Geography Reference
In-Depth Information
3000
Schärding
4000
3000
Hofkirchen
3000
2500
2000
2000
1500
1000
1000
0
Jun
Jul
Aug
Sep
Oct
Nov
Dec
500
12000
Kienstock
0
10000
Jun
Jul
Aug
Sep
Oct
Nov
Dec
8000
6000
4000
2000
0
Jul
Aug
Sep
Oct
Nov
Dec
Jun
40
Galtür
Golling
30
1200
1000
800
600
400
200
0
50 km
20
Brixlegg
10
800
0
600
Jun
Jul
Aug
Sep
Oct
Nov
Dec
400
Jun
Jul
Aug
Sep
Oct
Nov
Dec
200
0
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Figure 10.1. Runoff hydrographs observed at different points of the river network in Austria and Bavaria as used for the Danube runoff
forecasting system. From bottom-left clockwise: Galtür (98 km²), Hofkirchen (47 600 km²), Schärding (25 660 km²), Kienstock (95 970 km²),
Golling (3556 km²) and Brixlegg (8504 km²). Period shown is from June to December 2002. After Nester et al.( 2011 ).
need for distributed predictions at every point of the river
network, which Beven ( 2007 ) dubbed as
10.2 Runoff dynamics: processes and similarity
Figure 10.2 shows the hydrographs for two catchments of
very different size located in the USA and Austria. The
Vermilion catchment in Illinois (top row of Figure 10.2 )
is a very flat prairie catchment of 3341 km 2 almost
completely exploited for agriculture (see the top-left pic-
ture). The Gurk catchment, however, is a mountainous
catchment of 230 km 2 located in southern Austria. The
hydrograph of the Vermilion river is very flashy, i.e., the
catchment responds almost immediately to rain with
apparently no storage, while the Gurk river has a much
more dampened response to precipitation. At first this can
seem counterintuitive, since the more-than-ten-times
larger catchment is much flashier than the small one,
while one would expect higher response times for larger
catchments. From a hydrological point of view it is there-
fore of interest to understand why the Vermilion is flash-
ier than the Gurk, even though it is much larger.
Understanding the causal processes responsible for the
hydrograph shape is essential to define similarity and
dissimilarity between catchments.
'
predictions
everywhere
. The runoff predictions performed for the
Danube ( Figure 10.1 ) are an example of this, i.e., the
prediction of runoff at all points of the river network and
at all points in time.
This chapter is an extension of the material presented in
Chapters 5 to 9, in that the signatures covered in those
chapters are subsets of the full range of temporal variability
embedded in complete hydrographs. The understanding of
these signatures is essential for understanding the complete
runoff hydrograph (and vice versa, e.g., Claps and Laio,
200 3) and can influence the choice of methods for its
prediction. Indeed, the link between the various signatures
of variability and hydrograph predictions works in both
directions. On the one hand, the quality of hydrograph
predictions can be assessed by the ability to reproduce
each of the signatures, in order to ensure that the predic-
tions are right for the right reasons (Kleme
'
, 1986a ; Jothi-
tyangkoon et al., 2001 ). On the other hand, complete
hydrograph predictions remain one of the ways to predict
any of the signatures individually.
š
Search WWH ::




Custom Search