Lunar Periods as Grouping Variables for Temporally Fixed Sampling Regimes in a Tidally Dominated Estuary

Abstract

The complex interaction of a daily temporally fixed sampling design with seasonal, tidal, and diurnal cycles can result in serious aliasing and reduced usefulness of estuarine data. The choice of appropriate sampling windows for data grouping and analysis of tidally influenced data is complex and must take into account both solar and lunar influences and their interactions in order to represent actual estuarine conditions. A fixed modulus for grouping or time-series analysis introduces errors. Grouping data by calendar month can be misleading since calendar months may have as many as three or as few as one spring tide. Daily measurements of salinity and water clarity at a fixed time of day were analyzed to illustrate the effects of aliasing. Although the length of synodical (lunar) months is also variable, a seasonally adjusted grouping variable based on solar and synodical month (MO-DAY), created to partially correct for aliasing, was found to be an efficient way to remove short-term cycles for the identification of long-term trends, episodic events, and spatial patterns.