Title of Project
Machine learning approach to restoration, prediction and quality control of oceanographic data from IMOS Moorings
Paul Rigby, Oleg Makarynskyy, Ickjai Lee
College or Research Centre
Summary of Project
Reliable data on the state of the ocean and coastal areas are in growing demand. For the past 10 years, the Australian National Mooring Network, as part of the Integrated Marine Observing System (IMOS) has measured physical and biological parameters at over 50 sites in Australian coastal waters. The resulting data collection consists of a huge amount of time series information across many variables. Processing these data sets currently requires a significant amount of human-in-the-loop overheads for quality control and data processing. Furthermore, on occasion field instrumentation will fail due to the adverse ambient conditions and/or faulty manufacturing. The presence of gaps in data records may render the period of observation unsuitable for many practical purposes e.g. where a continuous record is required for numerical model validation or calibration. It has been demonstrated in a range of recent studies that meshless data–based methods of time series interpolation and expansion, which include stochastic models as well as artificial intelligence approaches such as genetic algorithms, fuzzy logics, artificial neural networks (ANNs) [2-4], may be beneficially used to estimate met-ocean parameters. Other recent studies have demonstrated that a Bayesian network can be trained to conduct quality control of real time measurements from IMOS oceanographic sensors . This project provides a candidate the opportunity to investigate a machine learning approach to increasing the value of oceanographic data. The full collection of IMOS Moorings data will be available for use in developing and training algorithms. Much of this data has already been flagged by heuristic quality control routines, and manually annotated by domain experts. This PhD project will contribute to the following tasks: • Investigate a machine learning approach to automate the quality control process of oceanographic data by flagging anomalies and outliers • Develop an approach and implement an artificial intelligence technique to the task of data gap interpolation; • Develop a forecasting methodology for extrapolating data on different timescales. • Analyse and understand the relationships between different oceanographic, water quality and ecosystem parameters in the tropics  http://imos.org.au/nationalmooringnetwork.html  Makarynskyy, O., Makarynska, D., Kuhn, M., Featherstone, W. E., 2005. Using artificial neural networks to estimate sea level in continental and island coastal environments. Hydrodynamics IV: Theory and Applications, L.Cheng and K.Yeow (eds.), Taylor & Francis Group, London, 451-457.  Makarynskyy, O., Makarynska, D., Rusu, E., Gavrilov, A., 2005. Filling gaps in wave records with artificial neural networks. Maritime Transportation and Exploitation of Ocean and Coastal Resources, C.Guedes Soares, Y.Garbatov and N.Fonseca (eds.), Taylor & Francis Group, London, 1085-1091.  Makarynskyy, O., 2005. Artificial neural networks for wave tracking, retrieval and prediction, Pacific Oceanography, 3 (1), 21-30.  Smith, D.; Timms, G.; De Souza, P.; D’Este, C. A Bayesian Framework for the Automated Online Assessment of Sensor Data Quality. Sensors 2012, 12, 9476-9501.
Machine Learning; artificial neural networks; artificial neural networks; algorithm development; IMOS data; artificial intelligence; statistics; interpolation; prediction
Would suit an applicant who
possess and be willing to further develop high-level quantitative analysis and machine learning techniques to progress this project. The PhD project proposal will be developed around analysis of IMOS oceanographic and water quality parameters. The candidate is expected to possess and be willing to further develop high-level quantitative analysis and machine learning techniques to progress this project. Demonstrated strong skills in programming and data science will be required. The candidate should have an interest in understanding and working with large collections of oceanographic data. Some previous exposure to oceanography would be highly regarded.
Updated: 1 year ago