Discretization issues in travel time calculation

Differences in the calculation of the time of concentration using the Natural Resources Conservation Service (NRCS) velocity method result from different degrees of discretization along the longest flow path in the watershed. We examine an idealized system for which an analytical solution can be derived. Next, we study a data set compiled from watersheds across the State of Maryland. Finally, the writers model a channel from a Vermont watershed for which field observations have been observed. In each case the writers show that the time of concentration estimate increases as the level of discretization increases. The writers develop a predictive equation that estimates the high discretization level time of concentration as a function of a single-segment time of concentration estimate and information derived from a high resolution longitudinal slope profile. The writers also show that observed travel times tend to fall in between the single-segment derived travel time and the travel time derived from high resolution data. By relaxing the uniform flow assumption and instead allowing for gradually varied flow, good agreement between modeled and observed travel times is achieved. These results have significant implications with regards to deriving time of concentration from increasingly greater resolution geographic information systems data.