Assessment of the Surface Water Quality of Ibrahim River (Lebanon): A Spatio-Temporal Analysis

Abstract

Monitoring surface water quality offers a clear understanding of its parameters over time and space. The Ibrahim River, one of the main rivers in Lebanon, was monitored over one hydrological year, from March 2021 to April 2022. Samples were collected from seven stations in the watershed, once every two weeks. A total of 504 samples were then analyzed for pH, conductivity, turbidity, total dissolved solids, dissolved oxygen, biochemical oxygen demand, dissolved nitrate, dissolved potassium, dissolved chloride, total alkalinity, fecal coliforms, and total coliforms. Principal Component Analysis (PCA) was able to highlight two principal components (PCs), representing spatial and temporal variations, identifying areas of pollution and the influence of flow on water quality. The adapted Water Quality Index (WQI) confirmed the PCA trend with an overall average for the entire watershed of 83.70 ± 4.97, indicating a “good” water quality.

1. Introduction

Surface water systems, especially rivers, are vital freshwater sources for drinking, irrigation and domestic and industrial use. They play a key role in maintaining ecological balance [1]. However, in recent years, factors such as population growth, industrialization, and urbanization have significantly impacted both the availability and the quality of water resources. This has led to notable water quality degradation, necessitating comprehensive surface water assessments [2,3].

To effectively evaluate surface water quality, it is essential to conduct spatio-temporal monitoring. This process involves carefully planning and defining a set of physical, chemical, and biological parameters. The challenge is to transform extensive parameter data into a single, interpretable metric. This is where the Water Quality Index (WQI) proves invaluable. The WQI consolidates complex water quality data into a single numerical value, typically ranging from 0 to 100, providing a clear and concise representation of overall water quality [4]. This tool not only simplifies the assessment process but also facilitates better understanding and decision-making in water resource management.

Various methods for WQI determination are used, such as the National Sanitation Foundation Water Quality Index (NSFWQI), Canadian Council of Ministers of Environment Water Quality Index (CCMEWQI) and Oregon Water Quality Index (OWQI). Each index is applied regionally, based on varying parameter sets, and serves a distinct assessment purpose, summarized in

Table 1. Comparison of water quality indices.

WQI	General Information	Parameters	Weights	Advantages	Limitations
NSFWQI	Developed by the National Sanitation Foundation in the U.S.; widely used globally	9 core parameters	Fixed weights	Summarized in a single index value in an objective, rapid, and reproducible manner	General water quality, therefore, does not represent specific use of water; loss of data during handling; lack of dealing with uncertainty and subjectivity present in complex environmental issues
CCMEWQI	Created by the Canadian Council of Ministers of the Environment; guideline-based	User-defined parameters	No weights	Easy to understand and calculate; tolerance to missing data; adaptability to different legal requirements and different water uses	Loss of information and interaction among variables; same importance given to all parameters
OWQI	Developed by Oregon Department of Environmental Quality; consumption-oriented	Specific to local use	Equal weights	Simple; availability of required quality parameters; formula is sensitive to changing conditions and to significant impacts on water quality	Exclude stressors; data is representative of sampling site

.

All the advantages and disadvantages of the different types of WQIs were reviewed in detail by Chidiac et al. in their 2023 study [5].

The primary objective of this research was to evaluate the water quality of the Ibrahim River watershed in Lebanon through a comprehensive spatiotemporal assessment. This was accomplished by monitoring water quality over a complete hydrological year across strategically selected sampling stations throughout the watershed. The integration of both WQI and PCA methodologies provided complementary analytical perspectives, enabling a robust characterization of water quality patterns and their driving factors across both space and time [6,7,8,9].

Given the unique ecological characteristics of the Ibrahim River, we adapted four Water Quality Indices (WQIs) for this study. We complemented this approach with Principal Component Analysis (PCA) to investigate both temporal and spatial variations in water quality parameters.

2. Materials and Methods

2.1. Study Area

The Ibrahim River, a major Lebanese river, flows through Mount Lebanon Province and is recognized for its ecological and hydrological significance [10]. The river has a 30-km length and a 16,106 m³/year average annual flow. It occupies a 327 km² area and is situated 20 km north of the capital of Lebanon [11]. Afqa (1200 m) and Roueiss (1300 m), located atop Mount Lebanon, are the two main sources sustaining year-round flow. For Afqa and Roueiss, the summertime water source debits are 0.75 and 0.4 m/s, respectively [12]. The Ibrahim River flows west draining into the Mediterranean Sea [13] and produces an average annual flow of 500 million cubic meters. The river is primarily fed by snowmelt, rainfall, and tributaries originating from various sources.

The river is bordered by Al Jaouz and Abou Ali basins to the north, El Kalb basin to the south, and Litani and Yammouneh basins to the east [11]. The river’s estuary is located about 25 km north of the capital Beirut. Its average altitude is 1576 m with an average slope of approximately 17% and a structure composed of limestones from the Jurassic and Cretaceous periods [14].

The industrial activity among Ibrahim River primarily refers to minor businesses including food and marble companies, woodworking shops, and repair shops [6]. With a catchment density of ~1700 residents/km², the river is impacted by human activities [15]. Most of the watershed consists of non-productive soil categories (43%), with 37% of forested area, mainly conifer, pistachio, and deciduous trees; 10% of herbaceous vegetation; and 2% of artificialized lands [11]. Agricultural zones cover ~8% of the watershed, consisting mainly of orchards, banana and olive plantations, vineyards, and tomato greenhouses. [11].

The area is characterized by a temperate Mediterranean climate described by light wet winters and hot dry summers [16], with an average annual precipitation of about 1220 mm, ranging between 900 and 1400 mm depending on the altitude, as described in Figure 1 [12]. On the other hand, the main annual air temperature is close to 16.2 °C, with a minimum of −1.45 °C and a maximum of 32.7 °C, both registered by the nearest recording station of Kartaba (at 1222 m of altitude), based on the 2000–2011 data [11].

2.2. Sampling and Pretreatment

From March 2021 to April 2022, water samples were taken every two weeks from seven locations delineated by Assaker [11]. Three basin axes were investigated: one near the Afqa source, one near the Roueiss tributaries, and five stations along the main watershed course. To locate the initial sampling station and avoid freshwater–seawater mixing zones, salinity measurements were frequently conducted at the watershed’s outlet during project initiation. At each site, samples were obtained from the center of the river, filling the bottles by hand in the middle of the water column facing the current and spaced equally from the two banks [6].

