According to Mustafa et al. (2017) who conducted a study on evaluating river water quality index while utilizing remote sensing and AI models measured 13 water quality parameters including turbidity, dissolved Oxygen, Potassium, Nitrate. The parameters were measured between two dates at a site in different dates.

Chabuk et al. (2020) conducted the quality of water on Tigris river using GIS and remote sensing where they tested 14 different locations measuring TDS, pH, Turbidity, chlorophyl A, Blue-green algae and dissolved oxygen were the testing parameters. They employed the use of spectral bands and indices in LASSO model where they found that the water quality Index (WQI) was low indicating poor water quality along the stream ( Jonnalagadda & Mhere, 2001 ).

Othe researchers have conducted a review on applicability of remote sensing and GIS in river relation studies. The authors conducted a review of sixty-nine research papers. Based on the scrutiny the authors found out that remote sensing data, popular ones being Landsat, sentinel and digital Elevation Model datasets, with the extracted information being indices for main parameter for understanding the rivers.

Adilakshmi & Venkasetan (2024) conducted a study on river Noyyal to understand the water quality nature. This research also utilized Lasso model for predicting the water quality index noting that the model provided a cost-effective method of monitoring the water quality in the river. The study also showed the importance of WQI in indicating the quality of water during the monitoring period showing the magnitude of pollution.

Water pollution has not only negative impacts on human but also agriculture, land and aquatic life below are some of the effects:

Water bodies are contaminated by industrial and household detergents, which have detrimental impacts on plants. Water becomes enriched in phosphates due to the high phosphate content of detergents. Through surface absorption or roots, phosphates enter plants and cause a variety of negative effects, including denaturation of proteins, destruction of chlorophyll and cell membranes, elongation of roots, carbon dioxide fixation, photosynthesis, cation uptake, pollen germination, and growth of pollen tubes ( Singh & Gupta, 2017 ).

Overuse of pesticides, herbicides, fertilizers, and other chemicals on crops results in runoff that carries the chemicals into the soil and eventually reaches water bodies. Fertilizer chemicals cause eutrophication by nutrient enrichment. Water becomes acidic due to the acidic nature of ammonium from fertilizers. In a similar vein, insecticides, herbicides, and pesticides similarly alter the pH of water bodies. The decrease in photosynthetic rate is these drugs’ most frequent side effect. Some could impede or decouple oxidative phosphorylation. Due to their limited solubility in water, these compounds have considerable capacities for absorption and bioaccumulation in macrophytes plants ( Singh & Gupta, 2017 ; Craswell, 2021 ; Watelet and Johnson, 1999 ).

Globally, an estimated 20 million hectares of arable land are irrigated with water that has been partially or completely treated by leftover amounts of household liquids. Because of the presence of heavy metals, which are deemed hazardous due to their potential toxicity and environmental persistence, there are dangers to ecosystems and public health. Since they fall into two groups, heavy metals and trace elements in the soil system are particularly interesting topics. In contrast, cadmium (Cd), lead (Pb), chromium (Cr⁶⁺), nickel (Ni), mercury (Hg), and arsenic (As) are not only not essential to plants but also toxic even at very low concentrations. Metals like copper (Cu), zinc (Zn), manganese (Mn), iron (Fe), and molybdenum (Mo) have been identified as essential micronutrients for plant growth. Heavy metal buildup in agricultural soils has the potential to have a detrimental impact on human health by moving up the food chain from soil to plants to people.

Some research on River Njoro has shown that there is a clear growing human activity that results in the discharge of organic wastewater which has a significant negative influence on the Njoro River. Organic waste discharges have led to a decline in water quality. The diversity and composition of macroinvertebrate species have changed as a result of changes in the physico-chemical water quality parameters. Thus, along the lower research sites, pollution-tolerant Polypedilum sp. became dominant. These findings unequivocally demonstrate how rising human activity levels in the Njoro River watershed, in particular, affect the quality of the water. The study’s findings show that a damaged aquatic ecosystem will alter the community structure of aquatic fauna and impede crucial ecological functions, such as the organism’s innate ability to purify itself. Therefore, it is imperative to preserve the ecological integrity of the Njoro River water resource by limiting human activity, protecting the river’s channel and basin, and raising public knowledge of environmental integrity through education. Therefore, all parties involved should prioritize developing a management plan and putting rules in place to reduce and regulate pollution in the Njoro River.

