Recycling of domestic wastewater treated by vertical-flow wetlands for watering of vegetables

Constructed treatment wetlands are engineered wastewater purification systems that encompass biological, chemical and physical processes, which are all similar to processes occurring in natural treatment wetlands. They are implemented for environmental pollution control to treat a variety of wastewaters including industrial effluents, urban and agricultural runoff, animal wastewaters, sludge and mine drainage (Scholz 2010; Sani et al. 2013). Recently, some large-scale wetland systems have also been successfully applied to treat domestic wastewater (Dong et al. 2011). However, there are few long-term and controlled studies involving domestic wastewater due to health and safety concerns.

The earliest documented sewage farms, where wastewater has been applied to land for disposal and for agricultural use, were operated in the 16th and 17th centuries in Bunzlau, Germany, and Edinburgh, Scotland (Shuval et al. 1986). The scientific basis for the acceptance of wastewater reclamation, recycling and reuse in agriculture has evolved from developments in water and wastewater engineering science coupled with an increasing pressure on water resources management. The evaluation of the effects of treated wastewater reuse on crops intended for human consumption is of particular interest (Asano & Levine 1996; Aiello et al. 2007; Cirelli et al. 2012). Unfavourable concentrations of certain nutrients and trace elements are a challenge to the growth of plants fed by recycled pre-treated wastewater. Moreover, traces of hydrocarbons from diesel spills associated with urban runoff or industrial effluent are a more recent challenge (Scholz 2010; García-Delgado et al. 2012).

Cirelli et al. (2012) presented the results of a reuse scenario where tertiary-treated municipal wastewater using a constructed wetland was supplied for irrigation of vegetables in Eastern Sicily, Italy. They found elevated levels of Escherichia coli in the irrigation water, which were frequently above the Italian limits of 50 colony forming units (CFU)/100 ml for secondary urban effluents.

García-Delgado et al. (2012) undertook a greenhouse study in Spain to assess the effect of treated urban waste water on soil and pepper quality. The wastewater application saved fertiliser (37% nitrogen, 66% phosphorus and 12% potassium). Total polyaromatic hydrocarbons and the heavy metals cadmium, lead and arsenic within the pepper fruits were low.

Morari & Giardini (2009) assessed the treatment efficiency of pilot-scale vertical-flow constructed wetlands on municipal wastewaters and their suitability for irrigation reuse in Italy. Only water quality parameters with high removal efficiencies fulfilled the Italian guidelines for irrigation reuse, whereas parameters with lower efficiencies such as suspended solids (SS) and total phosphorus limited the potential water reuse.

Vertical-flow constructed wetlands treating septic tank effluent in Guangzhou (China) achieved removal rates for chemical oxygen demand (COD), five-day biochemical oxygen demand (BOD), SS, total nitrogen and total phosphorus of 60%, 80%, 74%, 49% and 79%, respectively (Cui et al. 2003). After that the treated effluent was used for hydroponic cultivation of water spinach and romaine lettuce. It was found that using treated effluent for hydroponic cultivation of vegetables could reduce the nitrate content in vegetables.

Lopez et al. (2006) assessed constructed wetlands treating municipal effluents to be reused in agriculture. Recorded average removal efficiencies for SS, five-day BOD, COD, total nitrogen and total phosphorus were 85%, 65%, 75%, 42% and 32%, respectively.

Many vegetables have the potential to grow on recycled wastewater. However, there is the potential of some vegetables such as lettuce and cabbage to become contaminated by microbes, because their edible leaves are too close to the ground receiving the treated wastewater. Therefore, it makes sense to select vegetables where the edible fruit is located far away from the ground (Nickels 2012). This may include peppers, tomatoes, maize, eggplants, beans, lentils and peas. The next step in selecting suitable vegetables is to decide on easy-to-grow and relatively cost-effective plants with high nutritional value. However, the authors concentrated on two pepper types in this study, because they can also be grown in greenhouses in the UK (Nickels 2012; Jones 2013).