During monitoring, flow ranged from 0.015 m³/s in dry summer to 55.08 m³/s in high-flow spring, with an average of 15.99 ± 16.97 m³/s. These rates were measured at the hydrometric station operated by the Litani River Authority (LRA) already installed at the outlet of the river

2.2.1. Characteristics of the Stations

Before undertaking our sampling campaigns, we carefully defined the framework of the study with the main objective of evaluating the water quality of Ibrahim River to build a comprehensive spatio-temporal evaluation of the river. For this purpose, preliminary characterization was essential to define the nature and different characteristics of our study region. Secondly, coordinates were adjusted in the field based on accessibility, and terrain was examined for visible or concealed signs of pollution. Likewise, an investigation into all causes likely to cause contamination has been established among residents and villagers. Thus seven different stations were chosen with different diversifications. Moreover, three axes of the basin were examined, namely a first station at the level of the tributaries coming from the Afka source, a second at the level of those of the Roueiss source, and five stations all along the main course. The geographic coordinates were recorded using a GPS (latitude and longitude). The characteristics (geographical coordinates and elevation) of each station are summarized in

Table 2. Geographic DMS coordinates and elevation of the sampling stations from downstream (L1) to upstream (L6, L7).

Sampling Stations	Latitude	Longitude	Elevation (m)
L1	34°03′48.5″ N	35°38′42.4″ E	4
L2	34°04′58.0″ N	35°41′01.5″ E	111
L3	34°04′39.0″ N	35°43′26.7″ E	274
L4	34°04′41.2″ N	35°49′33.5″ E	743
L5	34°05′59.9″ N	35°51′35.1″ E	1063
L6	34°04′05.1″ N	35°53′33.1″ E	1213
L7	34°06′32.4″ N	35°54′29.1″ E	1268

.

Station L1 (L for level) was the first sampling site, where continuous salinity measurements were conducted. This station is relatively interesting since it controls materials carried from the catchment area to the sea as well as all runoff drained by the river and its tributaries. Station L2 is located after the Yahchouch dam surrounded by trees and lightly populated areas. Station L3 is located downstream of the Chouane dam in the industrial region of the Ibrahim River. There are large quarries used by the marble manufacturing companies. Furthermore, it is urbanized, and we observe the presence of five significant businesses in the food, pharmaceutical, and marble production industries. Station L4 is surrounded by bare rocks, agricultural strips, and wooded areas. Station L5 is located after the confluence of the two main sources, Afka and Roueiss, upstream of the river. On the flood-prone section of the proposed Jannah dam, this station was selected and is particularly surrounded by agricultural land. Station L6 is located where the Afqa spring emerges in a remote, inaccessible location covered by stark limestone cliffs. Station L7 is the highest accessible upstream point, located at a tributary fed by streams descending from the Roueiss source. The stream is now surrounded by areas left in their natural state and by terrain that is devoid of people or habitations (Figure 2).

Concerning the lithology of the different stations, we can notice the dominance of limestone. Its proportion goes from 70% downstream of the river along with small fractions of dolomitic limestone, basalt, quartzo-limestone sandstone, and colluvium, reaching 100% upstream of the river in stations L6 and L7 [11]. The sediments, on the other hand, of the different stations are mainly silt with varying percentages of clay (in April much higher than in December) [11].

2.2.2. Initial Processing

Before collecting water samples for physicochemical analysis, 1L polypropylene bottles (Nalgene™, Thermo Fisher Scientific, Rochester, NY, USA) were washed with a phosphate-free detergent, rinsed, and then repeatedly cleansed with 10% nitric acid (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) and Milli-Q ultrapure water (MilliporeSigma, Darmstadt, Germany). For microbiological analysis, borosilicate bottles (ISOLAB™, ISOLAB Laborgeräte GmbH, Eskişehir, Turkey) were autoclaved (Daihan Scientific Co., Ltd., Wonju, Republic of Korea) at 120 °C for 15 min after being rinsed using Milli-Q ultrapure water to remove any residues of antiseptic or detergent.

Two polypropylene bottles (2L) and one borosilicate amber glass (1L) (Simax, Kavalierglass, Sázava, Czech Republic) were each filled with water for a complete hydrological year at each station. The samples were immediately carried in a portable refrigerator (Rubbermaid, Atlanta, GA, USA) and kept at 4 °C in the laboratory under total darkness [17]. One polypropylene bottle was filtered using the membrane filtration method with a cellulose nitrate filter of 0.22 µm pore size (Sartorius Stedim Biotech GmbH, Göttingen, Germany), of which 100 mL were acidified with a few drops of highly pure nitric acid for further cation analysis. All alkaline and microbiological tests were carried out within 24 h [18].

2.3. Physicochemical Analysis

Twelve parameters were selected for analysis based on the study conducted by El Najjar et al. [6]. The latter initially used 23 physico-chemical parameters (flow, pH, temperature, electrical conductivity, turbidity, total suspended solids, total dissolved solids, dissolved oxygen, dissolved organic carbon, biochemical oxygen demand, chemical oxygen demand, chemical oxygen demand/biochemical oxygen demand, specific ultraviolet absorbance, dissolved cations (Na⁺, K⁺, Ca²⁺, Mg²⁺), total alkalinity, and dissolved anions (Cl⁻, NO₃⁻, NO₂⁻, PO₄³⁻ and SO₄²⁻)) and five microbiological tests (total germs, total coliforms, fecal coliforms, Escherichia coli, and Enterococcus). PCA and Pearson correlation were applied to identify the most valuable parameters and the least correlated to highlight the impact of anthropogenic inputs and flood events for a better monitoring of the river in the future.

Table 3. Tested physico-chemical and microbiological parameters with relevant analytical methods.

