Data Mining to Evaluate the Effect of Eichhornia crassipes and Lemna minor in the Phytoremediation of Wastewater in the Canton of Milagro

Abstract

The constant increase in industrialization and urbanization has led to the regular discharge of wastewater into the environment in excessive amounts, which has caused significant impacts on both human and wildlife ecosystems. The sustainable management and treatment of wastewater, whether of industrial or domestic origin, represents a crucial challenge in this century. In this study, phytoremediation was employed as a wastewater treatment strategy using two species of aquatic macrophytes: water hyacinth (Eichhornia crassipes) and duckweed (Lemna minor). The study was conducted over seven consecutive evaluation periods, with five-day intervals between each. The objective was to apply the multivariate HJ-Biplot methodology to evaluate the effects of phytoremediation of two species of aquatic microphytes on the physicochemical characteristics of wastewater from Milagro canton, Ecuador. Additionally, a microbiological analysis was conducted to determine the effectiveness of the floating macrophytes. The analysis was based on the measurement of various physicochemical parameters, such as pH, electrical conductivity (EC), dissolved oxygen (DO), oxidation–reduction potential (ORP), salinity, total dissolved solids (TDSs), biochemical oxygen demand (BOD), chemical oxygen demand (COD), hardness, and temperature. The results showed that the highest efficiency in pollutant removal was achieved with duckweed (Lemna minor) in five out of nine measured parameters, suggesting that this species was the most effective compared to the control sample and Eichhornia crassipes. The capacity of these macrophytes for wastewater treatment was confirmed by this study. To ensure effective water purification, timely extraction of aquatic macrophytes from water bodies is recommended. If this collection is not properly carried out, the nutrients absorbed and stored in the plant tissues may be released back into the aquatic environment due to plant decomposition.

1. Introduction

Water, as a vital element, is indispensable for all forms of life on the planet. Universal access to water under equal conditions and at an affordable cost is a human right and comprises the sixth sustainable development goal to be achieved by 2030 [1]. However, despite its bioavailability in ecosystems through various physical states, unequal distribution and growing global demand limit access to drinking water. Various international entities point out that the scarcity of potable water resources is not only due to the level of availability of the element but also due to factors such as poor governance, weak policies, and poor management [2]. The discharge of wastewater from various anthropogenic activities into water bodies without prior sanitation is one of the main causes that worsens the water crisis [3].

In Ecuador, poor wastewater management constitutes a serious environmental and public health issue. According to records from the National Institute of Census and Statistics, approximately 26.3% of the water distributed by Decentralized Autonomous Governments (GADMs) enters sanitation plants. The remaining water is discharged into water bodies, with rivers being the predominant final destination, receiving 53.3% of untreated wastewater [4]. One of the water bodies impacted by this problem is the Estero de Las Damas in the canton of Milagro, Ecuador. This estuary receives discharges of domestic origin, including untreated wastewater, urban solid waste, and decomposing organic matter. In addition, industrial and agro-industrial activities in the area contribute chemical pollutants, such as heavy metals, fertilizers, and persistent organic compounds.

Conventionally, various sanitation measures for wastewater have been applied, such as sedimentation, filtration, and chlorine disinfection; however, these purification systems require complex and specialized infrastructure which, after the purification process, generate polluting by-products that require additional processing before final disposal [5]. In addition, conventional methods such as biological treatments such as activated sludge are widely used, which employ microorganisms to decompose the organic matter dissolved in the wastewater, converting it into less harmful substances [6]. Likewise, anaero-biosystems are used, which have proven to be effective in reducing the organic load and generating usable by-products such as biogas, although their operation involves specialized management and considerable costs [7]. Faced with this problem, the use of aquatic plants or macrophytes with phytoremediation capacity emerges as an ecological solution.

In its broadest definition, phytoremediation refers to a strategy that uses the biochemical reactions of aquatic plants to reduce the in situ concentration of organic and inorganic pollutants in soils, sediments, water, and air [8]. Associated phytotechnologies are based on the basic physiological mechanisms that naturally occur in plant structures and symbiotic microorganisms, such as transpiration, photosynthesis, and nutrition [9]. The process is conditioned by numerous factors, the most important being the species type, contamination level, pollutant nature, and the quality of water under treatment [10]. In particular, the nature of the pollutant is a determining factor in bioremediation treatments, as each type of pollutant, organic or inorganic, has specific physical and chemical characteristics that influence its bioavailability and interaction with the plants, microorganisms, or enzymes involved [11].

The selection of floating macrophytes for phytoremediation in this study was based on their ability to absorb pollutants and improve water quality in contaminated environments [12]. Eichhornia crassipes (water hyacinth) and Lemna minor (duckweed) are among the most studied species due to their rapid growth, high pollutant uptake, and adaptability to a wide range of water conditions [13,14]. Previous studies have shown their efficacy in removing organic pollutants, period metals, and excess nutrients from wastewater [15,16]. In addition, their biomass can be harvested and reused for bio-energy production, making them a cost-effective and sustainable alternative to conventional treatment methods [17,18].

Data mining is a key phase within the comprehensive knowledge discovery process, focused on extracting valuable information from databases [19]. This approach is based on the identification of hidden patterns, models, trends, and sequences in data to represent large volumes of information in a reduced space, becoming an effective alternative to classical statistical classification and ordering methods as it allows for the graphical representation of large volumes of information in a reduced space [20]. Various algorithms and methodologies, such as the HJ-Biplot, are oriented toward relationships between multiple attributes. The application of these strategies has the potential to significantly improve wastewater sanitation [21].

