-
Abstract: Marine Heatwave (MHW) events refer to periods of significantly elevated sea surface temperatures (SST), persisting from days to months, with significant impacts on marine ecosystems, including increased mortality among marine life and coral bleaching. Forecasting MHW events is crucial to mitigate their harmful effects. This study presents a two-step forecasting process: short-term SST prediction followed by MHW event detection based on the forecasted SST. Firstly, we developed the “SST-MHW-DL” model using the ConvLSTM architecture, which incorporates an attention mechanism to enhance both SST forecasting and MHW event detection. The model utilizes SST data from the preceding 60 days to forecast SST and detect MHW events for the subsequent 15 days. Verification results for SST forecasting demonstrate a Root Mean Square Error (RMSE) of 0.64℃, a Mean Absolute Percentage Error (MAPE) of 2.05%, and a coefficient of determination (R²) of 0.85, indicating the model’s ability to accurately predict future temperatures by leveraging historical sea temperature information. For MHW event detection using forecasted SST, the evaluation metrics of “accuracy,” “precision,” and “recall” achieved values of 0.77, 0.73, and 0.43, respectively, demonstrating the model’s capability to capture the occurrence of MHW events accurately. Furthermore, the attention-enhanced mechanism reveals that recent SST variations within the past 10 days have the most significant impact on forecasting accuracy, while variations in deep-sea regions and along the Taiwan Strait significantly contribute to the model’s efficacy in capturing spatial characteristics. Additionally, the proposed model and temporal mechanism were applied to detect MHWs in the Atlantic Ocean. By inputting 30 days of SST data, the model predicted SST with an RMSE of 1.02℃ and an R² of 0.94. The accuracy, precision, and recall for MHW detection were 0.79, 0.78, and 0.62, respectively, further demonstrating the model’s robustness and usability.
-
Figure 1. Study area. (a) Topography of the ECS and its primary ocean currents, (b) SST variations in the ECS from 2016 to 2021.
Figure 3. Structural diagram of the proposed SST-MHW-DL model (the red box highlights our improvement and analysis part).
Figure 6. Comparison between the SST-MHW-DL model and several state-of-the-art models. (a) Comparison of study areas, (b) accuracy comparison for different forecast durations.
Figure 7. Spatiotemporal variations in SST forecast errors by the SST-MHW-DL model in the ECS region. (a) Difference and (b) RMSE of the predicted SST and ground truth.
Figure 9. Visualization of channel (temporal) attention weights in the SST-MHW-DL model.
Figure 10. Spatial attention in the SST-MHW-DL model. (a) Weight distribution of the SST-MHW-DL model, (b) spring and (c) autumn temperature.
Figure 11. Detection of MHW with different occurrence scales of 25%, 50%, and 75% in the ECS. TP and TN denote correctly predicted MHW sea areas.
Figure 12. Spatiotemporal statistics of MHW occurrence in the ECS from 2016 to 2021. (a) average intensity (℃), (b) occurrence duration of the MHW events (days), and (c) annual average occurrence frequency (counts/year).
Figure 13. Test area. (a) Topography of the NWA, (b) SST variations in the NWA from 2016 to 2021.
Figure 14. Spatiotemporal statistics of MHW occurrence in the NWA from 2016 to 2021. (a) average intensity (℃), (b) occurrence duration of the MHW events (days), (c) annual average occurrence frequency (counts/year).
Figure 16. Spatial distribution of (a) RMSE (℃) and (b) R² predicted by SST, as well as (c) the effectiveness evaluation of the model’s predictions.
Table 1. Comparison of the proposed SST-MHW-DL with basic ConvLSTM, U-Net, and Multi-attention ConvLSTM models.
Model RMSE/℃ MAPE/% R2 ConvLSTM 0.67 2.13 0.846 U-Net 0.73 2.35 0.823 Multi-Attention ConvLSTM 0.66 2.14 0.850 SST-MHW-DL 0.64 2.05 0.854 
下载: 导出CSV
Table 2. Diffusion matrix for MHW detection in the ECS.
MHWs predicted MHWs not predicted MHWs observed 614912 (TP)811248 (FN)MHWs not
observed224133 (FP)2837509 (TN）
下载: 导出CSV
Table 3. Diffusion matrix for MHW detection in the NWA.
