The review panel was tasked with assessing several methods tested and assumptions made in the GAM approach as implemented in an R package, with the goal of providing a temporally and spatially consistent standardized method for annual reporting of trends in tidal water quality.
During this review, the charge was to 1) address ten main questions on the scientific and management issues raised in the documentation, and 2) make recommendations for future work by the Chesapeake Bay Program (CBP) that would further strengthen this approach, or that are related to the scientific or management issues raised in the review.
A panel of four individuals with appropriate expertise in applied environmental statistics, long term water quality monitoring, and generalized additive models, was formed in September 2016, and the team conducted their review of the provided documentation and R package over the past few months.
The panel’s overall assessment finds that the addition of the Mixed GAM Computation Vehicle (mgcv) implementation of GAMs to the toolbox for tidal water quality trends analysis in the Chesapeake Bay is an excellent decision – making particular note that mgcv is a transparent, flexible, robust, and well supported approach. The panel identified key areas for exploration moving forward, as well as alternatives for consideration. Several suggestions also emerged for future analyses, including: