Science of the Total Environment 794 (2021) 148689 Contents lists available at ScienceDirect Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv Optimizing irrigation in urban agriculture for tomato crops in rooftop greenhouses Felipe Parada a, Xavier Gabarrell a,b, Martí Rufí-Salís a, Verónica Arcas-Pilz a, Pere Muñoz a, Gara Villalba a,b,⁎ a Sostenipra Research Group (SGR 1683), Institute of Environmental Sciences and Technology (ICTA-UAB), MdM Unit of Excellence, Universitat (CEX2019-000940-M) Autònoma de Barcelona (UAB), Campus UAB, 08193 Bellaterra, Barcelona, Spain b Department of Chemical, Biological and Environmental Engineering, School of Engineering, Building Q, Universitat Autònoma de Barcelona (UAB), 08193 Bellaterra, Spain H I G H L I G H T S G R A P H I C A L A B S T R A C T • Recirculation systems reduce the ma- rine and freshwater eutrophication. • Leachate as irrigation water is an alter- native for the reduction of irrigation water and maintains yields. • It is necessary to find ways to fertilizer substitution and irrigation optimization. • Water reduction management can in- crease in 7% the WUE of the tomato crop. • The operation phase shows improve- ments in all of the impact categories evaluated. a r t i c l e i n f o a b s t r a c t Article history: The rise of population in urban areas makes it ever more important to promote urban agriculture (UA) that is ef- Received 30 April 2021 ficient in terms of water and nutrients. How to meet the irrigation demand of UA is of particular concern in urban Received in revised form 11 June 2021 areas where water sources are often limited. With the aim of determining how to reduce water use for irrigation Accepted 22 June 2021 while maintaining productivity and reducing environmental impacts in UA, this study explores the agronomic Available online 29 June 2021 performance and environmental life cycle impacts and benefits of three different fertigation management prac- Editor: Damia Barcelo tices used in a rooftop greenhouse for tomato crop in Barcelona: 1) open management (OP); 2) recirculation (RC), in which 30% of the drained, unused water is used to irrigate the crop; and 3) the same recirculated man- agement of RC with a further reduction in fresh water input of 15%(RR). Despite the recirculation and reduction of Keywords: water and nutrients, all three irrigation management practices resulted in similar yields: 16.2, 17.9, and Life cycle assessment (LCA) 16.8·kg·m−2 for OP, RC, and RR, respectively. In terms of water-use efficiency, RR management was the most ef- Water-use efficiency (WUE) ficient, requiring 48.7·liters·kg−1 of tomato, followed by RC (52.4·L·kg−1) and OP (75.2·L·kg−1). RR presented Reduction water consumption an improvement of 7% in water-use efficiency. In terms of environmental performance, RC had the best perfor- Eutrophication reduction mance in almost all impact categories during the operational phase, especially in regard to marine and freshwater Leachate recirculation eutrophication, with 44% and 93% fewer impacts than OP due to the recirculation of nutrients and reduced nutri- Recirculation system ent loss through leachates. In terms of infrastructure, even though recirculation management requires additional equipment, the materials present better performance in the range from 0.2 to 14% depending on the impact cat- egory. This study can support evaluation of agricultural projects in the city, through yields and water consump- tion presented, incentivizing good practices aligned with the sustainability of UA. © 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). ⁎ Corresponding author at: Department of Chemical, Biological and Environmental Engineering, School of Engineering, Building Q, Universitat Autònoma de Barcelona (UAB), 08193 Bellaterra, Spain. E-mail addresses: Felipe.Parada@uab.cat (F. Parada), Xavier.Gabarrell@uab.cat (X. Gabarrell), Marti.Rufi@uab.cat (M. Rufí-Salís), Veronica.Arcas@uab.cat (V. Arcas-Pilz), Pedro.Munoz@uab.cat (P. Muñoz), Gara.Villalba@uab.cat (G. Villalba). https://doi.org/10.1016/j.scitotenv.2021.148689 0048-9697/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 1. Introduction countries was similar, Morocco had over three times the freshwater de- pletion. On the other hand, as a result of having more sophisticated Current trends in population growth lead to an increase in demand technologies and a cooler climate, French tomato production requires for food, water, and energy. This demand becomes a challenge, particu- more energy consumption, resulting in higher global warming and larly for urban areas where half of the world's population resides and is eutrophication potentials. He et al. (2016) were able to show a better expected to rise to more than 70% by 2050 (UNDESA, 2018). Urban ag- life cycle environmental performance by reducing chemical fertilizer riculture (UA) can meet part of that food demand, additionally provid- and pesticide consumption in organically grown tomatoes, albeit ing further advantages, such as reduction of environmental impacts more land was required to compensate for the lower yields. Rufí-Salís and food losses associated with transportation over long distances et al. (2020b) presented a study on an integrated rooftop greenhouse (Caldeira et al., 2019; Sanyé-Mengual et al., 2013). UA provides oppor- (i-RTG) with different crops grown using water recirculation manage- tunities to improve urban metabolism through the optimization of ment, identifying the best combination of crops in the greenhouse to urban cycles with agro-urban systems through the recovery of nutrients define the generated environmental impacts; nevertheless, their study from urban organic waste and wastewater (De Corato, 2020; Ulm et al., did not consider irrigation as a variable to be optimized. 2019) and the integration of buildings with greenhouses on roofs for en- To summarize, diverse authors have used LCA to evaluate crop pro- ergy reduction (Nadal et al., 2017), thereby promoting the circularity of duction in UA systems; however, few have explored and quantified resources. In other words, increasing food sovereignty in cities cannot (Parajuli et al., 2019) how various water and nutrient optimization come at the price of increasing environmental impacts because more re- strategies can reduce the impacts while maintaining profitable yields. sources need to be imported. This study aims to contribute to this research gap in UA systems by an- Traditionally, agriculture has been characterized by the inefficient alyzing alternatives for efficient water management strategies, such as use of resources, both in terms of water and nutrients. Currently, agri- the recirculation of water and nutrient flows and reduction of applied cultural practices consume more than 85% of available freshwater and water while maintaining yield. Furthermore, since an irrigation system 80% of the annual phosphate rock extracted globally (Shu et al., 2006; is used to fertilize crops, the reduction of water in recirculated irrigation van Schilfgaarde, 1994). Additionally, the water supply is scarce and un- management results in a reduced amount of fertilizers, which were also stable due to extended dry periods, heatwaves, and low pluviometry quantified by performing nutrient balances. In addition to yield, water (Schmidhuber and Tubiello, 2007). Today, there is scientific consensus efficiency, and nutrient balance, an environmental analysis of all three on the depleting nature of phosphorus (Rittmann et al., 2011), where irrigation strategies was performed (functional unit of 1 kg of tomato) phosphate rocks are the main source of phosphorus, and 80% of the and determines the effect of water recirculation management on the available stock is used in the production of fertilizers (Shu et al., yield and environmental burdens. The tomato crop was selected for 2006). The use of chemical fertilizers has increased up to 36% since three reasons: first, tomatoes are the most consumed horticultural 2002, indicating our dependence on a nonrenewable resource crop in Europe (European Commission, 2011), with 24.6 million (FAOSTAT, 2017a). Steen (1998) mentioned that mineral P resources tons per year, and are mainly produced in Spain and Italy (Cook have been depleted in the last century. The intensified use of fertilizers et al., 2018; FAOSTAT, 2017b). Second, tomatoes are traditionally results in eutrophication and other diffuse pollution problems (Chen grown in places with low precipitation and warm climates. An et al., 2017, 2021; Nagendran, 2011; Novotny, 1999). For these reasons, example of this is Almería (Spain), where the precipitation is near optimizing water and fertilizer management in agriculture should be a 218 mm per year (SIAR, 2019). Third, given its high water require- priority, mostly in cities, where these resources are limited or come ments, tomatoes are an excellent crop to study the benefits of from faraway places. producing them in urban areas