Parameters	Unit	Analytical Method
pH	-	Hanna Instruments Portable Analog pH Meter HI-8314 AOAC 973.41
Conductivity (EC)	μS/cm	Hanna Instruments Portable EC/TDS/Temperature Meter HI-99300 NF EN 27888
Turbidity (Turb)	NTU	Turbidimeter TB1 Velp Scientifica NF EN ISO 7027 (March 2007)
Total Dissolved Solids (TDS)	mg/L	Hanna Instruments Portable EC/TDS/Temperature Meter HI-99300 AOAC 920.193
Dissolved Oxygen (DO)	mg/L	Oxygen meter Lutron DO5510 NF EN 25814 (March 1993)
Biochemical Oxygen Demand (BOD5)	mg/L	Lovibond BD 600 OxiDirect NF EN 1899-2 (May 1998)
Dissolved Nitrate (NO3−)	mg/L	Spectrometry Helios Alpha Thermo NF T 90-045 ISO 7890-3:1988
Dissolved Potassium (K+)	mg/L	Flame photometry M410 Sherwood Scientific NF T 90-019
Total Alkalinity (TA)	mg/L CaCO3	Titrimetry NF T 90-036 ISO 9963-1:1994
Dissolved Chloride (Cl−)	mg/L	Titrimetry NF T 90-014 ISO 9297-1989
Total Coliforms (Tot Colif)	CFU/100 mL	NL ISO 9308-1:2012 and ISO 7899-2:2000(E)
Fecal Coliforms (Fec Colif)	CFU/100 mL	NL ISO 9308-1:2012 and ISO 7899-2:2000(E)

summarizes the twelve parameters selected for the study, the units of measurements, and the various analytical techniques.

During the sampling campaigns, several events (chemical, physical, or biological) could occur and significantly change the concentrations of some components. Events that dissolve carbon dioxide can change the pH and the conductivity (EC), while degassing processes can have an impact on the amount of dissolved oxygen. pH, EC, dissolved oxygen (DO), total dissolved solids (TDS) and turbidity (Turb) parameters are measured in situ, within the first 5 min after sampling [17]. All instruments used for in situ and laboratory measurements were calibrated according to manufacturer guidelines and international standards. Portable meters (pH, EC, DO, TDS) were calibrated before each sampling campaign using buffer solutions. Laboratory instruments, including spectrophotometers and titration setups, were calibrated weekly and verified with standard reference materials to ensure measurement accuracy throughout the study.

2.4. Microbiological Analysis

Total coliforms (CFU/100 mL) and fecal coliforms (CFU/100 mL) were examined using the membrane filtration technique. Water was introduced into a sterilized filtration unit that contained a gridded, sterile cellulose nitrate membrane filter of 0.45 μm diameter (Sartorius Stedim Biotech GmbH, Göttingen, Germany) [19,20]. The filter was placed on the appropriate culture media once germs were retained on it (Endo agar (Merck™, Cat. No. 104044, Merck KGaA, Darmstadt, Germany) for the total coliforms and Levine EMB (Biolife Italiana™, Milan, Italy) for the fecal coliforms). After incubation for 24 to 48 h at 37 °C (Prolabo™,CAP-LAB, France) for total coliforms and 44 °C for fecal coliforms (Memmert, Memmert GmbH + Co. KG, Schwabach, Germany), colonies were counted using a BOECO colony counting instrument (BOECO CC-1, Boeckel & Co. GmbH + Co. KG, Hamburg, Germany). Results are expressed as colony-forming units (CFU) per 100 milliliters (CFU/100 mL) of testing water [6,17].

2.5. Principal Component Analysis (PCA)

The interactions between the various parameters were investigated using a multivariate data analysis method, Principal Component Analysis (PCA), an explorative technique commonly used for the analysis of water quality data. PCA allows a thorough investigation of spatiotemporal trends in water quality and thus yields meaningful information. PCA converts numerous indicators into a smaller number of comprehensive factors using the concept of dimensionality reduction, which fosters a deeper understanding than what is possible when looking at variables individually [21,22].

Every multivariate data analysis begins with the definition of a data set or a data matrix, designated by X ($n\times k$). The samples in the table are the n rows, representing “observations” or “objects” as per the study strategy and design. The measurements taken on these observations are included in the k columns, which are referred to as “variables.” PCA works by projecting the hidden data structure enclosed in matrix X onto two subspaces, termed the “score space” and the “loading space,” respectively. The PCA equation can therefore be represented as follows:

$$X = T.P^T + E$$

(1)

where T is the score matrix, P is the loadings matrix, and E is the residual matrix. All PC directions are “linear combinations” of original k variables. They are calculated, ordered, and indexed, such that the first component PC1 explains the highest possible fraction of this sum, PC2 explains the second largest fraction, the third less than this, and so on, until the residual (E matrix) is left [23]. In this study, the columns refer to the 12 physicochemical and microbiological parameters evaluated from the Ibrahim River during different seasons and from different stations. The rows, on the other hand, refer to the total of 504 samples. Twenty-four sampling campaigns were carried out over seven chosen sites, with a frequency of three sampling replicates (7 sites × 24 sampling campaigns × 3 replicates = 504 samples). PCA was performed on the data set using MATLAB R2022a software, after column standardization (mean centering each variable and dividing the resulting values by the column’s standard deviation).

2.6. Water Quality Index (WQI)

One of the methods for describing water quality is the Water Quality Index, which summarizes a large amount of data into one value ranging between 0 and 100. Values between 0 and 25 represent a “very bad” water quality, 26 to 50 indicate a “bad” water quality, 51 to 70 reflect a “medium” quality, 71 to 90 represent a “good” water quality, and values between 91 and 100 indicate an “excellent” water quality [5].

While standard indices such as NSFWQI and CCMEWQI offer broad applicability, they present limitations in the context of the Ibrahim River (Table 1). Therefore, we adapted the WAWQI with adequate weights to each parameter depending on the ecological context, to better reflect the spatial and temporal variation of the watershed.

Four Water Quality Indices were thus developed, including an index for each sampling (done every two weeks), an index for each station, an index for low flow, and an index for high flow using the following equation:

$${WQI}_{objective}{\displaystyle \frac{\sum _{i=1}^{n}CiPi}{\sum _{i=1}^{n}Pi}}·$$

(2)

where n is the total number of parameters, C_i is the given value (%), and P_i is the weight given to each parameter. At all times, k = 1 was used to generate the objective Water Quality Index, which only considers fluctuations brought on by observed variables [5,24,25,26].

3. Results and Discussion

Table 4. The mean, standard deviation, and WHO standard values of the analyzed physical, chemical, and microbiological parameters.

