|
<br />...,.
<br />,- .. .,
<br />, ~""", l.:..... .....".,_., ,..... / " . ..,.' . ,-,'
<br />~'. .,~a~'l:.b'.-i.t~!"';').fl~-..,~.,~,,.I,,,~'l
<br />
<br />, :,,~",,",; ,.."._A,"_~,~ .,_",' ..,', .
<br />
<br />PRECIPITATION GAUGE SITES
<br />
<br />685
<br />
<br />For the assessment of a sites where direct measurements are missing or are no longer possible to obtain,
<br />a classification of GSE was suggested by Sevruk and Zahlavova (1992). It consists of four classes, 1-4, and
<br />three interim classes 1'5,2,5 and 3,5. The classes are characterized by (X values, as shown in Table I: normal
<br />scale. In Figure 2, the four classes are presented as a series of drawings, to enable a visual appreciation of
<br />how a particular GSE class should look. Consequently, (X can be used as a measure of GSE classes and
<br />VIce versa.
<br />The real problem is that there are no established classifications of GSE and they are not made routinely
<br />by meteorological services. Therefore, assessment of the GSE has to be made solely by the interested party.
<br />The present paper suggests a method of assessment of GSE classes from the station history records and
<br />compares the results with direct measurements in the field in order to assess the accuracy of GSE class
<br />estimates. In addition, applications of the method are presented. These applications concern the problem
<br />of detection of inhomogeneities in precipitation time-series due to changes of GSE during the period of
<br />observation, and the assessment and mapping of precipitation corrections for the wind-induced losses.
<br />
<br />METHODS
<br />
<br />The assessment of GSE classes is based on photographs, sketches, and written reports of gauge sites, as
<br />found in the archives of national meteorological services. The assessment was carried out by three people,
<br />separately: an inexperienced student, a partly experienced, young assistant, and an experienced, older
<br />assistant, two of them geographers and one a civil engineer trained in hydrology. All three received a brief
<br />introduction to the problem. The student started work immediately, without any more advice or supervision.
<br />In contrast, both assistants were carefully advised and trained on examples through the first day and
<br />consulted during the trial. The results for all three people have been compared with GSE classes as estimated
<br />from the measured values of (X for the same gauge sites and the same time period. The number of gauge
<br />sites included in the analysis was different for each person. This was partly due to the fact that the student
<br />was advised to work very quickly for a period no more than 2 days. In that short time he checked a lO-year
<br />period of 50 sites for GSE. In the case of the two assistants only the estimates of GSE classes were considered
<br />in the analysis, which were made after 2 weeks work for the younger assistant and 2 months work for the
<br />older assistant In total 40 gauge sites were analysed by the two assistants. The aim of using different
<br />approaches was to see if there is some dependency of error in the estimate of GSE classes on the degree
<br />of experience of a particular person. It is interesting to note that the normal scale as suggested in Table I
<br />
|