Laserfiche WebLink
<br />001466 <br /> <br />III.) <br /> <br />Consistency of records <br /> <br />, <br /> <br />The ideal temperature record is one that has been collected at a <br />representative site where there has been little or no change in exposure, <br />instrumentatiOn, or procedure in Observing. Very few are the stations <br />whose records satisfy these criteria, so that in 1II0St uses the records <br />first should be tested for consistency or homogeneity. Linsley, Kohler <br />and Paulhull (Applied Hydrology, 1959) describe a method for so testing <br />homogeneity. Briefly, observed values of temperature at the station in <br />question are plotted against the average of values at several nearby" <br />stations, with the plotted points identified according to year of <br />observatioJ;l. On such a plot, a change in slope or a break in continuity <br />of the re~ession line may indicate a marked change in exposure or in <br />some other. aspect of the station environment. Commonly the :full record <br />plots in a~series of regression lines that are en echelon but substantialRy <br />parallel. ,In such a situation, the early records can be adjusted to lllIltch <br />the current record by the algebraic addition of a constant or constantsi <br />Any such adjustments of a temperature record should be conservative. Under- <br />adjustJnent'is IIlUch to be preferred to over-adjustment. In general, an . <br />adjustJnent .' should be made only when justified by a kn01ll1 change in exposure <br />of the station, or in some other oompetent factor. <br /> <br />The Weather Bureauts "Key to Meteorological Records DocUlllentation, <br />No. 1.1, S*bstation Histories", published by states, summarizes infol'lllatibn <br />on location, altitude, exposure, kind of record, and observers for <br />climatOlogical substations. Although not all are complete, these histories <br />are ver.y ulleful when testing records for oonsistenct. <br /> <br />Use Of data <br /> <br />Determining averages or normals <br /> <br />In oOlllputing "normal" or average values of air temperature, care <br />should be taken that the measurements represent the same intervalof' time <br />and unchanged environmental conditions. As the standard periods for <br />temperature normals the Weather Bureau has adopted consecutive 30-year <br />periods as:folloW:SI 1 Januar.y 1901 to 31 December 1930,1 January 1931 <br />to 31 Decelllber 1960, etc. "Average" or "long-term mean" should be used <br />for reference to any other period. <br /> <br />I <br /> <br />Timing of observations <br /> <br />MaximUm and minilllUm daily temperatures are recorded and published as, <br />of the day ion which the observations were made. Thus, at a station llhere <br />the daily observations are take.n at 8:00 a.lII.,' the m.a:ximum temperature ,as <br />recorded n~arly lIlwaysoccurred on the previous day . The time of observation , <br />at each station is published in the station index of the IIIOhthly climatological <br />bulletins. <br /> <br />'f, . _,.'t <br /> <br />