Reducing Meteorological Effects on Monthly Mean Sea Levels

Rossiter (1958, 1960) has considered the errors in daily, monthly and annual mean sea levels, due to incomplete elimination of the tide. The purpose of this note is to point out that meteorological effects with periods less than a month are not adequately reduced by the direct averaging presently used to compute monthly mean levels, and to suggest the adoption of an improved averaging procedure. Monthly mean sea levels are being computed and published for several hundred ports, and the number is increasing each year. The main purpose in choosing an averaging period of one month is to reduce the effects of disturbances with periods less than one or two months, so that seasonal and secular changes can be studied. The shorter-period disturbances in sea level are usually of meteorological origin, and may be of greater amplitude than the seasonal oscillation. It has become traditional to think of a “ monthly mean ” as the simple arithmetic mean computed from the beginning to the end of a month. Such simple averaging however is not always the best — for example, there are serious difficulties in the use of a simple average of 24 hourly readings as a “ daily mean ” sea level, since tidal effects are not adequately reduced. In computing daily mean sea level, it is usual to use a numerical filter (Groves, 1955) specially designed to reduce the effects of tides as much as possible. These filters extend over more than 24 hours