View source for Chapter Three: Water Sources Analysis

Water source forms one single most important element and is a key to proper function and thus sustainability of any water supply project. Evidence shows that whenever proper water source analysis has not been adequately conducted, most of water supply projects fall into dysfunction. Thus, this chapter presents the analysis of water sources. It includes analysis of both surface and ground water. Further, the chapter gets into 'nitty gritty' of each of this category.

=Chapter Three: Water Sources Analysis=

==3.1	Availability of Water Resources in Tanzania Mainland==
Tanzania mainland is endowed with a wide range of water resources that includes the main drainage systems, river basins and natural wetlands that are identified in the next paragraphs. With respect to the drainage systems of water resources in Tanzania mainland, it is divided into five drainage systems that include:
* The Indian Ocean drainage system,
* The Internal drainage system to Lake Eyasi, Natron and Bubu depression,
* The Internal drainage systems to Lake Rukwa,
* The Atlantic Ocean drainage system through Lake Tanganyika, 
* The Mediterranean Sea drainage system through Lake Victoria.

The drainage systems in turn consist of nine river basins with some bearing names resembling the drainage systems. These nine basins are indicated in Figure 3.1.


 

Figure 3.1: The River Basins of Tanzania (Source: NWSDS, 2008)

Legend for Figure 3.1

(I) Pangani River Basin	(V) Lakes Nyasa Basin
(II) Wami/Ruvu River Basin	(VI) Internal Drainage Basin
(III) Rufiji River Basin	(VII) Lake Rukwa Basin
(IV) Ruvuma and South Coastal River Basin	(VIII) Lake Tanganyika Basin
(IX) Lake Victoria Basin
	
From geographical point of view, Tanzania is a party to at least eleven transboundary water resources in form of Lakes and Rivers (NWSDS, 2008). These include the following:

* Lake Victoria,
* Lake Tanganyika,
* Lake Nyasa,
* Lake Chala,
* Lake Jipe,
* Kagera River,
* Mara River,
* Pangani River,
* Umba River,
* Ruvuma River and
* Songwe River.

With its numerous water bodies, Tanzania is perceived to have abundant surface and groundwater resources for meeting its present consumptive and non-consumptive needs. However, the reality is that severe and widespread water shortages exist in many areas of Tanzania because of climate variability, poor distribution of the resource in terms of time and space, and inadequate management of the water resources (NWSDS, 2008). As a result, Tanzania experiences frequent and intense water shortages and some water use conflicts.

Furthermore, Tanzania is relatively dry with more than half of the country receiving, on average, less than 800 mm of rainfall per year depending upon air circulation patterns and the movement of the convergence zones in the region. The semi-arid Central and Northern parts of the country, including areas immediately South of Lake Victoria receive less than 700 mm of rainfall per annum and are dry for an average of seven consecutive months a year. River flows in these areas are intermittent. In the Southern, Western and Northern highlands, which receive more than 1,000 mm/year of rainfall, rivers are perennial, and some of these experience frequent floods.
 
As an example, in 1999 the availability of renewable freshwater resources, both surface and groundwater was estimated  to be about 2,700 m3/capita/year. By 2018, this estimate was reduced to 2,330 m3/capita/year due to increased population alone.  The average figure is significantly above the level of 1,700 m3/capita/year set by the United Nations as denoting water stress, or 1,000 m3/capita/year denoting water scarcity. Furthermore, due to the projected population growth alone, Tanzania's annual freshwater renewal rate is projected to drop to 1,500 m3/capita/year by 2025, thus categorising the country as being water stressed by then. 

On the whole, Tanzania has sufficient surface and ground water resource potential to meet most of her present needs.  However, differences in topography, rainfall patterns and climate account for the existing variation in the availability of water in different parts of the country.  In the densely populated Pangani and Rufiji Basins, these variations have already resulted into water stress.  It is estimated that the annual surface runoff from Tanzania to the world’s oceans is about 74 x 109 m3. The Rufiji, which drains a 177,000 km2 area, contributes over 50% of the runoff3. Typical annual runoffs are shown in the Table 3.1 for some of the major rivers of Tanzania. 

