Sediment yield estimation in small watersheds based on streamflow and suspended sediment discharge measurements

In the past, the watershed sediment yield estimation, in rivers where hydrometric information is available, has been done using the sediment rating curve, which is obtained fitting theoretical functions to the water and sediment discharge relationship. Once this relationship is defined, it can be used along with the flow duration curve of a given year (or runoff-frequency data) to obtain the annual sediment yield prediction. Sometimes, specially in recent years, a better estimation has been developed using the same procedure but replacing the flow duration curve for the daily flow series, transforming each daily flow in suspended sediment discharge, by means of the sediment rating curve, and accumulating this information to monthly and annual levels. This procedure is judged adequate for large watersheds, where the differences among the mean daily flow and the maximum and minimum streamflow during the day are not very large. Generally the equations that fit best the sediment rating curves are highly nonlinear, therefore, in small watersheds where the differences among mean daily flows and the extreme instantaneous flows during the day are important, working with mean daily flows (and even worse with flow duration curves) introduces large errors in sediment yield estimations. In addition to this, there also exists other problems related to the previous procedure, namely the extrapolation errors, that can be very important. Also the quality of the basic information is of paramount importance in this type of analysis. This paper discusses all the above problems analysing case studies related to two important Venezuelans reservoirs: Dos Cerritos in the Tocuyo River and Cumaripa in the Yaracuy river. An evaluation is made of the effects of these sources of errors in the sediment yield estimation of both cases.