© FEB 2026 | IRE Journals | Volume 9 Issue 8 | ISSN: 2456-8880

DOI: [Link]

Water Supply in Ibadan North Local Government

AYOADE, A. T.1, OYELOLA, I. B.2, OLABODE, V. T.3, ADESOGAN, S. O.4
1, 2, 3, 4
Department of Civil Engineering (Water Resources & Environmental Engineering), University of
Ibadan, Ibadan, Nigeria

Abstract- Access to reliable water supply remains a major variability, and poor planning. In Nigeria, less than
challenge in rapidly urbanizing cities of developing half of urban residents have reliable access to piped
countries. This project was undertaken to examine the water, forcing many households to depend on wells,
water supply situation in Ibadan North Local Government
boreholes, and water vendors.
Area, with the aim of understanding the sources of water
available to residents, the level of adequacy and reliability
of supply, and the challenges faced in meeting daily water Access to clean and sufficient water is one of the most
needs. The choice of the study area was informed by its fundamental needs of any urban population. Yet,
high population density and its importance as a residential across many cities in sub-Saharan Africa, including
and commercial hub within Ibadan metropolis. Structured Nigeria, this basic necessity is increasingly under
questionnaire were administered to residents to assess the threat. The high rate of urbanization, population
water supply, determine the water consumption of the study increase, shifting rates of consumption and variability
area. The result of the questionnaire were analyzed using in climatic conditions have been pooled together to
Microsoft Excel. A well yield test was carried out to know
exert colossal pressure on the existing water supply
how much water flows into the well from surrounding rock
infrastructure. Specifically, this challenge is
or soil ion a given time, the water supply for the area was
designed. For the prediction of household water demand, pronounced in such urban centres in Nigeria such as
Machine learning algorithms which includes linear Ibadan which is one of the largest and densely
regression, SVR, Neural Network, random forest populated cities. The demand of water in the city is
regression models were developed and evaluated using ever-increasing, whereas the supply systems cannot
metrics like MAE, RMSE, MAPE and R2 The study meet the demand because of the old-fashioned
combines field data, observations, and analytical methods infrastructure, lump-sum funding, and insufficiently
to provide a realistic picture of water supply conditions in grounded planning.
the area. Emphasis was placed on presenting findings in a
clear and practical manner so that the outcomes of the
Water supply is the provision of water by public
study can be useful not only for academic purposes but also
for policy makers, water supply agencies, and other utilities commercial organizations, community
stakeholders involved in urban water management. The endeavors or by individuals, usually through a system
random forest model outperformed the linear model, of pumps and pipes. The supply of fresh water in
achieving an R2 of 0.99 It is hoped that the findings and Nigeria for domestic and drinking purposes are from
recommendations presented in this project will contribute, three sources which include surface water ( river,
in a modest way, to ongoing efforts aimed at improving stream, pond, lakes,etc), groundwater or sub-surface
water supply planning and management in Ibadan North water (hand dug well, borehole) and rainwater. In
Local Government Area and similar urban settings.
Nigeria, groundwater is a major source of water used
for domestic purposes, while surface water is prone to
Keywords: Evaluation, Household Characteristics, Ibadan
North, Machine Learning, Water Supply pollution rainfall is the purest of natural water.

I. INTRODUCTION Despite the presence of surface and groundwater

resources, the existing water supply infrastructure is
Water supply is essential for public health, economic unable to meet growing demand, particularly in
densely populated areas such as Ibadan North LGA.
development, and environmental sustainability.
Conventional demand estimation methods, based
However, many urban centers in sub-Saharan Africa
continue to experience water shortages due to rapid largely on per capita assumptions and trend
population growth, aging infrastructure, climate

IRE 1714016 ICONIC RESEARCH AND ENGINEERING JOURNALS 7

 © FEB 2026 | IRE Journals | Volume 9 Issue 8 | ISSN: 2456-8880
DOI: [Link]
extrapolation, have proven inadequate in capturing the III. RESULTS AND DISCUSSION
complex drivers of urban water consumption.
Population study
Recent advances in machine learning provide The population estimation is a very important part of
opportunities for improved water demand forecasting the water demand analysis since the population size is
by integrating demographic, climatic, and socio- directly proportional to the current and prospective
economic variables. This study therefore assessed the water needs. Using exponential growth method. The
water supply situation in Ibadan North LGA and exponential growth model is expressed as:
explored the application of data-driven models to Pn = Po ert
support sustainable urban water management.
where: r=annual population growth rate t = time
II. MATERIALS AND METHOD interval in years =2026- 2006= 20years
Pn=projected population at present year (2026)
Study Area P0 = population sample (from 2026 census) e=
Ibadan North LGA is located at the core of Ibadan base of natural logarithms
metropolis, Oyo State, Nigeria. The area is
characterized by high population density, mixed land For this study, a moderate urban growth rate of 2.5%
use, and intense residential and commercial activities. per annum was assumed, which is consistent with
It experiences a tropical climate with distinct wet and values commonly applied in Nigerian urban
dry seasons, annual rainfall of about 1,200–1,500 mm, population studies. The projection period considered
and average temperatures ranging from 25–30°C spans 20 years, from 2006 to 2026.

