Artificial intelligence informed simulation of dissolved Inorganic Nitrogen from ungauged catchments to the Great Barrier Reef

Artificial intelligence informed simulation of dissolved Inorganic Nitrogen from ungauged catchments to the Great Barrier Reef

Land based sources of nutrient loads impact the health and resilience of the Australian Great Barrier Reef, however, the current methods used to quantify and prioritise nutrient reduction to the reef need improvement to increase certainty in estimates of Dissolved Inorganic Nitrogen (DIN) from ungauged catchments. Catchment Scale Water Quality Models are currently the primary tools used to quantify the influence of landscapes towards receiving waters and are effective for communication of the influences of the landscape and its management towards the Great Barrier Reef. The design and development of these models rely on extensive observed water quality data for development and calibration of the models, however, the collection of the data are both expensive and not possible in all areas. This PhD project has developed new knowledge in simulating DIN from ungauged catchments, to overcome the challenge and knowledge gaps associated with data voids that afflict water quality modelling. Research herein has coupled catchment classification, a method demonstrated by existing the literature to effectively overcome data voids for flows, with Artificial Intelligence pattern matching and techniques to identify corroborating catchment matches for both DIN patterns and spatial data. Additionally, this research, for the first time, has used spatial datasets for Original Vegetation, as a proxy dataset to the drivers of DIN. This research has found that the Original Vegetation data represents the variability in biological response to the drivers of heterogeneity in DIN patterns across the landscape. Explainable artificial intelligence approaches were then developed to identify landscape features most influential in the classification results. Development of these methods ultimately facilitated satisfactory simulation of DIN for a pseudo ungauged catchment as well as identifying catchments that are unsuitable to share data and others that need prioritisation for future gauging programs. Together, these approaches have enabled the development of knowledge to classify ungauged catchments of the Great Barrier Reef using spatial data as a proxy for absence of observed DIN data. The findings of this doctoral study have provided new insights into water quality modelling and the selection of catchments as well as classifying the catchments and performing DIN simulations.

File reproduced in accordance with the copyright policy of the publisher/author/creator.