Spatial and Temporal Estimation of Suspended Sediment Load in Aji-chay River Using Geostatistics and Artificial Neural Network

Sediment transport phenomenon in rivers, which has been under the consideration of specialists and water engineers, is one of the complicated problems in river engineering studies. Usually sediment transport and storage that threaten hydraulic structures in rivers are important problems. So presenting new and efficient approaches for accurate estimation of suspended sediment load at different scales will play very important role in river engineering studies. As in most of the sediment gauging stations of the country, sediment sampling is carried out daily and irregularly, if it is needed to know the suspended sediment load in a particular of river, it is necessary to utilize suitable temporal and spacial models. In this study, geostatistics and artificial neural network were used in order to combine time and space series analyses together to present a comprehensive model to estimate monthly suspended sediment load in Aji-chay river. Therefore, rational data has been produced with the aid of artificial neural network at monthly scale, then by both uni and multi-parametric estimators namely kriging and cokriging (in addition to suspended sediment load, water discharge is also used as a secondary variable) methods, monthly suspended sediment load was estimated along the Aji-chay river. Results showed that while both models were valuable in restricted area, the cokriging model in comparison with kriging model was more accurate.