638 GROEN AND GROVES [CHAP. 17 



(iii ) External surges 



An external surge is a free wave penetrating into the sea region under 

 consideration from an open boundary, where it is induced by a changing sea- 

 level disturbance of the adjacent open sea. Typical examples have been des- 

 cribed by Corkan (1948) for the North Sea and by Rossiter (1959a) for the 

 English Channel. 



The theory of external surges is very similar to the theory of co -oscillating 

 tides. Mathematical treatments have been given by Goldsbrough (1952), who 

 neglected the Coriolis force, Proudman (1954), who obtained solutions repre- 

 senting damped Kelvin- and Poincare-waves, and Crease (1956). The last 

 author considers a system of waves approaching a semi-infinite barrier. The 

 waves have transverse accelerations balancing the Coriolis force. The effect of 

 the barrier is to form a shadow zone on its lee side, but the transverse accelera- 

 tions on the edge of this zone cause Kelvin-waves to propagate at right angles 

 with the direction of the original waves, into the region behind the barrier. 

 The amplitude of these Kelvin-waves depends on the ratio of the given wave 

 period to the length of the pendulum day. Thig mechanism may account for the 

 properties of certain external surges running into a partly enclosed sea. 



d. Forecasting 



The main problem in forecasting sea-surface surges is to find the sea-level 

 disturbance heights from the wind field. In most cases the effect of atmospheric 

 pressure is of much less importance, except in those cases which are near 

 resonance ; most techniques of practically forecasting sea-level heights account 

 for it by taking the static pressure effect ( — 1 cm per millibar of atmospheric 

 pressure difference) multiplied by a statistical reduction factor of the order of 

 0.5 (e.g. Schalkwijk, 1947, p. 54). 



Some methods (Corkan, 1950; Leppik, 1952) use the idea of a travelling 

 sea-level disturbance by incorporating in the forecast for a certain place the 

 disturbance observed at some other place a certain time interval before. 



The various methods used for computing wind effects in partly or wholly 

 enclosed bodies of water from the wind field may be classified into three groups, 

 according to the main line of approach underlying a method, which may be 



(1) an empirical approach, 



(2) a semi-empirical-semi-theoretical approach, 



(3) a theoretical approach. 



Methods of the first class use direct empirical relations between the wind 

 effect at a certain place and the mean wind or the mean pressure gradient over 

 the sea region considered or over a few different parts of the region, either at 

 about the same time (for effects of nearby areas) or some time before (e.g. 3-9 h, 

 in the North Sea). These methods mainly suppose a quasi-stationary develop- 

 ment, non-equilibrium effects being only statistically (climatologically) in- 

 cluded in the formulas used. To this group belong the methods of (among 

 others) Corkan (1948, 1950) and Rossiter (1959) for the east and south coasts 



