A method and system for generating temperature gridded data for ocean profile observation

Abstract

This application provides a method and system for generating temperature gridded data for ocean profile observations. The method includes: acquiring multiple ocean temperature observation profiles from multiple sources; for each ocean temperature observation profile, interpolating the observed temperatures included in it to the target standard depth corresponding to the standard depth set according to an interpolation strategy to obtain a first observation temperature sequence of the ocean temperature observation profile; for ocean temperature observation profiles whose observation time falls within the target time period and whose geographical location falls within the target horizontal grid, averaging their first observation temperature sequences to obtain a second observation temperature sequence; and sequentially performing moving average processing on the second observation temperature sequence according to a preset time window and spatial window to obtain temperature gridded data. Thus, through vertical interpolation and spatiotemporal smoothing, the unevenly distributed and depth-varying ocean observation profile data is finally processed into high-quality ocean temperature gridded data with uniform spatial coverage.

Description

Technical Field

[0001] This application relates to the field of marine observation and data processing, and in particular to a method and system for generating temperature gridded data for marine profile observation. Background Technology

[0002] Reliable and accurate ocean observation gridded datasets (also known as objective analysis datasets) are crucial for assessing multi-scale ocean changes and climate change, and are widely used in the initial fields, constraint fields, and output validation of numerical models. Their accuracy directly impacts the reliability of model simulations and forecasts, providing a data foundation for scientific research and policy decision-making.

[0003] Currently, the main form of ocean temperature observation data is temperature profile data (also known as temperature profile data), which refers to the distribution of temperature variation with depth along the vertical direction at a specific geographical location in the ocean, from the upper layer to the lower layer (e.g., from the sea surface to the seabed). Most observation profiles are limited by observation methods and spatial distribution, resulting in insufficient vertical resolution and temporal coverage, especially in the deep sea, high latitudes, and historical periods before 2005, exhibiting problems such as sparse data, large interlayer spacing, and discontinuity.

[0004] Since the beginning of the 21st century, with the development of buoy observation systems, remote sensing technology, oceanographic instruments, and other modern ocean observation methods, the amount of ocean data has exploded. However, this massive amount of observation data comes from diverse sources, exhibits significant differences in spatial distribution and temporal frequency, and cannot intuitively present the temporal and spatial characteristics of observed variables, thus it cannot be directly used for scientific research and model forecasting. Therefore, it is necessary to process the raw ocean profile observation data into a gridded dataset with a consistent spatial structure through vertical and horizontal interpolation, as well as temporal and spatial smoothing.

[0005] Due to the aforementioned deficiencies in in-situ oceanographic data, directly gridding the raw data introduces systematic errors, severely impacting the accuracy, continuity, and physical consistency of the results. For instance, there are numerous international methods for vertical interpolation of temperature profile data, each with its own characteristics. Linear interpolation preserves the original profile features, but the interpolation segments are discontinuous, potentially producing stepped profiles. Furthermore, the reliability of the interpolation results decreases significantly when the number of data layers is sparse. Spline interpolation produces smooth and continuous results, providing good approximations in densely observed areas, but it may produce non-physical oscillations in areas of dramatic vertical temperature changes (such as thermoclines). The PCHIPs method, which constructs piecewise cubic Hermite polynomials for interpolation, maintains the shape characteristics and monotonic trend of the temperature profile, avoiding interpolation distortion in high-gradient regions. However, its smoothness is inferior to spline interpolation, and it is sensitive to sparse data.

[0006] Besides the limitations of vertical interpolation, the international practice of using globally consistent time windows or no-slip processing strategies in time smoothing is generally to ensure the uniformity of data processing and computational efficiency to a certain extent. However, these strategies lack the ability to adapt to problems such as differences in observation density, changes in depth, and uneven observations over historical periods. This can easily lead to excessive smoothing of shallow information and insufficient stability of deep sparse data, resulting in abrupt changes, missing measurements, or non-physical fluctuations, ultimately affecting the physical representativeness and temporal consistency of gridded data.

[0007] In summary, to obtain globally consistent gridded data from raw ocean temperature profile observations, and to ensure the accuracy, spatiotemporal continuity, and scientific usability of the gridded data, optimizing the vertical interpolation process and introducing a variable time window moving average method are crucial for improving the quality of gridded data. Summary of the Invention

[0008] This application provides a method and system for generating temperature gridded data for ocean profile observation, which can generate high-quality ocean temperature gridded data.

[0009] In a first aspect, embodiments of this application provide a method for generating temperature gridded data for ocean profile observation, the method comprising:

[0010] Acquire multiple ocean temperature observation profiles from multiple sources; the ocean temperature observation profile data includes the observation time, geographical location, at least one observation depth, and the observation temperature corresponding to each observation depth.

