Kindaruma and Kiambere Dams on Flood Frequency and Risk on River Tana, Kenya
Abstract
Floods are a pertinent problem whose efforts to solve should
consider tangible technical consideration based on fairly
accurate flood observations or forecasts.
The current study was done in the upper Tana basin which mainly
covers and with its headwaters in the south or on slopes of Mount
Kenya and the eastern parts of the Nyandarua Ranges. Despite its
high development potential for irrigated agriculture and
hydro power generation and tourism, the Tana system - the largest
river basin in Kenya, experiences extensiv0 flooding. Substantial
control efforts by means of damming have however been effected.
This study was set out to investigate the effect, if any, of the
Masinga, Kamburu, Gitaru, Kindaruma, and Kjambere dams on the
flood frequency and risk on the Tana river and the implications
of such effect on the future development of the floodplain below
the dams. The degree of success of such efforts could be of great
use in the stemming the flood problem elsewhere in the country.
Both river flow and rainfall data recorded in the basin were
analysed in this study. Annual flood max imo and mean annual flows
at four regular gauging stations (RGS) upstream of the dams and
two others downstream - including the control one - were used.
The control station, RGS 4G1 at Garissa, had a record period of
58 years (1933-90). The data was tested for homogeneity and
I
consistency using graphs and the t-statistic and then applied to
three approaches to Flood -Frequency and Ri.s k Analysis (FF & RA)
namely: Gumbel type I, 2-parameter Log-Normal and Log-Pearson
xi
type III, with a view of examining each ones adaptability to the
pre- and post- each of the damming floodflows on the Tana river.
Using the Mean Absolute Deviation Test (MADT), the 2 parameter
Log-Normal distribution was found to be the mare appropriate for
the purpose of flood data prediction, analysis and inferences in
this study. The Chi-Square statistic was utjlized to determine
the significance of the effect of each dam on the flood experiences
in the lower Tana basin.
Analysis of the generated data revealed that; , except for the
Kamburu/Gitaru, the dams have actually snpp ressed the various
pre-dam T-year floods with Kindaruma having the greatest effect
(~ 47%) followed by Masinga (~ 32 %). Kamburu/Gitaru was found
to have aggravated the flood problem cd ready buffered by
Kindaruma - by about 46%, bringing the net effect of Masinga to
a mere 1% with regard to the Kindaruma floods. By the completion
of Masinga the pre-dams floodflows of th(~ 'I'n na river had been
abated by approximately 48%. Though signif icant, Kindaruma' s
suppression of the floods was purely incid0nt~l, just as was the
Kamburu/Gitaru aggravation not totally unex pert ed because, except
Masinga, all the other dams considered were designed primarily
for hydro-electric power generation and th0rerore with secondary
consideration of flood control aspects.
From the study it was concluded that, should the predicted flood
abatement trend continue, then it is likely to appreciably influ-
I
ence the la~duse development patterns and policy downstream of
the reservoirs: one, development will be encouraged to very close
ranges of the river reach thereby exposing them to a lot of risk;
xii
the building of the dams, thus, has not completely eliminated the
dangers posed by the pre-dams floods. Two, the cost of flood
mitigation efforts will be much lower than in the pre-dams
period. Three, perennially swampy areas in the lower parts of the
floodplain - towards the delta, will easily be reclaimed and
communications networks improved but there will be an inevitable
ecological change to contend with.
Recommendations on the basis of the study, include:
building of minor floodwalls along the river stretches
franking urban areas like Garissa, Ho l.a , (;arsen etc.
identification and creation of "washland" areas upstream of
urban and agricultural areas like the northern outskirts of
Garissa town to hold local contribution of storm waters.
river channel improvement in the vic in i t.y of developments
already done
flood-zoning of the entire lower Tana floodplain to guide its
future development policy.
education of the floodplain popu:tace on the flood dangers they
are still ~xposed to.
employment of dam-based flood contra I ]n the other areas
similarly prone to flooding.
persuing further research in the current topic, using partial
flood series and examining other probability distributions not
considered.
Sponsorhip
The University of NairobiPublisher
University of Nairobi, Department of Science