Table 3.1 Mean annual river discharges for some of the principal rivers of Tanzania
River 	Mean Annual Discharge (Million m3/yr) 
Rufiji (at Steiglers Gorge) 	22,250 
Kilombero (at Swero) 	14,470 
Malagarasi (at Taragi Ferry) 	5,060 
Ruvu (at Moro Bridge) 	1,370 
Wami (at Mandera) 	3,280 
Ruhuhu (at Kikonge) 	5,600 
Kiwira (at Kyela) 	1,900 
Kagera (at Kyaka) 	7,064 
Mara (at Mara Mines) 	1,971 
Pangani (at Hale) 	627 
(Source: NWSDS, 2008)
 

The most abundant surface water resources exist in Lakes Victoria, Tanganyika, Nyasa, Chala and Jipe, as well as the Kagera, Mara and Songwe rivers, which are trans-boundary waters. The use of these abundant surface water resources for water supply, irrigation and other purposes is still very limited even today.

Tanzania is also rich in wetland systems that are areas which, for part of the year, have enough water to enable the development of types of plants and animals adapted to these conditions.  These include the lakes of the Western and Eastern Rift Valley system, Lake Victoria, numerous small lakes, riverine flood plains and permanent swamps, coastal mangrove and deltaic systems, and a number of artificial impoundments and reservoirs and fish ponds.  There are numerous permanent and seasonal freshwater swamps and flood plains distributed in almost all of the country's major drainage basins, which account for some 2.7 million hectares.  The largest in this category are found in the Rufiji/Ruaha river system and in the Malagarasi/Moyowosi system, while other river systems are the Kagera River, along with Ugalla River, Suiwe River, Mara River, Pangani, Wami and Ruvu Rivers. The principal wetlands of Tanzania constitute one of the country's richest and most durable resources. 

==3.2	Water Sources Available in Tanzania Mainland==
In Tanzania, there are three main categories of water sources available, namely rainwater, surface and groundwater. 

===3.2.1	Rainwater and fog harvesting===
One of sources of water include rainwater and fog that can generate limited amounts of very clean water if they are properly collected and stored. In an area where other water sources are not available, consideration should be given to harvesting rainwater and fog.

===3.2.2	Surface Water===
For design purposes, the surface water sources that can be considered include;
* Rivers or streams,
* Impoundments (Reservoirs and ponds),
* Springs,
* Lakes,
* Dams (charco, sand, earth etc).

A brief description of each water source is provided below.

3.2.2.1	 Rivers or streams 
Rivers and streams are water sources that originate from springs located in highlands which flow down to the end of the respective drainage basin which can be lakes, seas or oceans as depicted on the map of Tanzania in Figure 3.1. 
3.2.2.2	 Impoundments
Impoundments includes all types of reservoirs that emanate from road borrow pits, mining, human or natural activities that are utilised as sources of water for a formal water supply project. 
3.2.2.3	 Springs
Springs include artesian or freely flowing spring water that has been tapped by an intake structure to facilitate supply of water to a designated community. Spring is a point where groundwater flows out of the ground, and is thus where the aquifer surface meets the ground surface. A spring may be ephemeral (intermittent) or perennial (continuous). Springs can be developed by enlarging the water outlet and constructing an intake structure for water catchment and storage. 

3.2.2.4	Lakes 
Lakes found in Tanzania are either located at the end of drainage basins or are highland lakes and some of them are volcanic lakes. Tanzania is endowed with many small inland lakes apart from the third biggest lake in the world (Lake Victoria) as well as Lake Tanganyika which is the world’s deepest lake. Both lakes supply water to various localities around their respective catchments.
3.2.2.5	 Dams 
Dams are classified based on the availability of construction materials. Various types of dams can be built ranging from earth fill dams, concrete dams, sand dams and charco dams. These are purposely built structures that allow impoundment of river and/or rain water for various end uses. 

3.2.3	Groundwater
Groundwater is that portion of rainwater which has percolated beneath the ground surface to form an underground reservoir referred to as aquifer water. The upper surface of groundwater is the water table. Groundwater is often clear, free from organic matter and bacteria due to the filtering effect of the soil on water percolating through it. However, groundwater almost always contains dissolved minerals from the soil. Groundwater is often better in terms of quality than surface waters. It is less expensive to develop for use, and usually provides more adequate supply in many areas in the country. In semi-arid and the drier parts of the country,  groundwater has played and will continue to play a major role as the sole water source for various uses especially in the central and northern parts of the country and the drier regions of Dodoma, Singida, Shinyanga, Tabora, Mwanza, Mara, Arusha, Coast and Southern Kilimanjaro.

