The controlled temperature in S2 results in a doubled yield with respect to on-field cultivation, but still half of the one of vertical hydroponic . Having a high production in a limited space is one of the main qualities of vertical hydroponic systems. Moreover, since no soil is needed, the cultivation can take place in many different spaces: on rooftops, indoor, on abandoned industrial sites, on walls. An additional positive characteristic of hydroponics is that the quality and eventual contamination of the soil does not represent a risk for the products – simply because the two compartments are not in direct contact. This advantage is shown in the agricultural land occupation of scenario S1: by being soil independent and vertical, this type of hydroponics requires from four to 20 times less agricultural land than conventional agriculture. However, this means that hydroponic production needs to be supported by an external input of fertilizers to satisfy the nutrient requirement of the plants. The use of NPK fertilizers in the three scenarios is reported in Table 2. Scenario S1 needs more nutrient input than greenhouse cultivation, but less than on-field production, where part of the applied fertilizers is lost due to leaching processes. The different NPK proportion indicates that hydroponics allows an optimization of nutrient supply to support the growth phases of the plants. These results do not completely explain the impacts on freshwater eutrophication: in scenarios S1 the production of electricity and fertilizers are the main contributing processes, and in scenario S2 the production of heat is the dominant process. Scenario S3 has a negligible use of these inputs, but needs more fertilizers, which determines the emission of nutrients into the water streams.
It is important to distinguish where the emissions of nutrients take place, since eutrophication is a local phenomenon. In this sense, in scenario S1 and S2 the major risk of eutrophication is in the areas where fuels, hydroponic nft metals and fertilizers are extracted, i.e. mining sites distant from the place of final use. On the contrary, in scenario S3 the release of nutrients following the application of fertilizers represents a noteworthy impact on local water quality. About marine eutrophication, the actors in play are the same of freshwater eutrophication , but since scenario S3 has a high release of nitrogen , its impact is higher than scenario S1. Regarding eutrophication, a merit of hydroponics is the efficient use of nutrients, that, by being recycled with the water, are not released through the soil but remain available for the plants. The impact on freshwater ecotoxicity is due to the heat consumption in scenario S2, whereas for the other two scenarios several processes impact with the same order of magnitude: electricity production, fertilizers production, equipment production, pesticide production and transport. The combination of equipment and energy requirements is responsible of the impact of scenario S1 on fossil depletion. Scenario S2, due to the necessity to heat the greenhouse, has a performance more than ten times worse, while on-field cultivation , that is relatively low-input, has half the impact of S1. A benefit of urban agriculture is the shortening of the supply chain and the reduction of losses during the transport of the products. Thanks to the proximity of the farm to the consumers, the vegetables are fresher and do not need to be cooled during the delivery. For these reasons, we assumed zero losses in the transport phase. For conventional agriculture, on the other hand, the reported distribution losses are approximately 12%.The quality of the data influences the results. The primary data collected for scenario S1 refer to four months of production. Even if the farmers considered the seasonal variability for a better estimation of the annual consumptions, e.g. the water consumption, we recognise that the annual estimations could eventually not correspond to the actual values. Moreover, scenarios S2 and S3 are based on LCI datasets representative of the Integrated Production in Switzerland. Even though the authors of the datasets affirm that, most probably, their data are representative for similar cultivations in industrialised countries, these values cannot capture the peculiarities of country-specific cultivation practices. In conclusion, we acknowledge that the data do not have the same level of precision, and this affects the quality of the results.
To compare the performance of the systems, we chose the environmental indicators considered the most representative for LCA analyses of agricultural systems. We recognise that the agricultural sector can deliver other positive and negative effects, such as potential contribution to biodiversity, social issues and economic development. Consequently, integrating the analysis with other methodologies could give a broader perspective on the impacts of agriculture on an environmental, economic and social level.Constructed wetlands are commonly applied secondary or tertiary wastewater treatment systems . CWs exploit natural processes to remove pollutants via substrata, vegetation and microorganisms . CWs were traditionally employed to remove traditional pollutants, but are recently being used to remove micro-pollutants from wastewater. Numerous studies reported the adverse effects of micro-pollutants on human and environmental health . Effects include feminization of fish, short and long-term toxicity to biota, and the development of antibiotic resistance among natural and anthropogenic microbiomes . Currently, phyiscochemical technologies, such as ozonation, and membrane and activated carbon filters are implemented as tertiary treatment to remove micro-pollutants . However, such high-tech, expensive methods are less suitable in many emerging and developing economies . CWs are an attractive alternative: CWs are cost effective, eco-friendly and easy to operate and maintain . However, a key limitation in application of CWs is their high land area requirements, typically termed a land footprint. The CW substratum plays the key role in determining the size of a CW. Substrata remove various pollutants directly by sorption, precipitation, filtration and biodegradation . In addition, the substratum supports plant and microbial growth. Initially, soil was used as substratum. Recently, sand and gravel have been selected for their high hydraulic conductivity, resulting in improved performance of CWs. However, the application of these materials has limitations due to low removal of nitrogen and phosphorous, and more recently for micro-pollutants . Considering the importance of substrata for CW performance, selection of proper substrata could both reduce CW footprint while also removing micro-pollutants. Hydroponic substrata are good candidates for use in CWs.
