Introduction

Abstract

Hydrologic time series analysis plays an important role in managing water resources. Practical applications include the design and operation of hydraulic structures, flood control, water supply management, hydropower generation, and irrigation systems design among others. The task of modeling streamflow can be achieved by following either of two approaches. The first approach is deterministic in nature, which involves the physical simulation of the physical system through numerical models using rainfall as input. The second is stochastic in nature and involves the study of the important statistical properties of streamflow sequences. Besides the difficulties in modeling the relationship between rainfall and runoff, determining future flows requires forecasts of rainfall. One is thus forced either to resort to more complicated meteorological modeling, or to apply stochastic modeling to rainfall. On the other hand, river flows, as stochastic processes, suffer from a smaller degree of randomness than rainfall since the storage in river basins smooths out some of the randomness. Also, large fluctuations in rainfall are dampened by flow passing through river basins. It is therefore desirable to study streamflow series using the stochastic approach. Several attempts have also been made to establish some physical foundation for stochastic streamflow models of different types (see Salas et al., 1988, Chapter 1). KeywordsHydropower GenerationClimatic Time SeriesFractional Gaussian NoiseStreamflow SeriesWater Supply ManagementThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.