PUB in practice: case studies - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 297

evaluating water allocation permits, the ability to estimate
daily runoff at ungauged basins has led to additional
studies that have mapped water availability at approxi-
mately 1400 ungauged basins across Massachusetts. This
mapping provided a snapshot of water availability across
Massachusetts for a range of hydrological statistics
derived from the daily runoff estimates. Daily runoff
was also able to be estimated at ungauged basins that
coincide with fish-sampling locations to assess the rela-
tion between flow alteration and fish species and abun-
dance. Such information is providing insight into the
limits to which the ecology can tolerate changes in runoff
and, ultimately, will help to determine ecological flow
targets for rivers across the study area.

Description of study area
A total of 111 basins distributed across the Province of
Ontario with a total area of about 1 million km 2 and
complete runoff records between 1976 and 1994 were
selected for the study ( Figure 11.28 ). Basin areas range
between 100 and 100 000 km 2 . The selected basins cover
the range of observed basin attributes to meet model
regionalisation, calibration and validation requirements
and are representative of the diversity of basins across the
Province of Ontario (Samuel et al., 2011a ).
The climatology and landscape in the study area vary
along the north

-

south gradient. The annual mean precipi-

-

tation is ~ 800 1200 mm/yr in the south over an elevation

range of 300 600 mm/yr over an eleva-

tion range of 100 - 200 m in the northern region. In general,
northern basins consist of predominantly coniferous forest,
swamp, muskeg and small lakes, whereas basins in the
southern region are dominated by mixed forests (Atlas of
Canada, available at http://atlas.nrcan.gc.ca ) . The near-
surface geology of basins in the south is dominated by
gravel, sand and silt and less by rock compared to basins in
the north.
A total of 146 precipitation stations and 110 temperature
stations having less than 20% missing data for the 1960

- 500 m, and 400 -

~

11.8 CONTINUOUS SIMULATION OF
LOW FLOWS FOR HYDROPOWER
DEVELOPMENT IN ONTARIO,
CANADA
j. samuel, p. coulibaly and r. a. metcalfe

97
period were selected as base stations for the study ( Figure
11.28 ). Missing precipitation and air temperature data at
the base stations were spatially patched and interpolated
with data from all available climate stations using the
inverse square distance weighted (IDW) method. Daily
potential evaporation was estimated from daily tempera-
ture using a modified Thornthwaite equation (Samuel
et al., 2011a ).
Basin attributes used for the regionalisation include:
(i) latitude; (ii) longitude; (iii) mean percentage slope;
(iv) average elevation (m); (v) percentage of basin area
covered by rooting depth deeper than 150 cm; (vi) percent-
age of basin area covered by coniferous, deciduous and
mixed forest; and (vii) percentage of basin area covered by
glaciofluvial till (see Samuel et al., 2011a ). These attri-
butes were considered best for determining physical simi-
larity among basins and were selected using a cosine
pattern similarity procedure (Samuel et al., 2011a ).

-

The issue from societal and hydrological perspectives
Accurate baseflow estimation in ungauged basins is a
difficult task in many Canadian catchments because of
the large size of most basins, the high seasonal climatic
variability, and the heterogeneity of physiographic
conditions. The accuracy of baseflow estimates is par-
ticularly important for prescribing environmental flow
and potential ecosystem and economic trade-offs in
water development projects. Interest in hydropower
development has significantly increased in Northern
Ontario with the passing of the Ontario Green Energy
Act (2009 ). Most basins in remote northern regions of
Ontario are ungauged or poorly gauged. Estimating con-
tinuous flow series in these large ungauged basins
remains a challenging task for assessing aquatic ecosys-
tem effects of flow alteration as part of the approval
process for hydropower projects. This case study pre-
sents some of the results of the optimal combined region-
alisation and application of the rainfall runoff model

(MAC-HBV) identified for improved baseflow and run-
off estimation in ungauged basins across Ontario. Hence,
the method is particularly useful for estimating low flows
(see Section 8.4.2 and Chapter 10 ). A detailed model
description along with extended results is provided in
Samuel et al.( 2011b ). The proposed modelling tool is
now used by the Ontario Ministry of Natural Resources
and waterpower developers for continuous runoff simu-
lation in ungauged basins in Ontario.

-

Method
The combined modelling tool includes: (i) a physically
based rainfall runoff model (MAC-HBV); (ii) a dual

regionalisation method (i.e., combined inverse distance
weighted method (IDW) and physical similarity (PS)
approach, referred to as IDW-PS); along with (iii) a Monte
Carlo simulation approach (Samuel et al., 2011a ).
The MAC-HBV model (Samuel et al., 2011a ) is a vari-
ant of the original HBV model introduced by Bergström

-