CN114399073A - Ocean surface temperature field prediction method based on deep learning - Google Patents

Ocean surface temperature field prediction method based on deep learning Download PDF

Info

Publication number
CN114399073A
CN114399073A CN202111419175.2A CN202111419175A CN114399073A CN 114399073 A CN114399073 A CN 114399073A CN 202111419175 A CN202111419175 A CN 202111419175A CN 114399073 A CN114399073 A CN 114399073A
Authority
CN
China
Prior art keywords
temperature field
inversion
ocean
model
ocean surface
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202111419175.2A
Other languages
Chinese (zh)
Inventor
宋弢
魏伟
徐丹亚
王家荣
韩润生
孟凡
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China University of Petroleum East China
Original Assignee
China University of Petroleum East China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China University of Petroleum East China filed Critical China University of Petroleum East China
Priority to CN202111419175.2A priority Critical patent/CN114399073A/en
Publication of CN114399073A publication Critical patent/CN114399073A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Artificial Intelligence (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • Business, Economics & Management (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses an ocean surface temperature field prediction method based on deep learning, which belongs to the technical field of ocean observation, and comprises the steps of firstly obtaining a data sample library containing a plurality of characteristics of an ocean surface, preprocessing different characteristics and obtaining an ocean surface temperature field time-space sample library with time-space correlation; secondly, establishing an ocean surface temperature prediction model by utilizing a convolution long-term and short-term memory network based on the ocean surface temperature field space-time sample library: the input of the prediction model is a plurality of ocean surface environment characteristics at a plurality of past moments, including ocean surface temperature, ocean surface salinity and ocean surface height, and the output of the prediction model is temperature field data of the ocean surface at a future moment; the final output of the prediction model is the predicted temperature field at the target sea area t +1, t +2, t + 3. The invention utilizes the strong nonlinear mapping capability and the multi-mode fusion capability of deep learning to carry out the prediction research of the ocean surface temperature field, and realizes the prediction of the ocean surface temperature field in time, accurately and in light weight.

