FLUCTUATIONS AND SECULAR TREND OF NEW ZEALAND 



RAINFALL 



By C. J. Seelye, Victoria University College, New Zealand 



[Abstradi^)] 



A list of annual rainfall indices for both North and South Island is 

 given, the index being the average over a selection of stations of their 

 rainfall expressed as a percentage of the 1911-40 mean. Accurate indices 

 were obtained from 1911 to date, and with decreasing reliability back 

 to 1852. In the majority of years the dispersion of rainfall amongst New 

 Zealand, stations has a standard deviation from 14 to 19 per cent. 



The fluctuations from year to year are resolved into a longer period 

 by smoothing, and the residual gives a shorter period. The coefficient 

 of correlation of the former with sunspot number is 0-47 for the North 

 Island and 0-31 for the South and over an average cycle the rainfall 

 moves closely in phase with the sunspots through ranges of 13 and 4 

 per cent, in the two islands respectively. For the shorter fluctuation 

 there is an average period of. about three years and average range of 

 21 and 17 per cent. Both period and amplitude are irregular, and there 

 is little similarit}' between the two islands except at or immediately 

 after sunspot maximum. 



Using moving thirty year means from the 1863-1947 indices average 

 rates of decline of 4 per cent, and 2 per cent, per century for North and 

 South Island are obtained. Recognition of a trend requires more critical 

 examination of the indices before 1890 when isolated records cannot be 

 satisfactorily reduced to standard epoch. Assuming the average trend 

 of the island could be applied in such cases most of these earlier indices 

 should be increased by 1 or 2 per cent, and the revised trends become 

 6-7 and 3-3 per cent, per century. 



Further knowledge of the three-year cycle over the Pacific region seems 

 desirable if the irregular variations noted are to contribute to any method 

 of long range forecasting. 



Discussion 



In the discussion Captain Best raised the question of the significance 

 of the correlation coefficients given in this paper. Dr. Seetye replied that 

 the graphical representation in his paper gave one more confidence in 

 the correlation coefficients, and although they were not particular^ 

 high he thought that they were of significance. 



Mr. B. W. ColUns offered a ^'ery interesting contribution from the 

 point of view of a ground- water geologist. The figures given by Mr. 

 CoUins showed that the level of ground-water in Canterbury in recent 

 years tended to confirm the findings of Dr. Seelye. Dr. Berlage asked 

 if a similar study with regard to seasonal variation of rainfall should 

 not be undertaken. Dr. Seelye agreed that such a study would be well 

 worth while. 



In reply to a question by Mr. Mordy regarding the distribution of the 

 stations used to compute the rainfall indices. Dr. Seelye replied that they 

 were selected in the same way as those of the British Rainfall Organiza- 

 tion, in order to give an adequate coverage of the rainfall of the whole 

 countr^'. 



(1) The full paper will appear in the Xeiv Zealand Journal of Science and Technology. 



103 



