The Identification and Prediction of Mesoscale Eddy Variation via Memory in Memory With Scheduled Sampling for Sea Level Anomaly
Ocean mesoscale eddies are ubiquitous in world ocean and account for 90% oceanic kinetic energy, which dominate the upper ocean flow field. Accurately predicting the variation of ocean mesoscale eddies is the key to understand the oceanic flow field and circulation system. In this article, we propos...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2021-12-01
|
| Series: | Frontiers in Marine Science |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fmars.2021.753942/full |
| _version_ | 1818421534866276352 |
|---|---|
| author | Rui Nian Yu Cai Zhengguang Zhang Hui He Jingyu Wu Qiang Yuan Xue Geng Yuqi Qian Hua Yang Bo He |
| author_facet | Rui Nian Yu Cai Zhengguang Zhang Hui He Jingyu Wu Qiang Yuan Xue Geng Yuqi Qian Hua Yang Bo He |
| author_sort | Rui Nian |
| collection | DOAJ |
| description | Ocean mesoscale eddies are ubiquitous in world ocean and account for 90% oceanic kinetic energy, which dominate the upper ocean flow field. Accurately predicting the variation of ocean mesoscale eddies is the key to understand the oceanic flow field and circulation system. In this article, we propose to make an initial attempt to explore spatio-temporal predictability of mesoscale eddies, employing deep learning architecture, which primarily establishes Memory In Memory (MIM) for sea level anomaly (SLA) prediction, combined with the existing mesoscale eddy detection. Oriented to the western Pacific ocean (125°−137.5°E and 15°−27.5°N), we quantitatively investigate the historic daily SLA variability at a 0.25° spatial resolution from 2000 to 2018, derived by satellite altimetry. We develop the enhanced MIM prediction strategies, equipped with Gated Recurrent Unit (GRU) and spatial attention module, in a scheduled sampling manner, which overcomes the gradient vanishing and complements to strengthen spatio-temporal features for long-term dependencies. At the early stage, the real value SLA input guides the model training process for initialization, while the scheduled sampling intentionally feeds the newly predicted value, to resolve the distribution inconsistency of inference. It has been demonstrated in our experiment results that our proposed prediction scheme outperformed the state-of-art approaches for SLA time series, with MAPE, RMSE of the 14-day prediction duration, respectively, 5.1%, 0.023 m on average, even up to 4.6%, 0.018 m for the effective sub-regions, compared to 19.8%, 0.086 m in ConvLSTM and 8.3%, 0.040 m in original MIM, which greatly facilitated the mesoscale eddy prediction. This proposed scheme will be beneficial to understand of the underlying dynamical mechanism behind the predictability of mesoscale eddies in the future, and help the deployment of ARGO, glider, AUV and other observational platforms. |
| first_indexed | 2024-12-14T13:11:54Z |
| format | Article |
| id | doaj.art-689b6b65756c4bcf96dce9ed7b9da06e |
| institution | Directory Open Access Journal |
| issn | 2296-7745 |
| language | English |
| last_indexed | 2024-12-14T13:11:54Z |
| publishDate | 2021-12-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Marine Science |
| spelling | doaj.art-689b6b65756c4bcf96dce9ed7b9da06e2022-12-21T23:00:11ZengFrontiers Media S.A.Frontiers in Marine Science2296-77452021-12-01810.3389/fmars.2021.753942753942The Identification and Prediction of Mesoscale Eddy Variation via Memory in Memory With Scheduled Sampling for Sea Level AnomalyRui Nian0Yu Cai1Zhengguang Zhang2Hui He3Jingyu Wu4Qiang Yuan5Xue Geng6Yuqi Qian7Hua Yang8Bo He9School of Electronic Engineering, Ocean University of China, Qingdao, ChinaChina Unicom Software Research Institute, Beijing, ChinaPhysical Oceanography Laboratory, Ocean University of China, Qingdao, ChinaSchool of Electronic Engineering, Ocean University of China, Qingdao, ChinaSchool of Electronic Engineering, Ocean University of China, Qingdao, ChinaSchool of Electronic Engineering, Ocean University of China, Qingdao, ChinaSchool of Electronic Engineering, Ocean University of China, Qingdao, ChinaSchool of Electronic Engineering, Ocean University of China, Qingdao, ChinaSchool of Electronic Engineering, Ocean University of China, Qingdao, ChinaSchool of Electronic Engineering, Ocean University of China, Qingdao, ChinaOcean mesoscale eddies are ubiquitous in world ocean and account for 90% oceanic kinetic energy, which dominate the upper ocean flow field. Accurately predicting the variation of ocean mesoscale eddies is the key to understand the oceanic flow field and circulation system. In this article, we propose to make an initial attempt to explore