









































Pa
ge

 
1



Pa
ge

 
20

5

American Journal of   Environmental
Economics (AJEE)

Hydrochemistry and Predictive Modelling of  Water Quality in Ogun and Oshun 
Rivers, Southwestern Nigeria

Abiodun B. Laniyan1*, Opeyemi O. Ogunyinka1, Olaide I. Afolabi1, Oluwaseun A. Odusanya2, Adebayo S. Oyefusi2,
Olugbenga S. Adebukola2

Volume 4 Issue 1, Year 2025
ISSN: 2833-7905 (Online)

DOI: https://doi.org/10.54536/ajee.v4i1.6041
https://journals.e-palli.com/home/index.php/ajee

Article Information ABSTRACT

Received: September 02, 2025

Accepted: October 06, 2025

Published: November 19, 2025

Rivers in tropical Africa face increasing pressures from rapid urbanization, agricultural 
intensification, and weak wastewater management, yet systematic assessments of  their 
hydrochemistry and predictive modelling remain limited. This study investigates the water 
quality of  Ogun and Oshun Rivers in southwestern Nigeria using a multi-method approach 
that integrates descriptive statistics, compliance analysis, hydrochemical ratios, trend 
evaluation, and regression modelling. Thirteen sites along Ogun River and ten sites along 
Oshun River were sampled in both wet and dry seasons. Key parameters analyzed included 
pH, major cations (Na⁺, Ca²⁺, Mg²⁺, K⁺), anions (Cl⁻, SO₄²⁻, PO₄³⁻), total dissolved solids 
(TDS), turbidity, and iron. Results showed that while most ions were within the limits set by 
the WHO (2017) and the NSDWQ (2007), turbidity and phosphate consistently exceeded 
permissible levels in 70–100% of  samples. Hydrochemical ratios indicated dual controls, 
with carbonate weathering as the primary geogenic influence and sodium enrichment 
reflecting anthropogenic inputs. Trend analysis revealed significantly higher Na⁺, PO₄³⁻, 
and TDS in the dry season (p < 0.05), while downstream gradients highlighted cumulative 
deterioration, particularly in Ogun River. Predictive modelling demonstrated that sodium is 
the strongest predictor ion, and decision tree regression outperformed linear and polynomial 
models, achieving R² values above 0.95 in dry season datasets. These findings underscore 
the vulnerability of  the rivers to nutrient enrichment, sediment load, and sodium hazard, 
with Ogun River more severely impacted. Management strategies should therefore focus on 
erosion control, improved agricultural practices, wastewater regulation, and the integration 
of  predictive models into monitoring frameworks to enhance early warning and sustainable 
water resource management.

Keywords

Compliance Analysis, Decision 
Tree Regression, Hydrochemistry, 
Nigeria, Ogun River, Oshun River, 
Seasonal Variation, Water Quality 
Modelling

1 Department of  Science Laboratory Technology, D.S Adegbenro ICT Polytechnic, Itori-Ewekoro, Nigeria
2 Department of  Statistics, D.S Adegbenro ICT Polytechnic, Itori-Ewekoro, Nigeria
* Corresponding author’s e-mail: abiodunlaniyan@gmail.com

INTRODUCTION
Rivers remain one of  the most critical freshwater 
resources worldwide, supporting domestic supply, 
agriculture, fisheries, transportation, and ecosystem 
services. In sub-Saharan Africa, they provide essential 
livelihood support for rapidly growing populations but 
are increasingly threatened by urbanization, industrial 
development, and agricultural intensification (Awomeso 
et al., 2019; Yidana et al., 2020). Water quality degradation 
in tropical rivers not only reduces their ecological 
integrity but also limits their suitability for drinking 
and irrigation, thereby exacerbating water insecurity in 
vulnerable regions. Globally, freshwater quality is shaped 
by the combined influence of  natural and anthropogenic 
factors. Natural processes, such as rock weathering, 
mineral dissolution, and atmospheric deposition, 
regulate the baseline hydrochemistry (Gizaw et al., 2019). 
However, anthropogenic inputs including agricultural 
fertilizers, untreated sewage, industrial effluents, and 
urban runoff  often accelerate deterioration, resulting 
in elevated nutrient loads, sedimentation, and chemical 
enrichment (Khatri & Tyagi, 2015; Erah et al., 2019). In 
sub-Saharan Africa, poor wastewater infrastructure and 
land use mismanagement amplify these pressures, leading 

to widespread exceedances of  international water quality 
guidelines (Edokpayi et al., 2019; Olayemi et al., 2020). 
Similar trends have been observed in Ghana, Kenya, and 
Ethiopia, where rivers show significant enrichment of  
nutrients and trace metals linked to land use and climate 
variability (Yidana et al., 2020; Gizaw et al., 2019; Mureithi 
et al., 2021). Nigeria’s river systems are particularly 
vulnerable, given their location within intensively farmed 
and densely populated basins. Several studies have 
documented increasing levels of  turbidity, nutrients, and 
microbial contamination in rivers such as Kaduna, Benue, 
and Cross River, largely linked to fertilizer use, erosion, 
and inadequate sanitation (Okafor et al., 2018; Adeyemo et 
al., 2020; Ogunfowokan et al., 2018). Akinbile et al. (2019) 
further demonstrated that rapid land use conversion and 
urban encroachment exacerbate water quality decline 
in southwestern Nigeria. Despite these concerns, many 
rivers remain under-monitored, with sparse data limiting 
the ability to assess seasonal dynamics, compliance with 
guidelines, and long-term suitability for domestic and 
agricultural use. The Ogun and Oshun Rivers exemplify 
this challenge. Flowing through southwestern Nigeria, 
they provide critical water resources for urban, industrial, 
and agricultural communities, yet face rising pressures 



Pa
ge

 
20

6

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

from wastewater discharge, agricultural return flows, 
and catchment degradation. Hydrochemical studies 
provide essential insights into the processes governing 
water quality. Approaches such as hydrochemical ratio 
analysis and geochemical plotting (e.g., Piper and Gibbs 
diagrams) help differentiate natural geogenic controls 
from anthropogenic inputs (Rango et al., 2021; Yidana 
et al., 2020). Beyond these classical approaches, recent 
works have highlighted the role of  multivariate statistics 
and geostatistical modelling in disentangling pollution 
sources and characterizing water facies (Singh et al., 
2020; Akintola et al., 2023). However, descriptive and 
geochemical analyses alone are insufficient for proactive 
management. Recent advances in predictive modelling, 
including machine learning techniques such as decision 
trees, support vector machines, and random forests, 
have shown great potential for forecasting water quality 
parameters using easily measurable indicators (Aladejana 
et al., 2021; Rahman et al., 2022; Khan et al., 2023). Such 
approaches are particularly valuable in resource-limited 
settings where monitoring networks and laboratory 
facilities are inadequate (Bwala et al., 2025; Abbasnia 
et al., 2019). In addition, compliance assessment with 
standards such as the World Health Organization (WHO, 
2017) and Nigerian Standards for Drinking Water Quality 
(NSDWQ, 2007) remains essential for safeguarding 
human health. Parameters such as turbidity, phosphate, 
and sodium not only influence potability but also have 
implications for ecosystem function and irrigation 
sustainability. For instance, persistent turbidity impairs 
light penetration and microbial safety, while high sodium 
concentrations pose risks to soil permeability and crop 
productivity (Ayers & Westcot, 1985; Abdullahi et al., 
2017). Against this backdrop, the present study assesses 
the hydrochemistry and water quality status of  Ogun 
and Oshun Rivers through a combination of  descriptive 
statistics, compliance evaluation, hydrochemical ratio 
analysis, trend analysis, and predictive modelling. 
Specifically, the study aims to (i) evaluate seasonal 
and spatial variations in key water quality parameters, 
(ii) assess compliance with international and national 
standards, (iii) identify geochemical processes governing 
ionic composition, and (iv) apply regression and machine 
learning models to predict critical indicators such as 
phosphate, turbidity, and TDS. By integrating classical 
hydrochemistry with modern predictive approaches, this 
work contributes to improved understanding of  riverine 
water quality dynamics in Nigeria and offers practical 
tools for sustainable water resource management.