[0011] For each ocean temperature observation profile, at least one observation temperature included in it is interpolated to the target standard depth corresponding to the standard depth set according to the interpolation strategy to obtain the first observation temperature sequence of the ocean temperature observation profile data; the interpolation strategy is determined by the number of observation depths of the ocean temperature observation profile data.

[0012] For several ocean temperature observation profiles whose observation time falls within the target time period and whose geographical location falls within the target horizontal grid of the global ocean temperature horizontal grid set, the first observation temperature series is averaged to obtain the second observation temperature series:

[0013] The second observed temperature sequence is processed by a time-moving average according to a preset time window to obtain the third observed temperature sequence. The preset time window is determined by the time window radius in matrix form. The time window radius includes the time window radius corresponding to the South Pacific Ocean, the Arctic Ocean, and general sea areas.

[0014] The third observed temperature sequence is spatially averaged according to a preset spatial window to obtain temperature grid data.

[0015] Among them, the first observed temperature sequence has depth characteristics, while the second and third observed temperature sequences and the temperature grid data all have depth, location, and time characteristics.

[0016] Therefore, this application uses vertical interpolation and spatiotemporal smoothing to process unevenly distributed and varying depth ocean observation profile data into a high-quality ocean temperature grid dataset that is spatially uniformly covered, close to real observations, and spatiotemporally continuous.

[0017] Secondly, embodiments of this application provide a temperature grid data generation system for ocean profile observation, the system comprising:

[0018] The acquisition module is used to acquire multiple ocean temperature observation profiles from multiple sources; the ocean temperature observation profiles include the observation time, geographical location, at least one observation depth, and the observation temperature corresponding to each observation depth.

[0019] The interpolation module interpolates at least one observed temperature from each ocean temperature observation profile to the target standard depth corresponding to the standard depth set according to the interpolation strategy, thereby obtaining the first observed temperature sequence of the ocean temperature observation profile data; the interpolation strategy is determined by the number of observed depths of the ocean temperature observation profile data.

[0020] The averaging module averages the first observation temperature sequence of several ocean temperature observation profiles whose observation time falls within the target time period and whose geographical location falls within the target horizontal grid of the global ocean temperature horizontal grid set, to obtain the second observation temperature sequence.

[0021] The time-moving average processing module performs time-moving average processing on the second observed temperature sequence according to a preset time window to obtain the third observed temperature sequence. The preset time window is determined by the time window radius in matrix form. The time window radius includes the time window radius corresponding to the South Pacific Ocean region, the Arctic Ocean region, and general sea regions.

[0022] The spatial moving average processing module performs spatial moving average processing on the third observed temperature sequence according to a preset spatial window to obtain temperature grid data.

[0023] Among them, the first observed temperature sequence has depth characteristics, while the second and third observed temperature sequences and the temperature grid data all have depth, location, and time characteristics.

[0024] It is understood that the beneficial effects of the second aspect mentioned above can be found in the relevant descriptions of the first aspect mentioned above, and will not be repeated here. Attached Figure Description

[0025] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This application provides a framework diagram of a temperature grid data generation system for ocean profile observation, as illustrated in an embodiment of the present application.

[0027] Figure 2 This illustration shows a schematic diagram of a method for generating temperature gridded data for ocean profile observation provided in an embodiment of this application:

[0028] Figure 3 A flowchart illustrating a method for generating temperature gridded data for ocean profile observation, provided in an embodiment of this application, is shown below:

[0029] Figure 4 This illustration shows a dynamic window smoothing scheme provided in an embodiment of this application. Figure 1 South Pacific Ocean region;

[0030] Figure 5 This illustration shows a dynamic window smoothing scheme provided in an embodiment of this application. Figure 1 Arctic Ocean region;

[0031] Figure 6 This illustration shows a dynamic window smoothing scheme provided in an embodiment of this application. Figure 1 General sea areas;

[0032] Figure 7 This illustration shows a schematic diagram of a temperature grid data generation system for ocean profile observation provided in an embodiment of this application. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions in the embodiments of this application will be described below with reference to the accompanying drawings.

[0034] In the description of the embodiments in this application, any embodiment or design that is “exemplary,” “for example,” or “by way of example” should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as “exemplary,” “for example,” or “by way of example” is intended to present the relevant concepts in a concrete manner.

[0035] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and their variations all mean "including but not limited to," unless otherwise specifically emphasized. "Multiple" can refer to one or more, where multiple means two or more.

