Jökull - 01.01.2013, Page 72

P. Crochet
Glaciers cover almost 11% of the area of Ice-
land. More than 75% of the electricity production
comes from hydropower and the largest hydroelec-
tric power stations are fed by glacial rivers (e.g. Thor-
steinsson and Björnsson, 2011). Projected increase
in surface air temperature in coming decades is ex-
pected to deplete snow storage, reduce the volume
of Icelandic ice-caps and have an impact on stream-
flow volumes and seasonality in glacial and non-
glacial rivers (Jóhannesson 1997; Aðalgeirsdóttir et
al., 2006; Jóhannesson et al., 2007; Björnsson and
Pálsson, 2008; Jónsdóttir, 2008; Guðmundsson et al.,
2009; Einarsson and Jónsson, 2010; Thorsteinsson
and Björnsson, 2011). If realised, this development
would have a profound effect on water resources and
major implications for the design, management and
operation of reservoirs used to store water for hy-
dropower production and other uses.
Large uncertainties remain about future changes
of climate and their effects on hydrology. An exam-
ination of the hydrological response to observed cli-
mate variations improves our understanding of the re-
sponse of hydrological systems to climate forcing and
thereby improves our ability to assess the hydrolog-
ical impact of future climate changes. Here, the hy-
drological response of different types of catchments
in Iceland to past variations in temperature and pre-
cipitation is analysed considering a large collection of
variables describing various aspects of the hydrologi-
cal cycle.
DATA AND METHODS
Data
Eight watersheds, for which daily flow discharge mea-
surements are available for at least thirty years, were
selected for this study. They are relatively unaltered
by human influences such as regulation, drainage and
land-use (Figure 1 and Table 1). Rivers in Iceland are
frequently classified according to their source (Rist,
1990; Jónsdóttir et al., 2008; Jónsdóttir and Uvo,
2009): direct runoff (D), groundwater fed (L), glacial
(J) and whether they flow through lakes (S). Direct
runoff rivers are characterized by rather constant low
winter flow, when precipitation is stored in the snow-
pack, high flow in spring with snowmelt, low summer
flow and a secondary peak in autumn due to rainfall.
Glacial rivers display a similar hydrograph in winter
but high discharge is observed from spring to autumn
due to both snow melt in spring and glacier melt in
summer. Groundwater fed rivers are characterized by
fairly constant discharge during the year. The studied
river catchments are of mixed origin ( Table 1), three
of them are non-glacierized and five partly glacier-
ized, two of which receive a large groundwater contri-
bution to their flow (VHM-64 and VHM-66). Figure
2 presents the annual streamflow hydrographs, nor-
malized by the annual mean daily flow, for the period
1971–2000.
In order to describe and quantify climatic varia-
tions in the past decades and their effect on hydrology,
1-km gridded daily temperature (Crochet and Jóhann-
esson, 2011) and daily precipitation series (Crochet et
al., 2007; Jóhannesson et al., 2007) were used. The
temperature data set was obtained by gridding tem-
perature observations at meteorological stations with
a spline interpolation after elevation correction, us-
ing a fixed lapse rate of 6.5 ◦C/km. The precipitation
data set was obtained by dynamical downscaling of
ERA-40 precipitation (Uppala et al., 2005) with an
orographic precipitation model (Smith and Barstad,
2004) whose parameters were optimized to best sim-
ulate precipitation observations at meteorological sta-
tions and winter balance measurements on three large
glaciers.
Hydrological variables
A total of 22 variables were defined to analyse various
aspects of the hydrological response to climatic varia-
tions (Table 2). A hydrological year from Sept. 1st to
Aug. 31st was used in all cases. The role of snow stor-
age and snow and glacier meltwater runoff on stream-
flow variations was evaluated from the gridded daily
precipitation and temperature series. Daily precipi-
tation was split into rain and snow water equivalent
(SWE) with a fixed temperature threshold of 1 ◦C,
commonly used in glaciological studies in Iceland
(Jóhannesson et al., 1995; Jóhannesson, 1997; Aðal-
geirsdóttir et al., 2006; Guðmundsson et al., 2009).
Snowpack (accumulated SWE) and meltwater runoff
were estimated daily at each grid point, using a simple
72 JÖKULL No. 63, 2013