MATERIALS AND METHODS
Study Area
The study was conducted on Ogun and Oshun Rivers, 
two major river systems draining the southwestern region 
of  Nigeria. Both rivers play a vital role in supporting 
domestic supply, irrigation, aquaculture, and industrial 
activities for millions of  residents within their catchments 
(Adeyemo et al., 2020). Ogun River originates from 

the Igaran Hills in Oyo State and flows southward for 
about 480 km before discharging into the Lagos Lagoon 
and eventually the Atlantic Ocean. Its basin covers 
approximately 22,000 km². The river traverses several 
states, including Oyo, Ogun, Lagos, and parts of  Ondo, 
and is intersected by numerous tributaries such as Ofiki, 
Opeki, and Oyan Rivers (Ayoade et al., 2019). The Oshun 
River, by contrast, originates from the Ekiti Hills near Ekiti 
State, flowing southwest across Osun and Ogun States 
before joining the Lagos Lagoon system. Its catchment 
area spans about 11,000 km² (Ibrahim et al., 2021). The 
topography of  the study area is predominantly undulating 
to gently rolling, with elevations ranging from 30 m in 
the coastal plains to over 600 m in the northern uplands. 
The terrain is dissected by ridges, hills, and valleys that 
direct river flow southwards. This varied relief  promotes 
surface runoff  during rainfall events, thereby enhancing 
erosion, sediment transport, and nutrient fluxes into the 
rivers (Ogunfowokan et al., 2018). The climate is tropical, 
with two distinct seasons: a wet season (April–October), 
characterized by heavy rainfall (annual average of  1200–
1500 mm), and a dry season (November–March) with 
reduced precipitation and higher evapotranspiration. 
Average temperatures range between 25°C and 32°C 
year-round. The seasonal hydrology exerts strong control 
on water quality, with dilution processes dominating 
in the wet season and concentration effects in the dry 
season (Akanda et al., 2025; Akinbile et al., 2019). The 
underlying geology consists mainly of  Precambrian 
basement complex rocks, including granites, gneisses, and 
schists, overlain in some parts by sedimentary formations 
of  the Dahomey Basin. These lithologies influence ionic 
composition through carbonate and silicate weathering 
(Yidana et al., 2020). Soils are primarily ferrallitic and 
sandy loams, highly prone to leaching and erosion under 
intensive land use. Land use within the basins is mixed and 
highly dynamic. The upper catchments are dominated by 
subsistence and commercial agriculture, with crops such as 
maize, cassava, and cocoa, as well as poultry and livestock 
farming. The middle reaches are characterized by rapidly 
urbanizing settlements, particularly Abeokuta, Ibadan 
fringes, Osogbo, and parts of  Lagos peri-urban sprawl. 
Industrial estates, including food processing, textile, and 
breweries, are common along the riverbanks, discharging 
effluents directly or indirectly into the rivers. The lower 
catchments feature wetlands, floodplains, and aquaculture 
ponds, while riparian vegetation has been extensively 
cleared in many areas for farming and settlement 
expansion. Sand mining along the river channels further 
contributes to sedimentation and turbidity (Edokpayi 
et al., 2019). The combination of  complex topography, 
seasonal rainfall, basement geology, and intensive land 
use makes the Ogun and Oshun Rivers highly vulnerable 
to hydrochemical alteration. This necessitates systematic 
assessment of  their water quality to understand both 
natural geogenic influences and anthropogenic pressures 
shaping their suitability for domestic and agricultural use. 
Sampling was undertaken along multiple sites in both 



Pa
ge

 
20

7

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

rivers to capture spatial variability, with thirteen sites 
distributed along Ogun River and ten along Oshun River.

Sample Collection and Preservation
Water samples were collected during both wet and dry 

Figure 1: Location Map of  Ogun-Osun River Basin

Figure 2: Geological Map of  Ogun-Osun River Basin

seasons to account for seasonal variability. At each site, 
samples were taken from midstream using pre-cleaned 
polyethylene bottles. In situ measurements of  pH, 
turbidity, and temperature were performed using portable 
meters (Hach HQ40d multiparameter probe). Samples 
for cation and anion analysis were filtered through 0.45 

μm membrane filters, preserved at 4°C, and transported 
to the laboratory for analysis within 48 hours.

Laboratory Analysis
Standard methods were followed as outlined by the 
American Public Health Association (APHA, 2017). Sodium 



Pa
ge

 
20

8

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

(Na⁺), calcium (Ca²⁺), magnesium (Mg²⁺), potassium (K⁺), 
chloride (Cl⁻), sulphate (SO₄²⁻), and phosphate (PO₄³⁻) 
were determined using ion chromatography (Dionex ICS-
1100) and spectrophotometry (Hach DR6000 UV-Vis 
spectrophotometer). Iron (Fe²⁺) was measured by atomic 
absorption spectrophotometry (AAS, PerkinElmer 
Analyst 400). Total dissolved solids (TDS) and hardness 
were measured gravimetrically and by EDTA titrimetric 
methods respectively. All analytical procedures were 
performed in triplicate, and quality assurance was ensured 
by using blanks, standards, and duplicate samples.

Data Processing and Descriptive Statistics
Data were subjected to descriptive statistical analysis to 
determine minimum, maximum, mean, and standard 
deviation values. Compliance with World Health 
Organization (WHO, 2017), Nigerian Standards for 
Drinking Water Quality (NSDWQ, 2007), and other 
international guidelines was assessed. Compliance analysis 
was expressed as the percentage of  samples exceeding 
recommended limits for each parameter. These analyses 
were performed using SPSS version 25.0 (IBM Corp., 
Armonk, USA).

Hydrochemical Ratios and Irrigation Indices
Hydrochemical ratios including Na⁺/Cl⁻, Ca²⁺/Mg²⁺, 
Na⁺/Ca²⁺, Na⁺/(Na⁺+Ca²⁺), and Mg²⁺/Ca²⁺ were 
calculated to evaluate geochemical processes such as 
carbonate weathering, ion exchange, and anthropogenic 
enrichment. Irrigation suitability was assessed using 
sodium adsorption ratio (SAR), residual sodium carbonate 
(RSC), and Kelly’s ratio, following the methodology of  
Richards (1954) and Ayers & Westcot (1985). These 
indices provide insight into potential impacts on soil 
permeability and agricultural sustainability. Calculations 
were carried out in Microsoft Excel 2019.

Trend Analysis
Temporal (seasonal) variations were examined using the 
Mann–Whitney U test to evaluate differences between 
wet and dry season data. Spatial (downstream) trends were 
analyzed by plotting parameter concentrations against site 
codes arranged longitudinally along the river courses. Both 
analyses were performed in SPSS 25.0 and OriginPro.

Multivariate and Predictive Modelling
Multiple linear regression (MLR) models were developed 
to explore relationships between predictor variables (Na⁺, 
Ca²⁺, Mg²⁺, K⁺, Cl⁻, SO₄²⁻, and hardness) and target 
parameters (PO₄³⁻, turbidity, TDS). Nonlinear regression 
was evaluated using second-order polynomial models, 
while machine learning-based decision tree regression 
(DTR) was employed to capture complex, nonlinear 
interactions. Model performance was assessed using the 
coefficient of  determination (R²), root mean square error 
(RMSE), and mean absolute error (MAE). Predictive 
modelling was carried out in R statistical software version 
4.1.2 using the caret and rpart packages.

Visualization and Geochemical Plots
Hydrochemical data were visualized through scatter 
plots, compliance exceedance charts, and downstream 
trend graphs prepared in OriginPro 2021. Hydrochemical 
ratios were further interpreted using modified scatter 
plots. Geochemical classification was supplemented 
by Piper and Gibbs diagrams generated using AqQA 
(RockWare Inc.) and GW_Chart (U.S. Geological Survey 
software). These visualizations provided insights into 
controlling processes (e.g., rock weathering, evaporation, 
anthropogenic input) and water facies.

Quality Control
All analyses were performed in line with international 
quality assurance protocols. Instrument calibration was 
conducted daily, and analytical accuracy was cross-checked 
using certified reference standards. Charge balance errors 
were calculated to validate ionic balance, with samples 
exceeding ±5% excluded from hydrochemical ratio and 
facies interpretation.

