Abstract With a growing awareness of the importance of near-road air pollution and an increasing population of near-road pedestrians, it is imperative to “nowcast” near-road air quality conditions to the general public. This necessitates the building hourly predictive models that are both accurate and easy to use. This study demonstrates an approach to model the hourly near-road Black Carbon (BC) concentrations given on-road traffic information and current meteorological conditions using datasets from two urban sites in Seattle, Washington. The optimal set of prediction variables is determined with a Bayesian Model Averaging (BMA) method and three different model structures are further developed and compared by goodness-of-fit. An innovative approach is proposed to translate wind direction from numerical values to categorical variables with statistical significance. By modeling the autocorrelation within the BC time series using an AR(1) component, the model achieves a satisfactory prediction accuracy. The conditional heteroscedasticity and heavy-tailed distribution of the model residuals are successfully identified and modeled by the General Auto Regressive Conditional Heteroscedasticity (GARCH) model, which provides valuable insights to the interpretation of prediction results. The methodological procedure demonstrated in selecting and fine-tuning the model is computationally efficient and valuable for further implementation onto online platforms for near-road BC nowcasting. A comparison between the two sites also reveals the effectiveness of local freight regulation for mitigating the environmental impacts from a heavy truck fleet.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Coherent approach for modeling and nowcasting hourly near-road Black Carbon concentrations in Seattle, Washington


    Beteiligte:
    Yu, Runze (Autor:in) / Liu, Xiaoyue Cathy (Autor:in) / Larson, Timothy (Autor:in) / Wang, Yinhai (Autor:in)


    Erscheinungsdatum :

    2014-01-01


    Format / Umfang :

    12 pages




    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch





    T2007 - Seattle, Washington

    T2007 / International Council on Alcohol, Drugs and Traffic Safety / International Association of Forensic Toxicologists | TIBKAT | 2007

    Freier Zugriff

    Port of Seattle, Washington

    Engineering Index Backfile | 1932


    Port of Seattle Washington

    Engineering Index Backfile | 1938