- Title
- Evaluating the non-stationarity of Australian annual maximum flood
- Creator
- Ishak, E. H.; Rahman, A.; Westra, S.; Sharma, A.; Kuczera, G.
- Relation
- Journal of Hydrology Vol. 494, p. 134-145
- Publisher Link
- http://dx.doi.org/10.1016/j.jhydrol.2013.04.021
- Publisher
- Elsevier
- Resource Type
- journal article
- Date
- 2013
- Description
- Projections of future changes to flood risk resulting from the combined effects of natural climate variability and anthropogenic climate change typically have high uncertainty, due to uncertainty in both the atmospheric forcings and the processes by which rainfall is converted to runoff. In this paper we test for non-stationarity of the most extensive annual maximum (AM) streamflow database compiled in Australia to date, comprising records from 491 small to medium sized catchments with record lengths from 30 to 97. years, and with minimal regulation or land cover change. The data are analysed for trends using the Mann-Kendall (MK) test, for three study periods (1955-2004, 1965-2004, and 1975-2004). Three different approaches were used to account for the impact of serial correlation on the MK test, and a regional MK approach and a bootstrap resampling approach were used to account for the cross-correlation structure in the data. The outcome of the MK test indicates a significant downward trend in the AM flood data in the south-east and south-west regions of Australia. These downward trends are field significant at the 10% significance level, whereas any upward trends are not. In addition to the MK test, a partial MK analysis was performed to assess the influence of three indices representing large-scale climate variability: the Southern Annular Mode (SAM), El Nino Southern Oscillation (Nino 3.4) and the Interdecadal Pacific Oscillation (IPO). After accounting for these indices the total number of stations showing statistically significant trends is reduced, suggesting that much of the observed trend in AM floods may be associated with these climate modes. Given the significant covariability between these indices and the global warming trend, however, the extent to which observed changes are due to natural variability rather than anthropogenic climate change cannot yet be conclusively determined. 2013 Elsevier B.V.
- Subject
- trend analysis; floods; bootstrapping; climate change; climate variability; cross correlation
- Identifier
- http://hdl.handle.net/1959.13/1293454
- Identifier
- uon:18596
- Identifier
- ISSN:0022-1694
- Language
- eng
- Reviewed
- Hits: 1884
- Visitors: 2099
- Downloads: 0
| Thumbnail | File | Description | Size | Format |
|---|