with water and nutrient optimization For the development of UA, efficient use of water is essential (Tixier strategies. and de Bon, 2006), and some different technologies and management We hypothesize that the efficient use of water and nutrients through practices allow maximization of the use of water, such as drip irrigation recirculation management reduces the environmental impacts of to- systems, which enable reaching efficiencies in irrigation up to 95%; mato production in UA through the contrast of three management prac- added to other measures, as climatic predictions can improve the tices and the generation of real data. The three strategies include open amount of applied irrigation water. Mason et al. (2019) simulated dif- management and two types of recirculation management (with and ferent climates in the United States, showing the benefits of applying in- without restrictive water irrigation) during two cultivation periods of telligent irrigation systems, which consider climatic information to tomato production in an i-RTG. determine the amount of water to apply to crops. They found a 46% av- erage savings (ranging from 2 to 96%). Other studies have shown how 2. Materials and methods the implementation of efficient irrigation systems reduces the amount of water and nutrients applied (Contreras et al., 2017; Hooshmand 2.1. Case study: the integrated rooftop greenhouse (i-RTG) et al., 2019; Liu et al., 2019). Although there seem to be ample benefits from UA, it is crucial to This study was performed in the i-RTG located inside the Institute of analyze crop production from a systemic life cycle approach to avoid Environmental Science and Technology (ICTA-UAB) building on the counterproductive impacts and to improve system optimization. Universitat Autònoma de Barcelona (Catalonia, Spain) campus located Additionally, a widely used method to evaluate the environmental in the outskirts of Barcelona. The site is characterized by a Mediterra- performance of processes is life cycle assessment (LCA). This is used to nean climate with warm summers and rainy winters. assess the potential environmental impacts, both direct and indirect, The experiment was conducted in the southeast-facing corner of the associated with a product throughout its entire lifetime in a systemic i-RTG for the production of the tomato species Solanum lycopersicum L. approach and is useful in identifying opportunities to improve the pro- cultivar Arawak, with a total available area of 84.5 m2 and a functional cess and reduce impacts (ISO 14040, 2006). To summarize, diverse that harvesting area of 63.5 m2. The frame of the plantation was 0.33 × environmental impacts related to the operation stage are mainly associ- 1.1 m (Fig. 1) with a total of 171 tomato plants distributed in 57 perlite ated with fertilizers, diesel and emissions from land use change. substrate bags (40 L), making a plant density of 2.7 plant∙m−2. The study (Martínez-Blanco et al., 2011; Parajuli et al., 2019) Payen et al. (2015) took place during two consecutive years, 2018 and 2019, where the conducted a study on tomato production in two countries with con- tomato season lasted approximately 6 months each year. trasting climates (Morocco and France), showing that impacts depend A drip irrigation system was used with a 2 L∙h−1 water flow in which highly on water extraction and treatment for irrigation. Their study fertilizer was applied according to need throughout the cropping season showed that although the tomato crop water consumption in both (Table 1). 2 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 Fig. 1. Distribution of plants at the ICTA-UAB Laboratory of Urban Agriculture. This building has a rainwater harvesting system (RWHS) which con- applied irrigation water was made (Favati et al., 2009; Wang et al., sists of a 100 m3 tank buried under the building that is used to irrigate 2019). Irrigation was given in ten evenly distributed, daily doses. the crops inside the greenhouse. The rainwater used for irrigation was pumped from the RWHS to two containers of 300 L each inside the 2.2. Water-use efficiency (WUE) greenhouse on the top floor of the building. The leachate was collected in slightly-tilted aluminum trays where To relate the amount of biomass produced to the amount of water the crop bags were placed. These allowed collecting the excess irrigation used, the water-use efficiency was used as indicator (WUE) previously water (leachates) by gravity, towards a secondary container, for storage established in the literature (Green et al., 2010; Muñoz et al., 2008). and distribution. The WUE is defined as the rate of biomass accumulation per unit of The leachate collection system was carried out through aluminum water consumed and allows a simple comparison of various crop pro- trays where the crop bags were placed. These allowed collecting the ex- duction systems, such as greenhouse versus field production. For the cess irrigation water (leachates) by gravity, towards a secondary con- present work, the WUE was calculated as the relation between water tainer, to be later stored in a general container where it was stored to supplied (in liters) and yield (kg of tomatoes produced) for the entire be redistributed. crop cycle, as shown by Eq. (1). The agronomic results were focused on the water and yield relation- ship. For the environmental part, the results were centered on the opti- Liters of added water to the system, ½L WUE ¼ ð1Þ mization of fertilizers, energy, and analysis of infrastructure. The three Kilograms of Tomatoes produced by the system½Kg  irrigation managements were performed as follows and summarized in Table 2: (1) open management (OP): traditional drip irrigation where 30% of the water supply was drained and discharged to the 2.3. Nutrient balance wastewater sewer, implemented in 2018; (2) recirculation control (RC): traditional drip irrigation with an identical 30% of drainage, but The nutrient balance was estimated by determining the nutrient the drained water was collected and recirculated, implemented in input through the irrigation system and the nutrient output embodied 2019; and 3) leachate recirculation (RR): irrigation water volume was in the crop, residual biomass, and leachates. The difference was attrib- reduced by 15%, which was also recirculated and recycled, and imple- uted to the accumulation of nutrients in the perlite substrates and com- mented in 2019. The objective was to reduce water intake without pared with previous studies as a cross check. compromising production. In this sense, different authors have Eq. (2) was used to estimate the total amount of nutrients (nitrite, discussed that a reduction of approximately 20% in potentiation evapo- nitrate, phosphorus, and potassium) in both water flows: input (irriga- transpiration would not affect productivity. Considering that the hydro- tion) and output (leachates or drainage). ponic system requires a minimum leaching fraction, a 15% reduction in Total amount of nutrients from irrigation or leachates ½kg ∑X i ∗ ½Nci Table 1 ¼ ð2Þ Concentration of fertilizers used. 106 Fertilizers 2018 OP 2019 RC 2019 RR where i is a specific period of time, Xi represents the partial volume [L] [g∙m −3 a ] [g∙m −3 a ] [g∙m−3]a for period i, and Nci is the nutrient concentration [mg∙L−1]. In this way, the total amount of nutrients is the sum of the multiplication of KPO4H2 214 283 283 KNO3 104 138 138 the partial concertation by the volume. To obtain nitrite and nitrate K2SO4 277 367 367 concentrations, the irrigation and leachate samples collected directly Ca(NO3)2 403 533 533 from the dripper three times per week with ion chromatography were CaCl2 100 133 133 analyzed (ICS-1000 and AS-DV by Dionex), whereas nitrate, total Mg(NO3)2 134 178 178 Hortrilon 8 11 11 phosphorus, and potassium concentrations were obtained via atomic Sequestrene 8 11 11 spectroscopy (optima 4300DV by PerkinElmer). a Calculated on a basis of irrigation water applied. To prevent a low concentration of To estimate the nutrients embodied in the crop and residual bio- nutrients in the recirculated leachates and abrupt osmotic changes, the NPK concentration mass, Eq. (3) was used, where DMi represents the partial dry matter was increased. of the sample of tomatoes [g dry matter], and NcDM i is the nutrient 3 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 Table 2 Crops and treatment under assessment. Management Initial plants Daily drainage Year Start Ends Open (OP) 171 ~30% 2018 10th January 30th July Recirculated control (RC) 90 ~30% 2019 14th January 2nd August Recirculated reduction (RR) (~15% irrigation reduction of RC) 81 concentration [mg∙g −1 dry matter] obtained by gas chromatography for N The main function of the greenhouse is food production, 1 kg of to- (6890 by Agilent Technologies and 5973 by HP) and atomic spectros- matoes was determined as a descriptor. In the same way, previous re- copy (Optima 4300 DV by PerkinElmer) for P and