Sayantan Sengupta: During the last half of the 20th century, coastal ecosystems were subjected to different types of anthropogenic and climatic pressure such as nutrient enrichment, exploitation of coastal resources, overfishing, and destruction of habitats, chemical pollution and physical changes.
These phenomena have substantial effects on ecological system and human life. Coastal eutrophication, resulting from increased nutrient input, has been identified as the main driver of the deterioration of coastal ecosystems in Europe, North America, Asia and Oceania. Eutrophication has resulted in higher phytoplankton production, blooms of opportunistic algae, decreased light penetration, loss of underwater macrophytes and increased occurrence of hypoxia and anoxia. The first action plans in Denmark focused on reduction of nitrogen and phosphorus inputs to Danish waters to understand its effect on ecological status. Therefore, from the early 1990s, many efforts were made to find robust quantitative relations between nutrient loading and environmental quality.
Based on the EU water framework directive (WFD), the environmental quality or ecological status of aquatic ecosystems is categorized into five classes (Bad, Poor, Moderate, Good and High) and the goal of action plans is to achieve at least good ecological status in Danish water bodies. Today, the action plans aim at achieving good ecological status using nutrient reduction as the main measure for achieving this objective.
Project description
The SeaStatus project is a partnership formed by Department of Applied Mathematics and Computer Science at the Technical University of Denmark, the Department of Bioscience at Aarhus University, DHI, COWI, Rambøll, SVANA and the Danish Road Directorate. The project is part of work‐package‐1 (W.P1) of the SeaStatus collaboration. My PhD project concerns data assimilation, and
should provide the foundation for combining novel and traditional measurement techniques with ecosystem modelling, to improve the information basis for management, through establishing new routines for construction, standardization, integration and processing of large and differentiated data sets extracted from embedded information, for integration into model‐based decision support tools. The outcome of my project is to build an automated surveillance tool. The information comes in different forms, such as satellite images, remote sensing, ferry‐box data and underwater videos. As a starting point, I am planning to detect Eelgrass(a special kind of vegetation in the sea bed and is a good indicator of the excess nutrient presence in the water) from the underwater video files using image processing and Machine learning techniques. The next step is the analysing satellite images of the same region and use segmentation algorithms to find the regions with Eelgrass and keep a track of their growth over time and space and form some inference which will lead to further actions. There are many other indicators which shall lead to different inferences. Finally combining all of these inferences in a common framework is the goal of my thesis.
PhD project title: Big Data Processing and shaping in SeaStatus
Effective start/end date 15/08/2017 → 14/08/2020
DTU Supervisors: Anders Stockmarr and Bjarne Ersbøll from the section for Statistics and Data Analysis