Glacier Mass Balance and Regime: Data of Measurements and ...

decreasing number of measurements” (Jansson, 1999, p. 636). Thus one must be careful about acceptingrecommendations to substantially decrease the number of stations because of correlations which existbetween data at different sites.(2) It is important to keep the effort within reason. It has been shown most recently that very fewstakes are required to estimate mass balance components (results highly correlated), and these severalstakes should be placed along the elevation range of a glacier (Cogley, 1999; Fountain and Vechia,1999). WGMS proposed as the strategy for future mass balance monitoring only 2-3 stakes along themain stream of a glacier (Haeberli, 1995; Hoelzle and Trindler, 1998), which has never beenexperimentally verified.In connection with both of these requirements, two examples taken from published data may behelpful to illustrate the problem:(1) Mass balance of Vernagtferner. The mass balance has been calculated for three ice streamsemerging from the wide accumulation area. Mass balances calculated separately for these three partsshow very large differences for the 1964-96 period, such as: Brochkogel –7 m, Schwarzwand – 5 m,Taschachjoch – 11 m (data are taken from the Fig. 3 of Reinwarth and Escher-Vetter, 1999). This maybe typical for glaciers having several tributaries and local climate conditions. In such cases the numberof stakes must be determined for tributaries, separately.Mass balance measurements by stakes on Variegated Glacier. This surging glacier showsenormously high difference in mass balances between neighboring stakes and changes from year-to-year(FoG, v. 5, see also Appendix, 2). The local topography is very complex, which makes spatial variabilityin snow depth distribution very high, affecting snow and ice melting during a summer. This means thatresults are poorly correlated and many stakes are needed to get accurate estimates of mass balance overthe area.To calculate the effective number of stakes (or pits), a study of the spatial distribution of thevariable is required to determine the statistical structure (isotropy and homogeneity) of a field, whichcan be described by autocorrelation functions, (Gandin, 1963). This, possibly, is the most correctapproach, but it is labor extensive and time consuming.The accuracy of mass balance of entire glacier. It follows from the above that careful studiesare crucial for estimation of the real accuracy of seasonal components and annual mass balance of aglacier. We assume that these were used to get reasonably correct estimates of accuracy for severalindividual glaciers, e.g.: Storglaciären, ± 100 mm/yr (Jansson, 1999); Hintereisferner, ± 100 mm/ yr34