Stations	pH	EC (μS/cm)	Turb (NTU)	TDS (mg/L)	DO (mg/L)	BOD5 (mg/L)	NO3− (mg/L)	K+ (mg/L)	TA (mg CaCO3/L	Cl− (mg/L)	Tot Colif (CFU/100 mL)	Fec Colif (CFU/100 mL)
L1	8.13 ± 0.51	270.68 ± 36.69	12.83 ± 26.72	134.64 ± 17.71	13.32 ± 9.60	2.48 ± 3.61	1.54 ± 084	1.26 ± 2.31	182.47 ± 31.60	8.68 ± 3.23	4257.17 ± 2814.20	440.69 ± 568.13
L2	8.14 ± 0.41	280.18 ± 38.21	22.68 ± 57.86	138.71 ± 18.37	12.24 ± 7.95	1.98 ± 3.04	1.46 ± 0.59	1.31 ± 2.45	196.92 ± 38.86	8.25 ± 3.03	6152.42 ± 14,355.76	140.94 ± 118.87
L3	8.13 ± 0.43	282.90 ± 39.97	12.67 ± 26.77	139.63 ± 17.92	12.89 ± 9.69	2.27 ± 3.30	1.38 ± 0.77	1.19 ± 2.34	200.13 ± 39.98	7.86 ± 2.96	4300.52 ± 9238.88	117.35 ± 135.24
L4	8.13 ± 0.44	287.88 ± 42.36	7.19 ± 16.15	141.00 ± 20.39	13.99 ± 10.12	2.18 ± 3.17	1.57 ± 0.83	1.26 ± 2.41	202.28 ± 40.86	7.33 ± 3.02	5498.92 ± 14,404.14	127.21 ± 121.99
L5	8.06 ± 0.43	359.58 ± 66.88	1.33 ± 1.19	177.10 ± 29.78	12.89 ± 8.72	1.94 ± 3.29	1.53 ± 0.97	1.33 ± 2.57	269.79 ± 55.91	11.2 ± 4.82	12,308.40 ± 20,917.54	406.13 ± 1029.64
L6	8.04 ± 0.45	225.72 ± 37.67	0.88 ± 1.10	107.81 ± 9.47	13.43 ± 9.14	1.77 ± 2.71	1.48 ± 0.67	0.47 ± 0.93	151.71 ± 23.68	5.99 ± 2.70	4470.48 ± 10,312.26	60.83 ± 134.73
L7	8.00 ± 0.41	231.03 ± 45.38	0.81 ± 0.97	109.81 ± 10.08	13.24 ± 8.58	1.67 ± 2.67	1.09 ± 0.74	0.62 ± 1.17	162.40 ± 28.80	6.77 ± 3.78	13,911.7 ± 24,083.49	1.50 ± 4.40
WHO Standards	6.5–8.5	250	<5	<600	-	-	10	- *	200	<250	<10	0

Note: * Not of a health concern at levels found in drinking water.

presents the findings of the analyzed parameters per station, organized by season.

pH is an important physicochemical parameter influencing other water quality parameters such as alkalinity, solubility of metals, and hardness [27]. Across all the stations, pH levels remained slightly alkaline due to the contact with carbonate rocks in the specified locations [27]. EC variations, influenced by dissolved ion concentration, temperature, and ionic strength, were lowest at stations L6 and L7, indicating minimal contamination and reduced ion loads (Table 4). Turbidity, a measure of water cloudiness due to suspended particles [28], varied across stations. Station L7 exhibited the clearest water, while station L2 showed the highest turbidity, likely due to sediment load or suspended solids. TDS measurements were highest at station L5, reflecting elevated ion concentrations likely driven by anthropogenic inputs. DO levels were generally adequate for aquatic life, with fluctuations potentially linked to temperature changes and organic matter decomposition. Furthermore, lower BOD₅ values at stations L6 and L7 suggested lower organic pollution levels in these areas. BOD₅ is a critical indicator of biodegradable organic matter. Elevated levels suggest wastewater contamination, which can promote microbial growth, deplete oxygen, and disrupt aquatic ecosystems. Nitrate concentrations remained within acceptable limits, indicating minimal contamination. Potassium levels were relatively stable across stations, reflecting limited variation. TA values were within acceptable ranges, reflecting the water’s acid-neutralizing capacity influenced by carbonate, bicarbonate, and hydrogen sulfide. Chloride concentrations varied across stations, with L1, L2, and L5 showing elevated levels compared to L3–L7. These differences likely stem from industrial, agricultural, and natural sources. High total coliform counts at station L5 indicated contamination from anthropogenic sources. Elevated fecal coliform levels, especially at station L5, indicated significant anthropogenic contamination, needing further investigation and remediation efforts.

Thus, it is important to note that the high standard deviations observed for some parameters in Table 4, mainly turbidity, total coliforms, and fecal coliforms, indicate the variability both spatially and temporally linked to environmental conditions. For instance, elevated turbidity at certain stations during specific sampling periods resulted from rainfall events and surface runoff. Similarly, the large fluctuations in coliform counts reflect the contamination from point and non-point sources and the variation of the river’s discharge and dilution effect. A detailed explanation is presented in the following sections.

PCA was applied to analyze the spatio-temporal variation of the water quality along the Ibrahim River. This technique reduces dimensionality by replacing original indicators (variables) with new comprehensive indicators (PCs). The amount of information enclosed in each PCs is quantified by the explained variance percentages [21]. PCA was performed on the data matrix X (504 × 12), retaining two principal components (PC1 and PC2) that collectively explain about 45% of the original variance. Visualizing scores corresponding to the 504 samples helps in identifying potential spatial or temporal trends across river sampling stations and seasons. PCs loadings, derived from the 12 study variables (pH, EC, TDS, DO, BOD₅, NO₃⁻, K⁺, TA, Cl⁻, Tot Colif, and Fec Colif), reveal which variables drive the observed patterns in the scores.

3.1. Spatial Variation

Figure 3 illustrated the scores of the 504 samples and the loadings of the 12 parameters relative to PC1. PC1 demonstrated spatial variations, with stations L1 to L7 exhibiting different trends along PC1. These spatial trends could be linked to various parameters based on their loadings. Notably, EC, TDS, BOD₅, K⁺, TA, Cl⁻, and Fec Colif exhibited a significant positive correlation on one side of PC1, while pH, Turb, and Tot Colif showed minimal impact on the other side. Although total coliforms are important sanitary indicators, their low factor loading (<0.05) indicated minimal influence on the principal components. Therefore, they were excluded from the PCA. Conversely, DO and NO₃⁻ were inversely proportional to the other parameters. Due to their weak factor loadings below 0.05, pH, Turb, DO, and total coliform were disregarded in assessing water quality. Thus, the focus narrowed to the remaining eight parameters (EC, TDS, BOD₅, NO₃⁻, K⁺, TA, Cl⁻, and Fec Colif) to delineate spatial variations in the Ibrahim River. Score distribution revealed three spatial clusters: L5 (high pollution indicators), L6 and L7 (pristine source conditions), and L1–L4 (moderate anthropogenic influence).

