PUB in practice: case studies - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 300

Figure 11.31. Hydrographs of simulated daily flow (red lines) and observed daily flow (blue lines). Model prediction confidence limits are
shown in black dashed lines.

for Model (4) for both the 1976 - 85 and 1986 - 94 periods runoff in the identification of an appropriate hydrological

model for flow prediction in ungauged basins.
The optimal version of the MAC-HBV model combined
with the dual regionalisation method (IDW-PS) is now used
across Canada and in other countries for improved runoff and
baseflow estimation in ungauged basins. The model per-
forms well across heterogeneous landscapes, climatic condi-
tions, and in basins varying significantly in size. Significant
improvements (up to 50%) have been achieved for baseflow
estimates, which is particularly important for reliable envir-
onmental flow prescriptions to mitigate the impact of flow
regime alteration on river ecosystems, especially in the con-
text of increasing hydropower development in Ontario.

(Figures 11.29a and 11.29b, respectively).
Examples of simulated runoff and baseflows in ungauged
basins in different regions using Model (4) are shown in
Figure 11.31 . The model captures quite well both daily
runoff and baseflow variability. Basin 04FC001 shows
slower recession limbs (Figures 11.31b and 11.31d), which
is a characteristic of basins in northern regions due to flatter
topography and the presence of small lakes. Basin 02FC002
with steeper recession limbs (Figures 11.31a and 11.31c)
represents typical basins in southern regions where the
topography is steeper and small lakes are generally absent.

Discussion
The proposed combinedmodelling tool has been applied and
tested across the Province of Ontario for estimating continu-
ous runoff at ungauged basins with great success. In addition
to the results presented in this case study paper, more
detailed results can be found in Samuel et al.( 2011b ).
It appears that the good performance of the optimal
hydrological model identified for the Ontario basins can
be attributed to the fact that site-specific hydrological
attributes and hydrological processes have been captured
by the model structure. The non-linear storage discharge
relationship and the wide variability in model parameters
control the delay and rate of outflow by allowing deeper
soil to retain and release water through a non-linear func-
tion, especially during drier periods. Study results also
show the importance of optimising both baseflow and

11.9 ESTIMATING FLOW DURATION
CURVES FOR HYDROPOWER
DEVELOPMENT IN CENTRAL ITALY
a. castellarin

The issue from societal and hydrological perspectives
This case study reports on an extensive analysis performed
over a wide and hydrologically complex area in central Italy
(see also Castellarin et al., 2004a and 2007a ) . The analysis
was promoted by the R&D division of Enel S.p.A. (l

' Ente

Nazionale per l Energia Elettrica), an Italian energy provider,

which at that time needed to develop a parsimonious yet as
reliable as possible regional model for predicting long-term
flow duration curves (FDCs) in ungauged sites located

'