The purpose of this paper was to evaluate the effects of phytoremediation using two species of aquatic macrophytes on the physicochemical characteristics of wastewater from the Milagro canton, using the multivariate HJ-Biplot method, a statistical tool designed for the analysis of high-dimensional environmental data. This method has the ability to visually represent the relationships between multiple water quality parameters and treatment conditions in a reduced dimension space, which facilitates the interpretation of complex interactions [22]. Compared to traditional statistical approaches, the HJ-Biplot provides greater interpretability and a more complete overview of variations in water quality, making it a suitable option for assessing the phytoremediation potential of E. crassipes and L. minor in wastewater-impacted environments.

2. Materials and Methods

2.1. Study Area

This experimental study was conducted in the Milagro River, specifically in the “Estero Las Damas” (2°07′54″ S–79°35′22″ W), located in the city of Milagro, province of Guayas, Ecuador.

2.2. Selected Species

The species selected for remediation were water hyacinth (E. crassipes) and duckweed (L. minor). Both floating species have phytoremediative properties that allow them to absorb nutrients and pollutants from the water [23].

Water hyacinth (E. crassipes) is considered one of the most aggressive and prevalent invasive plant species, being a floating plant with rapid growth and a high capacity to absorb nutrients, heavy metals, and toxic compounds [24].

Duckweed (L. minor) is used in water quality research to control and analyze its ability to absorb heavy metals and certain types of pollutants [25].

2.3. Sample Collection

Wastewater samples were taken from the “Las Damas” estuary, near the Milagro municipal cemetery, following a systematic protocol to ensure representativeness. Sampling was carried out during a 31-day period between August and September, months characterized by low precipitation in the region. Samples were taken from stagnant water bodies within the estuary, which receive discharges from domestic and industrial sources without prior treatment. Due to the lack of water renewal in these sectors, pollutants accumulate, generating conditions typical of wastewater. The samples were stored in high-density polyethylene containers, previously sterilized, ensuring their preservation without alterations until laboratory analysis.

The macrophytes Eichhornia crassipes and Lemna minor were collected in the “Las Damas” estuary, in sectors with an accumulation of organic matter and domestic discharges, ensuring that the selected plants were adapted to the conditions of the stagnant water under study. To avoid interference in the experiment, the species were subjected to a decontamination process, which included washing with running water, rinsing with distilled water, and root treatment with acetone, eliminating residues and external microorganisms. Since both species are cata-logged as invasive in various aquatic ecosystems [26,27], preventive measures were taken to avoid their uncontrolled spread. During the study, macrophytes were confined within closed experimental units, without contact with natural water bodies. In addition, a periodic biomass removal protocol was implemented to control their growth and avoid obstructions in the water flow [13].

2.4. Experimental Design

The experimental setup consisted of selecting round plastic tubs with a capacity of 5 L. To start the experiment, the tubs were adequately washed and dried with a paper towel, and 5 L of wastewater was added to each tub. Tub 1 was set as a control and had no plants. E. crassipes was grown in tub 2, and L. minor was grown in tub 3.

The experimental tanks were placed outdoors, in an open area, exposed to natural sunlight, with an approximate photoperiod of 12 h of light and 12 h of darkness. The tanks were partially protected from direct rain by a transparent polyethylene cover to avoid dilution and maintain exposure to natural environmental conditions. No artificial lighting was used throughout the experiment, ensuring that the plants were exposed to conditions similar to those of natural aquatic environments.

2.4.1. Physicochemical Parameters Evaluated

Samples were taken individually from each of the three tanks every five days for a period of 31 days. The physicochemical parameters were analyzed using precision laboratory equipment, following standardized methods described by the American Public Health Association [28] and in accordance with the regulations of the Ecuadorian Institute of Standardization [29] for water quality analysis.

-

pH and temperature

The measurement of pH and temperature (°C) was performed with a digital potentiometer (Hanna HI 2020), calibrated according to the standard electrometric method [28].

-

Electrical conductivity and salinity

Electrical conductivity was determined using a digital conductivity meter (Thermo Scientific Orion Star A212, Fisher Scientific S.L., Madrid, España), following the 2520B method [28]. Salinity was calculated indirectly from conductivity using standard conversion tables [30].

-

Total dissolved solids (TDSs)

TDSs were quantified by using the 2540C gravimetric method [20], using a drying oven (Memmert UF55) at 180 °C until constant weight was reached, followed by weighing on an analytical balance (Ohaus Explorer EX224).

-

Biochemical oxygen demand (BOD)

BOD was determined by aerobic incubation at 20 °C (±1 °C) for five days in a laboratory incubator (Memmert IN30), following method 5210B [28]. Dissolved oxygen was measured using a portable oximeter (YSI Pro20).

-

Chemical oxygen demand (COD)

COD was analyzed by digestion with potassium dichromate in acid medium using a COD digester (Hach DRB200) and quantification by UV–Visible spectrophotometry with a spectrophotometer (Hach DR6000), following method 5220D [28].

-

Dissolved oxygen (DO)

The determination of dissolved oxygen was performed with a benchtop oximeter (Thermo Scientific Orion Star A213), equipped with an oxygen permeable membrane polarographic probe, following method 4500-O G [28].

-

Total hardness

Total hardness was evaluated by complexometric titration with EDTA, using an automatic titrator (Metrohm 877 Titrino plus, Metrohm, Madrid, España) and erythiochrome black indicator T. This procedure followed method 2340C and was verified according to Sawyer et al. (2003) [31].