MHWs predicted MHWs not predicted MHWs observed 2,063,321 (TP) 1,284,398 (FN) MHWs not observed 589,214 (FP) 811,248 (TN）
下载: 导出CSV
-
Bao Baoleerqimuge, Ren Guoyu. 2014. Climatological characteristics and long-term change of SST over the marginal seas of China. Continental Shelf Research, 77: 96–106, doi: 10.1016/j.csr.2014.01.013 Bian Ce, Jing Zhao, Wang Hong, et al. 2023. Oceanic mesoscale eddies as crucial drivers of global marine heatwaves. Nature Communications, 14(1): 2970, doi: 10.1038/s41467-023-38811-z Cao Lu, Tang Rui, Huang Wei, et al. 2021. Seasonal variability and dynamics of coastal sea surface temperature fronts in the East China Sea. Ocean Dynamics, 71(2): 237–249, doi: 10.1007/s10236-020-01427-8 Cao Lu, Tang Rui, Huang Wei, et al. 2021. Seasonal variability and dynamics of coastal sea surface temperature fronts in the East China Sea. Ocean Dynamics, 71(2): 237–249, doi: 10.1007/s10236-020-01427-8 (查阅网上资料,本条文献和标黄文献重复,请核对) Chen Zhuomin, Kwon Y O, Chen Ke, et al. 2020. Long-term SST variability on the Northwest Atlantic continental shelf and slope. Geophysical Research Letters, 47(1): e2019GL085455, doi: 10.1029/2019GL085455 Claar D C, Wood C L. 2020. Pulse heat stress and parasitism in a warming world. Trends in Ecology & Evolution, 35(8): 704–715, doi: 10.1016/j.tree.2020.04.002 De Burgh-Day C O, Spillman C M, Smith G, et al. 2022. Forecasting extreme marine heat events in key aquaculture regions around New Zealand. Journal of Southern Hemisphere Earth Systems Science, 72(1): 58–72, doi: 10.1071/es21012 Fan Luyi, Cao Yuhao, Huang Ningyuan, et al. 2024. OTCFM: a sea surface temperature prediction method integrating multi-scale periodic features. IEEE Access, 12: 108291–108302, doi: 10.1109/ACCESS.2024.3425514 Fredston A L, Cheung W W L, Frölicher T L, et al. 2023. Marine heatwaves are not a dominant driver of change in demersal fishes. Nature, 621(7978): 324–329, doi: 10.1038/s41586-023-06449-y Frölicher T L, Fischer E M, Gruber N. 2018. Marine heatwaves under global warming. Nature, 560(7718): 360–364, doi: 10.1038/s41586-018-0383-9 Gao Guandong, Marin M, Feng Ming, et al. 2020. Drivers of marine heatwaves in the East China Sea and the South Yellow Sea in three consecutive summers during 2016–2018. Journal of Geophysical Research: Oceans, 125(8): e2020JC016518, doi: 10.1029/2020JC016518 Giamalaki K, Beaulieu C, Prochaska J X. 2022. Assessing predictability of marine heatwaves with random forests. Geophysical Research Letters, 49(23): e2022GL099069, doi: 10.1029/2022gl099069 Han Mingxu, Feng Yuan, Zhao Xueli, et al. 2019. A convolutional neural network using surface data to predict subsurface temperatures in the Pacific Ocean. IEEE Access, 7: 172816–172829, doi: 10.1109/ACCESS.2019.2955957 Hobday A J, Alexander L V, Perkins S E, et al. 2016. A hierarchical approach to defining marine heatwaves. Progress in Oceanography, 141: 227–238, doi: 10.1016/j.pocean.2015.12.014 Hobday A J, Spillman C M, Eveson J P, et al. 2018. A framework for combining seasonal forecasts and climate projections to aid risk management for fisheries and aquaculture. Frontiers in Marine Science, 5: 137, doi: 10.3389/fmars.2018.00137 Holbrook N J, Scannell H A, Sen Gupta A, et al. 2019. A global assessment of marine heatwaves and their drivers. Nature Communications, 10(1): 2624, doi: 10.1038/s41467-019-10206-z Hövel L, Brune S, Baehr J. 2022. Decadal prediction of marine heatwaves in MPI-ESM. Geophysical Research Letters, 49(15): e2022GL099347, doi: 10.1029/2022GL099347 Hu Shijian, Li Shihan. 