Residues of organochlorine pesticides in aquatic ecosystems of Southwestern Ethiopia: implications for ecological risks

Abstract

Organochlorine pesticides (OCPs) remain a critical environmental concern due to their persistence and bioaccumulative potential. This study investigated OCP residues across three matrices—water, sediment, and macroinvertebrates—within the Awetu, Kito, and Boye aquatic systems in southwestern Ethiopia. Analysis using Gas Chromatography with Electron Capture Detector (GC-ECD) identified nine OCPs, with p, p′-DDT and its metabolite p, p′-DDE as the dominant contaminants. Concentrations varied across the three matrices (water: ng/L; sediment and macroinvertebrates: ng/g), and values are not directly comparable due to differences in units, matrix properties, and analytical methodologies. The screening-level Ecological Risk Assessment (ERA) indicated potential high ecological risk to aquatic biota. Risk Quotients (RQ) for macroinvertebrates showed extreme risk from heptachlor epoxide (RQ = 738.8) and high risk from dieldrin (RQ = 1.53), demonstrating that these legacy pollutants bioaccumulate to levels that may cause secondary poisoning to higher trophic levels. Statistical analysis confirmed significant variation in pesticide distribution across the three matrices (p < 0.001), with a contamination hierarchy of sediment > macroinvertebrates > water. This pattern is primarily driven by the high lipophilicity and environmental persistence of OCPs, which favor adsorption to organic-rich sediments and subsequent bioaccumulation in aquatic biota. In contrast, the low concentrations in the water column reflect the poor aqueous solubility of these hydrophobic compounds. These findings highlight the persistence of legacy OCPs in Ethiopian aquatic ecosystems despite existing bans, with the Boye-Awetu-Kito system acting as a long-term reservoir. The extreme ecological risk observed underscores the urgent need for targeted remediation of the Boye wetland and stronger regulatory enforcement to mitigate trophic transfer and protect the regional aquatic ecosystem.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-50749-y.

Introduction

Freshwater ecosystems are vital global resources, yet they are increasingly threatened by persistent organic pollutants like organochlorine pesticides (OCPs). In Ethiopia, wetlands in the southwest provide essential services including biodiversity conservation, medicinal resources, and clean water¹. However, the affordability and effectiveness of OCPs for agricultural productivity and malaria vector control in sub-Saharan Africa have led to their continued environmental presence². Due to their hydrophobic nature and low water solubility, OCPs resist degradation and preferentially partition into organic-rich sinks, such as sediments and aquatic biota³.

Despite the ecological sensitivity of southwestern Ethiopian water bodies like the Awetu, Kito, and Boye wetland systems, environmental monitoring remains limited and fragmented⁴. Most existing studies in the region focus on isolated matrices, such as water or sediment alone, failing to capture the complex partitioning and bioaccumulative potential of OCPs across the aquatic food chain⁵. Macroinvertebrates serve as critical biological integrators of these pollutants⁶ and yet data regarding their body burden and ecological risks for this indicator biota in relation to the surrounding abiotic environment remains scarce. There is a clear need for a comprehensive, multi-matrix approach to understand the true toxicological load within these ecologically linked systems.

This study addresses this gap by conducting a concurrent assessment of OCP residues across three interconnected matrices: water, sediment, and macroinvertebrates. By evaluating the distribution and partitioning of target pesticides, this research aims to identify contamination patterns and perform a screening-level ecological risk assessment (ERA). By utilizing macroinvertebrates as sentinel organisms, the study calculates Risk Quotients (RQ) to determine the toxicological pressure on the aquatic community and the potential for trophic transfer. Ultimately, these findings will serve to generate the baseline data essential for informing stricter regulatory oversight to promote sustainable pesticide management practices in the Ethiopian Rift Valley.

Methods

Study area description

The study was conducted in the Jimma zone of southwestern Ethiopia (7°41’ N, 36°50’ E; Study sites are illustrated in Fig. 1, located approximately 355 km from Addis Ababa. The region is characterized by a warm, humid climate with mean annual temperatures ranging from 11.5 °C to 27.1 °C and high annual rainfall (~ 1749.3 mm), making it one of the most productive agricultural highlands in the country⁷. The local hydrology is defined by the Awetu and Kito streams, which traverse Jimma town before converging southward into the Boye wetland⁸. Subsequently, their confluence joins the Gilgel Gibe River, home to a notable hydroelectric power.

Fig. 1.

Map of the study sites in the Jimma zone, southwestern Ethiopia. Sampling locations were distributed across the Awetu (urban), Kito (peri-urban), and Boye (wetland/sink) reaches.

Study design and strategy

A cross-sectional study was conducted to quantify persistent organochlorine pesticide (OCP) residues in the aquatic ecosystem. To ensure that the 42 samples (n = 14 per matrix: water, sediment, and macroinvertebrates) reflected the entire spatial gradient of the water sources, a stratified systematic sampling design was employed. The watershed was stratified into three reaches based on anthropogenic pressure: urban-industrial (Awetu), peri-urban (Kito), and wetland/sink (Boye), within these strata, probabilistic sampling points were selected to capture peak anthropogenic pressures. This approach ensures that the data represents the hydro-geographic gradient rather than a convenience-based selection, satisfying the requirement for robust interpretive inference. Sample collection and extraction were carried out following standard methods as outlined by the Environmental Protection Agency⁹. Proper labeling of sample identification codes, dates, water source names, and sampling locations was conducted before placing them into the sample containers, adhering to the specified sampling procedures for each matrix. A total of forty-two samples were collected, comprising fourteen samples from each matrix: water, sediment, and macroinvertebrates.

