- Title
- Adaptive waveform retracking of radar altimetry waveforms over hetreogeneous inland waters
- Creator
- Marshall, Andrew
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2021
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- To manage the pressure that population growth, human impact and climate change is having on the allocation of, and access to, water there is an increasing need to monitor the world’s water resources, independent of infrastructure and inter-government policies. Traditionally the realm of the hydrologist, this task has relied on the deployment of in-situ gauges and instruments. Recent focus has been on the capabilities of satellite-based technologies to augment the existing hydrology in-situ network with the aim of replacing it with a global water level monitoring tool for inland rivers, lakes and wetlands. This research has focussed on the satellite altimetry coverage of the middle Fly River floodplain as well as Lake Murray—both located in the Western Province of Papua New Guinea. The Fly River floodplain is a mine-impacted environment and monitoring of water level change through the various floodplain and wetland entities is required into the future. More than for other similar environments throughout the world there will become an increasing need to support Fly River local communities with information regarding predicted changes to inundation that may have impacts on their communities and subsistence livelihood. The current state-of-the-art satellite altimetry analysis methodologies over heterogeneous inland waters do not meet the accuracy and reliability requirements for water surface measurement. This is particularly relevant for the relatively small river and lake systems that contribute to a typical complex floodplain or wetland system. Methodologies developed in this study enable routine, accurate and reliable extraction of water surface elevations from nadir-looking pulse-limited radar altimeters over heterogeneous inland waters. This is achieved by deconstructing the shape and form of the recorded waveform and correlating that form against external inputs so that the environmental factors that have affected the shape and form of the waveform are understood and can be addressed. The external inputs comprise a range of supporting data, including information derived from satellite imagery as well as in-situ water level observations. A process of waveform footprint classification is developed with assessment of footprint inundation extent based on image analysis from both multi-spectral and synthetic aperture radar (SAR) imagery. The methodology is extended to include a full definition of the landform cover type as well as prediction capabilities for off-nadir calm water detection. A significant advancement over conventional processes is that waveforms, and the associated water surface elevations, are assessed based on an analysis of the waveform and adjacent waveforms as well as the nature of the altimetry footprint rather than solely on statistical agreement of the derived water surface elevation with that derived from adjacent waveforms. This facilitates the retention of water level estimates over relatively small water bodies, where multiple, statistically consistent, estimates would not be practical. The processes developed in this research offer a methodology for the extraction of reliable water surface estimates, in both a temporal and spatial context, over heterogeneous inland waters. An optimised adaptive threshold retracker, the Waveform Adaptive Threshold Retracker, is developed as part of this study with methodology and workflow detailed in the thesis. Methods for the accurate identification of waveforms impacted by hooking and other sources of contamination are developed, along with tools for the rectification of impacts and estimation of likely contamination magnitude. Optimised waveform retracking using the adaptive retracking methodology and workflow is validated at Envisat Radar Altimeter 2 (RA-2) and Satellite with Argos and AltiKa (SARAL/AltiKa) crossings of the Fly River and achieved by comparison of the altimetric time series with in-situ gauge data. Validation is also undertaken for floodplain sites where verified virtual in-situ gauges have been established for validation of both Envisat RA-2 and SARAL/AltiKa-derived elevations. This comparison has been undertaken for the 10 years of Envisat RA-2 data acquisitions and the pre-drifting phase cycles of SARAL/AltiKa data. Elevation profiles from Envisat RA-2, SARAL/AltiKa and Cryosat-2 SAR Interferometer Radar Altimeter (SIRAL) altimeters have been derived across both the Fly River floodplain and Lake Murray and used to assess the proposed retracking methodologies for the derivation of floodplain gradients and differential elevations between various floodplain water bodies. The methodologies developed offer potential for the reprocessing of a significant archive of data from nadir-looking pulse-limited radar altimeters as well as supporting analyses of data from currently operational altimeters into the future. The work undertaken in this study has facilitated tangible improvements in the quality and quantity of water level estimates across complex inland water environments.
- Subject
- satellite altimetry; waveform retracking; inland waters; remote sensing; Fly River
- Identifier
- http://hdl.handle.net/1959.13/1423399
- Identifier
- uon:37927
- Rights
- Copyright 2021 Andrew Marshall
- Language
- eng
- Full Text
- Hits: 5281
- Visitors: 6017
- Downloads: 898
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Thesis | 18 MB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Abstract | 386 KB | Adobe Acrobat PDF | View Details Download |