8 The Ocean 



lighter of the water masses must be 10 cm higher than the heavier. If the density differ- 

 ence changes with the depth the above equation will include the integral of 



Po — Pi 



dz 



taken from the surface to the depth /;. 



For practical purposes the mean water level is determined at coastal stations by a 

 tide gauge. Calculation of a mean value will eliminate the periodic factors (tides and 

 waves) but other factors will remain ; in the first place the aperiodic changes in mete- 

 orological factors such as the wind, barometric pressure, precipitation and evaporation 

 that can only be eliminated by taking a mean value over a number of years. However, 

 even this mean value cannot be taken as invariable. It will reflect secular (long period) 

 changes in meteorological factors and also slow deformations of the Earth and slow 

 changes in the total water mass of the oceans. For comparison and inter-relation of 

 mean sea-levels fixed at different places along a coast, precision levelling between these 

 points is essential. This must be taken over land and be independent of the conditions 

 in the sea in order to show whether the mean sea-levels are in one and the same or in 

 different niveaus. On the subject of precision levelling along the Baltic coast (1896-8) 

 see Westphal (1900), along the east coast of North America see Anvers (1927) and 

 Bowie (1936), and on the interpretation of these see Dietrich (1937). 



Sea-level at almost all coastal stations shows clearly an annual period which is 

 related principally to wind conditions along the adjacent sea coast; thus the sea-level 

 at Aden is connected with the monsoon in the Arabian Sea (Krummel, 1907), while in 

 Japanese waters the annual changes in barometric pressure and in density of the water 

 are of greater influence (Nomitsu and Okamoto, 1927). On the annual variation in the 

 sea-level along the Baltic coast see Hahn and Rietschel (1938), and Bergsten (1917). 

 Along the coasts of those seas where there are strong tides the determination of mean 

 sea-level is more complicated since the effect of the tides has first to be eliminated. 

 This is best done by subtracting the mean tide level calculated by means of the har- 

 monic tide constant from the actual change in water level as shown by the tide gauge. 

 The remaining part is the aperiodic deviation in water level (in addition possibly free- 

 oscillations of water masses) which must be related to meteorological factors (Marmer, 

 1927). If this ideal method is not possible the mid-point of each tide can be found by 

 taking an average of hourly readings over a full tide period and it can be assumed that 

 this value is reasonably free from any cosmic influence. An investigation of this type 

 has been carried out for the German Bay (North Sea) by Leverkinck (1915). 



The changes in sea-level recorded on a tide gauge can also be simulated by a rise or 

 a fall of the land on which the gauge stands. Movements of the coast line forming the 

 boundary between land and sea may be compounded of two movements, those of the 

 water and those of the land (Penck, 1934). As the ocean may be compared with a 

 large vessel filled with water, changes in the water surface may arise through changes 

 in the volume of water in the ocean or by alteration of either its size or the position 

 of the water surface in the vessel. All changes in sea-level that affect the entire ocean 

 surface in the same direction are termed, following Suess (1888), eustatic. This in- 

 cludes two very slow changes: the nomic and \h& juvenile motion. The first is due to the 

 slow erosion of the land that lifts the sea bottom, the second is due to the continuous 



