Optimising coagulation/flocculation using response surface methodology and application of floc in biofertilisation

2.1 Collection of wastewater and analysis of water quality parameters

The super-intensive shrimp farm in Bachok, Kelantan, Malaysia, provided the samples of aquaculture effluent. The Water Quality Laboratory at the Institute of Tropical Aquaculture and Fisheries (AKUATROP), Universiti Malaysia Terengganu, Malaysia, received these samples after they had been collected in plastic jerricans and transported on ice. The wastewater samples were received and kept at 4°C until further examination. Using the YSI ProDSS handheld multiparameter metre (SKU 626870-1), the initial water quality parameters of temperature, dissolved oxygen, total suspended solids, and pH were measured. A portable turbidity metre (HACH Model 2100P) was used to measure the wastewater’s initial and final turbidity; all the collected data are shown in Table 1. The percentage difference between the turbidity before and after coagulation/flocculation was used to calculate the percentage turbidity removal, as shown in Eq. 1. Using a handheld refractometer (Vee Gee 43036 STX-3), the salinity of the water was determined. A sample of water was placed on a glass plate for the test and the distance the light bent was measured by looking through the end of the droplet. This was carried out three times, and the average values were calculated.

Additionally, reagents, stock solutions, and standard solutions were made for the tests of NO₂, NH₃, and PO₄. A 0.2 mL sulphanilamide solution was introduced to a test tube containing 0 mL of samples, standard solutions, and NO₂ measurements. Then, 1 mL of NED reagent was added and instantly vortexed after 2–10 min. After 2 h, the sample combination was collected and analysed with a Shimadzu spectrophotometer (UV-1800, China) at 543 nm. To test for NH₃, a test tube containing 10 mL of samples and standard solutions was filled with 0.4 mL of phenol solution, 0.4 mL of sodium nitroprusside, and 1 mL of an oxidising solution. After that, parafilm was placed over the test tubes, and they were allowed to cool for an hour at room temperature (20–27°C). After that, sample mixes were examined using a Shimadzu spectrophotometer (UV-1800, China) at 640 nm. Eq. 2 was used to calculate the amount of NH₃ elimination. Similarly, PO₄ was evaluated by adding 1 mL of prepared mixed reagent to 10 mL of samples and standard solutions in a test tube. Using a Shimadzu spectrophotometer (UV-1800, China) and the cadmium reduction method, extinction was recorded at 885 nm after 5 min and ideally within the first 2–3 h.

2.2 Preparation of chitosan coagulant solution

Before initiating the coagulation/flocculation process, daily fresh stocks of chitosan solution were prepared. To produce suitable conditions for the dissolution of chitosan particles, 2 g of powdered chitosan was combined with 1 g of 1% acetic acid and 97 g of distilled water [20].

2.3 Experimental design and optimisation

RSM was used to analyse experimental data using multiple linear regression and linear and quadratic fitting using the commercial programme Statistica. A rotating and spherical design of Box–Behnken design (BBD) used to optimise the number of tests to run during an experiment design [21]. It is a cube whose centre and midway both completely encircle a sphere. The relationship between the independent and dependent variables is often best revealed by a regression model [22]. In this study, we optimised the process parameters for removing turbidity and NH₃ from shrimp aquaculture wastewater using RSM and an incomplete factorial design known as BBD. Chitosan dosage (mg·L⁻¹), particle settling time (min), and pH, which are the independent variables, were each altered at three distinct levels and designated as X ₁, X ₂, and X ₃, as presented in Table 2. Initial tests were conducted to establish the ideal values range for optimisation. The effluent’s pH levels were adjusted before adding the coagulant. This was performed by adding 0.1 M NaOH or 0.1 M HCl as necessary. To plan the experiment and examine the collected data, Statistica 12’s Design of Expert module was used. The results of a total of 15 experimental runs were documented.