Sample collection and in-situ measurements

Water samples

Composite water samples were collected using grab sampling (at a depth of 15–30 cm below the water surface) method into one-liter polyethylene bottles, ensuring minimal disturbance to aquatic organisms as stated by¹⁰. Prior to sampling, pre-cleaned the bottles (sequential wash: water, distilled water, and acetone), followed by thorough drying in an oven. In situ water quality parameters, including temperature, pH, electrical conductivity (EC), and dissolved oxygen (DO), were measured using a calibrated portable multiparameter probe (HACH HQ30d, USA). Turbidity was determined using a portable turbidity meter (HACH 2100P, USA). To ensure analytical precision and prevent cross-contamination, the sensors were rinsed with distilled water between measurements and allowed to stabilize before recording final values. The cuvette underwent a thorough cleaning process with distilled water and the water sample before being filled with the sample. After drying the cuvette’s outer surface, it was swiftly inserted into the turbidimeter, automatically measuring the Nephelometric Turbidity Unit (NTU) values.

Sediment samples

Sediment samples weighing 1 kg were precisely gathered using a stainless-steel grab from identical locations as the water samples. Sampling was conducted from a depth of about 0 to 5 cm beneath the surface. Post-collection, any pebbles and wood debris were manually eliminated.

To ensure an accurate portrayal of each sampling site, the collected sediment was meticulously mixed to blend its unique attributes. Subsequently, the samples were carefully placed in polyethylene bags, labeled, and secured in cold boxes. The samples were stored at −18 °C until extraction and further analysis, guaranteeing their preservation for accurate assessment.

Macroinvertebrates samples

Macroinvertebrate samples were obtained using a 10-minute kick sampling technique at each sampling site, following standardized protocols for rapid bioassessment^11,12. This method involved disturbing the substrate upstream of a standard kick net to ensure a representative collection of the benthic community across all available microhabitats. A D-frame net with a 250 μm diameter mesh size was employed for this purpose. The net was kicked and dragged across a 10 m distance, disturbing the bottom sediment to capture benthic macroinvertebrates. Upon collection, the macroinvertebrates were promptly sorted on-site and preserved in vials filled with 70% ethanol to maintain their integrity.

Subsequently, all water, sediment, and macroinvertebrate samples were immediately transported to the laboratory at the Department of Environmental Health Science and Technology, Jimma University. They were then stored in a deep refrigerator until extraction, ensuring their preservation for subsequent analysis.

Chemical and reagents

All chemicals and reagents utilized were of analytical grade, while the solvents met HPLC grade standards. Organic solvents, including acetone, n-hexane, and methanol, were sourced from BDH Chemicals Ltd (Poole, England). Sodium chloride (NaCl) and anhydrous sodium sulfate (Na2SO4) were acquired from BDH Chemicals Ltd and Sigma Aldrich (St. Louis, MO, USA), respectively. Ultrapure water was obtained through purification using the Mill-Q water purification system by Millipore (Bedford, France). For filtration, Whitman filter paper (grade 1, size 8.5 cm) and 3 μm nylon filters were procured from Whitman International Ltd (Maidstone, England). These filters were instrumental in separating particulate matter and impurities from water, sediment, and macroinvertebrate samples, ensuring cleaner samples for subsequent analysis. Analytical standards of various compounds were sourced from Sigma Aldrich (St. Louis, MO, USA) for use in the study.

Preparation of stock standard solution

To prepare the stock solutions for each pesticide, precisely 10 mg of each analyte was weighed and dissolved in 10 mL of methanol, yielding a concentration of 1000 mg/L. These stock solutions were then stored at 4 °C for future use. Following this, an intermediate standard solution with a concentration of 100 ppm was generated by combining appropriate volumes of each individual stock solution. This intermediate solution acted as a benchmark for further dilutions. Subsequently, a working standard solution of the desired concentration was prepared by diluting the intermediate standard solution. For the calibration curve, standard solutions with concentrations of 0.05, 0.1, 0.5, 2.5, 10, 50, and 100 µg/L were meticulously prepared. The linearity and accuracy of each pesticide’s calibration curve were confirmed by coefficients of determination (r²) exceeding 0.995, falling within the analytical range. These steps were crucial for the precise determination of physicochemical parameters in both water and sediment samples.

Physicochemical parameters determination for water and sediment samples

Water quality analysis

In situ parameters (temperature, pH, EC, and DO) were measured using a calibrated portable multiparameter probe (HACH HQ30d, USA). Turbidity was determined with a HACH 2100P turbidity meter. Sensors were rinsed with distilled water between measurements to prevent cross-contamination. Other physicochemical parameters, including total dissolved solids (TDS), total suspended solids (TSS), alkalinity, and acidity, were analyzed following standard gravimetric and titrimetric methods. TDS was measured by filtration and evaporation. TSS was determined by filtering 100 mL of sample through pre-weighed Whatman filter paper and drying at 105 °C. Alkalinity and acidity were estimated by titration with sulfuric acid and sodium hydroxide, respectively, using phenolphthalein and methyl red indicators.

Sediment characteristics

Sediment pH and electrical conductivity were measured in a 1:2.5 sediment-to-deionized water suspension after a 15-minute stabilization period. Moisture content and total organic carbon (TOC) were determined gravimetrically on 10 g homogenized sub-samples sieved through a 300 μm mesh. Moisture content was calculated from weight loss after drying at 105 °C overnight. TOC was determined using the loss-on-ignition (LOI) method on moisture-free samples.

Extraction and clean-up

Water samples

Samples were immediately transported to the laboratory in ice boxes and stored at − 18∘C. To maintain analyte integrity, water samples were processed within 7 days of collection, while sediment and macroinvertebrate samples were extracted within 14 days and stored in deep refrigerator (at − 18 °C) until preparation for the further experiments¹³.

Water samples were extracted using dispersive liquid-liquid microextraction (DLLME). Prior to extraction, samples were filtered through Whatman filter paper to remove particulates and homogenized. A 5 mL aliquot was rapidly mixed with 500 µL acetone (disperser solvent) and 50 µL n-hexane (extraction solvent). After 2 min, the cloudy solution was centrifuged at 4000 rpm for 4 min. Approximately 40 µL of the settled organic phase was collected, and 1 µL was injected into the GC-ECD¹⁴. A 5 mL water sample in a 15 mL conical glass flask was mixed with 50 µL n-hexane (extraction solvent) and 500 µL acetone (dispersive solvent) maintaining a 1:10 extractant-to-disperser ratio to maximize analyte recovery as described by Rezaee and his colleagues¹⁵.

