Process realism: flow paths and storage - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 68

related to either soils or geology. This choice should really
depend on where the flow paths are in the catchment of
interest

beneficial to take extra measurements if possible, particu-
larly measurements of runoff. Spot gauging, or perhaps
even installing a stream gauge to obtain a short runoff
record in the catchment of interest may provide very valu-
able information on the runoff signature of interest. Short
runoff records may be exploited by many methods for
predicting runoff in ungauged basins (see Sections 5.3.4 ,
6.3.4 e tc.)

through the soils or deeper in the geology. Soil
characteristics as predictors may be misleading and lead to
spurious correlations if the flow paths are actually much
deeper. (ii) Similarly, it is useful to interpret the coeffi-
cients in the relationships between runoff signatures and
catchment characteristics (e.g., regression coefficients in
terms of the sign of the coefficients depending on simpli-
fied concepts of the flow paths). For example, one would
expect floods to be negatively correlated with soil depth if
the saturation excess mechanism is the dominant flood-
producing mechanism, and positively correlated with the
percentage of clay if infiltration excess is the dominant
mechanism. One would expect low flows to be positively
correlated with the permeability of the bedrock as this
suggests higher storage, although this may depend on the
particular flow system. Similar considerations apply to
other runoff signatures, in particular flow duration curves
and hydrographs.

-

4.5.4 Regional interpretation and similarity
Spatial interpretation of all this information may always be
helpful. This means that rather than directly inputting the
information on flow paths and storage into a process-based
or statistical model, the information is first plotted on a
map with real landscape features. This allows the infor-
mation to be related to the landscape processes. It may
assist in understanding how the landscape is organised
from a hydrological perspective and enable patterns to be
detected, and highlight how the catchment of interest fits
into the regional pattern. In the second step, the informa-
tion can be entered into a quantitative model, either
process-based or statistical. Regional visualisation of
indices of flow paths and storage may involve maps of
recession parameters and baseflow indices estimated from
runoff. This may perhaps be assisted by residence times
estimated from tracers and other regional proxy data such
as the presence of springs (see e.g., Chapter 8 ), against the
backdrop of the regional hydrogeology and the climate.
The type of maps one chooses to draw may, again, depend
on the runoff signature of interest. For example, maps may
consist of flood response times against the backdrop of
soils, geology and mean annual precipitation in the case of
floods; and recession parameters and mean transit times
from tracers (if available) against the backdrop of geology
in the case of low flows. The purpose of such a regional
interpretation for comparing catchments across the land-
scape is to assist in the regionalisation step based on the
hydrological similarity between catchments (see Sections
5.2.2 , 6.2.2 etc.). In this comparative approach, qualitative
information on similarity can be made useful for quantita-
tive estimates. For example, the ungauged catchment of
interest may contain massive debris deposits near the
stream that are not present in the neighbouring (gauged)
catchment. From this comparison one can infer that the
ungauged catchment may respond more slowly during
floods because some of the rainfall will percolate deeper
in the ground and have deeper flow paths than in the
neighbouring catchment, resulting in smaller flood peaks
(Merz and Blöschl, 2008a , b ). Sections 5.2.3 , 6.2.3 etc.
provide quantitative methods of how this similarity or
dissimilarity can be exploited for estimating runoff signa-
tures in ungauged basins. If runoff model parameters are

4.5.3 Role of field visits, reading the landscape, photos
and other proxy data
Some of the information on flow paths and storage to be
used in ungauged basins may be available from existing
databases, be they global, regional or local (see Chapter 3 ).
This may include existing reports on the hydrological
processes in the catchment of interest derived from previ-
ous studies. Other information needs to be collected during
the study. It is particularly important to perform field trips
that allow extra information to be collected relevant to the
signature of interest. For example, an important question
on whether surface runoff occurs (and at which event
magnitudes) can be addressed through collection of proxy
data during field trips, e.g., by erosion marks (see Section
3.7 ) . Additional information can be obtained by

' reading

the landscape , to get an understanding of how the process

of co-evolution of climate, landform, vegetation, soils and
geology has shaped the landscape. The dynamics of flow
paths and storage are often reflected in visible geomorphic
features in the landscape. Examples are the existence of
highly permeable debris fans and deposits from landslides
that, once known, can be represented in the model. For
example, reading the geomorphic features of the landscape
assisted Rogger et al.( 2012a ) to set the storage parameters
for a runoff model for predicting floods. It is always a good
idea to create photo-documentation of the catchment to
record landscape features in order to better understand
how the catchment works hydrologically. The proxy data
obtained can be used by many of the different methods for
predicting runoff in ungauged basins (see Sections 5.4.3 ,
6.4.3 etc.). Additionally, and importantly,

'

it is always