Marine heatwaves are extreme warm water events that can cause devastating impacts on ecosystems and have complex socio-economic ramifications. Surface signals and drivers of marine heatwaves have been extensively investigated based on satellite observations, whereas their vertical structure in the global ocean remains unclear. In this study, we identify marine heatwave events in the epipelagic zone (0–200 m) using a four-dimensional spatio-temporal framework based on three ocean reanalysis datasets. We find that only about half of the marine heatwave events have continuous surface signals during their life cycles and around one-third always reside in the subsurface ocean without any imprint on sea surface temperature. The annual number of these subsurface marine heatwave events shows a significant increase in response to subsurface mean-state warming during the past three decades. Our findings reveal the limitation of identifying marine heatwaves solely based on the sea surface temperature and underscore the necessity of subsurface observations for monitoring marine heatwaves.

Jinyuan Li, Han Li, Zhuo Pan, Di Sun, Jiahao Wang, Wenkun Zhang, Gang Pan. Findings of the Association for Computational Linguistics: EMNLP 2023. 2023.

Jinyuan Li, Han Li, Di Sun, Jiahao Wang, Wenkun Zhang, Zan Wang, Gang Pan. Findings of the Association for Computational Linguistics ACL 2024. 2024.

Grounded Multimodal Named Entity Recognition (GMNER) task aims to identify named entities, entity types and their corresponding visual regions. GMNER task exhibits two challenging attributes: 1) The tenuous correlation between images and text on social media contributes to a notable proportion of named entities being ungroundable. 2) There exists a distinction between coarse-grained noun phrases used in similar tasks (e.g., phrase localization) and fine-grained named entities. In this paper, we propose RiVEG, a unified framework that reformulates GMNER into a joint modeling paradigm spanning MNER, VE, and VG perspectives by leveraging large language models (LLMs) as connecting bridges. This reformulation brings two benefits: 1) It enables us to optimize the MNER module for optimal MNER performance and eliminates the need to pre-extract region features using object detection methods, thus naturally addressing the two major limitations of existing GMNER methods. 2) The introduction of Entity Expansion Expression module and Visual Entailment (VE) module unifies Visual Grounding (VG) and Entity Grounding (EG). This endows the proposed framework with unlimited data and model scalability. Furthermore, to address the potential ambiguity stemming from the coarse-grained bounding box output in GMNER, we further construct the new Segmented Multimodal Named Entity Recognition (SMNER) task and corresponding Twitter-SMNER dataset aimed at generating fine-grained segmentation masks, and experimentally demonstrate the feasibility and effectiveness of using box prompt-based Segment Anything Model (SAM) to empower any GMNER model with the ability to accomplish the SMNER task. Extensive experiments demonstrate that RiVEG significantly outperforms SoTA methods on four datasets across the MNER, GMNER, and SMNER tasks. Datasets and Code will be released at https://github.com/JinYuanLi0012/RiVEG.