After 2 min, the cloudy solution was centrifuged at 4000 rpm for 4 min. Approximately 40 µL of the settled organic phase was collected, and 1 µL was injected into the GC-ECD for analysis.

Sediment samples

Sediment samples were air-dried for 72 h after removing visible debris. A 25 g sub-sample was ground, sieved through a 300 μm mesh, and homogenized with 20 g anhydrous sodium sulfate using a mortar and pestle. The mixture was extracted with 130 mL acetone: n-hexane (volume ratio not specified in original) by shaking on an orbital shaker for 2 h. The extract was dried over 50 g anhydrous sodium sulfate, filtered, and concentrated to 1 mL using a rotary evaporator. The concentrate then underwent column chromatography clean-up using activated silica gel (80–100 mesh)¹⁶ after activated for 16 h at 130 °C.

The silica gel was deactivated to 3.3% with reagent water in a 500 mL glass jar, mixed thoroughly, and allowed to equilibrate for 6 h. It was stored in a sealed glass jar. The chromatographic column (30 cm × 9 mm I.D.) was packed with 5 g deactivated silica gel slurry in 1.5% (v/m) n-hexane and topped with 3 g anhydrous sodium sulfate. The column was rinsed with 10 mL n-hexane. The extract was eluted with 80 mL n-hexane, concentrated to dryness using a rotary evaporator at 45 °C, and reconstituted in 2 mL n-hexane for GC analysis.

Macroinvertebrates samples

Macroinvertebrates (MI) samples were gently washed with distilled water to remove traces of ethanol for preservation. Soxhlet extraction was applied for the extraction of OCPs from MI. Macroinvertebrates were identified to family level using a stereomicroscope (10× magnification) and the identification key applied by¹⁷.

Macroinvertebrate samples were washed with distilled water followed by acetone and air-dried for 3 days at room temperature. The dried samples were ground and homogenized with anhydrous sodium sulfate. A 1 g sub-sample mixed with 5 g anhydrous sodium sulfate was placed in a cellulose thimble and extracted with 30 mL acetone: n-hexane (20:80 v/v) in a Soxhlet extractor at 45 °C for 4 h. The extract was filtered, concentrated to 1 mL using a rotary evaporator, and purified by column chromatography on a silica gel column layered with anhydrous sodium sulfate. The eluate (2 × 10 mL n-hexane) was concentrated to dryness at 40 °C and reconstituted in 2 mL n-hexane. A 1 µL aliquot was transferred to vials for GC-ECD analysis.

Instrumentation and GC-ECD analysis

The determination of organochlorine pesticides in water, sediment, and macroinvertebrate samples was performed using Gas Chromatography with an Electron Capture Detector (GC-ECD, Agilent Technologies 7890 A) equipped with an auto-sampler. An HP-5 capillary column (30 m × 0.25 mm i.d., 0.25 μm film thickness; 5% phenyl methyl siloxane) was employed. The oven temperature program and detailed chromatographic conditions are provided as Supplementary Table S1 (annexed). Nitrogen (99.999%) was used as the carrier gas at a flow rate of 20 mL/min. A 1 µL sample was injected in split mode (10:1) at an injection temperature of 280 °C. Each sample was injected in triplicate, and quantification was based on peak areas using calibration standards, while identification relied on retention times of known organochlorines.

Method validation and analytical performance

To satisfy environmental analytical standards, the method was validated as follows: Linearity, accuracy, and detection limits were assessed as follows. For individual OCP calibration, a seven-point external calibration curves (0.05–100 µg/L) were constructed. Linearity was confirmed with coefficients of determination (r2 ≥ 0.999). To evaluate the performance of the analytical procedures and the accuracy, recovery (% recovery), studies were undertaken for each pesticide¹⁸. Method blanks (solvent blanks) were run at the beginning of each batch and additionally between sample runs to check for carryover and contamination. No significant analyte signals were detected in the blanks. Recovery study were determined by spiking the previously analyzed samples with the pesticides standard¹⁹.

For water samples, 0.5 ppb of mixed standard solution was spiked in deionized water of blank matrices then, spiked samples were then analysed using the same analytical procedures as field sample. A mixed standard solution with concentration of 100 and 50 ppb was added in sediment and macro-invertebrates’ samples respectively to inquire the spike recovery percent of the nine organochlorine pesticides.

Where, CS₁ = concentration of pesticide residues in the sample, CS₂ = concentration of pesticide residues in the spiked sample, and CS = concentration of added pesticide standard.

The mean recoveries were Water (72.4–104.6%), Sediment (70.0–112.5%), and Biota (76.0–104.0%), all within the acceptable 70–120% range. affirming method accuracy. The limit of detection (LOD) and Limit of quantifications (LOQ) were determined as the lowest concentrations yielding a signal-to-noise (S/N) ratio of 3 and 10, respectively. One limitation of this study is that, internal standards were not employed due to resources limitations. Quantification was based on external standard calibration. While matrix-spiked recovery experiments demonstrated acceptable accuracy, the absence of internal standards may have limited full compensation for matrix effects in sediment and macroinvertebrate samples

The Limit of detections (LODs) ranged from 0.010 to 0.015 ng/L for water and 0.010 to 0.015 ng/g for sediment and macroinvertebrates. The Limit of Quantification (LOQs) ranged from 0.34 to 2.30 ng/L for water and 0.34 to 2.30 ng/g for sediment and macroinvertebrates. These results confirm the method’s reliability for pesticide detection and quantification within the study’s parameters. The reproducibility or analytical uncertainty was determined by calculating the percent Relative Standard Deviation (%RSD) of triplicate injections, with all values maintained at RSD < 20%.

Screening level Ecological Risk Assessment (ERA)

Hence the studied matrix is surface water, we have undertaken the ecological risk assessment by applying a standard approach for an ecological screening-level risk assessment to compare the Measured Environmental Concentration (MEC) to the Predicted No-Effect Concentration (PNEC). The calculation was done using the following formula:

MEC: Measured Environmental Concentration or mean concentration detected in water samples (µg/L or mg/L), PNEC: Predicted No Effect Concentration (The “safe level” for aquatic life).

