Prediction of seasonal runoff in ungauged basins - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 110

Figure 6.8. Runoff metrics used by
Czikowsky and Fitzjarrald ( 2004 )to
differentiate runoff behaviour before
and after vegetation activity increases
in spring. (a) The 30-day moving
average difference between
precipitation and runoff, P − R; (b)
more rapid recession dynamics due
to increased evaporation; (c)
increases in the amplitude of diurnal
runoff oscillations; (d) hydrological
predictions of spring date (based on
each of the three runoff metrics)
compared to standard meteorological
(Bowen ratio) estimates of spring
date from Fitzjarrald et al. (2001)
show a delay of 20

a)

Dormant
season

Growing
season

d)

200

Spring date
estimate

fit
P - R
Amplitude
Recession

Snow-
melt

180

x

0 365

b)

160

Before leaf
emergence

140

x

120

x

After leaf
emergence

x

100

t

c)

25 days that can
be attributed to the role of vegetation
phenology.

-

After leaf
emergence

80

60

80 100 120 140 160

Spring Date Estimate, Fitzjarrald 2001

Before leaf
emergence

t

snow, ice or soils in buffering these variations (Bower and
Hannah, 2002 ; Bower et al., 2004 ; Hannah et al., 2005 ).
The temporal stability of flow regimes has not been as
widely studied as the variation in flow regime between
catchments (but see Krasovskaia, 1997 ; Krasovskaia
et al., 2003 ; Bower et al., 2004 ; Hannah et al., 2005 ; Monk
et al., 2006 , 2008 ). The stability of flow regimes relates
directly to the reliability of seasonal patterns of runoff.
There is potential to directly include year-to-year variabil-
ity within classification schemes for regionalisation.
Figure 6.9 shows the inter-annual variability of the sea-
sonal runoff in four rivers in Switzerland from 1993 to 2006.
In the Alpine catchment of the Rhône (left panel), the
change in the seasonal pattern from one year to the next is
minimal (highly stable) due to the dominant influence of
accumulation and melting of snow and ice. The Pardé coef-
ficients (Equation 6.1) derived from the whole time series
are representative for almost all individual years. Seasonal
runoff estimates for catchments outside the Alpine zone are
much less stable, as the regular pattern of water availability
imposed by snow and ice diminishes, and the influence of
rain and evaporation variability increase.

usually associated with dams, dam construction and oper-
ation may be intended for hydroelectric power production,
water supply, flood mitigation, irrigation, navigation
(increase of low flows) or recreation ( Chapter 3 ). Hydro-
power production is particularly important in mountainous
areas and may significantly affect the flow regime (Zolezzi
et al., 2011 ). Figure 6.10 gives an example of the develop-
ment of reservoir capacity in the Rhône basin in Switzer-
land. During the 1960s, numerous reservoirs were
constructed. This led to a significant reduction in the
number of days with high flows ( > 400 m³/s) and low

flows ( 50 m³/s), reducing the amplitude and variability

of the flow regime. Weingartner ( 1999 ) and Holko et al.
( 2011 ) give further examples of

<

runoff modification

through reservoir operation.
Modifications to the flow regime may also occur due to
hydropower production, transfers of water from one catch-
ment to another and irrigation (Maheshwari et al., 1995 ;
Shao et al., 2003 ; Sauquet et al., 2008 ). Withdrawals for
irrigation frequently modify the natural flow regime in the
drier regions of the world. Global food production involves
a consumptive water use of 6800 km 3 /yr, of which
approximately 1800 km 3 /yr (57 000 m 3 /s) is supplied from
rivers and groundwater (Falkenmark and Rockström,
2005 ). Land use change and urbanisation also influence
the process of runoff generation, recharge of groundwater,
evaporation rates and the runoff pattern. Land use change
effects, typically, are most important in small catchments
because of their local nature (Blöschl et al., 2007 ).

Change (human impacts)
There are very few river systems worldwide that have not
experienced some change in the flow regime due to human
modification: for example, one study found that globally
172 out of 292 rivers investigated were regulated by dams
(Nilsson et al., 2005 ). While regime modification is