[0036] To address the problems in existing technologies, this application provides a method and system for generating gridded temperature data for ocean profile observations. It constructs a complete processing system encompassing data information extraction, spatial interpolation, gridded calculation, and spatiotemporal smoothing, enabling the conversion from multi-source observation profiles to unified spatial resolution and regular time series. Methodologically, it utilizes a layer-adaptive vertical interpolation strategy to improve the physical consistency and continuity of the data space; and sets differentiated temporal smoothing windows based on regional, depth, and observation period characteristics to enhance the temporal continuity and stability of the data. This system overcomes the shortcomings of previous data processing methods, such as vertical interpolation and temporal discontinuities, and possesses good modular scalability and processing versatility. It provides reliable technical support for constructing high-quality ocean temperature gridded data products with unified structure, physical rationality, and long-term applicability, and is suitable for various application scenarios such as climate analysis, model-driven processing, and data reconstruction.

[0037] Figure 1 This diagram illustrates a framework of a temperature gridded data generation system for ocean profile observation, provided in an embodiment of this application. Figure 1 As shown, the system 100 includes the following modules:

[0038] Input module: Receives high-quality ocean temperature observation profile data from multiple sources;

[0039] Preprocessing module: Extracts key variables and their attributes from the received profile data and serializes them into metadata;

[0040] Vertical interpolation module: Based on the number of observation layers in the profile, different interpolation methods are used to interpolate the observation data vertically to the standard depth layer;

[0041] Gridding module: Interpolates data horizontally to a standard grid and performs temporal and spatial moving average processing;

[0042] Output module: Outputs the final grid dataset in the specified format.

[0043] Therefore, by providing a vertical interpolation strategy that adapts to the number of layers and designing a spatiotemporal smoothing window with "region-depth-time" three-dimensional rules, this system effectively solves the problems of large errors and discontinuities caused by traditional gridding methods when observations are sparse in the deep sea, polar regions, and historical periods. This enhances the stability, continuity, and physical consistency of gridded data and can be widely applied to ocean dynamic analysis, initial field construction for climate models, and assessment of long-term change trends.

[0044] Figure 2 This illustration shows a schematic diagram of a method for generating temperature gridded data for ocean profile observation provided in an embodiment of this application. Figure 2 As shown, system 100 performs the following data processing:

[0045] Data acquisition using the input module is the starting point of the data processing flow, involving the collection of raw data from various sources.

[0046] Data preprocessing is performed using a preprocessing module, including determining the number of observation points that need to be interpolated in the vertical direction (number of profile observation points), and selecting an appropriate interpolation method based on data characteristics and requirements.

[0047] The vertical interpolation module is used to perform interpolation operations to generate a consistent data profile in the vertical direction.

[0048] A gridding module is used to interpolate the data horizontally to fill in spatially missing values ​​or generate a more complete dataset. The ocean area is divided into different regions based on geographical location, such as the Arctic Ocean, general ocean areas, and the South Pacific Ocean, with different time windows set for each region. Different time windows correspond to different time window radii, which are the neighborhood ranges considered both spatially and temporally. Depth and temporal information for different ocean areas are considered when determining the time window radius. Spatially, the data is smoothed to reduce noise and irregularities, improving data usability. Temporally, the data is smoothed to reduce short-term fluctuations and highlight long-term trends.

[0049] Finally, the output module is used to output the data.

[0050] Based on the above, this application provides a detailed description of a method for generating temperature gridded data for ocean profile observation.

[0051] Figure 3 A flowchart illustrating a method for generating temperature gridded data for ocean profile observation, provided in an embodiment of this application, is shown. Figure 3 As shown, the method mainly includes the following execution steps:

[0052] Step S301: Acquire multiple ocean temperature observation profiles from multiple sources; the ocean temperature observation profiles include observation time, geographical location, at least one observation depth, and observation temperature corresponding to each observation depth.

[0053] For example, ocean temperature observation profile data are acquired from multiple sources; where multiple sources refer to the variety of data sources, which can be from real-time observations in the field or from general datasets, such as the World Ocean Database (WOD), the Global Temperature and Salinity Profile Project (GTSPP), and the Global Ocean Science Dataset (CODC) of the Chinese Academy of Sciences Ocean Science Data Center; high quality refers to in-situ observation data that has undergone quality control and bias correction and can accurately describe the state of the ocean.

[0054] The acquired ocean temperature observation profile data were preprocessed, mainly extracting key information from the ocean temperature profile data and standardizing it into metadata.

[0055] Specifically, key parameters are extracted from ocean temperature observation data, including but not limited to: ocean temperature series (where ocean temperature corresponds one-to-one with at least one observation depth), observation depth series (composed of at least one observation depth), observation time, geographical location (longitude and latitude), and the type of equipment used for observation.