The PNEC values were obtained from the EPA Ecotoxicology database and Canadian water quality guidelines (CCME). Table 1 indicates the PNEC for highly detected pesticides in Water. To undertake screening level ecological risk assessment, we use the macroinvertebrates data for Risk Quotient (RQ) analysis because they act as biological indicators in the aquatic system, which reflect the long-term bioavailability and bioconcentration of organochlorine pesticides within the food web. In addition, their sedentary nature and close association with both the water column and benthos make them superior indicators for assessing the actual toxicological pressure on the aquatic community compared to non-living matrices (water and sediment).

Table 1.

The Predicted No effect Concentrations (PNECs) for the detected OCPs in water.

Pesticide	PNEC (µg/L)	References
p, p′-DDT	0.001	Canadian Water Quality Guidelines for the Protection of Aquatic Life20
p, p′-DDE	0.001	Aquatic Life Benchmarks and Ecological Risk Assessments (Standardized for DDT analogs)21
Aldrin	0.01	Aldrin and Dieldrin in Drinking-water; Environmental Health Criteria 9122
Dieldrin	0.01	Australian and New Zealand Guidelines for Fresh and Marine Water Quality23.
Heptachlor Epoxide	0.004	National Recommended Water Quality Criteria - Aquatic Life. And EU, 201321
Methoxychlor	0.03

PNEC Predicted No Effect Concentration.

Since our macroinvertebrate data is in ng/g (biota tissue), comparing it directly to a water-based PNEC (µg/L), would lead to significant overestimation because organisms bioconcentrate these chemicals from the water into their tissues. To correct this, we have derived Biota-based PNECs (PNECbiota​) in ng/g using the following standard approaches: for the correction of units, we have applied Official Environmental Quality Standards for Biota (EQSbiota​) such as the European Union Water framework directives (2013/39/EU) standards, which are specifically defined for concentration in fish/invertebrates to protect ecosystem. Therefore, we have converted the water based PNEC (µg/L) to a tissue-based value (ng/g) using the Bio-concentration Factor (BCF) using the formula below:

The conversion of PNECwater to PNEC biota using BCF is important for the interpretation of results. This conversion is by assuming equilibrium partitioning and the bioconcentration Factor (BCF) values previously derived from laboratory studies were applied in this risk assessment. While BCFs can vary with species, lipid content, and environmental conditions, this approach is widely applied in screening-level assessments when tissue concentrations are available as indicated in the USEPA guidance.

Statistical analysis and data interpretations

Data analysis was performed using IBM SPSS Statistics version 20.0. Before performing comparative tests, the Shapiro-Wilk test and Levene’s test were applied to evaluate the normality of distribution and homogeneity of variance, respectively. Since the pesticide residue data exhibited significant deviations from normality (p < 0.05), non-parametric statistics were employed for further inference. To address the special comparisons, the Kruskal-Wallis test was utilized to determine if statistically significant differences in OCP concentrations existed across the different sampling reaches. Where significant differences were found (p < 0.05), post-hoc analysis was applied to identify specific site pairs that differed while controlling Type I errors. For concentrations below the Method Detection Limit (MDL), a value of LOD/2 was used to facilitate statistical calculation, a standard practice in environmental trace analysis.

Quality appraisal

The methodological quality of the included studies was assessed using the Joanna Briggs Institute (JBI) critical appraisal checklist for cross-sectional studies. Each study was evaluated independently by two of the manuscript authors, with disagreements resolved through discussion. Overall, the included studies demonstrated good methodological strengths in clearly defined outcome measures and appropriate use of statistical analyses. However, common weaknesses were identified across the studies, including inadequate justification of sample size and poor reporting of response rates or handling of missing data. These limitations were considered when interpreting the findings of the review.

Results and discussions

Physicochemical parameters of water

Figure 2 shows the range of water quality parameters across all sampling sites. Water temperature (24–26 °C) remained within the optimal range for most tropical aquatic organisms. As temperature significantly influences metabolic functions and the physicochemical attributes of water, these favorable conditions likely facilitate the microbial and chemical degradation of pesticides in the system²¹. The mean dissolved oxygen (DO) was 5.20 ± 1.69 mg/L, which falls below 6 mg/L surface water quality standard required for safeguarding aquatic ecosystems⁸ and suggesting moderate organic loading and microbial oxygen demand likely from runoff-borne pollutants.

Fig. 2.

Physicochemical parameters of surface water from the study sites in southwestern Ethiopia.

Electrical conductivity (EC) ranged from 50.1 to 131.4 µs/cm (mean: 82.39 µs/cm), reflecting the presence of dissolved salts and organic materials. The pH levels ranged from 5.82 to 7.59 (mean: 7.03), falling within the WHO recommended range of 6.5–8.5. Occasional acidic values may be linked to the release of organic acids from decaying vegetation²².

Turbidity levels were significantly elevated, ranging from 26.4 to 104.33 NTU (mean: 54.14 NTU), which vastly exceeds the WHO limit of 5 NTU. Similarly, the mean Total Dissolved Solids (TDS) was 68.55 ± 16.39 mg/L. These high levels of suspended and dissolved solids are critical as they provide a greater surface area for the adsorption of hydrophobic pesticides, such as OCPs. Consequently, high turbidity likely facilitates the transport and persistence of pesticides within the water column, posing a prolonged risk to aquatic biota and consumers.

Physicochemical parameters of sediment

Table 2 outlines the sediment physicochemical characteristics across the sampling sites (n = 14). The pH of the sediment samples remained relatively stable (mean 6.70), while Electrical Conductivity ranged from 53.15 to 145.32 µS/cm (mean 85.65 ± 29.54 µS/cm). Although higher pH levels may favor the alkaline hydrolysis of certain pesticides, the observed mildly acidic to neutral sediments in this study likely contributed to the increased persistence of OCPs.

Table 2.

Minimum, maximum, mean, and standard deviation of pH, EC in (µS/cm), % moisture and TOC (%) in sediment (n = 14).