2.4.2. Microbiological Parameters Evaluated

The microbial load of treated water was evaluated by the quantification of total coliforms, fecal coliforms, Escherichia coli and fecal Staphylococcus. Coliform de-determination was performed by the Most Probable Number (MPN) method using multiple tubes. The presumptive test used Lauryl Sulfate Tryptose (LST) broth, incubated at 35 °C for 24–48 h, and positive samples were confirmed in Bile Brilliant Green Broth (BGB) under the same conditions. Identification of E. coli was performed by membrane filtration on EMB (Eosin Methylene Blue) Agar plates, incubated at 37 °C for 24 h, recognizing colonies by their characteristic greenish metallic coloration [28].

Quantification of fecal Staphylococcus was carried out by plate count on Baird–Parker Agar, incubated at 37 °C for 24–48 h. The presence of black colonies with a transparent halo, a product of lecithinase activity, indicated recent fecal contamination. This microorganism was used as a complementary indicator to assess sanitary risk and determine specific sources of bacterial contamination [32]. In this study, the initial values of the microorganisms analyzed were 7500 ± 30.5 MPN/mL for total coliforms, 1250.67 ± 15.2 MPN/mL for fecal coliforms, 1100.33 ± 10.5 CFU/mL for E. coli, and 600 ± 12.5 CFU/mL for fecal Staphylococcus.

2.5. Statistical Analysis

The evaluated physicochemical parameters were measured in triplicate, and the obtained data were subjected to data mining techniques, such as the HJ-Biplot, using R software version 4.4.1, a free software environment for statistical computing and graphics [33].

2.5.1. Biplot Graph

A biplot is a graphical representation of an N × M matrix, where N represents the number of individuals and M the number of variables [34]. This type of graph facilitates the analysis of data with multiple variables. Through multivariate analysis, data dimensionality is reduced by projecting them into a lower-dimensional space known as Euclidean space [35,36]. Biplots are a useful tool for visualizing the results of principal component analysis (PCA), showing statistical relationships such as distances between data points, groupings, variances, and correlations between variables or individuals [20].

In this study, the multivariate HJ-Biplot methodology was used because of its ability to simultaneously represent individuals and variables in a two-dimensional space, facilitating the interpretation of complex relationships in large datasets. This method projects variables and samples on a common factorial plane, making it possible to visualize associations between the physicochemical and microbiological variables evaluated and the plant species used in biological remediation. Its precision lies in the preservation of the original structure of the data, which enables a clear interpretation of patterns, correlations, and groupings [21,22].

In the interpretation of the HJ-Biplot, the arrows represent the variables (e.g., BOD, COD, salinity), and the length of each arrow indicates the contribution of that variable to the total variance explained by the components. The direction and angle between arrows reflect the correlation among variables: arrows pointing in similar directions suggest a positive correlation, while those pointing in opposite directions suggest a negative correlation. The data points (samples or treatments) are represented by labels (e.g., C1, E.c1, L.m1), and their proximity to a particular arrow implies a stronger association with that variable. Clusters of points indicate similarities among treatment groups, facilitating the identification of patterns in the dataset [37].

2.5.2. Descriptive Statistics

The results of the microbiological parameters were analyzed using an analysis of variance (ANOVA) to evaluate the significance of individual differences with a significance level of p < 0.05. The data were processed using R software version 4.4.1. Vertical error bars representing standard deviation were applied in all graphs to reflect the variability among the triplicate measurements obtained for each treatment and sampling point.

3. Results and Discussion

3.1. Physicochemical Analysis

Before starting the tests, physicochemical analyses of the wastewater were carried out in the laboratory of Milagro State University. pH, temperature, electrical conductivity (EC), dissolved oxygen (DO), biochemical oxygen demand (BOD), chemical oxygen demand (COD), salinity, total dissolved solids (TDSs), and hardness were analyzed.

pH: During the evaluation period, a reduction in pH was observed in the control sample as well as in the two treatments, E. crassipes and L. minor. In the control treatment, the pH decreased from 9.57 ± 0.04 to 8.41 ± 0.08 from the initial to the final analysis (Figure 1). The largest recorded decrease was 0.32 units between days 21 and 26 and 0.29 units between days 01 and 06 of the analyzed period, while a minimal decrease of 0.05 units was observed between days 11 and 16 during the evaluation period (Appendix A).

In the treatment containing E. crassipes, the pH varied from 9.59 ± 0.01 to 7.67 ± 0.04. The largest decrease recorded was 0.71 units between days 01 and 06 of the treatment, while the minimum decrease was 0.06 units between days 06 and 11, resulting in a total pH reduction of 1.92 units. Microbial degradation in the wastewater, which converts organic matter into acids [38], was the cause of the pH reduction in all treatments, with the L. minor treatment showing the highest microbial degradation. The L. minor treatment exhibited a pH reduction pattern similar to that of E. crassipes, with a variation from 9.53 ± 0.02 to 7.53 ± 0.06 during the evaluation period. The largest decrease in this treatment was 0.92 units between days 06 and 11, the smallest was 0.25 units between days 21 and 26, and the decrease was 0.24 units between days 26 and 31 of the analyzed period. However, between days 11 and 16, the pH increased by 0.13 units. Among all treatments, L. minor showed the greatest reduction in pH, followed by E. crassipes and the control treatment.