[0056] The extracted data and information are packaged into metadata units based on the observation profiles. Each metadata record fully documents the temperature, depth values, and attribute information of all measuring points within a profile. For any metadata record, its observation depth and temperature are denoted as follows:

[0057] Depth sequence: A = [a1, a2, a3, ..., a m ];

[0058] Corresponding ocean temperature values: B = [b1, b2, b3, ..., b m ];

[0059] Where m≥1, it represents the number of effective observation layers contained in the profile, i.e., the number of observation depths. All metadata is stored by month to provide a unified input interface for subsequent interpolation and gridding processing.

[0060] Step S302: For each ocean temperature observation profile data, interpolate at least one of its observed temperatures to the target standard depth corresponding to the standard depth set according to the interpolation strategy to obtain the first observed temperature sequence of the ocean temperature observation profile data; the interpolation strategy is determined by the number of observed depths of the ocean temperature observation profile data.

[0061] For example, different interpolation schemes are selected based on the number of observation layers in the metadata to interpolate the profile temperature values ​​vertically to a standard depth layer. The interpolation process includes:

[0062] S302-1, Introducing the vertical standard stratification sequence D:

[0063] D = [d1, d2, d3, ..., d z ],

[0064] Where z represents dividing the seawater vertically into z layers, and d represents the depth of the j-th layer from the sea surface. i The unit is meters, where j∈[1, z].

[0065] S302-2, Based on the observation layer number m of each metadata obtained in step S301, select the corresponding interpolation scheme:

[0066]

[0067] That is, depending on the different values ​​of m, the profile temperature b i Perform vertical interpolation for classification.

[0068] For each ocean temperature observation profile, if the number of observation depths in the ocean temperature observation profile is 1, determine whether there is a standard depth in the standard depth set whose difference from the observation depth is less than a threshold. If there is, use the standard depth as the target standard depth and interpolate the observation temperature corresponding to the observation depth to the target standard depth. If there is no, delete the ocean temperature observation profile from multiple ocean temperature observation profiles.

[0069] For each ocean temperature observation profile, when the number of observation depths in the ocean temperature observation profile is 2 or 3, a linear interpolation scheme is used to interpolate the two or three observed temperatures to the target standard depths corresponding to the standard depth set.

[0070] For each ocean temperature observation profile, if the number of observation depths in the ocean temperature observation profile is greater than 3, the Reiniger-Ross interpolation scheme is used to interpolate at least four of the observed temperatures to the target standard depths corresponding to the standard depth set.

[0071] Specifically:

[0072] (1) When m = 1, there is only one observation layer (a1, b1), and if the standard depth d j Very close to the depth a1 of the observation point (less than 10) -6 If the value is positive, it is considered that a value can be directly assigned; otherwise, the observation point is discarded.

[0073]

[0074] Where ε = 10 -6To ensure approximately equal tolerance, NaN indicates that the observation point should be removed.

[0075] (2) When m = 2 or m = 3, there exists at least one pair of observation layers (a i b i ) and (a i+1 b i+1 Using linear interpolation, the interpolated temperature value is:

[0076]

[0077] Wherein, the observation depth satisfies a i <d j <a i+1 .

[0078] (3) When m≥4, there are at least 4 observation points (a k b k ), (a k+1 b k+1 ), (a k+2 b k+2 ), (a k+2 b k+2 Using the RR (Reiniger & Ross) interpolation method, the following three linear interpolation values ​​are first calculated:

[0079]

[0080] The interpolation result is then calculated as follows:

[0081]

[0082] Among them, the four observation depths satisfy a k <a k+1 <d j <a k+2 <a k+3 .

[0083] S302-3, based on the interpolation scheme of S302-2, performs vertical interpolation on all metadata. The nth metadata item is located at the standard depth layer d1~d2. z The temperature value on it is expressed as:

[0084] H n = [T1, T2...T z ],

[0085] H n This is the first observed temperature sequence of the ocean temperature observation profile data corresponding to the nth metadata entry.

[0086] Step 8303: For several ocean temperature observation profiles whose observation time falls within the target time period and whose geographical location falls within the target horizontal grid of the global ocean temperature horizontal grid set, the first observation temperature sequence is averaged to obtain the second observation temperature sequence.

[0087] For example, ocean temperature observation profile data is interpolated onto a target horizontal grid, and gridded calculations are performed to prepare for its moving average in both temporal and spatial dimensions. The process includes:

[0088] S303-1 sets the standard horizontal resolution to χ°×ψ°, dividing the globe into X parts along the east-west direction (X = 360 / χ) and Y parts along the north-south direction (Y = 180 / ψ), resulting in X×Y grids for each depth layer in the horizontal direction. The location information of any grid point is denoted as (x, y, j), where longitude x ∈ [0, 360), latitude y ∈ [-90, 90)], and depth j ∈ [d1, d2, d3]. z ].

