An assessment of water quality in a watershed of 9600 ha in Guaíra, SP - Brazil, submitted to different occupations (rural and urban), was carried out during the period Jun/95 to Jun/96, using multivariate statistical techniques. Water samples were collected every 21 days at five sampling sites to evaluate ten water quality parameters (total and dissolved P, pH, dissolved O2, NH4, nitrate, electrical conductivity, turbidity, total suspended solids, and chlorophyll). A water quality index (WQI) was determined by the factorial analysis and the Bartlett method. The results point out a significant difference of WQI between the three upstream and the two downstream (in relation to Guaíra) sampling sites (mean values of - 1.76 and 2.35 respectively). The factorial analysis indicated that dissolved O2, total P, NH4, and electric conductivity were the most important parameters in determining the value of the WQI. The first three factors explained 71% of the variance of the data, whereas the first factor alone explained 47%. The influence of climate parameters was of little significance for the WQI, causing a minor water quality deterioration during the rainy season. The paper discusses the applicability of the factorial analysis as a means of assessing water quality alterations, as well as for the selection of appropriate indicators for the assessment of environmental impact in watersheds.
WQI; factorial analyses; monitoring; pollution; water resources; rivers