308 



NATURE 



[May 22, 1913 



In considering the structure of the atmosphere, as 

 it lias been revealed by the observations I have car- 

 ried out, principally at Ditcham on the South Downs, 

 we may divide the subject into two parts : first, the 

 wind structure in the lowest kilometre, and secondly, 

 the general wind distribution up to the greatest 

 heights reached by the balloons. 



It is a matter of common observation that the wind 

 increases above the surface, and in these days of 

 aerial navigation it is important to know the law of 

 this increase. It seems that at Ditcham, the increase 

 in velocity is at first linear or nearly so, and that the 

 fine representing the linear increase passes through 

 zero velocity at sea-level. That is to say, if we plot 

 the wind velocity at the surface and draw through it 

 a line from zero velocity at sea-level, the wind 

 velocities at other heights, up to half a kilometre to 

 one kilometre, will lie very nearly on this line ; this 

 approximately linear increase has been found to agree 

 with observations at several land stations, but over 

 the sea other conditions probably prevail. 



But there are occasions when this state of things 

 does not apply at all ; this is often the case in light 

 breezes, and at times when the surface wind is very 

 shallow, giving place to an entirely different wind 

 regime in the first kilometre of height. At such 

 times it often happens that the wind velocity is 

 greatest a very little way above the surface. The 

 fact that there are two separate conditions emphasises 

 the danger of taking means. By taking the mean 

 value of a number of separate observations we might 

 get as a result that the wind neither increased nor 

 decreased in the first kilometre of height, which in 

 reality is only true on very rare occasions. As has 

 been truly said, " La methode des moyennes c'est le 

 seul moyen de ne jamais connaitre le vrai ! " 



Another question of great importance to aviators 

 is the effect of hills upon the winds blowing over 

 them. The balloons used in my investigations ascend 

 at the rate of 500 ft. per minute, and in a few 

 minutes are carried beyond the reach of ground 

 eddies ; in some cases, however, I have found that a 

 balloon rose with more than its normal velocity when 

 passing over hills if a strong wind was blowing, and 

 the effect is visible sometimes even when the balloon 

 is more than a kilometre above the surface ; on other 

 occasions very little effect has been observed. More 

 light is being thrown on this question by the observa- 

 tions of Mr. J. S. Dines on slowly ascending balloons. 



The lower layers of the atmosphere up to one or 

 two kilometres are the most important to aviators. 

 To meteorologists the higher layers offer problems of 

 greater interest. In considering the winds in the 

 free air it is convenient to have some datum to which 

 to refer them. The observed surface wind is not 

 convenient for this purpose, being too much affected 

 by local conditions near the ground. A better datum 

 is what is known as the gradient wind. Under the 

 influence of the barometric gradient the air is being 

 pressed towards the areas of low pressure, but the 

 wind is actually blowing more or less along the 

 isobars at right angles to the force. In much the 

 same way, water in a basin, when allowed to escape 

 through a hole in the centre, and when given a slight 

 movement of rotation, moves round the basin at right 

 angles to the forces which are pressing it towards the 

 centre. In the case of the atmosphere the turning 

 movement is p-iven by the rotation of the earth under 

 the moving air. For any pressure condition to be 

 maintained the air must be moving with a certain 

 definite velocity, copending on the shape of the isobars 

 and the steepness of the barometric gradient. This 

 rate can be calcukted for the conditions obtaining at 

 the time, and the wind so calculated is called the 

 gradient wind. It has been found that there is a 

 NO. 2273. VOL. qi] 



fairlv good agreement between the wind so calculated 

 and the observed wind at a height of \ km. or so, 

 but owing to friction the surface wind is usually of 

 a smaller velocity, and directed more towards the low 

 pressure. 



In order to show in a clear manner the changes of 

 wind at different levels, I have prepared some models 

 which give a better mental picture of the conditions 

 than a diagram. The atmosphere is supposed to be 

 divided up into layers each 1 km. thick and the 

 average wind in each layer is represented by a 

 coloured card ; the length of the card represents the 

 velocity of the wind, 1 cm. representing 1 metre 

 per second, 1 metre per second being about i\ miles 

 per hour ; the direction of the card shows the direc- 

 tion of the wind, the arrow flying with the wind. 

 The red cards represent winds that may be supposed 

 to bring air from an equatorial direction, that is winds 

 from east-south-east through south to west-north- 

 west, the blue cards winds that may be supposed to 

 bring air from a polar direction. 



For convenience I have divided the wind structures 

 into five types ; they are perhaps rather artificial, as 



1 shall show later, but it is convenient to make some 

 sort of classification, even when further knowledge 

 must change it. In the first three types of wind 

 structure, the wind increases above the surface and 

 equals the gradient velocity at a height of 5 km. or 

 so ; above this in the first class the wind remains more 

 or less equal to the gradient velocity, up to a height 

 of 7 or 8 km. ; in the second class the wind in the 

 upper air greatly exceeds the gradient wind, and in 

 the third class it falls off again to a lesser value ; but 

 in all three classes the direction remains much the 

 same as that of the gradient wind. 



The first type may be called the solid current ; it 

 does not seem to be associated with any particular 

 type of isobars, but in a preponderance of cases the 

 wind is easterly, and the remaining cases are nearlv 

 all westerly ; it is rare to find the solid current with 

 winds from the north, or from the south. 



In rare cases there is scarcely any wind up to the 

 greatest heights reached, and the little wind there is 

 often blows from varying directions in different 

 layers ; this tvpe, which may be looked on as a sub- 

 class of the first type, sometimes occurs in still anti- 

 cyclonic conditions in summer. 



In the second class the gradient wind, after being 

 reached at a height of about \ km., is greatly ex- 

 ceeded in the upper air ; in some cases the wind at 



2 or 3 km. is double the gradient value, or even more. 

 This tvpe is likely to occur when there is a low 

 pressure to the north of the station and when there 

 is a strong temperature gradient, such that the low 

 temperatures correspond to the low pressures, and 

 vice versd ; such conditions should theoretically cause 

 an increase in wind velocity in the upper air, but it 

 is not possible to calculate what the effect should be 

 without knowing the temperatures, not only on the 

 surface, but in the upper air over the region in ques- 

 tion. One may, however, calculate what effect sur- 

 face temperatures would have on the isobars at, say, 



3 km., assuming that the vertical temperature 

 gradient is the same at every point ; a map con- 

 structed to show the isobars which have been thus 

 calculated must be looked on as a rough approxima- 

 tion only to the real conditions. A map of the isobars 

 at 3 km. for May 11, 1907, shows how much steeper 

 was the gradient on this day in the upper air than it 

 was on the surface, a fact which quite accounts for 

 the rapid increase in wind velocitv from 2 metres per 

 second at the surface to 19 metres per second at 

 3 km. 



Winds belonging to this class may come from any 

 point of the compass. 