[0089] S303-2, obtain the latitude and longitude information of N metadata entries, where the latitude and longitude of the nth profile is [P]. n Q n ], where P n ∈(0,360],Q n ∈[-90,90)], n∈[1,N]. Interpolate each data point to the corresponding target horizontal grid, following the interpolation rule:

[0090] If and only if and At that time, the data for the nth profile is divided into a grid with latitude and longitude of (x, y), where the operator... This indicates that the value will be rounded up.

[0091] S303-3, the averaging process includes taking the arithmetic mean of the first observed temperature series of several ocean temperature observation profiles in each target horizontal grid. If the target time period is set to one month, the observed data for the same month within each grid are arithmetically averaged to obtain the preliminary gridded data. The calculation method for ocean temperature data within the grid at position (x, y, j) is as follows:

[0092]

[0093] Among them, H n The variable represented is the ocean temperature value, and M is the number of data points mapped to all profiles in the latitude and longitude grid (x, y). M∈[0, N]. The current preliminary gridded global ocean temperature data has the dimension of X×Y×D.

[0094] S303-4, the time-based moving average is used because global observational data is not dense enough, especially in the deep sea and early historical data. Introducing a time window longer than one month can make the data coverage more complete. However, a smoothing window that is too large will affect the accuracy of the results. Therefore, it needs to be determined based on the characteristics of sea surface temperature changes and instrument observations. The preset time window includes the target period and two adjacent periods of the same size as the target period.

[0095] For the preliminary monthly gridded data obtained in S303-3, a fixed 3-month sliding window is used for time averaging to reduce the impact of sporadic anomalies and observation gaps.

[0096]

[0097] Where t represents the t-th month since January 1940, and t=1 represents January 1940. This is the second observed temperature sequence.

[0098] Step S304: Perform a time-based moving average on the second observed temperature sequence according to a preset time window to obtain a third observed temperature sequence; the preset time window is determined by the time window radius in matrix form; the time window radius includes the time window radii corresponding to the South Pacific Ocean region, the Arctic Ocean region, and general sea regions.

[0099] For example, the second observed temperature series is further processed by temporal moving average according to a preset time window. Considering the differences in spatiotemporal coverage of observation data in different regions, the observation data in the South Pacific and Arctic Oceans are particularly sparse due to harsh climate conditions and limited instrument deployment, and therefore are treated specially in the processing. In terms of temporal smoothing strategy, we combine the variation characteristics of different ocean depths: the temperature change in the deep sea layer is slow and stable, so a longer time window is used to enhance the extraction of long-term signals; while the upper ocean is significantly affected by the atmosphere and changes frequently, so a relatively shorter time window is used to retain its main characteristics. In addition, considering the differences in the application of different observation instruments in historical periods, the smoothing process is further divided into different time periods and depth layers. For example, the ARGO buoy has been widely deployed since 2005 and can obtain temperature profile data from 0 to 2000 meters; while the XBT mainly observed up to 500 meters from 1970 to 1990, and extended to 700 meters after 1990. Therefore, to ensure the rationality and consistency of data processing, we set smoothing windows for different observation periods and vertical depths.

[0100] Based on the observation characteristics of sea surface temperature data, the horizontal grid of the global ocean is divided into three sea areas: the South Pacific Ocean (SP), with a horizontal grid latitude and longitude range of x∈[1,360]&ye[1,120]; the Arctic Ocean (AR), with a horizontal grid latitude and longitude range of x∈[1,360]&ye[150,180]; and the general sea area, which refers to other sea areas outside the global center (GE), with a horizontal grid latitude and longitude range of x∈[1,360]&ye[1,180].

[0101] Introducing the time window radius of the SP sea area, such as Figure 4 The diagram illustrates a dynamic window smoothing scheme. Figure 1 In the South Pacific Ocean, the SP sea area is vertically divided into 8 depth layers, namely:

[0102] [1, 300), [300, 500), [500, 750), [750, 900), [900, 1500), [1500, 1950),

[0103] [1950, 2000], (2000, +∞); unit is meters;

[0104] The period from January 1940 to December 2005 is divided into eight time periods:

[0105] [1, 145], [146, 180], [181, 216], [217, 360], [361, 600], [601, 684]

[0106] [685,780], [781,792];

[0107] The time window radius of the SP sea area can be expressed as:

[0108]

[0109] Wherein, matrix W SP The row and column vectors in the diagram represent the depth and time-related characteristics of the SP sea area, respectively.

[0110] In the SP sea area, temperature data are applied using W in the following situations. SP :

[0111] (1) When t∈[1, 181], (x, y, j) satisfies

[0112] x∈[210,360], y∈[1,90], j∈[1,+∞);

[0113] (2) When t∈[182,601], (x, y, j) satisfies:

[0114]

[0115] (3) When t∈[602,793], (x, y, j) satisfies:

[0116]

[0117] In (x, y, j), x represents longitude, y represents longitude, and j represents depth. The operator ∪ represents the union of sets, and the same applies below.