2022. Progress and prospect of marine heatwave study. Advances in Earth Science (in Chinese), 37(1): 51–64, doi: 10.11867/j.issn.10018166.2021.121 Iskandar M R, Ismail M F A, Arifin T, et al. 2021. Marine heatwaves of sea surface temperature off south Java. Heliyon, 7(12): e08618, doi: 10.1016/j.heliyon.2021.e08618 Jahanbakht M, Xiang Wei, Azghadi M R. 2022. Sea surface temperature forecasting with ensemble of stacked deep neural networks. IEEE Geoscience and Remote Sensing Letters, 19: 1502605, doi: 10.1109/lgrs.2021.3098425 Kako S, Nakagawa T, Takayama K, et al. 2016. Impact of Changjiang River discharge on sea surface temperature in the East China Sea. Journal of Physical Oceanography, 46(6): 1735–1750, doi: 10.1175/jpo-d-15-0167.1 Lee M A, Huang W P, Shen Y L, et al. 2021. Long-term observations of interannual and decadal variation of sea surface temperature in the Taiwan Strait. Journal of Marine Science and Technology, 29(4): 7, doi: 10.51400/2709-6998.1587 Lee E Y, Park K A. 2021. Validation satellite sea surface temperature in the coastal regions. In: Proceedings of 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS. Brussels, Belgium: IEEE,7607–7610, doi: 10.1109/IGARSS47720.2021.9553695 Lian Peng, Gao Le. 2024. Impacts of central-Pacific El Niño and physical drivers on eastern Pacific bigeye tuna. Journal of Oceanology and Limnology, 42(3): 972–987, doi: 10.1007/s00343-023-3051-3 Marin M, Feng Ming, Phillips H E, et al. 2021. A global, multiproduct analysis of coastal marine heatwaves: Distribution, characteristics, and long-term trends. Journal of Geophysical Research: Oceans, 126(2): e2020JC016708, doi: 10.1029/2020JC016708 Oliver E C J, Benthuysen J A, Darmaraki S, et al. 2021. Marine heatwaves. Annual Review of Marine Science, 13: 313–342, doi: 10.1146/annurev-marine-032720-095144 Oliver E C J, Donat M G, Burrows M T, et al. 2018. Longer and more frequent marine heatwaves over the past century. Nature Communications, 9(1): 1324, doi: 10.1038/s41467-018-03732-9 Pearce A, Lenanton R, Jackson G, et al. 2011. The “marine heat wave” off Western Australia during the summer of 2010/11. Fisheries Research Report No. 222, Western: Department of Fisheries, 40 Pershing A J, Mills K E, Dayton A M, et al. 2018. Evidence for adaptation from the 2016 marine heatwave in the northwest Atlantic Ocean. Oceanography, 31(2): 152–161, doi: 10.5670/oceanog.2018.213 Ren Yibin, Li Xiaofeng. 2023. Predicting the daily sea ice concentration on a subseasonal scale of the Pan-Arctic during the melting season by a deep learning model. IEEE Transactions on Geoscience and Remote Sensing, 61: 4301315, doi: 10.1109/TGRS.2023.3279089 Sasaki Y N, Umeda C. 2021. Rapid warming of sea surface temperature along the Kuroshio and the China coast in the East China Sea during the twentieth century. Journal of Climate, 34(12): 4803–4815, doi: 10.1175/JCLI-D-20-0421.1 Shanks A L, Rasmuson L K, Valley J R, et al. 2020. Marine heat waves, climate change, and failed spawning by coastal invertebrates. Limnology and Oceanography, 65(3): 627–636, doi: 10.1002/lno.11331 Shi Xingjian, Chen Zhourong, Wang Hao, et al. 2015. Convolutional LSTM network: a machine learning approach for precipitation nowcasting. arXiv: 1506.04214, 2015 (查阅网上资料, 未能确认文献类型, 