Sinusoidal Modeling Applied to Spatially Variant Tropospheric Ozone Air Pollution

This paper demonstrates how parsimonious models of sinusoidal functions can be used to fit spatially variant time series in which there is considerable variation of a periodic type. A typical shortcoming of such tools relates to the difficulty in capturing idiosyncratic variation in periodic models. The strategy developed here addresses this deficiency. While previous work has sought to overcome the shortcoming by augmenting sinusoids with other techniques, the present approach employs station-specific sinusoids to supplement a common regional component, which succeeds in capturing local idiosyncratic behavior in a parsimonious manner. The experiments conducted herein reveal that a semi-parametric approach enables such models to fit spatially varying time series with periodic behavior in a remarkably tight fashion. The methods are applied to a panel data set consisting of hourly air pollution measurements. The augmented sinusoidal models produce an excellent fit to these data at three different levels of spatial detail.