Application of Functional Data Analysis in Streamflow Hydrograph
Streamflow data is often recorded at discrete time intervals, such as hourly, daily, monthly or annually and a hydrograph is normally used to represent the temporal variation of these flows in a graphical form. In practice, a river may have various shapes of flood hydrographs. The shape of a hydrograph varies in each river basin and each individual storm event. The aim of this study is to apply a functional data concept in hydrological applications using streamflow hydrographs as functional data. An entire hydrograph curve with respect to time can be considered as a single observation within the functional context. To analyse a streamflow hydrograph, functional descriptive statistics, functional principal components and functional outliers are among the functional data analysis tools introduced in this study. The functional principal component was adapted to find new functions that reveal the most important type of variation in the hydrograph curve, while the graphical methods (namely the rainbow plots) were used to visualise the functional data. The functional highest density region box-plot is employed to identify functional outliers. These methods were applied to the case study of flood analysis at the Sg. Kelantan River Basin, Malaysia. In conclusion, the functional framework is found to be more flexible in analysing the whole hydrograph and is able to make full use of information contained in the hydrograph.
KeywordsStreamflow Hydrograph Functional data analysis Rainbow plot Outlier
- 4.Ternynck, C., Alaya, M., Chebana, F., Dabo-Niang, S., Ouarda, T.B.M.J.: Streamflow hydrograph classification using functional data analysis. J. Hydrometeorol. (2015). https://doi.org/10.1175/jhm-d-14-0200.1