The problem of detecting emission sources and restoring the field of concentrations on a local scale was formalized using the data of observations (the inverse problem). The suggested statement supposes minimization of a squared discrepancy between the experimental and calculated data taking into account the balance between emissions of a pollutant over some period, its deposition on a local area, and the pollutant part transported over long distances. The observational data on chemical composition of the snow cover in industrial, agricultural, and background points of the region are used as initial ones. The observations were made at a checkpoint (Khomutovo village). The results allowed the surface, simulating the actual field of deposition, to be constructed and approximated by analytical function. It was established that for several sources of emissions the problem of calculating the coefficients is multi-extreme. The problem of searching for a global extreme of the function was solved numerically. Its real coefficients were calculated using the data on chemical analysis of the deposition (snow cover) at a reference site.