Extended Deployment of an Automated Coastal Environment Monitoring System

Abstract

About 600 million people reside along the coasts at an elevation of fewer than 10 meters above sea level; 2.4 billion live within 100km (60 miles) of coastline (UN, 2017). In addition to being critical to human society, the biodiverse coastal habitats such as estuaries, wetlands, coral reefs, and upwelling areas support and provide breeding grounds for organisms such as fish, marine mammals, sea turtles, and migratory birds. Given the role the coasts play in billions of lives, monitoring this highly dynamic environment is an urgent concern for safety and conservation efforts. To better understand anthropological impacts and natural fluctuations of coastal waters, it is necessary to develop a coastal monitoring system that can be deployed non-invasively and cost-effectively. However, surveying the coastal waters is not an easy task. Coasts are affected by a unique mix of environmental and human influences that vary over time. These influences range from natural erosion and accretion to substantial coastal engineering. Individual coastal sites can be challenging to reach and may have considerably different needs and vulnerabilities (Nichols et al., 2019, Kirwan & Guntenspergen, 2010, Janzen et al., 2019). The diverse conditions along the coasts create a complicated environment for any observation system to navigate. Within this environment, there is space for new and inexpensive approaches to coastal monitoring. Researchers need an affordable and reliable method of data collection to understand this critical area further. A source of high-frequency coastal environmental data provides an opportunity for an accurate analysis of the highly dynamic coastal ecosystem. Some of the most vital data is the hardest to obtain, that is, highly dynamic parameters that exist during transient events, such as during or shortly after a significant meteorological or oceanographic event. This kind of data is currently impractical to collect due to the risk it poses to data collectors. For instance, during extreme scenarios such as hurricane landfall, capturing transient event data is crucial for the proper development of local coastal models. The ideal data collection system needs to be economical, reliable, non-invasive, and capable of surviving highly dynamic coastal environments. The solution is a fully autonomous, remotely operated, cable-mounted mobile platform that will pull itself along a cable anchored between a land station and a point offshore anchored to the seabed in a non-invasive way. The platform locomotes under its own power along this cable to enter and exit the water. The robot traverses a fixed, partially underwater cable and autonomously collects data as it moves through the water. By simplifying the motion of the robotic system to a single degree of freedom, it can easily overcome the many challenges of navigating a coastal environment. Specifically, it can return to a predetermined location over time for long-term study and minimize the need for human supervision and maintenance, thus reducing costs while increasing sampling rates. The system will collect information regardless of weather and with much greater temporal resolution than manual methods. It is also highly modular; different sensor suites can be equipped depending on the application and deployment location. Mounted to this chassis was a commercially available In-Situ Aqua TROLL 600 sonde for data collection. The first prototype was constructed and field tested in the summer of 2020 (Bennett et al., 2020). Successful field testing and data collection proved the capabilities of the autonomous system, which collected water quality data of the same or better quality than “traditional” methods (human operator data collection w/ sonde). In the summer of 2021, a redesigned system completed a 48-hour deployment. The next challenge this system faces is long-term (>48hrs) deployment and consistent collection of high-quality data. The product goal of this system is to deliver water quality data to a remote operator, and therefore must operate in the field without a human operator present for extended periods of time. In this paper, the design and deployment of the 3rd generation platform will be discussed. Data from the extended (>48hr) deployment will be presented.