SECT. 5] SURGES 639 



of Great Britain, of Tomczak (1952, 1953) for the German Bight, of Verploegh 

 and Groen (1955) for the Dutch Wadden Sea, of Saville (1952) for Lake Okee- 

 chobee, of Miller (1958) for the coastal waters of New England, and of Donn 

 (1958) for some places on the east coast of the U.S.A. 



Methods of the second class, based on a semi-empirical-semi-theoretical 

 approach, are those of Schalkwijk (1947) and Weenink (1958) for the southeast 

 coast of the North Sea and of Hunt (1956) for shallow lakes, like Lake 

 Okeechobee. 



The Schalkwijk- Weenink method makes explicit use of the idea of the 

 equilibrium effect as distinct from the actual wind effect. From the basic 

 observational material the effects due to the non-stationarity of the wind fields 

 involved were eliminated in a more or less heuristic way so as to give the true 

 equilibrium effects. From these data, partly by direct correlation and partly by 

 theoretical calculation, graphs and formulas have been derived which give 

 the equilibrium effects at various places as functions of the gradient-wind 

 vectors over a number of sections of the North Sea (thus taking inhomogeneities 

 of the wind field into account) and over the Channel. The influence of air-mass 

 stability on the relation between the gradient-wind velocity and the wind stress 

 on the sea surface is also taken into account in this method of computation. 

 From the equilibrium effect the actual effect to be expected is then found by 

 applying a time lag and secondary corrections, which account for such effects 

 as "overshooting" or "undershooting" (resurging) and after-oscillation (see 

 section 3-B-c). For a concise description of this method see Groen (1961). 



Hunt's method for shallow lakes divides the lake into a number of "cells" 

 formed by a set of wind fetches and a number of their orthogonal trajectories. 

 The number of cells to be chosen is such that in each cell fairly homogeneous 

 conditions prevail as to depth and wind. In each cell the slope of the lake sur- 

 face is computed by assuming equilibrium within the cell. If now a provisional 

 line of zero disturbance is assumed, the heights along each fetch strip can be 

 determined. Any lateral height differences thus found are smoothed out by some 

 suitable procedure, which accounts for the current effect (see section 3-B-b), 

 the Coriolis force being left out of consideration. The provisional line of zero 

 disturbance and the elevations found are corrected until the condition of zero 

 mean elevation for the whole lake is satisfied. 



Wholly theoretical approaches, finally, have been proposed by Kivisild 

 (1954), Hansen (1956), Welander (1957), Fischer (1959) and Svansson (1959). 

 These methods come down to numerical integrations of the basic equations of 

 section 3- A, viz. the equations of motion and the continuity equation, either in 

 their vertically integrated form (Kivisild, Hansen, Fischer, Svansson) or in 

 the form of an integro-differential equation based on the well-known fact that, 

 in the absence of lateral stresses, the vertical velocity profile is uniquely 

 determined by the local time-histories of the wind stress and of the surface 

 slope (Welander), See also Lauwerier (1960) and Welander (1961). 



For the bottom stress the quadratic form may as well be used here as the 

 hnearized form. 



