A modified variational algorithm for data assimilation of measurements in models of transport of passive admixture

Abstract

In one-dimension transport model of passive admixture with the input parameters used in the model for the Azov sea is considered a modified variational algorithm for identifying the initial concentration fields of pollutants. The algorithm of data measurements assimilation based on the minimization of a quadratic coast functional of the prediction, characterizing deviation of the model solution from measurement data. Transport model of a passive admixture acts as limiters on variation of input parameters. The paper also compare the modified approach with the classical variational algorithm of identification of initial fields concentration. The field was found in good agreement with the known initial field as by location and total concentration of admixtures in the initial field. Modified assimilation algorithm of data measurement based on the evaluation method and the decision a series of adjoint tasks that can significantly reduce the processor time due to solving the problem of identification. This approach allows for the implementation of the algorithm in parallel. As a result of numerical experiments shown that the modified algorithm allows for more operational identify initial fields of concentration of pollutants. In General, numerical experiments have shown a reliable performance of the considered algorithms. The results can be used to solve various tasks of environmental directionally due to studying the effect of pollution sources of anthropogenic nature in the waters of the Azov and Black seas.