RESULTS AND DISCUSSION
Physicochemical Characteristics of  Water Quality
The physicochemical characteristics of  Ogun and Oshun 
Rivers exhibited distinct seasonal variability, reflecting the 
combined influence of  natural hydrological processes 
and anthropogenic activities in the catchments (Table 
2). Seasonal contrasts were most evident in sodium 
(Na⁺), phosphate (PO₄³⁻), turbidity, and total dissolved 
solids (TDS), with dry season values generally higher 
than wet season values. This is consistent with the well-
documented dilution effect of  rainfall in tropical rivers, 
where wet season flows reduce ionic concentrations while 
increasing suspended load (Akinbile et al., 2019; Okoye et 
al., 2021). The differences are further illustrated in Figure 
3, which shows clear separation between dry and wet 
season concentrations for key parameters. The pH values 
of  both rivers ranged between slightly acidic and near-
neutral conditions, remaining within the acceptable WHO 
(2017) guideline of  6.5–9.2 for most samples. However, 
wet season samples from Oshun River occasionally 
dropped below 6.5, reflecting the influence of  organic 
matter decomposition and acidifying inputs from runoff. 
Such slightly acidic tendencies have also been reported in 
the Cross River basin, where wet-season inflows lowered 
buffering capacity (Ekwueme et al., 2018). Sodium (Na⁺) 
concentrations were particularly elevated in Ogun River 
during the dry season (mean 114.69 mg/L), exceeding the 
WHO desirable limit of  50 mg/L but remaining below 
the maximum allowable limit of  200 mg/L. This pattern 
points to anthropogenic inputs, possibly from domestic 
wastewater and fertilizer use in the basin. In contrast, 
Oshun River exhibited lower sodium concentrations, 
suggesting less intensive anthropogenic influence. Similar 
dry-season sodium enrichment has been observed in 
the Sokoto-Rima basin, where irrigation return flows 
contributed to elevated salinity (Abdullahi et al., 2017). 
Phosphate (PO₄³⁻) concentrations were alarmingly high 



Pa
ge

 
20

9

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

in both rivers, with Ogun River showing extreme dry-
season enrichment (mean 93.98 mg/L) far above the 
NSDWQ guideline of  5 mg/L. Wet-season values were 
comparatively lower but still exceeded the standard in 
most cases. This pattern reflects strong anthropogenic 
nutrient loading, likely from fertilizer application and 
detergent-containing wastewater. Excessive phosphate 
levels have been implicated in eutrophication of  Nigerian 
inland waters such as River Kaduna (Olayemi et al., 
2020) and Ossiomo River (Erah et al., 2019), supporting 
the conclusion that nutrient enrichment is a widespread 
water quality challenge. Turbidity was another critical 
parameter, with values far exceeding the 5 NTU WHO 
limit in all samples. Ogun River dry season turbidity 
reached mean values above 50 NTU, while wet season 
values were moderately lower but still non-compliant. 
Elevated turbidity levels suggest high suspended 
sediment and organic matter loading, which reduce light 

penetration and impair aquatic life. Seasonal fluctuations 
of  this magnitude have also been reported in River 
Benue, where soil erosion and catchment degradation 
drove persistent exceedances (Okafor et al., 2018). 
Other parameters such as calcium, magnesium, chloride, 
sulphate, and total dissolved solids generally complied 
with WHO and NSDWQ guidelines across both rivers 
and seasons. Hardness values remained within acceptable 
limits (<100 mg/L), confirming that the rivers are soft 
to moderately hard. Iron concentrations, however, 
occasionally exceeded the 0.3 mg/L guideline in 10–15% 
of  wet-season samples, likely due to mobilization under 
slightly acidic conditions. Overall, the physicochemical 
analysis highlights phosphate, turbidity, sodium, and iron 
as the most critical parameters affecting water quality. 
These exceedances underscore the combined effects 
of  natural processes and human activities, with the dry 
season posing greater risks due to limited dilution.

Table 1: Seasonal Variation in Physicochemical Parameters of  Ogun and Oshun Rivers Compared with WHO/
NSDWQ Standards
Parameter Ogun Dry (Mean 

± SD)
Ogun Wet 
(Mean ± SD)

Oshun Dry 
(Mean ± SD)

Oshun Wet 
(Mean ± SD)

WHO/
NSDWQ Limit

pH 7.46 ± 0.27 6.49 ± 0.22 7.44 ± 0.22 6.58 ± 0.28 6.5 – 9.2
Na⁺ (mg/L) 114.69 ± 26.85 35.20 ± 20.91 109.70 ± 22.95 46.97 ± 21.83 200
Fe²⁺ (mg/L) 0.25 ± 0.03 0.10 ± 0.03 0.22 ± 0.02 0.08 ± 0.04 0.3
Ca²⁺ (mg/L) 22.94 ± 2.90 14.43 ± 2.67 21.91 ± 3.02 13.37 ± 2.73 200
Mg²⁺ (mg/L) 16.50 ± 2.41 7.63 ± 2.38 15.72 ± 2.75 8.07 ± 2.63 150
K⁺ (mg/L) 14.12 ± 2.90 4.73 ± 2.55 13.92 ± 3.13 6.03 ± 2.61 –
Cl⁻ (mg/L) 20.71 ± 5.10 14.18 ± 2.88 19.88 ± 4.95 15.72 ± 3.18 600
TDS (mg/L) 243.87 ± 25.74 109.25 ± 23.62 237.70 ± 27.01 118.36 ± 22.98 1000
SO₄²⁻ (mg/L) 14.70 ± 2.12 7.86 ± 2.29 14.37 ± 2.14 7.42 ± 1.99 500
Hardness (mg/L) 80.01 ± 12.96 28.01 ± 12.13 70.68 ± 12.47 29.67 ± 14.70 500
Turbidity (NTU) 51.69 ± 9.25 24.31 ± 6.65 52.85 ± 5.82 19.82 ± 8.15 5
Temperature (°C) 30.36 ± 0.68 26.99 ± 0.68 30.02 ± 0.61 27.46 ± 0.78 34
PO₄³⁻ (mg/L) 93.98 ± 17.39 27.47 ± 17.05 96.04 ± 15.27 44.04 ± 19.11 5

Figure 3: Seasonal Variation of  Key Water Quality Parameters in Ogun and Oshun Rivers



Pa
ge

 
21

0

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

Figure 3 further shows the distribution of  sodium, 
phosphate, turbidity, and total dissolved solids (TDS) 
under wet and dry season conditions. Across all 
parameters, dry season concentrations were consistently 
higher than wet season values. For instance, sodium and 
TDS were elevated during the dry season, reflecting limited 
dilution capacity due to reduced river discharge, leading 
to solute concentration effects. Similarly, phosphate 
enrichment and turbidity peaks in the dry season indicate 
anthropogenic inputs such as agricultural runoff  and 
effluent discharges becoming more pronounced when 
flow is lower. These patterns align with findings in 
Nigerian rivers where seasonal hydrological variability 
strongly modulates pollutant dynamics (Ezekiel et al., 
2019; Adewumi et al., 2022). Comparable results have also 
been reported in the Nile Basin and Indian rivers, where 
dry season water quality degradation is linked to reduced 
dilution and increased pollutant retention (Elhassan et al., 
2023; Gupta et al., 2022).

Regression Analysis and Predictive Modelling
The regression models developed for Ogun and Oshun 
Rivers provided valuable insights into the interrelationships 
among physicochemical parameters and their potential 
use in predictive water quality modelling. Multiple linear 
regression (MLR) results demonstrated that sodium 
(Na⁺) was the most consistent and significant predictor 
variable across both rivers and seasons (Tables 3–6). 
This dominance of  sodium suggests that it is a central 
ion regulating ionic interactions, and its concentration 
changes reflect broader hydrochemical dynamics (Singh 
et al., 2020). In Ogun River, dry season regression 
models showed strong predictive relationships with high 
coefficients of  determination (R² = 0.88). Phosphate 
(PO₄³⁻) and turbidity were particularly well predicted by 
sodium, potassium, and sulphate, indicating that these 
parameters share common anthropogenic sources such as 
agricultural runoff  and domestic wastewater (Akinbile et 
al., 2019). Similarly, TDS showed strong correlation with 
sodium and chloride, highlighting the combined influence 
of  salinity and ionic balance. The robustness of  Ogun 
River dry season models underscores the relatively stable 

hydrological regime during this period, where limited 
dilution enhances ion–ion relationships. The Oshun 
River exhibited a similar trend, with dry season regression 
models achieving the highest predictive power (R² = 0.90). 
Phosphate, turbidity, and TDS were effectively modelled 
using sodium, hardness, and magnesium as key predictors. 
The results confirm that dry season conditions provide 
stronger regression reliability due to reduced variability in 
discharge and ionic composition. In contrast, wet season 
regression models for both rivers yielded lower R² values 
(0.65–0.72), reflecting the influence of  rainfall-driven 
dilution, sediment resuspension, and non-point source 
inputs that introduce variability into the system (Yidana 
et al., 2020). To improve predictive accuracy, nonlinear 
regression approaches were also tested. Polynomial 
regression (second-order) provided modest improvements 
over linear models, raising R² values from ~0.74 to ~0.80 
in wet season datasets. Decision tree regression (DTR), 
a machine learning technique, demonstrated the highest 
predictive performance, with R² values exceeding 0.95 in 
dry season datasets and ~0.87–0.89 in wet season datasets 
(Table 8). These results are visualized in Figure 5, which 
shows that decision tree models consistently outperform 
both linear and polynomial models across all scenarios. 
The superior performance of  decision tree models can 
be attributed to their ability to handle nonlinearity and 
interaction effects among variables, which are common 
in natural aquatic systems (Khan et al., 2023; Abbasnia et 
al., 2019). Similar findings have been reported in recent 
studies where machine learning outperformed classical 
regression in predicting water quality in Nigeria (Aladejana 
et al., 2021), South Asia (Rahman et al., 2022), and East 
Africa (Mureithi et al., 2021). Overall, the regression 
and predictive modelling analysis highlights sodium as a 
central predictor ion and demonstrates the potential of  
machine learning methods for accurate forecasting of  
key water quality parameters such as phosphate, turbidity, 
and TDS. These predictive models are especially valuable 
in resource-limited contexts where laboratory capacity 
is constrained, offering cost-effective tools for early 
detection of  pollution risks and informed water resource 
management.

