| 103 | 0 | 151 |
| 下载次数 | 被引频次 | 阅读次数 |
通过在Argo浮标温度距平数据分析中引入多项约束条件，提出了一种基于多约束变分融合的海表温度数据重构方法，以此提升海表温度(Sea surface temperature, SST)的重构精度。首先，在Argo浮标观测点提取温度数据，并进行距平处理，得到离散的温度距平数据。然后，采用基于多约束条件的二维变分(Multi-constraint 2-dimensional variational, MC2D-Var)方法生成连续距平场，并在同化过程中分别引入海表面高度(Sea surface height, SSH)约束、东向海水流速和北向海水流速(Eastward sea water velocity and nnorthward sea water velocity, UV)产生的平流约束以及平流+SSH联合约束三种方案，以优化连续距平场生成效果。最后，将2018年世界海洋地图集(World ocean atlas 2018, WOA18)气候态SST叠加到连续温度距平场中，从而实现海表温度的重构。以2023年6月全球SST数据为例，采用随机抽样方式将Argo观测数据划分为训练集(70%)和验证集(30%),并基于三种约束方案生成0.25°×0.25°分辨率的SST场。验证结果表明，相较于逐日最优插值海表温度(Optimum interpolation sea surface temperature, OISST)数据(均方根误差(Root mean square error, RMSE)为0.902℃),本文提出的三种方案重构效果均有提升，其中方案1(SSH约束)和方案2(平流约束)的RMSE分别降至0.866和0.891℃,而方案3(联合约束)表现最佳，RMSE降至0.864℃。此外，对2023年1—5月的重构实验进一步验证了方案3的稳定性，与OISST相比，其RMSE分别降低5.55%、2.97%、7.26%、4.91%和11.46%,相较于WOA18数据，精度提升幅度分别达49.15%、42.09%、44.56%、47.30%和47.21%。结果表明，结合多重物理约束能够有效提升SST重构精度，为高精度海表温度重构提供了新的技术途径。
Abstract：By introducing multiple constraint conditions into the analysis of Argo buoy temperature anomaly data, a sea surface temperature(SST) data reconstruction method based on multi-constraint variational fusion is proposed to enhance the reconstruction accuracy of the SST field. First, temperature data are extracted at Argo buoy observation points and processed to obtain discrete temperature anomaly data. Then, a multi-constraint 2-dimensional variational(MC2D-Var) method is employed to generate a continuous anomaly field. During assimilation, three different schemes are implemented: sea surface height(SSH) constraint, advection constraint induced by eastward and northward sea water velocity(UV), and a combined advection+SSH constraint to optimize the generation of the continuous anomaly field. Finally, the WOA18 climatological SST field is superimposed onto the continuous temperature anomaly field to reconstruct the SST field. Using global SST data from June 2023 as an example, the Argo observation data were randomly divided into a training set(70%) and a validation set(30%). Based on the three constraint schemes, SST fields were generated with a resolution of 0.25°×0.25°. Validation results show that compared to OISST data(RMSE of 0.902 ℃), all schemes achieved improvements. Specifically, Scheme 1(SSH constraint) and Scheme 2(advection constraint) reduced RMSE to 0.866 ℃ and 0.891 ℃, respectively, while Scheme 3(combined constraint) performed best, lowering RMSE to 0.864 ℃. Additionally, reconstruction experiments from January to May further validated the stability of Scheme 3. Compared to OISST, its RMSE was reduced by 5.55%, 2.97%, 7.26%, 4.91% and 11.46%, respectively, while its accuracy improvement over WOA18 data reached 49.15%, 42.09%, 44.56%, 47.30% and 47.21%, respectively. These results indicate that incorporating multiple physical constraints can effectively enhance SST reconstruction accuracy, providing a new technical approach for high-precision sea temperature field construction.
[1]O’Carroll A G,Armstrong E M,Beggs H M,et al.Observational needs of sea surface temperature[J].Frontiers in Marine Science,2019(6):00420.
[2]Melet A,Teatini P,Le Cozannet G,et al.Earth observations for monitoring marine coastal hazards and their drivers[J].Surveys in Geophysics,2020,41(6):1-46.
[3]Wang X,Lu Q,Zhong S,et al.Differential impacts of the 2015-2020 El Niño/El Niño Modoki on seasonal ozone levels across China[J].Atmospheric Pollution Research,2025,16(5):102449.
[4]Zhang M,Cheng Y,Zhang H,et al.Spatiotemporal variability of air-sea CO2 fluxes in response to El Niño-related marine heatwaves in the tropical Pacific Ocean[J].Marine Environmental Research,2025,204:106949.
[5]Ruby S,Dibakar G.Analysis of pre-El Niño and La Niña events using climate network approach[J].Chaos,Solitons and Fractals,2025,191:115781.
[6]Mike C,Keith H.Altimetric assimilation with water property conservation[J].Journal of Geophysical Research:Oceans,1996,101(1):1059-1077.
