[0121] Step 5.2, data preparation.

[0122] 1) Topography data: Obtain the seafloor topography data of the study area from the ETOPO2 global topography dataset, with a spatial resolution of 2 arc minutes.

[0123] 2) Initial field and boundary field data: The initial conditions of the model and the lateral open boundary conditions evolving with time are both from the Hybrid Coordinate Ocean Model (HYCOM GLBu0.08) reanalysis data, which has a spatial resolution of and a time resolution of monthly. The model initialization uses the monthly mean field of January 2015, and the boundary conditions are provided by the linear interpolation of the monthly mean field of the data set from January 2015 to December 2017 to provide continuous forcing.

[0124] 3) Atmospheric forcing data: The sea surface forcing uses the Climate Forecast System Version 2 (CFSv2) atmospheric reanalysis data from January 2015 to December 2017. The data has a time resolution of 6-hour intervals, providing continuous heat flux, momentum flux (wind stress), sea surface air temperature, humidity, and precipitation variables. Its spatial resolution is about 0.5°, which is down-scaled to the model grid by bilinear interpolation.

[0125] 4) Tracer initial field preparation: The construction of tracer initial field strictly follows the release scheme of the field observation experiment (POSTRE). Specifically, at the initial time of turning on the global switch (ANA_STFLUX TPAS) of the passive tracer sea surface flux module in the open mode control, the initial field file of the CROCO model is preprocessed and rewritten by adding and defining a special tracer variable (tpas) using a MATLAB script. The core of the spatial setting is as follows:

[0126] a) Deployment location: Complete imitation of observation, injection at three selected latitude and longitude site locations on the continental slope of Peru (see Table 1 for site latitude and longitude).

[0127] b) Vertical distribution optimization: To avoid the generation of excessive vertical concentration gradient due to point source injection, leading to numerical instability or non-physical diffusion in subsequent simulation, the following two key measures are taken:

[0128] 1. The tracer is not placed directly on the seafloor, but in several vertical layers above the seafloor;

[0129] 2. At each site, the released tracer is not concentrated in a single grid point, but is distributed to 2 to 3 adjacent model layers in the vertical direction, and a weighted average algorithm centered on the target depth (about 250 meters) is used for mass distribution, so that a more realistic and vertically smooth "tracer group" is formed in the initialization stage (Table 1).

[0130] 5) Upload the completed grid, forcing field, initial field and boundary field files to the server.

[0131] Table 1. Parameters for setting the initial concentration field of passive tracer (TPAS) in CROCO model

[0132] Step 5.3, model configuration.

[0133] 1) Simulation area and grid: East Tropical South Pacific Ocean (

[0134] , ). The horizontal direction adopts orthogonal curvilinear grid with a resolution of about . The vertical direction adopts terrain following coordinates (S coordinates), which are divided into 40 layers, and the upper ocean is encrypted by setting parameters (theta_s=8, theta_b=1) to improve the resolution ability of the mixed layer and boundary layer processes.

[0135] ​2) Physical process parameterization: K-profile parameterization scheme is used for vertical mixing process; Smagorinsky nonlinear parameterization scheme is used for horizontal diffusion; advection terms are discretized by third-order upwind scheme. Tidal effect is not considered in this simulation.

[0136] 3) Computing resources and running setup: The simulation is run in parallel mode on a Linux high-performance computing cluster. The model integration starts from the initial field in January 2015, and is first run to September 30, 2015 without considering the tracer sea surface fluxes. Then, based on the output data at the last time step of 30 days, the initial concentration field of passive tracer is implanted. By enabling the switch ANA_STFLUX_TPAS in the preprocessor macro definition file (cppdefs.h), the sea surface flux module is activated, and the integration is then continued to December 2017 to complete the long-term simulation including the complete interface exchange process.

[0137] Step 5.4, model running and task submission.

[0138] 1) According to the grid scale of the simulation area, modify the model parameter file (param.h) to set the horizontal grid number (LLm0=371, MMm0=364), the vertical layer number (N=40), and the number of parallel computing cores (NP_XI=20, NP_ETA=18, total 360 computing cores).

[0139] 2) Modify the main input file (croco.in) to set the model time step to 60 seconds, define the S-coordinate parameters (theta_s=8, theta_b=1), and correctly specify the paths and names of the grid, forcing field, initial field, and boundary field input data files.

