We consider data of dissolved inorganic nutrient concentrations (ammonium (NH+4), ortho-Phosphorous (ortho-P)and nitrate (NO-3) concentrations from 24 locations for over sixteen years in the Elkhorn Slough National Estuary, located in the Monterey Bay area in Central California, USA. We develop a statistical model for the quantification of trends and seasonality. Initial exploratory data analysis revealed high variability among locations andthe presence of large outliers for some locations which are possibly due to observational errors. We propose a model consisting of a trend and a seasonal component where the errors follow an autoregressive process. An additive observational independent error is also included. We use a Bayesian approach to explore the posterior distribution of the parameters using WinBUGS. We found that some stations have a marked seasonal pattern, that neighboring stations can have substantially different behaviors and that some stations show significant increasing trends.
Keywords: Bayesian Modeling; Trend Analysis; Water quality data; Observational errors
Biography: Lelys Guenni has a Bachelor in Mathematics (1980) and a Master Degree in Water Resources Planning and Engineering (1982) from Universidad Simόn Bolívar, Caracas, Venezuela. She got a PhD (1992) taken at the Division of Environmental Sciences from Griffith University, Brisbane, Australia. Since 1994 she has been a full time professor at the Department of Scientific Computing and Statistics, Universidad Simόn Bolívar, Caracas, Venezuela.