A study of atmospheric diffusion from the LANDSAT imagery

Abstract

Detailed analyses of the LANDSAT multispectral scanner (MSS) data of the smoke plumes that originated in eastern Cabo Frio (22° 59′S; 42° 02′W) and crossed over into the Atlantic ocean are presented to illustrate how high‐resolution LANDSAT imagery can aid meteorologists in evaluating specific air pollution events. The eleven LANDSAT images selected are for different months and years. Conventional interpretation techniques are applied to analyze the images with a view to arrive at certain plume characteristics. The analysis of the visible smoke plumes revealed that the plume was 130 km long and attained a maximum width of 937 m, 10 km away from the chimney emitting the effluent. The results show that diffusion is governed primarily by water and air temperature differences. With colder water, low‐level air is very stable and the vertical diffusion is minimal; but water warmer than the air induces vigorous diffusion. The applicability of two empirical methods for determining the horizontal eddy diffusivity coefficient (Ky) in the Gaussian plume formula was evaluated with the estimated standard deviation of the crosswind distribution of material in the plume (σy) from the LANDSAT imagery. Most consistent estimates for Ky are obtained from the formula based on Taylor's theory of ‘diffusion by continuous moment’. Ky values of about 158 m2 s−1 in quasi‐neutral conditions and 49 m2 s−1 in stable conditions are obtained from a plot of σ2y as a function of distance from the source. The rate of kinetic energy dissipation (ε) is evaluated from the diffusion parameters σy and Ky. The ε value ranges from 0.1×10−5 m2 s−3 to 80.2×10−5 m2 s−3 in quasi‐neutral and stable stratifications. These results compare well with the previous experimental values obtained over water surfaces by various workers. They form a data base for use in the development and validation of mesoscale atmospheric diffusion models.