HYDROLOGICAL ASSESSMENT FOR THE

PROPOSED DAM CONSTRUCTION AT MALENGE

VILLAGE, KAHAMA DISTRICT

Draft Final Report

17th June, 2022

Prepared for Irrigation Commission Shinyanga Region by;

Salimu I. Lyimo,
Senior Hydrogeologist,
Internal Drainage Basin Water Board,
P.O. Box 2030,
Phone: +255 7127 66066, +255 7570 83267
Email: salimu.lyimo@maji.go.tz
Arusha, Tanzania
 HYDROLOGICAL ASSESSMENT FOR THE
PROPOSED DAM CONSTRUCTION AT MALENGE
VILLAGE, KAHAMA DISTRICT
Table of Contents

1. Introduction ............................................................................................................. 4

1.1. Background ........................................................................................................ 4

1.2. Scope of the Assignment.................................................................................... 4

1.3. The Study Area .................................................................................................. 5

2. Data Collection and Analysis ................................................................................. 6

2.1. Data Collection ................................................................................................... 6

2.2. Data Analysis ..................................................................................................... 7

2.2.1. Rainfall Data ................................................................................................. 7

2.2.2. River Flow Data .......................................................................................... 11

2.2.3. Weather Data .............................................................................................. 11

2.2.4. Spatial Data ................................................................................................ 12

3. Rainfall – Runoff Modelling .................................................................................. 14

4. Results and Conclusion ....................................................................................... 16

4.1. Results ............................................................................................................. 16

4.2. Conclusions ...................................................................................................... 16

 1. Introduction
1.1. Background
PANGAEA MINERALS LIMITED is aiming at fulfilling its social-cooperate responsibilities
to the nearby community where it operates. In doing so in Tanzania, the company plans
to build a water storage dam. The dam is expected to be constructed in Malenge Village,
Malenge Ward, Kahama District, about 10 kilometers from Bulyanhulu Gold Mine which is
owned by the company. The dam is expected to store water to supplement end-of-season
water shortage and if possible to sustain dry season farming by farmers of the area.

In order to construct a suitable and a long-lasting dam in the area, it is very important to
understand the hydrology regime of the catchment area upstream of the dam site. In a
simple language, understanding of hydrology regime of a catchment means to determine
and/or calculate hydrological factors such as daily, monthly and annual maximum,
minimum and average rainfall, its patterns and frequencies, peak floods and flood
frequency, etc.

The hydrological regime elements above are very important inputs during designing and
sizing a dam and its associated hydraulic structures such as crest and spill-way. It is for
that matter that this hydrological study/assessment is being carried.

1.2. Scope of the Assignment

The scope of this assignment is limited to determining all hydrological factors required for
designing and sizing of the dam to be constructed at Malenge Village, Malenge Ward in
Kahama District using satisfactory and acceptable hydrological methods and models. It
includes collecting and computing aerial rainfall, enumerating Long Term Average (LTA)
monthly and annual rainfall, examining rainfall patterns and analyzing rainfall intensities
and frequencies of the dam catchment.

The scope also covers determining rainfall–runoff relations and determining annual
renewable flow volume, calculate Probable Maximum Flood (probable maximum inflow
and outflow flood) with their return probabilities (frequencies) as stipulated by Tanzania’s
Ministry of Water Dam Safety Guidelines.
 1.3. The Study Area
The area where the proposed dam is going to be built is located in Lulenge Village, in
Kahama village at coordinates 9567499/190309 (UTM Zone 36S), along Lulenge River
(the river flows during wet period only). As the matter of hydrology is concerned, the study
area as referred in this study mean the catchment area responsible for flows at the dam
site and the surrounding area where hydrological data recording stations for the study are
found (rainfall, weather and river gauging stations). The study area is shown on the map
on Figure 1.

Figure 1: Location Map of a Study Area

Land use in the study area is mainly agriculture and grazing. Agriculture season begins in
November (farm preparations) when short rain starts and ends in May/June after rain
season ends. After the harvest, the farms are then used for grazing. Small scale mining is
practiced in few areas. There are no significant forest land or protected land (such as game
reserve) in the study area.
 2. Data Collection and Analysis
2.1. Data Collection
Hydrological assessments require sufficient amount of data. Major data sets required are
hydro-meteorological, hydrometric, spatial data sets and land use information. In order to
identify suitable hydro-meteorological and hydrometric data stations for the study, the
catchment area of the dam was first delineated and mapped and then overlaid with hydro-
meteorological and hydrometric station location layer/s. Eleven (11) station were found to
be at the appropriate for the study (Figure 2).

The identified stations are owned by different institutions which includes Lake Tanganyika
Basin Water Board (LTBWB), Lake Victoria Basin Water Board (LVBWB), Internal
Drainage Basin Water Board (IDBWB) and Tanzania Meteorology Authority (TMA) (Table
1). Each organization was consulted for data acquisition.

