Estimation of Data Uncertainty in the Absence of Repetition Experiments

Abstract It may not always be possible to replicate experiments to estimate data uncertainty due to considerable costs, especially in field conditions. This paper introduces a novel approach to estimate data uncertainty where replicate measurements are not available. The approach utilizes the residuals of regression models constructed using different independent variables. The modified residuals are aggregated to estimate the probability distribution of dependent-variable data uncertainty. The approach is applied to estimate erosion-extent measurement uncertainty. The results reveal that the uncertainty estimates are in good agreement with expert opinions and with values given in literature.