Trends in stratospheric ozone profiles using functional mixed models

Abstract

This paper is devoted to the modeling of altitude-dependent patterns of ozone variations over time. Umkher ozone profiles (quarter of Umkehr layer) from 1978 to 2011 are investigated at two locations: Boulder (USA) and Arosa (Switzerland). The study consists of two statistical stages. First we approximate ozone profiles employing an appropriate basis. To capture primary modes of ozone variations without losing essential information, a functional principal component analysis is performed as it penalizes roughness of the function and smooths excessive variations in the shape of the ozone profiles. As a result, data driven basis functions are obtained. Secondly we estimate the effects of covariates – month, year (trend), quasi biennial oscillation, the Solar cycle, arctic oscillation and the El Niño/Southern Oscillation cycle – on the principal component scores of ozone profiles over time using generalized additive models. The effects are smooth functions of the covariates, and are represented by knot-based regression cubic splines. Finally we employ generalized additive mixed effects models incorporating a more complex error structure that reflects the observed seasonality in the data. The analysis provides more accurate estimates of influences and trends, together with enhanced uncertainty quantification. We are able to capture fine variations in the time evolution of the profiles such as the semi-annual oscillation. We conclude by showing the trends by altitude over Boulder. The strongly declining trends over 2003–2011 for altitudes of 32–64 hPa show that stratospheric ozone is not yet fully recovering

Application of Functional Data Analysis for the Treatment of Missing Air Quality Data

“…The application of FDA for the air quality data has been identified. Among the areas of the application include the study on the trend, severity and the dynamic behaviour of a particular pollutant such PM 10 or ozone (Gao & Niemeier 2008;Park et al 2013;Shaadan et al 2012), the study on the extreme values and the prediction of pollutant curves (Quintela-del-Rio & Francisco-Fernandez 2011), as well as outliers or anomalies detection and assessment (Martinez et al 2014;Shaadan et al 2015;Torres et al 2011). Noticeably, the application of FDA for the treatment of missing values is rarely found.…”