Figure 2: Hydrometric and Hydro-meteorological Data Station Location Map

Online resources of acknowledgeable international organizations such as United Nation’s
Food and Agricultural Organization (FAO) and United State Geological Survey (USGS)
and few others were also consulted for spatial data sets such as soil data and land-use
maps. A high-resolution Digital Elevation Modal (12.5m resolution DEM) was downloaded
from https://search.asf.alaska (all data sources/institutions are listed on Table 1). On other
hand, field visit was done to familiarize with the area, identify key drainage channels and
take records of coordinates of all important features for the work.

Table 1: List of Data Source Institutions

Sn Institute Data
1 LTBWB, LVBWB and IDBWB River Flow
Daily Rainfall
2 TMA Hour Rainfall
4 AFS DEM
5 NASA Agro climatology Rainfall Intensity
6 USGS Landsat-(LULC)
7 FAO Soil data

2.2. Data Analysis

Data sets collected were scrutinized for errors, gaps, outliers and omission by using
appropriate methodology and/or tools such as spreadsheet (Microsoft Excel Spreadsheet)
and XLSTAT tool (for time-series data). Identified inaccuracies were rectified accordingly
(e.g. gap filling). Long data gaps were not rectified as doing that will lead to distortion of
data. Portion of timeline with long gap/s were removed from calculations.

2.2.1. Rainfall Data

Rainfall data were analyzed to (i.) determine aerial rainfall of the proposed dam catchment
and (ii.) calculate return probability of rainfall’s extreme events (rainfall frequencies
analysis).

i. Aerial Rainfall
Aerial rainfall is single most important input to the rainfall-runoff model that is used in this
kind of study (aerial rainfall is a point or specific location rainfall). The aerial rainfall of the
proposed Lulenge dam catchment was determined by interpolation from existing rainfall
stations using Kriging method. Kriging method was selected as it provides a measure of
the probable error associated with the estimates and it accounts geographical proximity
between stations. A total of 6 rainfall stations were used in Kriging method to come up with
aerial rainfall of the proposed Lulenge dam catchment.

The selected stations for Kriging method were those with more than 30 years of daily and
hourly intensities rainfall records in the same period (1971 to 2019). The resulting daily
and hourly aerial rainfall are attached as annex 1 and 2 respectively while monthly
averages are shown on Table 2 and Figure 3. Long Term Average (LTA) annual rainfall
was found to be 774mm.

NOTE: Long Term Average daily and hourly rainfall Intensity were calculated from days
with rainfall only.

Table 2. LTA (1971-2019) Monthly Aerial Rainfall of the Proposed Dam Catchment

September

November

December
February
January

October
August
Month
March

June
April

July
May

Rainfall (mm) 115 94 140 115 36 2 1 2 8 34 101 136

Figure 3: LTA (1971-2019) Monthly Aerial Rainfall of the Proposed Dam Catchment
Monthly Rainfall (mm)

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Month
 ii. Rainfall Frequency
Hourly rainfall intensity (rainfall depth in one (1) hour time; (mm/hour)) is fairly good
resolution and recommendable rainfall intensity for calculating “inflow floods” for dam
safety measures. This intensity resolution is backed by several guidelines including
Tanzania’s Dam Safety Guidelines (2020), made under Dam Safety Regulations GN 237
of 2013 and Florida Department of Transportation (FDOT, 2020).

The hourly aerial rainfall intensity data generated by Kriging method were analyzed to
determine their return probability (return probability or recurrence interval is an estimated
average time between such events to occur). The return probabilities of up to 40 years
were calculated directly from the data as they range for 40 years (1971 - 2019) while longer
periods (i.e. 50, 75 and 100years) were arrived at by plot of return time against the intensity
(Figure 4) which yielded the equation y=5.3831ln(x) + 15.022. The return time of 100years
is a recommended return time for Class D dams (very low risk dams) by Tanzania’s Dam
Safety Guidelines. The calculated return periods are presented on Table 3.

Table 3. Return Period

Return Period (Years)
Parameter
5 25 50 75 100
Hourly Rainfall Intensity (mm/hour) 23.7 32.3 36.1 38.3 39.8

Figure 4. Plot of Return Period and Hourly Rainfall Intensity

40
35
Hourly Rainfall Intensity

30
25
(mm/hr)

20 y = 5.3831ln(x) + 15.022
15 R² = 0.9176

10
5
0
0 5 10 15 20 25 30 35 40 45

Return Time (Years)

 iii. Design Flood / Probable Maximum Flood (PMF)
Design Flood or Probable Maximum Flood (PMF) corresponding to 100years return time
was calculated using rational method due to uncertainty of available flow data. Design
Flood is required for designing of dam’s spillway. Rational method predicts the peak runoff
using rainfall data and catchments characteristics such as vegetation, topography and
land use. Rational method has been used for over 150 years is widely acceptable method
of estimating runoff of small ungauged catchments as long as consistent units are
employed (Main Roads-Western Australia Publication, 2019). Rational method equation
is expressed as Q = C. I. A.
Where;
Q = Peak Flood of a catchment
I = Rainfall Intensity (of desired return period)
A = Catchment Area
C = Appropriate Runoff Coefficient.