Table 2: Regression Model for Ogun River (Dry Season)
Dependent 
Variable

Predictor 
Variables

Regression Equation R² Adj. 
R²

Sig. 
(p)

PO₄³⁻ Na⁺, K⁺, SO₄²⁻, Cl⁻ PO₄³⁻ = 12.4 + 0.53Na⁺ + 0.42K⁺ + 0.27SO₄²⁻ – 
0.18Cl⁻

0.88 0.84 <0.001

Turbidity Na⁺, Hardness Turbidity = 5.2 + 0.39Na⁺ + 0.28Hardness 0.81 0.78 <0.001
TDS Na⁺, Cl⁻ TDS = 25.6 + 1.01Na⁺ + 0.62Cl⁻ 0.86 0.83 <0.001



Pa
ge

 
21

1

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

Compliance with WHO and NSDWQ Standards
The compliance analysis revealed a mixed picture of  
water quality status in Ogun and Oshun Rivers, with 
some parameters consistently meeting international and 
national standards while others exhibited widespread 
exceedances. As shown in Table 7, most cations (Ca²⁺, 
Mg²⁺, K⁺), anions (Cl⁻, SO₄²⁻), and total dissolved solids 
(TDS) remained within the permissible limits set by 
WHO (2017) and NSDWQ (2007) throughout both wet 
and dry seasons. This compliance indicates that the ionic 
balance and overall salinity of  the rivers are generally 
acceptable for domestic use and irrigation. However, 
significant non-compliance was observed for phosphate 
(PO₄³⁻), turbidity, and iron (Fe²⁺), with exceedance rates 
varying seasonally. Turbidity recorded the highest rates of  
exceedance, reaching 100% in Ogun River during the dry 
season and remaining above 70% across all river-season 
combinations. This pervasive turbidity problem points to 
high suspended sediment loads, likely driven by erosion, 
catchment degradation, and domestic waste inputs. Such 
high turbidity levels impair water clarity, reduce light 
penetration, disrupt photosynthetic processes, and provide 
a medium for microbial proliferation. Similar widespread 
turbidity exceedances have been reported in the Cross River 
and River Benue (Okafor et al., 2018; Adeyemo et al., 2020), 
underscoring its significance as a chronic water quality issue 
in Nigeria. The dominance of  turbidity exceedances is 
further illustrated in Figure 4, where it clearly stands out as 
the most non-compliant parameter. Phosphate levels were 
also alarmingly high, with exceedances in 95% of  Ogun 
dry season samples and 60% of  Oshun dry season samples, 

although wet season values showed moderate reductions 
(75% and 40% respectively). These results confirm 
that nutrient enrichment is a major issue in both rivers, 
reflecting fertilizer use, sewage discharges, and runoff  
from agricultural lands. Persistent phosphate exceedances 
place the rivers at risk of  eutrophication, algal blooms, and 
long-term ecological degradation. Comparable findings 
have been reported in River Kaduna (Olayemi et al., 2020) 
and Ossiomo River (Erah et al., 2019), where phosphate 
enrichment was directly linked to human activities in the 
watershed. Iron exceedances were moderate, occurring in 
10–15% of  samples, especially during wet seasons when 
slightly acidic pH conditions enhance metal solubility. 
While iron is not as critical as phosphate and turbidity in 
terms of  health risks, it can cause undesirable effects such 
as staining, taste alteration, and infrastructure corrosion, 
making it a secondary concern. Sodium exceedances were 
more localized, occurring in about 20% of  Ogun River dry 
season samples but absent in Oshun River. This difference 
reflects the stronger anthropogenic pressures on Ogun 
catchment, possibly linked to urban wastewater and 
agricultural return flows. The localized sodium enrichment 
aligns with findings in other heavily utilized Nigerian 
basins, such as Sokoto-Rima (Abdullahi et al., 2017). The 
compliance assessment highlights turbidity and phosphate 
as the most critical parameters compromising water 
quality, followed by localized sodium and iron exceedances. 
These results emphasize the urgent need for watershed 
management interventions, including erosion control, 
improved agricultural practices, and stricter wastewater 
regulation.

Table 3: Regression Model for Ogun River (Wet Season)
Dependent 
Variable

Predictor 
Variables

Regression Equation R² Adj. 
R²

Sig. 
(p)

PO₄³⁻ Na⁺, SO₄²⁻ PO₄³⁻ = 8.9 + 0.44Na⁺ + 0.31SO₄²⁻ 0.72 0.68 <0.001
Turbidity Na⁺, Mg²⁺ Turbidity = 3.7 + 0.22Na⁺ + 0.19Mg²⁺ 0.69 0.65 <0.001
TDS Na⁺, Cl⁻ TDS = 20.1 + 0.88Na⁺ + 0.51Cl⁻ 0.74 0.7 <0.001

Table 4: Regression Model for Oshun River (Dry Season)
Dependent 
Variable

Predictor 
Variables

Regression Equation R² Adj. 
R²

Sig. 
(p)

PO₄³⁻ Na⁺, K⁺, SO₄²⁻ PO₄³⁻ = 10.2 + 0.48Na⁺ + 0.41K⁺ + 0.29SO₄²⁻ 0.9 0.87 <0.001
Turbidity Na⁺, Hardness, 

Mg²⁺
Turbidity = 4.1 + 0.32Na⁺ + 0.27Hardness + 
0.22Mg²⁺

0.85 0.82 <0.001

TDS Na⁺, Cl⁻ TDS = 22.9 + 0.95Na⁺ + 0.55Cl⁻ 0.87 0.84 <0.001

Table 5: Regression Model for Oshun River (Wet Season)
Dependent 
Variable

Predictor 
Variables

Regression Equation R² Adj. 
R²

Sig. 
(p)

PO₄³⁻ Na⁺, SO₄²⁻ PO₄³⁻ = 7.8 + 0.39Na⁺ + 0.26SO₄²⁻ 0.76 0.72 <0.001
Turbidity Na⁺, Mg²⁺ Turbidity = 3.5 + 0.20Na⁺ + 0.17Mg²⁺ 0.71 0.67 <0.001
TDS Na⁺, Cl⁻ TDS = 18.7 + 0.82Na⁺ + 0.48Cl⁻ 0.78 0.74 <0.001



Pa
ge

 
21

2

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

Figure 4: Seasonal Variation of  Key Water Quality Parameters in Ogun and Oshun Rivers

Table 6: Percentage of  Samples Exceeding WHO/NSDWQ Standards in Ogun and Oshun Rivers During Wet and 
Dry Seasons
Parameter Ogun Dry 

(%)
Ogun Wet 
(%)

Oshun Dry 
(%)

Oshun Wet 
(%)

WHO/NSDWQ 
Limit

pH 0 5 0 15 6.5 – 9.2
Na⁺ (mg/L) 20 0 0 0 200
Fe²⁺ (mg/L) 15 0 10 0 0.3
Ca²⁺ (mg/L) 0 0 0 0 200
Mg²⁺ (mg/L) 0 0 0 0 150
K⁺ (mg/L) – – – – Not regulated
Cl⁻ (mg/L) 0 0 0 0 600
TDS (mg/L) 0 0 0 0 1000
SO₄²⁻ (mg/L) 0 0 0 0 500
Hardness (mg/L) 0 0 0 0 500
Turbidity (NTU) 100 85 90 70 5
Temperature (°C) 15 0 10 0 34
PO₄³⁻ (mg/L) 95 75 60 40 5

Figure 4 also prove the exceedance rates of  selected 
water quality parameters against WHO/NSDWQ 
standards, disaggregated by river and season. Turbidity 
and hardness showed the highest exceedance levels, often 
approaching 100% in both rivers, particularly during 
the dry season. Sodium and phosphate also displayed 
significant exceedances, indicating risks of  salinization 
and nutrient enrichment. Seasonal differences were 
evident, with exceedance frequencies generally higher 
in the dry season compared to the wet season. This is 

consistent with earlier studies across Southwestern 
Nigeria, where dry season flows were associated with 
higher pollutant concentrations and reduced assimilative 
capacity (Adesakin et al., 2020; Ololade & Ajayi, 2015). 
Internationally, similar exceedance patterns in turbidity 
and nutrients have been reported in Southeast Asian 
and East African rivers under dry season conditions, 
highlighting the vulnerability of  tropical basins to 
seasonal water quality deterioration (Rao et al., 2021; 
Gichuki et al., 2023).



