WSP10386 - Laserfiche WebLink

Home Browse Search

00126Q -37- Consistency of records stability of site An ideally consistent snow-survey record shoUld result rrom measure- ments at regular intervals with no change in site, sampling'points, equipment, or technique of observation. All too frequently, however, changes have been made but not dOCUMented adequately. Accordingly, rec- ords never should be aSSUMed to be homogeneous; all records should be treated as though biased by some un-documented change along the course or in its vicinity. Physical changes at a snow-course site may be natural or man-caused. Natural changes are mainly an increase in tree canopy or growth of brush that changes the interception of snow and modifies the movement of wind. Man-caused changes usually result from road improvement, construction, or logging. Logging, in particular, can change the characteristics of a snow course so much that, unless they can be adjusted soundly, past records become useless for future analysis. Change in location and in number of sampling points Marry snow courses originally had from 20 to 40 sample points; some, as many as 100 points. Research and experience have shown, however. that the optimum number of points is from 10 to 15. The number of sample points on many snow courses has been reduced accordingly) generally by selecting points whose mean value for water equivalent of the snowpack as of April 1 is the same as the mean from the full array of the older data. Recent summaries of snow-course records include adjustments for those older courses whose number of sample points has been diminished. Such adjustments usually cover the full term of record and all months for which snow-survey information is available. Density estimates for aerial-marker observations Published records from aerial markers generally derive a value for water equivalent of the snowpack from an estimated density of snow. Provided that representative values of density are available from sur- rounding snow courses or stations, such assumption of a density factor to apply to aerial-marker readings generally does not introduce an error greater than five percent. However, in early-season aerial surveys there may be no data from nearby courses or stations and the estimated density factor may be in error as IllUch as 25 percent. Tests for consistency In addition to acknowledged statistical procedures and the much used double-mass plotting of records against one another, several methods are available to check the consistency of snow surveys. For example, plotting the historical records of all surveys at each sample point of a snow course will disclose a pattern characteristic of that course. Commonly, poor,or