This paper considers the problem of search and size estimation of local underwater heterogeneities in a given area. The authors introduce an approach allowing to localize such heterogeneities and to estimate their dimensions and the mass of a substance dissolved. Moreover, a task planning method is suggested for a group of autonomous underwater vehicles (AUVs) solving this problem. The corresponding discrete optimization problem is stated and two solution algorithms are compared. The authors also discuss replanning of group actions in case of emergencies during an operation. And finally, the results of simulation experiments for the proposed algorithms are provided.