Rob J Hyndman

John L Pearce^a, Jason Beringer^a, Neville Nich­olls^a, Rob J Hyndman^b, Pet­teri Uotila^a, and Nigel J Tap­per^a

^a School of Geo­graphy and Envir­on­mental Sci­ence, Mon­ash Uni­ver­sity, Mel­bourne, Aus­tralia
^b Depart­ment of Eco­no­met­rics and Busi­ness Stat­ist­ics, Mon­ash Uni­ver­sity, Mel­bourne, Australia

Atmo­spheric Envir­on­ment (2011), 45(1), 128–136.

Abstract
The influ­ence of synoptic-​​scale cir­cu­la­tions on air qual­ity is an area of increas­ing interest to air qual­ity man­age­ment in regards to future cli­mate change. This study presents an ana­lysis where the dom­in­ant syn­op­tic ‘types’ over the region of Mel­bourne, Aus­tralia are determ­ined and linked to regional air qual­ity. First, a self-​​organising map (SOM) is used to gen­er­ate a time series of syn­op­tic charts that clas­sify the annual daily cir­cu­la­tion affect­ing Mel­bourne into 20 dif­fer­ent syn­op­tic types. SOM res­ults are then employed within the frame­work of a gen­er­al­ized addit­ive model (GAM) to identify links between synoptic-​​scale cir­cu­la­tions and observed changes air pol­lut­ant concentrations. The GAMs estim­ate shifts in pol­lut­ant con­cen­tra­tions under each syn­op­tic type after con­trolling for long-​​term trends, sea­son­al­ity, weekly emis­sions, spa­tial vari­ation, and tem­poral per­sist­ence. Res­ults showed the aggreg­ate impact of syn­op­tic cir­cu­la­tions in the mod­els to be quite mod­est as only 5.1% of the daily vari­ance in O3, 4.7% in PM10, and 7.1% in NO2 were explained by shifts in syn­op­tic cir­cu­la­tions. Fur­ther ana­lysis of the par­tial resid­ual plots iden­ti­fied that des­pite a mod­est response at the aggreg­ate level, indi­vidual syn­op­tic cat­egor­ies had dif­fer­en­tial effects on air pol­lut­ants. In par­tic­u­lar, increases of up to 40% in NO2 and PM10 and 30% in O3 occur when a syn­op­tic con­di­tions res­ult in a north-​​easterly gradi­ent wind over the Mel­bourne area. Addi­tion­ally, NO2 and PM10 levels also showed increases of up to 40% when a strong high pres­sure sys­tem was centered dir­ectly over the Mel­bourne area. In sum, the uni­fied approach of SOM and GAM proved to be a com­ple­ment­ary suite of tools cap­able of identi­fy­ing the entire range syn­op­tic cir­cu­la­tion pat­terns over a par­tic­u­lar region and quan­ti­fy­ing how they influ­ence local air quality.

Keywords: air pol­lu­tion, gen­er­al­ized addit­ive mod­els, self-​​organizing maps, and syn­op­tic meteorology.

Work­ing paper