- Title
- A paleoclimate-informed examination of flood and drought epochs in the Murray-Darling Basin
- Creator
- Ho, Michelle
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2014
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- The Murray-Darling Basin (MDB) is Australia’s largest water catchment and the nation’s reputed ‘food bowl’. Climate, and consequently water availability, in the region is highly variable both temporally and spatially, as evident in the regular occurrence of floods and persistent droughts and the regionally distinctive impacts of such events. A key limitation to accurately quantifying flood and drought risks in the region is the relatively short instrumental records (approximately 100 years at best) of rainfall and stream flow. Furthermore, research over the past few decades has revealed that flood and drought risks across the MDB are modulated by a number of different large-scale climate drivers (e.g. El Niño/Southern Oscillation, Southern Annular Mode, Indian Ocean Dipole and Interdecadal Pacific Oscillation) on seasonal to multi-decadal timescales. These climate mechanisms influence the MDB hydroclimate both individually and in combination. Current assessments of flood and drought risk are based on relatively short instrumental records and are therefore inadequate for properly evaluating either multidecadal variability or the influence of numerous largescale climate drivers on MDB hydroclimatic variability. This thesis aims to improve understanding of long-term flood and drought risk in the MDB through the use of paleoclimate records of both large-scale ocean-atmospheric processes and continental Australian rainfall. The use of paleoclimate data will enable improved insight into pre-instrumental climate variability. The efficacy of using paleoclimate proxy records of large-scale climate drivers (e.g. the El Niño Southern Oscillation, the Indian Ocean Dipole, the Southern Annular Mode, the Interdecadal Pacific Oscillation and the Pacific Decadal Oscillation) to reconstruct MDB rainfall was examined. In order to reconstruct MDB rainfall using these relationships, both linear and non-linear relationships between MDB rainfall and different climate drivers and combinations of drivers were quantified. Importantly, it was found that the MDB rainfall response was markedly different when climate drivers were considered in combination compared to the response to a single climate driver. Currently, numerous multi-centennial paleoclimate records exist for the El Niño Southern Oscillation and the Pacific Decadal Oscillation. However, paleoclimate reconstructions of other climate drivers are less developed, limiting the feasibility of paleoclimate driver based reconstruction methods. Nevertheless, this work has highlighted significant potential for using paleoclimate proxy records of large-scale climate drivers to reconstruct MDB rainfall variability should multi-centennial records of Indian and Southern Ocean variability for key seasons, as well as the Pacific Ocean Basin-wide Interdecadal Pacific Oscillation, become available in the future. In addition to assessing the possibility of reconstructing MDB rainfall using paleoclimate proxy records of large-scale climate drivers, another approach was explored. This second approach attempted to utilise paleoclimate proxy records of rainfall that exist in Australia through the investigation of relationships between rainfall in the MDB and Australian rainfall outside the MDB. At present, only three continuous, high-resolution paleoclimate rainfall records exist in Australia, none of which are in the MDB. While in situ paleoclimate proxies would be ideal, there are currently no existing proxies in the MDB that provide continuous, high-resolution records of hydroclimatic variability. In addition, paleoclimate archives capable of sensing and recording rainfall variability at high resolutions are unlikely to be found in the MDB. Given the technical difficulty and costs involved in obtaining paleoclimate records, it is prudent to determine regions where the future assembly of these records would be of most use to reconstructing MDB rainfall (i.e. regions that would most accurately reconstruct MDB rainfall variability). This was achieved using an optimal interpolation procedure. Locations around Australia were sequentially selected to optimise the degree of MDB rainfall variability that could be resolved. The locations were then compared with sites that could potentially yield continuous, high-resolution, rainfall-sensitive paleoclimate archives, thereby providing an indication of where future paleoclimate research efforts could be concentrated to maximise the accuracy of MDB rainfall reconstructions. In order to demonstrate the utility of the existing Australian paleoclimate rainfall proxy records to remotely reconstruct MDB rainfall, reconstructions of rainfalls in four casestudy sub-catchments were made using the three available records. Four different reconstruction models were calibrated. The best model was able to resolve between 35% and 61% of rainfall variability in the four case-study sub-catchments when calibrated using instrumental data from the three proxy rainfall sites. The modelled results were then compared to rainfall in the four case-study MDB sub-catchments modelled using rainfall from the first three locations selected from the optimal interpolation procedure (i.e. the ideal locations). Rainfall from the first three optimised locations was able to resolve between 62% and 82% of MDB rainfall variability. This demonstrates the importance of obtaining additional paleoclimate data in optimal locations to more accurately reconstruct MDB rainfall. A key outcome of this work was the reconstructions of rainfall in four case-study MDB sub-catchments using existing high-resolution paleoclimate rainfall records around Australia. The reconstruction enabled an assessment of rainfall variability from 1685- 1981 using all three paleoclimate proxies and an extended reconstruction from 749 BCE to 2001 CE using the Wombeyan Cave record. The reconstructions showed that the risks of flood and drought have been higher prior to the instrumental records in both magnitude and persistence. A qualitative comparison was also made between the reconstruction of rainfall in the upper Murray catchment and previous paleoclimate reconstructions of hydroclimatic variability in the MDB. This work demonstrated that the realisation of a high-resolution paleoclimate rainfall network around Australia as determined in the optimal interpolation, would enable an increased degree of variance to be captured in reconstructions of MDB rainfall. In addition, it was revealed that extending the current paleoclimate records of Pacific, Indian and Southern Ocean variability spanning the Common Era will also enable MDB rainfall variability to be reconstructed. Reconstructions of MDB rainfall variability using different networks of paleoclimate data are expected to enable more accurate estimates of long-term flood and drought risks in the MDB. This would then provide a realistic assessment of the baseline risks, thus enabling the adoption of robust water resource management schemes capable of responding to the degree of natural variability identified from paleoclimate-based reconstructions. Such information will also enable future climate scenarios to be adequately constrained, validated and assessed using a multi-centennial or multi-millennial perspective of past hydroclimatic variability.
- Subject
- climate; paleoclimate; Murray-Darling Basin; flood; drought
- Identifier
- http://hdl.handle.net/1959.13/1048522
- Identifier
- uon:14919
- Rights
- Copyright 2014 Michelle Ho
- Language
- eng
- Full Text
- Hits: 1545
- Visitors: 2074
- Downloads: 663
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Abstract | 533 KB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Thesis | 14 MB | Adobe Acrobat PDF | View Details Download |