Parameters	Minimum	Maximum	Mean	SD
H	5.46	7.80	6.70	0.64
EC	53.15	145.32	85.65	29.54
Moisture	5.37	7.73	6.16	0.68
TOC	0.62	2.37	1.19	0.42

EC Electric Conductivity, TOC Total Organic Carbon.

Total Organic Carbon (TOC) levels ranged from 0.62% to 2.37%, with an average of 1.19 ± 0.42%. These levels are likely influenced by the influx of municipal waste and decaying detritus into the river system. The presence of TOC is a critical determinant for the environmental fate of hydrophobic contaminants. Specifically, the affinity of OCPs for sediment is directly related to the organic carbon content, as described by the organic carbon-normalized partition coefficient (KOC​). Elevated TOC values enhance the adsorption capacity of the sediment, effectively sequestering OCPs from the water column and increasing their accumulation in the benthic substrate²³. This sorption behavior suggests that the sediment acts as a long-term sink for pesticide residues in this aquatic ecosystem.

Concentration of pesticides in water, sediment and macroinvertebrate samples

Figure 3 illustrates the detection of nine organochlorine pesticides in water, sediment, and macroinvertebrate from various sites in southwest Ethiopia. Except endosulfan sulfate, all the studied pesticides such as Aldrin, γ-chlordane, p, p’-DDE, Dieldrin, p, p’-DDT, Methoxychlor, and Heptachlor epoxide, were detected in water. The significant accumulation of pesticides within the water column primarily associated with the particulate phase, is likely driven by high water turbidity. Although increased turbidity and the presence of suspended solids reduce the dissolved-phase concentration through adsorption, they facilitate a higher total OCP load being transported via suspended organic and inorganic matter. This disparity is primarily attributed to the high octanol-water partition coefficients (logKow​) and inherently low aqueous solubility of organochlorine pesticides. Furthermore, high water turbidity facilitates the partitioning of these hydrophobic compounds onto suspended particulate matter. While this adsorption mechanism effectively reduces the concentration of the dissolved-phase OCPs, it increases the total suspended load within the water column. As these pesticide-laden particles settle, they contribute to the significantly higher concentrations observed in the sediment and the subsequent bioaccumulation in benthic biota. These findings are strongly supported by the physicochemical results, which show, sites with elevated turbidity and total dissolved solids (TDS) serve as focal points for OCP sequestration. Spatial analysis revealed that the Boye-01 wetland site served as a significant hotspot for pesticide accumulation, with Aldrin reaching a peak mean concentration of 0.37 ± 0.36 µg/L. This localized elevation is likely due to the wetland’s role as a catchment sink for surrounding agricultural runoff. Similarly, the DDT metabolite, p,p′-DDE, exhibited concentrations ranging from < LOD up to 3.2 µg/L. These levels significantly surpass values reported in previous regional studies²⁴ and²⁵. This indicates a substantial environmental burden of organochlorine residues in the southwest Ethiopia Dieldrin in water samples was found at a mean concentration of 0.23 ± 0.22 µg/L, surpassing the World Health Organization’s (WHO) Maximum Residue Limit (MRL) guideline of 0.03 µg/L²⁶.The mean concentration of p, p’ DDT- in water samples was 4.92 ± 5.41 µg/L which is greater than the concentration observed from a study conducted by²⁴. Overall, a critical exceedance the WHO MRL, (0.03 µg/L) was observed at the Boye-01 site, where p, p′-DDT levels peaked at 11.14 µg/L. Even though, the observed concentration could also results from historical residues that have persisted due to slow degradation rates under local conditions or differences in metabolic breakdown among organisms, The elevated concentrations of DDT-related compounds in macroinvertebrates, particularly the presence of the parent compound p, p′-DDT in some samples, may indicate recent environmental inputs of DDT into the the study areat.

Fig. 3.

Mean concentrations of organochlorine pesticide (OCP) residues in water (ng/L), sediment (ng/g dry weight), and macroinvertebrates (ng/g wet weight) from the Jimma zone, southwestern Ethiopia. Error bars represent standard deviation (SD). Concentrations in water are expressed in ng/L (equivalent to ng/kg assuming 1 L water ≈ 1 kg).

Such findings align with observations in the Awash River basin, Ethiopia²⁷ and likely point to ongoing usage for indoor residual spraying (IRS) in malaria-endemic regions or agricultural applications. The mean concentration of Methoxychlor was 0.43 ± 0.32 µg/L and which is greater than from a study investigated by²⁸ and²⁵. Heptachlor epoxide was solely detected at the Boye-01 site, ranging from < LOD to 1.44 µg/L. The value is above the MRL of 0.03 µg/L for water²⁹. The results may indicate to the illegal use of this chemicals in the study area. Among the nine organochlorine pesticides analyzed across the 14 sites, the highest occurrence was observed for Aldrin and Methoxychlor, both with a detection frequency of 50%, followed by Dieldrin with a detection frequency of 35.71%.These moderate detection frequencies suggest that these three compounds are still present in a substantial proportion of the studied streams. However, the fact that none of the pesticides were detected at all sites indicates spatial variability in contamination levels, which may reflect differences in historical usage, local sources, or environmental fate processes.