Data collection Substituting into the equation:

The methods of data collection adopted for this project P2026 = 308,119 × e0.025×20
included P2026 ≈ 508,000
• Structured questionnaires administered to
households to assess water sources, daily Therefore, the estimated population of Ibadan North
consumption, and supply reliability. Local Government Area in 2026 is approximately
• Field hydrogeological investigations including a 508,000 people. This projected population forms the
well-yield test. basis for subsequent water demand forecasting and
• Population data and climatic records scenario analysis in this study. The present population
at 2026 is 508,000 to be projected to 2046. P2026 =
Tools and software used 837,550
The tools and software used in this research are
Microsoft excel, Google map and python Sources and Adequacy of Water Supply
Results show that the majority of households rely on
Data analysis groundwater sources, particularly hand-dug wells and
Questionnaire responses were analyzed using boreholes. Public piped water supply was found to be
descriptive statistics. Water consumption patterns irregular and unreliable, serving less than one-third of
were related to household size, income, and seasonal the population. During dry seasons, households
variations. For demand forecasting, selected machine experienced acute shortages and increased dependence
learning models (including regression and tree-based on private vendors.
algorithms) were evaluated using historical
consumption and explanatory variables such as Age Frequen
temperature, rainfall, and population. Group cy Percentage
<20 56 37.33333333
21-30 42 2
31-40 28 18.66666667

IRE 1714016 ICONIC RESEARCH AND ENGINEERING JOURNALS 8

 © FEB 2026 | IRE Journals | Volume 9 Issue 8 | ISSN: 2456-8880
DOI: [Link]
41-50 12 8 Water uses
51-60 9 6 Freq
61> 3 2 uenc Percentage (Based on
150 100 Water Usage y Data Set = 150 people)
Drinking 47 31.33333333
Cooking 110 73.33333333
Bathing 132 88
Household
Cleaning/Laun
dry 150 100

Water sources available

Water Frequenc
Source y Percentage

Well 110 73.33333333 The average daily water consumption by the residents
Stream/Riv in the area from the questionnaire analysis is 40litres
er 3 2 Water availability from source during dry season

Borehole 37 24.66666667 Freq

Rainwater 0 0 Water Availability from uenc
Total 150 100 source during dry season y Percentage

Yes 30 20

No 120 80

150 100

Water Consumption Patterns

Average daily household water consumption varied
significantly with household size and income level.
Higher-income households consumed more water due
to the use of water-intensive appliances, while lower-
income households relied on rationing and alternative
Demand forecasting
sources. Seasonal variations were evident, with higher
Well Yield
consumption during the dry season
Diameter of well= 40cm
At 6pm,water level was at a depth of 38cm
Volume of water =π x r² x h
= 3.142 x 40² x 38
=191,034cm³
=191litres

