Cesari R., Paradisi P., Allegrini P. A trajectory statistical method for the identification of sources associated with concentration peak events. In: HARMO15 - 15th International Conference on Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes (Madrid, Spain, 6-9 May 2013). Proceedings, article n. H15-191. Department of Environmental Science, Aarhus University (Denmark), 2013. |

Abstract (English) |
We briefly review the Trajectory Statistical Methods (TSMs) most used in literature for source identification, essentially based on the concept of Residence Time. Then, we introduce a statistical methodology that, starting from the Concentration Field method, takes into account only the peak values in the concentration time series measured at multiple receptor sites. We use virtual simulations to evaluate the performance of our approach. In order to derive concentration time series at multiple receptors, the Lagrangian Dispersion Model (LSM) FLEXPART is used, in the time forward mode, to simulate dispersion from a known emission source. Then, virtual concentration data are available in the receptor sites. As in many TSMs, Residence Times need to be computed and, to this goal, we use FLEXPART, but in the backward mode, that is, FLEXPART is applied to compute the backward trajectories from the receptor sites. Then, our proposed statistical method is applied to the computed Residence Times and to the concentration data to reconstruct the spatial distribution of emission sources. The numerical results show that our approach could overcome the problem of ghost sources. Further, the proposed method requires simulation times shorter that those required in other methods, since it makes use of a relatively small set of trajectories. This could be of some interest in the characterization of impact studies and local climatic scenarios. | |

URL: | http://www.harmo.org/Conferences/Proceedings/_Madrid/publishedSections/H15-191.pdf | |

Subject | Environmental pollution Source identification Trajectory-based statistical methods Backward trajectories Lagrangian dispersion models J.2 PHYSICAL SCIENCES AND ENGINEERING G.3 PROBABILITY AND STATISTICS I.5.4 PATTERN RECOGNITION. Applications I.4.6 IMAGE PROCESSING AND COMPUTER VISION. Segmentation 86Axx Geophysics 47N30 Applications in probability theory and statistics 46N55 Applications in statistical physics 62-xx Statistics |

1) Download Document PDF |

Open access Restricted Private