Prediction of annual runoff in ungauged basins - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 70

5 Prediction of annual runoff
in ungauged basins

Contributors: T. A. McMahon,* G. Laaha, J. Parajka,
M. C. Peel, H. H. G. Savenije, M. Sivapalan, J. Szolgay,
S. E. Thompson, A. Viglione, R. A. Woods and D. Yang

5.1 How much water do we have?
Human civilisation depends upon a reliable water supply.
One critical role of practical hydrology in delivering this
supply is estimating the reliability of water resources avail-
able for meeting human and environmental needs. Two
key elements of water resources planning are the long-
term mean rate at which river runoff is generated, and its
variability from year to year. For example, water supply
reservoirs are designed to smooth out fluctuations in
inflows and provide a reliable yield of water to sustain
human needs. Successful reservoir design must therefore
account for the mean and inter-annual variability of river
inflows. Inter-annual variability also provides a way to
quantify its sensitivity to variations in driving factors such
as climate. Fundamental understanding of the nature and
causes of variability of annual runoff is critical to assessing
how the reliability of water supplies will change when the
drivers of variability might change in the future, e.g.,
climate change, land use and land cover changes. Such
understanding is needed globally to improve water
availability and livelihoods for large human populations
worldwide, as well as protecting the natural environment.
Equally important, achieving an understanding of the aver-
age water flows and their variability on a continental scale
is an exciting aspect of earth system sciences per se, as the
water flows are intimately connected to many processes in
the oceans, in the atmosphere, on the land surface and in
the shallow subsurface.
This chapter focuses on prediction of annual runoff in
ungauged catchments. We define annual runoff as the total
volume of water discharging past a point of interest in a
river or stream in one year divided by the contributing
catchment area. Using this definition, the units of runoff
are usually mm/yr. If the total volume of water discharging
past a point is the variable of interest, volume units of m 3
year -1 or millions of m 3 year -1 and terms such as annual
runoff volume are adopted.

Mean annual runoff is the average of annual runoff
values estimated over many years. Its inter-annual variabil-
ity is usually quantified in terms of the standard deviation
(or coefficient of variation) of annual runoff. It can also be
expressed in terms of the growth-curve (i.e., cumulative
frequency distribution scaled by the long-term mean, see
Chapter 9 for examples in the context of floods) of the
annual runoff. Although commonly treated as constants
(i.e., stationary in the statistical sense) both the mean and
inter-annual variability of runoff may change over time as
a result of long-term (natural) changes in climate, catch-
ment characteristics or anthropogenic factors. For example,
Kuczera ( 1987 ) and Vertessy et al.( 2001 ) describe inter-
decadal to century-scale changes in runoff due to non-
stationary water use of Mountain Ash (E. regnans) forest
during regrowth following disturbance by fire.
Annual runoff is used in the preliminary design of water
supply systems (McMahon and Adeloye, 2005 ) involved in
allocating water for the environment, irrigation, industry,
human consumption, hydropower, navigation, recreation
and catchment management. Techniques for preliminary
analysis associated with the sizing of water supply systems
or estimating the annual yield from an existing system are
available and were reviewed by McMahon et al.( 2007a )
using a global database of annual (and monthly) runoff.
Estimates of mean annual runoff, its variability and auto-
correlation are needed for many of these techniques. Annual
runoff is also used to assess climate change impacts on
water resources (Arnell, 1999 ; Milly et al., 2005 ) and land
use change impacts on catchment yield (Bren et al., 2006 ;
Komatsu et al., 2011 ). Other uses of annual runoff include
analyses with respect to the global water crisis, water and
sustainability, global food production, and understanding of
the global water cycle (e.g., Vörösmarty et al., 2010 ).
Annual runoff and its inter-annual variability are import-
ant diagnostics of the surface water balance of a landscape,
especially at large spatial scales. Annual runoff variability
is one of several signatures of runoff variability (Sivapalan,
2005 , Wagener et al., 2007 ), the others being the
regime curve (see Chapter 6 ), the flow duration curve