2.4 Coagulation/flocculation process

At the Water Quality Laboratory of AKUATROP, Universiti Malaysia Terengganu, flocculation of aquaculture wastewater was conducted using a laboratory-scale jar test apparatus (JLT-6, Velp Scientifica, Italia) with 2 L square jars outfitted with six paddle stirrers. During this test, a coagulant made from naturally occurring chitosan from P. monodon was used. An automated controller was used to mix the sample and wastewater in the test jar quickly and slowly. Following dosing, the test apparatus was set for quick mixing for 2 min at 150 rpm and slow mixing for 30 min at 40 rpm to allow for flocculation. The upper portion of the treated water was taken and tested for ultimate turbidity and NH₃ at the predetermined period of settling (for example, 5, 17.5, and 30 min). At room temperature, the removal efficiencies of NH₃ (%) and turbidity (%) were assessed.

2.5 Fitness of mathematical model

By examining the importance and consistency of each model’s determination ratios (R ²), R ²-adjusted, and lack of fit, the current study investigated the suitability of the regression model. To evaluate the statistical significance of the fitness of the quadratic model as well as the relevance of the various response factors and their interactions, a statistical analysis of variance (ANOVA) was also performed. These investigations made use of the general quadratic equation model, denoted by the following equation:

where yᵢ represents the response, x _i is the input factors, β₀, βᵢᵢ (i = 1,2,…, k),ᵢⱼ (i = 1,2, …, k; j = 1, 2, …, k) are unknown parameters, and ε is a random error.

2.6 Validation of the developed model

The current study tested the validity of the constructed model by running three separate tests with the optimum values found throughout the procedure’s optimisation for turbidity removal and NH₃. Independent student t-tests with a 95% confidence level were used to examine replicate values from the experimental runs [23].

2.7 Nutritional analysis of planting soil and recovered solid

Following the methodology previously used by Saliu and Oladoja [13], nutrient element analysis of the flocs (Figure 1) and the planting soil was carried out at the Institute of Oceanography (INOS) and Faculty of Fisheries and Food Science (FPSM), Universiti Malaysia Terengganu laboratories. Next, a muffle furnace was used to heat a precisely measured crucible containing 0.5 g of the sample. The furnace’s temperature was raised gradually to 500°C and maintained there for 5 h. A 2 mL of concentrated HCl was then added once the crucible and its contents had cooled after heating. The ash and acid were then allowed to evaporate in a boiling pot that was heated to 70°C until dry. Then, 50 mL of 20% HNO₃ was added, and the mixture was heated to 70°C in a water bath for 1 h in a crucible. The mixture was then transferred to a 100 mL volumetric flask and adjusted to the necessary concentration using distilled water. Whatman No. 2 filter paper was used to filter the solution after it had been vigorously shaken. Finally, using an Octupole Reaction System (Agilent HP 7500c) and inductively coupled plasma mass spectrometry, Agilent, United States, analysed the nutrients in the filtrate.

Figure 1

Sample of floc recovered from aquaculture wastewater.

2.8 Experimental procedure and design for plant growth using the floc

At the Universiti Malaysia Terengganu, plant experiments were carried out in a greenhouse with a plastic top and sides covered in nets to control the weather. I. aquatica seeds were bought and transferred to the greenhouse from an agro-allied store in Kuala Terengganu, Malaysia. They were placed inside a peat moss substrate-filled germination tray that was 3.8 cm in diameter and 10 cm deep (Figure 2). In line with Fernández-Delgado et al. [24], germination trays were kept at room temperature, watered twice daily, and also provided adequate access to air and sunlight. After 15 days from the time of planting, three seedlings were transplanted into plastic pots with the following dimensions: 19 cm × 16.8 cm × 26 cm.

Figure 2

Water spinach, Ipomea aquatica seedlings on germination tray at day 7.

According to Table 3, a planting experiment utilising a completely randomised design allowing four treatments with five replications was carried out in the current study. T1 stood for “sand only,” T2 for “sand and floc,” T3 for “sand and liquid fertiliser from a commercial source,” and T4 for “mixed soil.” The sand was placed into each pot, which was then organised according to the various treatments (Figure 3). Each pot contained 1,500 g (1.5 kg) of the growing medium. For usage in this investigation, agro shops sold commercial liquid fertilisers (AB) (Figure 4). The two mixing elements that combine to form the secure chemical liquid fertiliser AB are known by the labels A and B. While iron and calcium nitrate make up the A portion, the B element is made up of potassium nitrate, magnesium sulphate, monopotassium phosphate, manganese sulphate, zinc sulphate, copper sulphate, boric acid, and ammonium molybdate [25].

