16. LONG-TERM VARIATIONS IN SEA-LEVEL 



J, R. ROSSITEE, 



1. Introduction 



The variations of sea-level dealt with in this Chapter can be said to fall into 

 two categories, periodic and secular. In the first category the lower limit of 

 period may be arbitrarily defined as the day, thereby eliminating the diurnal 

 and shorter-period oscillations of classical tidal theory but retaining the long- 

 period tides ; the upper limit is restricted only by the span of data available for 

 study. Secular variations in sea-level will be found no matter how short a span 

 of time is taken, and in view of the dominating tidal effects found in all oceans 

 and the majority of seas the minimum span of time adopted is the day. 



In this Chapter, therefore, it is sought to discuss variations in level represent- 

 ing time units of a day upwards, the most commonly used units being the day, 

 the month and the year. Despite the increasing attention being given to the 

 subject, it cannot be said that any revolutionary discoveries have been made 

 during the last few decades by research workers, though certain trends in the 

 types of research undertaken may be discerned. Fewer attempts are being made 

 to examine and explain variations at individual stations in favour of regional 

 and global investigations involving data from many stations. The greatest 

 difficulty encountered by those engaged in this type of work lies in the markedly 

 uneven distribution of tide gauges throughout the oceans and seas ; as might 

 be expected most gauges are concentrated in areas such as European, North 

 American and Japanese waters. Under the stimulus of the International 

 Geophysical Year (1957-58) a great effort has been made to improve the 

 situation, with considerable success, but much still remains to be done. 



2. The Determination of Mean Sea-Level 



The raw material for mean sea-level research is a continuous record of 

 heights of tide, accurately observed at fixed intervals of time, and referred to 

 a stable bench mark. No apology is made for stating these rather obvious facts. 

 Experience has repeatedly shown that much labour and money has been, and 

 continues to be, wasted in taking observations without scrupulous regard to 

 the above requirements. Modern demands of different scientific disciplines and 

 of industry in this field appear to be for longer series of observations, at more 

 stations and to a higher degree of accuracy, and it is a regrettable fact that 

 only by great efforts made now will data conforming to these standards be 

 available for research workers a generation hence. 



As remarked in the introduction, the astronomical tidal constituents with 

 periods up to and including the diurnal species must first be eliminated from 

 the observations. This may be achieved by various methods, the simplest being 

 a direct numerical averaging process, usually applied to hourly heights of tide. 

 More sophisticated numerical processes are frequently used, employing filters 

 which discriminate as required against any given species of tide. The best 



[MS received May, 1960] 690 



