J. Zhang, L. Zhang (PRC), and P.F. Ricci (USA)
Water quality; Watershed assessment; Statistic analysis.
This paper develops a statistical approach to retrospective assessment of the impacts of economic developments affecting a watershed in Fujian Province, People’s Republic of China. Using publicly available data, we study times series of several water quality parameters, link them via data analysis to several aspects of economic development, also based on time series, and assess the impact of time-varying economic changes on the watershed’s water quality. The methods used include time-series analysis (e.g., ARIMA), Monte Carlo simulations (e.g., bootstrap), Bayesian regression, trend analysis, and non-parametric correlations. We find that, although trend analysis and correlations can be appropriate, we prefer to develop causal associations that set the stage for more complete analyses done at a future stage of the research. Our results use alternative statistical methods, some of which are numerically intensive, to strengthen the study of causal associations. The portfolio of methods we develop yields results that are useful as a first step – hypothesis generation -- in the overall assessment of the sustainability of economic development of the Jiulong River Watershed characterized by similar relationships.
Important Links:
Go Back