Groundwater can be considered as either spring water or well (or borehole) water. Springs, offer excellent water supply opportunities, but are generally found in hilly or mountainous areas only. They may require long pipelines in order to bring the water to the demand area. This is a feasible source for larger and concentrated settlements but rarely for dispersed populations. For rural water supply systems, groundwater is generally preferred as a water source. 

The main sub-types of groundwater and extraction methods are as follows:

3.2.3.1	 Infiltration galleries/wells 
Infiltration galleries are horizontal wells, constructed by digging a trench into the water-bearing sand and installing perforated pipes in it. Water collected in these pipes converges into a “well” from which it is pumped out. 
3.2.3.2	 Well
This is a hole constructed by any method such as digging, driving, boring, or drilling for the purpose of extracting water from underground aquifers. Wells can vary greatly in depth, water volume and water quality. Well water typically contains more minerals in solution than surface water and may require treatment to soften the water by removing minerals such as arsenic, iron and manganese. Well water may be drawn by pumping from a source below the surface of the earth. Alternatively, it could be drawn up using containers, such as buckets that are raised mechanically or by hand.

Wells are various types of artificially constructed water production wells that are designated as shallow wells (up to 20 metres deep) or deep wells (more than 20 metres deep) as designated by the Ministry responsible for water from time to time.  Water is pumped out of the well into the end user containers or a storage tank using various types of pumps that can be driven manually or using various energies. Typical cross sections through such wells are given in Section 3.6.5.
3.2.3.3	Classification of wells based on the aquifer tapped
As mentioned, an aquifer contains a considerable amount of groundwater underground beneath layers of permeable soil material like sand or gravel. Aside from their water storage capacity, aquifers allow the underground flow of groundwater. Aquifers are recharged with rainwater that seeps down to the soil and through the permeable layers.
3.2.3.3.1	Shallow wells
Generally, a well is considered shallow if it is less than 20 metres deep. Shallow wells tap the upper water-bearing layer underground. This permeable layer, however, usually has limited safe yield due to its great dependence on seasonal rainfalls. Therefore, the supply capacity of shallow wells could be unreliable and are sometimes intermittent. Also, the water extracted from the upper strata is usually more affected by contamination since the aquifer being tapped is near the ground surface where possible sources of contamination are abound. Protection against contamination is therefore one of the main considerations in constructing a shallow well.
3.2.3.3.2	Deep wells
Deep wells, which are over 20 metres deep, tap the deeper unconfined aquifer. This aquifer is not confined by an overlying impermeable layer and is characterized by the presence of a water table. A deep well is less susceptible to surface contamination because of the deeper aquifer. Also, its yield tends to be more reliable since it is less affected by seasonal precipitation.

3.2.3.3.3	Artesian wells
Artesian wells are much like the deep wells except that the water extracted is from a confined aquifer. The confining impermeable layers are above and below the aquifer. Groundwater recharge enters the aquifer through permeable layers at high elevations causing the confined groundwater at the lower elevations to be under pressure. In some cases, the hydraulic pressure of the aquifer is sufficient for a well to flow freely at the well head.

3.3	Quality Suitability of Water Sources for Water Supply Projects
When considering the different water sources for water supply projects, it is necessary to ensure that the quality of the water source expected to be utilised is monitored well preferably for a period of not less than three years consecutively to ensure the variability of the quality is captured during the wet and dry seasons. When one looks at the list of the potential sources presented in the foregoing section, such a monitoring programme may not be necessary for rainwater and fog. Only short-term monitoring of the quality of these two sources should be undertaken.

3.4	Pilot Testing of Water Sources for Establishment of Appropriate Treatment

A decision on whether the water source needs to be subject to water treatment or otherwise will emanate from the results of the short term tests on the quality of the water which will in turn guide the decision of pilot testing the recommended flow sheets particularly for river/streams, lakes, impoundments and dams. Groundwater will usually need only a few unit operations for removal of the identified elevated impurities that may include Iron, Manganese or Fluoride and the need to disinfect water from shallow wells in addition to maintaining residual disinfectants for prevention of re-contamination. Rainwater and condensates from fog will not need to be pilot tested.