Figure 3

Newly transplanted Ipomea aquatica at day 15.

Figure 4

Commercial liquid fertiliser.

Prior to application, flocs and liquid fertiliser underwent tests for electrical conductivity (EC), salinity, pH, and total dissolved solids (TDS). To lower the EC to a level that was comparable to that of liquid fertiliser, flocs were 100% diluted first with tap water according to the method employed by Armenta-Bojórquez et al. [26]. Table 4 displays the initial characteristics of the liquid fertiliser and floc during the application process. Twice weekly until harvest, 100 mL of liquid floc and 100 mL of commercial liquid fertiliser were added to plastic containers with the markings T2 and T3. Ten fertilisations were carried out from day 20 to day 55, totalling 1,000 mL of the selected fertiliser in each pot. Plant irrigation was performed twice daily, at 8 a.m. and 5 p.m., and 100 mL of commercial liquid fertiliser was added every third day in line with the manufacturer’s instructions for applying liquid fertiliser to vegetable crops. The experiment took 55 days from seed to harvest (or 7 weeks and 6 days).

2.9 Plant analysis

Only four randomly chosen plants that accurately represented the treatments were sampled out of each treatment’s five replications. After watering the soil and moistening the plants, they were carefully plucked and placed into plastic bags with labels showing the treatments and replications they had received. They were then brought to the Water Quality Laboratory, AKUATROP for a growth performance characteristic examination. The morphological parameters of the plants were noted at both the early stage during transplantation (initial growth parameters) and upon harvesting after 55 days (final growth parameters). The plant height (measured from the plants’ stem bases to their peaks) and the root lengths (measured in cm using a table ruler) were the two morphological growth parameters examined in this study. Other growth indicators were also assessed, including root volume (calculated in cm³), plant fresh and dry mass, and leaf count (individual plant’s total leaves produced). As soon as the plant was removed from the earth, a digital weighing balance was used to determine its full mass. The plants were placed in a petri dish after measuring their fresh mass and were heated to 70°C to achieve complete drying. This finding was consistent with Acharya and Kumar [27]. The plants were weighed once more to determine their dry mass in grams, and the relationship outlined in Eq. 4 was utilised to calculate the percentage change in growth characteristics.

2.10 Statistical analysis

ANOVA, regression, and other analyses of the coagulation/flocculation experiment data were completed using Statistica version 12 and Minitab. The statistical evaluation of plant growth factors and nutrients was completed using SAS version 17. By using a one-way ANOVA, the impact of applying flocs on the development of I. aquatica was assessed. Also, a post hoc Tukey test with a 95% confidence interval was used to identify any significant differences between treatment means.

3.1 Removal of turbidity and NH₃ from aquaculture wastewater using chitosan coagulant

Model prediction uncertainty mainly depends on the sample size and the model coefficients [28]. A small sample size results in high variability and bias. If the sample size is too large, numerical overflow and high costs will occur. Therefore, it is essential to consider the optimal sample size when building reliable statistical models [29]. In this study, the models established using BBD were optimised for sample size, resulting in very low uncertainty bounds. The model coefficients, some of which were determined to be inconsequential, are the other crucial factors that contribute to uncertainty. The generated mathematical models with linear-linear interactions of operating factors produced poor adequacy with very low coefficients of determination and failed to describe the entire experimental data after numerous attempts at fitting the experimental data into regression equations. Eqs. 5 and 6 show the ideal regression model that was identified from the experimental data after several candidate models had been tested: a quadratic model with the linear interaction of operating parameters.

A model must have a coefficient of determination of at least 80%, a lack of fit that is not statistically significant, and a good R ²-adjusted (R ²) before it can be considered sufficient [30]. The models in the present study were sufficient, as shown by the turbidity and NH₃ removal R ² values and R ²-adjusted values, as well as their lack of fit as indicated in Tables 5 and 6. This further suggests that the reduction of turbidity and NH₃ from the wastewater was significantly impacted by the process variables and levels taken together. According to Ghafari et al. [31], a high value of R ² that should be close to 100% and in close range with the R ²-adjusted value is desired and necessary for a model to be appropriate. The statement went on to say that a high R ² ensures a good modification of the model data. High R ² also demonstrates that the chosen model adequately explained the majority of the overall variation in the responses received [28]. Table 5 demonstrates that there was no discernible interaction between the variable components. This indicates that changing one did not change the other. Table 6 shows that the significant effects of independent components were linear, quadratic, and interactive. Particularly, the only significant linear factor was pH. Furthermore, the dosage-dependent quadratic effect of chitosan was insignificant; nevertheless, its combination with pH had a substantial impact on the elimination of NH₃. Only 4.55% of the elimination percentage could not be explained by the model, according to the model’s R ² of 95.45% (Table 7).

