VII. THE CYCLE PROBLEM AND LONG-RANGE 

 FORECASTING 



THE PROBLEM 



In view of the recent great droughts and the need to control erosion and 

 generally to conserve natural resources, scientific thought should be directed 

 to the very important factor of climatic change and any possibility of fore- 

 knowledge or prediction of changes that might affect the value of lands and 

 alter economic conditions. 



The problem is this : Can we tell, years in advance, the coming of climatic 

 variations that will have an important economic effect? Can we even im- 

 prove our present fore-knowledge by ten per cent? The accepted procedure 

 today in attempting to arrive at such long-range prediction is merely to 

 determine an average climatic line (mean rainfall, for example) and the 

 average departure from that mean, and then expect in the future a similar 

 mean and similar departures from it, without knowing when the departures 

 are likely to occur. The question now is: Can we produce a climatic curve 

 of rainfall or temperature or other growing conditions that will reduce the 

 errors of the aforesaid expectation by a sensible amount? 



Many attempts to solve this problem have been made without success. 

 It is believed that such failures lie in part in three counts. First: Climatic 

 cycles have been assumed to be permanent, if existing at all, like the annual 

 or monthly cycles that are due to planetary revolutions, whereas, as every 

 investigator has found, climatic cycles do not act as if they were permanent 

 and should be investigated under this corrected view. 



Second: The usual methods of search for cycles after failing to recognize 

 this impermanence except in a very crude way, have without justification 

 assumed that the cycles are sine curves, as usually produced by planetary 

 movements, whereas we have no evidence that such is the case. 



Third: The mixture of cycles has been too great to be handled by ordinary 

 methods of analysis, which do not recognize the discontinuous period. 



The plan proposed below depends on the view that climatic cycles are 

 fragmentary or temporary variations and not necessarily of any definite 

 contour like a sine curve and that they must be extracted from a complex 

 mixture. 



In developing a plan we have three aids. First, a similar type of cycle 

 is found in the sunspot numbers and in certain irregular variable stars; sec- 

 ond, a different and far more adequate method of analysis has been found in 

 the cyclogram process, by which the length, the duration, and, to a great 

 extent, the contour or amplitude of the cycle may be read off at a glance; 

 and, last, the correctness of a given climatic theory involving cycle solution 



133 



