Author: Chanveasna Ly
Ly, Chanveasna, 2021 Detecting and mapping seasonal variations in water turbidity and mouth bar geometry along an asymmetrical delta using Normalized Different Turbidity Index (NDTI) with Sentinel 2 satellite imagery, Flinders University, College of Science and Engineering
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
Remote sensing studies for ocean monitoring first appeared in the 1960s and have subsequently increased in popularity (Potes et al. 2018). Satellite remote sensing compliments traditional in-situ data specimen and laboratorial examination of water quality variables, which is laborious and expensive and typically has a limited spatial and temporal range. In contrast to traditional field measurement techniques, satellite-based remote sensing methodologies are inexpensive and can cover relatively large spatiotemporal ranges. Satellite-based remote sensing techniques are effective for investigations that monitor river discharge into the coastal waters and subsequent reworking by near-shore basinal processes. This investigation focuses on suspended particulate matter (SPM) concentrations in rivers, their associated plumes, and the near-offshore. Monitoring of SPM concentrations is crucial for sediment transport and ecosystem modelling, and for understanding the morphology and evolution of marginal marine systems. Furthermore, the relative concentration of SPM has been treated as a proxy for turbidity and is used to investigate plume geometry and understand erosion and/or progradation of the coastline. This research aims to implement remote sensing to detect the relative suspended sediment concentration (SSC) for as a proxy for turbidity and erosional processes mapping the plume and mouth bar geometry for investigating two specific purposes: 1) to determine how fluvial discharge and basinal processes (waves, tides and longshore currents) influence mouth bar and sediment plume formation, and 2) determine how the interplay of sea level, sediment supply vs erosional activities influence the rates of localized progradation or erosion in the asymmetric, shallow basin, Mitchell River, Gulf of Carpentaria (GoC). The Mitchell River delta is complex delta system, which is an asymmetrical (sediment acquired from longshore drift trapping on the updrift delta as the mouth bar performs as a groyne), tide-dominated, wave-influenced, fluvial-affected (Twf) system (Nanson et al. 2013), with incredibly limited research where mostly about the lower delta plain and no research about near offshore system leading this research to be an understudied geographical regime that contribute to the understanding about the depositional sediment both onshore and offshore and enhance the understanding of delta morpho-dynamic. These understanding will become beneficial for delta ecosystem investigations in Australia, essentially in the GoC. Moreover, the most significant outcome of this research is to theoretically enhance understanding of processes affecting localized shoreline changes (either side of the asymmetrical delta). To detect and map the plume and mouth bar geometry, the Sentinel 2 L2A (atmospheric and geometric corrections) products were collected from January 2020 to August 2021 using automation process in Google Earth Engine (GEE) for analysis along the meteorological and hydrological statistics for the interpretation. The pilot process is performed to verify the best method in selecting sensitive bands included: (1) Red+Green+NIR, 2) Red+Green (NDTI), 3) Red+NIR+SWIR to get optimal result, which the NDTI is the most suitable, before the semi-automatically method is applied vii for this study including automation process: 1) water area extraction (image pre-processing) using band ratioing method of MNDWI (Modified Normalized Different Index), 2) turbidity detection using band ratioing method of NDTI; manual process: 3) plume pattern corrections, and 4) digitizing mouth bar geometry. This method is performed in the ArcGIS Pro version 2.8.2 for both the pilot and semi-automatically methods, yet the Erdas Imagine was assisted for visualizing the sensitivity of band combinations and spectral reflectance to enhance the confidence of choosing suitable bands in the pilot stage. Three significant inaccuracies of the algorithm revealed: 1) errors of MNDWI in masking out clouds where thin cloud and cloud shadow could be less masked out leading to confuse classified as plume and over masked out resulting in missing plume areas, 2) error of NDTI in differentiating pixels due to undistinguished between low concentration plume reflectance with water pixel leading to miss classified plume area, and 3) error of pixel-based classification in classifying plume boundary leading to assisting of the manual correction of plume pattern. The results revealed that the large-scale sediment plume orientated to southward (analogue to palaeo-flow orientation) in the wet season while it shifted its orientation in the dry season to northerly directed (analogue to longshore drift current) and parallel to the palaeo-channels orientation within the Mithcell River delta. Another discovery in this study showed that the high and very high turbidity constantly dominate in the vicinity of the north nearshore zone in both wet and dry seasons where severe coastline erosion has occurred. At the same time, the south nearshore zone of the Mitchell mouth bar, where the progradation coastline has prograded, experienced low turbidity in the dry season and moderate turbidity in the wet season.
Keywords: Water Turbidity, Mouth Bar, Plume Detection, Asymmetrical Delta, The Mitchell River Delta, Normalized Different Turbidity Index, NDTI, Sentinel 2 Imagery
Subject: Geography thesis
Thesis type: Masters
Completed: 2021
School: College of Science and Engineering
Supervisor: Tessa Lane