766 



NA TURE 



[November 24, 1923 



the shore, and conclusions have l)cen formulated that 

 the effect is due to the local wind blowing the water 

 towards the shore. This conclusion is not substantiated 

 by Fig. I, for the winds which raise sea-level on the 

 east coast of Britain are those which blow away from 

 the shore. A westerly wind therefore raises the water 

 of the whole of the North Sea in some degree or other, 

 and this effect must therefore be due to wind blowing 

 over a large area to the north of Scotland. The 

 direction of the most effective wind at Felixstowe has 

 a much larger northerly component than is present 

 at Dunbar. In other words, a northerly wind would 

 have little effect at Dunbar as compared with Felix- 

 stowe, the reason probably being that the sea becomes 



«-v.Cbwl*' 



J 



~x, .♦'^ 



•V^ 



V 



Iff- 



\ Fio. I.— The most effective winds for raising iea-lcvel round tlic Briiisli Islet 



shallower towards the south, agreeing with theoretical 

 conclusions that, apart from the effects of rotation, 

 wind operates more effectively in shallow water than 

 in deep water. 



The Irish Sea gives some interesting results. It 

 would appear that from Newlyn northwards the most 

 effective wind has a large southerly component. Local 

 influences are far more marked at Newlyn and Cork 

 than at Holyhead and Belfast, while the effect of the 

 broadening out of the Irish Sea is shown slightly 

 at Holyhead and still more at Liverpool, where the 

 westerly component of the wind shows its influence, 

 and again the shallower water of the upper part of the 

 Irish Sea helps the effect. 



Some of these conclusions could have been formulated 

 NO. 2821, VOL. 112] 



roughly from qualitative statements in seamen 

 almanacs, but what gives value to the results deal 

 with above is that they are expressed quantitative! 

 Further, qualitative statements are liable to give n< 

 the most effective wind for a given wind -strength, b 

 that wind which has happened to give a storm-effect 

 The predominating factor in the above results is tl 

 southerly wind operating on the Atlantic water souv 

 of Ireland. This conclusion has been verified 1 

 Liverpool by applying an extension of the fom. ' 

 as to include Atlantic winds (south of Ireland) . 

 as local winds. The results show that, for a given wijj. 

 strength operating in the most favourable direction : 

 each case, the Atlantic wind has 50 per cent 

 effect than a local wind, in spite of the «. 

 Atlantic water lx;ing less favourable to wii 

 effects. Further, the most effective Atlanti 

 wind blows from the south and the most effecti\ 

 local wind from almost due west. 



When we correlate the pressure system at 

 a fixed time with the mean sea-level at a varial>l 

 time we find that the correlation between tl 

 sea-level at Liverpool and an easterly graditnt 

 of pressure, corresponding roughly to a soutli 

 wind, is greatest when the mean sea-level : 

 taken about fifteen hours later than the corr' 

 spending pressure gradient. The corresptjndii 

 time for Newhn is nine hours. For a norther! 

 gradient, however, the time difference for max 

 mum correlation with mean level at Liverpool 

 practically zero. These results are in conformit 

 with those just discussed, for we should expet i 

 a large time-interval for setting up the circula- 

 tion of water from the Atlantic and a small 

 time interval for effects generated in the Irish Sea. 

 It can be deduced, therefore, that the mo-* 

 favourable conditions for giving exceptional effect 

 on sea-level are those in which a south wind blow - 

 for some hours, filling the Irish Sea as a whoh 

 and then changes to the west — the. rapidity with 

 which the west wind operates is appanntlv 

 favourable to storm-effects. 



The correlation between mean level at Li w - 

 pool and the fluctuation of the local atmospheri 

 pressure is greatest when the sea-level is taken 

 about three hours earlier than the pressure. For 

 ■{M Newl}Ti the time-advance is five hours. Thc>< 

 results are of very great interest : the anticifwticr; 

 in mean sea-level of changes in barometric pressur- 

 is probably due to the different rates of travti 

 of disturbances through air and through water. Fern 1 

 (U.S. Coast Survey, Report, 1871, p. 93) in 187 1 

 noted that the changes in sea-level in Boston 

 Harbour, U.S.A., appeared to anticipate the 

 barometric pressure. Anticipation of coming storms, 

 according to Dr. Bell Dawson (Trans. Roy. Soc. 

 Canada, 1909, pp. 186-188), is also shown in tb<. 

 currents off Newfoundland ; a change in magnitude 

 and direction is noticeable some twelve hours before thi. 

 onset of a storm, and generally (with some exceptions) 

 the current sets more strongly towards the direction 

 from which the wind is about to blow. This pheno- 

 menon is regarded by the local fishermen as an unfailing 

 indication of bad weather. These anticipator}' effects 

 are worthy of fuller investigations. 