K. Five tomato sam- search has used this functional unit as a reference in tomato (Piezer ples for each irrigation management were used. et al., 2019; Pineda et al., 2020) and other crops (Arcas-Pilz et al., 2021; Rufí-Salís et al., 2020a). Total amount of nutrients from biomass and tomatoes ½kg System Boundaries. ¼ ∑DM i ∗ ½NcDM i ð3Þ To better discuss the results, the assessment is separated into two systems, as shown in Fig. 2: 1) the infrastructure, which considers all To estimate nitrogen emissions to the atmosphere, the value pro- life cycle stages of the greenhouse structure, RWHS, auxiliary equip- posed by Llorach-Massana et al. (2017), which considers an emission ment, and recirculation system, as well as any materials that had more factor of 0.00785 kg∙N2O−1 per kg∙N−1, was used. than a 5-year lifespan; and 2) the operational system, which includes Finally, the nutrient balance was calculated with Eq. (4) (values all life cycle stages of the fertilizers, growth media, pesticides, nursery expressed in Kg), where XT represents the total mass of nutrient sup- plants, and energy (treatment, pumping, and transport). The auxiliary plied by the irrigation. XL is the amount of nutrients in the leachates. equipment considered in this work is crop trays, manometers, pumps, XY and XB represent nutrient uptake by tomatoes and the rest of the bio- water polyethylene tanks (2 × 300 L), and leachates polyethylene mass (leaves and stem). XEA represents the emissions to the atmo- tank (300 L). (For more details see the supplementary information sphere, which in our case is only applicable to N in the form of N2O. Appendix C). Waste management for the operation system considered the transport and landfilling of perlite after three years of use. Biomass X T ¼ X L þ X B þ X Y þ X EA ð4Þ obtained throughout the experiment due to pruning and at the end of the production season was composted, although this process was not considered within the environmental analysis. The impacts from 2.4. Life cycle assessment (LCA) transport to the distribution of the tomatoes to the consumers are not considered since the building personnel consumed the tomatoes. (See The LCA (ISO 14040, 2006) methodology was used in this study be- Fig. 3.) cause it provides a broad vision of the environmental impacts, allowing us to determine the particular contributions of each item considered in 2.4.1. Inventory our system. This provides a big picture on the performance of the differ- Previous inventories developed for this i-RTG were used for the ent water management and its implications at the productive level, con- RWHS and nursery plants (Sanjuan-Delmás et al., 2020; Sanyé- sidering all life cycle stages for tomato production. Mengual et al., 2015) and auxiliary equipment Rufí-Salís et al. (2020a). Fig. 2. The three irrigation schemes, including auxiliary equipment. OP management (red line), recirculated management (green line). 4 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 Fig. 3. Diagram of system boundaries, depicted by the dotted green line. The emissions in the leachate were considered to be directly emitted kg∙m−2 for open management (OP) to 17.9 kg∙m−2 for recirculation to the aquatic environment and were determined to be −NO2 and −NO3. control (RC), as shown in Table 3. These values are consistent with Air emissions from nitrogen fertilization were calculated using emission those achieved in other studies of tomato production in conventional factors by Montero et al. (2011) and the IPCC (De Klein et al., 2006). The greenhouses under similar ventilation conditions, such as Boulard life cycle inventory is available in the supplementary information (Ap- et al. (2011), who reported 15 kg∙m−2 in France, and Muñoz et al. pendix C.1 and C.2) for the infrastructure and operational subsystems, (2008), who reported a similar value of 16.5 kg∙m−2 in Spain. respectively. An impact allocation procedure based on rainwater vol- WUE was calculated to determine the biomass accumulation (edible ume consumed was applied to estimate the impacts of the RWHS attrib- yield) per liter of irrigated water to be able to compare productivity uted to the crops because this system is also used for irrigation of among the various irrigation systems explored. Here, leachate recircula- ornamental plants throughout the building, as had been done previ- tion management (RR) showed the best performance, with 48.7 L·kg−1, ously for other i-RTG studies (Rufí-Salís et al., 2020b). An allocation pro- follow by recirculation control (RC -52.4 L·kg-1) and open system (75.2 cedure was applied for the fertilizers, which were calculated in a linear L·kg-1) as shown in Table 3. Although RR obtained less production, in proportion concerning the irrigation water applied. In the same way for terms of WUE was approximately 35% more efficient than OP pesticides, the proportion was calculated in a linear proportion to the management. number of plants present at that moment in the field. Finally, impact assessment was performed with Simapro 9 software, 3.1. Yield and climatic variables and environmental information was acquired from Ecoinvent Database v3.5 (Wernet et al., 2016). The life cycle impact assessment (LCIA) Yield depends not only on water and nutrients but also on other fac- entailed the use of the Recipe 2016 method (Hierarchical) at the mid- tors, such as the amount of radiation received. Consequently, it is impor- point level (Huijbregts et al., 2017). The impact categories considered tant to determine to what degree radiation contributed to the yields in this study were global warming, (kg CO2 equivalent), terrestrial acid- obtained, rather than or in addition to the irrigation scheme chosen. ification (kg SO2 equivalent), freshwater eutrophication (kg P equiva- The total radiation (more details in supplementary information - lent), marine eutrophication (kg N equivalent), fossil resource scarcity Appendix E) during the crop season in 2018 and 2019 was very similar, (kg oil equivalent), cumulative energy demand (MJ), and ecotoxicity averaging 3610 MJ and 3988 MJ, respectively (~7% difference), thereby (kg 1,4-DB equivalent), which is the sum of marine, terrestrial, and allowing us to discard any hypothesis that similar yields were obtained freshwater ecotoxicities. A cutoff criterion was used to estimate the en- in the RR and RC systems as in the OP system due to more radiation vironmental impacts, and it was assumed that the secondary product compensating for the lack of water or nutrients. This situation contrasts receives the impacts and benefits of the recycling process. with the one presented by Rufí-Salís et al. (2020b), where important differences in terms of radiation above 60% were given in crop seasons 3. Results of 60 and 90 days long (green bean crop). Longer campaigns, such as the tomato cycle, are more stable in terms of accumulated radiation This section presents the experimental, analytical, and environmen- since a longer period of time allows climatic variability to be absorbed. tal results of the three irrigation management schemes. First, the water- use efficiency (WUE) is presented, where OP was the least efficient and RR was the most efficient. Next, the temperature and relative humidity Table 3 were similar for both 2018 and 2019, thereby affecting the yield equally Summary of agronomic variables (yield, water used, and WUE). for all experiments. Second, nutrient balance was performed to deter- Parameter Unit 2018 OP 2019 RC 2019 RR RC/OP RR/OP RR/RC mine the nutrient flows for an accurate accounting of the emissions to a −2 Yield kg∙m 16.2 17.9 16.8 110.2% 103.2% 93.7% the environment. Finally, the results of the life cycle analysis identify Water usedb L∙m−2 1220.7 936.8 815.7 76.3% 66.4% 87.1% the environmental benefits and costs for the infrastructure and the op- WUEc L∙kg−1 75.2 52.4 48.7 69.3% 64.4% 92.9% eration of these schemes. a Yield considering all tomatoes harvested, divided by effective harvest area. Despite the recirculation and reduction of water and nutrients, all b Calculated on a basis of irrigation water added to the system. three irrigation schemes obtained similar yields ranging from 16.2 c Water-use efficiency considering liters per kilogram of tomatoes. 