IRE 1714016 ICONIC RESEARCH AND ENGINEERING JOURNALS 9

 © FEB 2026 | IRE Journals | Volume 9 Issue 8 | ISSN: 2456-8880
DOI: [Link]
At 6am, water level in the well was at a depth of ALLOWANCE FOR SYSTEM LOSSES (NRW)
160cm, Assume 25% losses:
Volume of water =π x r² x h Qdesign=Qtotal × (1+0.25)
= 3.142 x 40² x 160 Qdesign=130.07×1.25= 162.59 ML/D
=804,352cm³ SOURCE ALLOCATION DESIGN
=805litres Surface Water Contribution (60%)
At 6pm,water level in the well was at a depth of 44cm Qsurface=0.60×162.59
Volume of water =π x r² x h Qsurface=97.56 ML/D
= 3.142 x 40² x 44 Groundwater Contribution (40%)
= 244litres Qground =0.40×162.59
Well recharge rate =
yield
=
(805−244)litres Qground = 65.04 ML/D
time 12
561
= = 46.75 MAXIMUM DAY AND PEAK HOUR DEMAND
12
= 47L/hr Maximum Day Demand (MDD)
Maximum Day Factor = 1.2
4.5 Design of Water Supply MDD = 1.2×Qdesign
1. Design period: 20 years (From 2026 – 2046) MDD=1.2×162.59 = 195.108 ML/D
2. Population growth rate: 2.5% per annum Peak Hour Demand (PHD)
3. Per capita water demand: 135 L/capita/day Peak Hour Factor = 1.8
4. Allowance for commercial, institutional & public PHD=(1.8×Qdesign)/24
use: 15% PHD=(1.8×162.59)/24= 12.2 ML/hr
5. Allowance for system losses (NRW): 25%
6. Supply system type: Hybrid surface + groundwater STORAGE RESERVOIR DESIGN
7. Storage provision: 35% of average daily demand Storage Requirement
8. Minimum supply hours target: 24 hours Provide 35% of average daily demand:
V=0.35×162.59
POPULATION PROJECTION V=56.91 ML
Base Population
Population Growth Formula DISTRIBUTION PIPE DESIGN
Exponential growth method: Pn = Po ert Design Flow Rate
P20=508,000e(0.025∗20) Q=162.59/(24×3600)
P20≈837,550 persons Q=1.88 m³/s
Pipe Diameter (Using Velocity Criterion)
AVERAGE DAILY WATER DEMAND Adopt: 1300 mm diameter ductile iron trunk main
Domestic Demand PRESSURE REQUIREMENTS
Qd=P20×q Minimum pressure: 10–15 m
Qd =837,550 persons ×135 L/capita/Day Maximum pressure: ≤60 m
Qd = 113,069,250 L/day= 113.1 ML/D Achieved through: Zonal reservoirs and Pressure
reducing valves
ALLOWANCE FOR COMMERCIAL &
INSTITUTIONAL USE PUMP CAPACITY (GROUNDWATER)
Qci=0.15×113.1 Assume:
Qci=16.97 ML/D • Borehole yield = 10 L/s
Total Demand Before Losses Number of boreholes required:
Qtotal=113.1+16.97 Q=65.04 ML/D
Qtotal=130.07 ML/D =752.8 L/s
N=752.8/10=75.28 boreholes
Provide:
• 85 boreholes (including standby)

IRE 1714016 ICONIC RESEARCH AND ENGINEERING JOURNALS 10

 © FEB 2026 | IRE Journals | Volume 9 Issue 8 | ISSN: 2456-8880
DOI: [Link]
IV. SUMMARY OF DESIGN OUTPUTS Pipe Sizing
Pipe Length Diameter Flow
Component Designed Value ID (m) (mm) (L/s)
Design P1 300 300 80
0.84 million P2 250 250 60
population
P3 180 200 45
Design demand 162.59 ML/D
P4 210 200 40
Treatment 97.56 ML/D surface + P5 150 160 30
capacity 65.04 ML/D groundwater
P6 190 160 28
Storage volume 56.91 ML P7 140 110 18
Trunk main 1300 mm P8 170 110 16
Boreholes ~85 P9 200 90 12
P10 220 90 10
Distribution Network
The distribution network was modelled using Head Loss Calculation
[Link] distribution system is designed as a All pipes were sized to ensure:
looped network (recommended for urban areas like • Head loss < 10 m/km
Ibadan North) to ensure: Pressure stability, Reliability • Pressure at nodes ≥ 15 m (minimum WHO
during pipe failure, Reduced head loss standard)
• Max pressure ≤ 60 m (to avoid pipe bursts)
System Components
Component Description Simulation Results (Expected Output)
Source Borehole Parameter Result Status
Reservoir Elevated service tank Min pressure 18 m Acceptable
Main pipes Trunk & distribution mains Max pressure 52 m Safe
Nodes Junctions at streets & zones Max velocity 2.1 m/s Acceptable
Valves Isolation & control Headloss < 8 m/km Good
Fire hydrants At 150–200 m spacing Supply coverage 100% Satisfactory

Parameter Value

Elevation 235 m

Storage capacity 6,000 m³

Min level 2m

Max level 6m

Pipe Network Design Model results: what the numbers actually mean
Pipe Material Having trained and tuned a number of various
• PVC / HDPE (recommended for Nigerian soil algorithms, I can now have a good idea of their relative
conditions) performance. The way they worked on the test set is as
• Hazen-Williams coefficient: C = 130 follows:
i. Linear Regression (The Baseline)
• Performance: MAE ~47.45L R² 0.78.
• My Notes: This was an acceptable starting point,
however, the R 2 of 0.78 is obviously below the
target. It did not fare well at picking up on the non-

