AUTONOMOUS MONITORING OF LAKES AND OTHER WATER BODIES USING MICRO UNDERWATER VEHICLES

: (Alternative Format Thesis)

Abstract

The process of manually collecting lake water quality (WQ) data contains numerous potential errors and biases. Such errors and biases can be due to process deviations, spatial, and temporal effects. Some of these are understood and can be mitigated or accounted for. However, human variability can introduce such errors into lake monitoring programmes. As an exemplar, the influence of weather on the data collection in a professional, manual lake monitoring programme in Maine, USA was investigated. The investigation revealed higher average water temperatures and chlorophyll values with lower dissolved oxygen (DO) values compared to true values as a statistically significant result of weather influencing the humans collecting the data. The magnitude of the weather biases found were significant. Whilst the exemplar was a narrow single issue, it raises questions about the validity of lake WQ data from manual monitoring programmes. For the use of manually collected WQ data to be credible, weather and all other biases, need to be investigated and taken into account. As used to gain the water temperature "truth" in the case study, a better alternative to manual monitoring may be to embrace automation.

Comparing mobile autonomous methods which capture data over time and space, with manual methods, various emulated uncrewed underwater vehicle (UUV) travel paths were compared and contrasted with manual. The work showed the WQ data that manual and autonomous methods collected are statistically equivalent and both can identify the same temporally and spatially separated temperature changes in a lake. However, the different methods do not necessarily measure the same physical piece of water. There are many other additional considerations such as finance, equipment, training and time when considering autonomous lake WQ monitoring methods. Autonomous methods can gather more data than manually deployed sensors, and this could be their key advantage.

Unlike in oceanography, lake science has not widely adopted UUV to collect WQ data. This work explores and identifies the barriers to the adoption of mobile automated lake monitoring. These barriers include UUV weight, size, and cost and to address them a solution of reducing the UUV component count was proposed. A minimal configuration, single actuator, micro-underwater glider-type UUV model was developed and its design space was explored to produce design guidelines. This design work was partially validated in a successful prototype lake trial. Whilst further work will be required to allow such a design to be used commercially, this work provides a sound engineering basis to begin the lake WQ automation journey.

Date of Award	7 May 2025
Original language	English
Awarding Institution	University of Bath
Sponsors	Engineering and Physical Sciences Research Council
Supervisor	Alan Hunter (Supervisor), Lee Bryant (Supervisor) & Danielle Wain (Supervisor)