spatio-temporal predictability of mesoscale eddies, employing deep learning architecture, which primarily establishes Memory In Memory (MIM) for sea level anomaly (SLA) prediction, combined with the existing mesoscale eddy detection. Oriented to the western Pacific ocean (125°−137.5°E and 15°−27.5°N), we quantitatively investigate the historic daily SLA variability at a 0.25° spatial resolution from 2000 to 2018, derived by satellite altimetry. We develop the enhanced MIM prediction strategies, equipped with Gated Recurrent Unit (GRU) and spatial attention module, in a scheduled sampling manner, which overcomes the gradient vanishing and complements to strengthen spatio-temporal features for long-term dependencies. At the early stage, the real value SLA input guides the model training process for initialization, while the scheduled sampling intentionally feeds the newly predicted value, to resolve the distribution inconsistency of inference. It has been demonstrated in our experiment results that our proposed prediction scheme outperformed the state-of-art approaches for SLA time series, with MAPE, RMSE of the 14-day prediction duration, respectively, 5.1%, 0.023 m on average, even up to 4.6%, 0.018 m for the effective sub-regions, compared to 19.8%, 0.086 m in ConvLSTM and 8.3%, 0.040 m in original MIM, which greatly facilitated the mesoscale eddy prediction. This proposed scheme will be beneficial to understand of the underlying dynamical mechanism behind the predictability of mesoscale eddies in the future, and help the deployment of ARGO, glider, AUV and other observational platforms.https://www.frontiersin.org/articles/10.3389/fmars.2021.753942/fullMemory In Memoryscheduled samplingmesoscale eddysea level anomalyspatio-temporal predictiongated recurrent unit |
| spellingShingle | Rui Nian Yu Cai Zhengguang Zhang Hui He Jingyu Wu Qiang Yuan Xue Geng Yuqi Qian Hua Yang Bo He The Identification and Prediction of Mesoscale Eddy Variation via Memory in Memory With Scheduled Sampling for Sea Level Anomaly Frontiers in Marine Science Memory In Memory scheduled sampling mesoscale eddy sea level anomaly spatio-temporal prediction gated recurrent unit |
| title | The Identification and Prediction of Mesoscale Eddy Variation via Memory in Memory With Scheduled Sampling for Sea Level Anomaly |
| title_full | The Identification and Prediction of Mesoscale Eddy Variation via Memory in Memory With Scheduled Sampling for Sea Level Anomaly |
| title_fullStr | The Identification and Prediction of Mesoscale Eddy Variation via Memory in Memory With Scheduled Sampling for Sea Level Anomaly |
| title_full_unstemmed | The Identification and Prediction of Mesoscale Eddy Variation via Memory in Memory With Scheduled Sampling for Sea Level Anomaly |
| title_short | The Identification and Prediction of Mesoscale Eddy Variation via Memory in Memory With Scheduled Sampling for Sea Level Anomaly |
| title_sort | identification and prediction of mesoscale eddy variation via memory in memory with scheduled sampling for sea level anomaly |
| topic | Memory In Memory scheduled sampling mesoscale eddy sea level anomaly spatio-temporal prediction gated recurrent unit |
| url | https://www.frontiersin.org/articles/10.3389/fmars.2021.753942/full |
| work_keys_str_mv | AT ruinian theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT yucai theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT zhengguangzhang theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT huihe theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT jingyuwu theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT qiangyuan theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT xuegeng theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT yuqiqian theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT huayang theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT bohe theidentificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT ruinian identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT yucai identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT zhengguangzhang identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT huihe identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT jingyuwu identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT qiangyuan identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT xuegeng identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT yuqiqian identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT huayang identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly AT bohe identificationandpredictionofmesoscaleeddyvariationviamemoryinmemorywithscheduledsamplingforsealevelanomaly |