Abstract. In this paper, an automated algorithm is
developed, which is used to identify the spectral gap during the heavy haze
pollution process, reconstruct acquired data, and obtain pure turbulence
data. Comparisons of the reconstructed turbulent flux and eddy covariance
(EC) flux show that there are overestimations regarding the exchange between
the surface and the atmosphere during heavy haze pollution episodes. After
reconstruction via the automated algorithm, pure turbulence data can be
obtained. We introduce a definition to characterize the local intermittent
strength of turbulence (LIST). The trend in the LIST during pollution
episodes shows that when pollution is more intense, the LIST is smaller, and
intermittency is stronger; when pollution is weaker, the LIST is larger, and
intermittency is weaker. At the same time, the LIST at the city site is
greater than at the suburban site, which means that intermittency over the
complex city area is weaker than over the flat terrain area. Urbanization
seems to reduce intermittency during heavy haze pollution episodes, which
means that urbanization reduces the degree of weakening in turbulent exchange
during pollution episodes. This result is confirmed by comparing the average
diurnal variations in turbulent fluxes at urban and suburban sites during
polluted and clean periods. The sensible heat flux, latent heat flux,
momentum flux, and turbulent kinetic energy (TKE) in urban and suburban areas
are all affected when pollution occurs. Material and energy exchanges between
the surface and the atmosphere are inhibited. Moreover, the impact of the
pollution process on suburban areas is much greater than on urban areas. The
turbulent effects caused by urbanization seem to help reduce the consequences
of pollution under the same weather and pollution source condition, because
the turbulence intermittency is weaker, and the reduction in turbulence
exchange is smaller over the urban underlying surface.