|
Abstract:
|
Understanding long-term patterns in water quality data is challenging due to both variability and trends being influenced by climatic, biological, and anthropogenic factors, as well as changing technology. In Chesapeake Bay, the largest estuary in the United States, a collaborative partnership of stakeholders want to know if water quality is responding to a multi-state effort to reduce nutrient pollution. For this purpose, our team of estuary scientists and statisticians have been using Generalized Additive Models (GAMs) to analyze changes over a 30 year period for 10 different water quality parameters at 140 locations. We will describe some of the details of our implementation, including flexibility to model changing seasonal cycles, using a maximum likelihood approach for censored data, an intervention approach for major method changes, and challenges with concurvity between explanatory variables. This year we are at a key moment to be able to influence revisions of pollution management plans, so in this talk we will touch on our current efforts to generate actionable insights from our statistical analyses that are helpful to water quality managers and decision makers.
|
