This paper summarizes the results of four statistical approaches for the estimation of long-term trends (1983-92) in sulfate concentration data from 90 monitoring sites across the United States. Least squares regression models and nonparametric techniques were applied to these data. Sulfate concentrations were found to be generally decreasing on the order of 0-4% at most sites. There was general agreement that trends were significant in the Great Lakes, Pacific northwest, and southwest regions. Although strengths and weaknesses are described for each approach, all of these approaches are useful for long-term trend estimation. Visualization techniques are recommended for displaying trend patterns and associated levels of statistical significance.
- Long-term trends
- statistical models