The high pH value (9.57) recorded in the initial characterization of the water sample can be attributed to several factors. First, alkaline industrial discharges from agro-industrial activities, detergents, and chemical effluents in the area could have contributed to the increase in pH. In addition, the high photosynthetic activity of microalgae during the day may cause CO₂ depletion in the water, which subsequently raises the pH. The presence of carbonates and bicarbonates in the water, influenced by dissolved inorganic carbon, could also contribute to the maintenance of elevated pH levels [39].

Electrical Conductivity (EC): During the evaluation period, electrical conductivity (EC) varied from 4012.33 ± 3.06 to 3807.67 ± 8.33 µS/cm in the control; from 4011.00 ± 3.46 to 2722.00 ± 9.54 µS/cm in E. crassipes; and from 4002.00 ± 2.00 to 1926.33 ± 49.24 µS/cm in L. minor (Figure 2). In the control, the largest reduction in EC was 40.67 µS/cm between days 06 and 11, while the smallest was 28.00 µS/cm between days 21 and 26 of the treatment.

The E. crassipes treatment showed notable fluctuations in EC reduction, with a maximum decrease of 290.00 µS/cm between days 11 and 16, and a minimum reduction of 130.00 µS/cm between days 21 and 26. The average EC reduction in this treatment was 214.83 µS/cm per day during the 30-day evaluation period. Finally, the L. minor treatment showed a maximum decrease of 821.0 µS/cm between days 26 and 31, and a minimum reduction of 185.00 µS/cm between days 01 and 06. The average EC reduction was 24.46 µS/cm per day in this treatment.

Among the three samples, L. minor showed the greatest decrease in EC, followed by E. crassipes and the control. The reduction in EC is due to nutrient absorption by the macrophytes, suggesting that L. minor has the highest nutrient absorption potential among the evaluated plants [40].

Total Dissolved Solids (TDSs): The concentration of total dissolved solids (TDSs) decreased during the treatment period, from 3068.33 ± 44.09 to 2187.00 ± 115.86 mg/L in the control, from 3130 ± 57.82 to 1099.00 ± 42.32 mg/L in the treatment with E. crassipes, and from 3063.67 ± 45.76 to 1348.67 ± 33.08 mg/L in the L. minor treatment (Figure 3). The treatment with E. crassipes achieved the highest TDS removal between days 11 and 16 of the evaluation, while the control showed an increase in TDSs between days 01 and 06.

In the control treatment, TDS removal fluctuated between 17.33 and 451.67 mg/L between consecutive evaluations, with the greatest reduction occurring between days 11 and 16, and the smallest between days 26 and 31. In the case of E. crassipes, there were larger fluctuations in TDS removal, with a maximum decrease of 929.67 mg/L between days 11 and 16, and a minimum of 0.33 mg/L between days 01 and 06 of the evaluation. L. minor showed an increase in TDSs between days 01 and 06, resulting in a rise of 60.67 mg/L between consecutive evaluations. The greatest reductions occurred between days 06 and 11 and between days 11 and 16. The smallest reduction of 83.00 mg/L was observed between days 16 and 21 of the treatment.

TDSs represent the amount of dissolved solutes in the wastewater, and plants absorb these dissolved salts through their roots as they grow, which decreases the concentration of TDSs over time [41]. The treatment with E. crassipes showed the highest efficiency in reducing TDSs, highlighting its potential for salt absorption in wastewater.

Salinity: Salinity refers to the total amount of dissolved salts present in wastewater. During the analysis period, salinity concentration decreased from 798.00 ± 7.94 to 755.00 ± 10.39 mg/L in the control, from 770.67 ± 12.86 to 337.00 ± 12.12 mg/L in the treatment with E. crassipes, and from 745.00 ± 6.25 to 395.33 ± 9.45 mg/L in the L. minor treatment (Figure 4 and

Table A1. Effect of floating macrophytes on wastewater.

