Figure 6: PGRB SWE and albedo DA mean, control, and ensembles for the
four melt seasons at the highlighted glacier ice HRU in Figure 1.
3.4. Albedo DA Differences
Between Snow and Ice HRUs
A prolonged spring and summer snowcover over glacier ice is not the only
cause of decreased streamflow in these basins. Because streamflow at the
outlet of glacierized basins is an integrated response of mainly snow,
firn, and ice melt, the melt at higher elevation snow and firn HRUs is
also relevant to the overall streamflow contribution. Figures 7 and 8
show the modelled albedo during wildfire (2018) and heatwave (2021)
conditions for snow and glacier ice HRUs in AGRB and PGRB, respectively.
Figure 7 illustrates that although AGRB DA albedo is lower for glacier
ice during wildfires and heatwaves, it is higher in snow-dominated
regions. The larger albedo in AGRB snow compensates for the smaller
albedo in glacier ice in years (e.g., 2018) when DA improved streamflow
predictions. Figure 8 shows that albedos were commonly smaller with DA
in PGRB snow, but in ice, they were only persistently lower than CTRL
during heatwaves. A larger DA PGRB ice albedo explains why streamflow
prediction improvements are better for this basin since the positive
bias in CTRL is decreased even further. During the 2021 heatwaves, DA
could not substantially improve streamflow predictions in either basin.
However, AGRB DA was closer to observed streamflows when compared to
PGRB. This difference can be explained by the longer persistence of DA
snowcover over ice in PGRB than AGRB. DA snow albedo in PGRB was closer
to CTRL, thus contributing to an even smaller streamflow for that year
(Figure 8). The results suggest that the albedo decay algorithm was
capable of simulating albedo in a heatwave, but could not predict the
lower albedos due to soot from wildfires.