5 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 The temperature and relative humidity during the experimental pe- on the top floor of the building, where it was distributed to the plants riods of the two years. In terms of temperature, 2018 was initially through the irrigation system. The reduced energy requirement due to slightly colder, but March onwards, the temperature inside the urban less volume of water being used in the RR and RC management was agriculture laboratory was similar for both years. Regarding outdoor enough to offset the energy required for the additional pumping during temperatures, 2019 was slightly colder until April, after which there recirculation, resulting in overall reduction in the global warming cate- were no significant differences in temperature. The internal relative gory of 28% and 19% for RR and RC, respectively, compared with OP. humidity was higher in 2018. The external relative humidity was very Since all three schemes used water from the same rainwater harvesting similar in both years. A summary of the temperature, relative humidity, system, there were no energy savings associated with fewer water treat- vapor pressure deficit, and radiation for both years can be found in ment requirements for RR and RC. Pesticides, substrates and nursery supplementary information - Appendix D. plants represented less than 5% of the impact in this analysis. The main factor was the substrate, and its impacts were associated with transport 3.2. Nutrient balance from its production site. The impact from pesticides can be explained by the implementation of integrated pest control. This type of control re- Nutrient balance calculations were performed for N, P, and K, and duces to a minimum the application of chemical products for pest and dis- the results are shown in Table 4. The calculations based on measured ease control. As mentioned above, the greenhouse is connected to the concentrations were able to total account for 77 to 84% of N, 59 to 69% building, the application of chemicals is very restricted. Organic products of P, and 86 to 92% of K. The remaining amounts of nutrients are as- are in low concentrations and are used to control pests and diseases. In sumed to be accumulated in the perlite bags, since the conditions this sense, the risks of generating an impact are minimum, both for health were similar to Sanjuan-Delmás et al. (2020) for the same substrate and for the environment. (with values of 5% of N, 6% of P, and K values were marginal), and the The infrastructure category includes the greenhouse structure, rest of the values are possibly attributable to dissipative losses. Despite RWHS, and auxiliary equipment, which are all applicable to all three a reduction in nutrients supplied in the RC and RR systems, all three sys- systems, and consequently, all three have similar impacts in all catego- tems showed similar assimilation rates of N in the tomatoes, ranging be- ries. Differences were only due to the small variability in the yield. tween 22 and 24%, indicating that the recirculation schemes did not In particular, it is important to mention that for the RWHS, the items cause an insufficient nutrient supply. K accounted for more in the RC that presented the highest impacts in all categories were the production and RR systems (20 and 22%, respectively) than in the OP management of the glass fiber tank and injection molding. Auxiliary equipment (18%). The results were opposite for P assimilation: RC and RR assimi- exerted high relative impacts on terrestrial acidification, freshwater eu- lated less P (10 and 11%, respectively) than OP (15%). In terms of nitro- trophication, and ecotoxicity. These impacts are associated with the use gen in the tomatoes, all management practices showed similar values, of aluminum and injection molding. ranging from 22% to 24% (Table 4). In terms of phosphorus accounting in the biomass, there was a difference of 12% between OP and RC (36 4. Discussion and 24%). Potassium accounted in biomass present a maximum differ- ence between RR and OP (40% and 30%, respectively). This section discusses the results in light of previous literature, focusing first on the effect of water management on the obtained yields, 3.3. Life cycle assessment in addition to the influence of other variables such as greenhouse materials. Second, the effect of water and nutrient management was Life cycle assessment was performed on all three irrigation manage- examined on environmental performance through life cycle analysis ment practices, and the results were disaggregated into the operational and how the various schemes affected the nutrient-uptake capacity of phase (use of fertilizers, substrate, pesticides, and nursery plants) and in- the crop. Last, through a sensitivity analysis, the following question frastructure (auxiliary equipment, greenhouse structure, and RWHS), as was answered: what is the minimum yield that still provides environ- shown in Fig. 4. In terms of the operational stage, freshwater and marine mental benefits? Finally, we analyze the potential optimization of the eutrophication impacts were reduced by 59% and 98%, respectively, in the different elements within the operational subsystem and provide rec- RC management due to the avoided leaching of nitrogen and phosphorus. ommendations for greenhouse infrastructure. In general, energy and fertilizer use were the highest contributors in all impact categories, ranging from 35% in fossil resource scarcity to 99% in 4.1. Effect of water management on crop yield marine eutrophication for the OP management. The energy used within the RWHS system was to pump water from the 100 m3 tank buried un- Water reduction and recirculation did not limit productivity of the derground to the 2 containers (300 L each one) inside the greenhouse tomato plants. The yields obtained in the configurations with Table 4 Mass balance of nutrients by management. Nutrients Input [kg] Output [kg] Input/output [%] Total Nutrient-uptake Nutrient-Use efficiency Efficiency Irrigation Leachates Air Biomass Tomato Leachates/Other Air Biomass Tomato [%] emission water uses emission N 6.38 1.39 0.05 2.34 1.55 22 0.8 37 24 84 0.61 162 2018 P 2.34 0.42 0.85 0.36 18 – 36 15 69 0.52 441 OP K 16.72 6.57 4.96 2.93 39 – 30 18 87 0.47 62 N 2.5 0.41 0.02 0.95 0.55 16 0.8 38 22 77 0.60 221 2019 P 1.27 0.32 0.31 0.13 25 – 24 10 59 0.35 435 RC K 5.73 1.72 2.07 1.14 30 – 36 20 86 0.56 96 N 2.08 0.35 0.02 0.87 0.5 17 0.8 42 24 84 0.66 239 2019 P 1.06 0.27 0.28 0.11 25 – 27 11 63 0.32 470 RR K 4.75 1.45 1.89 1.05 30 – 40 22 92 0.67 105 *Absorbed macronutrient (N-P-K) was determined from the biomass and tomatoes. We extend the equation of Albornoz et al. (2020) for nitrogen for all macronutrients to determine uptake efficiency and use efficiency. The last two columns are as follows: nutrient-uptake efficiency determined as nutrients absorbed by the plant [kg]/nutrients supplied through irri- gation [kg]; and nutrient-use efficiency calculated as the mass of tomatoes harvested [kg]/nutrients supplied from planting until harvest [kg]. 6 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 Fig. 4. Environmental performance per kg of tomato crop. Open management (OP), recirculated management control and reduction (RC and RR). Global warming (GW), terrestrial acidification (TA), freshwater eutrophication (FE), marine eutrophication (ME), fossil resource scarcity (FRS) and ecotoxicity (ET). The numerical data are presented in supplementary information - Appendix A and B. recirculation (RC) and reduction (RR) were similar and even slightly research conducted on tomato crops in the same climate (Muñoz et al., higher than the yields obtained in open management (OP), which was 2008) has shown lower WUEs than those in this study, ranging between irrigated conventionally for the tomato variety under study. To analyze 30.2 and 36.2 L·kg−1, approximately 20 L·kg−1 less in comparison with water consumption further, the WUE was calculated as indicator, which our results, indicating that there is still margin in terms of water reduc- indicates water consumed per biomass accumulation and was calcu- tion. To ascertain to what degree irrigation can be reduced without af- lated. The RR strategy had the lowest WUE because it used 13% less fecting the yield, some considerations must be taken into account. It is water, while its yield was slightly lower than that of RC (6.3% lower). important to understand that the intensified salinization of the sub- The WUE values were similar to those found in previous studies, strate due to a decrease in applied water can reduce the accumulation such as Chen et al. (2018), who presented an experiment of irrigation/ of biomass due to the increase in osmotic potential in the substrate so- aeration levels in a solar greenhouse with tomato production (in soil) lution. This can be alleviated through irrigation management that favors in a semiarid region, obtaining a WUE from 35.7 L·kg−1 to 65.3 L·kg1. the removal of excess salts at the same time that water is reincorporated Furthermore, our results confirm previous studies that found that limit- into the substrate. It is also important to maintain the matrix potential ing the water supply actually improved the nutrient metabolism of the of the substrate, defined as the force with which water is held by parti- plant. For example, previous studies (Favati et al., 2009; Wang et al., cles and pore space (Yadvinder-Singh et al., 2014). In our situation, to- 2019) found that reducing irrigation applied to tomato crops resulted matoes have been estimated to be −10 and − 40 kPa (Baudoin et al., in better water-use efficiency and that appropriate