Artificial Intelligence and Freshwater Modelling

To protect our freshwater for future generations, we develop a framework enabling an understanding of how environmental factors impact our water quality and how mitigation strategies can help.

Our project aims to develop a novel framework for freshwater modelling that integrates across process-based and machine learning models, augmented by more qualitative expert judgement, that can be used to address policy-relevant questions in freshwater quality well beyond the current state-of-the-art. Current models lack the means to estimate uncertainty, limiting the evaluation of their suitability for particular use cases. Furthermore, state-of-the-art systems lack the ability to model the interconnectivity of the myriad of real-world factors affecting freshwater quality.

We will develop machine learning methods to provide model interconnectivity, enabling models to share information and learn from each other. In the context of the National Freshwater Accounting System, we will provide a standardised interface for integrating and evaluating data and models, novel domain-specific metrics, and machine learning methods to quantify uncertainty in freshwater models and in the data used to drive those models. Our framework will not only simulate the temporal evolution of the system under prescribed conditions but also report uncertainty bounds on that simulation as well as monitoring needs and knowledge gaps.

The simulations will enable a better understanding and quantification of the impact of environmental factors (e.g., land use, climate change) on freshwater quality and be used to evaluate and propose mitigation strategies for pollution at different spatiotemporal scales, contributing to Te Mana o te Wai. A reflexive layer of the project queries the values embedded within existing models for freshwater and, in partnership with iwi and hapū, will consider how the new framework better reflects Māori values and priorities.

Our transdisciplinary team includes leading experts in Computer Science, Artificial Intelligence, Environmental and Freshwater Science, Agricultural and Catchment Systems Modeling, Māori Environmental Science, Policy, and Environmental Social Science.

For more information, get in touch.