Study Question 9. Is the sampling frequency appropriate (temporal optimization )?

C.9 Study Question 9: Is the sampling frequency appropriate (temporal optimization)?

Optimization and design of the monitoring program must assure sample independence while covering the site sufficiently and collecting adequate data over an appropriate time period for proposed statistical evaluations. If the monitoring program is in the early stages, statistical design options should be considered such that an adequate number of samples are collected (Section 3.6). For sites with existing long term monitoring data sets, sampling frequency can often be reduced while still providing adequate data for evaluation. The required frequency of sampling can be evaluated with statistical methods that assess whether there is redundancy of sample results for a particular well. You can also apply spatial statistics to evaluate sampling frequency among a set of wells. For an overview of spatial optimization methods, see Study Question 10. For effective optimization, you must establish the goal of the long-term monitoring program and identify an acceptable length of time to determine a change.

This question can be relevant in all stages of the project life cycle: release detection, site characterization, monitoring, remediation, and closure. Although it is more likely that there is enough information to conduct optimization at later stages in the project life cycle.

Selecting and Characterizing the Data Set

Verify that the data set can support optimization techniques. Refer to Section 3.4: Common Statistical Assumptions for further discussion of how the following requirements may impact statistical analysis results.

Check for outliers using box plots, probability plots, Dixon’s Test, and Rosner’s Test.
Check for autocorrelation between successive sampling events.
Check significant temporal trends using time series plots.
See also Section 4.1: Considerations for Statistical Analysis.

Statistical Methods and Tools

Using the results of the above plots and tests as a guide, you can use more sophisticated statistical methods to evaluate the redundancy of sample results for a particular well. These methods can also be applied to a network of wells. The two approaches highlighted for this question are an iterative thinning analysis or cost effective sampling (CES). There are also some other optimization methods including the modified CES method and genetic algorithms. Be aware that in some cases where the uncertainty is determined to be high, additional sampling may be recommended. See Appendix D for software packages.

Iterative Thinning

This test analyzes a large data set and trims results.
If trends or seasonality exist, then the performance metric is based on replicating these temporal trends with the subset of samples.
If there are no trends, then trimmed data are compared to the stationaryA distribution whose population characteristics do not change over time or space (Unified Guidance). summary statistics.

Cost Effective Sampling (CES)

Base "the sampling frequency on the changes in concentration at a given well, rather than the well's location with respect to the plume" (Ridley and McQueen 2005; Ridley et al. 1995).
"CES calculates quantitative measures of the trend and variability of important COCs [chemicals] at each monitoring location and interprets this information by means of decision trees to arrive at a recommended sampling frequency" (Ridley and McQueen 2005).
"An essential aspect of the CES program has been to use simple statistics within a decision-logic framework to provide information that can be easily understood" (Ridley and McQueen 2005).

Interpretation of Results and Associated Uncertainty

Groundwater monitoring well network optimization often works best when the network is evaluated as a unit. Therefore, there is greater potential for project benefits when both spatial and temporal information is considered. However, there are cases where a project could benefit by eliminating redundant sampling events or by adding sampling events to reduce uncertainties.