A considerable number of papers have been written about the possibilities and even more about the "inpossibilities" of measuring precipitation at high alpine regions. High wind speeds, low temperatures, a high percentage of solid precipitation and an increased probability of roar-frost are the limiting factors for precipitation measurements in such regions.
In Austria three different types of precipitation measuring systems (manned rain gauges, totalisers -shielded or unshielded, automatic systems) have been in use since many years. The following investigation focuses on one selected region in the Goldberggruppe / Hohe Tauern and dealing with problems and intercomparisons of different precipitation measuring systems at the Sonnblick Observatory located on a mountain top in 3106 m asl.

The database of the investigation is shown in Table 1 and Figure 1 respectively. All instruments are located in a rectangle between 46°40'- 47°30' N and 12°40'-13°15'E.

In Auer, 1992 a comparison of the rain gauge South (OS) and the nearby weighing gauge (WG1) showed a surplus of precipitation for the weighing gauge of 33 mm for a period of three months (September - November 1991). However, this study clearly showed that this surplus was due to 33 days where the weighing gauge registered precipitation although there was no precipitation at all. In the meantime much effort has been undertaken to rise the quality of the measurements. A comparison between the weighing gauge and the manual gauge for the period January 1998 - September 1999 is shown in Figure 2. During these 17 months the weighing gauge provided reliable results in respect to amount and time resolution. The correlation coefficients of the daily values between the two gauges exceed 0.7 most of the time (comp. Fig. 2). There is also a good agreement between weighing gauge and the totaliser TG4 (comp. Fig. 3). The correlation coefficient based on monthly sums between WG1 and TG4 come to r= 0.79 for the period January 1988 - September 1999. Moreover the measurements of the weighing gauge are able to confirm the measurements of the totalisers with annual precipitation amounts of 2500 mm.

In Auer, 1993 a long-term series for grid point 13 deg E and 47 deg N was evaluated from homogenised long term precipitation series. The deviations to this grid point of some selected stations are shown in Figure 4. Out of these the valley station Bad Gastein fits best to this grid point. On the contrary, the long term series of rain gauge North (ON) gives a picture far from any reality and unfortunately, cannot be used for any statistical time series analyses. The rain gauge South (OS), and best of them the totaliser TG4 seem to provide better results.

Table 2 shows the correlation coefficients between precipitation series in the Sonnblick region for the period 1961-1990, split into two profiles north and south of the alpine main crest including the valley stations which normally show a higher accuracy of the measurements. The highest correlation overall occurs between TF1 and TF3 between May and September and TG4 and TG5 in the winter season.
The valley station Rauris (profile north) is best correlated to the totalisers TG1 and TG3 from October to April. Concerning the measurements on the mountain top the totalisers give better results than the two rain gauges. Due to a higher percentage of convective precipitation events the correlation coefficients are in summer smaller than in winter. Nevertheless the totalisers are higher correlated to the valley station than the rain gauges ON and OS. On the contrary, the valley station Heiligenblut (south profile) shows in winter the best agreement to the rain gauge on the north side of the observatory. In summer a higher correlation appears to the totalisers TF1 and TF3 compared to the rain gauge ON.

Since 1986/87 winter mass balance measurements have been carried out on the glacier Goldbergkees (Schöner, 1995), which may be interpreted as an "independent, external rain gauge". Two of the profiles Lieslstange und Bockpalfen will be compared with the precipitation measurements on Sonnblick. Lieslstange is situated a little bit lower, but quite near to the Observatory with an altitude off 3000 m, Bockpalfen more centralised on Goldbergkees in an altitude of 2680 m asl. Additionally to profiles from the nearby glacier Wurtenkees (Chemieprofil in 3000 m asl and P5 in 2650 m asl., Auer et al., 1995) are included as well. Fig.5 shows these comparisons, and Tab.3 lists the correlation coefficients.

