Filters
Availability
Publication Year
Resource Type
Faculty
School/Institute
Publication Search Results
Sort
Results per page
1 - 3 of 3
-
(2009)Conference PaperA raft of models and definitions of SoTL exist and the best appear to transcend disciplinary contexts, and are sufficiently robust for academics to measure scholarly practices. Critical engagement with the scholarly literature is necessary for academics to gain a realistic view of where their work practices are situated within the scholarly domain. Because academic staff are disciplinary experts they are best placed to comment on whether the models of scholarship describe the scholarship of learning and teaching within the context of their own disciplines as well as within the confines of the Australian higher education sector. This paper pushes the existing debates on reconciling what evidence of scholarship in the disciplines actually is and what is considered valid, and in doing so uncovers why the process of reconciliation, between current practice and supporting evidence, remains elusive. Higher education academics need to identify and reconcile tacit disciplinary knowledge with their SoTL approach in order to unpack the complexity and value of their practices. Enabling academic staff to annotate their activities, roles and accomplishments and then map these items onto the various models of scholarship will enrich the status of scholarship of teaching and learning within the higher education sector.
-
(2023)ThesisShifting patterns in consumption and the inadequate disposal of wastes has led to the escape of anthropogenic debris into the marine environment. The growing volume of debris, both within and entering coastal and marine areas, has prompted global concern over the risks they may pose to environmental and human health. Responses to curb further entry and address debris already within the environment include several management interventions, informed by policies and legislation. Effective debris management requires an understanding of potential sources, subsequent dispersion and an estimate of the risks posed to habitats and biological assemblages, which could be attained through environmental monitoring. Monitoring across relevant spatio-temporal scales, however, is often outside the reach of formal government and research programs and there is a growing recognition of the role citizen science data may play in debris management and decision making. This thesis aims to bridge environmental monitoring with policy and decision making, combining citizen science with other data into an evidence-base for management. The thesis assesses several citizen science datasets from a local to Federal scale to identify debris trends and their drivers (Local: four estuaries; State: Queensland; Federal: Australia). Further, I combine expert elicitation and empirical debris data to assess the risk posed by debris. I provide a framework for linking debris identified within the environment to economic sectors, as part of a formal accounting framework. The thesis also provides methodological guidance to refine citizen science sampling during monitoring programs, to improve the accuracy and reliability of resulting datasets. Through careful application and consideration of data quality, citizen science data could be used to supplement formal monitoring efforts to better understand and address the challenge of marine debris. This thesis advances the role of citizen science beyond environmental monitoring to inform management efforts at scale.
-
(2024)ThesisRecent applications of Remote Sensing have tended to implement more automated approaches, for example with Machine Learning (ML) or Artificial Intelligence (AI) approaches that are implemented without, or with limited geovisualization of the multivariate data. This poses challenges to the interpretability of patterns and the explainability of classifications. Although such approaches can work consistently over large areas, limitations on interpretation have led to perceptions of the approaches as black boxes that typically require very large amounts of training data and processing power. Conversely, approaches founded on operator-interpreted ecological relationships, such as traditional Aerial Photography Interpretation (API), lack reproducibility and cannot be implemented across large areas. These trade-offs have stimulated the development of complementary approaches, with the development of explainable ontologies for landscape geovisualization. In an era increasingly dominated by AI derived solutions, this thesis aims to formalize Satellite Imagery Interpretation (SII) as a quantitative multispectral extension to API and to demonstrate its application for environmental mapping. The motivation is to develop a Remote Sensing approach built on the understanding of ecological relationships that is both communicable and logistically feasible across scales for global monitoring efforts. SII places emphasis on extracting existing knowledge from landscape ecology in the form of indices for environmental gradients and colourimetric indicators of physiognomic properties with the evidential reasoning from perceptual cues, Gestalt principles and cognitive attributes. These indices are intended to enable interpreters to classify land classes in their simplest semantic form (such as forest, water or snow) to facilitate efficient classifications with holistic reduction analyses that are intended to test and help mitigate the effects of the Modifiable Spatial Temporal Unit Problem (MSTUP). The MSTUP combines the scaling effects that may result from