Air pollution prediction models of particles, As, Cd, Ni and Pb in a highly industrialized area in Castellón (NE, Spain)

Abstract

ABSTRACT:The objective of this study was to elaborate a series of mathematical models with the aim of short-term prediction of TSP, PM10, As, Cd, Ni and Pb in air ambient. These pollutants depend on some known variables (meteorological variables). The goal is to provide a useful instrument to alert the population facing possible episodes of high concentrations of atmospheric pollutants. The study was carried out in a highly industrialized area in the ceramic cluster of Castellón during five years (2001)(2002)(2003)(2004)(2005). The origin of the contamination in this area is both natural and anthropogenic. The natural origin is due to the resuspension of mineral materials from the surrounding mountains and from the long-range transport of materials from North Africa. The anthropogenic contamination sources that stand out include the non-metallic mineral material industries (ceramic production), chemical industries (color, frit and enamel manufacturing) as well as vehicular traffic. Once the particle samples were 2 collected in quartz fiber filters, the concentration levels of TSP and PM10 were determined gravimetrically. The chemical analysis of the filters was carried out by ICP-MS. Predictive models have been constructed by using Multiple Regression Analysis together with Time Series Models (ARIMA). The SPSS 14.0 statistical software has been employed to analyse the obtained experimental data.

Environmental surveillance modeling: A predictive respiratory alert model for the Shenandoah Valley, Virginia

“…The team utilized multiple linear regression analysis and autoregressive integrated moving average (ARIMA) time series models to generate the alerting system [10]. Similarly, the University of Virginia developed a criminal analysis program to monitor anomalies in criminal data built around a zero modified Poisson linear model [11].…”

Time series analysis of wintertime O3 and NOx formation using vector autoregressions

“…Existing studies involving time series analysis of O 3 and NO x have generally used univariate approaches (e.g., Kumar and Jain, 2010;Kumar et al, 2004;Slini et al, 2002;Robeson and Steyn, 1990), meaning that precursors relevant to formation chemistry are excluded from the analysis. Likewise, studies on time series analysis from particulate matter (PM) mass and/or it components (e.g., Jian et al, 2012;Vicente et al, 2012;Diaz-Robles, 69 2008;Vana et al, 1999) have focused on a single variable approach. For this study, we completed real-time measurements of gas-phase organic and inorganic species during a winter field campaign in Utah, then applied a vector autoregression to examine precursor species that are important to the formation of O 3 and NO x .…”

Estimation of PM10-Bound As, Cd, Ni and Pb Levels by Means of Statistical Modelling: PLSR and ANN Approaches

“…The state-space model performed better than the ARX model. On the other hand, Vicente et al (2012) developed predictive models based on multiple regression analysis together with time series (ARIMA) models to predict the concentration of total suspended particles (TSP), PM 10 , As, Cd, Ni and Pb in the ambient air of Castellón (Spain). Furthermore, in a previous study conducted by Arruti et al (2011), estimations of As, Cd, Ni and Pb levels in Cantabria (Spain) by means of statistical MLR and PCR models have been conducted.…”