Jones (2013) discussed the positive and negative impacts on key nutrients and minerals on plants. The major minerals impacting on the growth of plants are nitrogen (predominantly ammonium), phosphorus, potassium, calcium, magnesium and sulphur. Micro-nutrients that are beneficial in small amounts are (in no particular order) boron, copper, manganese, molybdenum, potassium, iron, zinc, iodine and chlorine. Copper, manganese, molybdenum, iron, zinc and aluminium (in no particular order) are often described as heavy metals. They may be toxic in high inflow water concentrations for peppers. Under acid conditions (soil with a pH of less than 7), heavy metals could be a problem to sensitive plants (FAO 2003).

Pescod (1992) and FAO (2003) also recommended limits for trace minerals in reclaimed water use for irrigation. Long-term (for water used continuously on all soils) and short-term (for water used for a period of up to 20 years on fine-textured neutral alkaline soils) threshold values for 18 elements have been listed. Recommended maximum concentrations for the trace elements that are often exceeded (see results below) are 5.0 mg/l for iron, 0.2 mg/l for manganese and 2.0 mg/l for potassium.

FAO (1994) classified the suitability of treated wastewater for recycling in terms of pollutants. Acceptable ranges for ammonia-nitrogen, ortho-phosphate-phosphorous and potassium were between 0 mg/l and 5 mg/l, between 0 mg/l and 2 mg/l, and between 0 mg/l and 2 mg/l, respectively. Furthermore, Pescod (1992) stated that for irrigation water there is no restriction for its reuse if nitrate-nitrogen values are <5.0 mg/l. Slight to moderate constraints exist for the range between 5 and 30 mg/l. Severe recycling restrictions are usually imposed for values of more than 30 mg/l.

The overall aim is to assess if vegetables can be grown successfully on recycled domestic wastewater treated by constructed wetlands. The corresponding key objectives related to the growing of Sweet Pepper and Chilli are to assess (a) the suitability for growth when using recycled wastewater, (b) the impact of different treated wastewaters as a function of the wetland type, (c) the volume of treated wastewater for irrigation, (d) the suitability of different growth media, (e) the effect of a diesel oil spill on the suitability of the recycled wastewater, and (f) the economic viability of different experimental set-ups.

The vertical-flow wetland system is located within a greenhouse (door left open) on top of the roof of the Newton Building, The University of Salford, Greater Manchester, UK (Sani et al. 2013). Wetland filters were operated between 27 June 2011 and 4 June 2014. The set-up includes two filters that are essentially controls receiving clean de-chlorinated tap water. Table 1 indicates an overview of the statistical experimental set-up used to test the impact of four variables. Filters 1 and 2 compared to Filters 3 and 4 test the influence of a larger aggregate diameter. Filters 5 and 6 compared to Filters 3 and 4 check the impact of a higher loading rate. The application of a lower contact rate is tested if Filter 7 is compared with Filters 3 and 4. Finally, a lower resting time is the difference between Filters 7 and 8.

The 10 laboratory-scale wetland filters were constructed from Pyrex tubes with an inner diameter of 19.5 cm and a height of 120 cm. The filters were filled with siliceous (minimum of 30%) pea gravel up to a depth of 60 cm and planted with Phragmites australis (Cav.) Trin. ex Steud. (Common Reed). Dead macrophyte plant material was harvested in winter and returned to the corresponding wetland filters by placing it on top of the litter zone (Sani et al. 2013; Al-Isawi et al. 2015a).

The main outlet valve is located at the bottom of each constructed wetland system. Eight further valves (used to test for clogging) are located on the sidewall of each wetland column. The sidewall valves were located at heights of 10, 20, 30, 40, 45, 50, 55 and 60 cm from the bottom of each column (Sani et al. 2013).

