New classification scheme for ozone monitoring stations based on frequency distribution of hourly data

O Tapia; M Escudero; Á Lozano; J Anzano; E Mantilla

doi:10.1016/j.scitotenv.2015.11.081

New classification scheme for ozone monitoring stations based on frequency distribution of hourly data

Sci Total Environ. 2016 Feb 15:544:1-9. doi: 10.1016/j.scitotenv.2015.11.081. Epub 2015 Dec 1.

Authors

O Tapia¹, M Escudero², Á Lozano¹, J Anzano³, E Mantilla⁴

Affiliations

¹ Centro Universitario de la Defensa (CUD) de Zaragoza, Academia General Militar, Ctra. de Huesca s/n, E-50090 Zaragoza, Spain.
² Centro Universitario de la Defensa (CUD) de Zaragoza, Academia General Militar, Ctra. de Huesca s/n, E-50090 Zaragoza, Spain. Electronic address: mescu@unizar.es.
³ Laser Laboratory and Environment, Department of Analytical Chemistry, Faculty of Sciences, University of Zaragoza, Pedro Cerbuna 12, E-50009 Zaragoza, Spain.
⁴ Instituto Universitario Centro de Estudios Ambientales del Mediterráneo (CEAM-UMH), Parque Tecnológico C/ Charles R. Darwin 14, E-46980 Paterna, Spain.

PMID: 26654968
DOI: 10.1016/j.scitotenv.2015.11.081

Abstract

According to European Union (EU) legislation, ozone (O3) monitoring sites can be classified regarding their location (rural background, rural, suburban, urban) or based on the presence of emission sources (background, traffic, industrial). There have been attempts to improve these classifications aiming to reduce their ambiguity and subjectivity, but although scientifically sound, they lack the simplicity needed for operational purposes. We present a simple methodology for classifying O3 stations based on the characteristics of frequency distribution curves which are indicative of the actual impact of combustion sources emitting NO that consumes O3 via titration. Four classes are identified using 1998-2012 hourly data from 72 stations widely distributed in mainland Spain and the Balearic Islands. Types 1 and 2 present unimodal bell-shaped distribution with very low amount of data near zero reflecting a limited influence of combustion sources while Type 4 has a primary mode close to zero, showing the impact of combustion sources, and a minor mode for higher concentrations. Type 3 stations present bimodal distributions with the main mode in the higher levels. We propose a quantitative metric based on the Gini index with the objective of reproducing this classification and finding empirical ranges potentially useful for future classifications. The analysis of the correspondence with the EUROAIRNET classes for the 72 stations reveals that the proposed scheme is only dependent on the impact of combustion sources and not on climatic or orographic aspects. It is demonstrated that this classification is robust since in 87% of the occasions the classification obtained for individual years coincide with the global classification obtained for the 1998-2012 period. Finally, case studies showing the applicability of the new classification scheme for assessing the impact on O3 of a station relocation and performing a critical evaluation of an air quality monitoring network are also presented.

Keywords: Classification; Frequency distribution; Gini index; Monitoring stations; O(3).

Publication types

Research Support, Non-U.S. Gov't