deficit irrigation 2017; Buttaro et al., 2015); if water is reduced to the point where the can improve fruit quality in terms of nutritional characteristics. Zhang matrix potential is below this value, the assimilation of CO2 by the et al. (2017) also found that reducing irrigation to 80% of evapotranspi- plant is reduced by closing the stomata as a defense mechanism. ration did not reduce yield. Another strategy to optimize water is to vary the distribution of irriga- The question that arises is to what degree can water be reduced tion throughout the day. In our experiments, ten irrigations were applied while maintaining yield in this hydroponic urban agricultural setup? A per day, increasing the quantity towards solar zenith, and reducing it in drastic reduction in irrigation can have detrimental effects concerning the afternoon (in accordance with water demand). Other strategies are the crop's final yield. Therefore, it is necessary to reach a balance that re- worth exploring, such as more frequent irrigation times with less volume duces water inputs while maintaining satisfactory production. Previous or increasing irrigation to favor more leaching and water reuse. The aim of 7 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 the latter strategy is to promote transpiration at times of increased water to be 45% for an entire year. It is relevant to emphasize the importance demand, to avoid physiological limitations, and favor adaptive behavior of of designing rooftop greenhouses to maximize radiation further. the crop, to minimal hydric and nutritional requirements (Li et al., 2017; Madrid et al., 2009; Ullah et al., 2017). Water demand models, such as the 4.2. Effect of water and nutrient management on environmental and Penman-Monteith fixed equation (Gong et al., 2019; Qiu et al., 2013), can agronomical performance also aid in more precisely determining the amount of water to apply. However, these models need to be adapted to urban agricultural technol- There are two main environmental improvements derived from the ogies to be applied to these systems. application of recirculation strategies. First, the release of nutrients to In addition to water, yield is also influenced by temperature and the aquatic environment is minimized, thus considerably reducing the radiation. Both years of our experiments had very similar radiation, contribution to freshwater and marine eutrophication. Second, the nu- relative humidity, wind speed, and temperature. Consequently, evapo- trients that are recovered through the recirculation system are used transpiration rates were comparable; therefore, the impact of climatic again instantly, maximizing the efficiency of the use of resources and variables on the yield in the various schemes was reviewed. However, avoiding the additional impact generated by new fertilizers and their it is necessary to emphasize the importance of selecting greenhouse transport. Considering that some macronutrients, such as P, are nonre- materials that allow a high transmittance rate to obtain optimal produc- newable and have negligible recycling rates (Villalba et al., 2008), im- tive conditions. The i-RTG is composed of polycarbonate (with a lifespan proving use efficiency is critical in current nutrient-intensive of 10 years) that allows 88% theoretical transmissivity (Model Marlon agriculture. The reduction of nutrients has resulted in the most signifi- CS - Brett Martin), is resistant to impacts and has an intermediate cant life cycle benefits. These benefits are appreciated in the reduction level of insulation. While other materials with higher transmittance of marine and freshwater eutrophication, as well as the reduction of en- may be employed, it is important to consider lifespan and resistance be- ergy required for fertigation, as shown in Fig. 5. OP presents a greater cause that will directly affect the life cycle impacts of the infrastructure impact in terms of energy for all impact categories. The remaining (Parajuli et al., 2021). Muñoz-Liesa et al. (2021) suggested that both items (substrates, fertilizers, pesticides and nursery plants) in the glass and glass films have a similar transmittance of 90% and are envi- three management systems have similar impacts. ronmentally better than polycarbonate. Glass has a long lifespan (15 Other research has shown similar results concerning the impact of years) in contrast to film (3–5 years) (Antón et al., 2012). However, fertilizers during the operational stage. Muñoz et al. (2017) confirmed the glass is rigid and has a heavy weight (1400 g·m−2), requiring a our findings that recirculation systems can reduce impacts. Similarly, greater structure to support it, in contrast to plastic film with greater Rufí-Salís et al. (2020a) showed how open systems have a high impact flexibility and lower weight (230 g·m−2) (Castilla Prados, 2004). The in the freshwater and marine eutrophication categories, with 90% for i-RTG setup had a satisfactory balance between transmittance, flexibil- both impact categories, due to the leachate emission of phosphorus ity, weight, and environmental impacts. and nitrogen to the environment. In addition to the transmissivity of the material, it is also important In addition to reducing nutrients through recirculation and reduction, to consider the effective radiation (interior radiation/external radiation) impacts can be further minimized by substituting chemical fertilizers, that reaches the plants. In our case, the structure and configuration of such as calcium nitrate and potassium sulfate. Determining appropriate our greenhouse generates shadows and opaque walls, which reduce substitutes requires understanding the various combinations of NPK. For the amount of radiation inside the greenhouse. To compensate for the example, the use potassium nitrate (with an N:P:K ratio of 13-0-45) in- shadows inside the greenhouse and to maximize light throughout the stead of potassium sulfate (0-0-50) to deliver potassium to the plants, day, the rows of tomato plants have a north-south orientation. Even the first fertilizer also provides nitrogen in the form of nitrate, which so, the maximum effective radiation in the i-RTG has been estimated could highly affect the nutrient dynamics at the crop, plant and substrate Fig. 5. Life cycle impacts of the Recirculation and Recovery (RR), Recirculation (RC), and Open (OP) systems during the operational stage. 8 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 levels. In this way, fertigation schemes that constantly evaluate the nutri- efficient control of humidity and irrigation (Zotarelli et al., 2009), and ent dynamics of the crop and that also consider environmental aspects through nutrient availability sensors, the management of both irrigation can be highly effective in reducing the impact of agriculture. water and recirculation water quality can be better adjusted. To evaluate the agronomical performance of the various schemes, the nutrient uptake and use efficiency was analyzed, which are shown 4.3. Effect of radiation on yield and life cycle impacts in Table 4. The nitrogen-uptake efficiency, defined as the nitrogen absorbed to the nitrogen supplied to the irrigation system, was similar Since the environmental impacts determined through LCA are de- in the three treatments, ranging between 0.60 and 0.66. This is an im- pendent directly and linearly on the yield, a logical next step is to deter- portant finding of the experiments: the recirculation and reduction mine the minimum yield that still provides environmental benefits. schemes of nutrients and water did not affect the nutrient-uptake ca- First, to need consider the variability of yield related to the radiation re- pacity of the crop. Additionally, the nitrogen-use efficiency calculated ceived by the crop. To do so, the model proposed by Montero et al. as the mass of total tomatoes harvested (in kg) divided by the mass of (2017) was adapted to obtain a theoretical yield based on the potential nitrogen supplied to all tomato plants for the entire experimental radiation range during a standard growth period (195 days) and the cycle (kg) was lowest for the OP system, further indicating that the re- radiation-use efficiency of the tomato plant of 8.77 g·MJ−1 (Montero circulation and reduction schemes resulted in higher nitrogen-use effi- et al., 2017). Radiation data were obtained from a nearby weather sta- ciency. Other studies have also shown that reductions in supplied tion 8 km northeast (Ruralcat, 2019) from the laboratory for eleven con- nitrogen have increased the efficiency of the use of this resource (Min secutive years (2009 to 2019). During the standard growth period, the et al., 2011), and the RC and RR management practices present 37% - maximum and minimum radiation values obtained were 3541 and 48% more efficiency than OP. In this sense, this value can be explained 3904 MJ (accumulated per season), respectively; therefore, obtaining by the effect of the recirculation system, which reduces the amount of the minimum and maximum theoretical yield ranges. nitrogen applied, causing the biomass-nutrient ratio to increase, in con- Next, based on the LCA results of the RC treatment, the life cycle inven- trast to the nutrient-uptake efficiency. tory was adjusted to determine the LCA for both the minimum and max- In contrast, phosphorus-uptake efficiency showed a reduction in the imum yield that could occur due to radiation variability. The amount of recirculated managements compared with the OP system, from 0.52 water used for each theoretical yield was adjusted by means of the (OP) to 0.35 (RR) kg of P absorbed per kg supplied. One potential expla- WUE (water potentially consumed = yield [kg]/WUE [L·kg−1]). Based nation is that due to recirculation, sulfate ions accumulate over time, cre- on the water used, the new impacts of energy and RWHS were estimated. ating high concentrations of sulfate anions that compete with phosphates The fertilizers varied proportionally to the variation in yield (for a more for root uptake, as has been seen in other studies (Marcelis and Heuvelink, detailed review, see supplementary information - Appendix E and F). 