Wetland columns received 6.5 l of inflow water during the feeding mode, which was different between several filters (Table 1). Columns 1– 6 were sampled after 72 h contact time and then left to rest for 48 h, while columns 7 and 8 were sampled after 36 h contact time and left to rest for 48 h and 24 h, respectively. All water quality parameters discussed in this paper were determined during or directly after sampling. The preliminary treated urban wastewater used for the inflow water was obtained from the Davyhulme Sewage works, one of the largest waste water treatment plants in Europe (http://en.wikipedia.org/wiki/Davyhulme), operated by the water utility company United Utilities in Greater Manchester. Fresh wastewater was collected approximately once per week, and was stored and aerated by standard aquarium air pumps in a cold room before use. The wastewater quality was highly variable, and comprised domestic and industrial wastewater as well as surface water runoff.

To simulate a one-off Diesel fuel (100% pure; no additives) spill, 130 g (equivalent to an inflow concentration of 20 g/l) of diesel fuel (100% pure; no additives) were poured into Filters 1, 3 and 5, and into one of the two columns (Control A) on 26 September 2013 (Table 1) as discussed by Al-Isawi et al. (2015b). The fuel was obtained from a petrol station operated by Tesco Extra (Pendleton Way, Salford, UK).

Aqua Medic Titan chillers (Aquacadabra, Bexleyheath, UK) were used to maintain the root system and debris layer of all wetland systems at semi-natural below-surface temperatures of about 12 °C. This temperature simulates the temperature of the upper earth layer where the root system of the wetland plants of a real treatment system would be located (Sani et al. 2013).

The COD was used as an indicator to differentiate between low and high loads (Table 1). An inflow target COD of about 285 mg/l (usually between 122 and 620 mg/l) was set for wetlands with a high loading rate (Filters 5 and 6). The remaining Filters 1, 2, 3, 4, 7 and 8 received wastewater diluted with de-chlorinated tap water. The target inflow COD for these filters was approximately 138 mg/l (usually between 43 and 350 mg/l).

Routine water quality sampling was carried out according to APHA (2005), unless stated otherwise, to monitor long-term and seasonal treatment performance. The spectrophotometer DR 2800 Hach Lange (www.hach.com) was used for standard water quality analysis for variables including COD, ammonia-nitrogen, nitrate-nitrogen, ortho-phosphate-phosphorus and SS.

Total petroleum hydrocarbons (TPH) were determined by gas chromatography and flame ionization externally by Exova Health Sciences (Montrose Ave, Hillington Park, Glasgow, UK) according to their own TPH in Waters (with Aliphatic/Aromatic Splitting) Method (Exova Health Sciences 2014), which is accredited to the British Standard (BS) method BS EN ISO IEC 17025 by the United Kingdom Accreditation Service and compatible to the International Organization for Standardization (ISO) standards (e.g., ISO17025), BS method BS DD 220 1994, and American Standard methods (United States Environmental Protection Agency (US EPA) Method 3510C and US EPA SW846 Method 8015).

The five-day BOD was determined in all water samples with the OxiTop IS 12-6 system, a manometric measurement device, supplied by the Wissenschaftlich-Technische Werkstätten (WTW), Weilheim, Germany. Nitrification was suppressed by adding 0.05 ml of 5 g/l N-Allylthiourea (WTW chemical solution No. NTH600) solution per 50 ml of sample water.

The third planting (i.e., second and final replanting) took place 28 days after the first replanting on 8 November 2013. Table 2 outlines the experimental set-up. The Sweet Pepper and Chilli were planted individually into 10-litre round plastic plant pots provided by scotplants (Hedgehogs Nursery, Crompton Road, Glenrothes, UK). The plant pot dimensions were as follows: height of 22.0 cm, bottom diameter of 22.0 cm and top diameter of 28.5 cm. All plant pots remained indoors (laboratory) under controlled environmental conditions. The experimental set-up was chosen to allow for the statistical assessment of different types of treatment such as the impact of organic media and nutrients in the wastewater.

