Presenters
Source
Navigating the Digital Ocean: How DIGI4ECO is Revolutionizing Marine Monitoring 🌊💻
The vast, mysterious ocean holds secrets that are crucial for our planet’s health. But how do we truly understand and protect it when traditional monitoring methods are costly and limited? Enter DIGI4ECO, a groundbreaking project spearheaded by experts like Jacopo Aguzzi from the Marine Science Institute of Barcelona and Enoc Martinez from the Technical University of Catalonia. They’re pioneering the use of robotic networks and advanced data management to create a digital twin of the ocean, offering unprecedented insights into marine ecosystems. Let’s dive in! 🏊♀️
The Urgent Need for Ocean Monitoring 🚨
Our oceans are facing unprecedented pressures from human activities. Jacopo highlights two critical examples:
- Industrial Trawling: This destructive practice sweeps across continental margins daily, decimating entire ecosystems. 💥
- Plastic Pollution: The omnipresent threat of plastic waste continues to choke marine life and habitats. 🐢
These impacts necessitate robust monitoring to assess the Good Environmental Status (GES) of our marine environments. The European Community, through its Marine Strategy Framework Directive, aims to achieve this using 11 key descriptors or ecological indicators. These indicators, covering everything from seabed health to commercially exploited species, require a deep understanding of ecological and biological data across various spatial and temporal scales.
The High Cost of Ocean Exploration 💰🚢
Traditionally, collecting ocean data relies heavily on research vessels. Jacopo points out the staggering cost: a single vessel like the Sarmiento de Gamboa can cost up to 25,000 euros per day to operate. This financial burden, coupled with limited vessel availability, makes continuous, large-scale monitoring a significant challenge. Data truly equals money in this field!
Enter Robotic Networks: The Future of Monitoring 🤖📡
DIGI4ECO is shifting the paradigm by embracing robotic networks. This means moving scientists from the decks of expensive vessels to their offices, leveraging autonomous platforms to gather data. The project utilizes three main types of robotic platforms:
- Cabled Observatories: These platforms are connected to shore via internet or telephone cables, enabling continuous, real-time data streaming and year-long power supply. They offer constant insights but are geographically fixed. 💡
- Landers: These are standalone versions of cabled observatories, offering flexibility for strategic redeployment. However, their autonomy is limited by battery packs, typically lasting up to one year. 🔋
- Data-Read Docked Crawlers: To overcome the limitations of fixed platforms, DIGI4ECO employs crawlers – trackable vehicles that can autonomously sweep the environment around observatories, covering areas up to an hectare. These robots are designed to automatically count animals and gather data across wider zones. crawling
The Digital Twin of the Ocean: A Virtual Mirror 🪞🌐
The ultimate goal of DIGI4ECO is to create a digital twin of the ocean. This involves building virtual replicas of marine ecosystems, powered by the vast amounts of data collected by these robotic networks. The more data we feed into these twins, the more accurate and insightful they become.
The data streams into the digital twin come from multi-parametric sensors and fall into three main categories:
- Biological Data: This includes AI-processed images and acoustic outputs from listening to the “soundscape” of the ocean. 🎶
- Geochemical Data: Sensors characterize parameters like dissolved oxygen, methane, CO2, nitrates, phosphates, pH, and turbidity. 🧪
- Oceanographic Data: Classical sensors measure salinity, temperature, and water speed and direction for currents. 🌡️💨
These real-time updated replicas allow us to precisely understand the current status of ecosystems and run _ “what if” scenarios_ using artificial intelligence. This is crucial for predicting the biological responses of marine life to changing environmental conditions like rising temperatures and salinity, as highlighted by the IPCC.
DIGI4ECO’s Ambitious Aims 🎯
The DIGI4ECO project has set out to achieve several key objectives:
- Virtual Recreation: To virtually recreate digital replicas of four distinct coastal ecosystems: two in the Atlantic and two in the Mediterranean Sea. These coastal areas are targeted because they are the frontline of human impact. 🌍
- Master Platforms: Establishing master robotic platforms for these four environments, including two cabled observatories (Galway Bay, Ireland, and near Barcelona, Spain), crawlers in Kristineberg, Sweden, and landers in the Eastern Mediterranean Sea off the Italian coast in Ancona.