With regards to the detection of OCPs in macroinvertebrates samples, nine organochlorine pesticides were detected in macroinvertebrates, with varying concentrations. The concentration of Aldrin ranged from below the limit of detection (< LOD) to 4.43 ng/g, with a mean concentration of 2.21 ± 1.74 ng/g. The detection of γ-chlordane ranged from below the LOD to a maximum of 8.22 ng/g, with a mean concentration of 1.69 ng/g. The mean concentration of Endrin in the macroinvertebrates was determined to be 0.34 ng/g. This value is comparatively lower than the concentrations of other organochlorine pesticides detected in this study. The significantly lower levels of Endrin (mean 0.34 ng/g) can be explained by its lower logKow​ (approx. 5.2) relative to DDT. This lower lipophilicity results in a reduced affinity for adipose tissue. Furthermore, unlike the highly stable DDT, Endrin is more susceptible to metabolic biotransformation and rapid elimination by aquatic invertebrates. While DDT residues persist for years, Endrin is more readily degraded and excreted, leading to the lower body burdens observed in this study. The high incidence of p, p′-DDT (85.71%) is particularly significant; the presence of the parent compound alongside its metabolite suggests that these organisms may exposed from recent inputs, likely from malaria control or agricultural runoff. From a trophic perspective or from historical applications, macroinvertebrates serve as the primary link for transferring these residues to higher-trophic-level consumers, such as fish and insectivorous birds. Comparing these findings to regional standards, the observed levels in Boye-01 exceeds FAO/WHO Maximum Residue Limits (MRLs) for aquatic biota, which can lead to biomagnification and long-term reproductive impairment in the local food web.This result is supported by a study done in Rift Valley region, Ethiopia, p,p’-DDE was the most abundant compound and had significantly a higher burden in birds tissues³⁰. Following these, Aldrin and Dieldrin were detected with a frequency of 57.14%. In contrast, Endosulfan sulfate exhibited a lower frequency of detection, being found in only 21.43% of the samples.

The mean concentration of p, p’-DDT was 45.67 and ranged from < LOD to 116.34 ng/g with incidence of 85.71% this may indicate there may be a recent use for agricultural and malaria control and slow degradability in the environment. From a study done by³¹, the area is a hub to produce coffee, corn, and various vegetables, where OCPs like DDT and Aldrin were historically used for soil treatment and until recent years as pest control due to their low cost and effectiveness¹. Additionally, Southwestern Ethiopia has a high malaria cases were the detection of high levels of p, p’-DDT, alongside its metabolite p, p’-DDE, suggests its ongoing or historical use for indoor residual spraying (IRS), despite official restrictions, a practice noted in other parts of Ethiopia³². The concentration of Aldrin was recorded with range from lowest < LOD to highest 4.63 ng/g with mean concentration of 2.21.

With regards to the ecological relevance of macro-invertebrate identification our study has done macro-invertebrate taxa at family level identification indicates high richness in less OCPs concentration areas such as Kito 1 to 3 (Fig. 4). The collected macroinvertebrates predominantly belonged to families such as Baetidae, Hydropsychidae, Chironomidae, and Libellulidae, which are well-established bioindicators of water quality and pollutant stress in freshwater ecosystems. The richness of macroinvertebrates (i.e., the number of taxa at each site), a key ecological indicator of ecosystem health, was relatively higher at the Kofe stream sites compared to other streams with lower pesticide concentrations in both water and sediment. This may be attributed to minimal agricultural and human activities in the area, likely because the stream is located near Jimma Airport, where access is restricted for security reasons. High richness can be an indicator for low pesticides residues.

Fig. 4.

Composition and relative abundance of macroinvertebrate taxa collected across all sampling sites in the studied streams and wetland, southwestern Ethiopia.

The findings from the studied sediment samples, all pesticides were detected in the two streams and the Boye wetland at varying concentrations. Among the pesticides examined, p,p’-DDE, was detected at a mean concentration of 8.76 ng/g and also identified in the natural environment through chemical and biological processes³³. Aldrin was found in sediment samples ranging from 0.14 to 19.86 ng/g, with a mean value of 4.35 ± 4.86 ng/g, surpassing concentrations reported by Chang and Doong (2006). Studies in Lake Hawassa, Ethiopia, and the Akaki River also identified Aldrin and p, p’-DDE. Gamma-chlordane levels averaged 5.27 ± 5.42 ng/g, while Endrin and Endosulfan sulfate averaged 1.64 ng/g and 1.71 ng/g, respectively. In contrast, p,p’-DDT levels ranged from 23.83 to 409.64 ng/g, notably higher than studies done by Vu Duc’s (2015) and Javedankherad et al.‘s (2013). This may be due to low degradation process and may be due to a recent input in the study area. Boye wetland site exhibited significantly higher pesticide residues, likely due to receiving pollutants from surrounding areas and also probably due to the convergence of streams and drainage channels at the Boye Wetland. Overall, the order of percent detection for the studied organochlorine pesticides in sediment samples was as follows: Aldrin and p, p’-DDE (100%), p, p’-DDT (78.57%), Heptachlor (78.57%), Methoxychlor (71.43%), and Heptachlor epoxide (71.43%). The detection of organochlorine pesticides in water and macroinvertebrate tissues may primarily reflect historical applications, although recent inputs for agricultural activities or public health purposes cannot be ruled out. The co-occurrence of both parent compounds and their metabolites complicates source attribution, as this pattern may result from either persistent legacy residues or slower degradation rates under local environmental conditions. Further investigations, such as sediment core analysis or current usage surveys, would be necessary to better elucidate the relative contributions of historical versus recent sources.

Spatial analysis (Fig. 5) shows that the Boye-01 wetland generally exhibited higher OCP concentrations compared to other sites. As a convergence point for the Awtu and Kofe streams, this wetland is characterized by lower flow velocities and deeper sediment beds, which may promote the deposition and accumulation of particle-bound pesticides. In contrast, the macroinvertebrate samples from the upstream reaches of the Awtu stream exhibited notable peaks of Heptachlor-epoxide. This localized elevation may be due to direct runoff from adjacent agricultural fields where heptachlor may have been applied as a soil insecticide. These spatial trends, where wetlands act as regional reservoirs while upstream reaches reflect immediate land-use impacts, align with observations in other Ethiopian aquatic systems, such as the Rift Valley and Lake Hawassa³⁴.

Fig. 5.

Concentrations of organochlorine pesticides (ng/g) across different sampling sites in streams and wetlands of the Jimma zone, southwestern Ethiopia.

According to³⁵, the elevated concentrations of organochlorine pesticides in sediment samples stem from their hydrophobic nature, causing them to settle and accumulate in sediments. Sediments serve as the primary reservoir for these pesticides in aquatic environments due to their low solubility and affinity for suspended particulate matter³⁶. This accumulation explains the typically higher levels of organochlorine pesticide residues in sediment compared to water and macroinvertebrates³⁷.