[0118] Introducing the time window radius of AR sea areas, such as Figure 5 The diagram illustrates a dynamic window smoothing scheme. Figure 1 In the Arctic Ocean region, the depth of the AR zone is divided into four layers:

[0119] [1,750), [750,1950), [1950,2000], (2000,+∞); unit is meters.

[0120] All months since January 1940 are divided into five periods:

[0121] [1, 216], [217, 360], [361, 600], [601, 792], [793, +∞);

[0122] The time window radius of the AR sea area can be expressed as:

[0123]

[0124] Wherein, matrix W AR The row and column vectors in the diagram represent the depth and time-related characteristics of the AR sea area, respectively.

[0125] In the AR sea area, for (x, y, j) satisfies:

[0126] When using W AR .

[0127] Introducing the time window radius for the GE sea area, such as Figure 6 The diagram illustrates a dynamic window smoothing scheme. Figure 1 In general sea areas, the vertical depth of the GE sea area is divided into 7 layers, namely:

[0128] [1,300), [300,500), [500,750), [750,1450), [1450,1950), [1950,2000], (2000,+∞); Units are meters;

[0129] All months since January 1940 are divided into five periods:

[0130] [1, 216], [217, 360], [361, 600], [601, 792], [793, +∞);

[0131] The time window radius for the GE sea area can be expressed as:

[0132]

[0133] Wherein, matrix W GE The row and column vectors in the graph represent the depth and time-time characteristics of the GE sea area, respectively. Except for the grid points of the SP and AR sea areas, all other grid points use W. GE ;

[0134] Based on the given time window radius W, the ocean temperature at location (x, y, j) in the t-th month is calculated. Perform a time-moving average:

[0135]

[0136] The gridded data after time smoothing is the third observed temperature sequence. x,y,j，t It is the offset relative to the current time point t, 2*w x,y,j，t +1 represents the size of the time window, which includes the current time point and the time before and after W. x,y，j，t A specific point in time.

[0137] Step S305: Perform spatial moving average processing on the third observed temperature sequence according to the preset spatial window to obtain temperature grid data; wherein, the first observed temperature sequence has depth features, and the second observed temperature sequence, the third observed temperature sequence and the temperature grid data all have depth, location and time features.

[0138] For example, before performing spatial moving average processing on the third observed temperature sequence according to a preset spatial window, the method further includes: using an ensemble optimal interpolation method to perform spatial interpolation on other horizontal grids outside the target horizontal grid in the global ocean temperature horizontal grid set, and adding the spatial interpolation result to the third observed temperature sequence.

[0139] The preset spatial window includes all horizontal grids within a three-dimensional spatial region centered on the target horizontal grid, within the global ocean temperature horizontal grid set.

[0140] Specifically, a climate model is introduced, and the ensemble optimal interpolation method proposed by the Institute of Atmospheric Physics, Chinese Academy of Sciences, is used to perform spatial interpolation on the data. The gridded data obtained from S304 processing is then processed into gridded data covering global ocean temperatures. The ocean temperature at location (x, y, j) in the t-th month is denoted as T. x,y,j，t ;

[0141] Perform local median smoothing on the spatial dimension of the spatially interpolated data:

[0142] To extract temperature values ​​within a neighborhood of 3×5 grid points horizontally and 3 vertical layers above and below the current grid point, and to calculate the median of the temperature values ​​within this neighborhood as a replacement value for the target grid point:

[0143]

[0144] in, The spatial domain centered at the grid point (x, y, j) is a 3×5×3 grid window, where median represents the median value. The temperature values ​​are spatially smoothed, ultimately generating a three-dimensional grid of global ocean temperatures.

[0145] Finally, output the data file according to the specified data format (such as NetCDF, CSV, txt, Matlab, etc.).

[0146] Therefore, the scheme in this application utilizes a layer-adaptive vertical interpolation strategy to improve the physical consistency and continuity of the data space; and sets differentiated temporal smoothing windows based on regional, depth, and observation period characteristics to enhance the temporal continuity and stability of the data. This scheme has the following beneficial effects and advantages:

[0147] 1. By adopting a "layer-adaptive" vertical interpolation strategy, the optimal interpolation method is adaptively selected based on the vertical structural characteristics of different observation profiles. This effectively avoids problems such as non-physical oscillations, profile steps, or thermocline distortion caused by unreasonable interpolation method selection in the vertical direction when the observation data is sparse, the structure is complex, or there are abrupt changes. It balances interpolation accuracy, profile continuity, and algorithm stability, thereby improving the quality and reliability of gridded data products.

