68o 



NA TURE 



[November i 8, 1922 



The Origin of Atmospherics. 1 



By R. A. Watson Watt. 



'"THAT abnormal outbursts of atmospherics — the 

 -*■ " x storms ' ' of the radio-telegraphist — were asso • 

 ciated with convective weather was indicated by the 

 work of the British Association Committee on Radio- 

 telegraphic Investigations in K114-15. That actual 

 thunderstorms could be located by direction-finding 

 on atmospherics was established early in a Meteoro- 

 logical Office investigation begun in 1915. But it 

 has not yet been shown whether the fully developed 

 thunderstorm is the only, a main, or merely a sub- 

 sidiary source of atmospherics. 



A critical examination of the .data obtained in 

 the Meteorological Office investigation referred to 

 promises to throw some light on the question, and 

 it has been thought desirable to give a preliminary 

 indication of the evidence which is emerging. 



The coastal direction-finding stations of the 

 Admiralty co-operated in the investigation by 

 reporting the apparent direction from which atmo- 

 spherics were arriving whenever such a direction 

 was observed, and when pressure of traffic permitted 

 an observation. 



Some twelve stations took part in this scheme, 

 which began in March 191 6. The examination of 

 the results has been made for two years, April 1916 

 to May 1918. During this period there were ap- 

 proximately 1000 occasions on which three or more 

 stations observed a direction of arrival of atmospherics 

 within the same hour. On plotting these approxi- 

 mately simultaneous observations on a gnomonic chart, 

 it is found that on almost exactly half of the occasions 

 the three or more bearings gave an intersection in a 

 point of a limited area, indicating a source of 

 atmospherics at a point so determined. Actually 

 the distribution of these intersections, according to 

 the number of participating stations, is as follows : 

 Six stations giving bearings meeting in a 

 point (within the limits of accuracy 



of observation) 2 



Five stations giving bearings meeting in a 



point (within the limits of accuracy 



of observation) . . . . . 15 



Four stations giving bearings meeting in a 



point (within tin- limits of accuracy 



of observation) 68 



Three stations giving bearings meeting 



exactly in a point 231 



Three stations giving bearings meeting in 

 a point after adjustment within the 

 limits of accuracy, assumed ± 5 



degrees no 



Three stations giving bearings failing to 

 meet in a point but delimiting a small 

 area as source 62 



Total number of locations . . 488 



The geographical distribution of these apparent 

 sources of atmospherics is : 



The comparison of these locations with meteoro- 

 logical data is a somewhat extensive investigation 

 rid is still in progress. In 288 cases, however, the 



• 1 276S, VOL. I 10] 



immediately available data seemed to provide an 

 adequate basis for discussion. In a relatively small 

 number of cases only were thunderstorms found to 

 have occurred in the region located as a source of 

 atmospherics, and near the time of observation. 

 Since the thunderstorm is a known source of atmo- 

 spherics, it is not proposed to detail here these cases, 

 particularly as it is necessary to search more closely 

 for records of thunderstorms, which are notoriously 

 sporadic phenomena, liable to slip unobserved 

 through the open mesh of the network of observing 

 stations. 



Comparison was also made with the map in the 

 British Daily Weather Report, which shows the area 

 over which rain has fallen during the 24 hours, 

 7 a.m. to 7 a.m. In 239 out of the 288 cases, the 

 apparent source of atmospherics was definitely 

 associated with the rainfall area for the corre- 

 sponding 24 hours. In 105 cases the source was on 

 the advancing edge of the rain area, in 75 within 

 that area, and in 59 cases it was on the rear edge. 

 Of the remaining 49 locations, 30 were found to 

 fall in places where thunderstorm or squall pheno- 

 mena had been reported about the same time. 



Thus in only 19 of the 288 cases, i.e. in 6i per cent., 

 has no meteorological relation with the source of 

 atmospherics yet been traced, while it is also note- 

 worthy that 10 of these 19 fell in the least trust- 

 worthy class of location, the three bearings delineat- 

 ing an area not negligibly, small. 



One is therefore faced with the alternative 

 conclusions that — 



(1) Rainfall, without the occurrence of a fully 



developed thunderstorm, is an important 

 source of atmospherics. 



(2) The climate of south-west Europe is so wet, 



that there is an extremely high probability 



of rain in a random 24 hours at a random 



point. 



To test the validity of conclusion (2) the maps 



for the same two-year periods were used. Four 



individuals (two without knowledge of the nature 



of the test) were asked to carry out a blind spotting 



game by placing a random dot on each chart, without 



seeing its detail. Again, many of these dots were 



beyond the range of adequate data, but 335 out of 



732 could be compared with data. The results show 



that the chances are nearly even for or against rain, 



the distribution of the random points being : 



In rain area 73 



On advancing edge 46 



On rear edge 62 



Total associated with rainfall . . . 181 



Total definitely not so associated . . 154 



It appears, therefore, to be established conclusively 

 that a very high proportion of sharply defined 

 sources of atmospherics are to be found in areas in 

 which rain is falling, and particularly on the advanc- 

 ing edge of such areas, more than 90 per cent, in 

 the present series being in rain areas, and 36 per 

 cent, on the forward edge of the 24 hours' rain area. 

 It is perhaps a permissible inference that, were one 

 able to deal with the instantaneous advancing edge 

 instead of the edge of the area for the day, the latter 

 figure would be increased. 



The results of this investigation fall into line with 

 modern views of the mechanism of rainfall and 

 thunderstorm phenomena. The separation and ac- 

 cumulation of charges caused by ascending currents 

 may be readily believed to be more pronounced on 



