Rhodes>IWR>Projects>Uncertainty

Uncertainty

Student: Mr T Sawunyama
Supervisor: Prof DA Hughes
Degree: PhD (Hydrology)


The research project is on identifying and quantifying sources of modelling uncertainties (input data, parameters and model structure) and suggesting alternative options of reducing hydrological uncertainties in water resource estimation in Southern Africa. The modelling procedure is through application of a revised version of Pitman hydrological model developed at the Institute for Water Research. Options for managing uncertainty are suggested and under investigation. The progress so far has been to identify and quantify uncertainties in model input (rainfall and evaporation) using simple approaches in selected regions within South Africa and their potential effects on simulated flows. This work was published in International Association of Hydrological Sciences (IAHS) red book, in addition to a presentation at international conference in July 2007 in Perugia, Italy. There is significant progress on the use of alternative data sets to reduce model input uncertainties (through use of satellite-derived rainfall), to provide spatial rainfall estimates in ungauged basins. A paper related to this topic was accepted for publication in WaterSA and will be available in January 2008. The findings were presented at the SANCIAHS conference in September 2007 in Cape Town, South Africa. Further work has also looked at developing a strategy for parameter sensitivity and uncertainty analysis, which is simple and applicable to the model structure we are using. A few case studies in South Africa were tested. The approach allows parameter exploration, giving ensemble model predictions allowing assessment of propagation of parameter uncertainties to prediction uncertainty. The research is promising in giving water resource practitioners guidelines and confidence in the application of the regional water resource model. Future work will include an assessment of the collective effects of the different sources of hydrological uncertainties on estimating sustainable yields. This will provide valuable information for decision-making purposes in water allocation in ungauged basins.

Last Modified: Fri, 14 Aug 2015 10:21:56 SAST