3.5	General Considerations for Selection of Water Sources 
In the selection of a source or sources of water supply, adequacy and reliability of the available supply can be considered as the overriding criteria. Without these, the water supply system cannot be considered viable. These, together with the other factors that should be considered (and which are interdependent), are as follows:

3.5.1	Adequacy and reliability
Adequacy of water supply requires that the quantity of water flow of water source be large enough to meet present and future water demand. On the other hand, source reliability can be expressed by how frequently a water system expects normal demand to go unmet, such as a one-in -25 year or even a one-in-50 year drought. Safe yield is a 1-day low flow rate that is exceeded for 96 percent of the period of record and that can be related to the determined average daily water demand in order to establish the reliability of a water source. For a river/stream, safe yield represents the minimum flow rate that will guarantee no risk to the river hydrology and its surroundings. Safe yield is estimated so as to check whether the planned withdrawal for water supply purposes will be met. To determine the safe yield of a river or stream, a flow -frequency/probability analysis presented in section 3.1.5.5 should be performed. From the analysis, the determined 96 % low flow index should be taken as the safe yield of the river or stream and thus considered as the water source reliability.

3.5.2	Quality of water sources
The assessment of water quality of a water source is important to establish the suitability of water source for human consumption. The quality of surface water is determined by the amount of pollutants and contaminants picked up by the water in the course of its travel. While flowing over the ground, surface water collects silt, decaying organic matter, bacteria and other micro-organisms from the soil. Sources which require little or no treatment of the water should be chosen in the first instance, provided the required quantity of water can be obtained. Hence springs and ground water resources should always be exploited in the first hand. Surface water from rivers, streams and lakes will almost always require some treatment to render it safe for human consumption. However, for large supplies, surface water will often still be the most economical alternative. Rivers which have the bulk of their catchments in forest areas should be preferred.

Thus, all surface water sources should be presumed to be unsafe for human consumption without some form of treatment. The option to treat surface water to make it safe for human consumption in compliance with the latest edition of Tanzania potable water standards (TBS, TZS 789) has to be evaluated to decide on the feasibility of the water supply project.

3.5.3	Technical Requirements
The development of the source should be technically feasible, the operation and maintenance requirements for the source abstraction and supply system should be appropriate to the resources available.

3.5.4	Cost implications to develop a water source
The assessment of investment costs to develop a given water source including operation and maintenance costs has a bearing in the selection of the water source for development. Affordability of investment costs is an important factor to be considered in the selection of the water source.

3.5.5	Protection of water sources
The location of a water source is a key factor in securing the highest quality water source. In analysing a source location, the design engineer should consider the measures necessary to protect the water source from human excreta, from industrial discharges and from agricultural run-off. In addition, measures to establish and maintain watershed control, physical protection and barriers to contamination have to be considered to ensure sustainable quantity and quality of the raw water.

3.5.6	Legal and management requirements
Ownership of the land and the legal requirements of obtaining permission to abstract are also factors to consider when selecting a source. Sources on private land may cause access problems.

3.5.7	Distance of water supply source
The source of the water supply must be situated as near to the demand area as possible. Hence, less length of pipes needs to be installed and thus economical transfer and supply of water. The source(s) nearest to the demand area is usually selected.

3.5.8	Topography of the project area and its surroundings
The area or land between the source and the area to be served by water supply system should not be highly uneven, i.e., it should not have steep slopes because cost of construction or laying of pipes is very high in such areas.

3.5.9	Elevation of a source of water supply
The source of water must be on a high elevation than the demand area so as to provide sufficient residual pressure in the water for daily requirements. When the water is available at lower levels, then pumps are used to pressurize water. This requires an excess developmental and operational tasks and costs.

3.6	Determination of water source yield
Sources which require little or no treatment of raw water such as springs, wells and boreholes should be given the highest selection priority provided their yields are sufficient to meet the water demands of the water supply scheme. For large supplies, surface water will continue to be the most economical alternative water source. In selecting surface water sources, rivers with upland and mostly forested catchments should be given preference. Sub-surface water drawn from a riverbed or river bank can sometimes be a viable alternative in dry areas with only seasonal flows in the river, or in rivers with a high silt load.

Sources from which water can be supplied by a gravity system are particularly more favourable than those which require pumping with significant energy costs. For household and small community water supplies, rainwater harvesting will be the most appropriate in most medium and high potential areas in Tanzania that receive sufficient rains.


3.6.1	Rainwater and fog harvesting
3.6.1.1	 Rainwater harvesting
Rainwater harvesting is a technique of collection and storage of rainwater into natural reservoirs or tanks, or the infiltration of surface water into subsurface aquifers (before it is lost as surface runoff). Types of rainwater harvesting systems are described in the sections below.

3.6.1.1.1	Types of rainwater harvesting
Two types of rainwater harvesting should be considered:

•	Land catchment,
•	Roof catchment.