Description

Ocean surface temperature field prediction method based on deep learning
Technical Field
The invention relates to the technical field of marine observation, in particular to a marine surface temperature prediction method based on deep learning.
Background
Sea Surface Temperature (SST) refers to the Temperature of water at the Surface of the ocean. It is a vital parameter of the world's oceans, playing a fundamental role in the exchange of energy, momentum and moisture between the ocean and the atmosphere, which can affect the distribution of precipitation, which can lead to extreme weather events such as drought and flooding, ultimately affecting various environmental conditions and dynamics. Therefore, the prediction of the future sea temperature is of great significance for better understanding of the dynamic change of the climate.
However, due to the large variation of heat flux, radiation and solar wind near the sea surface, the sea temperature prediction has a high degree of uncertainty and difficulty. The current ocean temperature forecast is mainly divided into two categories: one is a physical based numerical model. The change rule of the ocean temperature is described by a series of complex physical equations. These equations are very complex and usually require a lot of computational effort. The other is a data-driven model based on data. It attempts to learn the laws of temperature change from the ocean data and further uses a learning model to infer future ocean temperatures. They are far less complex than numerical methods and are suitable for predicting sea temperature at a particular location. Such methods include traditional statistical methods and the latest machine learning and artificial intelligence methods. The current machine learning method mostly stays in a single-point prediction stage and is low in efficiency, so that the method for using the ocean surface temperature field based on deep learning contributes to a new quick and accurate selection for the field.
Disclosure of Invention
The invention discloses an ocean surface temperature field prediction method based on deep learning, which belongs to the technical field of ocean observation, and comprises the steps of firstly obtaining a data sample library containing a plurality of characteristics of an ocean surface, preprocessing different characteristics and obtaining an ocean surface temperature field time-space sample library with time-space correlation; secondly, establishing an ocean surface temperature prediction model by utilizing a convolution long-term and short-term memory network based on the ocean surface temperature field space-time sample library: the input of the prediction model is a plurality of ocean surface environment characteristics at a plurality of past moments, including ocean surface temperature, ocean surface salinity and ocean surface height, and the output of the prediction model is temperature field data of the ocean surface at a future moment; the final output of the prediction model is the predicted temperature field at the target sea area t +1, t +2, t + 3. The invention utilizes the strong nonlinear mapping capability and the multi-mode fusion capability of deep learning to carry out the prediction research of the ocean surface temperature field, and realizes the prediction of the ocean surface temperature field in time, accurately and in light weight.
The embodiment of the invention provides a method for predicting an ocean surface temperature field based on deep learning, which realizes timely and accurate prediction of the ocean surface temperature field through multi-dimensional factor analysis and is used for making up the defects of the prior art.
The following presents a simplified summary in order to provide a basic understanding of some aspects of the disclosed embodiments. This summary is not an extensive overview and is intended to neither identify key/critical elements nor delineate the scope of such embodiments. Its sole purpose is to present some concepts in a simplified form as a prelude to the more detailed description that is presented later.
The invention provides a deep learning-based ocean three-dimensional temperature field inversion method, which comprises the following specific steps of:
s1, obtaining observation data of the ocean surface by re-analyzing the data set, wherein the data is real data actually measured on the ocean surface;
s2, combining the multiple environmental characteristic data actually measured on the ocean surface at the time t … t-14 with a convolution long-term and short-term memory network to predict and obtain the temperature data of the ocean surface at the time t +1, t +2 and t + 3;
s3, adopting a convolution long-short term memory network method to carry out deep information mining on the data of the ocean surface layer, and remarkably improving the prediction efficiency and the time-space relevance of the data;
s4, comparing the ocean surface temperature data at the t +1, t +2 and t +3 moments in S3 with the actually measured temperature;
s5, optimizing the network model according to the comparison result of S4 to make the network model reach the optimal state;
further, in the step S1, the data preprocessing adopts the following mechanism:
the original data is obtained through multiple reanalysis data sets such as Argo, SODA and REDOS, and meanwhile, preprocessing is conducted on missing values and abnormal values of the data. The missing value processing method is a global average interpolation method, and the abnormal value processing method is to discard the data at the current time point.
Further, in the step S2, the convolutional long-short term memory network satisfies the following condition:
(1) the input of the network model is the past t.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cndot.cnto 14, the sea surface temp. of sea surface temp. at 14 times, sea surface temp., sea surface salinity, sea surface height and other physical characteristic parameters of sea surface height, sea surface height and other 14 time point, etc. are normalized and unified into a certain range, and normalized and then passed through specific hyper-parameter combination calculation, and output as t +1, and output as sea level data of t + 1.
(2) The model architecture consists of five layers, 1 input layer, 3 ConvLSTM layers and one output layer, with input dimensions (samples, time _ steps, longitude, latitude, features), and weights are passed between each layer.
Further, in the step S3, an Attention mechanism is introduced:
considering that the contribution values of different sea surface features to the sub-surface temperature field inversion are different, we introduce an Attention mechanism to assign importance weights for different time intervals, and establish an Attention assignment mechanism on each step output of ConvLSTM: a higher attention value is distributed to a time step with a large contribution to a prediction result, a smaller attention value is distributed to a time step with a small contribution, the attention value is obtained by automatic learning of a model in a training process, and differential contributions of different time steps to sea surface temperature field prediction can be described;
further, in step S4, comparing the temperature field value predicted by the model with the observation value of the reanalysis data, feeding back the result to the model, debugging the model parameters, optimizing the model, and finally obtaining an accurate three-dimensional temperature field inversion model; two metrics are set to quantitatively measure the effect of the ocean surface temperature prediction, the first being the Root Mean Square Error (RMSE) of the predicted temperature field and the reanalyzed data set temperature field:
Figure BDA0003376247820000031
wherein, ypreTemperature value, y, representing model outputobsRepresenting the temperature values in the actual observation, and m representing the number of temperature fields input in one test set.
