Agro-Industrial Effluent Phytoremediation with Lemna gibba and Hydrocotyle ranunculoides in Water Recirculating Mesocosms

2.1 Sampling and cultivation of Lemna gibba and Hydrocotyle ranunculoides

Lemna gibba specimens were collected in the reservoir Ing. Carlos F. Roggero (34° 41′ 11.32″ S, 58° 51′ 59.12″ W). Hydrocotyle ranunculoides specimens were collected from cultivation ponds located at the Museo Argentino de Ciencias Naturales “Bernardino Rivadavia.”

2.2 Bioassay with continuous recirculation of wastewater

For acclimatization purposes, individuals of both species were placed for a week in containers filled with effluent from a poultry industry discharged into the La Choza stream (General Rodriguez, Buenos Aires, 34° 39′ 48.64″ S, 58° 56′ 44.43″ W). After the acclimatization period, individuals of similar size and weight grown in greenhouse were selected. Nine independent continuous flow reactors were installed each filled with 9 L undiluted wastewater (Fig. 1). A submersible pump was installed in each container for the continuous recirculation of the wastewater at a rate of 0.6 mL/s, in order to recreate the low current velocities of constructed wetlands for effluent treatment 12. An equivalent amount of biomass of 1.21 ± 0.04 g (dry weight, DW) of L. gibba (Lg treatment), H. ranunculoides (Hr treatment) or both species together (Lg + Hr treatment, 50% w/w of each species) was placed in each randomly selected mesocosm. The reactors were placed in greenhouse with natural and controlled artificial light photoperiod of 16:8 (light/dark) for 6 days.

2.3 Analytical determinations

Water samples were collected from each reactor at the beginning and at the end of the assay, determining the physical and chemical variables listed in Table 1 by means of standardized methodologies 13-17.

Given the effects of evapotranspiration 18, 19, the removal percentage of each analyte was calculated according to Eq. 1:

(1)

where R% is the removal percentage of the variable considered, v_i and c_i are the initial volume and concentration in each reactor, v_f and v_f the final values and the product v c is the load (L) or mass of the analyte in each reactor.

2.4 Count of total coliforms and Escherichia coli

Determination of total coliform bacteria and E. coli was carried out by plate count technique to assess water quality. Briefly, 1 mL of decimal serial dilutions of the sample were inoculated in selective and differential medium CHRO ECC Agar (CHROMagar^©). Plates were incubated at 35°C for 48 h 13.

2.5 Allium cepa genotoxicity test

A wastewater mutagenicity test was performed, following the methodology described by Matsumoto et al. 20. A homogeneous seed lot of A. cepa grown under organic conditions (variety Valcatorce) was used for the test. The seeds of this species are preferred because of their genetic and physiological homogeneity and their availability throughout the year. A total of 100 seeds were placed in Petri dishes containing filter paper inoculated with 5 mL of sample or control. Distilled water was used as a negative control and methyl methanesulfonate (MMS) 2 × 10⁻⁴ M as a positive control. Wastewater samples were collected at the beginning and at the end of the assay for each treatment (Lg, Hr, and Lg + Hr). The plates were kept at 22–24°C for 4 days. The seeds with their roots were fixed for 24 h in acetic Carnoy, then washed with distilled water and preserved in 70% ethanol for later observation 21. Chromosomes of meristematic root cells were stained with orcein in 2% acetic acid. A total of 5000 cells were counted in at least five slides. The mitotic index (MI) in 5000 cells and the frequency of chromosomal aberrations (CA) in metaphase and anaphase (bridges, fragments, delayed chromosomes, and others) were analyzed for a total of at least 200 anaphases. The frequency of micronuclei (MN) in 2000 interfaces was analyzed as well.

2.6 Statistical analysis

The existence of statistically significant differences (p < 0.05) between the values of concentrations or loads of variables were determined by analysis of variance (ANOVA) test with post hoc Tukey or Dunnett comparisons as appropriate. Non-normal variables and/or heteroskedastic were previously transformed 22. Genotoxicity data from A. cepa test were analyzed using the Kruskal–Wallis test 20.

3.1 Physical and chemical variables

During the test, the water temperature in the reactors was 25 ± 5°C. The pH increased throughout the test, reaching a peak of 8.79 on day 5 of treatment with Lg + Hr (Fig. 2). The increase in all treatments could be associated to the notorious periphyton growth on the walls of all containers (data not shown), which promoted photosynthesis and thus increased pH 23.

