Limited Power Analysis and Monitoring Design Optimization for Sampling in the Wood River Basin
- Brett Marshall
- May 5
- 1 min read

I was recently asked to conduct a statistical power analysis to optimize the assignment of statistical replicate samples throughout the Wood River Basin. The current sampling plan uses five statistical replicates at six sites throughout the Wood River Basin. The fixed count subsampling protocol requires >200 animals to be identified from benthic samples. Thus, a total of > 1000 invertebrates are identified from each site. Due to the cost of processing samples in the lab, an increase in the subsampling target to > 500 invertebrates identified per sample will significantly reduce the number of statistical replicates that may be collected from each site. These results were intended to address the effects of reducing field replication on the ability of the monitoring program to detect changes in the taxa richness (an easily digestible expression of community diversity) at each of the sites from year to year--given the assumptions of alpha, beta, sigma, n, and delta. We found that decreasing the replication level sufficiently to offset the cost of increasing the fixed-count subsample-target detrimentally reduced the ability of the monitoring program to detect change.
The attached PDF includes the rationale, methods, and results (graphs & summary). Technical Memo RCC-2025-02.
Comments