Time	Samples	pH	Temp °C	EC (µS/cm)	TDS (mg/L)	Salinity (mg/L)	BOD (mg/L)	COD (mg/L)	DO (mg/L)	Hardness (mg/L)
Day 01	Control	9.57 ± 0.04	25.20 ± 0.00	4012.33 ± 3.06	3068.33 ± 44.09	798.00 ± 7.94	355.33 ± 5.51	557.00 ± 4.36	2.25 ± 0.02	334.67± 4.51
Day 01	E. crassipes	9.59 ± 0.01	25.23 ± 0.06	4010.33 ± 3.46	3052.67 ± 57.82	793.00 ± 12.86	359.67 ± 6.03	555.33 ± 0.58	2.24 ± 0.04	334.67 ± 4.36
Day 01	L. minor	9.53 ± 0.02	25.27 ± 0.00	4009.67 ± 2.00	3088.67 ± 45.76	776.00 ± 6.25	356.33 ± 6.51	557.67 ± 3.21	2.25 ± 0.03	335.67 ± 4.04
Day 06	Control	9.58 ± 0.06	25.27± 0.32	4011.00± 0.00	3130.00 ± 19.08	770.67 ± 0.00	356.67 ± 8.08	555.33 ± 3.79	2.25 ± 0.03	336.00 ± 3.79
Day 06	E. crassipes	9.53 ± 0.06	25.23 ± 0.40	4008.67± 35.60	3141.00 ± 18.18	756.67 ± 3.46	353.00 ± 5.51	557.67 ± 4.04	2.27 ± 0.02	334.00 ± 4.04
Day- 06	L. minor	9.28± 0.03	25.20 ± 0.21	4005.67± 21.50	3120.67 ± 7.37	754.67 ± 8.62	355.67 ± 8.08	514.67 ± 3.06	2.25 ± 0.04	334.67 ± 3.06
Day 11	Control	9.53 ± 0.03	25.20 ± 0.17	4002.00± 7.23	3063.67 ± 64.63	745.00 ± 8.89	358.67 ± 10.00	557.67 ± 3.79	2.26 ± 0.03	337.67 ± 3.79
Day 11	E. crassipes	9.28 ± 0.02	25.53 ± 0.26	3996.67± 7.02	3065.67 ± 81.50	761.67 ± 7.64	357.00 ± 3.21	514.67 ± 4.93	2.18 ± 0.02	305.67 ± 4.93
Day 11	L. minor	8.88± 0.21	25.70 ± 0.15	3990.00± 10.54	3087.67 ± 88.15	774.33 ± 6.43	358.33 ± 3.51	481.00 ± 5.69	2.14 ± 0.05	269.00 ± 5.69
Day 16	Control	9.28± 0.03	26.07 ± 0.35	3984.00 ± 5.03	3130.00 ± 59.50	788.00 ± 11.59	354.33 ± 0.00	514.67 ± 4.62	2.06 ± 0.04	232.67 ± 4.62
Day 16	E. crassipes	8.88 ± 0.06	25.90 ± 0.10	3927.67± 28.01	3144.00 ± 33.62	779.00 ± 5.51	348.00 ± 3.46	481.00 ± 2.08	2.02 ± 0.10	226.00 ± 2.08
Day 16	L. minor	9.13 ± 0.05	26.03 ± 0.29	3891.00± 14.53	3135.00± 127.58	768.00± 9.50	336.67 ± 4.51	478.00 ± 4.58	1.98 ± 0.03	223.33 ± 4.58
Day 21	Control	8.88 ± 0.05	26.10 ± 4.35	3856.00± 15.04	3129.67 ± 65.91	759.00 ± 6.81	331.67 ± 3.21	481.00 ± 4.00	1.96 ± 0.03	217.67 ± 4.00
Day 21	E. crassipes	9.13 ± 0.09	26.20 ± 0.12	3840.33± 20.42	3123.00 ± 87.64	753.33 ± 9.02	336.00 ± 6.03	478.00 ± 4.00	2.03 ± 0.02	219.67 ± 4.00
Day 21	L. minor	9.22 ± 0.02	26.30 ± 0.19	3800.67± 20.98	3123.33 ± 73.73	744.00 ± 5.57	340.00 ± 4.73	510.33 ± 7.77	2.10 ± 0.02	221.33 ± 7.77
Day 26	Control	9.13 ± 0.07	26.23 ± 0.17	3773.33± 18.01	3124.33 ± 79.76	734.67 ± 8.02	345.67 ± 5.51	478.00 ± 4.36	2.18 ± 0.04	226.33 ± 4.36
Day 26	E. crassipes	9.22 ± 0.05	26.97 ± 0.36	3833.33± 20.88	3060.33± 63.89	749.33 ± 7.77	343.00 ± 5.19	510.33 ± 2.00	2.13± 0.01	225.33 ± 2.00
Day 26	L. minor	8.28 ± 0.07	27.47 ± 0.39	3897.33± 24.99	2960.33 ± 64.86	765.33 ± 6.51	339.67 ± 2.08	428.33 ± 7.64	2.05 ± 0.02	223.33± 7.64
Day 31	Control	9.22 ± 0.08	28.10 ± 0.15	3943.33 ± 8.33	2889.33 ± 115.86	778 ± 10.39	332.00 ± 5.86	510.33± 4.58	1.97 ± 0.04	222.67 ± 4.58
Day 31	E. crassipes	8.28 ± 0.04	28.10 ± 0.21	3827.33 ± 9.54	2676.33 ± 42.32	762.33 ± 12.12	324.33 ± 4.16	428.33 ± 8.00	1.95 ± 0.04	215.00 ± 8.00
Day 31	L. minor	8.21 ± 0.06	28.07 ± 0.15	3713.67 ± 49.24	2557.33 ± 33.08	756.00 ± 9.45	315.33 ± 7.77	437.67 ± 1.53	1.94 ± 0.04	207.33± 1.53

Note: Control = treatment without floating macrophytes. The values presented are averages with their respective standard deviation.

).

In the control treatment, the largest salinity reduction was 10.00 mg/L, observed between days 01 and 06 and between days 06 and 11. However, the control treatment showed an increase in salinity between days 16 and 21, while the smallest decrease occurred between days 21 and 26 of the evaluation. The treatment with E. crassipes showed significant fluctuations in salinity removal, with a maximum reduction of 198.67 mg/L between days 16 and 21, and a minimum of 11.33 mg/L between days 06 and 11 of the treatment.

L. minor treatment also showed significant results, very close to those of E. crassipes, with salinity removal fluctuating between 10.33 and 86.33 mg/L between consecutive evaluations, reaching its greatest reductions between days 11 and 16 of the treatment.

Although this study focuses on the observed decrease in salinity over time, occasional fluctuations were observed. These increases in salinity could be due to natural evaporation, which concentrates dissolved salts in the water. In addition, the degradation of organic matter and the release of ions from sediments could contribute to temporary increases in salinity. The limited inter-exchange of water in the experimental system could also have influenced salt accumulation during specific periods [30].

