In geophysics, especially river engineering, any processes may be considered occurring in several time scales. For example, while sediment or contaminant transport is studied, it is preferred to follow the problem in each second. On the other hand, the unstable morphology of a river is a problem of decades and centuries. Identification of the presence of scaling in the river flow processes has been a challenging problem in hydrology. While large number of unidentified dominant variables affect river flow, it is common to consider it as a stochastic process. However, recent studies indicate that by employing nonlinear deterministic and chaotic dynamic concepts, the river flow process could easily become an outcome of a deterministic system with only a few dominant variables. This paper attempts to investigate river flow in four time scales (15 minute, hourly, daily and weekly) utilizing the flow data recorded for the Uncompahgre River in Colorado, USA. The correlation dimension was used as an indicator to distinguish between chaotic and stochastic behavior. It is concluded that the flow was completely chaotic in small time scales (15 min), while it tended to be more stochastic in large time scales.
)
