SNOW SEASON FRACTIONAL FLOW

1

• Understanding the sensitivity of Eurasian Arctic
  runoff changes to snow cover-related surface
  energy flux variations
• Amanda Tan1, Jennifer Adam2, Mark Serreze3,
  Dennis P. Lettenmaier1
• Department of Civil and Environmental
  Engineering, Box 352700, University of
  Washington, Seattle
• Department of Civil and Environmental
  Engineering, Washington State University,
  Pullman, WA,Department of Civil and Enviromental
  Engineering
• National Snow and Ice Data Center, University of
  Colorado, Boulder, CO, National Snow and Ice Data
  Center, 449 UCB University of Colorado Boulder

THE DOMAIN
3
ABSTRACT
1
6
SNOW SEASON FRACTIONAL FLOW
Pronounced land surface process changes have
occurred in the Arctic and sub-Arctic in recent
decades. Satellite data have confirmed that
average snow cover has decreased, especially in
the spring and summer. One implication of the
ablation of snow cover earlier in spring is the
potential for a positive snow albedo feedback,
where patchy snowcover leads to reduced effective
albedo, increasing net radiation, and thereby
making energy available for advection from snow
free to snow covered areas, further accelerating
melt. Furthermore, some of this increased surface
energy is available to melt permafrost. Changes
in timing of snowmelt and permafrost melt have
been linked to increased riverine discharge to
the Arctic Ocean which has the potential to
affect the thermohaline circulation and in turn
the formation of North Atlantic Deep Water and
the northward-flowing Gulf Stream. To date,
however, causality for observed hydrologic trends
and changes in climatic forcings to the land
surface system remain elusive, primarily because
linkages between hydrologic and climatic
sensitivities are not well established. This
study explores the role of snow on the
sensitivities of annual and seasonal Arctic river
discharge through analysis of in-situ and
satellite-based observations and a large-scale
hydrologic model applied over Lena River basin in
the Eurasian Arctic. Through analysis of
satellite-based snow cover data for the period of
1972 2001 and observed river discharge for 1958
1999, we attempt to draw associative
relationships between changes in snow cover over
time and their effect on runoff timing.
Streamflow gauges chosen for analysis are
relatively uninfluenced by anthropogenic changes
and we find that while average annual snow cover
has decreased slightly, this apparently does not
account for observed discharge trends winter
discharge increase in 75 of gauges, and June
discharge decrease for 82 of the gauges.
The Eurasian Arctic region was chosen because the
most significant increases in river discharge to
the Arctic Ocean have been observed there
(Peterson et al. 2002). We chose the Lena River
basin because 80 of its drainage area is
underlain by permafrost (Brown et al.,1998
Zhang et al., 1999) and is therefore more
sensitive to changes in temperature and
precipitation. The Lena basin outline based on
drainage characteristics is shown in green,
whereas the R-ArcticNet definition of the Lena
basin is shown in black lines. Some of the
stations selected therefore fall outside of, but
are close to the Lena basin boundary, because of
inconsistencies in the two basin definitions.
(b)
(a)
The Lena Basin

• May
• June
• July
• Seasonal Fractional Flow during Snowmelt (May,
  June, July)

(d)
(c)
4
SNOW COVER CHANGES
Snowmelt occurs during May and June in the
Arctic, with peak discharge usually occurring in
mid-June. Analysis of fractional
(a)
(c)
(b)
Figure 4(a-f) show anomalies relative to the
long-term mean over the entire basin (i.e. in
space and time). The average snow cover
disappearance day starts on Day 99 of the
calendar year (April 1), snow cover onset is Day
228 (August 15). Snow cover disappearance dates
correlate significantly (r 0.1 0.5) with
spring pulse onset dates at 36 of the gauges.
streamflow for the snowmelt period of May and
June show an increase in May monthly fractional
flows at 79 of the gauges, whereas June
discharge is decreasing for 82 of the gauges.
Significant trends are detected on the higher
latitudes and towards the eastern part of the
basin. This is the Aldan basin which is underlain
by 90 continuous permafrost (Brown et al.,
2000). Summer flow is also decreasing with some
gauges registering as much as a 20 decrease in
discharge. The spring pulse onset and snowmelt
season flow is correlated at 65 of gauges. The
snowmelt season fractional flow is decreasing in
86 of the gauges with some of the highest
numbers reaching 30 decreases, particularly in
the eastern basin of Aldan at Lena (which is
underlain by continuous permafrost).
(f)
(e)
(d)
5-year anomalies of number of snow cover days
over the Lena basin (a) 1972 1976 (b) 1977
1981 (c) 1982 1986 (d) 1987 1991 (e) 1992
1996 (f) 1997 2001.
2
METHODOLOGY
WINTER SEASON FRACTIONAL FLOW
7
5
CHANGES IN STREAMFLOW TIMING

