Experiences of running a hydroponic system in a pilot scale for resource-ef ﬁ cient water reuse

Within the research project HypoWave, a hydroponic system for plant production was investigated. The hydroponic system was fed with wastewater that had undergone specially adapted treatment. The principal aim was to develop a combined system for water treatment and hydroponic plant production, where water and nutrients were reused ef ﬁ ciently to produce marketable food products. Another goal was to ﬁ nd out whether the reuse of pre-treated wastewater for plant growth in a hydroponic system could also present an additional alternative wastewater treatment step for enhanced nutrient removal. A pilot plant, consisting of various treatment steps such as activated sludge process, ozonation and biological activated carbon ﬁ ltration, was used to produce lettuce with irrigation water of different qualities. The hydroponic pilot plant was operated in two different modes – ﬂ ow-through and feed & deplete. This paper focuses on the in ﬂ uence of the various modes of operation and accordingly varying nutrient concentrations (N, P, K) on plant growth. Furthermore, heavy metal content in the various types of treated wastewater and in the produced plants was investigated. In addition, the results of the different modes of operation were veri ﬁ ed by mass balances for N, P and K.


GRAPHICAL ABSTRACT INTRODUCTION
Agriculture, which consumes 70% of the fresh water resources utilized by mankind, is the greatest water user of all (FAO ). Climate change, growing population and urbanization are expected to increase the pressure on water resources but also on the availability of arable land and nutrients (WHO a, b; Jiménez & Asano ). This makes it increasingly difficult to guarantee food security and to prevent the deterioration of ecosystems.
Hence, the investigation of alternative water resources for agricultural irrigation and production is necessary.
Countries suffering from significant water stress already use wastewater for irrigation purposes. A great advantage of this option is that not only the water itself is reused but also the nutrients contained in the treated water. However, even treated wastewater contains problematic substances, such as heavy metals, salts, micropollutants or human pathogens. Therefore, treated wastewater could cause sali- all these investigations were carried out in small-scale systems with the focus on treating wastewater with the production of plants as an inevitable byproduct. Therefore, the approach of regarding the processes of wastewater treatment as a very specific production of high-quality irrigation water, which is used in hydroponic systems for the production of products of high quality, is new.
This special focus of the research project, discussed in this paper, is already mirrored in its title 'HypoWave', a combination of the words 'hydroponic' and the official This paper will give an overview from the results of the pilot plant of the research project HypoWave and will focus on the data of the growing periods 2017 and 2018. First, a detailed description of the pilot plant is presented and second, the nutrient ratios of the differently treated wastewaters are described. In addition, the use of wastewater treated especially for the reuse in a hydroponic system for the cultivation of lettuce is discussed, taking into account the impact of heavy metal concentrations. Finally, the used hydroponic system is verified by mass balances together with a recommendation on which type of operation (flowthrough, feed & deplete system) is suitable for which type of application, combining adapted wastewater treatment and hydroponic systems. Here, the first mode of operation was a flow-through system. It is a classic open hydroponic system without recycling of the nutrient solution (¼treated wastewater) (Maucieri et al. ). In order to operate the hydroponic system as a recirculation system with treated wastewater (second mode of operation), a circulation system (feed & deplete system) especially optimized for the HypoWave project was used.
In addition to the investigations of the nutrients and heavy metals in the irrigation water and the plants, further investigations regarding pathogens, antibiotic resistances and organic micropollutants were carried out (Blau et al. in preparation; Kreuzig et al. in preparation; Mohr et al. submitted The integration of the results into an effective quality management system will be part of a follow-up project.

METHODS
In order to analyse the suitability of various types of treated wastewater for growing lettuce in a hydroponic system installed in a greenhouse, a modularly structured pilot plant was set up at the municipal WWTP Wolfsburg-Hattorf, Germany. The pilot plant consisted of three main modules: basic treatment, quality and hygiene and the hydroponic system (greenhouse). The modules and technologies used have been presented below in more detail as well as the various treatment combinations and modes of operation of the hydroponic system.

Basic treatment
The aim of this module was to reduce the amount of organic compounds of the mechanically pre-treated wastewater by using aerobic and anaerobic processes. In order to obtain a nutrient-rich effluent with low COD concentrations, an SBR was installed down-stream of the EGSB. The exchange volume of the SBR was 240 L per cycle (40-60 L h À1 ; equal to 1/3 of the volume). By combining the process steps of aeration, settling and decanting, the SBR makes it possible to convert ammonium into nitrate during the nitrification process. The conversion is necessary, since nitrate is more readily available for plants then ammonium (Tochobanoglous et al. ). At the same time, elimination of phosphorus and denitrification are avoided, thus preserving these nutrients in the irrigation water.

Quality and hygiene
The effluent of the conventional biological treatment process (secondary sedimentation tank (SST)) may still contain organic micropollutants or pathogens. These pollutants in treated wastewater could have a negative effect on the quality of the cultivated product. Therefore, further advanced treatment processes are required. In the Hypo-Wave pilot plant, ozonation and BACF were used as advanced treatment processes. Ozone is one of the principal disinfectants for treating wastewater and a strong oxidant pumps (year 2018). Outer lines on each side (two on the south side, one on the north side) served as border lines, which were used to ensure that conditions (e.g. shading, neighbourhood) were identical for all lines/plants. Four lines were operated with the various treated wastewaters and one line was operated as a reference line. In the reference line, a nutrient solution was used with a nutrient concentration corresponding to a 50% Hoagland nutrient solution (Epstein & Bloom ). Preliminary tests had shown that 100% Hoagland nutrient solution was too highly concentrated and could damage the plants. Therefore, a diluted 50% solution was used to simulate conventional hydroponic lettuce cultivation as reference.

