Recently, manual observation sequence has been gradually replaced by automatic observation sequence. The difference between manual observation sequence and automatic observation sequence is somewhat inevitable. This challenges the the homogeneity and the continuity of historical weather data, and influences atmospheric researches and applications. Therefore, based on the understanding of the influence caused by the two observation sequences, how to modify the data sequence of manual observation to automatic observation sequence has become a problem. In this paper, a model, which is a neural network based on the particle swarm optimization technique (PSONN), is established to modify the wind speed data sequence from manual observation to automatic observation. The proposed model achieves 15.6% in mean absolute percentage error (MAPE) compared to manual observation data sequence. For wind speed, it could be a promising candidate for modifying manual observing data sequence to automatic observing data sequence.

An atmospheric tracer dispersion study known as Urban Meteorological and Tracer Experiment was carried out in Yuzhong County, Gansu province in northwest of China, during October of 2011. By using the observational data, the applicability of CALPUFF model based on Gaussian formula was verified within urban area. The results show that (i) CALPUFF model can reasonably reflect the actual observations on the whole, and the high value centers of tracer concentration was changed with wind direction over the urban canopy layer. (ii) The simulated plume central axis has a bias of 10-15° with observations, and the simulation error is very much big at the single point, especially at the ground.