摘要

This paper describes a simulation tool to aid the design of nutrient monitoring programmes in coastal waters. The tool is developed by using time series of water quality data from a Smart Buoy, an in situ monitoring device. The tool models the seasonality and temporal dependence in the data and then filters out these features to leave a white noise series. New data sets are then simulated by sampling from the white noise series and re-introducing the modelled seasonality and temporal dependence. Simulating, many independent realisations allows us to study the performance of different monitoring designs and assessment methods. We illustrate the approach using total oxidised nitrogen (TOxN) and chlorophyll data from Liverpool Bay, U.K. We consider assessments of whether the underlying mean concentrations of these water quality variables are sufficiently low; i.e. below specified assessment concentrations. We show that for TOxN, even when mean concentrations are at background, daily data from a Smart Buoy or multi-annual sampling from a research vessel would be needed to obtain adequate power.

  • 出版日期2010-2