2019; Pardossi et al., 2002). Aulakh and Pasricha (1977) presented a test Fig. 6 shows that impacts increased by approximately 8% when the of nutrient assimilation rates and different concentrations of phosphorus minimum yield was considered. The highest increase was observed in cu- and sulfur in Phaseolus aureus L. They mentioned that an antagonistic ef- mulative energy demand (9.7%), and the lowest was observed in fect between these ions could be explained by competition at the root ab- ecotoxicity (5.1%). In contrast, when analyzing the maximum yield, im- sorption sites or for the same uptake pathways. This could explain the pacts were only reduced on average by a value close to 1%, even with pos- lower rate of phosphorus assimilation in the recirculation treatments. itive values (terrestrial acidification). The greatest reduction was seen in However, in terms of nutritional value, phosphorus-use efficiency was ecotoxicity (4.6%). Since water-use efficiency presents an average value comparable among all three treatments, with a range from 441 to higher than the one obtained within the present study, although there 470 kg of tomatoes produced per kilogram of phosphorus applied. Re- is a rational use, productivity (yield increase) tends to be above efficiency; garding potassium, both uptake and use efficiency showed the same in- this situation generates an overestimation in water consumption, which creasing trend from OP to RC to RR. The highest value observed in RR translates into a more significant impact. The theoretical maximum per- could be explained by an increase in potassium retention in the tissues formance obtained with the model is 11% higher than the yield obtained of a plant associated with a water deficit (De Luca et al., 2021). in RC. Despite the yield increase, environmental performance does not Analytical validation is required to utilize the full potential of fertiliza- show improvements due to the increase in the operational values. tion management strategies by improving nutrient retention through di- lution or increasing the concentration if required. The application of 5. Final remarks and conclusions sensors would allow fertilization management to be better adjusted to the needs of the crop while maximizing the efficiency of irrigation strate- Our main conclusion is that the implementation of water and gies. Through the measurement of moisture in the substrate to have more nutrient recirculation and reduction in the i-RTG for tomato plants did Fig. 6. Sensitivity analysis with the minimum and maximum theoretical yields. 9 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 not affect the yield, while it minimized eutrophication impacts related Acknowledgments to nutrient discharge and increased water-use efficiency. The methodol- ogy applied has proven to be consistent for the environmental analysis We would like to acknowledge the following funding sources for this of an i-RTG as in other research (Rufí-Salís et al., 2020b; Sanjuan- work: the European Commission's ERC Consolidator grant 818002- Delmás et al., 2018). It may be replicable for different types of URBAG led by G. Villalba and the National Commission for Scientific production. and Technological Research (Chile) for the grant awarded to F. Parada This work provides information on the environmental impacts asso- [grant number PFCHA-CONICYT 2017 – Folio 72180248]. ciated to food production and how they can be solved through the im- plementation of the recirculation system in an i-RTG. This kind of Appendix A. Supplementary data system allows the reuse leachates and nutrient recovery, in contrast open systems, is usually extensive, irrigation control is more complex Supplementary data to this article can be found online at https://doi. and less efficient. Additionally, the characteristics of the substrate in org/10.1016/j.scitotenv.2021.148689. soilless systems are homogeneous in all crop bags used in the green- house, unlike what occurs in open systems where the soil can present References high spatial variability in the physical and chemical characteristics. In Albornoz, F., Nario, A., Saavedra, M., Videla, X., 2020. Rootstock x environment interac- this sense, fertilization plans adjusted by mass balance, which consider tions on nitrogen-use efficiency in grafted tomato plants at different phenological the availability of soil nutrients, together with the removal of the crop, stages. Agronomy 10, 350. https://doi.org/10.3390/agronomy10030350. are an important tool for managing the impacts of open systems. Antón, A., Torrellas, M., Montero, J.I., Ruijs, M., Vermeulen, P., Stanghellini, C., 2012. Envi- ronmental impact assessment of Dutch tomato crop production in a Venlo glass- In terms of water-use efficiency, RR management was the most effi- house. Acta Hortic., 781–791 https://doi.org/10.17660/ActaHortic.2012.927.97. cient, requiring 48.7 L per kg of harvested tomato, followed by RC (52.4 Arcas-Pilz, V., Rufí-Salís, M., Parada, F., Gabarrell, X., Villalba, G., 2021. Assessing the envi- L·kg1) and OP (75.2·L·kg-1). Among recirculation management prac- ronmental behavior of alternative fertigation methods in soilless systems: the case of tices, irrigation reduction (RR) presented an improvement of 7% in Phaseolus vulgaris with struvite and rhizobia inoculation. Sci. Total Environ. 770, 144744. https://doi.org/10.1016/j.scitotenv.2020.144744. water-use efficiency. In terms of environmental performance, RC Aulakh, M.S., Pasricha, N.S., 1977. Interaction effect of sulphur and phosphorus on growth shows the best performance in almost all impact categories during the and nutrient content of moong (Phaseolus aureus L.). Plant Soil 47, 341–350. https:// operational phase, especially in marine and freshwater eutrophication, doi.org/10.1007/BF00011493. Baudoin, W., Nersisyan, A., Shamilov, A., Hodder, A., Gutierrez, D., De Pascale, S., Nicola, S., with 44% and 93% fewer impacts. For the infrastructure phase, the re- Gruda, N., Urban, L., Tanny, J., 2017. Good Agricultural Practices for Greenhouse Veg- placement of materials such as aluminum with lower impact recycled etable Production in the South East European Countries - Principles for Sustainable plastics. Intensification of Smallholder Farms. This study can support the decision-making process of the design of Boulard, T., Raeppel, C., Brun, R., Lecompte, F., Hayer, F., Carmassi, G., Gaillard, G., 2011. En- vironmental impact of greenhouse tomato production in France. Agron. Sustain. Dev. agricultural projects in the city, through yields and water consumption 31, 757. https://doi.org/10.1007/s13593-011-0031-3. obtained per square meter, in order to have a basis to contrast for urban Buttaro, D., Santamaria, P., Signore, A., Cantore, V., Boari, F., Montesano, F.F., Parente, A., agriculture projects. Similarly, to have more information about the envi- 2015. Irrigation management of greenhouse tomato and cucumber using tensiome- ter: effects on yield, quality and water use. Agric. Agric. Sci. Procedia 4, 440–444. ronmental impacts generated within a recirculating crop, in order to https://doi.org/10.1016/j.aaspro.2015.03.050 Efficient irrigation management and compare other UA typologies. As well as can help the activity of incen- its effects in urban and rural landscapes. tivizing good practices aligned with the sustainability of urban agricul- Caldeira, C., De Laurentiis, V., Corrado, S., van Holsteijn, F., Sala, S., 2019. Quantification of ture. At the domestic or private level, the idea of promoting the use of food waste per product group along the food supply chain in the European Union: a mass flow analysis. Resour. Conserv. Recycl. 