1921.] Birks and Bates.—Annual Rainfall Fluctuations. 
173 
ANNUAL RAINFALL FLUCTUATIONS. 
By L. Birks. B.Sc., and D. C. Bates. 
There are several important industries, particularly fanning and the 
production of hydro-electric power, in which the question of the probable 
annual fluctuations in the rainfall are of the utmost importance. In the 
case of the farmer, he stands to lose a large value of his product if the 
rainfall drops for the season below a certain minimum or if it rises above 
a certain maximum, and in forecasting the prospects of his returns and 
in determining the amount of the fund which he must maintain in liquid 
form to cover a bad year he has to decide in his own mind what are his 
chances of loss from these causes. The engineer responsible for the design 
of hydro-electric works is required to provide storage suitable for carrying 
supply over the year of lowest rainfall, or else is required to expend 
additional capital in seeking another more distant or otherwise more 
expensive water-power site, or in the installation of fuel-reserve plants. 
In order to determine whether such additional capital expenditure is 
justified he must again assume some frequency at which the years of 
reduced rainfall are likely to occur, and the expenditure of very large sums 
will depend upon the assumption made. In both cases the forecast of the 
future must still remain merely a working hypothesis, and no meteorological 
skill is likely to enable a definite forecast of this description to be based 
upon any scientific determination of what will actually occur in the future. 
But the most that can be done, and the least that should be done, towards 
making such an estimate is to review carefully what has occurred in the 
past, and with this object a preliminary analysis has been made of the 
records available in New Zealand. 
The matter has been brought to our notice recently very forcibly by 
the period of three years of deficient rainfall in many parts of New 
Zealand, which has been broken by this winter’s rains. As a result of 
this, combined with a drop in the market prices for wool and meat, many 
sheep-farmers have had to meet very serious losses this year, which would 
have been much less serious had the rainfall last summer been normal. 
And at the same time many rivers which have been considered good for 
substantial hydro-electric-power developments have fallen so low that the 
estimates of the power available have had to be entirely recast. The 
subject of the probable frequency of future recurrence of such periods 
of low rainfall is thus obviously of the utmost importance. 
For the past fifty years, since the establishment of the Meteorological 
Department, these records have been excellently kept for a number of 
stations throughout the Dominion under various geographical and climatic 
conditions. But the only figures available for a more extended period 
are those of the rainfall in Auckland, Wellington, and Dunedin. In 
Auckland the records go back to 1853, and have been continuous since 
then, a period of sixtv-eight years. In Wellington the records have been 
continuous since 1863, a period of fifty-eight years. In Dunedin records 
are available since 1855, but have been continuous only since 1863, a 
period of fifty-eight years. In Hokitika the records have been taken 
since 1866, but are only continuous since 1894, a period of twenty-seven 
years. Records have been kept for other stations for lesser periods. 
Taking first the three stations for which the longest records are avail¬ 
able—viz., Auckland, Wellington, and Dunedin--fig. 1 gives a graphical 
record of returns from these stations for the last fifty-eight years, showing 
the fluctuations from year to year. The average rainfall over the whole 
