Forecasting of Ozone Exposure using Temperature Data


  • Author: Xinyi Lu, Alan E. Gelfand and David M. Holland
  • Date: 07 June 2019
  • Copyright: Image copyright of Patrick Rhodes

Rigorous and rapid assessment of ambient ozone exposure is important for informing the public about ozone levels that may lead to adverse health effects. In a paper published in Environmetrics, the authors use hierarchical modeling to enable real‐time forecasting of 8‐hr average ozone exposure. This contrasts with customary retrospective analysis of exposure data. Specifically, their contribution is to show how incorporating temperature data in addition to the observed ozone can significantly improve forecast accuracy, as measured by predictive performance and empirical coverage.

The paper is available via the link below and the authors explain their findings in further detail below:

Local real‐time forecasting of ozone exposure using temperature data

Xinyi Lu, Alan E. Gelfand and David M. Holland

Environmetrics, Volume 29, Issue 7, November 2018, e2509

thumbnail image: Forecasting of Ozone Exposure using Temperature Data

There is now growing interest in providing rapid access to real-time air pollution monitoring data and using these data to provide air quality forecasts for local communities. Real-time forecasts and personalized smartphone notifications could allow people to modify their behaviour to help reduce pollutant exposures. In the past, continuous, long-term measurement of ambient air pollution has been limited by monitoring restrictions and resource constraints. Recently, there have been monitoring advances in the development of stationary air measurement systems that can provide real-time data collection and can be deployed in community environments. To improve the public understanding of local air pollution, the U.S. Environmental Protection Agency (EPA) developed the Village Green Project ( The primary monitoring goal is to provide communities with previously unavailable current pollution information about air quality as well as to promote community awareness of air pollution. This paper presents single site real-time forecasting using hourly ozone data collected at the Village Green Durham Station during May and June, 2015.

We focus on forecasting 8-hour average ozone concentrations (ppb) defined as the average of the previous four hourly concentrations, the current hour, and the next 3 hours. This statistic is chosen in accord with its interest in the environmental and epidemiological communities. The EPA's National Air Quality Standard is based on 4th maxima of 8-hour ozone averages ( and 8-hour averages are used in the computation of EPAs air quality index (AQI) ( We provide forecasting through the development of appropriate hierarchical models, specifically two-stage autoregressive models. We also examine incorporation of periodicity and heterogeneity of variance and compare a variety of candidate models by evaluating their predictive performance. Models including temperature as a covariate are shown to outperform autoregressive models with just ozone. All of these models can be implemented in real-time, thus providing almost immediate forecasts of 8-hour average ozone concentrations at a Village Green site.

Related Topics

Related Publications

Related Content

Site Footer


This website is provided by John Wiley & Sons Limited, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ (Company No: 00641132, VAT No: 376766987)

Published features on are checked for statistical accuracy by a panel from the European Network for Business and Industrial Statistics (ENBIS)   to whom Wiley and express their gratitude. This panel are: Ron Kenett, David Steinberg, Shirley Coleman, Irena Ograjenšek, Fabrizio Ruggeri, Rainer Göb, Philippe Castagliola, Xavier Tort-Martorell, Bart De Ketelaere, Antonio Pievatolo, Martina Vandebroek, Lance Mitchell, Gilbert Saporta, Helmut Waldl and Stelios Psarakis.