Polluted water has detrimental effect on the aquatic life as illustrated by Malik et al. (2024) .

A high concentration of suspended water contaminants can interrupt the normal behavior of fish populaces. Many fish species, such perch and brown trout, are highly sensitive to elevated suspended solids and exhibit severe avoidance behavior ( Malik et al., 2024 ). These species rely on sight to quickly capture their meal. When a fish species lives in an environment with murky water, suspended materials may damage or block the gill aperture, decreasing the fish’s susceptibility to certain diseases and parasites. Fish species may also ingest these suspended particles, which exposes them to possible poisons or pathogens on the sediment and can cause disease.

Whether they are submerged, floating, or emergent, macrophytes are aquatic plant species that grow in water. Because they react swiftly to changes in the pace and variety of several environmental factors, including as water flow, alkalinity, substrate, shade, and nutrient concentrations, macrophytes are considered bioindicators. The total amount of suspended sediment has a negative effect on aquatic macrophytes and algae by reducing the amount of light that reaches the water column, which in turn reduces the frequency of photosynthesis. Fast flowing rates frequently carry large amounts of suspended particles, which also brush aquatic macrophytes and algae away from bed substrates, damaging their photosynthetic structures. The submerged flora may be choked by the sedimentation process, drastically slowing down photosynthesis. While certain plants thrive in water with high DO and little dissolved oxygen, others do well in water with more nutrients. Because of this, aquatic macrophytes might be ranked and scored based on how well they adapted to different chemical and physical environments. Aside from that, macrophytes are robustly predisposed by soil type and geology, and they can withstand unpredictable pollution. Furthermore, the organization of macrophyte communities is often determined by a number of interrelated factors, which can make it challenging to attribute a species’ presence or absence to a particular contaminant in any body of water, macrophytes are an essential bio indicator of ongoing pollution issues ( Malik et al., 2024 ).

Our study area is along River Njoro/Ndarugu River, a third order stream in Kenya, Nakuru county. River Njoro watershed is located Kenya’s southwestern Rift Valley at 0˚30' South, 35˚20' East. The river is approximately 50 km in length with an estimated 270 km² contributing source area, originating in the Eastern Mau Escarpment at approximately 3000 m. River Njoro winds through forested and agricultural lands before serving several urban settlements and finally emptying in Lake Nakuru, a shallow salty lake. Figure 1 below shows the study area.

The main economic activities in the area are agriculture-based industries including vegetable and milk processing, large-scale wheat, maize, barley farming and green house flower farming.

For the assessment of water pollution using GIS and Remote sensing methods we incorporated a number of data. Data were acquired for the years 2000, 2010 and 2020. Nitrogen, Phosphorous and Biological Oxygen Demand were calculated from satellite images from different spectral reflectance using the raster calculator tool. Satellite images were obtained from the US Geological Survey (USGS) Earth Explorer Landsat 7 for 2000 processes, Landsat for 2010 processes and Landsat 8 for 2020 data processing. Land Use Land Cover maps for the corresponding years were generated using satellite images obtained from USGS. Table 1 above shows the data used.

The study area was divided into five zones;

1) near the source, Eastern Mau Forest.

2) around Egerton dams and Njoro caves.

3) around Njoro near Kenyatta stage.

4) around Kaptembwa.

5) around Nakuru sewage treatment plant.

Calculate water quality indices that are correlated with Total Nitrogen (TN), Total Phosphorous (TP) and Biological Oxygen Demand (BOD) using Landsat imagery.