IRE 1714016 ICONIC RESEARCH AND ENGINEERING JOURNALS 11

 © FEB 2026 | IRE Journals | Volume 9 Issue 8 | ISSN: 2456-8880
DOI: [Link]
linear trends I observed during EDA and in The performance comparison is summarized below:
particular when it was time to determine the effects
MA R²
of household size and income on demand. RMS MAP
Model E Scor
ii. Random Forest (The most successful) E (L) E
(L) e
• Performance: MAE ~5.48L R² 0.99.
• Best Params: 200 trees, Max Depth: 20. Linear
• My Observation: This model is way to be the 47.4 63.0 0.779 117.8
Regressi
winner. It is a terrific R 2 of 0.99 and an average 5 1 9 %
on
error ( MAE ) of only 5.5 Litres is truly exquisite
when it comes to such survey data. It did an Random
exquisite job of dynamic interplay between family 14.1 0.989 5.99
Forest 5.48
size, type of building, and amount of water people 0 0 %
(RF)
are actually paying.
iii. Support Vector Regressor (SVR) 11.2 27.8 0.957 17.9
• Performance: MAE ~11.20L | R² 0.96. SVR
0 6 0 %
• My Observation: SVR performed quite well, but it
was noticeably slower to tune and train than the ANN
other models. While a 0.96 R² is strong, it still (Neural 12.9 22.9 0.970 26.5
couldn't quite catch up to the Random Forest's Network 5 3 8 %
accuracy. )
iv. Artificial Neural Network (ANN)
• Performance: MAE ~12.95L | R² 0.97. Voting
19.0 28.7 0.954 42.1
• My Observation: I’m really happy with how the Ensembl
0 7 1 %
ANN turned out. Even though it's a bit of a "black e
box," the loss plot showed really quick
convergence. Thanks to the dropout layers and
The Random Forest model emerged as the superior
early stopping I set up, I'm not seeing any signs of
predictor for this tabular survey data. With an R² of
overfitting.
0.99, it effectively captured the complex, non-linear
v. Voting Ensemble
interactions that the Linear Regression baseline
• Performance: MAE ~19.00L | R² 0.95.
struggled to manage.
• My Observation: I actually expected the ensemble
to be my "ultimate" model by combining the CONCLUSION
strengths of RF, SVR, and LR, but it actually
performed worse than the Random Forest on its This\ study highlights the persistent challenges of
own. It’s balanced, but the weaker models (like urban water supply in Ibadan North LGA, where
Linear Regression) likely dragged down the high groundwater sources have become the primary means
precision of the RF. of meeting domestic demand due to unreliable public
supply. Water consumption is strongly influenced by
The Random Forest model is the clear winner for me. socio-economic and climatic factors, underscoring the
It offers the best mix of accuracy, processing speed, need for improved forecasting tools.
and interpretability. Because it handles this tabular
survey data so effectively, this is the model I’m The integration of machine learning-based demand
moving forward with for all my final forecasting and forecasting offers a practical approach for enhancing
scenario analysis. water resource planning in Ibadan and other rapidly
urbanizing cities. Adoption of such tools by water
authorities could improve infrastructure investment
decisions, reduce shortages, and support sustainable
urban development. Water utilities should adopt data-

IRE 1714016 ICONIC RESEARCH AND ENGINEERING JOURNALS 12

 © FEB 2026 | IRE Journals | Volume 9 Issue 8 | ISSN: 2456-8880
DOI: [Link]
driven demand forecasting tools to support planning
and operations. Investment in upgrading and
maintaining public water infrastructure is urgently
needed. Groundwater abstraction should be regulated
to prevent long-term depletion and quality
deterioration. Household-level water conservation
awareness programs should be promoted.

REFERENCES

[1] Adeleye, B. N., Osabuohien, E. S., & Bowale, E. I.

K. (2019). Water infrastructure and sustainable
development in Nigeria: The role of governance
and institutions. Utilities Policy, 58, 1–10.
[Link]
[2] Agbede, O. A. (1991). Engineering geology of
Nigeria. Rock View (Nig.) Ltd., Jos, Nigeria.
[3] Akintola, F. O. (1982). Hydrological atlas of
Nigeria. University of Ibadan Press, Ibadan,
Nigeria.
[4] Ayoade, J. O. (1988). Introduction to climatology
for the tropics. Spectrum Books Ltd., Ibadan,
Nigeria.
[5] Breiman, L. (2001). Random forests. Machine
Learning, 45(1), 5–32.
[Link]
[6] Cortes, C., & Vapnik, V. (1995). Support-vector
networks. Machine Learning, 20(3), 273–297.
[Link]
[7] Dalhuisen, J. M., Florax, R. J. G. M., de Groot, H.
L. F., & Nijkamp, P. (2003). Price and income
elasticities of residential water demand: A meta-
analysis. Land Economics, 79(2), 292–308.
[8] Donkor, E. A., Mazzuchi, T. A., Soyer, R., &
Roberson, J. A. (2014). Urban water demand
forecasting: Review of methods and models.
Journal of Water Resources Planning and
Management, 140(2), 146–159.
[9] Fetter, C. W. (2001). Applied hydrogeology (4th
ed.). Prentice Hall, Upper Saddle River, NJ,
USA.
[10] Ghimire, S., & Barkdoll, B. D. (2020). Issues in
municipal water demand forecasting and
analysis. Journal of Water Supply: Research and
Technology—AQUA, 69(2), 97–113.

IRE 1714016 ICONIC RESEARCH AND ENGINEERING JOURNALS 13