Calm analysis using a robust method

Abstract

La Plata city and surroundings (Argentina) -around 800 000 inhabitants- possesses high industrial activity and intense vehicular traffic but there is no governmental air pollution network and basic meteorological parameters have been scarcely studied. This paper focuses on wind calm structure and wind patterns after the end of calms in order to assess the potential accumulation and fate of released airborne pollutants. A robust correlation method using an M-estimator is employed to compare seasonal observations at three monitoring sites located in key parts of the city covering 1995- 2006. Results show that calms are on average 14.7% in summer, 19.1% in autumn, 12.8% in winter and 11.6% in spring. The three sites under study were highly correlated indicating a generalized calm pattern for the city and surroundings. The robust coefficient allows inferring the lineal character of dependence among observations at the three sites. Calm structure revealed that approximately 50.6 % of the calms lasted 1 hour while calms lasting 5 hours or less constitute around 90% of calm occurrences. Calms lasting two hours or more are more probably to occur during wee and evening hours indicating when the accumulation of released air pollutants is likely to happen. Wind roses named "outgoing of calm wind roses" (time consuming to estimate) showed the wind direction patterns more probable to occur after a calm. Overall correlations between seasonal averaged wind roses and those corresponding to the outgoing of calms showed strong linear correlation for all the seasons. This allows concluding that winds after calms may be well represented by their corresponding seasonal averaged patterns (easy to compute).