The analysis of the data was carried out in ENVI, ArcGIS software and quantifying the output using Microsoft excel. Several steps in image processing and analysis were carried out in order to obtain the final results. Processing and integration were done using ArcMap software and ENVI tools. Integrating remote sensing data with GIS allows spatial analysis and mapping of water quality parameters. GIS provides tools to overlay and visualize water quality data, analyze trends, identify hotspots of pollution, and assess changes over time ( Chabuk et al., 2020 ). Land use land cover maps were generated for the three epochs (2000, 2010, 2020) to see the influence or impacts it has to the water quality parameters. The maps were generated using supervised classification, a process where you select representative samples for each land cover class in the satellite image to identify land cover classes in the entire image. To assess water pollution, it’s essential to identify the relevant water quality parameters to be measured. In this study, we focused on the Common parameters; Total Phosphorous (TP), Total Nitrogen (TN), and Biological Oxygen Demand (BOD).

The methodology ( Figure 2 ) employs the utilization of remotely sensed satellite data which undergoes image enhancement for utilization in subsequent image processes. The chosen image for the year 2000, 2010 and 2020 undergoes processing for pollution mapping for the study river with the three indices being utilized for the investigation generated the Nitrogen Differential Nitrogen Index (NDNI), Normalized Difference Phosphorus Index (NDPI), Normalized Difference Water Index (NDWI). An overlay analysis is carried out and finally a validation and correlation carried out the corresponding statistics for the degree of impact and are coverage are also generated for each year.

The technology ( Figure 2 ) utilizes remotely sensed satellite data, which is enhanced for subsequent image treatments. The selected image for the years 2000, 2010, and 2020 is processed to create pollution maps for a study on a river. Three indices are used for the investigation: The Nitrogen Differential Nitrogen Index (NDNI), the Normalized Difference Phosphorus Index (NDPI), and the Normalized Difference Water Index (NDWI). An overlay analysis is conducted, followed by a validation and correlation process. Statistics regarding the degree of impact and coverage are generated for each year.

Nitrogen (N) in the form of nitrate is a common pollutant in surface and ground waters. The source of Nitrate-N is leach down beyond the root zone in agricultural soils and reach the ground and surface waters ( Jonnalagadda & Mhere, 2001 ). Nitrogen Differential Nitrogen Index (NDNI) is helpful in identifying areas affected by stressors such as pollution, disease, or physical damage. It’s calculated from the reflectance values of shortwave infrared (SWIR) and red (R) bands;

$\text{NDNI}=\left(\text{SWIR Band}-\text{RED Band}\right)/\left(\text{SWIR Band}+\text{RED Band}\right)$

2000

$\text{NDNI}=\left(\text{Band}5-\text{Band}3\right)/\left(\text{Band}5+\text{Band}3\right)$

NDNI for 2000 ranged from very low to very high. The areas with very low NDNI values were majorly observed on the forested areas while very high NDNI values were majorly on the urban areas as shown in Figure 3 and Table 2 .

2010

$\text{NDNI}=\left(\text{Band}6-\text{Band}4\right)/\left(\text{Band}6+\text{Band}4\right)$

NDNI for 2010 ranged from very low to very high ( Figure 4 ). The areas with very low NDNI values were majorly observed on the forested areas while very high NDNI values were majorly on the urban areas ( Table 3 ). The NDNI values were increasing as compared to 2000 NDNI values.

2020

$\text{NDNI}=\left(\text{Band}6-\text{Band}4\right)/\left(\text{Band}6+\text{Band}4\right)$

NDNI for 2020 ranged from very low to very high ( Figure 5 ). The areas with very low NDNI values were majorly observed on the forested areas while very high NDNI values were majorly on the urban areas ( Table 4 ). The NDNI values were increasing as compared to 2010 and 2000 NDNI values as shown in Table 5 and graphically in Figure 6 below.