The highest scores on PC1, particularly exhibited by L5, signaled a distinct spatial pattern justifying deeper investigation concerning PC1 loadings. This anomaly could signify potential pollution. Parameters such as EC, TDS, BOD₅, NO₃⁻, K⁺, TA, Cl⁻, and Fec Colif notably influenced the water quality at station L5 compared to pH, Turb, DO, and Tot Colif. Station L5, located at the convergence of sources L6 and L7, is bordered by agricultural fields and other regions, including 39.48% industrial zone, 63.29% rocky outcrop, and 13.1% herbaceous and shrubby vegetation [11]. Table 4 highlights station L5’s higher values for EC, TDS, K⁺, TA, and Cl⁻ compared to other stations. Previous studies have established the link between these parameters. Emenike et al. [29] demonstrated the interaction of elements using Cluster Analysis (CA) and PCA techniques, exposing statistical strong relationship between TDS and EC, indicating that the increased concentrations of both EC and TDS can be linked to heavy application of agrochemicals, rainwater percolation, ion exchange, and sediment dissolution. EC and TDS also showed moderate correlation with K⁺ and Cl⁻ and alkalinity showing high values as a result of industrial and agricultural activities within this zone (39.48% of industrial zone and surrounded by agricultural fields). Moreover, Kothari et al. [30] proved significant positive linear correlation of TDS with total alkalinity (0.8) and conductivity (0.9) as well as a moderate positive correlation with chloride (0.7), indicated by correlation coefficients (R-values) close to one.

PC1 scores revealed a distinct trend, indicating the relationship among water quality parameters: EC, TDS, TA, and Cl⁻. EC indicated elevated scores, within the acceptable ranges (Table 4), mainly due to agriculture and geological weathering processes [31,32]. High TDS scores aligned with EC, potentially resulting from both natural geological and human activities. TDS typically measures dissolved ions content, while EC indicates electrical charges influenced by factors like temperature, ion concentration, and ionic strength, illustrating their close relationship [33]. Moreover, TDS exhibited a correlation with TA. TA refers to its capacity to neutralize acids originating primarily from sources like hydrogen sulfide, carbonate, and bicarbonate [34]. Although station L5 exceeded recommended TA limits (200 mg/L), it caused no health concerns according to WHO 2022 standards [35]. Analysis of the 504 water samples revealed the combined influence of anthropogenic activities and natural processes. The strong association between Cl⁻ and K⁺ likely stems from rock and soil weathering, contributing to secondary mineral formation [36,37]. Their presence in water sources may result from natural geochemical processes, agricultural runoff, and sewage contamination [30]. While Cl⁻ and K⁺ concentrations stayed within acceptable ranges (Table 4), preventive measures were essential to prevent potential future increases. Nitrate (NO₃⁻) levels were consistently low and within permissible limits (Table 4). PC1 analysis highlighted their limited influence on spatial patterns, suggesting an inverse relationship with other water quality parameters. These observations were endorsed by the bacteriological results. Station L5 was severely discriminated against due to an increase in Fec Colif (Table 4), primarily linked to anthropogenic inputs as previously mentioned. Daou et al. [16] similarly reported mineral and bacterial contamination near sites adjacent to station L5, primarily due to anthropogenic discharges from nearby villages and urban areas. Comparable patterns were observed in the Kadisha River, which flows through densely populated regions and exhibits elevated bacterial contamination [13].

The second evolution trend clearly identified the two sources, L6 and L7, located in Afqa and Roueiss, respectively. Both sources exhibited similar distributions, with L6 and L7 displaying low scores, and consequently good parameter values. Similarly, the sources of the Kadisha and Damour rivers formed clusters with low scores and high water quality, unaffected by pollution indicators [13]. Remarkably, both sources disclosed nearly identical average values for the selected parameters with the lowest values recorded for EC, TDS, Turb, BOD₅, NO₃⁻, K⁺, TA, and Cl⁻. However, Fec Colif showed higher levels in L6 compared to L7. According to Assaker [11], the low concentration of Fec Colif in L7 was due to its natural state and the absence of habitation. Bou Saab [12] supported these findings, highlighting higher Fec Colif levels in Afqa (L6), attributing this observation to wastewater discharges into the river and soil leaching.

Stations L1 to L4 showed intermediate values for the eight selected parameters, positioned between the trends of L5 and the sources L6 and L7. Station L4 displayed the highest NO₃⁻ reading and an elevated BOD₅ reading, linking the level of NO₃⁻ in water to bacterial contamination. A similar pattern appeared in station L1, pointing to possible contamination via sewage systems, drainage, animal waste, etc. [30]. These sites of the river (station L4) were surrounded by agricultural strips, boulders, and trees, emphasizing anthropogenic activities, particularly in agricultural areas, contributing to elevated nitrate concentrations [6,11]. Exceeding the acceptable EC limit in all four sites (

Table 5. Mean and standard deviation values of the parameters during the different seasons.

Seasons	pH	EC	Turb	TDS	DO	BOD5	NO3−	K+
Spring 2021	7.84 ± 0.46	258.11 ± 76.52	4.62 ± 7.05	117.48 ± 26.40879	7.86 ± 3.23	0.55 ± 0.70	1.41 ± 0.55	0.02 ± 1.39 × 10−17
Summer 2021	8.25 ± 0.22	267.29 ± 49.97	0.64 ± 0.74	133.86 ± 2497	8.01 ± 3.87	6.78 ± 2.61	1.11 ± 0.48	3.66 ± 2.84
Fall 2021	8.00 ± 0.48	297.39 ± 61.17	5.14 ± 30.56	148.59 ± 69.14	17.92 ± 9.59	0.70 ± 1.26	1.17 ± 0.73	0.47 ± 1.00
Winter 2021	8.37 ± 0.46	284.41 ± 38.58	3.63 ± 1.79	142.27 ± 19.22	11.44 ± 4.18	0 ± 0	2.37 ± 1.03	0.02 ± 1.04 × 10−17
Spring 2022	8.21 ± 0.08	278.57 ± 33.84	60.86 ± 72.36	138.78 ± 17.14	30.21 ± 6.55	0 ± 0	2.03 ± 0.53	0.02 ± 1.04 × 10−17