请确认) Shi Benyun, Feng Liu, He Hailun, et al. 2024. A physics-guided attention-based neural network for sea surface temperature prediction. IEEE Transactions on Geoscience and Remote Sensing, 62: 4210413, doi: 10.1109/TGRS.2024.3457039 Smith K E, Burrows M T, Hobday A J, et al. 2021. Socioeconomic impacts of marine heatwaves: global issues and opportunities. Science, 374(6566): eabj3593, doi: 10.1126/science.abj3593 Smith K E, Burrows M T, Hobday A J, et al. 2023. Biological impacts of marine heatwaves. Annual Review of Marine Science, 15: 119–145, doi: 10.1146/annurev-marine-032122-121437 Tseng C, Lin Chiyuan, Chen Shihchin, et al. 2000. Temporal and spatial variations of sea surface temperature in the East China Sea. Continental Shelf Research, 20(4-5): 373–387, doi: 10.1016/S0278-4343(99)00077-1 Wang Haoyu, Hu Shineng, Guan Cong, et al. 2024. The role of sea surface salinity in ENSO forecasting in the 21st century. npj Climate and Atmospheric Science, 7(1): 206, doi: 10.1038/s41612-024-00763-6 Wang Haoyu, Li Xiaofeng. 2024. DeepBlue: advanced convolutional neural network applications for ocean remote sensing. IEEE Geoscience and Remote Sensing Magazine, 12(1): 138–161, doi: 10.1109/mgrs.2023.3343623 Wang Fan, Li Xuegang, Tang Xiaohui, et al. 2023a. The seas around China in a warming climate. Nature Reviews Earth & Environment, 4(8): 535–551, doi: 10.1038/s43017-023-00453-6 Wang Yunhe, Yuan Xiaojun, Ren Yibin, et al. 2023b. Subseasonal prediction of regional antarctic sea ice by a deep learning model. Geophysical Research Letters, 50(17): e2023GL104347, doi: 10.1029/2023gl104347 Wang Zihan, Zeng Cong, Cao Ling. 2023c. Mapping the biodiversity conservation gaps in the East China sea. Journal of Environmental Management, 336: 117667, doi: 10.1016/j.jenvman.2023.117667 Wei Yi, Ding Ruibin, Huang Daji, et al. 2023. The weakened upwelling at the Upstream Kuroshio in the East China Sea induced extensive sea surface warming. Geophysical Research Letters, 50(1): e2022GL101835, doi: 10.1029/2022gl101835 Woo S, Park J, Lee J Y, et al. 2018. CBAM: convolutional block attention module. In Proceedings of the 15th European Conference on Computer Vision. Munich, Germany: Springer,3–19 Xiao Changjiang, Chen Nengcheng, Hu Chuli, et al. 2019. A spatiotemporal deep learning model for sea surface temperature field prediction using time-series satellite data. Environmental Modelling & Software, 120: 104502, doi: 10.1016/j.envsoft.2019.104502 Yu Haiqing, Wang Hui, Yuan Chunxin, et al. 2024. Assessing the predictability of the marine heatwave in the Yellow Sea during the summer of 2018 based on a deterministic forecast model. Weather and Climate Extremes, 44: 100663, doi: 10.1016/j.wace.2024.100663 Zhang Kun, Geng Xupu, Yan Xiaohai. 2020. Prediction of 3-D ocean temperature by multilayer convolutional LSTM. IEEE Geoscience and Remote Sensing Letters, 17(8): 1303–1307, doi: 10.1109/LGRS.2019.2947170 Zhang Xiaoyu, Li Yongqing, Frery A C, et al. 2022. Sea surface temperature prediction with memory graph convolutional networks. IEEE Geoscience and Remote Sensing Letters, 19: 8017105, doi: 10.1109/lgrs.2021.3097329 Zhang Qin, Wang Hui, Dong Junyu, et al. 2017. Prediction of sea surface temperature using Long Short-Term Memory. IEEE Geoscience and Remote Sensing Letters, 14(10): 1745–1749, doi: 10.1109/LGRS.2017.2733548 -
点击查看大图
计量
- 文章访问数: 25
- HTML全文浏览量: 10
- 被引次数: 0
