Inverse estimation of finite-duration source release mass in river pollution accidents based on adjoint equation method

Obtaining the pollutant release mass at a timely manner is crucial in emergency response for river pollution accidents. However, compared to the instantaneous source, release mass estimation of finite-duration source has been rarely studied. In addition, few studies involve the influence of partial observation data and observation data with different levels of noise on inversion results. Based on the adjoint equation method (AEM), this study developed a new release mass estimation model to make up the above deficiencies. In this model, one-dimensional physical transport advection-dispersion equation was used as governing equation to describe pollutant transport and the finite-duration sources and instantaneous sources were both considered. Two synthetic experiments and two field experiments were used to evaluate this model. In synthetic experiments, detailed analysis of the influence of observation errors and incomplete concentration data due to equipment failure was conducted. Results indicate that the effect of observation errors on the inverse estimation results was within the relative error of 12%; the incomplete concentration data could also be used to obtain inverse estimation results. The two field experiments gave confidence to the application of this model in release mass estimation in actual pollution accidents with a relative error within 10%. These findings will assist in the decision-making for dealing with actual river pollution accidents.