Remote Sensing of Temporal and Spatial Variations of Suspended Sediment Concentration in Bahmanshir Estuary, Iran

Suspended Sediment Concentration (SSC) in surface waters affects directly on the water quality, phytoplankton's fertility, pollution distribution and redistribution. In this study, temporal and spatial variations of SSC at Bahmnshir Estuary (BE) in the southwest of Iran were investigated using five field campaigns and Moderate Resolution Imaging Spectroradiometer (MODIS) sensor images of a nine-year time series from 2003 to 2011. An Artificial Neural Network (ANN) model with one hidden layer that had good simulation performance (training: R² = 0.84; N = 23; validation: RMSE of 59 mg L^-1, N = 6) was used against regression analysis (modeling: R² = 0.63; N = 23; validation: RMSE of 261 mg L^-1, N = 6). Results indicated ANN model has higher accuracy than regression analysis to model the SSC because it's higher capability in optimization of non-linear problems. R² and RMSE were improved, 25 and 77% respectively when using ANN model. It was found that average annual SSC for years 2008 to 2011 was 10% more than years 2003 to 2007 and the highest SSC accrued at 2008 with a noticeable decrease in the flow discharge of Karun River in this year. Minimum SSC occurs, at the point in 3.2 km before the Estuary Mouth (EM). This point was defined as the boundary of impact of upstream and downstream on the SSC in the BE. High SSCs occur at EM, at the muddy shore of the Abadan Island on right bank and at the Khoure Mousa on the left bank with 2.8 and 1.8 km distance from the EM respectively. This study demonstrated ANN model and MODIS MOD09GQ products are appropriate tools for monitoring surface SSC dynamics within coastal water environments such as BE.