[7]Paul C.Fiedler,surface manifestations of subsurface thermal structure in the California Current[J].Journal of Geophysical Research:Oceans,1988,93(5):4975-4983.
[8]王喜冬,韩桂军,李威,等利用卫星观测海面信息反演三维温度场[J].热带海洋学报,2011,30(6):10-17.Wang D,Han G J,Li W,et al.Inversion of three-dimensional temperature fields using satellite-observed sea surface information[J].Journal of Tropical Oceanography,2011,30(6):10-17.
[9]Ali M,Swain D,Weller A R.Estimation of ocean subsurface thermal structure from surface parameters:neural network approach[J].Geophysical Research Letters,2004,31(20):L20308.
[10]Wu B,Yan X H,Young-Heon J,et al.Estimation of subsurface temperature anomaly in the North Atlantic using selforganizing map neural network[J].Journal of Atmospheric and Oceanic Technology,2012,29(11):1675-1688.
[11]Fujii Y,Kamachi M.reconstruction of observed profiles in the sea east of Japan using vertical coupled temperature-salinity EOFmodes[J].Journal of Oceanography,2003,59(2):173-186.
[12]Helber R,Townsend L,Barron C,et al.Validation test report for the improved synthetic ocean profile (ISOP) system,partⅠ:Synthetic profile methods and algorithm[J].Naval Research Laboratory,2013,7320-13-9364.
[13]He K,Wang X D,Wu X R,et al.Projecting three-dimensional ocean thermohaline structure in the North Indian Ocean from the satellite sea surface data based on variational method[J].Journal of Geophysical Research:Oceans,2021,126(1):e2020JC016759.
[14]高志斌,王云涛,柴扉基于海表温度遥感资料重构的热带不稳定波特征分析[J].厦门大学学报(自然科学版),2024,63(3):512-525.Gao B,Wang Y T,Chai F.Analysis of tropical instability wave characteristics reconstructed from sea surface temperature remote sensing data[J].Journal of Xiamen University (Natural Science),2024,63(3):512-525.
[15]Yan Z,Yingying G,Junxiao L.Research on Sea Surface temperature Reconstruction from long-term MODIS data[J].IOP Conference Series:Earth and Environmental Science,2021,631(1):012032.
[16]范冬林,杨鑫,曾优,等Himawari-8卫星云下海表温度反演的机器学习方法比较研究[J].遥感信息,2023,38(5):39-48.Fan L,Yang X,Zeng Y,et al.Comparative study of machine learning methods for sea surface temperature retrieval under cloudy conditions using Himawari-8 satellite data[J].Remote Sensing Information,2023,38(5):39-48.
[17]Chao Y,Li Z,Farrara D,et al.Blending sea surface temperatures from multiple satellites and in situ observations for coastal oceans[J].Journal of Atmospheric and Oceanic Technology,2009,26(7):1415-1426.
[18]Chen L,Zhang H,Gao C.Effects of temperature and salinity on surface currents in the equatorial pacific[J].Journal of Geophysical Research:Oceans,2022,127(4):e2021JC018175.
[19]Marcelja S.The timescale and extent of thermal expansion of the global ocean due to climate change[J].Ocean Science,2010,6(23):179-184.
[20]Dimet F,马怀存.气象观测资料分析和同化的变分方法:理论部分[J].气象科技,1987(5):14-22.Dimet F,Ma H C.Variational methods for the analysis and assimilation of meteorological observations:Theoretical aspects[J].Meteorological Science and Technology,1987(5):14-22.
[21]Philippe L,Angela B,Peter B,et al.Experimental 2D-Var assimilation of ARM cloud and precipitation observations[J].Quarterly Journal of the Royal Meteorological Society,2006,132(617):1325-1347.
[22]Lv S R,Lin W M,Wang Z X,et al.Blending sea surface winds from the HY-2 satellite scatterometers based on 2D-Var method[J].Remote Sensing,2022,15(1):193.
[23]Barth A,Beckers M,Troupin C,et al.DIVAnd-1.0:n-dimensional variational data analysis for ocean observations[J].Geoscientific Model Development,2014,7(1):225-241.
[24]Tan T,Cheng L J,Gouretski V,et al.A new automatic quality control system for ocean profile observations and impact on ocean warming estimate[J].Deep-Sea Research Part I,2023,194:103961.
[25]Huang B,Liu C,Banzon E,et al.Improvements of the daily optimum interpolation sea surface temperature (DOISST) version2.1[J].Journal of Climate,2020,34:2923-2939.
基本信息:
DOI：10.16441/j.cnki.hdxb.20250123
中图分类号:P715.2;P714.1
引用信息:
[1]王敏,谷文杰,郭晓峰,等.基于多约束变分融合的海表温度数据重构[J].中国海洋大学学报(自然科学版),2026,56(05):1-10.DOI:10.16441/j.cnki.hdxb.20250123.
基金信息:
国家自然科学基金项目(41775165,41775039); 安徽省高校杰出青年科研项目(2023AH020022)资助~~
2025-04-08
2025
2025-04-29
2025-05-08
2025
1
2026-04-27
2026-04-27