Peak timing

As our world warms, our snowpack thins and our snow season grows shorter, but is that all simple and uniform or is there more to it? Those who have done the work¹² to study and model the physical processes of snow accumulation, consolidation and melt have an answer; perhaps a surprising one:

CSIRO/BOM model snow depths at Mt Hotham, now and at 2050

CSIRO/BOM model snow depths¹ at Mt Hotham, historical and at 2050 (©CSIRO)

The black line is the average depth at 1990 (modelled snow depth for Mt Hotham, Victoria, Australia) and the coloured lines are various projections for 2050. Seems there’s much more decline on the melt side of the curve than on the accumulation side, and the greater the warming effect³ the greater the asymmetry.

Surely that’s obvious though — warmer weather means more and earlier melt. But wait a bit; if you look closely, the melt rate — the slope of the decline — doesn’t change all that much, at least for the smaller declines. The big change is in the timing of the onset of melt. Just the model? No, much the same effect is apparent in the recorded data (this is Snowy Hydro’s 60-year Spencers Creek record, from midway between Perisher Valley and Thredbo, NSW):

Spencers Creek snow depths by decade

Spencers Creek snow depths by decade


Warming affects our snow in multiple ways: by decreasing the winter proportion of snow vs rain precipitation, by decreasing total winter precipitation (overall drying in south-eastern Australia), and by increased and earlier melt. What I think is going on is that the roll-over from net weekly snow accumulation to net weekly loss is happening earlier and earlier, because warmer weather means the rain proportion of precipitation⁴ is increasing rapidly in the later months of the season. Unfortunately good data on that seems hard to come by.

There is another change which ought to attach to this asymmetrical reduction. The timing of the peak snow depth should change, becoming earlier as the climate warms and the snow depth declines. In fact that is clear in the Spencers Creek data, and has been for a long time:

Spencers Creek peak snow depth timing


Peak snow depths used occasionally to occur in October; we now see them as early as July (in 2011). What of 2014? Tomorrow’s measure at Spencers Creek is unlikely to beat the 168.5 cm recorded on 24 July, so the first chance of a new peak will probably be Thursday 7 August. On the trend stats (and nothing else), that means the probability that we’ve already seen the 2014 peak has now risen to about 16%.



1. Climate change impacts on snow in Victoria (2012), J Bhend, J Bathols and K Hennessy; Centre for Australian Weather and Climate Research (a CSIRO / Bureau of Meteorology partnership)

2. The impact of climate change on snow conditions in mainland Australia, K Hennessy, P Whetton, I Smith, J Bathols, M Hutchinson and J Sharples, CSIRO Atmospheric Research, August 2003

3. The spread of curves comes from the spread of local climate change effects inferred from the 18 different global climate model outputs used to drive the snowpack model. I suspect that it derives as much from differing precipitation reductions as from differing amounts of local warming.

4. Total precipitation is also predicted to fall, though that is not yet obvious in the (very noisy) data for south-eastern Australia in my opinion (it is obvious in south-western Australia).