Developed for hydroponic plant growth in horticulture, hydroponic substrata have attractive properties, such as higher oxygen and water holding capacity, large surface area and retain nutrients . In this study, the use of four commonly used hydroponic substrata were compared: mineral wool, pumice, wood fibre and coconut fibre. To date, a limited number of studies were conducted on their usability in CWs and on their potential to remove micro-pollutants from wastewater . Yang et al. reviewed the studies which focused on the use of emerged substrata in CWs to remove conventional pollutants. However, suitability of the emerged substrata to remove micro–pollutants was not reported. Furthermore, Wang et al. articulated that there is a lack of comprehensive studies available on the suitability of hydroponic CW substrata to remove various pollutants. Therefore, there is an urgent need to study the potential of the four substrata to remove micro-pollutants in CWs. Generally, sorption is one of the initial removal processes of a pollutant in the filter bed of CW. Sorption is an umbrella term covering both absorption and adsorption processes, and is especially used when these two mechanisms cannot be distinguished . Isotherm models, such as Freundlich and Langmuir models can be used to unravel the sorption mechanisms from experimental data to distinguish surface related or absorption related mechanisms.These include physico-chemical properties of a sorbent and sorbate, such as acidity and hydrophobicity of a pollutant, and organic matter content of a substratum . These factors influence the affinity between a sorbent and sorbate and transport of the sorbates from bulk solution to sorption sites, influencing extent and kinetics of sorption. Thus, hydroponic channel understanding sorption mechanisms is crucial for successful applications of a substratum in CWs treating micro-pollutant-containing wastewater. This study aims to investigate the applicability of four hydroponic substrata for efficient removal of micro-pollutants from wastewater. We determine sorption affinities and kinetics, and interpret this data using isotherm models and physico-chemical properties of substrata and micro-pollutants.
These results are used to assess the suitability of the substrata in CWs, both for micro-pollutant removal and for reducing CW footprint.Sorption kinetics were studied using the sorption of the micro-pollutants into the substrata over the contact time . Fast initial sorption to organic substrata was observed for most of the studied micro-pollutants during the first 6 h, followed by a slower sorption phase. This suggests diffusion into internal substratum was limiting: rapid sorption to the outer layer of the substrata is followed by slower diffusion in an apparent first order process. Naproxen and ibuprofen exhibited distinctly slower kinetics during sorption to the organic substrata. For inorganic substrata, even slower kinetics were observed for the studied micro-pollutants . A thorough and further mechanistic interpretation of the sorption kinetic data using pseudo-first and second-order models was not possible due to insufficient R2 values resulting from fitting data with these models . The profile of the sorption kinetics was used to identify equilibrium. The sorption curves levelled off between 24 and 72 h, indicating that sorption reached a status of equilibrium. Based on this, the concentration obtained at 72 h was taken as an estimate of the equilibrium concentration Ce.Organic substrata wood fibre and coconut fibre sorbed micro-pollutants more than the inorganic substrata mineral wool and pumice . The observed sorption affinity of the organic substrata with the micro-pollutants followed the order: trimethoprim>carbamazepine >caffeine>sulfamethoxazole>ibuprofen= naproxen . A similar order of sorption affinity for trimethoprim, carbamazepine and sulfamethoxazole was found for soils by Kodešová et al. . Ibuprofen and naproxen, containing -COOH groups, and sulfamethoxazole, containing -SO2-NH moieties, have acidic protons which can dissociate and form anionic species . Subsequently, a strong repulsion occurs between these anionic species and negatively charged surface of the organic substrata. Although sulfamethoxazole has a basic NH2 group , it has acidic protons due to the presence of an acidic -NHand an electron-withdrawing -SO2- in the vicinity. The presence of two basic NH2 groups in trimethoprim explains its highest sorption affinity. Carbamazepine has one amido group and the basicity of an amido group is lower than of an amino group.
This could explain the sorption order: trimethoprim is followed by carbamazepine. Caffeine has a lone pair of electrons at the non-methylated N site . Therefore, caffeine acts as a proton-acceptor, basic, so it is positively charged and attracted to the negatively charged surface. Caffeine is hydrophilic whereas carbamazepine is hydrophobic due to its dibenzoazepine structure . The hydrophobicity positively affects the sorption . Therefore, carbamazepine is followed by caffeine in the sorption order. The sorption of micro-pollutants on organic rich materials appears to be a trade-off between electrostatic interaction and hydrophobic interactions between the organic matrices and the micro-pollutants. This is further depicted in a four quadrant matrix . In the high pKa and high Log Kow quadrant, both electrostatic interactions and hydrophobicity positively affect the sorption, as indicated by high KF values for carbamazepine on the organic substrata. For the low pKa and high Log Kow quadrant, the positive effect of hydrophobicity of iburprofen and naproxen appears to be largely counteracted by strong electrostatic repulsion due to deprotonation . At the quadrant of high pKa and low Log Kow, an intermediate sorption on the organic substrata was observed for caffeine and sulfamethoxazole. In this study, no compounds falling into the low pKa and low Log Kow quadrant were investigated. We speculate based on our results, that even lower sorption is to be expected for compounds in this quadrant on our selected substrata. Inorganic substrata mineral wool and pumice generally exhibited much lower sorption and slower sorption kinetics towards the studied micro-pollutants. Mineral wool is a non-reactive neutral material. Therefore, unlike wood fibre and coconut fibre, mineral wool has no extensive net negative surface charge. Furthermore, mineral wool has the lowest surface area among all the substrata . These properties together explain the low sorption of the micro-pollutants.