The second measure is the average absolute value error (MSE) between the temperature value predicted by the model and the actual observed value, and is used for measuring the average modular length of the error of the predicted value, and the calculation formula is expressed as:
Figure BDA0003376247820000032
wherein, ypreAnd yobsRespectively representing the predicted value and the real observed value of the model.
The technical scheme provided by the embodiment of the invention has the following beneficial effects:
the strong nonlinear mapping capability and the multi-modal data fusion capability of the deep learning technology are utilized to intelligently process the space-time sequence in the ocean atmosphere field, and the accurate prediction of the ocean surface temperature field is realized. A plurality of sea surface environmental characteristics influencing an ocean surface temperature field are selected in a targeted manner, and a multichannel convolution mechanism is adopted to assist a long-term and short-term memory network, so that depth information mining of time and space correlation degree is realized. In addition, the invention also introduces an attention mechanism to distinguish the differential contributions of the predicted temperature field aiming at different time steps.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
FIG. 1 is a flow chart illustrating method steps according to an exemplary illustrative method of deep learning based ocean surface temperature prediction.
FIG. 2 is a classical architecture diagram of a convolutional long-short term memory network model.
Detailed Description
The following description and the drawings sufficiently illustrate specific embodiments of the invention to enable those skilled in the art to practice them. Other embodiments may incorporate structural, logical, electrical, process, and other changes. The examples merely typify possible variations. Individual components and functions are optional unless explicitly required, and the sequence of operations may vary. Portions and features of some embodiments may be included in or substituted for those of others. The scope of embodiments of the invention encompasses the full ambit of the claims, as well as all available equivalents of the claims. Embodiments may be referred to herein, individually or collectively, by the term "invention" merely for convenience and without intending to voluntarily limit the scope of this application to any single invention or inventive concept if more than one is in fact disclosed. Herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, or apparatus. The use of the phrase "including a" does not exclude the presence of other, identical elements in a process, method or device that includes the recited elements, unless expressly stated otherwise. The embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. As for the methods, products and the like disclosed by the embodiments, the description is simple because the methods correspond to the method parts disclosed by the embodiments, and the related parts can be referred to the method parts for description.
Fig. 1 shows an alternative implementation architecture of an ocean surface temperature field prediction model based on deep learning.
In this alternative example, the system includes a ConvLSTM model, which is a variant based on a conventional long-short term memory network (LSTM) having three gates, i.e., an input gate, an output gate, and a forgetting gate, and ConvLSTM mainly changes the weight of W into a convolution operation, so that the features of the two-dimensional spatial image can be extracted. The input of the model is that 3 characteristic data such as sea surface temperature, sea surface salinity, sea surface height and the like at the t-1 … … t-13 moment are obtained from various reanalysis data and are used as the input of the model, and the final output is the sea surface temperature at the t +1, t +2 and t +3 moments.
Optionally, the reanalysis data set comes from products of well-known scientific research institutions at home and abroad, including data sets of Argo, redox, SODA and the like, which are generally stored and distributed in a NetCDF format, read and preprocess the data set by using a Python language supporting multiple extensions, and perform difference and coverage on land and island points to enable the land and island points to have complete plane values, so that subsequent convolution operation is facilitated. And simultaneously, after interpolation is completed, normalization methods are used for normalizing the ocean surface features of different categories to be (0, 1), and the ocean surface features are integrated into a uniform 5D tensor to prepare for input of a neural network model.
Optionally, as shown in fig. 2, the ConvLSTM unit performs sufficient feature extraction on a multivariate time and space series (multivariate time and space series), and continuously learns a long-term dependency relationship of the multivariate time and space series, which specifically includes: ConvLSTM replaces the matrix multiplication with a convolution operation of each gate in the LSTM cell. Thus, it captures the underlying spatial features by performing convolution operations in the multidimensional data. The main difference between ConvLSTM and LSTM is the input dimension. Since the LSTM input data is one-dimensional, it is not suitable for spatial sequence data, such as video, satellite, radar image datasets. ConvLSTM is designed for 3D data as its input, and may well take into account the spatiotemporal correlation inside the sea surface temperature data.
Optionally, considering that the contribution values of different time steps to the ocean surface temperature prediction are different, we introduce an Attention mechanism, assign importance weights of different time steps, and establish an Attention assignment mechanism on each step output of ConvLSTM: the features which contribute greatly to the target prediction result are assigned with higher attention values, the features which contribute less are assigned with smaller attention values, the attention values are obtained by automatic learning of the model in the training process, and the differential contribution of different time steps to ocean surface temperature prediction can be described.
Optionally, the model output by the convolution long-term and short-term memory network may cover a plurality of time steps, and at different corresponding times, a corresponding missing value processing mechanism may be selected, the prediction result is returned to a state where lands and islands are removed, and meanwhile, the prediction result is visually analyzed by using a technology based on the GIS and the matplotlib.
The invention utilizes the artificial intelligence technology to carry out intelligent processing on the space-time sequence in the field of ocean atmosphere, can avoid the difficult problems that a plurality of computing resources are consumed in the earth numerical mode and real-time prediction cannot be carried out, can quickly and effectively predict the distribution of the ocean surface temperature field, can assist in early warning of disasters such as 'Elnino' and 'Ranina' and guiding of various offshore operations, has high processing speed and small computing resources, and is beneficial to integration and large-scale application.
Those of skill in the art would appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention. It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the embodiments disclosed herein, it should be understood that the disclosed methods, articles of manufacture (including but not limited to devices, apparatuses, etc.) may be implemented in other ways. For example, the above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form. The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment. In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
It should be understood that the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. The present invention is not limited to the procedures and structures that have been described above and shown in the drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the invention is limited only by the appended claims.