[0140] 3) In the compilation script (jobcomp), configure the model source code path, compiler (such as Intel Fortran), parallel library directory, and other dependent library paths, then submit the compilation task to generate the executable file.

[0141] 4) After successful compilation, submit the job script on the computing cluster to start the numerical simulation task.

[0142] Step 5.5, result output.

[0143] After the model runs, output the result file in NetCDF format. The file contains three-dimensional concentration field, temperature field, salinity field, density field, three-dimensional flow field, and sea surface height field, etc. The output time frequency is instantaneous data every 3 days, and this high time resolution output facilitates detailed analysis of the transient characteristics of tracer diffusion, circulation evolution, etc., and provides a data basis for subsequent quantitative verification and visualization analysis.

[0144] Step six: simulation result verification and evaluation.

[0145] Step 6.1, quantitative verification.

[0146] To quantitatively evaluate the simulation accuracy of the improved model, the three-dimensional concentration field output by the model is interpolated to the precise positions (longitude, latitude, depth) of the 132 observation sites by the bilinear interpolation method. The regional mean absolute relative error (MRE) is used as the evaluation index. It is calculated that the MRE of the improved model is 32%, which is significantly lower than the set acceptable threshold of 40%, indicating that the method effectively improves the simulation accuracy.

[0147] Step 6.2, tracer simulation effect comparison.

[0148] The core effect of model improvement is intuitively reflected in the simulation of tracer concentration distribution. Figure 3 The figure intuitively shows the instantaneous distribution difference of tracer concentration in the meridional-vertical section (horizontal axis: longitude 100°W-70°W, vertical axis: depth 0-500 m) before and after the model improvement, and shows the spatial characteristics of the isodensity line (black line) and tracer concentration (color filling, right color bar, unit: fmol / kg). The most significant difference between the two figures is reflected in the distribution characteristics of the tracer in the surface layer (0~50 m), which directly reflects the correction of the model improvement to the "non-real surface tracer concentration accumulation".

[0149] As shown in (a) of FIG. 6, Figure 3 The simulation result before improvement shows that in the surface ocean with a depth of less than 50 m, a large range of red high value area (concentration > 2 fmol / kg) appears, forming an obvious "non-real surface tracer concentration accumulation". This phenomenon is seriously inconsistent with the fact that the sea surface tracer concentration is close to 0 in the POSTRE observation, which is a systematic bias caused by the original model ignoring the gas exchange at the sea-air interface.

[0150] As shown in (b) of FIG. 6, Figure 3 The simulation result after improvement shows that the false high concentration accumulation in the surface layer completely disappears, and the tracer concentration in the depth range of 0~50 m is generally reduced to below 1 fmol / kg, which is consistent with the observation characteristics, indicating that the improved model can more realistically reproduce the vertical transport and stratification adjustment process of the tracer in the ocean interior, effectively eliminating the "non-physical surface accumulation of tracers".

[0151] Through the above analysis, the improved CROCO model successfully eliminates the problem of false high surface concentration caused by ignoring air-sea exchange, and the simulation results are significantly improved in accordance with the observation data. The application has a clear physical mechanism, stable numerical implementation and good portability, and is suitable for the application of various stable gas tracers in the regional to global scale numerical simulation of the CROCO model in the ocean.

[0152] The computer readable storage medium of the application stores a computer program, and the computer program is executed by a processor to realize the passive tracer diffusion simulation method based on the CROCO model.

[0153] The computer program product of the application comprises a computer program, and the computer program is executed by a processor to realize the passive tracer diffusion simulation method based on the CROCO model. The computer readable storage medium can include RAM, ROM, EEPROM, CDROM or other optical disk storage devices, magnetic disk storage devices or other magnetic storage devices, flash memory or any other media that can be used to store program codes or data structures in the form of instructions and can be accessed by a computer.

[0154] The processor is used to execute the computer program stored in the memory to realize each step in the method related to the above-mentioned embodiments.

