Laserfiche WebLink
<br />Section 2.02. Estimating Missing Precipitation at a Point <br /> <br />It is often necessary or desirable to supplement incomplete precipi- <br /> <br /> <br />tation records by estimating values that are missing at one or more <br /> <br /> <br />stations. One procedure that can be used is to estimate the precipitation <br /> <br />from that observed at stations as close to and as evenly spaced as possible <br /> <br />around the station with the missing record. Experience has indicated that <br /> <br /> <br />at least three stations should be used. If the average annual precipitation <br /> <br /> <br />at each of these stations is about equal, a simple arithmetic average of the <br /> <br /> <br />precipitation at the three selected stations provides the estimated amount. <br /> <br /> <br />If there is a significant difference in the average annual precipitation <br /> <br /> <br />of the stations, then further adjustments should be made or another techni- <br /> <br /> <br />que should be used. <br /> <br /> <br />Another method of estimating missing data is by means of regional <br /> <br /> <br />analysis, as discussed in Volume 2 of this report. This method determines <br /> <br /> <br />by multiple-linear regression the correlation between the station with the <br /> <br /> <br />missing data and all other nearby stations for the period when they have <br /> <br /> <br />coincident records. By using the resultant regression equation and adding <br /> <br /> <br />a random component, the missing record can then be estimated. <br /> <br /> <br />A more commonly used technique is the mass precipitation curve, illus- <br /> <br /> <br />trated in fig. 2.01. Mass precipitation curves are determined by plotting <br /> <br /> <br />accumulative precipitation versus time. Curves are first plotted for con- <br /> <br /> <br />tinuous recording stations within and around the study area, and then simi- <br /> <br /> <br />lar curves are constructed for nonrecording stations,in the same area, using <br /> <br />the recording station curves as a guide. Stations should be compared for <br /> <br /> <br />possible similarities of topographic influences and meteorological condi- <br /> <br /> <br />tions during the storm and grouped accordingly. The mass precipitation <br /> <br /> <br />curves should be completed in accordance with these groupings by interpo- <br /> <br /> <br />lating the curves between established points in such a manner as to reflect <br /> <br /> <br />reasonable consistency with the period of precipitation at neighboring <br /> <br /> <br />stations. Ineremental values can then be extracted from the curves for <br /> <br /> <br />use in determining mean precipitation over an area. Care must be used in <br /> <br /> <br />this procedure, however, because minor smoothing of mass curves could <br /> <br /> <br />change intensities greatly. <br /> <br />2-02 <br />