Prediction of low flows in ungauged basins - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 164

below a runoff threshold. The duration or frequency of
dry spells (or no flow) may be a particularly useful index
in arid regions.
The third group of characteristics represents how fast the
low flow runoff decreases with time and includes the base-
flow index, which represents the proportion of runoff that
originates from deep storage in the catchment (Institute of
Hydrology: IH, 1980 ). Another characteristic is the reces-
sion parameter, which measures the rate of decay of the
runoff hydrograph during dry periods and is related to the
outflow of stored water from within the catchment (Tal-
laksen, 1995 ; Eng and Milly, 2007 ; Gustard and Demuth,
2009 ).
Low flow characteristics are best estimated from long
runoff records at the site of interest, and a minimum record
length of about 20

detail and to understand what makes two catchments simi-
lar in terms of the driving processes and the low flow
variability. Figure 8.1 presents runoff hydrographs of two
catchments in the UK along with the photographs of the
streams. The low flow behaviours of the two catchments
are completely different. The South Tyne at Featherstone
( Figure 8.1 top) is a quickly responding (flashy) stream,
which can be inferred also by the evident bank erosion in
the picture, while the Kennet at Theale ( Figure 8.1 bottom)
has more damped dynamics, reflected in the gentler land-
scape. As a consequence, the South Tyne experiences more
frequent low flow periods than the Kennet, and low flow
runoff, relative to the mean runoff, is also much lower.
From a hydrological point of view it is of interest to
understand why the low flow behaviour of these two
streams is different (i.e., why is the South Tyne so much
flashier than the Kennet), to identify the causal processes,
and explore how similarity and dissimilarity between
catchments can be defined in terms of low flow processes.

30 years is usually recommended (e.g.,
DVWK, 1983 ; Hisdal et al., 2004 ). In ungauged basins
where such data are unavailable, low flow characteristics
need to be estimated from regional information and from
auxiliary, local data. This chapter deals with the prediction
of low flows in ungauged basins with a main focus on low
flow indices (i.e., Q x , MAM d , Q d,T ) as they are the most
important runoff signatures for low flows from a practical
perspective. Also, these low flow indices are a subset of the
full spectrum of variability embedded in the complete run-
off hydrograph and are therefore a key signature of natural
runoff variability, which is interesting to understand per se
from a process perspective.
Low flows are runoff extremes and therefore share some
common features with floods ( Chapter 9 ), although they
are at the other end of the spectrum of variability. Both
may be more variable than other runoff signatures as they
are extremes, and both may be difficult to estimate from
short records. However, floods can tend to be very local-
ised, in particular flash floods caused by convective pre-
cipitation (Borga et al., 2010 ), while low flows caused by
droughts may extend over much larger spatial and tem-
poral scales.

-

8.2.1 Processes
What are the processes causing low flows? Precipitation
that falls on a catchment undergoes several transform-
ations: it is partitioned at the surface into overland flow
and infiltration, and the infiltrated water is stored in the soil
and the aquifer below and eventually released or drained as
streamflow. River flow falls because release or drainage
stops or becomes low for prolonged periods of time. There
are therefore two main component processes driving low
flows, those related to precipitation and other climate vari-
ables, and those related to catchment processes in the soil
and aquifers.

Climate
Climate controls the water fluxes at the land surface of
catchments through the variation of precipitation and evap-
oration over the year, which in turn determine the fluxes of
water into the streams. With respect to generating pro-
cesses and seasonal occurrence, two types of low flows
can be distinguished ( Figure 8.2 ). The first type occurs as a
consequence of persistent dry and warm weather periods,
when evaporation exceeds precipitation. This leads to
depletion of subsurface storages and causes runoff reces-
sion. This is the case in most arid places, or even in humid
places under summer, dry conditions, when there is very
little precipitation and the evaporation rates are very high.
These are often called summer low flows ( Figure 8.2a ).
The second type of low flows occurs in cold regions,
where low flows are caused by the freezing of water. This
causes precipitation to be temporarily stored in the snow or
ice cover, again causing runoff recession. Subsurface path-
ways may also be frozen,

8.2 Low flows: processes and similarity
Low flows are the result of a complex combination of
processes related to the climate and the catchment. They
are a manifestation of the interactions between climate
inputs during particular critical periods of the year and
complex flow pathways that result from the heterogeneous
patterns of geology, soils and vegetation cover in the
landscape. Understanding the subsurface flow paths
( Chapter 4 ), in addition to the climatic conditions, is there-
fore particularly important for predicting low flows in
ungauged basins. In order to predict low flows in ungauged
basins one needs to understand these processes in some thereby preventing water