Claims

1. A passive tracer diffusion simulation method based on the CROCO model, characterized in that, The method comprises the following steps: Obtaining observation data of passive tracer content in a sampling grid of a target sea area; the passive tracer is a stable gas compound type; Simulating passive tracer diffusion by using a CROCO model to obtain a concentration distribution of the passive tracer; in a convection diffusion equation of the CROCO model, a gas exchange flux of the passive tracer at a sea-air interface is considered. The convection diffusion equation is as follows: ； where is the surface source / sink term, , is the gas exchange flux, is the Dirac function, is the sea surface height, is the passive tracer concentration, , and are the flow velocities in , and directions, respectively, is the eddy diffusivity, is the molecular diffusivity, is the vertical diffusion term, is the source / sink term, is the dissipation term; The gas exchange flux ; wherein, is the gas transfer velocity, is the surface ocean gas concentration, is the atmospheric gas concentration, is a dimensionless solubility coefficient; set empirically and .

2. The CROCO mode based passive tracer diffusion simulation method of claim 1, wherein, The gas transfer rate was set to 10 cm / hour, the atmospheric gas concentration was set to 0.

3. The CROCO mode based passive tracer diffusion simulation method of claim 1, wherein, The method further comprises: processing the observation data into grid data, interpolating the concentration distribution of the passive tracer to spatial positions of the grid data, and calculating an average relative error between an observation value in the grid data and a simulation value in the concentration distribution. if the average relative error exceeds its corresponding threshold value, modifying and / or the values of the numerical passive tracer diffusion simulation.

4. A CROCO mode based passive tracer diffusion simulation system based on the method of claim 1, characterized in that, The method comprises the following steps: A data acquisition unit is configured to obtain observation data of passive tracer content in a sampling grid of a target sea area; the passive tracer is a stable gas compound type; The passive tracer is a stable gas compound type; A simulation unit is configured to simulate passive tracer diffusion by using a CROCO model to obtain a concentration distribution of the passive tracer; in a convection diffusion equation of the CROCO model, a gas exchange flux of the passive tracer at a sea-air interface is considered.

5. A passive tracer diffusion simulation method based on the CROCO model, characterized in that, The method comprises the following steps: Obtaining observation data of passive tracer content in a sampling grid of a target sea area; the passive tracer is a stable gas compound type; Simulating passive tracer diffusion by using a CROCO model to obtain a concentration distribution of the passive tracer; The CROCO model comprises an analytical field calculation module, a boundary condition processing module and a three-dimensional advection term time integration solving module; The convection diffusion equation for simulating passive tracer diffusion by using the CROCO model is as follows: ； where is the surface source / sink term, , is the gas exchange flux, is the Dirac function, is the sea surface height, is the passive tracer concentration, , and are the flow velocities in the , and directions, is the eddy diffusivity, is the molecular diffusivity, is the vertical diffusion term, is the source / sink term, is the dissipation term; Simulating passive tracer diffusion by using the CROCO model comprises the following steps: In the analytical field calculation module, the gas exchange flux is calculated by traversing each sampling grid; In the boundary condition processing module, a boundary condition of the gas exchange flux is calculated; In the three-dimensional advection term time integration solving module, the concentration distribution of the passive tracer is solved by applying the boundary condition of the gas exchange flux to sampling grids on the surface of the target sea area; The gas exchange flux ; wherein, is the gas transfer velocity, is the surface ocean gas concentration, is the atmospheric gas concentration, is the dimensionless solubility coefficient; set by experience and ; The boundary condition of the gas exchange flux is as follows: ； where, is the gas exchange flux, is the eddy diffusivity, is the passive tracer concentration, is the ocean directional sea surface height.

6. The CROCO mode based passive tracer diffusion simulation method of claim 5, wherein, The method further comprises the following steps: A macro definition switch module is configured to configure a file macro definition in the macro definition switch module, which is used to activate a macro of a condition compilation branch for calculating the gas exchange flux; After being activated, the gas exchange flux is calculated by traversing each sampling grid in the analytical field calculation module; if not activated, the flux data is not read.

7. A computer-readable storage medium storing a computer program, wherein the computer program comprises the following steps of: receiving a request for a resource from a client; determining whether the client is authorized to access the resource; and if the client is authorized to access the resource, providing the resource to the client. The computer program is executed by the processor to implement the passive tracer diffusion simulation method based on the CROCO model according to any one of claims 1-3, 5-6.

8. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the passive tracer diffusion simulation method based on the CROCO model according to any one of claims 1-3, 5-6.

Images