• There are 417 streamflow gauges in the Lena basin
  of which only 28 had full-length, uninterrupted
  records for 1958 1999.
• The 28 gauges in the Lena were analyzed over the
  same 1958 1999 period of record. The 28
  stations were selected to be relatively
  uninfluenced by land-use changes and
  anthropogenic factors including effects from
  the Vilyui reservoir.
• Since only monthly streamflow data were
  available, the Variable Infiltration Capacity
  (VIC) hydrologic model was used to generate daily
  streamflow through use of the models ration of
  daily to monthly streamflow, which was applied to
  the observed monthly streamflows to produce
  reconstructed daily discharge time series for
  each gauge, which are constrained to the sum of
  monthly observed totals.
• Forcings for VIC Temperature - Wilmott
  Matsura, 2005 Precipitation - Adam et al., 2007
  Groisman, 2005
• Daily streamflow was analyzed for three measures
  outlined in Stewart et al., 2005
• Spring Pulse Onset The first day of spring or
  snowmelt season
• Centroid of Timing (CT) Center mass of annual
  flow
• Fractional flows Ratio of monthly or seasonal
  flow to the annual flow
• Each measure was tested for linear trends and
  significance using Seasonal Mann Kendall test.
• Correlation analysis between snow cover and
  spring pulse onset and snowmelt was conducted
  using the Log-Pearson method and Kendalls Tau
  method.
• Correlation analyses between different months and
  interseasonal flow was also conducted.

(b)
(a)

• Spring Pulse Onset (Fig a)
• 0 16 day shifts towards earlier dates in 61 of
  stations
• Trends occur largely towards eastern part of
  basin
• Due to interannual variability, no trends are
  significant at p 0.10
• Stations with increasing trend are linked largely
  to increased snow cover (see Section 4).

(b)
(a)
7

• December
• January
• February
• Seasonal Fractional Flow during Winter (Dec,
  Jan, Feb)

(c)
(d)

• Centroid of Timing (Fig b)
• Equal shifts in earlier and later trends, no
  coherent trends
• CT is significantly correlated with spring pulse
  onset for 36 of gauges
• A shift of CT later in the year may indicate that
  the melt season is being prolonged while the
  maximum peak discharge is decreasing.

The combined seasonal fractional flow for the
winter season is taken as December, January and
February. The trends show an increase in 75

• Summary
• There is modest evidence that snowmelt is
  shifting earlier in the water year (about 60 of
  the gauges)
• Winter season fractional flow is increasing in 90
  of the gauges in the basin
• Summer season flow is decreasing at most stations
• The most significant changes occur during the
  months of May, July and all through winter
• Snow cover disappearance influences spring pulse
  timing, but the linkage is not as strong as might
  be expected.

of the gauges for all of the winter months and
the winter seasonal flow as a whole. Trends are
statistically significant at roughly half of the
gauges. The correlation between spring pulse
onset and winter discharge is significant at 30
of the gauges. A recent study of the Arctic and
particularly the Lena basin have shown that
winter temperatures are increasing (Yang et al,
2002). Furthermore, the cold and threshold
basins generally have positive correlation
coefficients, indicating that as temperature
increases, streamflow also increases, possibly
due to the melting of ground ice (Adam et
al.,2007).

• Adam, J.C., and D.P. Lettenmaier, 2008
  Application of new precipitation and
  reconstructed streamflow products to streamflow
  trend attribution in Northern Eurasia J. Climate
  21(8) 1807-1828
• Stewart I., D. R. Cayan, and M. Dettinger, 2005
  Changes toward earlier streamflow timing across
  western North America. J. Climate, 18, 11361155.