Microsoft Excel (www.microsoft.com) was used for general data analysis. IBM SPSS Statistics Version 20 (www.ibm.com) was applied to calculate the correlation between variables and statistical differences between treatments.

Findings are shown in Tables 3,⁴⁵–6. Table 3 outlines the inflow water quality received by the plants. Note that the wetland effluent was used as the influent for the vegetable pots. Highly fluctuating values for TPH were observed in the outflow waters obtained from Filters 1, 3, 5, and 8, and Control A. The TPH concentrations followed this order: Control A > Filter 8 > Filter 1 > Filter 3 > Filter 5. However, these TPH concentrations were in compliance with the threshold set by the Chinese standard for irrigation water quality (SEPA 2005) highlighting a maximum allowable value of 1 mg/l. Note that the Chinese standard was used, considering that China produces about 54% (estimated in 2008) of the Sweet Peppers in the world (ERS/USDA 2008).

COD values were the highest for raw wastewater followed by those for Filter 5, which was fed with a high inflow load in terms of COD (Table 1). In contrast, the lowest values were noted for Control B. Filters 1, 3 and 8 had relatively similar COD concentrations. Control A had higher COD values than Filter 6. No differences in COD values were noted for Filters 2, 4 and 7. This observation helps to explain why the inflow load of wetland filters directly impacts on the COD values of the outflow water rather than design variables such as aggregate size, contact time and resting time. High rate filters are likely to be overloaded as discussed by Sani et al. (2013).

The five-day BOD was high for raw wastewater followed by wastewater samples, which were diluted with 80% of tap water. In comparison, the lowest five-day BOD was observed for tap water with fertiliser, tap water and deionized water.

High concentrations of ammonia-nitrogen, which exceeded the threshold of 5 mg/l (FAO 1994), were noted for both Filters 5 and 6 (filters with a high loads; Table 1), followed by those for Filters 1 and 2 (filters with large aggregate sizes). This confirms results findings by Sani et al. (2013) that high rate filters tend to overload. Table 3 shows that nitrate-nitrogen for all filter outflow waters was less than the maximum thresholds value of 30 mg/l (Pescod 1992).

Based on the recommended threshold of 2 mg/l for ortho-phosphate-phosphorus (FAO 1994), the outflow waters from all wetland filters (with the exception of Controls A and B) were associated with too high ortho-phosphate-phosphorus concentrations. In general, phosphorus is one of the most difficult pollutants to be removed by mature constructed wetlands (Pant et al. 2001). This can be explained by the fact that phosphorus is usually present in particulate form, and does not dissolve well in filters that are not yet saturated by phosphorus or other compounds competing for adsorption sites (Scholz 2006, 2010).

The highest value for SS was noted for raw wastewater followed by that for wastewater, which was diluted with 80% tap water. In contrast, the lowest values were observed for Filter 7 outflow water and tap water with fertiliser. Turbidity was high for raw wastewater. Filter 7 had the lowest turbidity values. The pH values for all filter outflows were within the normal range; i.e., between 6.0 and 8.5 (Pescod 1992).

The statistical experimental set-up as specified in Table 2 was chosen for the second replanting stage. Table 4 shows visual growth problems and indicates possible reasons. Figures 1(a) and 1(b) provide an overview of total irrigation water volume consumed by Sweat Pepper and Chilli plants, respectively, growing in different media. All plants grown in organic media consumed more water than those grown in inorganic media leading to better overall plant development.

In countries, where wastewater is seen as a resource, low wastewater consumption by plants is an advantage. However, high wastewater use by plants is seen as an advantage in temperate regions. The productivity of plants in terms of harvest is, however, independent of the wastewater consumption (Figure 1(b)). Nevertheless, a higher foliage production requires more water.

Correlation analysis results show that the fruit weights were significantly positively correlated with the total water volume used for irrigation (r = 0.821, p =0.000). Potential water stress might have reduced cell division and caused cell enlargement to cease. This could have led to a slowdown of the growth rate and might have been the reason for the relatively low weight, diameter and length of Pepper fruits (Tadesse 1997).

