FIGURE LEGENDS
Figure 1: Innere Ölgrube rock glacier and its catchment area. (a) Location of Innere Ölgrube within Austria relative to all 5769 rock glaciers mapped in the Austrian Alps (Wagner et al., 2020b,c).(b) Topographic map of the Innere Ölgrube rock glacier and its catchment (red polygon) including the extents of two cirque glaciers for the years 2006 and 2015. Location of the gauging station, of the precipitation and the glacier meltwater samples and the two artificial tracer injection points are depicted. Rock glacier and catchment extents from Wagner et al. (2020b,c); glacier extent from the Austrian glacier inventories GI3 (Fischer et al., 2015) and GI4 (Buckl & Otto, 2018). (c) Field impression of the two tongues of the active Innere Ölgrube rock glacier seen from the western side of the Kauner valley viewing towards east. Note the active (unweathered) steep slope of the fronts; the creek (“Schiltibach”) which emerges below is the result of several springs at the base of the rock glacier front.
Figure 2: Data overview. Precipitation and air temperature data (from the Weißsee station (TIWAG) and corrected to the average elevation of the catchment), isotopic data sampled at the gauging station as well as within the spring catchment, electrical conductivity and discharge from the gauging station.
Figure 3: Model structure of the rainfall-runoff model GR4J+ with a daily time step and an additional ice store (modified after Wagner et al., 2016). P = precipitation; T = air temperature; Ts = temperature at which snow starts to fall; Tm = temperature at which snow starts to melt; Cm = melt factor that allows a certain amount of snowmelt per degree temperature increase; Tim = temperature at which ice starts to melt (if snow is absent); Cim = melt factor that allows a certain amount of ice melt per degree temperature increase; Re = extraterrestrial solar radiation; x1 = maximum capacity of the production store; F = groundwater exchange term acting on the fast and slow flow components; x2 = water exchange coefficient; x3 = maximum capacity of the routing store; x4 = time parameter; UH1 and UH2 = unit hydrographs to account for the time lag between rainfall and resulting streamflow that depend on the time parameter x4; Q = runoff simulated by the rainfall-runoff model.
Figure 4: Master recession curve and the interpreted recession coefficients shown as straight lines in a semi-log plot and the spring hydrograph as an inset. Modified after Wagner et al. (2020a) and extended with more recent data (a time period in which also natural tracer data became available).
Figure 5: Artificial tracer breakthrough curve of (a) 2015 and (b) 2017 with a range of tracer recovery reflecting uncertainties in the runoff computation. Note the different scale of the tracer concentration for (a) and (b); tracer recovery and discharge are displayed at the same scale.
Figure 6: Natural tracer data. Runoff separation based on EC related to longer stored / higher mineralized groundwater. (a) Snowmelt period; (b) summer to autumn period with diurnal variations that are shown in more detail in (c) including air temperature, precipitation and isotopic data. (d) Observed hysteresis in EC versus isotopic data indicates dynamic event water composition (ice melt, rainfall, and groundwater).
Figure 7: Observed and modelled specific discharge of the Innere Ölgrube spring catchment. Visual fit between observed and modelled discharge and related precipitation input data. Differentiation of rainfall, snowmelt and ice melt input as well as snow accumulation over the winter periods.
Figure 8: Monthly contribution of “recharge” (input into the rainfall-runoff model without considering potential loss due to evapotranspiration) and discharge components (based on event water analysis); absolute values normalized to catchment area. Shaded area indicates total flux. Dashed lines indicate assumed groundwater/event water contribution during periods of low discharge (no EC record available). Numbers next to bars indicate the number of months analysed (see Figure 2).
Figure 9: Observed discharge (dark blue line), modelled discharge (green line) without the input from the ice store, assuming cirque glaciers have vanished; and modelled discharge (purple line) without the input from the ice store and applying model parameters from the relict Schöneben rock glacier (Table 2; cf. Wagner et al., 2016); assuming the cirque glaciers as well as all the permafrost ice have vanished. Red line refers to the simulated runoff as in Figure 7.
Figure 10: Master recession curves (MRC) of observed as well as simulated runoff; in addition the simulated MRCs without the ice melt and the hypothetic scenario of a relict rock glacier in the (far?) future.