where X ₁ = A = dosage, X ₂ = B = pH, and X ₃ = C = settling time.

Note: only terms with * in the equation are significant.

where X ₁ = A = dosage, X ₂ = B = pH, and X ₃ = C = settling time.

Note: only terms with * in the equation are significant.

The present study’s maximum turbidity removal (92.16%) occurred at a chitosan dose of 16.25 mg·L⁻¹, a pH of 7.5, and a settling period of 17.5 min (Figure 5). This was comparable to the earlier study on stabilised leachate (94.0%) and textile industrial wastewater (94.9%) using chitosan as the flocculant and coagulant [31]. Other investigations utilising different coagulants or flocculants have also been carried out [12]. In the present study, it was discovered that removal effectiveness declined as one moved away from the process parameters’ optimum values. This indicates that the reaction declines as any of the examined factors rise or fall. From the aforementioned, it should be clear that the established turbidity removal model equation can be utilised to calculate the specified percentage of removal without the non-significant terms. Using a Pareto chart, how each impact affects a response can be seen as how important each impact is [30]. In the present investigation, dose and settling time had less of an impact on turbidity removal than did the linear effect of pH (Figure 6). The fitted surface plot (Figure 7) further demonstrated that a higher pH will result in more turbidity removal.

Figure 5

Predicted and desirable % turbidity removal.

Figure 6

Pareto chart for standardised effects of the independent factors and interactions for % turbidity removal.

Figure 7

Fitted surface plots of % turbidity removal in 3D.

According to Xiao et al. [32], the two main components of aquaculture wastewater are ammonia (NH₃) and nitrite (NO₂), which are metabolic byproducts of leftover feed and fish faeces. Aquaculture wastewater has been shown to contain up to 103.7 mg·L⁻¹ of NH₃ at a pH of 8.1 and a salinity of 28.6 ppt [33]. According to Wang et al. [34], NH₃ from aquaculture wastewater is naturally removed from mangrove wetlands at a rate of 65%. They also noted that pH, temperature, amount of dilution, and salinity have a significant impact on this rate of removal. This investigation shows that, at a dosage of 5 mg·L⁻¹ and a pH of 7.5, chitosan effectively eliminated 84.42% NH₃ from aquaculture effluent after 30 min of settling (Figure 8). Chitosan has been shown in related research to be capable of eliminating and absorbing up to 7.6 mg·g⁻¹ of NH₃ after 180 min at pH 8. Additionally, it showed that clearance rates rise as pH rises [35]. The results of this investigation corroborated those of Alvarez and Otero [36], who showed that chitosan effectively removed 84.9% of NH₃ from the effluent of marine food canning businesses. Additionally, Kumar et al. [37] could only remove 49.1% of NH₃ from shrimp cultivation effluent, which is a lower percentage than the current study. The Pareto chart’s findings also demonstrated that pH had a dominant impact on NH₃ elimination in this investigation (Figure 9). Response surface plots, which make the relationship between the experimental data and the response simple to visualise, were used to display the constructed model graphically [38]. So, for pH and settling time, the optimal ranges for NH₃ removal were 7.0–7.6 and 26–30 min, respectively (Figure 10).

Figure 8

Predicted and desirable % NH₃ removal.

Figure 9

Pareto chart for standardised effects of the independent factors and interactions for % NH₃ removal.

Figure 10

3D fitted surface plots of NH₃ percentage removal.

