The Use of Control Charts to Interpret Environmental Monitoring Data

1 Corresponding author: [email protected]; 417-836-3119 ABSTRACT: Many different methods of synthesizing and analyzing environmental monitoring data exist. Given the diversity of current environmental monitoring projects, and the large number of scientists and policy-makers involved, there is a critical need for a universal format that both summarizes data sets and indicates any potential need for management action. Control charts, originally developed for industrial applications, represent one way of doing this. Control charts indicate when a system is going ‘out of control’ by plotting through time some measure of a stochastic process with reference to its expected value. Control charts can be constructed for many different types of indicators, whether univariate or multivariate. Control charts are simple to interpret, and can easily be updated whenever additional data become available. The relative risks of Type I (i.e., concluding meaningful change has occurred when actually it has not) and Type II (i.e., concluding meaningful change has not occurred when in fact it has) errors are intuitive and easily adjusted, and one may define a threshold for action at any desired level. Control charts may often be more informative than traditional statistical analyses such as regressions or parameter estimation with confidence intervals. The primary challenge in most situations will be determining a stable or baseline state for the ecological indicator in question.