Available quantity of phosphorous controls the pace at which algae and aquatic plants are produced. In appropriate quantities, phosphorus can be used by vegetation and soil microbes for normal growth. However, in excess quantities, phosphorus can lead to water quality problems such as eutrophication and harmful algal growth. Phosphorus generally occurs in small quantities in the natural environment so even small increases can negatively affect water quality and biological condition. The main sources of phosphorus pollution are sewage effluent (primarily from water industry sewage treatment works) and losses from agricultural land.

Water quality index that can be correlated with Phosphorous is the Normalized Difference Phosphorous Index (NDPI). NDPI is sensitive to changes in Phosphorous content, making it suitable for water quality assessments. NDPI is calculated using Red Band and the Near Infrared band;

$\text{NDPI}=\left(\text{RED Band}-\text{NIR Band}\right)/\left(\text{RED Band}+\text{NIR Band}\right)$

2000

$\text{NDPI}=\left(\text{Band}3-\text{Band}4\right)/\left(\text{Band}3+\text{Band}4\right)$

NDPI for 2000 ranged from very low to very high as shown in Figure 7 and Table 6 . The areas with very low NDPI values were majorly observed on the forested areas while very high NDPI values were majorly on the urban areas.

2010

$\text{NDPI}=\left(\text{Band}4-\text{Band}5\right)/\left(\text{Band}4+\text{Band}5\right)$

NDPI for 2010 ranged from very low to very high (see Figure 8 and Table 7 ). The areas with very low NDPI values were majorly observed on the forested areas while very high NDPI values were majorly on the urban areas. The NDPI values were increasing as compared to 2000 values.

2020

$\text{NDPI}=\left(\text{Band}4-\text{Band}5\right)/\left(\text{Band t}4+\text{Band}5\right)$

NDPI for 2020 ranged from very low to very high ( Figure 9 and Table 8 ). The areas with very low NDPI values were majorly observed on the forested areas while very high NDPI values were majorly on the urban areas. The NDPI values were increasing as compared to 2010 and 2000 NDPI values as shown in Table 9 and Figure 10 below.

Biological oxygen demand is the measure of the amount of oxygen required to remove waste organic matter from water in the process of decomposition by aerobic bacteria. The presence of high BOD may indicate contamination or increases in particulate and dissolved organic carbon from non-human and animal sources that can restrict water use and development, necessitate expensive treatment and impair ecosystem health. We computed BOD by correlating with Normalized Difference Water Index (NDWI). NDWI is sensitive to changes in water content, making it suitable for water quality assessments;

$\text{NDWI}=\left(\text{G}-\text{NIR}\right)/\left(\text{G}+\text{NIR}\right)$

2000

$\text{NDWI}=\left(\text{Band}2-\text{Band}4\right)/\left(\text{Band}2+\text{Band}4\right)$

NDWI for 2000 ranged from very low to very high (see Figure 11 and Table 10 ). The areas with very high NDWI values were majorly observed on the urban areas while very low NDWI values were majorly on the forested areas.

2010

$\text{NDWI}=\left(\text{Band}2-\text{Band}4\right)/\left(\text{Band}2+\text{Band}4\right)$

NDWI for 2010 ranged from very low to very high as shown in Figure 12 and Table 11 . The areas with very high NDWI values were majorly observed on the urban areas while very low NDWI values were majorly on the forested areas.

2020

$\text{NDWI}=\left(\text{Band}3-\text{Band}5\right)/\left(\text{Band}3+\text{Band}5\right)$

NDWI for 2020 ranged from very low to very high (see Figure 13 and Table 12 ). The areas with very low NDWI values were majorly observed on the forested areas while very high NDWI values were majorly on the urban areas. The NDWI values were changing as compared to 2010 and 2000 NDWI values as shown in Table 13 and Figure 14 below.