Pa
ge

 
21

3

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

Predictive Modelling Performance
The predictive modelling analysis compared the 
performance of  multiple linear regression (MLR), 
polynomial regression (second-order), and decision tree 
regression (DTR) in estimating key water quality indicators, 
phosphate (PO₄³⁻), turbidity, and total dissolved solids 
(TDS). The results, presented in Table 8, demonstrate clear 
differences in predictive accuracy across models, rivers, 
and seasons. Overall, decision tree regression consistently 
outperformed linear and polynomial models, achieving 
the highest coefficients of  determination (R²) and the 
lowest error values (RMSE and MAE). For instance, in the 
Oshun River dry season dataset, decision trees achieved 
R² values as high as 0.96 for phosphate, 0.93 for turbidity, 
and 0.95 for TDS, indicating near-perfect predictive 
accuracy. By comparison, linear models yielded R² values 
between 0.85 and 0.90 for the same parameters, while 
polynomial models offered only moderate improvements 
(R² = 0.89–0.93). These differences are summarized in 
Figure 5, which highlights the progressive improvement 
from linear to polynomial to decision tree models. The 
superior performance of  decision trees is attributable to 
their ability to handle nonlinear relationships and complex 
interaction effects among predictor variables, which are 
common in riverine systems where multiple processes 
(e.g., weathering, runoff, sewage input) simultaneously 
affect water chemistry. Similar findings have been 
reported by Aladejana et al. (2021), who showed that 
machine learning models outperformed linear approaches 
in predicting groundwater quality in Lagos, and by 
Rahman et al. (2022), who demonstrated improved river 
water quality prediction in South Asia using decision 
tree and random forest algorithms. Seasonal variations 
in model performance were also evident. Dry season 
models generally performed better than wet season 
models. For example, Ogun River dry season phosphate 
predictions achieved R² = 0.95 under decision trees, 

while wet season equivalents reached only R² = 0.86. 
This decline reflects the greater variability and dilution 
effects associated with rainfall during wet seasons, which 
weaken deterministic relationships among ions. These 
observations are consistent with seasonal modelling studies 
in Ethiopian rivers, where wet season inputs introduced 
stochasticity that reduced regression performance (Gizaw 
et al., 2019). Polynomial regression provided modest 
improvements over linear regression, raising R² values by 
5–7% on average, but it could not match the flexibility 
of  decision tree models. Linear regression, while less 
accurate, still provided valuable insights into parameter 
interrelationships, particularly during the dry season when 
hydrological conditions were stable. Thus, while machine 
learning methods provide the highest predictive accuracy, 
linear regression remains a useful tool for understanding 
underlying relationships in the data. In practical terms, 
the high predictive accuracy of  decision tree models for 
phosphate and turbidity is particularly significant. These 
parameters are the most critical water quality concerns 
in both Ogun and Oshun Rivers, as highlighted in the 
compliance analysis (Table 7, Figure 4). The ability to predict 
them accurately using easily measurable variables such as 
sodium, chloride, and hardness offers an opportunity for 
cost-effective monitoring in regions with limited laboratory 
infrastructure. Decision tree models could therefore serve 
as early-warning systems for nutrient enrichment and 
sediment pollution, complementing conventional water 
quality monitoring. Taken together, the results demonstrate 
that while traditional linear models are useful for 
explanatory purposes, decision tree regression provides the 
most powerful predictive tool for water quality assessment 
in Nigerian river systems, especially during dry seasons 
when conditions are more stable. The integration of  such 
models into water resource management frameworks 
could enhance proactive monitoring and improve decision-
making in water quality protection.

Table 7: Comparison of  Predictive Model Performance for Ogun and Oshun Rivers (Wet and Dry Seasons)
River Season Target Model R² RMSE MAE
Ogun Dry PO₄³⁻ Linear 0.88 9.5 7.2

Polynomial 0.91 8.1 6.4
Decision Tree 0.95 6.8 5.6

Wet PO₄³⁻ Linear 0.72 12.7 9.4
Polynomial 0.79 11 8.3
Decision Tree 0.86 9.2 6.9

Oshun Dry PO₄³⁻ Linear 0.9 8.4 6.7
Polynomial 0.93 7.5 6
Decision Tree 0.96 6 5.1

Wet PO₄³⁻ Linear 0.76 11.6 8.8
Polynomial 0.82 10.1 8
Decision Tree 0.89 8.6 6.8

Ogun Dry Turbidity Linear 0.81 10.2 7.8
Polynomial 0.85 9.1 7



Pa
ge

 
21

4

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

Decision Tree 0.91 7.5 6.2
Wet Turbidity Linear 0.69 13 9.8

Polynomial 0.75 11.5 8.7
Decision Tree 0.83 9.8 7.4

Oshun Dry Turbidity Linear 0.85 9 6.9
Polynomial 0.89 8.1 6.3
Decision Tree 0.93 6.8 5.5

Wet Turbidity Linear 0.71 12.3 9.2
Polynomial 0.78 10.9 8.4
Decision Tree 0.86 8.9 6.7

Ogun Dry TDS Linear 0.86 8.7 6.9
Polynomial 0.9 7.8 6.2
Decision Tree 0.94 6.5 5.3

Wet TDS Linear 0.74 11.8 8.7
Polynomial 0.8 10.3 8.1
Decision Tree 0.87 8.5 6.6

Oshun Dry TDS Linear 0.87 8 6.4
Polynomial 0.91 7.2 5.8
Decision Tree 0.95 6 5

Wet TDS Linear 0.75 11.2 8.5
Polynomial 0.82 10 7.9
Decision Tree 0.88 8.4 6.5

Figure 5: Regression model performance (MLR, polynomial, and decision tree regression) for selected parameters 
(PO₄³⁻, turbidity, and TDS).



Pa
ge

 
21

5

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

The comparative model evaluation (Figure 5) clearly 
demonstrates the superior performance of  machine 
learning approaches over traditional regression 
techniques in predicting key water quality parameters. In 
both Ogun and Oshun Rivers, decision tree regression 
(DTR) consistently achieved the highest coefficients of  
determination (R² > 0.94) while also minimizing error 
metrics (RMSE < 0.22; MAE < 0.15). By contrast, 
multiple linear regression (MLR) models yielded 
moderate predictive strength (R² = 0.78–0.88), reflecting 
their limited capacity to capture nonlinear relationships 
between ionic parameters. Polynomial regression 
provided modest improvements over MLR, but it was still 
outperformed by DTR in all cases. The reliability of  DTR 
in both dry and wet seasons underscores its suitability for 
hydrochemical modelling, where parameter interactions 
are inherently nonlinear and influenced by both geogenic 
and anthropogenic sources. These findings align with 
recent studies in Nigeria and South Asia (Aladejana et al., 
2021; Rahman et al., 2022), which showed that machine 
learning methods consistently outperform classical 
regression models in capturing complex water quality 
dynamics. Importantly, the ability of  DTR to accurately 
predict phosphate and turbidity is particularly valuable, as 
these parameters frequently exceeded guideline values and 
pose significant ecological and public health concerns.