[0148] 2. It can set targeted "region-depth-time" three-dimensional rule smoothing windows according to the ocean temperature characteristics of different geographical regions, depth levels and observation periods. While maintaining the ability to handle rapid changes in shallow layers, it enhances the temporal continuity and stability of deep or historical data. It avoids abrupt changes and extreme values ​​caused by sparse observations and uneven spatial distribution of observation data. It effectively balances the dual goals of "fidelity preservation" and "noise reduction". It is a key technological innovation to solve the limitations of the smoothing mechanism of similar international data products, and improves the temporal comparability of data and the ability to analyze climate trends.

[0149] 3. It has strong scalability and adaptability, and can be widely applied to the construction of gridded data of different instrument types and different observation variables. It also supports the construction of gridded data of various vertical resolutions (such as 41 layers and 119 layers) and horizontal resolutions (such as 1°×1° and 0.5°×0.5°) in the ocean, and has the potential to be extended to multi-source and multi-scale ocean observation data processing platforms.

[0150] It is understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. Furthermore, in some possible implementations, each step in the above embodiments may be selectively executed according to actual circumstances; it may be partially or fully executed, without limitation here. Additionally, all or part of any feature in the above embodiments can be freely and arbitrarily combined without contradiction. The combined technical solutions are also within the scope of this application.

[0151] Figure 7 This illustration shows a schematic diagram of a temperature gridded data generation system for ocean profile observation provided in an embodiment of this application. Figure 7 As shown, the temperature grid data generation system 700 includes:

[0152] The acquisition module 710 is used to acquire multiple ocean temperature observation profile data from multiple sources; the ocean temperature observation profile data includes the observation time, geographical location, at least one observation depth, and the observation temperature corresponding to each observation depth.

[0153] The interpolation module 720 interpolates at least one observed temperature from each ocean temperature observation profile to the target standard depth corresponding to the standard depth set according to the interpolation strategy, thereby obtaining the first observed temperature sequence of the ocean temperature observation profile data; the interpolation strategy is determined by the number of observed depths of the ocean temperature observation profile data.

[0154] The averaging module 730 averages the first observation temperature sequence of several ocean temperature observation profiles whose observation time falls within the target time period and whose geographical location falls within the target horizontal grid of the global ocean temperature horizontal grid set, to obtain the second observation temperature sequence.

[0155] The time-moving average processing module 740 performs time-moving average processing on the second observed temperature sequence according to a preset time window to obtain the third observed temperature sequence; the preset time window is determined by the time window radius in matrix form; the time window radius includes the time window radius corresponding to the South Pacific Ocean region, the Arctic Ocean region, and general sea regions.

[0156] The spatial moving average processing module 750 performs spatial moving average processing on the third observed temperature sequence according to a preset spatial window to obtain temperature grid data.

[0157] Among them, the first observed temperature sequence has depth characteristics, while the second and third observed temperature sequences and the temperature grid data all have depth, location, and time characteristics.

[0158] Based on the methods in the above embodiments, this application provides an electronic device. The electronic device may include: at least one memory for storing a program; and at least one processor for executing the program stored in the memory. When the program stored in the memory is executed, the processor executes the methods described in the above embodiments. Exemplarily, the electronic device may be a mobile phone, tablet computer, desktop computer, laptop computer, handheld computer, server, ultra-mobile personal computer (UMPC), netbook, cellular phone, personal digital assistant (PDA), or artificial intelligence (AI) device. This application does not impose any special limitations on the specific type of the electronic device.

[0159] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0160] It is understood that the various numerical designations used in the embodiments of this application are merely for descriptive convenience and are not intended to limit the scope of the embodiments of this application. It should be understood that in the embodiments of this application, the order of the process numbers does not imply the order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0161] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of this application should be included within the scope of protection of this application.