Claims (5)

1. A marine three-dimensional temperature field inversion method based on deep learning is characterized in that a data sample library containing marine surface features is obtained, different features are preprocessed, and a marine three-dimensional temperature field sample library with time-space correlation is obtained; secondly, establishing an ocean three-dimensional temperature-salt field inversion model by utilizing a convolution long-term and short-term memory network based on the ocean three-dimensional temperature field sample library: the input of the inversion model is marine environment characteristics including sea surface temperature, sea surface salinity and sea surface wind, and the output of the inversion model is temperature field data of a sea subsurface; the final outputs of the inversion model are 100, 200, 500, 1000 and 2000 meters inversion temperature fields below the sea surface. According to the invention, the inversion analysis of the ocean three-dimensional space-time temperature field is carried out by utilizing the strong nonlinear mapping capability and the multi-mode fusion capability of deep learning, and timely, accurate and lightweight inversion of the subsurface temperature field is realized.
2. The method of claim 1, comprising the steps of:
(1) ocean reanalysis data-based construction of ocean three-dimensional temperature field sample library
Acquiring a marine three-dimensional initial data set which integrates satellite observation and reanalysis processing, preprocessing a sample, and extracting to obtain a corresponding three-dimensional temperature field characteristic sample library;
(2) construction of ocean three-dimensional temperature field inversion model
Based on the three-dimensional temperature field characteristic sample library, establishing an ocean three-dimensional temperature and salt field inversion model by utilizing a convolution long-term and short-term memory network: the input of the inversion model is environmental characteristics including sea surface temperature, sea surface salinity and sea surface wind, and the output of the inversion model is temperature field data of the sea subsurface; the method comprises the following steps:
s1: carrying out normalization processing on the sea surface environment characteristics, inputting different input factors into a normalization network, and finally fusing to obtain a five-dimensional tensor at the time t as the input of a ConvLSTM neural network model;
s2: considering that the contribution values of different sea surface features to the sub-surface temperature field inversion are different, we introduce an Attention mechanism to assign importance weights of different input features, and establish an Attention assignment mechanism on each step output of ConvLSTM: the features which contribute greatly to the target inversion result are assigned with higher attention values, the features which contribute less are assigned with smaller attention values, the attention values are obtained by automatic learning of the model in the training process, and the differential contributions of different sea surface features to the sub-surface temperature field inversion can be described;
s3: and (2) constructing a marine three-dimensional temperature field sample library as a training data training model according to the step (1), inputting sea surface environment characteristics with parameters at the time t, performing inversion prediction on the temperature field of the sea subsurface at the time t corresponding to different depths, and outputting the temperature field of the subsurface at a certain depth at the time t.
3. The marine three-dimensional temperature field inversion method of claim 1, further comprising the step of (3) marine three-dimensional temperature field inversion model validation and optimization: and carrying out verification on the constructed three-dimensional temperature field model by using the independent verification data set, adjusting model parameters according to a verification result, continuously optimizing the temperature field inversion model, and finally obtaining the high-precision ocean three-dimensional temperature field inversion model.
4. The marine three-dimensional temperature field inversion method according to claim 1, wherein in the step (1), raw data is obtained by multiple reanalyzing data sets such as Argo, SODA and redox, and preprocessing is performed on missing values and abnormal values of the data. The missing value processing method is a global average interpolation method, and the abnormal value processing method is to discard the data at the current time point.
5. The method of inverting a three-dimensional temperature field of the ocean of claim 1, wherein in step (3), two indicators are set to quantitatively measure the results of the prediction of the propagation of internal waves, the first indicator being the Root Mean Square Error (RMSE) of the inverted temperature field and the temperature field of the reanalyzed data set:
Figure FDA0003376247810000021
wherein, ypreTemperature value, y, representing model outputobsRepresenting the temperature values in the actual observation, and m representing the number of temperature fields input in time sequence in one test set.
The second measure is the average absolute value error (MSE) between the temperature of the inversion field and the actual observed value, and is used for measuring the average mode length of the predicted value error, and the calculation formula is expressed as:
Figure FDA0003376247810000022
wherein, ypreAnd yobsRespectively representing the inversion predicted value and the real observed value of the model.
CN202111419175.2A 2021-11-26 2021-11-26 Ocean surface temperature field prediction method based on deep learning Pending CN114399073A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111419175.2A CN114399073A (en) 2021-11-26 2021-11-26 Ocean surface temperature field prediction method based on deep learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111419175.2A CN114399073A (en) 2021-11-26 2021-11-26 Ocean surface temperature field prediction method based on deep learning

Publications (1)

