Abstract
Since 1948, when the Weather Bureau discontinued the practice of publishing interpolated precipitation data, monthly and annual totals have been omitted from its publications whenever any portion of the record was missing. In view of objections by the users of precipitation data, this policy was reviewed for possible revision. A series of tests involving some 1200 storms was conducted to determine if missing records could be satisfactorily estimated by simple procedures which could be applied in the limited time available. Results of these investigations are presented and two methods of interpolation, namely, the normal-ratio and 3-station-average, me selected for use by the Weather Bureau. Procedures to be followed in applying these methods are outlined.
Because of the early deadlines for the issuance of monthly Climatological Data, estimated data, properly identified, will appear in the annual number only, beginning with the 1952 issue. In order to limit the magnitude of the errors of interpolation, annual totals including more than one-third estimated precipitation will not be published.
