Geoscience Reference
In-Depth Information
Table 1.
Average annual water balance (1948-1970) for Lake
Albert.
Vol. (10
6
m
3
) l w(m
3
/
s)
Parameter
Total %
Rainfall on Lake
4
121
9
Inflow from
3
147
11
Semiliki, other
Inflow from
30
954
73
Kyoga Nile
Land catchment
3
86
7
contribution
Total input
40
1308
100
Evaporation from
8
263
20
the Lake
Outflow from
34
1070
82
Albert Nile
Storage
−
2
−
25
−
2
Total output
40
1308
100
Autogressive Moving Average ARMA (4) model
3
for missing data in a
period more than 12 months. The reasons for using the ARMA model were
firstly: because it gives the closest approximation to the regional series of
flows; secondly, the model incorporates the variation in flows that are man-
ifested as a result of the seasonality factor; and lastly previous success in
applying these models on catchments in Uganda.
3
2.2.
Frequency analysis
From the data, the annual maximum and annual minimum series for both
inflows and outflows were derived. The effective inflow of Lake Albert was
obtained as the sum of all major inflows for the respective time intervals.
The return periods for 2, 5, 10, 25, 50, 100, 150, and 200 years for the
maximum inflow and outflow, minimum inflow and outflow and average
inflow, were estimated by applying the method of moments. The year of
average flow was determined by selecting the year whose annual maximum
inflow was closest to the average of all the effective maximum inflows. It
was assumed that these data series followed an Extreme Value Type One
distribution given as
F
(
q
)=
e
−e
−
((
q
−
u
)
/α
)
,
(1)
where
q
is the flow whose probability of occurrence is being determined
α
=(
√
6
/π
)
σ
,and
σ
is the standard deviation and
µ
the mean of the data
set of being analyzed
u
=
µ
+
αγ
,where
γ
is a constant and
γ
is 0.5772.