Publication Number Publication Date
CN114399073A true CN114399073A (en) 2022-04-26

Family

ID=81225817

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111419175.2A Pending CN114399073A (en) 2021-11-26 2021-11-26 Ocean surface temperature field prediction method based on deep learning

Country Status (1)

Country Link
CN (1) CN114399073A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115147239A (en) * 2022-06-10 2022-10-04 自然资源部第一海洋研究所 Calculation method for environmental background water temperature of newly-built coastal power plant temperature rise area and temperature rise calculation method
CN115307780A (en) * 2022-09-29 2022-11-08 中国海洋大学 Sea surface temperature prediction method, system and application based on time-space information interaction fusion
CN116822710A (en) * 2023-05-24 2023-09-29 国家海洋环境预报中心 Coral reef whitening hot spot prediction method, calcification rate prediction method and electronic equipment
CN116933666A (en) * 2023-09-19 2023-10-24 深圳康普盾科技股份有限公司 Thermal management optimization method, system and medium for container energy storage system

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115147239A (en) * 2022-06-10 2022-10-04 自然资源部第一海洋研究所 Calculation method for environmental background water temperature of newly-built coastal power plant temperature rise area and temperature rise calculation method
CN115147239B (en) * 2022-06-10 2023-04-07 自然资源部第一海洋研究所 Method for calculating environmental background water temperature and calculating temperature rise of newly-built coastal power plant temperature rise area
CN115307780A (en) * 2022-09-29 2022-11-08 中国海洋大学 Sea surface temperature prediction method, system and application based on time-space information interaction fusion
CN115307780B (en) * 2022-09-29 2023-01-06 中国海洋大学 Sea surface temperature prediction method, system and application based on time-space information interaction fusion
CN116822710A (en) * 2023-05-24 2023-09-29 国家海洋环境预报中心 Coral reef whitening hot spot prediction method, calcification rate prediction method and electronic equipment
CN116933666A (en) * 2023-09-19 2023-10-24 深圳康普盾科技股份有限公司 Thermal management optimization method, system and medium for container energy storage system
CN116933666B (en) * 2023-09-19 2023-12-26 深圳康普盾科技股份有限公司 Thermal management optimization method, system and medium for container energy storage system

Similar Documents

Publication Publication Date Title
CN112099110B (en) Ocean internal wave forecasting method based on machine learning and remote sensing data
CN114399073A (en) Ocean surface temperature field prediction method based on deep learning
CN110442666B (en) Mineral resource prediction method and system based on neural network model
CN102222313B (en) Urban evolution simulation structure cell model processing method based on kernel principal component analysis (KPCA)
CN114417693A (en) Ocean three-dimensional temperature field inversion method based on deep learning
CN115393540A (en) Intelligent fusion method and system of three-dimensional marine environment field based on deep learning
CN110516856B (en) Method for estimating ocean subsurface temperature based on convolutional neural network
CN115730684A (en) Air quality detection system based on LSTM-CNN model
Contractor et al. Efficacy of Feedforward and LSTM Neural Networks at predicting and gap filling coastal ocean timeseries: Oxygen, nutrients, and temperature
CN115204032A (en) ENSO prediction method and device based on multi-channel intelligent model
CN201716727U (en) Geographical simulation system based on remote sensing and GIS
González‐Abad et al. Using explainability to inform statistical downscaling based on deep learning beyond standard validation approaches
Fu et al. Simulation and forecasting of fishery weather based on statistical machine learning
CN113590971A (en) Interest point recommendation method and system based on brain-like space-time perception characterization
CN117253344A (en) Seawater acidification early warning and forecasting method, system and electronic equipment
CN116822381A (en) Marine temperature salt structure inversion method based on artificial intelligence
CN112668606A (en) Step type landslide displacement prediction method based on gradient elevator and quadratic programming
CN116522648A (en) Lake algae state prediction method and application
Lyu et al. A deep information based transfer learning method to detect annual urban dynamics of Beijing and Newyork from 1984–2016
CN115545159A (en) Average sea surface temperature forecasting method of deep neural network
CN114782740A (en) Remote sensing water quality monitoring method combining genetic optimization and extreme gradient promotion
Krapivin et al. An Expert Systems for the Aquatic Systems Investigation
CN111325384B (en) NDVI prediction method combining statistical characteristics and convolutional neural network model
Xiang et al. Simulation of Remote-Sensed Chlorophyll Concentration with a Coupling Model Based on Numerical Method and CA-SVM in Bohai Bay, China
Li et al. Deepphysinet: Bridging deep learning and atmospheric physics for accurate and continuous weather modeling

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination