Methodology for development of drought severity-duration-frequency (SDF) Curves

Rahmat, S 2015, Methodology for development of drought severity-duration-frequency (SDF) Curves, Doctor of Philosophy (PhD), Civil, Environmental and Chemical Engineering, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Rahmat.pdf Thesis Click to show the corresponding preview/stream application/pdf;... 6.51MB
Title Methodology for development of drought severity-duration-frequency (SDF) Curves
Author(s) Rahmat, S
Year 2015
Abstract Drought monitoring and early warning are essential elements impacting drought sensitive sectors such as primary production, industrial and consumptive water users. A quantitative estimate of the probability of occurrence and the anticipated severity of drought is crucial for the development of mitigating strategies. The overall aim of this study is to develop a methodology to assess drought frequency and severity and to advance the understanding of monitoring and predicting droughts in the future. Seventy (70) meteorological stations across Victoria, Australia were selected for analysis. To achieve the above objective, the analysis was initially carried out to select the most applicable meteorological drought index for Victoria. This is important because to date, no drought indices are applied across Australia by any Commonwealth agency quantifying drought impacts. An evaluation of existing meteorological drought indices was first conducted to assess their suitability for the determination of drought conditions. The use of the Standardised Precipitation Index (SPI) was shown to be satisfactory for assessing and monitoring meteorological droughts in Australia.

Temporal changes in historic rainfall variability and the trend of SPI were investigated using non-parametric trend techniques to detect wet and dry periods across Victoria, Australia. The first part of the analysis was carried out to determine annual rainfall trends at five selected meteorological stations with long historical records, as well as a short sub-set period (1949-2011) of the same data set. Temporal trends in the rate of occurrence of drought events (i.e. inter-arrival times) were examined. Most of the stations showed negative slopes indicated that the intervals between events were becoming shorter and the frequency of events was temporally increasing. Based on the results obtained from the preliminary analysis, the trend analyses were then carried out for the remaining 65 stations. The main conclusions are the trend analysis was observed to be highly dependent on the start and end dates of analysis and from the SPI and inter-arrival drought trends, it was observed that some of the study areas in Victoria will face more frequent dry period leading to increased drought occurrence.

A novel concept centric on drought severity-duration-frequency (SDF) curves was successfully derived for all the 70 stations using an innovative threshold approach. Using regionalisation techniques, the study area was separated into homogenous groups based on rainfall characteristics. A set of mean SDF curves was developed for each cluster to identify the frequency and severity of the risk of drought events for various return periods in each cluster. Non-homogeneous Markov Chain modelling was used to estimate the probability of different drought severity classes and drought predictions 1, 2 and 3 months ahead. Overall, this model predicted drought situations 1 month ahead well. However, predictions 2 and 3 months ahead should be used with caution.

Many parts of Australia including Victoria have experienced their worst droughts on record over the last decade. The information on the probability of occurrence and the anticipated severity of drought will be helpful for water resources managers, infrastructure planners and government policy-makers.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Civil, Environmental and Chemical Engineering
Keyword(s) Drought frequency analysis
Markov Chain modelling
Standardised Precipitation Index (SPI)
Trend analysis
Version Filter Type
Access Statistics: 418 Abstract Views, 1454 File Downloads  -  Detailed Statistics
Created: Fri, 26 Jun 2015, 09:10:32 EST by Denise Paciocco
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us