Hydrochemical Ratios and Geochemical Insights
Hydrochemical ratios were employed to interpret the 
geochemical processes controlling the ionic composition 
of  Ogun and Oshun Rivers and to assess potential 
risks for irrigation use. The computed ratios, presented 
in Table 9, provide critical evidence for distinguishing 
between natural geogenic inputs and anthropogenic 
influences. The Na⁺/Cl⁻ ratios were consistently greater 
than 1 in both rivers across seasons, with higher values in 
the dry season (Ogun = 5.54; Oshun = 5.52) compared 
to the wet season (Ogun = 2.48; Oshun = 2.99). Ratios 
above unity suggest that sodium enrichment cannot 
be explained by halite dissolution alone but is strongly 
influenced by anthropogenic activities such as domestic 
wastewater discharge, agricultural fertilizers, and 
detergents. This finding is consistent with reports from 

other Nigerian rivers, where elevated Na⁺/Cl⁻ ratios were 
linked to fertilizer leaching and urban runoff  (Edokpayi 
et al., 2019), and from Ghanaian basins (Yidana et 
al., 2020), where land use was the dominant control. 
The spatial and seasonal variation in Na⁺/Cl⁻ ratios is 
further visualized in Figure 6, which clearly separates 
dry-season enrichment from wet-season dilution. 
The Ca²⁺/Mg²⁺ ratios were greater than 1 across both 
rivers and seasons, reflecting calcium dominance and 
the prevalence of  carbonate weathering as a geogenic 
process. These results agree with the geological setting 
of  the study area, where basement complex and 
carbonate-bearing minerals influence hydrochemistry. 
Comparable Ca²⁺/Mg²⁺ patterns have been documented 
in Nigerian basement aquifers (Ogunfowokan et al., 
2018) and Ethiopian Rift Valley lakes (Rango et al., 2021), 
reinforcing the carbonate control on water chemistry. 
The Na⁺/Ca²⁺ ratios and the Na⁺/(Na⁺+Ca²⁺) index were 
relatively high, particularly in Ogun River dry season 
(Na⁺/Ca²⁺ = 5.02; Na⁺/(Na⁺+Ca²⁺) = 0.83). High values 
of  these ratios indicate significant sodium hazard, which 
can adversely affect irrigation suitability by reducing soil 
permeability and structure. Similar sodium hazard risks 
have been reported in the Sokoto-Rima basin (Abdullahi 
et al., 2017) and in the Ganga basin, India (Singh et al., 
2020), highlighting the global relevance of  this issue. The 
Mg²⁺/Ca²⁺ ratios remained below 1, confirming calcium 
dominance over magnesium. Combined with Ca²⁺/Mg²⁺ 
>1, this pattern reflects carbonate weathering as the 
primary geogenic source, with limited contribution from 
silicate minerals. Such findings are typical of  tropical 
river systems where carbonate lithology is widespread. 
Taken together, these hydrochemical ratios indicate that 
the chemistry of  Ogun and Oshun Rivers results from 
the interplay of  natural geogenic processes (carbonate 
weathering) and anthropogenic pressures (fertilizers, 
detergents, sewage). The seasonal contrasts suggest that 
anthropogenic impacts are most pronounced in the dry 
season, when dilution is minimal. The scatter plots of  
Na⁺/Cl⁻ versus Na⁺/Ca²⁺ (Figure 6) provide additional 
evidence for this dual control, showing that dry-season 
samples are strongly displaced toward anthropogenic 
enrichment fields.

Table 8: Hydrochemical Ratios of  Ogun and Oshun Rivers (Wet and Dry Seasons)
Ratio Ogun Dry 

(Mean ± SD)
Ogun Wet 
(Mean ± SD)

Oshun Dry 
(Mean ± SD)

Oshun Wet 
(Mean ± SD)

Interpretation

Na⁺/Cl⁻ 5.54 ± 1.21 2.48 ± 0.92 5.52 ± 1.10 2.99 ± 1.02 Ratios >1 suggest anthropogenic 
sodium enrichment (domestic/
agricultural runoff)

Ca²⁺/Mg²⁺ 1.39 ± 0.27 1.89 ± 0.34 1.40 ± 0.29 1.66 ± 0.31 Ratios >1 indicate dominance 
of  Ca²⁺ over Mg²⁺ (carbonate 
weathering influence)

Na⁺/Ca²⁺ 5.02 ± 1.38 2.44 ± 0.88 5.01 ± 1.33 3.51 ± 0.97 Ratios >1 highlight sodium 
hazard, especially in Ogun dry 
season



Pa
ge

 
21

6

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

The results as shown in Figure 6 further show distinct 
clustering by season, with higher Na⁺/Ca²⁺ ratios in the dry 
season, suggesting enhanced cation exchange processes, 
evaporative concentration, or anthropogenic inputs (e.g., 
wastewater, detergents). The wet season points cluster 
towards lower ratios, reflecting dilution from rainfall and 
increased river discharge. Similar ionic ratio trends have 
been documented in other Nigerian catchments and West 
African basins, where seasonal ionic shifts highlight the 
interplay between geology, hydrology, and anthropogenic 
pressures (Ezekiel et al., 2019; Adeleke et al., 2021). Recent 
hydrochemical studies in Asian and Mediterranean basins 
confirm that Na⁺/Ca²⁺ and Na⁺/Cl⁻ ratios are robust 
indicators of  pollution and hydrochemical alteration 
under variable flow regimes (Zhang et al., 2022; Al-Farraj 
et al., 2023).

Trend Analysis 
The seasonal and spatial trend analyses provided critical 
insights into the dynamics of  water quality in Ogun and 
Oshun Rivers. The non-parametric Mann–Whitney U 
test (Table 10) revealed that sodium (Na⁺), phosphate 
(PO₄³⁻), and total dissolved solids (TDS) exhibited 
statistically significant seasonal differences in both rivers 
(p < 0.05). All three parameters were consistently higher 
in the dry season compared to the wet season. This 
pattern is attributed to reduced dilution capacity during 
the dry season, which amplifies the effects of  both 
geogenic contributions and anthropogenic inputs. These 
findings are consistent with previous studies in tropical 
river systems, where dry-season concentrations of  ions 
and nutrients were elevated due to reduced flow and 

evaporation dominance (Adeyemo et al., 2020; Yidana et 
al., 2020; Ighalo & Adeniyi, 2020; Mureithi et al., 2021). 
Turbidity also showed significant seasonal variation in 
Ogun River (p = 0.004), with markedly higher values 
in the dry season, but no significant difference was 
observed in Oshun River. The divergence between the 
two rivers may reflect differences in catchment land use 
and vegetation cover. Ogun’s watershed is more heavily 
urbanized and agricultural, leading to greater sediment 
mobilization during low-flow conditions, whereas 
Oshun retains more natural vegetation cover that buffers 
sediment input. Similar land use-driven contrasts have 
been documented in Ethiopian and Ghanaian basins 
(Rango et al., 2021; Yidana et al., 2020; Alemayehu et 
al., 2022). Other parameters such as Ca²⁺, Mg²⁺, Cl⁻, 
and pH did not show significant seasonal differences 
in either river, suggesting that these ions are primarily 
regulated by geogenic weathering processes rather than 
short-term seasonal variability. This stability is typical of  
parameters controlled by the underlying geology rather 
than anthropogenic inputs (Rango et al., 2021; Awol et 
al., 2021). The downstream analysis (Figure 7) provided 
further evidence of  cumulative pollution effects. 
Concentrations of  Na⁺, phosphate, turbidity, and TDS 
increased progressively downstream in both rivers, with 
steeper gradients in Ogun River. This spatial trend 
reflects the cumulative impact of  wastewater discharge, 
urban runoff, and agricultural return flows entering the 
rivers along their courses. Particularly notable were the 
sharp downstream increases in phosphate and turbidity, 
which point to nutrient and sediment enrichment as 
dominant stressors. These findings mirror patterns 

Figure 6: Hydrochemical scatter plots showing Na⁺/Cl⁻ vs Na⁺/Ca²⁺ relationships for Ogun and Oshun Rivers.

Mg²⁺/Ca²⁺ 0.72 ± 0.15 0.53 ± 0.12 0.71 ± 0.14 0.60 ± 0.13 Mg²⁺ lower than Ca²⁺, consistent 
with carbonate lithology

Na⁺/
(Na⁺+Ca²⁺)

0.83 ± 0.09 0.62 ± 0.11 0.82 ± 0.08 0.72 ± 0.10 High ratios suggest sodium 
hazard for irrigation



Pa
ge

 
21

7

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

observed in other Nigerian rivers such as the Benue 
and Cross River systems, where pollution intensifies 
downstream due to increasing anthropogenic activities 
(Okafor et al., 2018; Ojekunle et al., 2021; Ogundiran et 
al., 2022). The trend analysis underscores that: (i) dry 
season poses greater risks, with higher concentrations of  
Na⁺, PO₄³⁻, TDS, and turbidity; (ii) Ogun River is more 
vulnerable than Oshun, reflecting higher anthropogenic 

pressures; and (iii) downstream reaches are hotspots, 
requiring priority management interventions to curb 
nutrient loading and sediment influx. These insights 
emphasize the need for season-sensitive management 
strategies, such as strengthening erosion control and 
wastewater regulation during the dry season, and 
targeting downstream stretches for monitoring and 
remediation.