) may be attributed to river runoff, extensive application of agrochemicals, rainwater percolation, ion exchange, and sediment dissolution [29]. Moreover, nitrate elevation increased water salinity and dissolved ions concentration, directly impacting the increase in EC [21]. Total alkalinity approached the permitted limit and exceeded it in stations L3 and L4 without posing immediate health risk (Table 5). TDS showed high values within the permissible range due to its correlation with EC and TA (Table 5). Chlorine concentration decreased from L5 to L4 but increased in L3 due to pharmaceutical, agri-food, and other industrial activities. Further increases from L3 to L2 and L1 were linked to residential areas and some agricultural land use, aligning with findings from Assaker [11] in the same locations. Nitrate and potassium followed a similar ascending pattern from L3 to L2 to L1. The presence of residential sewage, indicated by high BOD₅ content, possibly resulted from municipal waste and organic pollution from homes and businesses [38,39]. Furthermore, elevated BOD₅ reflected the degree of organic pollution associated with fecal sources [40]. The high level of fecal coliforms in all the stations, especially L1, exceeded permissible limits, causing public health issues, making the water unfit for drinking due to waterborne diseases [30]. Thus, the regions constituting the third trend (L1, L2, L3, and L4) with habitation and agricultural lands necessitate close monitoring of all parameters, including those within permissible ranges.

It can be concluded that PC1 reflected a spatial variation, especially station L5, demonstrating elevated concentrations of EC, TDS, and Fec colif, implying pollution issues related to the effects of both anthropogenic activities such as industrial and agricultural practices along with natural geological processes on water quality. The correlation between EC, TDS, TA, and Cl⁻ indicates the need for monitoring and management to eliminate potential contamination. While NO₃⁻ levels remained within acceptable limits, the increased presence of Fec Colif at station L5 raised concerns about water safety, necessitating further investigation. Similarly, Daou et al. [13] reported a spatial variation of water quality when going from the spring to the outlet of both Damour and Kadisha rivers due to numerous anthropogenic and natural inputs.

A distinct temporal trend appeared in PC1, seen in the scores, particularly those of L5, which increased significantly during summer and fall. This increase was primarily attributed to increased anthropogenic, agricultural, and recreational activities during summer season.

Additionally, summer and fall seasons coincided with low-water events, leading to higher concentrations of contaminants. Conversely, during the winter and spring seasons, characterized by high-water events, contaminant concentrations decreased due to dilution effects caused by increased water flow within the watershed. A more detailed understanding will be provided in Section 3.2.

3.2. Temporal Variation

Figure 4 illustrates the scores of the 504 samples and the loadings of 12 parameters on PC2. PC2 revealed a temporal pattern, with all stations’ scores following the same trend along a time axis from spring 2021 to spring 2022. This temporal trend demonstrated an inverse relationship between EC, Turb, TDS, DO, and NO₃⁻ on one side and pH, BOD₅, and K⁺ on the other side, all exhibiting significant loadings. However, TA, Cl⁻, Tot Colif, and Fec Colif did not exhibit important seasonal fluctuations. Consequently, the analyzed parameters taken into consideration fell within a range of ±0.1 loading unit (pH, EC, Turb, TDS, DO, BOD₅, NO₃⁻, and K⁺). Seasonal variation was categorized into three groups: summer 2021 (red dots), spring 2022 (green dots), and a mixed season group encompassing spring, fall, and winter 2021 (blue, yellow, and purple dots).

Between June 21 and 22 September 2021, all five stations clustered within the summer group, distinguished by higher PC2 scores and low flow conditions averaging 2.10 ± 1.92 m³/s. This period exhibited elevated BOD₅ values, signaling increased organic pollutant oxidation, in accordance with Maity et al.’s [41] findings. These increased BOD₅ levels coincided with lower concentrations of DO (Table 5), aligning with Kumar [38]. PC2 loadings further confirmed this inverse relationship among BOD₅, DO, and various other parameters. Seasonal DO variation was influenced by reduced water levels and flow rates during summer (0.89 ± 1.10 m³/s), which limited oxygen dissolution and amplified anthropogenic impacts [6]. Conversely, DO levels typically increased during periods of high-water flow when the river discharged 25.71 ± 15.41 m³/s, elevated water turbulence, and overflow. Remarkably, nitrate concentrations were reduced during the summer low-flow periods. This decline was linked to low flow, minimal runoff, and natural weathering typical of the summer season, considering that nitrate enters streams through various pathways including groundwater, surface water, and subsurface flow [42]. Interestingly, conductivity levels did not significantly increase during summer. Contrarily, the highest potassium levels recorded during this period might be attributed to anthropogenic and agricultural sources, aligning with the findings of Maity et al. [41], which reported a significant increase in potassium levels during the summer due to non-point source pollution.

The spring group denoted the spring season from 20 March 2021, until April 2022. During the spring season, the watershed experienced its peak flow levels, primarily driven by the snowmelt events. Snowmelt from higher elevations contributed to increased flow rates, with an average discharge of 32.16 ± 14.51 m³/s. The gradual melting of snow accumulated at higher elevations contributed to these increased flow rates. This increase in water flow served a dual purpose: it replenished the various water bodies within the watershed and influenced the overall water quality. Consequently, PC2 displayed decreasing scores in this group, indicating improved water quality. The stability of pH and EC, with minimal seasonal fluctuations, indicated a consistently slightly alkaline nature of the water throughout the year. Similarly, TDS levels exhibited no significant temporal variation. In contrast, turbidity peaked during spring, driven by runoff, erosion, and flood events associated with seasonal discharge. Moreover, loading analysis demonstrated microbial activity. BOD₅ exhibited negligible counts, suggesting minimal microbial influence during this season. Conversely, DO levels peaked during the rainy season, when increased water flow transported pollutants and enhanced oxygenation, consistent with findings by Chen et al. [43]. Similarly, nitrate levels also exhibited an increase during spring, driven by the high-water flow observed during this period. Potassium concentrations in the Ibrahim River, already low across all seasons, reached their lowest point during spring. This low concentration was intensified by increased precipitation and high freshwater discharge, as previously noted by Akhtar et al. [44].

