How isotopic fractionation of snowmelt affects hydrograph separation

We present the isotopic composition of meltwater samples from four seasonal snowpacks: a warm, maritime snowpack in California; a temperate continental snowpack in Vermont; a cold continental snowpack in Colorado; and an Arctic snowpack in Alaska. Despite the very different climate conditions the delta(18)O of meltwater from all four snowpacks increased as melting progressed. This trend is consistent with theoretical results that model isotopic exchange between water and ice as meltwater percolates through a snowpack. We have estimated the systematic error in the hydrograph separation if the isotopic composition of a snow core were used in place of that of meltwater. Assuming no error in the old water or stream water values, the error in the new water fraction depends on: (1) the isotopic difference between the snow core and the old water; (2) the isotopic difference between the snow core and the meltwater; and (3) the new water fraction contributing to the stream flow during a spring melt event. The error is large when snowmelt contributes a dominant fraction of the stream flow, which may be expected where infiltration of snowmelt is limited (e.g. permafrost, urban areas). A particular challenge will be how to incorporate the changes in isotopic composition of meltwater measured at a point into hydrograph separation models conducted at the watershed scale. Published in 2002 by John Wiley Sons, Ltd.