The decrease in salinity in wastewater is due to the absorption of dissolved salts by the plants through their roots which use these salts as macronutrients [42]. In conclusion, the E. crassipes treatment achieved the greatest salinity reduction, demonstrating its high capacity to absorb salts from wastewater.

Biochemical Oxygen Demand (BOD): The values of biochemical oxygen demand (BOD) decreased during the treatment period in the evaluated systems, except for the control treatment, which increased by 6.67 mg/L between days 16 and 21. In the control sample, the BOD value decreased from 355.33 ± 5.51 to 308.33 ± 5.86 mg/L; in the E. crassipes treatment, it decreased from 356.67 ± 6.03 to 233.33 ± 4.16 mg/L; and in the L. minor treatment, it decreased from 358.67 ± 6.51 to 229.67 ± 7.77 mg/L (Figure 5 and Table A1). According to Deshmukh et al. [43], similar trends in the reduction in chemical oxygen demand (COD) were observed.

In the control treatment, the highest BOD removal was 22.33 mg/L between days 06 and 11, while the smallest reduction of 1.00 mg/L occurred between days 01 and 06. Meanwhile, the E. crassipes sample showed considerable variability in the rate of BOD removal, with a maximum of 27.67 mg/L between days 11 and 16 and a minimum of 14.67 mg/L between days 26 and 31.

The L. minor treatment proved to be the most efficient, with BOD removal fluctuating between 10.33 and 52.00 mg/L between consecutive periods. The maximum BOD removal, 32 mg/L, was observed between days 21 and 26 of the treatment, while the minimum, 10.33 mg/L, occurred between days 11 and 16. Compared to the control and E. crassipes samples, L. minor showed greater tolerance to the stress caused by pollution, reflected in its sustained growth and the generation of new leaves.

Biochemical oxygen demand (BOD) is a measure of the amount of oxygen that microorganisms require to break down organic waste present in wastewater and is an indicator of the level of contamination. A high BOD indicates high oxygen consumption, which can lead to anaerobic conditions. Microorganisms that degrade organic matter release enzymes that convert complex compounds into simple nutrients, which are absorbed by the plants [44].

BOD reduction in phytoremediation systems is mainly due to microbial biodegradation in the rhizosphere of E. crassipes and L. minor, which favors aerobic decomposition of organic matter [45]. In addition, macrophytes can absorb certain compounds, contributing to the removal of organic load [16]. The decrease in BOD in the control suggests that other processes, such as sedimentation and natural degradation, also influenced the observed reduction [46]. These combined mechanisms highlight the efficiency of phytoremediation in improving water quality.

Chemical Oxygen Demand (COD): Chemical oxygen demand (COD) represents the amount of oxygen required to oxidize organic and inorganic compounds present in wastewater [44]. During the control treatment, COD decreased from 557.00 ± 4.36 to 204.00 ± 4.58 mg/L over the evaluation period. Although it decreased, it was observed that from day 21 until the end of the evaluation, there was a 0.00 mg/L reduction (Figure 6 and Table A1). The maximum COD removal occurred between days 01 and 06 of the treatment, while the lowest value was observed between days 21 and 31.

On the other hand, the E. crassipes treatment showed a COD decrease from 555.33 ± 0.58 to 178.00 ± 8.00 mg/L. The maximum COD removal of 337.67 mg/L occurred between days 01 and 03, while an increase of 2.00 mg/L was observed from day 26 to day 31.

In the L. minor treatment, COD was reduced from 557.67 ± 3.21 to 174.33 ± 1.53 mg/L, representing the highest COD removal rate among the treatments, with a total reduction of 68.6%. COD removal fluctuations varied between 0.67 and 331.33 mg/L between consecutive periods, with the maximum value (331.33 mg/L) being reached in the first five days of treatment, while the minimum removal value (0.67 mg/L) was recorded between days 11 and 16 of treatment.

Previous studies, such as that by Deshmukh et al. [43], reported similar trends in COD reduction. Additionally, the presence of plants in wastewater can reduce dissolved CO₂ during periods of high photosynthetic activity. This process increases dissolved oxygen concentration, creating aerobic conditions that favor aerobic bacterial activity, which is crucial for the reduction in both BOD and COD [47].

Dissolved Oxygen (DO): Dissolved oxygen (DO) values showed a significant decrease across the different treatments, from 2.25 ± 0.02 to 1.26 ± 0.04 mg/L in the control sample, from 2.25 ± 0.04 to 1.65 ± 0.04 mg/L in E. crassipes, and from 2.27 ± 0.03 to 1.65 ± 0.04 mg/L in the L. minor treatment (Figure 7 and Table A1). The total DO decrease was recorded at 44.0%, 26.6%, and 27.3% in the control, E. crassipes, and L. minor treatments, respectively, over the treatment period. The greatest reduction in DO was observed in the control sample, followed by L. minor and E. crassipes.

In the control treatment, the greatest reduction in DO occurred between days 16 and 21, with a value of 0.26 mg/L, and the smallest reduction was between days 11 and 16, at 0.07 mg/L. In the E. crassipes treatment, there was an increase of 0.01 between days 21 and 26, as well as a maximum decrease of 0.29 mg/L in the first five days and a minimum decrease of 0.04 mg/L between days 06 and 11 of the evaluation period. In the L. minor sample, the greatest reduction occurred between days 16 and 21, and the smallest reduction was 0.04 mg/L between days 06 and 11 of the treatment.