Important part in using rational method is assigning appropriate runoff coefficient. Runoff
coefficient (C) is unit-less. Runoff coefficient is the representation of catchment
characteristics being vegetation, geology, geomorphology and land use. Runoff coefficient
of the proposed dam site (C = 0.43) was estimated with the aid of pre-developed table of
runoff coefficient presented on Figure 5.

Figure 5. Runoff Coefficient (C) Table

With all variables for the Rational Method generated, the computation of Design Flood was
carried as follows;

Design Discharge (m3/sec) = Rainfall Intensity (mm/hr.) x Area (km2) x Runoff Coefficient
= (39.8mm/hr.) x (19.43km2) x (0.43)
= (39.8mm/hr.) x (1.943 x1013 mm2) x (0.2713)
= 2.098 x 1014mm3/hr.
= 95,112.2711m3/hr.
= 26.42m3/sec.

2.2.2. River Flow Data

The river channel on which the proposed dam is expected to be built is not gauged. Despite
this limitation, flow data had to be sought from a nearby gauged catchment as the flow
data are necessary for calibration and validation of the rainfall-runoff model selected for
this assessment. The nearby gauged river is Ntambalale River gauged at Kinabili Village,
around 16km South-East from the study catchment. The station is operated by Internal
Drainage Basin Water Board.

The available Ntambalale River data spans for 28years from 1990 to 2017. After the data
were secured from the IDBWB, were examined for possible errors such as outliers and
gaps by using spreadsheet tool (data plots). The plots shown many short and long period
data gaps, outliers and unjustifiable trends. Identified short gaps were filled while no
attempt was made to fill longer gaps to avoid skewing the data. Outliers identified with
good confidence were rectified using moving average methods. Original data set is
attached on annex 2.

2.2.3. Weather Data

Weather data (temperature, humidity, wind speed and sunshine) for use in rainfall-runoff
modelling were sourced from Kahama Met Station (operated by Lake Tanganyika Basin
Water Board). The station is located around 23km north-west of the study site. The data
spans for 33 years from 1989 to 2021. The data were examined for errors (outliers and
gaps) and rectified accordingly then the monthly averages were prepared. The Monthly
Long Term Averages of these parameters are shown on Table 4.
Collection and use of as many as possible weather data was not possible due to limited
availability and poor data quality of weather stations data. However, the available data
were satisfactory for the study.

Table 4. Monthly Long Term Average Weather Data

Evaporation Average Humidity Wind Speed Sunshine
Month
(mm) Temp (○C) (%) (km/day) (Hrs)
January 4.4 25.5 68.7 0.4 6.8
February 5.6 25.1 68.4 0.4 6.9
March 5.8 24.8 71.3 0.4 7.4
April 5.6 24.3 65.5 0.5 7.4
May 6.2 25.5 57.6 0.7 7.7
June 6.4 24.3 54.7 0.7 7.8
July 6.9 23.6 53.0 0.7 8.0
August 7.9 23.0 50.0 1.2 8.0
September 8.5 25.3 53.8 1.1 8.3
October 7.6 25.3 53.8 0.9 9.3
November 8.1 26.6 56.2 1.1 9.1
December 5.9 24.7 65.7 0.6 8.1

2.2.4. Spatial Data

A high-resolution Digital Elevation Modal (12.5m resolution DEM) was downloaded from
https://search.asf.alaska for use in delineating the catchment upstream of the dam site in
cooperation with Topo maps. The topo map of the area was not as useful as required
because the tail-end area of the catchment is very flat such that the height difference
between lowest point (dam site, 1,199m a.m.s.l.) and highest point (water divide 1,248m
a.m.s.l.) fall in the middle of two contour lines.

Important details of the delineated catchment useful for this assessment are summarized
on Table 5 while the catchment map is shown on Figure 6.
Table 5. Summary of Important Details of the Delineated Catchment
Sn Item Measure/Details
1 Catchment Area 19.43km2 (1,943 ha)
2 Height Difference between highest and lowest point 47.85 meters
3 Longest water channel (river/stream) 7.36km
4 Drainage Pattern Dendritic
5 Flow direction W-E

Figure 6. Map of the Catchment Upstream of the Proposed Dam

 3. Rainfall – Runoff Modelling
The river channel on which the proposed dam is expected to be built, is not gauged just
as many small rivers in Tanzania. Under such circumstances rainfall-runoff model should
be developed to simulate the natural hydrological processes to estimate the runoff from
the catchment. Rainfall-runoff models are mathematical model describing the rainfall–
runoff relations of a catchment area. In simple words, the model carries out the conversion
of rainfall into runoff. The models should be tested and the generated runoffs should be in
good agreement with actual runoff hydrograph before the model results can be used.