The average electrical conductivity (EC) at the start of the trial was within the range of 1388–1394 μS/cm, increasing in all treatments towards the end of the trial by the effect of evapotranspiration, with a maximum value of 1430 μS/cm in the Hr treatment. This effect was less noticeable in the Lg treatment.

The value of dissolved oxygen (DO) concentration was 0 mg/L in all reactors during the first 2 days, and can be attributed to anoxic fermentation and remineralization resulting from bacterial activity. Then, DO increased gradually until day 5, with a maximum average value of 12.73 mg/L in the Lg + Hr treatment; most variability between replicates was observed in this treatment (Fig. 2). In sparse vegetation, oxygen production may exceed respiration resulting in an oxygen surplus, whereas the opposite may occur in dense vegetation 24. Subsequently, at the end of the test a slight decrease in DO concentration was observed in all treatments, although values remain close to saturation. It should be noted that measurements were performed daily at the same time to avoid changes by photosynthetic activity occurring during the day. The mean concentration of total suspended solids (TSS) in wastewater at the beginning of the assay was 141 mg/L. Similar reductions of TSS concentration were observed in all treatments, with mean final values between 13 and 18 mg/L.

In all treatments, ammonium (NH₄⁺-N) was the primary nitrogenous form at the start and at the end of the assay while nitrite (NO₂⁻-N) and nitrate (NO₃⁻-N) concentrations were always very low (<0.05 mg/L). The final concentrations of NH₄⁺-N and DIN were significantly lower than the initial ones (Fig. 3). The assimilation of ammonium nitrogen by macrophytes and periphyton along with denitrification that occurred in the anoxic environment of the early stage of the experiment, resulted in a removal of 74.1% in the Lg treatment, 65.7% in the Hr treatment, and 77.9% in the Lg + Hr treatment (Table 2). The greater efficiency in nutrients uptake in mixed cultures could be associated with the largest area to hold periphyton. Studies showed that this community plays an important role in nutrients uptake 25. In addition, Levi et al. 26 proposed that the coexistence of different species of macrophyte increases the efficiency of ammonium uptake, thus higher concentrations may be processed before it is saturated. Cardinale 27 argued that systems with more biological diversity can better exploit niche opportunities because they can develop different strategies that optimize the process of resources uptake, in this case N.

The initial concentration of SRP in the wastewater was 3.35 mg/L and the concentration of TP was 5.93 mg/L, with a SRP/TP ratio of 0.565. Concentrations of SRP and TP in the wastewater were significantly reduced in all treatments after 6 days (Fig. 3). Values of the SRP/TP ratio at the end were higher than at the beginning of the experiment for all treatments; the values were 0.752, 0.748, and 0.752 in Lg, Hr, and Lg + Hr, respectively. These values indicate the progress of bacterial degradation and deposition of particulate matter throughout the test, increasing the dissolved inorganic fraction of P. At the start of the assay, the particulate fraction of P was 43.5% of TP and at the end it was reduced to 10%, which reveals the importance of adsorption, complexation, and sedimentation indicated by Brix 28 as the main factors in the decrease of P in artificial wetlands. Co-precipitation of P is favored in the Pampas streams where the substrates have high content of CaCO₃ 29. Ansari et al. 11 found that the removal of N and P was up to 92 and 78%, respectively, in cultures of more than one species of macrophytes (Eichhornia + Spirodela + Lemna) and was higher than in monocultures of each single species.

