Evaluation of PM_(10), NO_X, SO_2, CO and O_3 Ambient Monitoring Networks of Vitoria, Brazil

摘要

Statistical analyses, the entropy concept and the Shannon information index were used astools to evaluate the performance of ambient automatic monitoring networks in Vitoria, ametropolitan region with 2,000,000 inhabitants, located in Southeast Brazil. The region isimpacted by emissions from an industrial district located up wind, when considered thepredominant wind directions, and by vehicle emissions including heavy traffic roads. Theautomatic networks comprise of seven PM_(10), five Nitrogen Oxides (NOx), six SulfurDioxide (SO_2), four Carbon Monoxide (CO) and four Ozone (O_3) analyzers collecting datasince 2001. The evaluation was done using hourly values and covered the aspects related tothe merit of monitoring for epidemiological studies purposes, the information exchangeamong the stations and the importance of generated information. The results express thatPM_(10), NOx and CO networks are yielding very useful data, pointing to well definedprobability density functions with a minimal loss of information when one station ishypothetically turned-off at a time. The O_3 network is over dimensioned for the features ofthe region as a whole and very influenced by local chemical mechanisms and certainstations must be spatially rearranged. The Sulfur Dioxide network was found to berepresentative of concentrations ranging over two magnitudes, due to different patterns ofindustrial and heavy vehicle emissions, but a great loss of data was verified by seriesanalysis due to changes in optical systems of equipment.