PUB in practice: case studies - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 272

estimates for improved performance. Takeuchi et al. apply a
grid-based distributed process model across the
entire Mekong River basin, involving six countries and 90
million inhabitants. Arheimer and Lindström apply a dis-
tributed process model for the entire country of Sweden and
adopt several strategies to improve prediction performance.

Yet it is refreshing to note that it is under these circum-
stances that we found the use of innovative and non-
standard approaches (e.g., India, Zambia, Ghana), albeit
in an ad hoc manner. This raises the question of how we
can encourage and formalise these approaches, to empower
local practitioners through targeted activities that generate
understanding and develop home-grown solutions to local
problems. These are the situations where one needs
hydrologists to make genuine efforts to read the landscape,
make inferences, aspire to come up with reasonable esti-
mates, and then find independent information to validate
them.

Insights gained from the comparative assessment
Interestingly, in at least 7 of the 19 studies the authors
report that their PUB work has been highly beneficial and
has been well received in the community (Archfield,
Samuel et al., Merz et al., Arheimer and Lindström, Cas-
tellarin, Winsemius and Savenije). This is a remarkable
testament to the societal relevance of PUB.
Interestingly also, it appears that in developed countries
(e.g., North America, Europe and Australia) the driver for
the application of PUB is governmental or regulatory, or an
industry. Examples include the EU Water Framework Dir-
ective (Arheimer and Lindström, Sweden) and the EU
European Flood Directive (Merz et al., Austria), the focus
on hydropower in Italy (Castellarin) and Canada (Samuel
et al.), and the national push to update the Australian
Rainfall and Runoff (Rahman et al.) and the Australian
National Water Audit (Viney). The regularity of govern-
mental push provides the motivation for much of the
progress that is being made.
In contrast, the case studies reported from developing
countries are being carried out either through individual
efforts locally or by external players supported by foreign
aid funds, in both cases on an ad hoc basis. There is no
mention of an organised push for improved predictions at a
regional or national scale. This has implications for PUB.
Lack of organisation or direction also means that the stud-
ies can tend to be local, uncoordinated, unfunded and
unappreciated. Consequently, there is much less scope for
accumulation of knowledge and experience: if this situ-
ation is true as inferred and is widespread, as we think it
could be, then it cannot be good for the practice of PUB.
Second, it appears that in developed countries, where
there is much more data available, standard approaches
(both statistical and process-based) are being used and
the focus on PUB is helping to generate further improve-
ments. Furthermore, in the humid countries amongst these,
there is much more widespread use and trust in process-
based methods as well. This is consistent with one of the
outcomes of the synthesis that with increasing humidity or
wetness performance of process methods tends to increase.
In contrast, in most developing countries data are very
scarce, and many of the countries happen to be located in
more arid parts of the world as well. Thus, there is not
much scope for either statistical or process-based methods,
unless significant investments are made to make advances
on both fronts.

11.2 HYDROLOGICAL INSIGHTS
FROM LONG-TERM RUNOFF
PATTERNS ACROSS KRISHNA
BASIN, INDIA
t. biggs

The issue from societal and hydrological perspectives
Peninsular India supports a large population whose liveli-
hoods depend on water. Irrigated agriculture, in particular,
is credited with increasing agricultural output and farmers

'
incomes. Irrigation depends on the availability of surface
and groundwater, both of which vary in space, seasonally,
and inter-annually. Despite the importance of water
resources for the economy and society in India, conceptual
and mathematical models on the climatic and land surface
controls on runoff are not well developed. Inter-state con-
flict over scarce water resources has resulted in limited
availability of stream discharge data, even among govern-
ment agencies, so relatively little data on stream discharge
is available. In response to the need for regional assess-
ments of water and scenarios of change under different
land cover and climate change, the International Water
Management Institute, funded by various international
organisations and in collaboration with universities in
India, Australia, the USA and Europe, has performed inte-
grated hydrological and economic analyses of the Krishna
Basin, one of India s largest basins with acute surface

water scarcity problems (Bouwer et al., 2006 ; Biggs
et al., 2007 ; Immerzeel and Droogers, 2008 ; Immerzeel
et al., 2008 ; Bouma et al., 2011 ). The hydrological models,
which are calibrated to existing rainfall and runoff data, are
useful for developing scenarios of water availability under
land use and climate change for a particular catchment, but
the understanding of the regional controls on runoff and
how they respond to change is limited. Due to legal and
political concerns of the three Indian states that share the
basin,

'

the models have only been applied to smaller