The Kruskal-Wallis test results, depicted in box and whisker plot, demonstrate a significant distinction (p-value < 0.001) between the water and the sediment and macroinvertebrate matrices. This may be due to the physicochemical properties of the studied analytes, OCPs are highly lipid soluble and intends to adsorb in sediment and consequently accumulate Biota.

Findings from the screening level ecological risk assessment using macroinvertebrate data

The Screening level Ecological Risk Assessment was undertaken by considering the macroinvertebrate as representative matrix, and helps to see the direct effect on aquatic biota. From our findings in Table 3, the ecological risk characterization reveals a high-to-extreme toxicological pressure on the macroinvertebrate communities within the Awetu, Kito, and Boye aquatic systems. The most alarming finding in this study is the extreme risk posed by Heptachlor epoxide (RQ = 738.8) and Dieldrin (RQ = 1.53). This may be due to the high concentration detected in the matrix. From our findings, the mean concentration of Heptachlor epoxide (4.95 ng/g) exceeded the European Union Environmental Quality Standard (EQS) for Biota (0.0067 ng/g) by several orders of magnitude. With regards to those OCPs of RQ > 1, are significant because Heptachlor epoxide is a potent neurotoxin with a high bioconcentration factor (BCF). According to Directive 2013/39/EU (2013) indicates the toxic effect may occur along the trophic level (fish and birds). Even though, the conversion of water-based PNEC to PNECbiota using BCF relies on assumptions of steady-state conditions and the transferability of previously laboratory-derived BCFs to field macroinvertebrates. Uncertainties may arise from inter-species differences, variable lipid content, and site-specific environmental factors. However, the consistently observed high RQ values underscore a substantial ecological risks posed by the detected OCPs in the study area. These result highlights the value of biota monitoring and call for follow-up higher-tier assessments to confirm and quantify impacts on aquatic communities. The high risk quotients calculated for organochlorine pesticides suggests potential ecological concern. Macroinvertebrates in the studied Ethiopian streams accumulated measurable concentrations of these persistent compounds. Because of their central position in the food web, these organisms may play a role in the transfer of bioaccumulated pesticide residues to higher trophic levels, including fish and other predators.

Table 3.

Results of the screening level ecological risk assessment for the commonly detected pesticides in macroinvertebrates.

Pesticide	Mean MEC (ng/g)	PNEC biota​ (ng/g)	Risk Quotient (RQ)	Risk Level
Heptachlor Epoxide	4.95	0.0067	738.8	Extreme
p, p′-DDT	45.67	50.0	0.91	Medium
Dieldrin	7.67	5.0	1.53	High
Aldrin	2.21	40.0	0.05	Low
Methoxychlor	9.98	240.0	0.04	Low

MEC Measured Environmental concentration, PNEC Predicted No Effect Concentration, RQ Risk Quotient.

While p, p′-DDT exhibited the highest absolute mean concentration (45.67 ng/g), its calculated RQ (0.91) remains below unity (RQ < 1). This indicates that while the current body burden is approaching the limit of chronic toxicity for the invertebrates themselves, it poses a “medium” risk rather than an acute one. However, the presence of such high residues in primary consumers indicates recent exposure or high historical loading in the sediment. According to US EPA (2022), the cumulative effect of multiple OCPs (DDT, Dieldrin, and Aldrin) can lead to additive toxicity, where the combined chemical stress reduces the reproductive success and diversity of sensitive macroinvertebrate taxa (such as Ephemeroptera and Trichoptera), even if individual RQs are near 1. The use of macroinvertebrates in this risk assessment was crucial for capturing the bioavailable fraction of these pollutants. Unlike water samples, which fluctuate with seasonal runoff, the tissue concentrations in these organisms provide a time-integrated measure of pollution. The disparity between low water solubility and high tissue accumulation confirms the high lipophilicity of these compounds. This observation aligns with the findings of Imo and colleagues³⁸, who argued that in tropical aquatic systems, sediment-bound pesticides are readily remobilized and sequestered by benthic organisms, making them the most vulnerable link in the ecosystem.

Comparison of the detected pesticides in sediments with standards

Organochlorine pesticide concentrations in sediment samples were assessed against Canadian sediment quality guidelines (Table 4) designed to safeguard aquatic life and evaluate sediment quality in streams. Dieldrin levels ranged from < LOD to 15.39 ng/g, surpassing both the Interim Sediment Quality Guidelines (ISQGs) of 2.58 ng/g and the probable effect limit (PEL) of 6.67 ng/g. For p, p’-DDT, concentrations ranged from 23.845 ng/g to 409.644 ng/g, notably exceeding the ISQGs value of 1.19 ng/g and the PELs value of 4.77 ng/g. Heptachlor epoxide levels ranged from < LOD to 13.988 ng/g, surpassing the ISQGs value of 0.6 ng/g and the PEL value of 2.74 ng/g. The Canadian sediment quality guideline establishes the PEL as the threshold for frequent adverse effects on aquatic biota³⁹. Concentrations of Dieldrin, p,p’-DDT, and Heptachlor epoxide in most sediment samples exceeded the interim sediment quality guidelines (ISQGs) and probable effect limits (PEL) set by the Canadian Sediment Quality Guidelines (CSQGs). This poses a concern as benthic macroinvertebrates, residing in the sediment-water interface with restricted mobility, face heightened exposure to persistent organic pollutants through sediment and water pathways. Their proximity to both bottom sediments and the water column amplifies their susceptibility to these stressors, including the aforementioned organic pollutants⁴⁰. Accumulated pollutants in sediment can harm the health and survival of diverse organisms, such as benthic macroinvertebrates and other aquatic species.

Table 4.

Comparison of Organochlorine Pesticide Concentrations in Sediment with Canadian Sediment Quality Guidelines.

