14 CLIMATIC CYCLES AND TREE GROWTH 



trees; it may extend over considerable areas; it is obviously not accidental, 

 but comes from climatic factors in the environment common to the trees of 

 the area. The study of cycles given below seeks to extend this investigation 

 of similarities in the environment by picking out the slower and more pro- 

 longed reactions in tree growth to climate by the aid of a new method of 

 analysis. At the same time, the expression of ring records and other forms of 

 long records in terms of cycles puts their details in a form more compact and 

 more readily available for human use. 



RAINFALL CORRELATIONS IN NORTHERN ARIZONA 



The powerful cross-identity over the plateau area of northern Arizona and 

 northwestern New Mexico has raised the question : What are the climatic or 

 other elements that make it possible? 1 Extended discussion of that im- 

 portant problem will be reserved for a future volume. It is proposed here 

 to give only a very few of the facts. Before 1919 it was found that in the 

 time from 1867 to 1910 rainfall and ring growth at Prescott, Arizona, had a 

 correlation coefficient of about 0.52 ± 0.05. This rose to well over 70 per 

 cent when a conservation correction was introduced. Rainfall and tree 

 growth near Flagstaff gave an unsatisfactory correlation which was somewhat 

 improved by using winter rainfall only. The trees used in this comparison 

 were from the forest interior and so were less sensitive than those used in the 

 Prescott tests. 2 



In 1933 Dr. Glock and I attacked the problem of ring growth and rainfall 

 once more, using a number of our most important growth curves representing 

 some sixty trees from various parts of the area centering on the Chuska 

 Mountains. Rainfall curves were constructed from the longer records at 

 Flagstaff, Prescott, and Natural Bridge. It was found that the direct correla- 

 tion was again between 50 and 55 per cent. The introduction of a conserva- 

 tion factor based on a lag in smoothed curves observed by Dr. Glock, raised 

 the correlation coefficient to between 70 and 75 per cent. When, however, 

 the rainfall curve with conservation and tree-growth curves were smoothed in 

 the fashion used for many years in cycle analysis (using a running mean of 

 three with double weight at center), the mean correlation coefficient between 

 rainfall and tree growth rose to 80 per cent. In figure 3b, the curves showing 

 this relationship are reproduced. Figure 4 extends the comparisons by giv- 

 ing the runoff of the Rio Grande and the Colorado River near the northern 

 boundaries of New Mexico and Arizona, and corresponding curves of ring 

 growth. Figure 5 shows similar relation, less close, between southern Cali- 

 fornia rainfall and tree growth in the San Bernardino Mountains and in the 

 giant sequoias. 



1 Carnegie Inst. Wash. Year Book No. 32, 1932-33, p. 209. 



* Mr. W. E. Davis, at Berkeley, made in 1933 a series of measures upon increment 

 cores secured by Mr. Pearson at the Fort Valley Forest Experiment Station, Flagstaff. 

 His results were not at all promising for a relation between ring growth and winter pre- 

 cipitation. On examining the specimens used by him, I found them all giving com- 

 placent records, and that the better ring records had been discarded on the ground that 

 the ring identity was uncertain. 



