Spatial and temporal trends of physicochemical parameters in the water of the Reconquista river (Buenos Aires, Argentina).

The Reconquista river is one of the most polluted watercourses in Argentina. More than 3 million people and over 10,000 industries are settled on its basin. The available data show that pollution is mainly related to the discharge of domestic and industrial liquid wastes that are poured into the river almost untreated. At present no site of the river can be characterized as free of pollutants. Samples were taken monthly from 5 sites of the river; some 18-20 physicochemical parameters were determined in each sample. Analyses revealed significant differences in the degree of deterioration between sites. Two of them, close to the source of the river (Cascallares-S(1) and Paso del Rey-S(2)) resulted less polluted than the two points located close to the mouth of the river (San Martin-S(4) and Bancalari-S(5)). The worsening of the water quality in S(4)-S(5) was attributed to the discharges of the Moron stream, a tributary that flows into the main course of the river a complex mixture of non treated waste waters. PCA was used in the ordination of samples (sites, season and physicochemical parameters). In the PCA performed using all variables, the first principal component showed positive correlation with N-NH(4) (+), conductivity, orthophosphate, BOD(5), COD and alkalinity, and negative correlation with DO. The second principal component was positively correlated with pH, temperature and chlorophyll a and negatively with phenols and hardness. In respect to the spatial distribution, the plot of the scores for the first two components of samples taken in each sampling station showed S(1) and S(2) values displayed farthest at the left side of the X axis with high DO. In contrast, S(4) and S(5) values stayed at the right side of this axis with high N-NH(4) (+), conductivity, orthophosphate, BOD(5), COD and alkalinity; data of S(3) were "intermediate". In order to identify seasonal trends in the concentration of contaminants scores of cases labeled by season were plotted. The line drawn on the ordination plane showed that summer samples tended to converge to the upper right portion of the graph where pollution variables had more importance on the first axis, and along the second axis with high correlation with pH, temperature and chlorophyll a. Winter scores were settled in the left lower part of the plot with minor contribution of pollution parameters and more importance of DO on the first axis. Samples of autumn and spring did not fit a clear cut pattern.