Important data for design of rainwater harvesting systems:
•	Rainfall data
•	Catchment/Surface Area 
•	Run-off Coefficient

To accurately estimate the potential rainwater supply from a catchment, reliable rainfall data for a 10-year period is required . The Hydrology Section, Tanzania Meteorology Agency, and Agriculture Departments should be contacted for rainfall data wherever rainwater-harvesting technology is proposed. 

The amount of rainfall collected depends on the surfaces where rain falls and the runoff coefficient K of the surface. The runoff coefficient varies with topography, land use, vegetation cover, soil type and moisture content of the soil. In selecting run off coefficients the future characteristics of the water shed are considered. If land use varies within a water shed consider the segments individually and use a weighted coefficient value to determine the total runoff for the watershed. Practically annual rainfall falls during the rainy seasons between October and May.

(i) Run-off Coefficients
Table 3:2 shows the runoff coefficients for various surfaces. They should be used for calculating the fraction of the rainfall which can be harvested.

Table 3.2: Run-Off Coefficients for Different Surfaces
SURFACE	RUN-OFF
COEFFICIENT (K)
1.	Roof catchments	
	•	Roof tiles	0.8 to 0.9
	•	Corrugated sheets	0.7 to 0.9
2.	Ground surface covering
	•	Concreted	0.6 to 0.8
	•	Bitumen, plastic sheeting, butyl rubber	0.8 to 0.9
	•	Pavement of stone, bricks with open joints	0.5 to 0.6
	•	Pavement of stone, bricks with tightly cemented joints	0.75-0.85
3.	Compacted and smoothened soil	0.3 to 0.5
4. Lawns, sandy soil
	•	2% slope	0.05-0.10
	•	2.7% slope	0.10-0.15
	•	>7%	0.15-0.20
5. Lawns, heavy soil
	•	2% Slope	0.13-0.17
	•	2.7% slope	0.18-0.22
	•	>7%	0.25-0.35
6.	•	Uncovered surface, flat terrain	0.3
	•	Uncovered surface, slope less than 10%	0.0 to 0.4
	•	Rocky natural catchments	0.2 to 0.5

3.6.1.1.2	Components of rainwater harvesting system
a) Catchments Area: The catchment of a water harvesting system is the surface which directly receives the rainfall and provides water to the system.
b) Coarse mesh at the roof to prevent the passage of debris.
c) Gutters to collect and transport rainwater to the storage tank. Gutters can be semi-circular or rectangular and could be made using:

•	Locally available material such as plain galvanised iron sheet (20 to 22 gauge), folded to required shapes 
•	Semi-circular gutters of PVC material can be readily prepared by cutting those pipes into two equal semi-circular channels.
•	Bamboo trunks cut vertically in half

The size of the gutter should be according to the flow during the highest intensity rain. It is advisable to make them 10 to 15 per cent oversize.

d) Conduits/pipeline that carry rainwater from the catchment or rooftop area to the harvesting system. Conduits can be of any material like polyvinyl chloride (PVC) or galvanized iron (GI), materials that are commonly available.

e) First Flush pipe to separate first rainwater contaminants namely debris, dirt, and dust.

3.6.1.1.3	Estimation of the yield
A first estimate of the average yield of a catchments area can be found using the following expression.
S = K × I × A						(3.1)
Where: 
S = Yield in m3 / annum 
A = Area of catchment/surface, m2 
I = Average annual rainfall m/annum 
K = runoff coefficient 

Determination of average runoff coefficient for the entire catchment area composed of different surfaces can be calculated as follows:
 +...					(3.2)
The required capacity of the collection facility should be calculated using available meteorological data showing the rainfall pattern of the area. However, for rough calculations the storage tank, capacity may be calculated as follows:
C = D × T × 10-3						(3.3)

Where:
C = Capacity of tank in m3 
D = Total water demand in litres / day
T = Longest dry spell in days

3.6.2	Hydrological Analysis of Surface Waters
The Design Manual for Water Supply projects is expected to reflect the best concepts on what constitutes the basis for designing a safe, reliable and sustainable water system. Hydrological principles must be taken into consideration during the feasibility and preliminary design stages of the water supply system to ensure that from the outset, design and construction of the system is done right. The design engineer must apply hydrological principles during design of the water system to ensure that the system being designed does not result in exhausted water supply sources and empty reservoirs after construction of the project is completed. If it is recognised from the beginning that there is water deficiency from the source, then the water source should not be considered for development.

