Geography Reference
In-Depth Information
Figure 11.58. The disaggregated
model units with their dominant land
cover characteristics (hydrotopes).
26
°
E
8
°
S
28
°
E
30
°
E
32
°
E
34
°
E
36
°
E
Riverine
Dambos
Forested
H
i
g
hlands
10
°
S
12
°
S
14
°
S
16
°
S
evaporation estimates were derived using the SEBAL
method (Bastiaanssen et al.,
1998
). Satellite rainfall esti-
mates (SRE) considered were product 3B42 of the Tropical
Rainfall Measuring Mission (
Huffman et al., 2007
) and the
CPC/Famine Early Warning System (FEWS) daily esti-
mates (Herman et al.,
1997
). A 16-year data set of dis-
charges was available at the outlet of the catchment,
collected largely in the 1970s. Some of the challenges
faced in this case study were:
much lower and, given the presence of wetland vegetation in
the
, these areas have a higher capability of retaining
moisture during the dry season than the lower savanna
regions. The average rainfall in the catchment is around
1000 mm/yr. Rainfall is concentrated in one wet season from
November until April. Based on this impression and a time
series of maps of normalised difference vegetation index
(NDVI), a subdivision into hydrotopes was made and a simple
hydrological model with a limited number of parameters was
set up.
Figure 11.58
shows the subdivision into hydrotopes.
The issue of validation material was also tackled during
field visits. To ensure that (as far as possible) an unbiased
and concomitant time series of rainfall and discharge was
available, measurement equipment was installed at Mfuwe
(location is indicated in
Figure 11.57
) to collect water
levels over one wet season. These water levels were rated
based on old rating data at the site. To bias correct SRE
over the period for which validation data were collected,
time series of local rainfall from lodges in the region were
also collected (see
Figure 11.57
for the locations).
'
dambos
'
typical globally available satellite rainfall estimates
(SRE) are non-concomitant with typically available cali-
bration series (see
Chapter 3
);
all information available is error prone: SRE are for
instance typically biased, satellite evaporation is noisy
and old discharge records may have gaps;
validation material is not present.
and to
recover some of the main features of the basin (see
Chapter 3
).
The Luangwa Basin is a relatively pristine and remote area of
about 150 000 km
2
, located in Zambia, Southern Africa. It
consists partly of flash-flood generating mountainous areas
and lower-lying floodplain and wetland areas with tropical
savanna vegetation and some agriculture. The north-eastern
boundary (the Muchinga escarpment) is densely forested and
interspersed by pristine wetland areas and large basalt lava
rocks. This area has a different hydroclimatology from the
low-lying savannas. Temperatures on the escarpment are
A field visit was conducted to
'
read the landscape
'
Method
Framework to combine different sources of information
into a calibrated model
Winsemius et al.(
2009
) developed a calibration frame-
work to profit from any available data in a catchment, in
this case signatures from the old discharge time series and
Search WWH ::
Custom Search