Forest canopy effects on sample size requirements in snow accumulation and melt comparisons

Forest management regulations in British Columbia attempt to minimise changes in the hydrologic regime by limiting the extent of forest removal. Considerable effort is currently being focused on the quantification of snow pack differences between a variety of forest treatments and stand types. This paper assesses what size of difference can reliably be detected by the methodologies used in these studies. We examine the sample sizes required to detect specified differences in April 1 st. snow water equivalent and in the average melt rates resulting from harvesting an Engelmann spruce -subalpine fir forest. Snow water equivalent was measured at 64 points in a clearcut, in thinned and unthinned juvenile pine stands, and in a mature forest. A power analysis was used to determine sample sizes necessary to detect specified differences. The standard 10 point sampling approach was found to be adequate to detect differences of greater than 6 cm in snow water equivalent, and 1 cm d-l in the mean daily snowmelt rate calculated over a six day period, for a= 0.05 (Type I error) and power = 0.9 (Type n error). Increasing sample size to 30 points substantially decreased the size of the difference in snow water equivalent that could be reliably measured. However, 50 samples were required to determine differences in melt rate between the forest and clearcut of 0.5 mm d-l.