Sediment transport process in mountain streams has been studied chiefly in the development of theoretical as well as experimental bed load equations. Difficulties associated with field observation have been a major obstacle in testing and applying bed load equations properly. Direct sediment sampling is often impossible; and its effective is limited due to its durability even if it is applicable. Attempts have been made to utilize more indirect but stable monitoring methods (hereafter �gindirect method") in recently years such as a hydrophone sediment discharge measuring system (hereafter �ghydrophone system"). Hydrophone systems count the times that bed load sediments strike the acoustic sensor of the system (hereafter �gpulses") upon proper electric amplification. Early observation brought light on small to medium discharge non�]equilibrium sediment transport phenomena which have not been adequately studied either theoretically or experimentally. Non�]equilibrium states are influenced by many hydro�]sediment factors, among which flow discharge is not always dominant. Therefore, sediment discharge as a functional form needs to be regarded as dependent on multiple variables. Expanding bed load equations and building ones from many primary hydro�]sediment factors at one time seem to be beyond our reach. Thus, a step�]wise approach is taken by introducing and adding to flow discharge an intermediary and secondary hydro�]sediment factor called �gsediment�]related quantity," which is observed by indirect methods.
In this study, hydrophone systems have been installed in 100 and 200km2�]scale river basins in order to observe hydrophone pulses that are used as sediment�]related quantity. Sediment transport, water flow, and acoustic phenomenon are mutually intercorrelated. The phenomena were described by each corresponding observational variables, whose correlations were analyzed statistically. Hydrophone pulses have correlation with and are dependent upon both bed load discharge and flow discharge. Therefore, pulses alone can not provide suitable estimates of sediment discharge. In order to search analytical estimation equations of bed load, therefore, pulses and flow discharges were combined as an initial step. Additive forms both with and without interaction terms, and multiplicative forms were introduced and calibrated for the observation. Each bed load analytical estimation method was statistically assessed. The best method was a linear additive form with no interactive term. The introduction of sediment�]related quantities facilitates us to construct a viable analytical estimation method of non�]equilibrium sediment discharge. Its own dependency on other hydro�]sediment factors needs to be examined for further physical understanding of the sediment discharge observation. Indirect method with the suggested forms of analytical estimation equations, with due care for its insufficient physical understanding, seems to be applicable for further development to analyze bed load discharge consecutively in the scale of full river section.


Key words�Fsediment discharge, hydrophone, bed load estimation, direct sampling