Two Different Modeling Approaches to Predict the Biological Contaminations in Aliso Creek, California

Five physical and chemical parameters that can be measured instantaneously were tested in a study focused on Aliso Creek, CA to determine if these could be used to predict bacterial indicators and parameters showed potentials as surrogates for indicator bacterial concentrations during the dry season flow throughout incorporating into two models. Error values of estimations from two modeling approaches stayed in 5–10% of observations and both models could capture high peaks of observations which might be more important than lower peaks of observations because high concentrations of pollutants may affect the public health. The study applied a principle component regression (PCR) model and a classification and regression tree (CART) to the data and estimations from those two approaches were very promising for predicting.