The impact of snow survey frequency on Australian snow depth records

Regular manual measurements of snow depth have been made in the Australian Alps since the 1950s and snow courses such as Spencers Creek continue to provide a primary record of snowpack variability in this region, being used for a range of scientific studies while also attracting considerable public interest. In this study, we investigate how the frequency of snow surveys within a season affects uncertainty in the representation of snow depth as well as in commonly derived seasonal metrics. Automatic measurements of snow depth at two locations, one with relatively persistent snow cover and the other more marginal, were used to simulate sequences of snow surveys at sampling intervals of between 1 and 4 weeks. Uncertainty was related to how well these sequences approximated actual temporal variability and, in each metric, increased with an increasing sampling interval. However, we show that the increase in uncertainty with time is greatest in the first week following a survey, suggesting that even weekly snow surveys should not be assumed to provide a complete representation of actual snowpack variability. The impact of the sampling interval on the estimation of snow depth varied intra-annually and was greatest during periods of rapid change, most notably later in the season as the snowpack underwent ablation. An awareness of the factors influencing uncertainty in manual snow depth records can be used to guide decisions relating to operational measurement programs and will also be beneficial to end users of these datasets.