The following steps should be followed when undertaking hydrological analysis for the water supply projects:

Step 1: Measurement of the quantity of surface water sources,
Step 2: Low flow assessment of surface water sources,
Step 3: Flood flow estimation,
Step 4: Rainfall analysis,
Step 5: Water permit application,
Step 6: Data to support hydrological analysis.

Step 1: Measurement of River Discharge
In order to assess the amount of water available from the identified surface water source, a discharge measurement must be carried out during both the dry and wet seasons. Measurement of discharge will highlight the production capacity of a water source, the information which is important in the planning of a water system. An estimate of the quantity of water that can be reliably produced by a water source gives the planner a basis to decide for or against its development. For the sources to be considered adequate, they must at least satisfy the average day water demand of the area to be served by a water system. The average daily water demand is calculated from estimated average water requirements for domestic, commercial, industrial, public institutions and livestock. The following methods can be used to measure discharge.

a) Volumetric Method
This method is appropriate for measuring small quantities of flow from small streams and springs. Flow can be measured by measuring the volume. The equipment required are a stop watch and a bucket or drum of known volume. The method consists of determining the time required to fill the bucket or drum. For more accurate results, the measurement is repeated several times, and the average time of these trials is taken.

b) V-Notch Weir Method
A weir is an overflow structure built across an open channel for the purpose of measuring the rate of flow. Weirs may be rectangular, trapezoidal or triangular in shape. The triangular or V-Notch Weir is a flow measuring device particularly suited for small flows. The V-Notch Weir often used in flow measurements is the 90° V-Notch that is placed in the middle of the channel and water is allowed to flow over it. The water level in the channel is then measured using a gauging rod. The zero point in the rod should be level with the sill or crest of weir/notch. For a known height of water above the zero in the rod, the flow in cumecs for the 90° V-Notch can be obtained by using the formula:

 							(3.4)

 								(3.5)

Where,
Q = Discharge in m3/sec
H =Height of Water level above the crest of the weir in meters
In this case, the discharge coefficient of the weir is approximated to be equal to 0.58.

c) Current meter Measurement
The current meter is an instrument that is used to measure relatively larger quantities of flow from streams and rivers. The instrument consists of a propeller rotating freely on a well-lubricated shaft. The device is lowered into the water and the rate of revolution of the impeller is directly proportional to the velocity of the water flow. A small magnet is usually built into the shaft of the instrument and a coil detects the passage of the magnet and allows the number of revolutions of the shaft in the given time to be counted. Once the rate of revolution of the impeller is known the water velocity can be calculated using the calibration equation for the instrument, which is expressed as follows:

 									(3.6)

Where, 
V is the water velocity in meters per second
n is the number of revolutions of the impeller per second
a, b are constants for the particular instrument.

The discharge measurement using the current meter at the time of measurement, Q,  is  determined by multiplying the velocity of flow and water flow cross-section area.

Step 2: Low flow assessment
The assessment of low flow magnitudes of streams/rivers or springs in hydrology is important in the planning of a water supply system in view of the fact that it reflects on the water source adequacy and reliability to meet the consumer demand. In low flow hydrology, two questions are asked about a particular river identified to be a water source for a given water supply system:

(i)	Does the river supply a particular water demand at all times?
(ii)	If not, how much water must be stored in order to meet any deficiency which may arise?

Techniques of flow duration curves, annual minimum flow analyses and annual drought volumes are applied to address the two questions.

a) Flow duration curve
The flow-duration curve (FDC) is defined as a cumulative frequency curve that shows the percent of time specified discharges were equalled or exceeded during a given period. It combines in one curve the flow characteristics of a stream throughout the range of discharge, without regard to the sequence of occurrence. To prepare a flow-duration curve, the daily, weekly, or monthly flows during a given period are arranged according to magnitude, and then percent of time during which specified flow values are equalled or exceeded are computed. 

A flow duration curve once it is prepared, is used to determine the indices of low flow magnitudes; for example, the 96-percentile flow (Q96), is the flow that is exceeded for 96 percent of the period of record. This discharge value is a useful index of low flow that is related to the quantity of water that can be available for water supply in the dry season.

The following steps are followed to construct the FDC:

(i)	Rank the observed stream flows in descending order (from the maximum to the minimum value).
(ii)	Calculate exceedance probability (P) of each flow as follows:

		 						(3.7)

		Where,
			P is the probability that a given flow will be equalled or exceeded 			(% of time),
			m is the ranked position of a given flow value on the list,
			n is the length of the sample.

