View Abstract

Version: 1.2 Status: Approved A new version will be created automatically if this content is modified.
Fábio L. Matos, CESAM, University of Aveiro, Portugal
Steve W. Ross, University of North Carolina, Wilmington, US
Veerle A.I. Huvenne, National Oceanography Centre, Southampton, UK
Jaime S. Davies, Plymouth University, UK
Marina R. Cunha, CESAM, University of Aveiro, Portugal
Exploring the submarine canyons' scientific landscape
Submarine canyons are among the most studied geomorphologic structures in the deep sea. Around 9500 canyons have been identified worldwide to date. In a fraction of those, numerous studies across disciplines have been conducted, covering several structural and functional aspects of their nature and revealing their remarkable influence on the surrounding areas. Despite the vast literature published, and the valuable information that the study of past and present knowledge trends can yield, the canyons’ scientific landscape delimitation through ‘knowledge clusters’ is still missing. Most scientific and technical literature is available in electronic format, allowing its analysis by powerful text retrieval techniques. Text mining offers powerful tools to organize, classify, label and retrieve previously unknown data patterns from large text collections. The INCISE canyons database, recently developed by Working Group 2, was structured into a textual corpus and updated with the corresponding abstracts to extract the most relevant terms in literature. Extraction of terms (single- and multiword) was carried out using natural language processing (NLP) methods, and the subsequent epistemic analyses were performed based on the frequency of terms occurrence. All the NLP tasks and co-word analyses were performed using the tools available in the open access platform CORTEXT Manager. In addition, the Gephi program was used to map the canyons scientific landscape while complementary analyses and data visualization processing were run using R software. In this study we addressed three main issues: 1) determine the main knowledge clusters of canyon research by scientific subject area and how they are interconnected; 2) identify which are the canyons and scientific subject categories well-covered in published literature and where are the main knowledge gaps in canyon research; and 3) reveal the dynamics and historic evolution of the canyons' science landscape. Although language ambiguity in multidisciplinary co-word analysis may be criticized, these studies provide systematized and useful information to scientists and science managers. The scientific landscape mapping and its complementary results will be available online in the near future, as an open interactive platform that can be used by canyon stakeholders as a tool for better planning of future research, management actions, and funding allocation.
Theme 2: New ways to study submarine canyons: integrated programs, new technologies and coordinated monitoring efforts
Oral Presentation
Scientific landscape, Knowledge clusters, Co-word analyses, Network analysis