Dissolved oxygen is absorbed by microorganisms during the breakdown of organic compounds present in wastewater. The continuous decrease in DO levels suggests intense microbial activity during the experiment [47]. However, the reduction in DO was lower in the treatments with macrophytes, suggesting that they contributed to the introduction of oxygen into the wastewater through photosynthesis.

The observed decrease in dissolved oxygen (DO) levels over time (Figure 7) correlates with the increase in temperature recorded in all treatments. This is consistent with the inverse relationship between temperature and DO solubility, where higher temperatures reduce the ability of water to retain oxygen [48]. Therefore, it is likely that the decrease in DO levels was primarily due to thermal effects, rather than the exclusive action of macrophyte systems.

However, while temperature plays an important role, the presence of macrophytes could have influenced local oxygen dynamics. Root zones of E. crassipes and L. minor are known to favor microbial activity, which consumes oxygen by degrading organic matter, which could contribute to the reduction in DO [46]. In addition, the oxygenation effects of macrophyte photosynthesis may have varied as a function of shading, biomass growth, and diurnal fluctuations.

Hardness: The hardness in the wastewater also decreased across all treatments during the evaluation period. In the control sample, hardness was reduced from 334.67 ± 4.51 to 204.00 ± 4.58 mg/L; in E. crassipes, from 336.00 ± 4.36 to 178.00 ± 8.00 mg/L; and in L., from 337.67 ± 4.04 to 174.33 ± 1.53 mg/L (Figure 8 and Table A1). The total decrease in hardness was recorded at 39.04%, 47.02%, and 48.37% in the control, E. crassipes, and L. minor treatments, respectively, throughout the treatment period. The greatest reduction in hardness was observed in the L. minor sample, followed by E. crassipes and the control. The variation in hardness removal between consecutive evaluation periods ranged from 0.00 to 102.00 mg/L in the control, from −2.0 to 118.33 mg/L in E. crassipes, and from 0.67 to 111.33 mg/L in L. minor.

The mechanism of hardness removal is attributed to the absorption of calcium and magnesium salts by the plants’ roots [49]. Previous research by has also reported similar trends in the reduction in hardness in wastewater.

The results obtained are in agreement with previous studies that have shown that floating macrophytes, such as E. crassipes and L. minor, can significantly reduce BOD and COD in wastewater due to their ability to absorb organic matter and nutrients [16]. However, the low reduction in electrical conductivity and salinity indicates that these treatments are not fully effective in removing dissolved inorganic compounds, such as nitrate and phosphate salts, which can contribute to eutrophication of receiving water bodies [50].

Since these parameters are critical for improving the quality of treated water, it is recommended to complement the use of floating macrophytes with additional treatments, such as activated carbon filtration or reverse osmosis membranes, to enhance the reduction in inorganic pollutants and ensure effluent quality [39].

3.2. A Statistical Algorithm for the Physicochemical Parameters in the Different Samples

Figure 9 shows the formation of three groups in relation to nine variables, which include pH, temperature (°C), electrical conductivity (µS/cm), total dissolved solids (mg/L), salinity (mg/L), biochemical oxygen demand (BOD, mg/L), chemical oxygen demand (COD, mg/L), dissolved oxygen (DO, mg/L), and hardness (mg/L).

Figure 9a presents the replicas corresponding to the first 16 days of the evaluation process, showing inertia of 88.8%. On the other hand, Figure 9b indicates that the inertia reached 95.4% from day 16 until the end of the evaluation period.

Figure 9a presents six replicas of the analyzed samples, grouped into three clusters. The green cluster includes the control sample replicas (C1, C2, C3), followed by the blue cluster, which groups the replicas of E. crassipes (E.c1, E.c2, E.c3), and finally, the red cluster corresponds to L. minor (L.m1, L.m2, L.m3). During the first 16 days of treatment, it was observed that the control sample had a higher correlation with the evaluated parameters, with salinity being the parameter with the greatest influence on dimension 1.

Figure 9b illustrates the distribution of six replicas of the different samples analyzed from day 16 to the end of the treatment. In cluster 1 (green), the three replicas of the control sample (C1, C2, C3) are found, showing a stronger association with pH, COD, electrical conductivity (EC), total dissolved solids (TDSs), hardness, BOD, and salinity parameters. Cluster 2 (blue), corresponding to E. crassipes (E.c1, E.c2, E.c3), shows a higher correlation with dissolved oxygen (DO). On the other hand, cluster 3 (red), corresponding to L. minor (L.m1, L.m2, L.m3), is mainly associated with temperature and dissolved oxygen, due to the influence of temperature on dissolved oxygen and aquatic life.

These changes provide a favorable response to the study, as they indicate that floating macrophytes are effective in the phytoremediation process. The significant reduction in key parameters, such as biochemical oxygen demand (BOD), chemical oxygen demand (COD), total dissolved solids (TDSs), and hardness, among others, suggests that floating aquatic plants like E. crassipes and L. minor not only tolerate contaminated environments but also actively participate in the removal of pollutants.

The HJ-Biplot has been widely used in environmental studies to analyze water quality and evaluate remediation treatments. Silva Viamonte (2022) applied this methodology to examine the distribution of environmental indicators in different municipalities, demonstrating its effectiveness in identifying multi-variate patterns and relationships between water quality variables [37]. Similarly, Osorto Núñez et al. (2023) used multivariate techniques to evaluate water quality in the estuaries of the Gulf of Fonseca, Honduras, classifying water bodies according to their level of contamination and anthropogenic activity [51]. These findings support the application of this technique in the present study, since it provides a clear visualization of the effects of phytoremediation on the physicochemical and microbiological parameters of wastewater.