The mixed season group included spring, fall, and winter 2021, reflecting the overall water quality profile of the Ibrahim watershed during these periods. Scores across these seasons consistently ranged between −3 and 1, indicating stable water quality and minimal seasonal variation. This uniformity was demonstrated by eight parameters displaying similar average scores. Additionally, EC, Turb, TDS, DO, and BOD₅ parameters showed comparable patterns. Nevertheless, nitrate displayed a slight increase in winter compared to spring and fall due to increased water discharge. Conversely, DO exhibited a slightly lower value in spring 2021 compared to fall and winter, marking a reverse trend when compared to spring 2022. Turb and DO exhibited a relationship. Higher water discharge in spring 2022 led to elevated turbidity scores compared to spring 2021 (27.12 ± 18.29 vs. 17.64 ± 13.68). pH value displayed negligible variations throughout the seasons, indicating a mild alkaline pH in the Ibrahim watershed, serving as a reliable indicator of aquatic ecosystem diversity and pH-adapted species [45]. Unlike the mixed season group’s stability, summer 2021 and spring 2022 exhibited contrasting water quality patterns, reflecting seasonal extremes. Even within the summer and spring groups, pH and EC exhibited no visible seasonal fluctuations.

In conclusion, PC2 revealed an inverse relationship between key parameters, particularly in summer, where elevated BOD₅ coincided with reduced DO, highlighting organic pollution under low-flow conditions. In spring, water quality improved due to peak flow levels caused by snowmelt, resulting in enhanced DO levels and decreased BOD₅ counts. This season also showed increased NO₃⁻ concentrations and turbidity, primarily due to elevated flow and runoff. On the other hand, spring, fall, and winter 2021 were characterized by stable scores among the different parameters, exhibiting a constant water quality profile. These findings align with patterns observed in the El-Kalb River, where bacterial contamination was more pronounced in summer than in spring and winter—likely due to reduced dilution during dry, high-temperature conditions (~30 °C) [15].

3.3. Water Quality Index

Table 6. Weights given to the parameters used for the WQI calculation.

Parameter	Weight (Pi)
pH	2
EC	3
Turb	2
TDS	2
DO	4
BOD5	3
NO3−	1
K+	2
TA	3
Cl−	3
Tot Colif	3
Fec Colif	3

provides a comprehensive overview of the assigned weights for each selected parameter to calculate four distinct indices relevant to Water Quality Index analysis. The weights were assigned based on extensive literature reviews, previous methodologies, and the ecological sensitivity of the Ibrahim River. The parameters were assigned weights according to their importance, ranging from 1 to 4, with 4 indicating the most important parameter and 1 the least important. For example, DO, BOD₅, total coliforms, and fecal coliforms were assigned higher weights due to their role directly affecting the aquatic ecosystem. On the other hand, EC, TA, and Cl− reflect the ion concentration and salinity, highlighting the areas influenced by runoff and discharges [24,25,26]. These indices included WQI per sampling event, WQI per monitoring station, WQI during low flow conditions, and WQI during high flow periods. Each index focused on a particular aspect of water quality assessment.

The analysis of Water Quality Indices across various sampling dates (

Table 7. WQI values per sampling, month, and station and during high and low flows.

Date	WQI per Sampling Campaign	Months	WQI per Month	Stations	WQI per Station	Date	WQI During High Flow	Date	WQI During Low Flow
21 March 2021	87.73 ± 3.62	Mar-21	87.73 ± 3.62	L1	80.10 ± 6.37	21 March 2021	87.73 ± 3.62	26 May 2021	85.36 ± 3.66
8 April 2021	94.73 ± 2.32	Apr-21	91.56 ± 3.11	L2	82.69 ± 5.24	8 April 2021	94.73 ± 2.32	13 June 2021	82.06 ± 3.52
25 April 2021	88.39 ± 3.44	May-21	83.37 ± 3.64	L3	83.16 ± 5.09	25 April 2021	79.63 ± 3.44	27 June 2021	75.48 ± 3.26
9 May 2021	81.38 ± 2.35	Jun-21	78.77 ± 4.72	L4	83.90 ± 5.73	9 May 2021	71.98 ± 2.35	12 July 2021	71.90 ± 1.91
26 May 2021	85.36 ± 3.66	Jul-21	74.64 ± 5.59	L5	80.62 ± 7.33	9 January 2022	89.62 ± 2.82	25 July 2021	77.37 ± 1.91
13 June 2021	82.06 ± 3.52	Aug-21	79.75 ± 4.92	L6	87.30 ± 5.54	13 February 2022	90.61 ± 1.68	4 August 2021	78.97 ± 6.68
27 June 2021	75.48 ± 3.26	Sep-21	78.98 ± 3.97	L7	86.90 ± 5.25	27 February 2022	90.51 ± 1.86	28 August 2021	80.54 ± 4.86
12 July 2021	71.90 ± 1.91	Oct-21	83.67 ± 3.50		-	27 March 2022	84.55 ± 5.87	11 September 2021	76.94 ± 4.17
25 July 2021	77.37 ± 1.91	Nov-21	84.09 ± 4.57		-	3 April 2022	85.05 ± 5.71	26 September 2021	81.01 ± 2.49
4 August 2021	78.97 ± 6.68	Dec-21	85.49 ± 6.20		-		-	10 October 2021	84.22 ± 3.63
28 August 2021	80.54 ± 4.86	Jan-22	89.62 ± 2.82		-		-	24 October 2021	83.12 ± 3.37
11 September 2021	76.94 ± 4.17	Feb-22	90.56 ± 1.75		-		-	7 November 2021	83.23 ± 4.19
26 September 2021	81.01 ± 2.49	Mar-22	84.55 ± 5.87		-		-	21 November 2021	84.95 ± 4.86
2 October 2021	84.22 ± 3.63	Apr-22	85.05 ± 5.71		-		-	5 December 2021	86.10 ± 5.12
24 October 2021	83.12 ± 3.37		-		-		-	19 December 2021	84.88 ± 7.20
7 November 2021	83.23 ± 4.19		-		-		-		-
21 November 2021	84.95 ± 4.86		-		-		-		-
5 December 2021	86.10 ± 5.12		-		-		-		-
19 December 2021	84.88 ± 7.20		-		-		-		-
9 January 2022	89.62 ± 2.82		-		-		-		-
13 February 2022	90.61 ± 1.68		-		-		-		-
27 February 2022	90.51 ± 1.86		-		-		-		-
27 March 2022	84.55 ± 5.87		-		-		-		-
3 April 2022	85.05 ± 5.71		-		-		-		-
Average	83.70 ± 4.97		84.13 ± 4.28		83.53 ± 5.79		86.05 ± 6.84		81.08 ± 4.17

) revealed intriguing insights into the environmental health of the studied area.

