Prediction of floods in ungauged basins - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 197

Figure 9.7. Map of mean annual
flood runoff normalised to a
standard catchment area of 100
km² and mean annual precipitation
for the Kamp and Pulkau regions
in northern Austria. From Merz
and Blöschl ( 2008b ).

change was more pronounced in the lowland and hilly
regions than in the mountains.

reasons why mean annual precipitation is a useful similar-
ity measure for the flood frequency curve. First, mean
annual precipitation is usually highly correlated with event
precipitation as it is an aggregated measure of the actual
events occurring in the catchment. Second, mean annual
precipitation is an important control of soil moisture at the
seasonal scale, and soil moisture itself is an important
control of floods, in particular if saturation excess runoff
generation dominates, but is also important for other mech-
anisms (e.g., Zehe and Blöschl, 2004 ). Third, because of
the co-evolution of climate, vegetation, soils and land-
forms, there is often a close linkage between the rainfall
regime and the soils and landform, which themselves are
important controls of floods at the event scale (see Chapter
2 for a discussion of co-evolution). Mean annual precipita-
tion is therefore a similarity index that may capture the
results of catchment co-evolution relevant to floods.

Climate similarity
Climate similarity in the context of floods can be quanti-
fied by the similarity of extreme rainfall. Intensity

- dur-

ation frequency (IDF) curves represent the cumulative

distribution of extreme rainfall for a given aggregation
interval. The flood frequency curve is a non-linear trans-
formation of the IDF curve to capture the effects of runoff
generation (e.g., by choosing a runoff coefficient that cor-
responds to the return period of interest) and routing (e.g.,
by choosing a storm duration that resonates with catchment
mean response time). Climate similarity can also be quan-
tified by the seasonality of extreme rainfall and by atmos-
pheric circulation patterns (Petrow et al., 2007 , 2009 ;
Pa rajka et al., 2010a ) .
Many studies have shown that mean annual precipitation
is an excellent similarity measure for flood frequency (see
e.g., Madsen et al., 1997 ; Reed et al., 1999 ; Merz et al.,
Merz and Blöschl, 2008a , b , 2009b and references therein).
An example is shown in Figure 9.7 , where both the mean
annual flood and mean annual precipitation have a ten-
dency to decrease from west to east. The eastern part of the
region differs from the west not only in terms of precipita-
tion but also in terms of soils and the efficiency of the
drainage network. For example, the average event runoff
coefficients in the west are around 0.25 while they are less
than 0.1 in the east. There are therefore a number of

-

Catchment similarity
Catchment area is one of the most important similarity
measures for a number of reasons. At the basic level, flood
peaks will be larger for larger catchments, simply due to
the fact that the total rainfall volume is larger. Of course,
this will be moderated by the heterogeneity of rainfall and
runoff processes within the catchment. Increasing catch-
ment area leads to a reduction in the average amount of
precipitation the catchment receives. This is because rain-
storms are likely to affect only a portion of the large
catchments, while they may affect the complete area of a