The maximum COD reduction was 70.6%, observed in the Lg + Hr treatment, with a final concentration of 160 mg/L (Fig. 4). From stoichiometry calculations, biodegradable organic matter should become microbial biomass or mineralized in 5 days. However, the nature of the organic matter and certain environmental conditions such as lighting can alter this simple stoichiometric relationship. Given the test results, conditions were favorable for attenuating BOD₅ levels, which dropped as much as 96.9% in the Lg treatment, with a final value of 10.4 mg/L (Fig. 4). This decrease was mainly due to bacterial degradation and deposition of particulate matter on the surface of containers and plants 28. Depletion of DO in the first two days of the test (deoxygenation step) was achieved due to active bacterial degradation. However, from then until the end of the trial DO levels increased up to supersaturation (Fig. 2). Possibly, aerobic and facultative bacterial activity was very intense and efficient during the first 2 days rapidly consuming the DO incorporated by diffusion and water movement and depleting the amount of biodegradable material. Then, bacterial activity decreased and DO levels grew by photosynthetic activity of periphyton, oxygenation from the roots of the macrophytes 30 and the positive balance in oxygenation by diffusion from the atmosphere by the movement of water circulation 8. The BOD₅/COD ratio was 0.62 at the start and decreased over time, with a minimal final value of 0.06 (Lg treatment) and a maximum final value of 0.12 (Hr treatment). The initial value of the ratio is as expected (0.6) in wastewater with a high proportion of readily biodegradable organic matter, such as municipal wastewater 31. Initial concentrations of BOD₅ and nutrients were similar to those reported by Basílico et al. 32.

Loads of nutrients were calculated for each reactor at the start and at the end of the assay (Table 2). Statistically significant differences (p < 0.05) between the final loads of NH₄⁺-N, DIN, TOC, and POC and highly significant differences (p < 0.01) between final loads of SRP and TP were found for the different treatments (Table 2). In the case of NH₄⁺-N, DIN, SRP, TP, and POC final loads (L) were in the order of L_Hr > L_Lg > L_Lg+Hr, while for TOC of L_Hr > L_Lg+Hr> L_Lg. While an important removal for these variables was achieved in all treatments, the performance of the Hr treatment was lowest in most cases, except for TSS (Table 2). DIN and TP %R were higher than those reported by Basílico et al. 13 in the effluent bio-treatment with S. intermedia and periphyton.

The initial value of total coliforms and E. coli was 210 CFU/mL and both indices dropped to 5 CFU/mL (Hr and Lg + Hr treatments) and 9 CFU/mL (Lg treatment) (Fig. 5). While the initial load of coliforms indicators in wastewater was low compared to other effluents of similar characteristics, a reduction of 97.6% was reached in the Hr and Lg + Hr treatments. This value is similar to that obtained in wastewater pathogen removal treatments using constructed wetlands 33 but lower than the typical 99% reduction expected in a chemical disinfection treatment 31. The decrease of total coliforms and E. coli determined after the test reveals the efficiency of the treatment in wastewaters with moderate or low pathogen load as the one used in the assay. However, this decrease cannot be attributed to a particular species or to the presence of both. The need for subsequent disinfection of the effluent prior to its discharge into a water body should be pointed out.

3.2 Plant biomass production

The final biomass of L. gibba (Lg treatment) was 1.61 ± 0.29 g DW, with a relative growth rate RGR of 0.051 day⁻¹. For H. ranunculoides the final biomass was 1.95 ± 0.14 g DW, with an RGR of 0.075 day⁻¹ and in the Lg + Hr treatment the final biomass of both species was 1.87 ± 0.26 g DW and the global RGR in this case was 0.071 day⁻¹. Concentrations of nutrient elements in the assay exceeded those in conventional cultures 34. However, the growth rates in the assay were much lower than the growth rate in the culture media. The effect probably resulted from the gradient change occurring in the recirculating mesocosm, which might induce stress in plants. Most Lemnaceae species are sometimes sensitive to changes in the environment 34. Expected growth of macrophytes would be higher in a full-scale application of effluent phytoremediation; in that case, the nutrients supply is higher as it takes place in a continuous flow system or in batch where the surface/volume ratio of ponds is smaller compared to that of the assay.

3.3 Allium cepa genotoxicity test

The A. cepa test allows analysis of the cytotoxic and genotoxic effects of environmental samples on meristematic root cells. Thus, the cytotoxicity is evaluated from the analysis of MI, and genotoxicity from CA and MN frequencies. Both the initial sample of the effluent from the poultry industry and the samples of the three treatments showed a significant higher MI value over the negative control (Table 3). This would indicate the absence of cytotoxic agents. However, an increase in cell division could be dangerous due to the possible proliferation of cells that could result in unordered tumor tissues 5. According to the values of CA frequencies, the initial sample and the three treatments showed no significant differences with respect to the control. However, higher values were observed in MN frequencies in the initial sample and in treatments Lg and Hr, indicating the presence of mutagens. The Lg + Hr treatment was effective in eliminating or reducing the genotoxicity in wastewater, since no chromosomal aberrations or micronuclei were recorded.