Análise exploratória de dados de monitoramento de efluentes urbanos

Abstract

Aiming the development of knowledge on the temporal variation of urban wastewater quality, this work was based on the analysis of raw sewage quality data from samples collected from the East Interceptor of the sewerage system of Campina Grande City, Paraíba state, northern Brazil. Data on the indicators biochemical oxygen demand (BOD5), ammonia nitrogen, thermotolerant coliforms, total alkalinity and suspended solids, obtained in the period 1977-1989, at the Experimental Station of Biological Treatment of Sewage (EXTRABES), then a research station at the Federal University of Paraíba, were analysed through Cluster Analysis and Mann-Kendall and Box-Jenkins methods. Exploratory analysis sought to establish seasonal patterns of relationships between sewage quality and environmental factors, to model temporal variation trends of urban sewage quality indicators and to evaluate the performance of models interpreting this variation. Time series analysis for data on wastewater quality was able to demonstrate significant trends over the studied period, but did not show the same performance for data on both precipitation and temperature. The application of the Box-Jenkins method allowed modelling the behaviour of both the sewage quality variables and the hydrological variables, with the models being considered satisfactory based on the evaluation of errors and residues.