Central Water and Power Research Station, Poona.
For fixing the requisite number of raingauges in any river basin for the purpose of its hydro-logical study it is the usual practice, apart from topographical considerations, to compute the mean precipitation depths of the existing numbers of raingauges for the relevant period of interest. Such means are computed separately for a number of years and their stability or year to year variational extents studied.
Since the standard deviation has come into vogue as a measure of this stability, its highest observed value from all the recorded numbers of years is picked out and set against the desired control magnitude for determining the increased number of raingauges necessary. In other words the relative magnitudes of all the other lower yearly observed values of the standard deviation are not assigned any useful role.
The paper outlines the use of the frequency analysis method of all the standard deviation values or the coefficient of variation (standard deviation/mean) values as better appropriate. The requisite estimates so derived are also shown to be relatively better consistent irrespective of the lengths of the data availed for the purpose, against the large inconsistencies obtaining of the estimates otherwise. This consistency aspect obtaining particularly from the use of the frequency analysis method thus immediately warrants greater confidence in the fair accuracy of the estimates as well as of the procedure.