Claims

1. A temperature grid data generation method for ocean profile observation, characterized by, The method comprises: obtaining a plurality of ocean temperature observation profile data from multiple sources; the ocean temperature observation profile data comprises observation time, geographical position, at least one observation depth, and observation temperature corresponding to the observation depth; for each ocean temperature observation profile data, at least one observation temperature included in the ocean temperature observation profile data is respectively interpolated into a target standard depth corresponding to the standard depth set according to an interpolation strategy, to obtain a first observation temperature sequence of the ocean temperature observation profile data; the interpolation strategy is determined by the number of observation depths of the ocean temperature observation profile data; wherein, for each ocean temperature observation profile data, when the number of observation depths of the ocean temperature observation profile data is 1, determine whether there is a standard depth in the standard depth set whose difference with the observation depth is less than a threshold value; if there is, the standard depth is taken as the target standard depth, and the observation temperature corresponding to the observation depth is interpolated into the target standard depth; if there is not, the ocean temperature observation profile data is deleted from the plurality of ocean temperature observation profile data; for each ocean temperature observation profile data, when the number of observation depths of the ocean temperature observation profile data is 2 or 3, two or three observation temperatures included in the ocean temperature observation profile data are respectively interpolated into a target standard depth corresponding to the standard depth set by using a linear interpolation scheme; for a plurality of ocean temperature observation profile data whose observation time falls into a target time period and whose geographical position falls into a target horizontal grid in the global ocean temperature horizontal grid set, the first observation temperature sequence of the plurality of ocean temperature observation profile data is averaged to obtain a second observation temperature sequence; the second observation temperature sequence is processed by time sliding average according to a preset time window to obtain a third observation temperature sequence; the preset time window is determined by a time window radius in matrix form; the time window radius comprises a time window radius corresponding to each of the South Pacific Ocean, the Arctic Ocean, and the general sea area; the third observation temperature sequence is processed by space sliding average according to a preset space window to obtain temperature grid data; wherein, the first observation temperature sequence has depth characteristics, and the second observation temperature sequence, the third observation temperature sequence, and the temperature grid data all have depth, position, and time characteristics.

2. The method of claim 1, wherein, The method comprises: for each ocean temperature observation profile data, when the number of observation depths of the ocean temperature observation profile data is greater than 3, at least four observation temperatures included in the ocean temperature observation profile data are respectively interpolated into a target standard depth corresponding to the standard depth set by using a Reiniger-Ross interpolation scheme.

3. The method of claim 1, wherein, The averaging process comprises performing arithmetic averaging on the first observation temperature sequence of a plurality of ocean temperature observation profile data in each target horizontal grid.

4. The method of claim 1, wherein, The preset time window further comprises the target time period and each time period adjacent to the target time period and having the same size as the target time period.

5. The method of claim 1, wherein, The row vectors and the column vectors of the matrix respectively represent depth features and time period features of the South Pacific Ocean region, or the North Polar Ocean region, or the general sea region.

6. The method of claim 1, wherein, Before the third observation temperature sequence is subjected to the spatial moving average processing according to the preset spatial window, the method further comprises the following steps of: using a set optimal interpolation method, performing spatial interpolation on other horizontal grids outside the target horizontal grid in the global ocean temperature horizontal grid set, and adding the spatial interpolation result to the third observation temperature sequence.

7. The method of claim 1, wherein, The preset spatial window includes all horizontal grids in a three-dimensional spatial region centered on the target horizontal grid in the global ocean temperature horizontal grid set.

8. A temperature grid data generation system for ocean profile observation, characterized by, The system comprises: an acquisition module configured to acquire a plurality of ocean temperature observation profile data from multiple sources; the ocean temperature observation profile data comprises observation time, geographical position, at least one observation depth, and observation temperature corresponding to the observation depth; an interpolation module configured to, for each piece of ocean temperature observation profile data, interpolate the at least one observation temperature included in the ocean temperature observation profile data to a corresponding target standard depth in a standard depth set according to an interpolation strategy to obtain a first observation temperature sequence of the ocean temperature observation profile data; the interpolation strategy is determined by the number of observation depths of the ocean temperature observation profile data; wherein, for each piece of ocean temperature observation profile data, when the number of observation depths of the ocean temperature observation profile data is one, it is determined whether there is a standard depth in the standard depth set that has a standard depth difference less than a threshold value from the observation depth; if there is, the standard depth is taken as a target standard depth, and the observation temperature corresponding to the observation depth is interpolated to the target standard depth; if there is not, the ocean temperature observation profile data is deleted from the plurality of ocean temperature observation profile data; for each piece of ocean temperature observation profile data, when the number of observation depths of the ocean temperature observation profile data is two or three, linear interpolation is adopted to interpolate the two or three observation temperatures to the corresponding target standard depths in the standard depth set; an average processing module configured to, for a plurality of ocean temperature observation profile data whose observation time falls within a target time period and whose geographical position falls within a target horizontal grid in a global ocean temperature horizontal grid set, perform average processing on the first observation temperature sequences of the plurality of ocean temperature observation profile data to obtain a second observation temperature sequence; a time moving average processing module configured to, according to a preset time window, perform time moving average processing on the second observation temperature sequence to obtain a third observation temperature sequence; the preset time window is determined by a time window radius in matrix form; the time window radius comprises a time window radius corresponding to each of a South Pacific Ocean region, a North Polar Ocean region, and a general sea region; a spatial moving average processing module configured to, according to a preset spatial window, perform spatial moving average processing on the third observation temperature sequence to obtain temperature grid data; wherein the first observation temperature sequence has depth features, and the second observation temperature sequence, the third observation temperature sequence, and the temperature grid data all have depth, position, and time features.

Images