PUB in practice: case studies - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 323

15 ° E 20 ° E 25 ° E 30 ° E 35 ° E 40 ° E

5 ° S

Catchment boundary

Rivers

CONGO

Water bodies

Survey routes

Places

10 ° S

Kalene Hills

ANGOLA

Mutinondo

Mfuwe

Kabompo

Lukul u

Kabwe

Bridge East Rd.

ZAMBIA

15 ° S

Lusaka

Sioma Falls

Victoria Falls

NAMIBIA BOTSWANA ZIMBABWE

20 ° S

Figure 11.57. Luangwa Basin, located in Southern Africa. The study area is plotted in red. The large water bodies and larger wetlands
(called ) are indicated in blue. The tarred road network is also plotted to emphasise the remoteness of the area. All smaller untarred

roads inside the basin are only accessible in the dry season. At the bridge on Great East Road, only an old time series of daily discharge data
is available.

' dambos '

This case study focuses on a real-world example of
an ungauged basin. It demonstrates and applies a frame-
work to integrate both hard and soft unconventional infor-
mation in the form of signatures, delivered by available
hydrological observations of poor quality, as mentioned
above. The framework makes use of generalised likeli-
hood uncertainty estimation, or GLUE (Beven and Bin-
ley, 1992 ), within the limits of acceptability approach
(Beven, 2006 ) and has been presented earlier by Winse-
mius et al.( 2009 ).

estimation of river flows at the daily time scale (see Chapter
10 ). This would be particularly useful for flood prediction
and operation of the downstream reservoir, Lake Cahora
Bassa, which is visible just below the Great East Road
Bridge in Figure 11.57 . The choice of a real-world example
forced us to use unconventional methods to perform model
inference.
The essence of this case study is that, rather than ques-
tioning whether the setup of such a model is possible and
can be validated, we focus on whatever few data are
available and extract the most out of these data. The crux
is to identify small pieces of information from the few data
available and combine these pieces of information into a
scientifically sound model inference framework. The focus
in this case study is on signatures from available discharge
data and the decay of evaporation over

Description of the study area
The Luangwa River basin is a real-world PUB case (see
Figure 11.57 ). The aim of the study was to calibrate a
rainfall - runoff model that should be applicable for time. The