There are several rainfall-runoff models in use; they vary in nature, complexity and
purpose (Shoemaker et al., 1997). This study used Soil and Water Analysis Tool (SWAT
Model). The SWAT Model was selected because it accounts for varying soil, land use,
and management conditions and is also compatible with the available data.

Input data for the model includes land use map (generated by ArcSWAT tool which
integrates remote sensing techniques), soil data obtained from Global Soil Data by FAO,
a high resolution (12.5m DEM downloaded from USGS website) and time series data
(rainfall, relative humidity, temperature, solar radiation and wind speed) and discharge
data for model calibrations.

Aerial rainfall time series used are those generated by Kriging method in section 2.2.1.
River flow data were obtained from IDBWB’s Ntambalale River at Kinabili Gauging Station
(see section 2.2.2) Relative humidity, temperature, solar radiation and wind speed time
series are those obtained from LTBWB’s Kahama Met Station (see section 2.2.3).

The model was calibrated and validated using the observed flow (Ntambalale flow data).
Simulation period for the model was fifteen (15) years (1999-2013) and warm-up period
was three (3) years (1999-2001). Model calibration was done by SWAT-CUP using two-
third (2/3) of the dataset which is from 2002 – 2006 while one-third (1/3) of the dataset
from 2006-2008 were used for data validation.

Assessment of the model performance was done by using hydrographical approach and
statistical methods including Coefficient of Determination (R2) and Nash-Sutcliffe
Efficiency (NSE). This helped in coming up with best fit scenario. The results of modelling
are summarized on Table 6.

Table 6. Summary of SWAT Model Results for the Proposed Dam Catchment
Sn Parameter Depth (mm)
1 Precipitation 774
2 Evaporation 381
3 Potential Evapotranspiration 1,638.5
4 Evaporation from shallow aquifer 30.93
5 Percolation to shallow aquifer 64.87
6 Surface Runoff 7.46
7 Lateral Flow 15.02
8 Return Flow 40.3
9 Recharge to deep aquifer 3.24
10 Average Curve Number (CN Number) 45.35

The results of the modelling shows that only 7.46mm out of 774mm of total precipitation
in this catchment becomes surface runoff while other either evaporates or recharges to
shallow and deep aquifers. This runoff depth translates to runoff volume by using normal
formula of calculating volume of an object; Volume = Area x. Depth. Recalling section 2.2.4
and figure 7 where the area of the catchment was found to be 19.43km2, then;

Runoff Volume = Area of the catchment x. Runoff Depth

= 19.43km2 x. 7.46mm
= 19,430,000,000,000mm2 x. 7.46mm
= 144,947,800,000,000mm3
= 144,947,800Litres
= 144,948m3

The calculations above concludes that the renewable annual runoff volume that can be
harvested at the proposed dam site in Lulenge Village is estimated at 144,948m3
((144,947,800) One Hundred Forty Four Million, Nine Hundred Forty Seven Thousands
and Eight Hundred Liters).
 4. Results and Conclusion
4.1. Results
The study has established that there are just 19.43km2 (1,943 hectares) of catchment area
that contributes to the flows passing at the proposed dam site. The aerial rainfall in the
catchment was found to be 774mm a year. Rainfall-runoff modeling of the catchment has
indicated that a total of 144,948M3 of surface runoff can be harvested annually. Probable
Maximum Flood required for designing of dam’s spillway (one of very important dam safety
structure) was calculated and found to be 26.42m3/second.

Summary of all results for this hydrological assessment is presented on table 7.

Table 7: Summary of Hydrological Assessment Results

Sn Item/Parameter Unit Results
1. Catchment Area Km2 19.43
Hectares 1,943
2. Catchment Max and Min Point Height Difference Meter 47.85
3. Catchment’s Longest River Channel Km 7.36
4. Long Term Average Annual Aerial Rainfall Mm/Year 774
5. Long Term Average Daily Rainfall Intensity Mm/Day 10.7
6. Long Term Average Hourly Rainfall Intensity Mm/Hour 3.6
7. 100 Years Return Time Hourly Rainfall Intensity Mm/Hour 39.8
8 Run-off Depth MM 7.46
9. Probable Maximum Flood (Design Flood) M3/Second 26.42
10. Renewable Runoff Annual Volume M3/Year 144,948

4.2. Conclusions
The ultimate goal of this hydrological assessment to determine hydrological factors
required for designing and sizing of the proposed dam to be constructed at Malenge
Village, Malenge Ward in Kahama District has been completed successful.