Highest correlation values to the winter mass balance measurements on Goldbergkees occur at OS and TG1 followed by TG4 on the mountain top and TG3 Rojacherhütte. The worst representation of the water equivalent of the accumulated snow on the glacier is given by TG2 and ON. The Chemieprofil correlates best with OS, P5 with the TG1 Kolm-Saigurn. Besides the correlations also the total amounts are of interest. Inspite of the high correlation between OS and the accumulated winter snow OS shows a great deficiency. Concerning precipitation amounts the totalisers provide better results.

Including or excluding the values of totalisers in regard to precipitation maps makes a great difference in the final outcome by using the same methods of map construction. This was shown in Auer et. al, 2000, where precipitation maps for the Sonnblick region were computed for the period 1961-1990 using or excluding the totaliser values. The maps were produced in Arcview, GIS using a Wedge-filter application along the alpine main ridge..
Below 1500 m altitude only small differences occurred, whereas big differences characterised the alpine main ridge itself. For the annual precipitation amount the deficiency exceeded more than 1000 mm in the summit region and higher parts of the northern inclines, the southern region was less affected. The deficit during summer season exceeded 700 mm.

The paper intended to compare the results of different methods of precipitation measurements. Advantages as well as disadvantages of the various instruments were pointed out. For precipitation information with a high resolution in time the weighing gauge gives the most reliable information. The ombrometer on the south side of the Sonnblick Observatory provides good correlation results to the independent winter mass-balance measurements but the total amount of precipitation is underestimated. Totalisers give best results of monthly amounts, but the time resolution is insufficient.

Auer, I., 1992: Precipitation Measurements in a high Alpine Region- inter-comparisons of different measuring systems. TECO 92,WMO/TD No.462, 251-256, Vienna.

Auer, I., 1993: Niederschlagsschwankungen in Österreich seit Beginn der instrumentellen Beobachtungen durch die Zentralanstalt für Meteorologie und Geodynamik. Österr. Beitr. zu Meteorologie und Geophysik, Heft 7, 73 S , Wien.

Auer, I., R. Böhm, N. Hammer, W. Schöner, T. Wiesinger, W. Winiwarter, 1995: Glaziologische Untersuchungen im Sonnblickgebiet: Forschungsprogramm Wurtenkees. Österr. Beitr. zu Meteorologie und Geophysik, Heft 12. Zentralanstalt für Meteorologie und Geodynamik, Wien.

Auer, I. und R. Böhm, 1998: Schneepegel und Totalisatoren im Sonnblickgebiet. Jahresbericht des Sonnblickvereines für die Jahre 1996 und 1997, S 42 - 87, Eigenverlag des Sonnblick-Vereines, Wien.

Auer, I., R. Potzmann, W. Schöner, 2000: Welchen Beitrag leisten Totalisatoren für die Klimaforschung im Hochgebirge? - gezeigt mit Hilfe des Totalisatorenmessnetzes im Sonnblickgebiet. Jahresbericht des Sonnblickvereins für die Jahre 1998 und 1999, S 22 - 30, Eigenverlag des Sonnblick-Vereines, Wien.

Schöner, W., 1995: Schadstoffdeposition in einer hochalpinen winterlichen Schneedecke am Beispiel von Wurtenkees und Goldbergkees (Hohe Tauern). Dissertation Universität Salzburg, Salzburg.

Schöner, W., I. Auer, und R. Böhm, 2000: Klimaänderung und Gletscherverhalten in den Hohen Tauern. Salzburger Geographische Arbeiten, Bd. 36, Salzburg.

Skoda, G., 1993: Pilotstudie zu I. Korrekturmöglichkeiten von Fehlern bei der Niederschlagsmessung und II. Möglichkeiten der Bestimmung der Höhenabhängigkeit des Niederschlages, Anhang, Teil E: Erfassung und Auswertung des Niederschlages. Hydrographisches Zentralbüro im BmfLuF-, Wien.

Zentralanstalt für Meteorologie und Geodynamik, 1999: Bericht über den Sonnblick-Workshop Umweltforschung im Hochgebirge - Ergebnisse von GAW-DACH und verwandten Projekten, 5. bis 6. Oktober 1999, Wien. Österr. Beitr. zu Meteorologie und Geophysik Heft 21. Zentralanstalt für Meteorologie und Geodynamik, Wien.