Analytes	This study		Canadian sediment quality guideline
	Range	Mean	ISQG	PEL
Aldrin	0.6–19.86.6.86	4.68	n.a.	n.a.
g-chlordane	< LOD-15.76	5.84	n.a.	n.a.
4,4’-DDE	0.18–69.25	8.76	n.a.	n.a.
Endrin	< LOD-14.91	4.56	45	62.4
Endosulfan sulfate	< LOD- 10.72	3.00	n.a.	n.a.
Dieldrin	< LOD-15.39	2.41	2.58	6.67
p, p’-DDT	23.85–409.64.85.64	104.11	1.19	4.77
Methoxychlor	< LOD- 78.07	22.64	n.a.	n.a.
Heptachlor epoxide	< LOD-13.99	6.18	0.6	2.74

Source⁴¹, ISQGs Interim sediment quality guidelines, PEL probable effect limits, n.a not available.

The exceedance of maximum residue limits indicates illegal use of pesticides and a lack of adherence to Good Agricultural Practices (GAP) by farmers in and around the study sites.

Pearson correlation coefficients between OCPs concentrations and physicochemical parameters in water, sediment, and macroinvertebrate samples Pearson’s correlation analysis was used to explore the connection between physicochemical parameters and pesticide concentrations in the samples and presented as supplementary table S2 in annex. Turbidity in water exhibited strong positive correlations with γ-chlordane (r = 0.618), Methoxychlor (r = 0.564), and Heptachlor epoxide (r = 0.568), and weaker positive correlation with Aldrin (r = 0.529). They may suggest that in turbid waters, suspended particles provide additional surface area for pesticide adsorption. This can increase the exposure to filter-feeding organisms and facilitate the transport of pesticides, ultimately leading to deposition and accumulation in sediments. The correlation with Total Organic Carbon (TOC) indicates that organic-rich sediments act as a major sink for these hydrophobic pesticides, prolonging their persistence in the ecosystem and increasing the exposure risk for benthic-dwelling organisms.

These findings align with prior studies emphasizing turbidity’s impact on the presence and endurance of pesticide residues in water. A study by⁴², a positive correlation was discovered between turbidity and organochlorine pesticide concentrations in a river system. This relationship may be attributed to turbidity increasing the available surface area for pesticide adsorption, resulting in higher concentrations in turbid waters.

Our findings also indicate a strong correlation between Total Suspended Solids (TSS) in water and γ-chlordane (r = 0.513), p, p’-DDE (r = 0.549), Methoxychlor (r = 0.558), and Heptachlor epoxide (r = 0.573) (P < 0.05). These robust positive correlations suggest that turbidity and TSS in water could enhance the adsorption of organochlorine pesticides, aligning with prior research highlighting the impact of TSS and turbidity on pesticide adsorption in aquatic settings. Similarly, a study done by⁴³, revealed a strong relationship was noted between turbidity and pesticide adsorption in a lake ecosystem. Positive correlations between Total Suspended Solids (TSS) and pesticide concentrations imply that suspended solids offer surfaces for pesticide adsorption in waters with higher TSS levels. Our research also revealed a significant positive correlation between pH and p, p’-DDT pesticide concentration (p < 0.01). Moreover,

Furthermore, a strong positive correlation was observed between electrical conductivity and p, p’-DDE and Endrin pesticides (p < 0.05). These results are consistent with prior studies emphasizing the impact of pH and electrical conductivity on pesticide behavior in aquatic settings. For instance, a recent study by⁴⁴, A recent study by Tian et al. (2009) highlighted a positive correlation between pH and DDT concentrations in water, indicating that pH and electrical conductivity can impact the solubility, transport, and persistence of pesticides. The pH changes can influence pesticide ionization, affecting their availability and adsorption potential, while electrical conductivity can influence pesticide mobility and distribution in water bodies. Our research unveiled a strong positive correlation (p < 0.01) between total organic carbons and the concentrations of Aldrin, DDE, Methoxychlor, and Endrin. Additionally, a positive correlation (p < 0.05) was found between total organic carbons and Heptachlor epoxide, suggesting that organic pollutants like organochlorine pesticide residues tend to adsorb onto organic carbon.

Various studies have emphasized the significant role of organic carbon in influencing the behavior and distribution of organic pollutants. From a recent study by⁴⁵, a substantial positive correlation was identified between organic carbon levels and organochlorine pesticide concentrations in sediments and a significant positive correlation between organic carbon and the levels of organochlorine pesticides in sediments. Overall, the physicochemical parameters of the environmental matrices affect the accumulation of organochlorine pesticides.

Conclusions

This study reveals the presence of nine organochlorine pesticides (OCPs) in water, sediment, and macroinvertebrate samples from the Boye, Awetu, and Kito aquatic systems in southwest Ethiopia. Aldrin and p, p′-DDT were among the frequently detected compounds, and the occurrence of parent compounds alongside their metabolites may suggest possible recent inputs in addition to historical residues, although further studies are needed to clarify the relative contributions of current versus legacy sources. The screening-level ecological risk assessment indicates that macroinvertebrates in these aquatic systems accumulate notable concentrations of OCPs. Given their central role in the aquatic food web, these organisms may serve as a vector for the transfer of bioaccumulated contaminants to higher trophic levels, including fish and other predators. Risk quotients (RQ) for some pesticides substantially exceeded 1, with the highest value reaching 738.8 for heptachlor epoxide. While these elevated RQs point to potential ecological concern, they should be interpreted in a way considering the assumptions and uncertainties associated with the BCF-based conversion from water PNEC to biota concentrations in the screening level ecological risk assessment.

Overall, the results highlight the persistence of legacy organochlorine pesticides in the studied aquatic environment and the need for continued monitoring. To support the long-term sustainability of water resources in southwest Ethiopia, it is recommended that existing regulations be strengthened through the establishment of routine monitoring programs, particularly in ecological hotspots such as the Boye wetland. Additional measures, including community education on the environmental behavior of persistent pesticides and the promotion of safer alternative pest management practices, could help reduce future inputs and mitigate potential risks.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Seblework Mekonen and Tamirat Beyene, conceived, designed methodology, data collection, laboratory analysis and writing the manuscript, Professors Argaw Ambelu and Abera Gure designing the method, result analysis and drafting and editing the manuscript, Dr. Higemengist Astatkie data collection, laboratory analysis and writing the manuscript. All authors have approved the submitted version of the manuscript.

Funding

This work has no funding source.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.