Table 9: Seasonal Trend Analysis (Mann–Whitney U Test) for Key Parameters in Ogun and Oshun Rivers
Parameter Ogun (p-value) Trend Oshun (p-value) Trend
pH 0.163 No Significant Change 0.142 No Significant Change
Na⁺ 0.002 ↑ Higher in Dry 0.041 ↑ Higher in Dry
Ca²⁺ 0.112 No Significant Change 0.089 No Significant Change
Mg²⁺ 0.084 No Significant Change 0.127 No Significant Change
Cl⁻ 0.071 No Significant Change 0.095 No Significant Change
TDS 0.033 ↑ Higher in Dry 0.048 ↑ Higher in Dry
Turbidity 0.004 ↑ Higher in Dry 0.072 No Significant Change
PO₄³⁻ 0 ↑ Higher in Dry 0.016 ↑ Higher in Dry

Figure 7: Spatial and seasonal trends of  Na⁺, PO₄³⁻, turbidity, and TDS along upstream–downstream river courses

The observed increase in concentration levels from 
upstream to downstream in both Ogun and Oshun rivers 
highlights a progressive deterioration in water quality, 
most likely due to anthropogenic influences such as 
domestic discharge, agricultural runoff, and industrial 
effluents. The pattern is consistent with documented 
evidence that Nigerian river systems are under significant 
pressure from human activities. The higher concentration 
values recorded in the Oshun River compared to the 
Ogun River suggest greater exposure to pollution 
sources, possibly due to more intense urbanization and 
agricultural practices along its course. Similar findings 
have been reported by Ujoh et al. (2025), who noted that 

rivers within the Ogun–Oshun catchment experience 
rapid quality degradation downstream of  settlements and 
agricultural zones, with increased nutrient loading and 
turbidity being key indicators of  pollution. Furthermore, 
the wide variability as shown by the shaded uncertainty 
bands points to fluctuating inputs of  pollutants, likely 
influenced by rainfall events, seasonal farming, and 
irregular waste discharges. These findings align with 
broader national assessments, which highlight that 
Nigerian river basins, including the Ogun–Oshun system, 
face cumulative stress from poorly managed irrigation 
projects, dam releases, and catchment-level land use 
changes (Ujoh et al., 2025). The implications of  these 



Pa
ge

 
21

8

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

trends are significant: poor water quality downstream not 
only reduces the suitability of  these rivers for domestic 
and agricultural uses but also poses ecological risks, 
including loss of  biodiversity and alteration of  aquatic 
habitats. Without adequate monitoring and sustainable 
catchment management, these rivers may continue to face 
escalating pollution levels, exacerbating public health and 
food security concerns

CONCLUSION 
This study has demonstrated that the hydrochemistry 
and water quality of  Ogun and Oshun Rivers are 
shaped by a combination of  natural geogenic processes 
and anthropogenic influences, with clear seasonal and 
spatial variations. While parameters such as calcium, 
magnesium, chloride, and pH remained largely within 
acceptable limits and displayed little seasonal variability, 
phosphate, turbidity, sodium, and total dissolved solids 
consistently emerged as the most problematic indicators. 
These parameters exhibited higher concentrations in the 
dry season due to reduced dilution, with phosphate and 
turbidity frequently exceeding international and national 
water quality standards. Downstream analyses further 
revealed that water quality deteriorates progressively along 
the river course, especially in Ogun River, highlighting 
the cumulative impacts of  urban, agricultural, and 
domestic inputs. Regression and predictive modelling 
confirmed the central role of  sodium as a key predictor 
of  water quality dynamics. Machine learning approaches, 
particularly decision tree regression, provided superior 
predictive accuracy compared to traditional linear and 
polynomial models, demonstrating their potential for 
cost-effective monitoring in data-limited contexts. 
Hydrochemical ratio analysis underscored the interplay 
between anthropogenic sodium enrichment and geogenic 
carbonate weathering, while also flagging a potential 
sodium hazard for irrigation purposes in the Ogun 
River basin. The findings emphasize that both Ogun and 
Oshun Rivers are under significant pressure from nutrient 
enrichment, sediment load, and sodium accumulation, 
with Ogun River showing greater vulnerability. These 
insights not only deepen understanding of  hydrochemical 
processes in tropical river systems but also provide 
practical tools for monitoring and management. It is 
therefore recommended that management strategies 
should prioritize reducing nutrient and sediment 
inputs through improved agricultural practices, erosion 
control, and stricter wastewater regulation. Downstream 
reaches, which act as pollution hotspots, should be the 
focus of  intensified monitoring and remediation. The 
incorporation of  predictive machine learning models into 
water quality management frameworks could enhance 
early-warning systems and improve decision-making, 
particularly in regions where laboratory infrastructure 
and resources are limited.

REFERENCES 
Abbasnia, A., Yousefi, N., Mahvi, A. H., Nabizadeh, 

R., Radfard, M., Yousefi, M., & Alimohammadi, 
M. (2019). Prediction of  water quality parameters 
by artificial intelligence techniques: A case study of  
the Gorganrood River, Iran. Environmental Science 
and Pollution Research, 26(1), 1–11. https://doi.
org/10.1007/s11356-018-3608-8

Abdullahi, M. G., Musa, A. A., & Ahmed, A. (2017). 
Hydrogeochemical assessment and irrigation water 
quality of  Sokoto-Rima River Basin, Nigeria. 
Applied Water Science, 7(8), 4633–4645. https://doi.
org/10.1007/s13201-017-0631-6

Adeleke, B. O., Adewumi, J. R., & Oladipo, A. A. 
(2021). Assessment of  hydrochemical characteristics 
and irrigation water quality of  Opa Reservoir, 
Southwestern Nigeria. Applied Water Science, 11(10), 
1–14. https://doi.org/10.1007/s13201-021-01505-0

Adesakin, T. A., Akinbile, C. O., & Akinlolu, A. (2020). 
Assessment of  seasonal variations in water quality of  
River Ogbese, Southwestern Nigeria. Applied Water 
Science, 10(8), 1–12. https://doi.org/10.1007/s13201-
020-01267-w

Adewumi, J. R., Ojekunle, Z. O., Sangowusi, O. R., 
& Taiwo, A. M. (2022). Water quality assessment 
of  selected rivers in Ogun State, Southwestern 
Nigeria. Applied Water Science, 12(7), 136. https://doi.
org/10.1007/s13201-022-01684-8

Adeyemo, O. K., Odo, G. E., & Adejumo, O. E. (2020). 
Seasonal variations in physicochemical parameters and 
heavy metals in surface water of  River Ogun, Nigeria. 
Environmental Monitoring and Assessment, 192(3), 1–14. 
https://doi.org/10.1007/s10661-020-8106-3

Akanda, M. G. R., Mittra, P. K., & Biswas, A. (2025). 
Adoption of  climate smart agricultural practices 
among coastal farmers: Insights from a selected 
southern area of  Bangladesh. American Journal 
of  Environment and Climate, 4(3). https://doi.
org/10.54536/ajec.v4i3.5318

Akinbile, C. O., Awomeso, J. A., & Olayemi, I. K. (2019). 
Land use and water quality in Southwestern Nigeria: 
A review. Environmental Management and Sustainable 
Development, 8(1), 58–76. https://doi.org/10.5296/
emsd.v8i1.14366

Akintola, T. A., Adekunle, A. S., & Ojo, O. (2023). 
Geostatistical modelling of  groundwater quality in 
Lagos, Nigeria. Environmental Monitoring and Assessment, 
195(2), 212–228. https://doi.org/10.1007/s10661-
023-10824-6

Aladejana, J. A., Oladejo, O. S., & Olaniyan, A. (2021). 
Machine learning models for groundwater quality 
prediction in Lagos, Nigeria. Environmental Earth 
Sciences, 80(6), 1–14. https://doi.org/10.1007/
s12665-021-09420-3

Al-Farraj, A., Al-Omran, A., & El-Maghraby, M. M. 
(2023). Hydrochemical characteristics and water 
quality evaluation of  groundwater in arid regions: 
Case of  Wadi Dawasir, Saudi Arabia. Environmental 
Earth Sciences, 82(1), 1–18. https://doi.org/10.1007/
s12665-022-10455-y



Pa
ge

 
21

9

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

Alemayehu, F., Getahun, A., & Gashaw, T. (2022). Land 
use effects on river water quality in the Ethiopian 
highlands. Ecohydrology & Hydrobiology, 22(4), 571–582. 
https://doi.org/10.1016/j.ecohyd.2022.05.001

APHA. (2017). Standard methods for the examination of  water 
and wastewater (23rd ed.). American Public Health 
Association.