Normalized Difference Moisture Index (NDMI) is a remote sensing technique used to assess vegetation moisture content and monitor changes in water availability over large areas. It is calculated by taking the difference between the near-infrared (NIR) and short-wave infrared (SWIR) bands of satellite imagery, normalized by their sum. Higher NDMI values indicate higher moisture content, while lower values suggest drier conditions showing pollution effects;

$\text{NDMI}=\left(\text{NIR}-\text{SWIR}\right)/\left(\text{NIR}+\text{SWIR}\right)$

2000

$\text{NDMI}=\left(\text{Band 4}-\text{Band 5}\right)/\left(\text{Band 4}+\text{Band 5}\right)$

NDMI for 2000 ranged from very low to very high (see Figure 15 and Table 14 ). The areas with very high NDMI values were majorly observed on the forested areas while very low NDMI values were majorly on the urban areas.

2010

$\text{NDMI}=\left(\text{Band 5}-\text{Band 6}\right)/\left(\text{Band 5}+\text{Band 6}\right)$

2020

$\text{NDMI}=\left(\text{Band 5}-\text{Band 6}\right)/\left(\text{Band 5}+\text{Band 6}\right)$

NDMI for 2020 ranged from very low to very high. The areas with very high NDMI values were majorly observed on the forested areas while very low NDMI

values were majorly on the urban areas as shown in Figure 16 and Table 15 . The NDMI values were increasing as compared to 2010 (shown in Figure 17 and Table 16 ) and 2000 NDMI values as shown in Table 17 and annual NDMI analysis Figure 18 below.

Pollution map for 2000 ( Figure 19 ) ranged from very low to very high as shown in Table 18 . The areas with very low pollution values were majorly observed on the forested areas while areas with very high pollution values were majorly on the urban areas.

Pollution map for 2010 ranged from very low to very high (see Figure 20 and Table 19 ). The areas with very low pollution values were majorly observed on the forested areas while areas with very high pollution values were majorly on the urban areas. The pollution values were increasing as compared to 2000 pollution values.

Pollution map for 2020 (see Figure 21 ) ranged from very low to very high as shown in Table 20 . The areas with very low pollution values were majorly observed on the forested areas while areas with very high pollution values were majorly on the urban areas. The study demonstrated that the level of pollution to Njoro river has been increasing from 2000 to present as illustrated in Table 21 and Figure 22 below. This is due to the increase in Nitrogen, Phosphorous and BOD levels overtime as shown in Figure 24 .

River Njoro flows from the source in the eastern Mau into Lake Nakuru. We selected the sampling points along river Njoro by taking into consideration the land use land cover along the river and the accessibility to the rivers, that is, the way/route to the sampling points through either the main roads or the feeder roads and we ensured that the samples were uniformly distributed along the river. We ensured that we covered all the LULC classes, that is the forested area which was mainly along the source of the river in the Eastern Mau region, the croplands area which is majorly around Nassuit ward, and Njoro, the grassland area which were somewhere near the mouth of the river in Mwariki area, Bare land areas which were in areas near Kaptembwa, and Kapkures bridge and the built-up areas which were areas around Njokerio, Njoro and the outskirts of Nakuru. We visited sites near the source, Eastern Mau Forest around Tiritagoi, Nassuit, Sigotik, Kapkatet, Mwigito, Ahero, Kenyatta stage, Njoro, Kaptembwa, Rhoda, Kapkures, Baruti, Mogoon, Mwariki and Park View, near the mouth as shown below.

The observation showed an increase in pollution as we approached the mouth, as shown in Figures 23-27 , and Figure 28 showing higher pollution as compared to Figure 19 , Figure 20 , and Figure 21 . Which were areas showing lower levels of pollution near the source of the river.

During the site validation visit, several observations were made. Figure 29 shows the local laundry works Kaptembwa bridge, Figure 30 shows heavy sewage effluent contaminating the river at Rhoda area, Figure 31 and Figure 32 show water stagnating at Kapkures bridge and sedimentation and garbage disposal to the river respectively. Finally, eutrophication caused by sewage contamination with river indicating presence of Nitrogen and Phosphorus (see Figure 33 ).