3.2 Validation of the adequacy of predicted models

The validation experiments using the optimised conditions were carried out in triplicate to minimise error. As shown in Table 8, the experimentally derived and model-estimated removal efficiencies were in close agreement for all response parameters. These results demonstrated the validity of the reaction model. Moreover, the errors for these validation runs ranged from 0% to 4.0% and are in good agreement with the predicted values with less than 10% error. This submission concurs with Priyatharishini and Mokhtar [39], who said that where the experiment involves many numbers of variables, a percentage error of less than 10% is typically acceptable. Additionally, experimental data must be contained within the model’s 95% prediction interval. In the validation experiment, this was likewise met. The resulting model response predictions were regarded as being of high quality because the discrepancies between the projected and experimental results fell within an acceptable range.

3.3 Nutrient content of recovered floc

The amount of nutrients in the recovered solid from aquaculture wastewater as well as the growing media were tested to ascertain their suitability for plant growth. Generally, the use of recovered solids from wastewater as organic fertiliser and soil conditions has been confirmed to vary with the composition of substances being applied and the pH, with recommended ranges between 6.5 and 7.5 [13]. Following this recommendation, the pH value of the floc (6.91) in the present study was considered suitable for the application. Fish excretion and the rotting of meal residues in the water have both been linked to pollutants in aquaculture wastewater. The macronutrients (C, Ca, H, K, Mg, N, O, P, and S) and micronutrients (B, Cl, Co, Cu, Fe, Mn, Mo, Ni, Se, Si, and Zn) required for plant growth are abundant in these pollutants, which may be collected in the form of suspended particles [40]. Additionally, Al, As, Ba, Bi, Cd, Cs, and Pb present in wastewater-suspended particles are implicated in having a harmful effect even at minuscule amounts of less than 100 ppm, according to Akter et al. [41].

Macronutrients are crucial components that plants need in considerable amounts (over 10 moles per kg dry matter) to promote healthy growth and development. In contrast, lower levels of micronutrients (below 10 moles/kg dry matter) are required for a number of essential plant processes [16]. The main nutrients found in this study were Ca, K, Mg, N, and P, but minor amounts of micronutrients like Co, Cu, Fe, Mn, Ni, Se, and Zn were also present. Additionally, the presence of trace amounts of As, Al, Cd, Cr, and Pb was also noted. N is a key nutrient for plants and is essential for their growth, somatic embryogenesis, chlorophyll synthesis, photosynthesis, cell division/differentiation, and electron transport [42]. According to Effendi et al. [43], P is mostly linked to improving water usage efficiency, encouraging leaf expansion, and increasing axillary bud and shoot canopy development. As opposed to this, K plays a major role in protein synthesis, water balance, and plant growth [44]. For sand, floc, and mixed soil, the study reported N values of 0.01%, 5.94%, and 0.09%, respectively. The flocs are suitable as growth enhancers because of their comparatively high N content [44]. According to Kala et al. [45], an organic compound needs to have a N content of between 5% and 10% to qualify as fertiliser. According to the results of the P content analysis, floc (41,895 mg·kg⁻¹), mixed sand (2,486 mg·kg⁻¹), and soil (4,810 mg·kg⁻¹) had the highest P concentrations. The reported P content of 54 mg·kg⁻¹ in silty loam sand by Bender and van der Heijden and in farmyard manure by Dutta and Banik [44] was exceeded by these values. The excess excrement and feed from the cultivated shrimps, which contributed to the wastewater’s composition, may be blamed for the floc’s high P content (Alnawajha et al., [40]). According to Jana et al. [46] and Dutta and Banik [44], normal wastewater K content values fall within 1.5–3.0% and 10–40 mg·L⁻¹ (K). To encourage better crop growth, Bender and van der Heijden [47] suggested a K level of 391 mg·kg⁻¹ in organic fertilisers. In this investigation, the mixed soil (4816.3 mg·kg⁻¹) had the highest K values, followed by the floc (4,424.5 mg·kg⁻¹) and the sand (761.11 mg·kg⁻¹).

