Establishing multiple regression models for ozone sensitivity analysis to temperature variation in Taiwan

Pao Wen Grace Liu, Jiun Horng Tsai, Hsin Chih Lai, Der Min Tsai, Li Wei Li

Research output: Contribution to journalArticlepeer-review

14 Citations (Scopus)

Abstract

Sensitivity of meteorological variation to air quality has attracted people's attention since climate change became a world issue. The goal of this study is to investigate the sensitivity of ground-level ozone concentrations to temperature variation in Taiwan. Several multivariate regression models were built based on historical data of ozone and meteorological variables at three cities located in northern, mid-western, and southern Taiwan. Results of descriptive statistics indicate that the severe pollution from the highest to the minor conditions following by the order of the southern (Pingtung), mid-western (Fengyuan), and the northern sites (Hsichih). Multiple regression models containing a principal component trigger variable effectively simulated the historical ozone exceedance during 2004-2009. Inclusion of the PC trigger were improved R2 from the lowest 0.38 to the highest 0.58. High probability of detection and critical success index (mostly between 85% and 90%) and low false alarm rates (0-2.6%) were achieved for predicting the high ozone days (≧100ppb). The results of sensitivity analysis indicated that (1) the ozone sensitivity was positively correlated with the temperature variation, (2) the sensitivity levels were opposite to that of the ozone problem severity, (3) the sensitivity was mostly apparent in ozone seasons, and (4) the sensitivity strongly depended on the seasonality in the urban cities Hischih and Fengyuan, but weakly depended on seasonality in the rural city Pingtung.

Original languageEnglish
Pages (from-to)225-235
Number of pages11
JournalAtmospheric Environment
Volume79
DOIs
Publication statusPublished - 2013 Nov

All Science Journal Classification (ASJC) codes

  • Environmental Science(all)
  • Atmospheric Science

Fingerprint Dive into the research topics of 'Establishing multiple regression models for ozone sensitivity analysis to temperature variation in Taiwan'. Together they form a unique fingerprint.

Cite this