The monthly WQI analysis (Table 7) accentuated the temporal variations in water quality across different months. From March 2021 to April 2022, the water quality maintained a “good” classification, with WQI values exceeding 80. While overall conditions remained favorable, a decline was observed between June and September 2021, with WQI values falling below 80. This decrease in water quality could be attributed to intensified anthropogenic activities or other external influences. July 2021 recorded the lowest WQI value of 71.90 ± 1.91, signifying a period of lower water quality. In contrast, April 2021 exhibited the highest WQI of 94.73 ± 2.32, attributed to increased flow rates due to precipitation and snowmelt. The increased water flow during April contributed to a better water quality condition.

The compatibility between the Water Quality Index outcomes and the temporal Principal Component Analysis findings strengthened the validity of the water quality. WQI results and temporal PCA outcomes supported the interaction between water quality parameters and their temporal trends. Interestingly, the WQI outcomes reflected patterns observed in temporal PCA trends. Specifically, the WQI data highlighted favorable water quality during the summer season, such as the summer group in the temporal PCA analysis. Moreover, both spring and autumn seasons exhibited a “good” water quality status, exceeding that of summer, similar to the scores represented by the spring group in the temporal PCA results. These results reflected observations made by Assaker [11], who observed low contamination rates in the Ibrahim watershed from December to April, contrasting with elevated contamination rates from May to November. Chen et al. [46] observed similar seasonal trends in Erhai Lake, attributing water quality fluctuations to human and agricultural activities. Additionally, Qi et al. [47] demonstrated that during a 12-month sampling period, August 2018 and January 2019 had the lowest and highest monthly average WQI values, respectively, and that the WQI values decreased from spring to summer, increased from summer to autumn, and then fell again in winter.

Regarding the WQI for each station (Table 7), the overall WQI values per station exhibited a range from 80.10 ± 6.37 to 87.30 ± 5.54, falling within the “good” WQI category. This uniformity indicated a spatial conformity in water quality distribution across the stations. Stations L6 and L7 recorded the highest WQIs (87.30 ± 5.54 and 86.90 ± 5.25), indicating minimal pollution due to limited human activity and absence of agricultural or residential development. Similar results were reported by Xiao et al. [48], where the Beichuan River’s upstream area had a higher water quality compared to midstream and downstream sections because of less anthropogenic activity. However, the water quality degraded gradually from upstream to downstream. Stations L1 and L5 both marked a “good” level of Water Quality Index, lower than the other stations, with recorded values of 80.10 ± 6.37 and 80.62 ± 7.33, respectively. This lower WQI was linked to extensive agricultural activities in region L5. Additionally, L1 at the outlet received all the flows coming from the river, which explains the lower WQI. Conversely, stations L2, L3, and L4 presented comparable WQIs, with values of 82.69 ± 5.24, 83.16 ± 5.09, and 83.9 ± 5.73, respectively. In addition, Maity et al. [41] found that the first sites had similar and comparable WQI values compared to other sites. However, a contrast was recorded comparing the WQI to the spatial PC of neighboring stations. This contrast indicated an initial sign of pollution, due to the environmental conditions such as residential areas and agricultural lands. This proves the effect of habitation and agricultural activities on the water quality. Calmuc et al. [49] also explained that the stations located near the agricultural and industrial areas had a lower water quality.

The analysis of WQI during both low flow (0.015 m³/s) and high flow (55.08 m³/s) is illustrated in Table 7. Clearly, the average WQI during high flow conditions (86.05 ± 6.84) exceeded the WQI during low flow (81.08 ± 4.17), indicating a “good” water quality during both high and low flow events. The enhanced WQI during high flow can be assigned to the dilution process. This process led to an overall improvement in water quality during high flow conditions. Conversely, pollution concentrations increased during low flow periods, mainly due to anthropogenic activities that contributed to higher pollutant discharge during low flow events. When comparing our results with the water quality of the Verdinho River in Brazil, the values obtained for the WQI allowed classifying the water body in the category of good quality due to the dilution of the potential pollutants during the rainy season. On the other hand, compared to the rainy season, the average water quality during the dry season was lower and was categorized as regular and good [50].

These results provided a comprehensive view of water quality within the watershed. The average WQI for the entire watershed was 83.70 ± 4.97, indicating a “good” water quality. The fluctuations in WQI ranged from 71.90 ± 1.91 to 94.73 ± 2.32. Furthermore, these findings correlated with El Najjar et al. [6] in their study during 2016 and 2017. WQI classification demonstrated a positive correlation with flow conditions, displaying a “good” WQI during high flow and a “medium” WQI during low flow periods.

4. Conclusions

The comprehensive spatio-temporal study carried out on the Ibrahim River across seven distinct stations throughout the hydrological year 2021–2022, through the measurement of various physico-chemical and microbiological parameters, provided an overall watershed assessment.

The application of PCA explained the interaction between these parameters. PC1 highlighted spatial variations; in particular, station L5 presented a remarkable deviation, indicating potential early signs of pollution, due to specific anthropogenic activities in the area. Furthermore, stations L6 and L7 demonstrated low scores, suggesting better water quality conditions. The other stations in the mixed season group, influenced by the presence of habitation and agricultural lands, marked the necessity for continuous monitoring of parameters, even those falling within permissible ranges. PC2 highlighted temporal variations within the Ibrahim River’s water quality. The summer season showed a distinctive trend, displaying elevated scores, indicating a decline in water quality. This showed a potential sign of pollution during this season, linked to human activities and increased water resources demand. Conversely, the springtime trend presented lower scores, thus better water quality conditions during this period. The mixed seasons group summarized the river’s quality across seasons, reflecting good water quality. Remarkably, these PCA findings aligned with the WQI assessment of the entire river, validating both analytical approaches. Consequently, spatial and temporal analyses and the WQI calculated potential pollution sources; therefore, there is a need for strategies to preserve the watershed.

For better watershed monitoring and assessment, we recommend targeted monitoring efforts at station L5, where elevated pollutant levels result from anthropogenic impact. Special attention should be given during the dry season, when reduced flow increases contamination. Implementing buffer zones, regulating agricultural runoff, and conducting seasonal assessments will preserve water quality across the river.