Modes of operation of the hydroponic system
The experiments were carried out in 2017 and 2018 from April to November. Every growing period lasted about 35 days. Within the project, two modes of operation were investigated. In 2017, the hydroponic system was operated as a flow-through system without any further addition of nutrients. The hydroponic system was operated at an average flow rate of 23.5 L h À1 and with 68 lettuce plants per line.
During this test period, research investigations covered the analysis of the effluent from the treatment lines A to D.
In a second test period in 2018, a feed & deplete system was used. The treated wastewater (V start ¼ 185 L) circulated In the case of the harvested lettuce, fresh and dry matter, the concentration of macronutrients and micronutrients, as well as heavy metals, was examined. Table 1 shows the methods used for analysis.
After the roots were cut off, the fresh weight of the lettuce was measured on the day of the harvest. After weighing, the lettuce was slowly dried at a temperature of 70 C, to make sure that the material was not destroyed by the heat. The drying process took 3 days. The subsequent analyses were carried out either by the Abwasserverband Braunschweig (Braunschweig, Germany), the Wolfsburger Entwässerungsbetriebe (Wolfsburg, Germany), the TU

Braunschweig (Institute of Sanitary and Environmental
Engineering TU Braunschweig, Braunschweig, Germany) or the University of Hohenheim (Stuttgart, Germany).

Mass balances
Mass balances were used to check the validity of the system and the plausibility of achieved results. Mass balances were drawn up for the greenhouse covering the load of each parameter in the influent of the hydroponic lines, the effluent of the hydroponic lines and in the harvested lettuce. Loads were calculated on the basis of medians (concentration, mass).  Former research has shown that mass balances of wastewater treatment systems are considered to be closed, if the balance gap is less than or equal to 10% (Mieske ).
Due to additional sources of error, in this research project, mass balances for hydroponic systems were considered to be closed if the balance gap was less than or equal to 20%.
Reasons for this assumption were that the period of nursing (20 days) and the nutrient composition of the roots after harvesting were not considered in the mass balancing. Further, the structure of lettuce is not homogeneous and random samples were taken instead of composite samples on account of the equalization of the treated waters during the various treatment processes.

Treated wastewater
In order to be able to use a larger database, the mean aver- A similar tendency could be observed for zinc. The zinc concentration in the water of treatment line C corresponded to

Fresh mass
For both RUN I and RUN II, the fresh mass of the shoots was measured after harvesting along the hydroponic line. In contrast to these results, the weight of the lettuce grown in lines C and D was similar along the entire line. It is noticeable that the lettuce plants in line D (150 g) were about 20% lighter than those in line C (183 g). This difference was attributed to the nutrient composition of the water in lines C and D. Treatment line C and D differ only in the biological activated carbon filter. As already discussed above, the biological activated carbon filter reduced some metals required for plant growth. Lower yield seems to be a consequence of these micronutrient deficiencies. The weight of the lettuce plants from lines D and C corresponded to the lettuce weight in the reference line. Thus, it can be deduced from these investigations that water used in treatment lines D and C is suitable as irrigation water in a hydroponic system operated as a flow-through system. The results of RUN II are presented in Figure 3 is sufficient for achieving the required yield (275 g).

Quality of the lettuce
The N content of the shoot biomass corresponded to the biomass production presented above within and between the lines over the two experimental years. In the first year, the shoot N content was between 40 and 50 g kg À1 dry matter in the front of all lines and fell to 20 g kg À1 towards the end of lines A and B. This suggests that in RUN I, the depletion of N towards the end of these lines A and B limited plant growth. A small, but significant N decrease (less than 5 g kg À1 ) in the shoot biomass was also observed in the reference line and line D. Accordingly, the shoot P content decreased from 6 to 4 g kg À1 along lines A and B. K was higher in the front section of the reference line and lines A and C. However, the K content of all shoots was in the magnitude of 60 g kg À1 . In RUN II, the reference shoot biomass contained 65 g N kg À1 in the front and 50 g N kg À1 in the end of the line, thus distinctly more than in 2017. In all lines fed with treated wastewater, the N content was between 37 and 48 g kg À1 without any remarkable difference between the front and end of the lines. The highest P content was measured in the plants from the front of the reference line ( μg kg À1 respectively. In RUN II, the maximum As value detected was 90 μg kg À1 , Co 190 μg kg À1 and Cd was below the detection limit in all cases. The highest Cr content was 3,200 and that of Pb 110 μg kg À1 .

Mass balances
In order to validate the system, mass balances were carried out exemplarily on the hydroponic line, operated with the effluent of treatment process D. The results of the mass balances are presented in Table 3. It is apparent that the nitrogen, phosphorus and potassium load in the influent, effluent and the lettuce in RUN I are distinctly higher than in RUN II, although the concentrations were similar (see Figure 2).  suggest that these heavy metal contents are of no concern. Previous studies did not validate the hydroponic system run with wastewater as irrigation water by mass balances.
Therefore, to complete the investigations mass balances were performed for two runs (RUN I flow-through and RUN II feed & deplete) of the hydroponic lines operated with treated wastewater from line D in 2017 and 2018.
The mass balances were calculated for irrigation water and grown biomass and verified the plausibility of achieved results, thus validating the hydroponic system used.
As a conclusion, it can be stated that it is generally possible to operate hydroponic systems using treated wastewater.
Our research showed that wastewaters with high and with