Table 7 provides summaries of the bud, flower and fruit development for Sweet Pepper and Chilli plants. The overall growth development of Sweet Peppers was rather disappointing, possibly due to the high concentrations of nutrients and trace minerals, and adverse environmental boundary conditions in the laboratory (Jones 2013).

Chilies did reasonably well but the growth of foliage was excessive and the harvest was delayed. High numbers of buds were recorded for plants growing in an organic media compared to those growing in an inorganic media. However, most of the buds associated with the organic media fell down before reaching the flowering stage. Most flowers either also fell down or died before producing any fruits, possibly due to the high nutrient concentrations (especially nitrogen) supplied to those plants by rich organic media and irrigation (pre-treated) wastewater (Haifa Chemicals 2014). Plants growing in an inorganic media produced lower numbers of buds compared to those Chillies grown in organic media. Nevertheless, most of these buds successfully reached the flowering and fruiting stages. This is possibly due to a better balance in nutrients supplied to those plants by the corresponding irrigation water. Low fruit numbers correlate well with inorganic growth media (Table 7).

Figures 2,³–4 summarize a comparison of maximum fruit lengths, widths and total weights harvested from plants grown in different media subjected to different irrigation water types. Generally, fruits harvested from plants grown in organic media had diameters, lengths and weights greater than those from plants raised in inorganic media. These results in addition to findings based on other research studies undertaken in greenhouse conditions to assess the effect of different growth media on Pepper growth rates and yields indicated that seedlings benefited from peat moss media (Rahimi et al. 2013).

Another study was undertaken to determine the effects of peat and sand on variables such as fruit length, diameter and weight, as well as the total fruit number per plant and yield. Results showed that peat significantly increased length, diameter and weight of fruits in all cultivars grown in comparison to sand (Gungor & Yildirim 2013).

Furthermore, organic substrates decompose over time, and subsequently release nutrients. The rate of decomposition and the physical conditions of the media vary with the parent material. That in turn will enhance crop growth and development. Moreover, better aeration of peat promotes vigorous root growth, which allows rapid development of foliage and therefore increases the whole plant yield (Olle et al. 2012). In contrast, for inorganic media such as sand, nutrient provision to the crops is limited to the nutrients that are part of the irrigation water resulting in a delay of plant foliage growth with a subsequent poor yield (Olle et al. 2012).

Figure 5 summarizes differences in fruit mean diameter, length and weight harvested from plants irrigated with different water types and grown in organic media. Regarding the comparison of mean fruit lengths and widths of plants grown in organic media and irrigated with different wetland outflow waters, statistical results show that harvested fruits were not significant different from each other (P value greater than 0.05). However, fruits harvested from plants grown in organic media and irrigated with Filter 2 outflow water were very close to those harvested from Control B outflow water, tap water and deionized water in which plants depended mainly on nutrients provided by the organic growth media, confirming that nutrients provided to plants by the treated wastewater and the compost were very high, produce good yield quality.

Regarding a comparison of mean fruit weights harvested from plants grown in organic media and irrigated with different water types (Figure 5), results showed that plants irrigated with water harvested from Filters 2 and 4 produced fruits of mean weight, which were significantly greater than those harvested from plants irrigated with outflow water from Filter 7 (P value of 0.008 and 0.027, respectively). These result can be explained by a higher contact time (Filters 2 and 4 compared to Filter 7; Table 1) leading to more favourable nutrient quantity distributions (e.g., lower nitrate-nitrogen in the irrigation water for Filters 2 and 4 compared to Filter 7; Table 3) provided by Filters 2 and 4, which subsequently positively impacted on fruit diameters, lengths and weights (Bar-Tal et al. 2001). However, since the plants irrigated with tap water and grown in organic media produced the highest number of fruits (23), this explains why the total weight of harvested fruits was linked to plants irrigated with tap water associated with low nutrient loads (Figure 4).