3.3. The Microbial Load of the Effluent

Table 1. Microbiological analysis of wastewater during treatment of the different samples.

Samples	Total Coliforms (NMP/mL)	Fecal Coliforms (NMP/mL)	E. coli (UFC/mL)	Fecal Staphylococcus (UFC/mL)
Control	3535.00 a ± 20.5	523.67 b ± 11.5	522.33 b ± 1.5	299.00 c ± 7.5
E. crassipes	3821.33 a ± 25.7	629.33 b ± 7.1	517.00 c ± 15.1	251.33 d ± 14.1
L. minor	3928.67 a ± 12.7	552.67 b ± 9. 2	413.67 c ± 3.1	274.00 d ± 4.6

Note: Control: sample without floating macrophytes. The values presented are averages with their respective standard deviation. Different letters indicate significant differences between treatments.

shows the levels of total coliforms, fecal coliforms, E. coli, and fecal Staphylococcus in the different water treatment samples.

The results indicate that no significant differences (p < 0.05) were observed in total coliforms (a) and fecal coliforms (b) among the samples analyzed. However, significant differences were found in E. coli and fecal Staphylococcus, particularly between the control sample and the treatments with floating macrophytes, which showed no significant differences between them. The highest concentration of E. coli was observed in the control sample (522.33 CFU/mL), followed by the sample treated with E. crassipes (517.00 CFU/mL), while the lowest concentration was recorded in the sample with L. minor (413.67 CFU/mL). Likewise, the highest concentration of fecal Staphylococcus was detected in the control sample (299.00 CFU/mL), followed by the treatment with L. minor (274.00 CFU/mL), and the lowest concentration was observed in the sample with E. crassipes (251.3 CFU/mL).

It is important to mention that the concentrations recorded in all samples evaluated far exceed the permissible values established by current Ecuadorian regulations, specifically according to the Unified Text of Secondary Legislation of the Ministry of the Environment (TULSMA), which establishes maximum limits of 1000 NMP/100 mL for total coliforms and 200 NMP/100 mL for fecal coliforms in treated effluents [52]. These results highlight the need to improve or complement the applied methodology, possibly by combining it with other physical or chemical treatments, in order to achieve adequate microbiological levels for safe discharge into receiving natural bodies and prevent risks associated with public health [53,54].

At the end of the experiment, both E. crassipes and L. minor showed an increase in biomass, indicating their adaptation to the experimental conditions and their active role in phytoremediation. However, biomass was not quantified in this study. Future research should consider biomass monitoring to evaluate the correlation between plant growth and pollutant removal efficiency.

4. Limitations and Future Research

Despite the interest in using robust statistical methods such as the HJ-Biplot, a significant limitation of this study is the absence of nitrate and phosphate concentration measurements in the treated wastewater. These parameters are essential for comprehensively assessing effluent quality before discharge into natural water bodies, as high concentrations can lead to eutrophication, significantly reducing water quality and negatively impacting aquatic biodiversity [28,39]. Future studies should include the determination of nitrates and phosphates using widely recognized standardized methods to provide a more rigorous and comprehensive evaluation of the phytoremediation efficiency of E. crassipes and L. minor in wastewater. This will yield more robust results aligned with current environmental regulations.

Another important limitation of this study is the short experimental duration, which was restricted to 31 days. This timeframe was selected considering the rapid response of floating macrophytes, such as E. crassipes and L. minor, in influencing water quality and the need to maintain controlled experimental conditions. Previous studies have shown that phytoremediation can generate significant changes in a short period, particularly in the reduction in biochemical oxygen demand (BOD), chemical oxygen demand (COD), and microbial load [37,51]. However, a longer evaluation period would allow for a more in-depth analysis of the long-term effects of phytoremediation, potential contaminant accumulation in plants, and the stability of water quality parameters after treatment.

5. Conclusions

This study confirmed that phytoremediation is an efficient technology for wastewater treatment. The statistical results demonstrated that duckweed (L. minor) was particularly effective in significantly reducing the parameters of pH, electrical conductivity (EC), biochemical oxygen demand (BOD), chemical oxygen demand (COD), and hardness. Regarding salinity and total dissolved solids (TDSs), a reduction was observed in all treatments, with the maximum decrease occurring in the E. crassipes treatment.

The absorption of nutrients and dissolved salts by the roots of macrophytes is the primary mechanism for reducing solutes in wastewater, not only lowering the pollutant load but also improving overall water quality by reducing salinity and dissolved solids. This process is essential for natural purification. However, it is important to harvest the macrophytes after they reach maturity, as plant decomposition in water would release the absorbed nutrients back into the water, compromising the purification process.

Furthermore, the reduction in BOD and COD in the treatments with E. crassipes and L. minor highlights the ability of these plants to enhance microbial activity and improve the efficiency of the organic compound degradation process. It is noteworthy that the control treatment, which did not include floating macrophytes, showed significant differences in the levels of E. coli and fecal Staphylococcus, attributed to the absence of phytoremediation processes.

In contrast, the samples treated with floating macrophytes did not show significant differences between them, indicating similar efficiency in reducing pathogenic bacteria and other contaminants, regardless of the species used. This underscores the positive impact of macrophytes in wastewater treatment through the removal of pathogenic bacteria.