(iii)	 A FDC is obtained by plotting each ordered observed streamflow value versus the corresponding calculated exceedance probability.
(iv)	 Read the indices of low flow magnitudes from the FDC corresponding to 90%, 95% and 99% probabilities of exceeded.

b) Low flow frequency analysis
The frequency analysis of low river flows is performed by analyzing 1-day or 7-day or 10-day annual minimum flow series obtained by selecting the lowest flow values occurring in each year of record. The set of observed annual minimum flow values recorded at any gauging station is assumed to be a random statistical sample from the population of all possible annual minima at the given site.

The selected set of observed annual minimum flow values is fitted to the Gumbel statistical distribution and then the annual minimum flow magnitudes (QT) corresponding to the design probability of failure (1/T) is then estimated from the Gumbel prediction equation:

 								(3.8)

Where,
QT = Low flow magnitude
T = Return period of one failure in T years
  and  = Gumbel Parameters 
KT = Frequency factor, obtained as:

 							       (3.9)

Results from low flow frequency analysis:
(i)	If the value of QT is large in comparison to QD, the average day water demand, then the river can be considered to be able to supply the demand satisfactorily.
(ii)	On the other hand if QT is less than or of the same order of magnitude as, QD, ,then the river alone without some form of flow regulation could not be considered satisfactory for supplying the demand. 

c) Annual drought volumes Analysis
On the basis of the results obtained from low flow frequency analysis, in case the annual minimum flow magnitudes (QT) corresponding to the design probability of failure (1/T) is found to be less than QD, demand flow, then storage will be required to meet the established water demand. The required storage is determined by carrying out deficiency/drought Volumes Analysis.

The storage required on a river to meet a specific demand depends on the following factors:

•	Variability of the river flow
•	Magnitude of the demand
•	Degree of reliability of meeting the demand

The capacity of the reservoir required to augment the river flow in any year can  be determined from the analysis of the series of annual maximum deficiencies (drought volumes) as follows:

Drought volumes V1, V2, V3,…, Vn are computed from a hydrometric record of the river flow (Qi), with reference to the demand flow (QD), i.e., (Vi = Qi-QD). 

The set of observed annual maximum deficiencies at any gauging station is assumed to be a random statistical sample. The annual maximum deficiency (VT) corresponding to the probability of failure (1/T) is estimated from the series of annual maximum deficiencies using a statistical distribution, e.g., Gumbel distribution as illustrated in previous section. The design storage of the reservoir can be made equal to the volume VT corresponding to a risk of one failure in T years. 

Step 3: Dependable Rainfall Analysis
Rainfall analysis is carried out when the need arises to determine dependable rainfall in a given area for the purpose of designing a rainwater harvesting system for domestic use. Frequency analysis of recorded annual rainfall data from a given area, enables the determination of the 90% dependable annual rainfall. This is the value of rainfall magnitude that will be exceeded 90% of the time. In the design of Rainwater Water Harvesting system, catchment (i.e. roof) area and depth of rainfall are important parameters for estimation of optimal storage size. Taking note of the fact that rainfall amounts vary on a year to year basis, the computed rainfall magnitude that is exceeded 90% of the time, is taken as the value of annual rainfall depth that can be expected to occur with some degree of certainty and thus used in the design. The exceedance probability is determined by ranking the observed annual rainfall in ascending order (from the minimum to the maximum value) and then calculating non-exceedance probability (P) as follows:

 								(3.10)

Where,
P is the probability that a given rainfall will be equalled or not exceeded (% of time),
m is the ranked position of a given rainfall value on the list,
n is the length of the sample.

The probability of dependable rainfall is obtained by calculating  the value of exceedance probability (1 – P).

Step 4: Flood flow estimation for intakes and small dams' spillways
The need to estimate flood peaks or design floods arises where it is required to design a spillway of a dam proposed for water storage and also the design of water intake structures. Water intakes and spillway of small dams are designed to accommodate the 100-year flood. 

Frequency analysis of observed Annual Maximum streamflow records from a gauging station enables the estimation of flood peaks. The statistical distribution namely the Gumbel distribution or other statistical distribution used in Tanzania such as Pearson type 3, Log-Pearson type 3, Log-Normal and General Extreme Value can be used to carry out frequency analysis in order to determine the magnitude of flood peak of 100-year required for the design. The estimation of design flood peak magnitudes for specified return periods using the Gumbel Distribution is illustrated below.