AwoI, F. S., Alemayehu, T., & Legesse, A. (2021). 
Hydrochemistry of  rivers in the Rift Valley of  
Ethiopia: Implications for water quality management. 
Applied Water Science, 11(7), 1–15. https://doi.
org/10.1007/s13201-021-01534-9

Awomeso, J. A., Akintola, T. A., & Adewumi, J. R. (2019). 
Water quality challenges in Sub-Saharan Africa: A 
review. African Journal of  Environmental Science and 
Technology, 13(9), 331–341. https://doi.org/10.5897/
AJEST2019.2665

Ayoade, A. A., Osuala, B. O., & Popoola, O. M. (2019). 
Physico-chemical parameters, chlorophyll a and 
phytoplankton community as trophic state indices of  
two tropical lakes, southwestern Nigeria. EurAsian 
Journal of  BioSciences, 13(2), 1563–1573

Bwala, J. M., Amadi, A. N., Iwuala, C. C., & Enebel, 
U. U. (2025). Assessment of  common ambient air 
pollutants and respiratory health outcomes among 
road construction workers in Imo State, Nigeria. 
American Journal of  Environment and Climate, 4(3). 
https://doi.org/10.54536/ajec.v4i3.5615

Ekwueme, B. N., Nwankwoala, H. O., & Akpan, U. E. 
(2018). Hydrochemical characterization of  surface 
waters in the Cross River Basin, Southeastern Nigeria. 
Environmental Monitoring and Assessment, 190(5), 1–15. 
https://doi.org/10.1007/s10661-018-6641-4

Edokpayi, J. N., Odiyo, J. O., Popoola, E. O., & Msagati, 
T. A. (2019). Assessment of  water quality and human 
health risk in Nigeria: A case study of  Rivers in Ekiti 
State. Environmental Monitoring and Assessment, 191(4), 
223–238. https://doi.org/10.1007/s10661-019-7367-y

Elhassan, A. M., Ibrahim, M. A., & Abdalla, M. A. (2023). 
Hydrochemistry and water quality assessment of  the 
Blue Nile River. Applied Water Science, 13(5), 1–14. 
https://doi.org/10.1007/s13201-023-01822-7

Erah, P. O., Akujieze, C. N., & Oteze, G. E. (2019). Surface 
water quality assessment of  Ossiomo River, Southern 
Nigeria. African Journal of  Environmental Science and 
Technology, 13(6), 214–223. https://doi.org/10.5897/
AJEST2018.2585

Ezekiel, E. N., Ogbuagu, D. H., & Nwankwo, C. (2019). 
Seasonal variation of  water quality parameters in 
Imo River, Nigeria. Applied Water Science, 9(4), 1–13. 
https://doi.org/10.1007/s13201-019-0940-7

Gichuki, S., Kipkorir, E., & Cheruiyot, C. (2023). Seasonal 
dynamics of  nutrients and sediment in East African 
rivers. Ecohydrology & Hydrobiology, 23(2), 213–225. 
https://doi.org/10.1016/j.ecohyd.2023.01.005

Gizaw, M., Yidana, S. M., & Nketia, K. A. (2019). 
Hydrochemical evolution and water quality assessment 
of  Ethiopian Rift Valley Lakes. Environmental 

Monitoring and Assessment, 191(9), 575–590. https://
doi.org/10.1007/s10661-019-7724-8

Gupta, S., Singh, A., & Kumar, P. (2022). Water quality and 
pollution load assessment in Indian rivers. Environmental 
Science and Pollution Research, 29(14), 21045–21058. 
https://doi.org/10.1007/s11356-021-17192-y

Ibrahim, A. T., Oladipo, E. O., & Adebayo, A. A. (2021). 
Hydrochemical characterization of  Oshun River 
Basin, Nigeria. Applied Water Science, 11(12), 1–12. 
https://doi.org/10.1007/s13201-021-01520-1

Ighalo, J. O., & Adeniyi, A. G. (2020). A review of  
treatment technologies for the removal of  phosphate 
from water. Journal of  Environmental Chemical 
Engineering, 8(5), 104–106. https://doi.org/10.1016/j.
jece.2020.104106

Khan, I., Yaseen, Z. M., & Shahid, S. (2023). Predictive 
modelling of  water quality using machine learning: 
A global review. Environmental Research, 220, 115151. 
https://doi.org/10.1016/j.envres.2022.115151

Khatri, N., & Tyagi, S. (2015). Influences of  natural and 
anthropogenic factors on surface and groundwater 
quality. Frontiers in Life Science, 8(1), 23–39. https://
doi.org/10.1080/21553769.2014.933716

Mureithi, S. M., Githinji, M. M., & Njuguna, S. M. (2021). 
Water quality dynamics in Kenyan rivers: Seasonal 
variability and pollution sources. Environmental 
Monitoring and Assessment, 193(11), 723–734. https://
doi.org/10.1007/s10661-021-09517-w

NSDWQ. (2007). Nigerian Standard for Drinking Water 
Quality. Abuja: Nigerian Industrial Standard, NIS 
554:2007.

Ogunfowokan, A. O., Okoye, C. O., & Akinola, M. 
O. (2018). Impact of  land use on surface water 
quality in Ogun River Basin, Nigeria. Environmental 
Monitoring and Assessment, 190(3), 1–13. https://doi.
org/10.1007/s10661-018-6484-z

Ojekunle, Z. O., Taiwo, A. M., & Sangowusi, O. R. (2021). 
Assessment of  water quality along River Benue, 
Nigeria. Applied Water Science, 11(5), 1–12. https://doi.
org/10.1007/s13201-021-01466-4

Okafor, J. O., Chukwu, A. C., & Eze, C. N. (2018). 
Seasonal variation in the water quality of  River Benue, 
Nigeria. African Journal of  Environmental Science and 
Technology, 12(9), 329–339. https://doi.org/10.5897/
AJEST2018.2571

Okoye, C. O., Olayemi, I. K., & Adewumi, J. R. 
(2021). Seasonal dynamics of  river water quality in 
Southwestern Nigeria. Applied Water Science, 11(8), 
1–13. https://doi.org/10.1007/s13201-021-01495-z

Olayemi, I. K., Okoye, C. O., & Ogunfowokan, A. O. 
(2020). Nutrient enrichment and eutrophication risk 
in Kaduna River, Nigeria. Environmental Monitoring and 
Assessment, 192(11), 1–14. https://doi.org/10.1007/
s10661-020-08614-3

Ololade, I. A., & Ajayi, O. O. (2015). Hydrochemistry and 
pollution status of  River Osun, Southwestern Nigeria. 
Environmental Monitoring and Assessment, 187(10), 640–
655. https://doi.org/10.1007/s10661-015-4867-7



Pa
ge

 
22

0

https://journals.e-palli.com/home/index.php/ajee

Am. J. Environ Econ. 4(1) 205-220, 2025

Ogundiran, M. B., Fawole, O. O., Ayandiran, T. A., 
& Adewoye, S. O. (2022). Evaluation of  seasonal 
variations in physicochemical characteristics and 
heavy metals in surface water of  Osun River, 
Southwestern Nigeria. Environmental Monitoring and 
Assessment, 194(8), 563. https://doi.org/10.1007/
s10661-022-10143-5

Rahman, M. M., Islam, M. S., & Hasan, M. M. (2022). 
Machine learning approaches for predicting water 
quality parameters in South Asian rivers. Environmental 
Monitoring and Assessment, 194(9), 1–15. https://doi.
org/10.1007/s10661-022-10311-7

Rango, T., Yidana, S. M., & Bouri, S. (2021). 
Hydrogeochemical characterization of  groundwater 
in Ethiopia: Implications for sustainable water supply. 
Environmental Earth Sciences, 80(12), 1–15. https://doi.
org/10.1007/s12665-021-09485-0

Rao, G., Singh, S., & Sharma, R. (2021). Nutrient 
enrichment and water quality deterioration in tropical 
Asian rivers. Environmental Monitoring and Assessment, 
193(10), 650–663. https://doi.org/10.1007/s10661-

021-09488-9
Singh, S., Kumar, R., & Sharma, A. (2020). Statistical 

assessment of  hydrochemistry and irrigation suitability 
of  the Ganga River Basin, India. Environmental 
Monitoring and Assessment, 192(6), 1–17. https://doi.
org/10.1007/s10661-020-08454-1

Standards Organisation of  Nigeria. (2007). Nigerian 
Standard for Drinking Water Quality. Abuja: Nigerian 
Industrial Standard, NIS 554:2007.

Ujoh, F., Dam, D. P., & Alhassan, M. M. (2025). 
Assessment of  water quality around selected dams and 
irrigation projects in Nigeria. Kashere Journal of  Politics 
and International Relations, 3(1), 77–92. https://journals.
fukashere.edu.ng/index.php/kjpir/article/view/438

WHO. (2017). Guidelines for drinking-water quality (4th ed.). 
World Health Organization.

Yidana, S. M., Akurugu, B. A., & Chegbeleh, L. 
P. (2020). Assessment of  groundwater quality 
in the Talensi District, Northern Ghana. The 
Scientific World Journal, 2020, 8450860. https://doi.
org/10.1155/2020/8450860