- Standardized Digital Twin: Building a standardized digital twin for each of these four diverse environments.
- Online Data Bank: Creating a central, online data bank that remotely stores, pre-processes, and standardizes all collected data. This bank will feature embedded AI routines for data quality flagging, outlier elimination, and gap-filling due to sensor malfunctions. 💾
The Frontend: Interacting with the Digital Twin 🖱️✨
The user-facing aspect of DIGI4ECO, the “phenotype,” will be an interactive web framework. Users will be able to:
- Select Areas: Choose specific geographical regions for analysis. 🗺️
- Select Platforms & Sensors: Specify the robotic platforms and sensors they wish to examine. 🤖
- Select Time Windows: Define the temporal scope for their analysis. ⏳
- Receive AI Guidance: Get recommendations from AI on the best statistical tools for their analysis, based on data quality screening. 🧠
The end result will be a synthesized output, similar to the ecological descriptors used for policy, providing meaningful ecological information for citizens and policymakers.
AI in Action: Processing the Data Deluge 🧠👁️
The sheer volume of data generated by marine observation platforms presents a significant challenge. Jacopo emphasizes the need to overcome the human bottleneck in processing terabytes of information weekly. Artificial intelligence is the key:
- Image Processing: AI is being used to automatically classify animals from images, even in complex scenarios with multiple species present. This is vital for overcoming the limitations of manual image analysis. 📸
- Acoustic Analysis: AI algorithms process acoustic outputs to identify biological sounds, providing insights into the marine soundscape. 🔊
- Object Detection: In the OBSEA observatory, yellow object detection models are used to identify fish species, allowing for time-series extraction and analysis of individual detections. This helps in comparing textures to identify potential prey-predator interactions. 🐠
Facing the Harsh Realities of the Ocean Environment 🌊🛠️
The ocean is a formidable environment, and maintaining monitoring equipment is a constant battle. Enoc highlights the challenges:
- Physical Issues: Corrosion, storms, and wave damage can break cables and degrade sensors. ⛈️
- Biofouling: Marine organisms like shellfish can attach themselves to sensors, distorting measurements. This can impact expensive sensors, requiring regular cleaning and redeployment. 🐚
To combat these issues, Grafana is employed not only for visualizing scientific data but also for monitoring sensor health. By tracking parameters like salinity and observing significant drops when shellfish are present, scientists can identify when sensors need maintenance.
Managing Sensor Lifecycles with Grafana 📊
Grafana plays a crucial role in tracking sensor history and metadata. By storing information about sensor deployments, operational lifespans, and images before and after deployment, researchers can quickly get an overview of sensor performance. This helps in identifying sensors that have failed multiple times, prompting decisions about repair or replacement. Dashboards can be created for individual sensors or groups of sensors, crucial for managing the tens or hundreds of devices involved.
Monitoring the IT Infrastructure Itself 🖥️📈
Even the IT infrastructure that supports these operations requires monitoring. Grafana, integrated with tools like Zabbix, visualizes server status, disk space, and request loads, ensuring the smooth functioning of the entire data pipeline.
The Vision: A Comprehensive Digital Twin 🚀✨
The DIGI4ECO project aims to build a digital twin of the ocean, a virtual representation powered by multi-parametric data streams and sophisticated underwater asset management. While many of the tools used are common in IT, their application in data simulation, modeling, and AI-based products within this specific marine context is revolutionary.
Grafana serves as an excellent open-source frontend for this vision, allowing the visualization of acquired ocean data in a user-friendly format. The team is particularly excited about Grafana’s capabilities for visualizing real-time and historical scientific data, operational activities, and infrastructure health.
The Future of Ocean Observation with Grafana 🌟
A key aspiration for scientists working in earth observation is the ability to display four-dimensional data (time, latitude, longitude, and depth) directly within Grafana. This would allow for the visualization of gridded data like satellite imagery or model outputs, making Grafana a truly complete frontend for digital twins of the ocean.
The DIGI4ECO project, now in its middle stage, has focused on data housekeeping and harmonization. The next two years will be dedicated to creating scientific products, generating statistics, and enabling “what if” scenarios. Stay tuned for more updates as they work towards delivering a complete digital virtualization of our oceans! 🌊💻✨