Prediction equation

 								(3.11)

Where,
QT = Flood peak magnitude
T = Return period of one failure in T years
  and  = Gumbel Parameters
KT = Frequency factor, obtained as:

 							(3.12)


Estimation of Gumbel parameters by Method of Moments (MoM)

Mean,  							(3.13)
Standard deviation,  α						(3.14)		

The mean, μ and Standard Deviation values, σ are computed from observed annual maximum streamflow records. Note that other frequency distributions (Pearson type 3, Log-Pearson type 3, Log-Normal and General Extreme Value) have different expressions for estimating the distribution parameters.

Step 5: Application for water permit
Water abstraction for water supply from a river or spring requires a permit from the respective water basin office. The planner of the water supply project must apply for the water permit abstraction early in the project design because it can affect the viability of a project. The design engineer must seek the water permit if the project involves a new, replacement, increased withdrawal from a source or an increase in the water system’s physical capacity.

Step 6: Environmental flow considerations
Environmental flow may be computed in terms of magnitude, timing of low flow in the dry month, duration of low flow in days, frequency of occurrence of the low flow event (return period) and rate of change of low flow over time (m3/day of flow recession). The recommended environmental flow varies for individual rivers and streams and therefore to determine its flow value, a comprehensive Environmental Impact Assessment (EIA) should be conducted and approved by NEMC. Also, there are some guidelines and procedures for environmental flow assessment for specific catchments in Tanzania developed by NEMC. Accordingly, designers need to consult NEMC for environmental flow information in their project areas (https://www.nemc.or.tz/).

Step 7: Data to support hydrological analysis
Hydrological data is invaluable for planning of water supply systems. For example, water source adequacy and reliability can be determined from analysis of streamflow data which is important hydrological data. Hydrological data expected to be collected by water basin offices in Tanzania to support the planning of water supply systems, specifically to answer questions related to the following:

•	Water availability in terms of quantity and quality
•	Frequency of occurrence of low flows and flood flows
•	Variability of flow regime in terms of quantity and quality

Important data to be collected include the following:

(i)	Streamflow – required to quantify available water and estimate flood peaks and low flow magnitudes,
(ii)	Rainfall – required to determine 90% dependable rainfall,
(iii)	Sediment - Sediment deposition affects the water carrying capacity of rivers and the useful life of reservoirs. Sediment data is required to determine the useful reservoir capacity and the life span of the reservoir.
(iv)	 Climate data - (Evaporation, Temperature, Wind speed, sunshine hours, radiation and humidity) – required to estimate water loss from reservoirs.

During the feasibility and preliminary design stage, the design engineer must look for streamflow records from stream gauging stations located at or near water intakes and dam sites to support the design work. In a situation where there are no gauging stations at or near water intakes or dam sites, two options may be considered to get flow data to be used in the design.

Option 1: Transfer data from adjacent or neighbouring drainage areas that have comparable or similar characteristics. The same applies to a situation where rainfall data is missing, rainfall data from adjacent or similar catchments is used to derive flow frequency/probability curves required in the design of water supply system.
Option 2: Install permanent or temporary gauging stations and start recording flow data at the earliest possible time during the planning steps of the water supply project.

3.7	Other considerations for various water sources
3.7.1	Water permits Considerations
During the course of implementation of water supply projects, designers will need to work with the relevant Water Basin Authorities and relevant catchment and sub-catchment committees to ensure all water users with water withdrawal permits are considered during the course of sizing the projects to ensure no developmental constraints are faced as a result of ignoring other users. It will be necessary to consult the updated water permits registers maintained by each Basin Water Board prior to planning expansion of any new water supply project.

3.7.2	Conservation of water sources
In line with NAWAPO, the protection and conservation of water sources is one of the main duties of all the Basin Water Boards.  Intuitively, for national water resources the MoW also has the responsibility to deal with resolution of all water use conflicts. It will ensure that the WRM Act No.11 of 2009 as well as the associated regulations are fully observed by all parties. Other relevant laws such as those associated with pollution coordinated by other agencies or bodies like the National Environment Management Council (NEMC) are observed with respect to water resources. Where transboundary water resources are involved, the MoW has to ensure the protection roles that are expected of Tanzania are properly fulfilled in line with the relevant international laws, agreements or conventions stated in section 1.1.4.




3.8	References