According to the study, the sources had falling Ca levels, with concentrations of 18,128 mg·kg⁻¹ for floc, 17,461 mg·kg⁻¹ for mixed soil, and 7,834.7 mg·kg⁻¹ for sand, respectively. Other types of shell deposits that were present in the wastewater and likely settled next to the suspended materials during coagulation and flocculation may be to blame for the higher Ca levels in the floc. It is well known that the bases of Mg and Ca are frequently present in soil as oxides and carbonate hydroxides. Their availability and use in organic fertilisers contribute to improving plant nutrient accessibility and reducing soil acidity [45]. The findings reported by Bender and van der Heijden [47] are consistent with the high Ca content seen in this investigation. However, compared to Ca, Mg concentrations in the soil and floc were noticeably lower. The highest Mg values were found in the floc (6,798.9 mg·kg⁻¹), then in the mixed soil (1,747.8 mg·kg⁻¹), and lastly in the sand (370.85 mg·kg⁻¹). The seawater used for shrimp culture, from which the flocs were derived, is where the magnesium in the flocs originated. It has been demonstrated that elements like pH, EC, and soil organic matter have a significant impact on the dynamics and transformation of micronutrients in the soil. In plant nutrition, their role, which primarily entails acting as co-factors of several enzymes responsible for the metabolism of organic molecules (carbohydrates, nucleic acids, proteins, and lipids), is essential [48].

According to Wang et al. [49], organic fertilisers contain trace metals such as Pb (3.5 mg·kg⁻¹), Cd (0.3 mg·kg⁻¹), and Cr (55.9 mg·kg⁻¹), as well as micronutrients like Cu (2.9 mg·kg⁻¹), Zn (2.8 mg·kg⁻¹), Fe (16.3 mg·kg⁻¹), and Mn (6.2 mg·kg⁻¹). The essential micronutrients for healthy plant development and growth, according to Kaur et al. [42], are Fe, Zn, Mn, Cu, B, Cl, Mo, and Ni. In the current study, the majority of the trace elements and micronutrients (Fe, Cd, Pb, Al, Cr, and Co) were higher in the sand than in the floc and mixed soil, except for Zn, Ca, Cu, Mn, and As, which were higher in the floc. while Se could only be found in mixed soil. The levels of micronutrients and trace elements found in this study were greater than those suggested for organic fertilisers by Wang et al. [49] and Dhaliwal et al. [48].

In general, the floc’s composition under ideal conditions compared favourably with the requirements of the European Parliament and Council’s Regulation (EU) 2019/1009 on CE-marked fertiliser goods (European Commission, 2019). The substance must satisfy at least one of the three minimum nutritional concentrations (2% TN, of which 0.5% is organic nitrogen, 2% P as P₂O₅, and/or 2% w/w K as potassium oxide (K₂O), according to the EU Regulation). Additionally, the level of pollution cannot exceed the restrictions outlined in the EU Regulation. Quantities of 0.4 mg Cd per kg, 0.02 mg Hg per kg, 9.2 mg Ni per kg, and 6.6 mg Pb per kg were recommended within the confines of the EU Regulation. Heavy metals, including Ni, Cd, Cr, and Pb, are often best absorbed under acidic or neutral (pH 7) conditions. In the current investigation, floc had decreased levels of Pb and Ni but no Hg or Cd. In an alkaline medium, plants are unable to absorb these metals. As a result, I. aquatica tissues may not have absorbed or bioaccumulated Pb and Ni due to the floc’s somewhat basic pH. Due to its European Conformity (CE) certification, the floc developed in this study might be considered fertiliser [24]. The fact that the nutrient element N is the most limited in marine water may account for N’s typically low level in the floc found in this current investigation. This is consistent with Effendi and Giri [50]. The sludge from aquaculture wastewater, according to Fu et al. [51], contains up to 7–32% of the total nitrogen and 30–84% of the total phosphorus. In addition, P was found in the floc at the highest level compared to every other nutrient element in this investigation.

This is because effluent from aquaculture contains more sources of the element, such as uneaten feed and decomposing organic material. The amount of P found in aquaculture wastewater as measured in this study is consistent with those of Delaide et al. [19], who found that sludge had a greater content of P. However, according to Goddek and Körner [52], among other macronutrients, aquaculture sludge contains 6% nitrogen, 18% phosphorus, 6% potassium, 16% calcium, and 89% magnesium. In addition (based on dry matter), 24% iron, 86% manganese, 47% zinc, and 22% copper were reported for micronutrients. The management strategies used during the fish production period may have had an impact on the variability in the nutrient levels in floc obtained in the current study and earlier investigations (Table 9).