149, 479–488. https://doi.org/10.1016/j. closed production systems, in order to reduce nutrient emissions to resconrec.2019.06.011. the environment, and their derived detrimental effects, can be Castilla Prados, N., 2004. Cápitulo 4: El invernadero de plástico, in: Invernaderos de highlighted. Plástico. Ediciones Mundi-Prensa. Chen, Y., Wen, X., Wang, B., Nie, P., 2017. Agricultural pollution and regulation: how to For further research it is necessary to consider: 1) developing ways subsidize agriculture? J. Clean. Prod. 164, 258–264. https://doi.org/10.1016/j. to optimize irrigation distribution which ultimately reduces overall jclepro.2017.06.216. water and fertilizer consumption because uptake efficiency is im- Chen, H., Hou, H., Hu, H., Shang, Z., Zhu, Y., Cai, H., Qaisar, S., 2018. Aeration of different proved; 2) choosing highly transmissive materials while safeguarding irrigation levels affects net global warming potential and carbon footprint for green- house tomato systems. Sci. Hortic. 242, 10–19. https://doi.org/10.1016/j. low life cycle impacts and long lifetimes; and 3) finding ways to ensure scienta.2018.07.021. high effective radiation for greenhouses, especially those that are ac- Chen, Y., Hu, S., Guo, Z., Cui, T., Zhang, L., Lu, C., Yu, Y., Luo, Z., Fu, H., Jin, Y., 2021. Effect of commodated to already existing buildings that were not originally de- balanced nutrient fertilizer: a case study in Pinggu District, Beijing, China. Sci. Total Environ. 754, 142069. https://doi.org/10.1016/j.scitotenv.2020.142069. signed to maximize radiation on the entire surface of the rooftop. Contreras, J.I., Alonso, F., Cánovas, G., Baeza, R., 2017. Irrigation management of green- house zucchini with different soil matric potential level. Agronomic and environmen- CRediT authorship contribution statement tal effects. Agric. Water Manag. 183, 26–34. https://doi.org/10.1016/j. agwat.2016.09.025 Special Issue: Advances on ICTs for Water Management in Agriculture. Felipe Parada: Conceptualization, Methodology, Formal analysis, In- Cook, E., European Commission, Statistical Office of the European Union, 2018. Agricul- vestigation, Data curation, Visualization, Writing – original draft. Xavier ture, Forestry and Fishery Statistics: 2018 Edition. Gabarrell: Conceptualization, Methodology, Writing – review & editing, De Corato, U., 2020. Agricultural waste recycling in horticultural intensive farming sys- tems by on-farm composting and compost-based tea application improves soil qual- Supervision, Project administration, Funding acquisition. Martí Rufí- ity and plant health: a review under the perspective of a circular economy. Sci. Total Salís: Methodology, Data curation, Visualization, Writing – review & Environ. 738, 139840. https://doi.org/10.1016/j.scitotenv.2020.139840. editing. Verónica Arcas-Pilz: Methodology, Investigation, Writing – De Klein, C., Novoa, R.S.A., Ogle, S., Smith, K., Rochette, P., Wirth, T.C., Mcconkey, B.G., Mosier, A., Rypdal, K., 2006. N2O emissions from managed soils, and CO2 emissions review & editing. Pere Muñoz: Conceptualization, Methodology, from lime and urea application. 2006 IPCC Guidel. Natl. Greenh. Gas Invent, p. 4. Writing – review & editing, Supervision. Gara Villalba: Conceptualiza- De Luca, A., Corell, M., Chivet, M., Parrado, M.A., Pardo, J.M., Leidi, E.O., 2021. Reassessing tion, Methodology, Writing – review & editing, Supervision, Funding the role of potassium in tomato grown with water shortages. Horticulturae 7, 20. acquisition. https://doi.org/10.3390/horticulturae7020020. European Commission, 2011. Food: From Farm to Fork Statistics, 2011 Ed. Ed, Eurostat Pocketbooks. Publications Office of the European Union, Luxemburg. Declaration of competing interest FAOSTAT, 2017a. Food and Agriculture Organization of the United Nations FAO, Statistical Database. Fertilizers by Nutrient. [WWW Document]. URL. http://www.fao.org/ The authors declare that they have no known competing financial faostat/en/#data/RFN (accessed 5.12.20). FAOSTAT, 2017b. Food and Agriculture Organization of the United Nations FAO, Statistical interests or personal relationships that could have appeared to influ- Database. Crops. Tomatoe. [WWW Document]. URL. http://www.fao.org/faostat/en/ ence the work reported in this paper. #data/RFN (accessed 5.12.20). 10 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 Favati, F., Lovelli, S., Galgano, F., Miccolis, V., Di Tommaso, T., Candido, V., 2009. Processing Pardossi, A., Malorgio, F., Incrocci, L., Campiotti, C.A., Tognoni, F., 2002. A comparison be- tomato quality as affected by irrigation scheduling. Sci. Hortic. 122, 562–571. https:// tween two methods to control nutrient delivery to greenhouse melons grown in doi.org/10.1016/j.scienta.2009.06.026. recirculating nutrient solution culture. Sci. Hortic. 92, 89–95. https://doi.org/ Gong, X., Liu, H., Sun, J., Gao, Y., Zhang, H., 2019. Comparison of Shuttleworth-Wallace 10.1016/S0304-4238(01)00292-8. model and dual crop coefficient method for estimating evapotranspiration of tomato Payen, S., Basset-Mens, C., Perret, S., 2015. LCA of local and imported tomato: an energy cultivated in a solar greenhouse. Agric. Water Manag. 217, 141–153. https://doi.org/ and water trade-off. J. Clean. Prod. 87, 139–148. https://doi.org/10.1016/j. 10.1016/j.agwat.2019.02.012. jclepro.2014.10.007. Green, T.R., Yu, Q., Ma, L., Wang, T.-D., 2010. Crop water use efficiency at multiple scales. Piezer, K., Petit-Boix, A., Sanjuan-Delmás, D., Briese, E., Celik, I., Rieradevall, J., Gabarrell, X., Agric. Water Manag. 97, 1099–1101. https://doi.org/10.1016/j.agwat.2010.03.018 Josa, A., Apul, D., 2019. Ecological network analysis of growing tomatoes in an urban Crop water use efficiency at multiple scales. rooftop greenhouse. Sci. Total Environ. 651, 1495–1504. https://doi.org/10.1016/j. He, X., Qiao, Y., Liu, Y., Dendler, L., Yin, C., Martin, F., 2016. Environmental impact as- scitotenv.2018.09.293. sessment of organic and conventional tomato production in urban greenhouses Pineda, I.T., Lee, Y.D., Kim, Y.S., Lee, S.M., Park, K.S., 2020. Review of inventory data in life of Beijing city. China. J. Clean. Prod. 134, 251–258. https://doi.org/10.1016/j. cycle assessment applied in production of fresh tomato in greenhouse. J. Clean. Prod. jclepro.2015.12.004 Special Volume: Transitions to Sustainable Consumption 124395. https://doi.org/10.1016/j.jclepro.2020.124395. and Production in Cities. Qiu, R., Song, J., Du, T., Kang, S., Tong, L., Chen, R., Wu, L., 2013. Response of evapotranspi- Hooshmand, M., Albaji, M., Boroomand Nasab, S., Alam Zadeh Ansari, N., 2019. The effect ration and yield to planting density of solar greenhouse grown tomato in northwest of deficit irrigation on yield and yield components of greenhouse tomato (Solanum China. Agric. Water Manag. 130, 44–51. https://doi.org/10.1016/j.agwat.2013.08.013. lycopersicum) in hydroponic culture in Ahvaz region, Iran. Sci. Hortic. 254, 84–90. Rittmann, B.E., Mayer, B., Westerhoff, P., Edwards, M., 2011. Capturing the lost phospho- https://doi.org/10.1016/j.scienta.2019.04.084. rus. Chemosphere. The Phosphorus Cycle 84, 846–853. https://doi.org/10.1016/j. Huijbregts, M.A.J., Steinmann, Z.J.N., Elshout, P.M.F., Stam, G., Verones, F., Vieira, M., Zijp, chemosphere.2011.02.001. M., Hollander, A., van Zelm, R., 2017. ReCiPe2016: a harmonised life cycle impact as- Rufí-Salís, M., Parada, F., Arcas-Pilz, V., Petit-Boix, A., Villalba, G., Gabarrell, X., 2020a. sessment method at midpoint and endpoint level. Int. J. Life Cycle Assess. 22, Closed-loop crop cascade to optimize nutrient flows and grow low-impact 138–147. https://doi.org/10.1007/s11367-016-1246-y. vegetables in cities. Front. Plant Sci. 11. https://doi.org/10.3389/ ISO 14040, 2006. ISO 14040: Environmental Management - Life Cycle Assessment - Prin- fpls.2020.596550. ciples and Framework. International Standard Organization. Rufí-Salís, M., Petit-Boix, A., Villalba, G., Ercilla-Montserrat, M., Sanjuan-Delmás, D., Li, Y., Sun, Y., Liao, S., Zou, G., Zhao, T., Chen, Y., Yang, J., Zhang, L., 2017. Effects of two Parada, F., Arcas, V., Muñoz-Liesa, J., Gabarrell, X., 2020b. Identifying eco-efficient slow-release nitrogen fertilizers and irrigation on yield, quality, and water-fertilizer year-round crop combinations for rooftop greenhouse agriculture. Int. J. Life Cycle As- productivity of greenhouse tomato. Agric. Water Manag. 186, 139–146. https://doi. sess. https://doi.org/10.1007/s11367-019-01724-5. org/10.1016/j.agwat.2017.02.006. Ruralcat, 2019. Weather data - Ruralcat [WWW document]. URL. https://ruralcat.gencat.cat/ Liu, H., Li, H., Ning, H., Zhang, X., Li, S., Pang, J., Wang, G., Sun, J., 2019. Optimizing irrigation web/guest/agrometeo.estacions?p_auth=OKK5BXY6&p_p_id=AgrometeoEstacions_ frequency and amount to balance yield, fruit quality and water use efficiency of WAR_AgrometeoEstacions100SNAPSHOT&p_p_lifecycle=1&p_p_state=normal&p_p_ greenhouse tomato. Agric. Water Manag. 