Correction of measurement errors on sediment concentration sampled by stirring-sampling method from traditional runoff collection tanks

Stirring-sampling method is a widely adopted method to measure sediment concentrations in collection tanks of runoff plots, but with high systematic measurement errors. This research aimed to advance an approach for building correction equations to remove measurement errors in designed sediment concentration range. Experimental data of sediment measurement from the stirring-sampling method, with four representative soils, under the designed sediment concentrations (1, 2, 5, 8, 10, 20, 50, 80, 100, 200, 500, 800, and 1000 kg/m3) were used to demonstrate the correction methodological process. Two correction methods (step-wise correction and universal correction) were suggested for the trial in this study based on the distribution of measurement errors. In the step-wise correction, the correction equations were made with a series of linear functions without intercept for the low concentration group (0-20 kg/m3), a series of linear functions with intercept for the high (20-200 kg/m3) and extremely high (200-1000 kg/m3) concentration groups, consecutively. The correction equations were a series of power functions in the universal correction. For the step-wise correction, most of the relative errors of correction sediment concentrations were smaller than 15% and 10% under high and extremely high concentration groups, but the corrected accuracy was not good in the sediment concentration of 1, 2, 5 kg/m3 with the corrected relative errors of 0.20%-206.07%. For the universal correction, the corrected relative errors (0.19%-31.81%) of the four soils were low under the condition of extremely high sediment concentrations, but other corrected accuracies weren’t good with the corrected relative errors of 0.68%-1154.71%. The corrected accuracy of step-wise correction is higher than that of the universal correction, but the universal correction is more convenient. These results indicated that the correction equations could efficiently revise the measurement errors of the tested soils and that this method can be generalized to other soil types and was meaningful in monitoring soil erosion.