3.4 Growth analysis of I. aquatica

I. aquatica, a free-floating tropical herb commonly found growing in marshy areas and canals, is thought to be subaquatic [17]. According to Jampeetong et al. [18], it is renowned for its high tolerance and capacity to both use and eliminate significant amounts of Cu, Cd, N, and P from soil and water. This action therefore prevents Cu toxicity in the soil by absorbing and storing excess Cu in the plant [53]. Generally, sandy soil is considered poor in terms of plant nutrients due to the high level of leaching associated with it. Therefore, the addition of flocs to the sand in this study may have been an efficient strategy for enhancing the nutrient availability of the mix for improved growth of I. aquatica.

In terms of root volume and length, plant height, leaf count, fresh and dry mass, floc treatment, or no floc treatment units, the plant yield was measured at 55 days. Figure 11 depicts how the plants look after being harvested. Final growth parameter results showed that higher values were obtained from plants treated with flocs, while the lowest values were recorded from experimental units with sand only (Table 10). The highest percentage differences were also recorded in plants treated with floc. The obvious fact of better mineralisation of nutrients in the flocs from the aquaculture wastewater may have been responsible for the advantageous growth and yield of I. aquatica grown on them. According to Acharya and Kumar [27], the growth of Chinese garlic plants and the number of leaves did not significantly increase after the application of organic fertiliser made from chicken dung containing 3.03% N. In disagreement, this study recorded the highest percentage change in plant height from plants treated with floc (78.67%), followed by sand-commercial liquid fertiliser mix (74.95%), and the least obtained from sand medium only (63.41%). Similarly, differences in leaf count were also observed to be highest in floc and least in plants treated with only sand (55.26%) (Figure 12).

Figure 11

I. aquatica after five treatments at harvest (55 days).

Figure 12

Percentage change in I. aquatica’s growth metrics at day 55.

T1 plants displayed nutritional deficiencies as most plants have stunted growth. Commercial liquid fertiliser-treated T2 plants showed good growth performance with a modest yellowing of the leaves and lower numbers. The T3 (applied floc) plants had more leaves and a longer stem, but the leaves had a minor yellowing. Plants grown in T4 (mixed soil) were short and yellowish. Plants showed a higher growth response in the shoots as compared to the roots. This is evident from the display of higher histogram bars for the plant height.

According to Thomas et al. [54], while applying organic manure, the plant’s production cycle length is frequently taken into consideration, especially for crops with short lifespans like vegetables. The I. aquatica belongs to the group of plants with a brief life cycle. Previous research has shown that harvesting I. aquatica after 30 days of organic fertiliser administration led to greater percentage growth in shoots than in roots, with an impressive 50% increase in plant weight [55]. These results are consistent with those of the current study, which likewise noted greater variations in percentage growth for leaf count and plant height as opposed to root length and root volume. Another study by Zhao et al. [56] found that using biochar as an organic fertiliser as opposed to NPK inorganic fertiliser resulted in significant variations with regard to the growth performance of I. aquatica after 33 days. Likewise, Ruhunuge et al. [57] provided evidence in favour of the hypothesis that organic fertilisers increased soil fertility through nutrient recycling and mineralisation, resulting in early yields of I. aquatica within 30–45 days after sowing. Although the specific reaction of I. aquatica to various types of N has not yet been fully determined [18], this crucial macronutrient can be adjudged as the primary factor that makes plants treated with floc perform better, which includes a larger percentage of N.

The efficiency of nutrient absorption by plants and eventually their production is often determined by the EC in the soil, which may be changed by the application of organic fertiliser [48]. Therefore, the EC could be decreased to a level that immature I. aquatica plants could tolerate by diluting the initial floc obtained for this investigation by 100%. Additionally, the high TDS level that had been previously found in the flocs was a sign of a high level of plant nutrients [58]. Although nutrients in suspended solids in wastewater are present in high concentrations, Feria-Díaz et al. [59] found that, unless they are first solubilised, they might not be easily available for plant uptake when applied as organic manure. Though nutrient solubilisation was not done in the current investigation, I. aquatica performed noticeably better than the control treatment. This implies that the macro- and micronutrients found in the floc were absorbed and effectively used for the observed plant growth performances.