Prediction of seasonal runoff in ungauged basins - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 104

locations and severity of events such as floods, droughts
and low flow conditions. Inferences drawn from the flow
regime curves are particularly valuable if these curves are
highly reproducible from year to year, especially where
there is a strong seasonality in the runoff. These charac-
teristics generally indicate locations with simple hydro-
logical responses that have a clear connection to dominant
climatic drivers.
There are several ways of characterising the flow
regime (both quantitative and qualitative), as a way to
assist in the regionalisation of seasonal runoff in
ungauged basins. Quantitatively, it is common to express
the flow regime in a non-dimensional way as the
monthly sequence of

foundation for its regionalisation through similarity
indices, and to provide an overview and comparative
performance assessment of different methods used to
predict it in ungauged basins.

6.2 Seasonal runoff: processes and similarity
Examples of seasonal runoff and the flow regime for two
very different catchments in Austria are shown in Figure
6.3 . The photographs show the contrasting landforms and
vegetation of the catchments. The catchment in the top row
of Figure 6.3 , the Lafnitz River, is a mid-elevation catch-
ment in the eastern part of Austria, characterised by mod-
erate annual precipitation and subject to human regulation.
The flow regime of the Lafnitz exhibits little seasonality,
and high relative variability (from year to year). The catch-
ment in the bottom row of Figure 6.3 , the Lech River, is a
small, pristine Alpine catchment located at a higher eleva-
tion and receiving higher mean annual precipitation includ-
ing significant snow. The seasonality in the flow regime is
in this case very pronounced and, in comparison to the
Lafnitz River, there is relatively less variability from year
to year. It is of interest to understand why the seasonality is
so strong in one river but not in the other and to explore
why runoff in the catchment that has high seasonality is
more consistent between years than in the catchment that
shows little seasonality. Understanding the causes of these
differences is essential if one wants to select methods for
predicting flow regimes in ungauged basins and interpret
the results of these methods.
This section reviews the dominant processes responsible
for the shapes of flow regimes, and identifies similarity
indices that can be used to relate known flow regimes to
those of ungauged catchments. As in the case of annual
runoff ( Chapter 5 ), regionalisation and extrapolation of
seasonal flow regimes relies heavily on grouping methods.
Specific grouping techniques relevant to the flow regime
curve are also reviewed.

the Pardé coefficients (Pardé,

1933 ):

Q i
Q A

PK i ¼ ð 6 : 1 Þ

where PK i is the Pardé coefficient of month i ( ), Q i is

the mean monthly runoff (ensemble averaged over a
number of years) in month i (m 3 /s), and Q A is the mean
annual runoff (averaged over the same years) (m 3 /s). The
Pardé coefficient allows for a quantitative inter-
comparison of catchments with different absolute mag-
nitudes of runoff. A more qualitative way to characterise
seasonal flow regime, however, is through the use of
'

-

, which is usually based on a description
of the long-term seasonality of runoff. Regime types
maybedefinedonthebasisoftheseasonsoftheyear
in which runoff maxima or minima occur, the climatic
and catchment characteristics, or the causal processes
driving the seasonal runoff. For example, the regime
typology identified in the Hydrological Atlas of Switzer-
land (Weingartner and Aschwanden, 1992 ) describes
three main regime types: glacial (ice-fed), nival (snow-
fed) and pluvial (rain-fed) regimes. Both the quantitative
(e.g., Pardé coefficients) and qualitative (e.g., regime
type) approaches are widely used in the regionalisation
of seasonal runoff behaviour and will be reviewed later
on in this chapter.
The seasonal flow regime builds a bridge between
rapid variations in runoff, for example, captured by the
flow duration curve ( Chapter 7 ), while also directly
impacting variability on long time scales, e.g., the annual
and inter-annual ( Chapter 5 ). Seasonality and variation in
storage impact the occurrence of low flows ( Chapter 8 )
and determine antecedent conditions, thus impacting the
flood frequency curve ( Chapter 9 ). The flow regime
therefore informs runoff predictions over the full range
of time scales addressed in this topic ( Chapters 5

regime type '

6.2.1 Processes
The seasonal flow regime is also a basic hydrological
fingerprint of a catchment reflecting the interplay of
climate, geology, land use, vegetation cover and human
modification. As in the case of annual runoff ( Chapter 5 ),
seasonal runoff variations are driven by the relative sea-
sonality in precipitation and potential evaporation. At
seasonal time scales, however, catchment storage pro-
cesses often play a relatively more important role in
driving runoff fluctuations than they do at annual time
scales. Total continental water storage consists of water
on vegetation surfaces, in the biomass in the unsaturated

10).
The aim of this chapter is to review the processes that
control

-

seasonal runoff variability, to develop the