Submitted to: Federal Interagency Sedimentation Conference Proceedings
Publication Type: Proceedings
Publication Acceptance Date: January 6, 2006
Publication Date: April 2, 2006
Citation: Dabney, S.M., Locke, M.A., Steinriede Jr, R.W. 2006. Turbidity sensors track sediment concentrations in runoff from agricultural fields. Federal Interagency Sedimentation Conference Proceedings, April 2-6, 2006, Reno, Nevada. CDROM.
Interpretive Summary: Monitoring soil loss from agricultural fields is an expensive undertaking that involves analysis of numerous samples. We conducted laboratory and field experiments to determine if optical backscatter (OBS) turbidity measurements could provide improved estimates of erosion processes for three 15-ha fields in the Mississippi Delta. Our results demonstrate that OBS data elucidated the dynamics of the erosion process of a Sharkey silty clay soil and permitted prediction of sediment yield from the fields within approximately 0.5 t per ha per year. Our results are important to scientists and agencies interested in monitoring sediment erosion and sediment yield both from the standpoint of erosion and TMDL studies.
Optical backscatter (OBS) turbidity sensors offer the opportunity to obtain a continuous record of soil loss if turbidity is well correlated with sediment concentration. The relationship between turbidity and sediment concentration depends on several factors, including: particle size, particle shape, and particle color. As watershed size is reduced to the scale of individual fields or plots, variability in particle composition is reduced and the relationship between turbidity and concentration may become stable. We undertook studies to determine this relationship for several fields using OBS-3 turbidity sensors. We compared concentration - turbidity relationships for both suspensions of dried soil samples and sequential natural rainfall runoff samples from three 15-ha fields. We obtained strong correlations between turbidity and concentration, but the relationships differed between natural runoff and resuspended samples. Overall, turbidity explained 95% of the variability in sediment concentration observed in natural runoff samples from one of the fields that was used for calibration. When the resulting regression relationship was applied to two adjacent but independent fields, accuracy of prediction was similar. Prediction accuracy was only marginally improved by consideration of additional parameters including discharge, rate of change of discharge, and time within runoff event, indicating little hysteretic behavior. Peak turbidity usually preceded both peak flow rate and the time of collection of the first sequential sample. We conclude that calibrated optical backscatter turbidity sensors placed in edge-of-field grade control pipes have good potential for continuously monitoring soil loss and improving field-scale soil loss estimates with an expected accuracy estimated to be about 0.5 t ha-1 y-1. When combined with measured concentrations, turbidity data they may provide an indication of the particle size distribution of sediment in transport.