inadequate spatial extents and temporal ranges for analysis. By applying gradient theory, geospatial holistic reduction is essentially chorologic, deducing the causal relations between biophysical geographic phenomena occurring within, but also across regions and modelling them systematically. Combining holistic and parametric procedures is possible since gradient analyses provide scalability by avoiding the need to redefine entities and interactions across spatial and temporal scales which is important for global monitoring. SII achieves communicability and replicability through three key steps: 1) development of false colour RGB composite interpretation keys, 2) colourimetric ontologies for training references and/or direct measurement when applicable, and 3) chorologic classification to infer and validate causality of the classes created. Taking these approaches, gradient-directed sampling was applied with ecological colourimetric deductions of pseudo invariant features from gradient extremes to significantly reduce sampling intensity without compromising classification performance. Following an introductory literature review and background to the concepts that comprise the SII approach, this thesis has four core chapters that develop the approach and methods with case studies: Chapter 2 (A Colourimetric Approach to Ecological Remote Sensing: Case Study for the Rainforests of South-Eastern Australia) introduces colourimetric ontologies, holistic analysis, Satellite Imagery Interpretation keys and eco-regionalization. The direct measurement of the hue angle from a new band ratio RGB produced the best overall accuracy across a diverse range of ecoregions, only slightly higher than a new conventional index based on a band ratio subtraction. These results demonstrate how the visual association of features from existing reference maps to recognizable ecological features in false colour images can be used to determine classification thresholds for a specific forest type without any new training data; and that the visually intuitive direct measurement of colour space metrics can yield comparable performance to conventional indices. Chapter 3 (Holistic Reduction to Compare and Create New Indices for Global Inter-Seasonal Monitoring: Case Study for High Resolution Surface Water Mapping) introduces the application of a geospatial decision matrix to provide a simple holistic analytical framework to guide the creation and comparison of new indices with existing ones in order to rationalize the best index that would satisfy a set of commission and omission criteria compiled from the literature. It systematizes holistic reduction with Satellite Imagery Interpretation keys and a decision matrix, emphasizing the need to test for inter-seasonality and proposes a global zonation to facilitate ubiquitous measures. Holistic reduction proved to be an effective approach for deducing and validating indices to separate water as a reduced land class uniquely from other unrelated land cover classes across the seasons in a large, diverse and therefore highly representative study zone. Chapter 4 (High Resolution Forest Masking for Seasonal Monitoring with a Regionalized and Colourimetrically Assisted Chorologic Typology) consolidates the prior concepts and introduces the concepts of Chorologic Typologies and phenologic Moments of Maximum Separability (MoMS) in combination with regionalization as ways of mitigating the MSTUP. The study applied colourimetric conditioning to demonstrate how single indices can be refined with additional variables as transparent and explainable chorologic typologies. It applied a comparison with two well-recognized global products at a comparable spatial resolution for assessment. The chorologic typology was composed of two high resolution tree cover indices conditioned by colour space metrics to indicate insolation, surface NIR saturation and water extent in order to consistently distinguish tree cover from other spectrally similar classes identified from the literature and visual inspection across the four seasons. Phenologic Moments of Maximum Separability (MoMS) are proposed to promote optimal data selection and facilitate single image analyses of annual or interannual median-based composites that served to account for temporal variability in order to focus classifications on temporal aggregates when classes of interest display the greatest spectral contrast with other classes. The results demonstrate how regionalized chorologic typologies can be developed to provide continual inter-seasonal monitoring for forest change and when single date drone surveys would be most effective. Chapter 5 (High Resolution Snow and Ice Mapping for Global Inter-seasonal Monitoring) applies all of the former concepts to a highly spatio-temproally variable land cover class. An RGB saturation index and an existing snow/water index were identified as masks to facilitate ubiquitous global measures based on the averages of thresholds for an ecologically wide range of study zones. The interpretation keys and methods for creating colourimetric ontologies, new environmental indices and chorologic typologies presented herein provide a comprehensive foundation to formalize SII as a visual analytic framework for the simplification and communication of effective remote sensing classifications for environmental mapping. Each chapter will demonstrate how spatial-temporal scaling effects can be mitigated in different ecoregions by applying inter-seasonal conditioning, and regionalizations or zonation when necessary to standardize concerted global monitoring efforts.