Sediment-yield prediction using map correlation method

Abstract

Sediment yield is of utmost importance in estimating soil erosion, reservoir sedimentation, aquatic life protection, change in river cross-section, and many more. Daily sediment time series can be readily obtained from sediment gauging stations; the problem lies at ungauged sites where information on daily sediment data is unavailable. At ungauged catchments, the most common methods to estimate sediment time series involve the use of a reference sediment gage, which shifts the properties of the sediment flow time series at a reference station to the ungauged catchment. The appropriate selection of reference stations is challenging in the estimation of daily sediment flows at ungauged watersheds. At most times in conventional methods, the parameters at the ungauged sites are determined by assuming the nearest gage station as a reference station. In the present paper, it is established that the near neighbour catchment may not always be the most correlated one. One of the techniques termed as map correlation method is undertaken to identify the most correlated sediment gage station as a reference to an ungauged station. To assess the efficacy of the present method daily sediment data for two years at forty sediment gauging stations of the USA watershed are utilized. Sample and model variograms were used to model the spatial structure of Pearson's r correlation coefficient. Observed and estimated correlations resulting from the cross-validation of the variograms are in line with each other for most of the stations. This method selects the most related reference station for an ungauged sediment site.