Continuous rainfall simulation in a warmer climate

Download files
Access & Terms of Use
open access
Embargoed until 2018-08-31
Copyright: Wasko, Conrad
Altmetric
Abstract
Continuous sub-daily precipitation sequences are required for many hydrological applications. Unfortunately sub-daily precipitation data is often unavailable due to the paucity of measurements. To overcome this, statistical methods are used to synthetically generate continuous precipitation sequences. These methods generally assume the climate is stationary, that is, the future climate will behave in the same way as the past climate. Anthropogenic climate change implies that the assumption of stationarity is no longer valid. Sub-daily precipitation is expected to change with greater precipitation intensities associated with warmer temperatures, causing greater flood related extremes and disasters. This thesis examines the relationship between extreme precipitation and temperature and proposes to use the relationship between precipitation and temperature to simulate sub-daily precipitation for a future warmer climate. Quantile regression is presented as a superior alternative to current binning techniques in quantifying the relationship between precipitation and temperature. It is found the relationship between precipitation intensity and temperature is modulated by storm duration. Using precipitation from an accumulation of differing storm duration results in a different relationship to when individual storm durations are considered. It is shown that, at higher temperatures, storm patterns are temporally less uniform with more precipitation occurring in a shorter duration. Likewise, the spatial pattern of precipitation also changes with more moisture concentrated in the storm centre at higher temperatures. The results suggest a change to flood peaks at higher temperatures, however, an investigation of the scaling relationship of streamflow and temperature presented little evidence of greater discharges at higher temperatures. It is concluded that antecedent conditions are likely to dominate flooding in a future climate with only the most extreme storms dominated by changes to the flood causing precipitation. Two non-stationary Poisson process continuous sub-daily precipitation models are presented. The first is conditioned on climatic state and the second on temperature, presenting methodologies that can be used to generate precipitation sequences that better reflect the future climate. The thesis concludes by arguing for the use of the alternatives presented here as a basis for planning and designing water resources infrastructure in future settings.
Persistent link to this record
Link to Publisher Version
Link to Open Access Version
Additional Link
Author(s)
Wasko, Conrad
Supervisor(s)
Sharma, Ashish
Mehrotra, Rajeshwar
Creator(s)
Editor(s)
Translator(s)
Curator(s)
Designer(s)
Arranger(s)
Composer(s)
Recordist(s)
Conference Proceedings Editor(s)
Other Contributor(s)
Corporate/Industry Contributor(s)
Publication Year
2016
Resource Type
Thesis
Degree Type
PhD Doctorate
UNSW Faculty
Files
download public version.pdf 10.61 MB Adobe Portable Document Format
Related dataset(s)