Findings indicate that nutrient concentrations supplied to the Chillies by a combination of compost and treated waste water are usually too high to produce a good harvest. However, as the compost is depleted of nutrients after about 8 months, the harvest increased for pots that received pre-treated wastewater in comparison to those pots depending only on the nutrients associated with the compost. A high yield is rather related with the most suitable provision of nutrients and trace elements.

Table 5 provides an overview of the ICP–OES analysis for selected elements in the irrigation water. High concentrations of iron, which exceeded the threshold of 5 mg/l (Pescod 1992; FAO 2003), were noted for both raw wastewater and tap water. Based on the recommended threshold of 2 mg/l for potassium (FAO 1994), the outflow water from all wetland filters (with the exception of Controls A and B) was associated with too high potassium concentrations. Furthermore, high concentrations were also observed for raw wastewater, and wastewater samples, which were diluted with 80% of tap water. Results show for Filters 3 and 5 relatively high manganese concentrations, which exceeded the threshold of 0.2 mg/l (Pescod 1992; FAO 2003).

Table 6 shows the proposed novel harvest classification scheme for Chilli. However, the estimated prices are dependent on the current market value. Only the following numerical and objective variables were used to classify fruits for the purpose of this study: length, width, weight and bending. The lowest variable class entry for any individual fruit assessment determined the final class. For example, if a fruit is categorized as class A in terms of length, class B in terms of diameter and E in terms of weight, then the final class for this fruit would be class E, and accordingly the corresponding price for this Chilli would be zero pence (Table 8).

The monetary value of the harvest was only calculated for the example plant. A similar assessment could also be undertaken for Sweet Peppers at the end of the harvest. Table 7 showed that the highest number of fruits categorized as Class A were harvested from plants grown in organic media and watered with tap water. However, tap water was also associated with the highest fruit numbers categorized as Class E. The highest mean price of harvested fruits is also associated with tap water (76.7 pence) followed by those harvested from plants growth in organic media and watered with Filter 2 (25.1 pence). A low mean price for harvested fruits is associated with filters contaminated with diesel (Table 1). The lowest mean price is linked to plants growing in inorganic media (Table 8).

In comparison to Chillies, the overall growth development (Table 7) of Sweet Pepper was rather disappointing. This is possibly due to the high concentrations of nutrients and trace minerals, and adverse environmental boundary conditions in the laboratory.

The experiment shows that Sweet Peppers and Chillies can be grown successfully using wastewater treated by constructed wetlands. However, the yield of Sweet Peppers was insignificant in contrast to that of Chillies. The overall growth development of Sweet Peppers was rather disappointing, possibly due to the high concentrations of nutrients (particularly phosphorus and nitrogen) and trace minerals (iron, potassium and manganese). In contrast, chilies did reasonably well but the growth of foliage was excessive due to high nitrogen concentrations in the inflow water and the harvest was delayed.

The highest number of fruits was associated with tap water and an organic growth medium. However, the best fruit quality in terms of length, width and weight was observed for plants grown in organic media and irrigated with outflow water from wetlands with large aggregate size indicating that nutrient concentrations supplied to the Peppers by a combination of compost and treated wastewater were usually too high to produce a good harvest. A high yield was related to the most suitable provision of nutrients and trace elements. However, as the compost got depleted of nutrients such as nitrogen after about 14 months, the harvest increased for pots that received pre-treated wastewater in comparison to those pots depending only on the nutrients associated with the compost. Filters associated with hydrocarbon contamination were commonly associated with a poor harvest.

The productivity of Chillies in terms of harvest was independent of the wastewater consumption. Nevertheless, higher foliage production due to excess nutrients and trace minerals required more water. A high yield was related with the most suitable provision of nutrients and trace elements.

The current research will be continued with the same plants to assess if further harvests are economic and determine when the nutrients within the compost are fully depleted. Moreover, the accumulation of metals and their toxicity in the soil as well as microbiological contamination will also be studied.