Hydroclimate change that alters runoff seasonality is an increasing concern
in snowmelt-influenced watersheds throughout the western United States. De
tecting seasonal runoff change is complicated by the interaction of atmosph
eric and topographic influences and by human alteration of streamflow. Reco
nstructed or unimpaired discharges for 1921 to 1994 for the Sacramento Rive
r are employed as an indicator of natural discharge and are analyzed for ev
idence of seasonal change. Rescaled adjusted partial sums (RAPS) improve vi
sualization of trends and fluctuations in annual discharge and the seasonal
fractions of annual discharge. The statistical significance of trends indi
cated by the RAPS is evaluated using nonparametric tests. Small increasing
trends beginning in the late 1940s in annual unimpaired discharges and the
seasonal fractions for fall, winter, and summer are not statistically signi
ficant. In contrast, the spring fraction of annual discharges shows a decre
asing trend beginning in the late 1940s that is statistically significant.
The decline in spring fraction discharges when annual discharges are increa
sing is consistent with a greater proportion of annual precipitation in the
watershed occurring as rain rather than snow. Reduced spring discharges wi
ll impose new limits on water-management alternatives for maximizing use of
the water supply.