226, 105787. https://doi.org/10.1016/j. mode=view&p_p_col_id=column-2&p_p_col_pos=1&p_p_col_count=2&_Agrometeo agwat.2019.105787. Estacions_WAR_AgrometeoEstacions100SNAPSHOT_action=goEstacion. Llorach-Massana, P., Muñoz, P., Riera, M.R., Gabarrell, X., Rieradevall, J., Montero, J.I., Sanjuan-Delmás, D., Llorach-Massana, P., Nadal, A., Ercilla-Montserrat, M., Muñoz, P., Villalba, G., 2017. N2O emissions from protected soilless crops for more precise Montero, J.I., Josa, A., Gabarrell, X., Rieradevall, J., 2018. Environmental assessment food and urban agriculture life cycle assessments. J. Clean. Prod. 149, 1118–1126. of an integrated rooftop greenhouse for food production in cities. J. Clean. Prod. https://doi.org/10.1016/j.jclepro.2017.02.191. 177, 326–337. https://doi.org/10.1016/j.jclepro.2017.12.147. Madrid, R., Barba, E.M., Sánchez, A., García, A.L., 2009. Effects of organic fertilisers and ir- Sanjuan-Delmás, D., Josa, A., Muñoz, P., Gassó, S., Rieradevall, J., Gabarrell, X., 2020. Apply- rigation level on physical and chemical quality of industrial tomato fruit (cv. Nauti- ing nutrient dynamics to adjust the nutrient-water balance in hydroponic crops. A lus). J. Sci. Food Agric. 89, 2608–2615. https://doi.org/10.1002/jsfa.3763. case study with open hydroponic tomato crops from Barcelona. Sci. Hortic. 261, Marcelis, L., Heuvelink, E., 2019. Achieving Sustainable Greenhouse Cultivation. Burleigh 108908. https://doi.org/10.1016/j.scienta.2019.108908. Dodds Science Publishing Limited https://doi.org/10.1201/9780429266744. Sanyé-Mengual, E., Cerón-Palma, I., Oliver-Solà, J., Montero, J.I., Rieradevall, J., 2013. Envi- Martínez-Blanco, J., Muñoz, P., Antón, A., Rieradevall, J., 2011. Assessment of tomato Med- ronmental analysis of the logistics of agricultural products from roof top greenhouses iterranean production in open-field and standard multi-tunnel greenhouse, with in Mediterranean urban areas. J. Sci. Food Agric. 93, 100–109. https://doi.org/ compost or mineral fertilizers, from an agricultural and environmental standpoint. 10.1002/jsfa.5736. J. Clean. Prod. 19, 985–997. https://doi.org/10.1016/j.jclepro.2010.11.018. Sanyé-Mengual, E., Oliver-Solà, J., Montero, J.I., Rieradevall, J., 2015. An environmental and Mason, B., Rufí-Salís, M., Parada, F., Gabarrell, X., Gruden, C., 2019. Intelligent urban irriga- economic life cycle assessment of rooftop greenhouse (RTG) implementation in Bar- tion systems: saving water and maintaining crop yields. Agric. Water Manag. 226, celona, Spain. Assessing new forms of urban agriculture from the greenhouse struc- 105812. https://doi.org/10.1016/j.agwat.2019.105812. ture to the final product level. Int. J. Life Cycle Assess. 20, 350–366. https://doi.org/ Min, J., Zhao, X., Shi, W.-M., Xing, G.-X., Zhu, Z.-L., 2011. Nitrogen balance and loss in a 10.1007/s11367-014-0836-9. greenhouse vegetable system in southeastern China. Pedosphere 21, 464–472. Schmidhuber, J., Tubiello, F.N., 2007. Global food security under climate change. Proc. Natl. https://doi.org/10.1016/S1002-0160(11)60148-3. Acad. Sci. 104, 19703–19708. https://doi.org/10.1073/pnas.0701976104. Montero, J.I., Antón, A., Torrellas, M., Ruijs, M.N.A., Vermeulen, P.C.M., 2011. Environmen- Shu, L., Schneider, P., Jegatheesan, V., Johnson, J., 2006. An economic evaluation of phos- tal and Economic Profile of Present Greenhouse Production Systems in Europe - phorus recovery as struvite from digester supernatant. Bioresour. Technol. 97, EUPHOROS Deliverable N 5 Final Report (No. Deliverable 5 Final Report). European 2211–2216. https://doi.org/10.1016/j.biortech.2005.11.005. Commission, Cabrils/Wageningen. SIAR, 2019. Sistema de información Agroclimática para el Regadío. [WWW Document]. Montero, J.I., Baeza, E., Heuvelink, E., Rieradevall, J., Muñoz, P., Ercilla, M., Stanghellini, C., 2017. Productivity of a building-integrated roof top greenhouse in a Mediterranean URL. http://eportal.mapama.gob.es (accessed 11.21.19). climate. Agric. Syst. 158, 14–22. https://doi.org/10.1016/j.agsy.2017.08.002. Steen, P., 1998. The most appropriate health care in the right place, and the right Muñoz, P., Antón, A., Paranjpe, A., Ariño, J., Montero, J.I., 2008. High decrease in nitrate time. The importance of clinical decision support criteria in home health care. leaching by lower N input without reducing greenhouse tomato yield. Agron. Sustain. Home Care Provid. 3, 25–27 quiz 28–29. https://doi.org/10.1016/s1084-628x Dev. 28, 489–495. https://doi.org/10.1051/agro:2008024. (98)90152-6. Muñoz, P., Flores, J.S., Antón, A., Montero, J.I., 2017. Combination of greenhouse and open- Tixier, P., de Bon, H., 2006. Urban horticulture. In: Veenhuizen, R. van (Ed.), Cities Farming field crop fertigation can increase sustainability of horticultural crops in the Mediterra- for the Future : Urban Agriculture for Green and Productive Cities. IDRC, Ottawa, ON, nean region. Acta Hortic., 627–634 https://doi.org/10.17660/ActaHortic.2017.1170.78. CA. Muñoz-Liesa, J., Toboso-Chavero, S., Mendoza Beltran, A., Cuerva, E., Gallo, E., Gassó- Ullah, I., Hanping, M., Chuan, Z., Javed, Q., Azeem, A., 2017. Optimization of irrigation and Domingo, S., Josa, A., 2021. Building-integrated agriculture: Are we shifting environ- nutrient concentration based on economic returns, substrate salt accumulation and mental impacts? An environmental assessment and structural improvement of urban water use efficiency for tomato in greenhouse. Arch. Agron. Soil Sci. 63, 1748–1762. greenhouses. Resour. Conserv. Recycl. 169, 105526. https://doi.org/10.1016/j. https://doi.org/10.1080/03650340.2017.1306641. resconrec.2021.105526. Ulm, F., Avelar, D., Hobson, P., Penha-Lopes, G., Dias, T., Máguas, C., Cruz, C., 2019. Sustain- Nadal, A., Llorach-Massana, P., Cuerva, E., López-Capel, E., Montero, J.I., Josa, A., able urban agriculture using compost and an open-pollinated maize variety. J. Clean. Rieradevall, J., Royapoor, M., 2017. Building-integrated rooftop greenhouses: an en- Prod. 212, 622–629. https://doi.org/10.1016/j.jclepro.2018.12.069. ergy and environmental assessment in the mediterranean context. Appl. Energy UNDESA, 2018. Press Release World Urbanization Prospects 2018 - Population Division. 187, 338–351. https://doi.org/10.1016/j.apenergy.2016.11.051. United Nation -Dep. Econ. Soc. Aff. Nagendran, R., 2011. Chapter 24 - agricultural waste and pollution, in: Letcher, T.M., van Schilfgaarde, J., 1994. Irrigation — a blessing or a curse. Agric. Water Manag. 25, Vallero, D.A. (Eds.), Waste. Academic Press, Boston, pp. 341–355. doi:https://doi. 203–219. https://doi.org/10.1016/0378-3774(94)90061-2. org/10.1016/B978-0-12-381475-3.10024-5. Villalba, G., Liu, Y., Schroder, H., Ayres, R.U., 2008. Global phosphorus flows in the indus- Novotny, V., 1999. Diffuse pollution from agriculture — a worldwide outlook. Water Sci. trial economy from a production perspective. J. Ind. Ecol. 12, 557–569. https://doi. Technol. 39, 1–13. https://doi.org/10.1016/S0273-1223(99)00027-X Integrated Man- org/10.1111/j.1530-9290.2008.00050.x. agement of Water Quality: The Role of Agricultural Diffuse Pollution Sources. Wang, X., Yun, J., Shi, P., Li, Z., Li, P., Xing, Y., 2019. Root growth, fruit yield and water use Parajuli, R., Thoma, G., Matlock, M.D., 2019. Environmental sustainability of fruit and veg- efficiency of greenhouse grown tomato under different irrigation regimes and nitro- etable production supply chains in the face of climate change: a review. Sci. Total En- gen levels. J. Plant Growth Regul. 38, 400–415. https://doi.org/10.1007/s00344-018- viron. 650, 2863–2879. https://doi.org/10.1016/j.scitotenv.2018.10.019. 9850-7. Parajuli, R., Matlock, M.D., Thoma, G., 2021. Cradle to grave environmental impact evalu- Wernet, G., Bauer, C., Steubing, B., Reinhard, J., Moreno-Ruiz, E., Weidema, B., 2016. The ation of the consumption of potato and tomato products. Sci. Total Environ. 758, ecoinvent database version 3 (part I): overview and methodology. Int. J. Life Cycle As- 143662. https://doi.org/10.1016/j.scitotenv.2020.143662. sess. 21, 1218–1230. https://doi.org/10.1007/s11367-016-1087-8. 11 F. Parada, X. Gabarrell, M. Rufí-Salís et al. Science of the Total Environment 794 (2021) 148689 Yadvinder-Singh, Kukal, S.S., Jat, M.L., Sidhu, H.S., 2014. Chapter four - improving water Water Productivity to Ensure Food Security under Changing Environments Overseen productivity of wheat-based cropping systems in South Asia for sustained productiv- by: Brent Clothier. ity. In: Sparks, D. (Ed.), Advances in Agronomy. Academic Press, pp. 157–258 https:// Zotarelli, L., Dukes, M.D., Scholberg, J.M.S., Muñoz-Carpena, R., Icerman, J., 2009. Tomato doi.org/10.1016/B978-0-12-800131-8.00004-2. nitrogen accumulation and fertilizer use efficiency on a sandy soil, as affected by ni- Zhang, H., Xiong, Y., Huang, G., Xu, X., Huang, Q., 2017. Effects of water stress on process- trogen rate and irrigation scheduling. Agric. Water Manag. 96, 1247–1258. https:// ing tomatoes yield, quality and water use efficiency with plastic mulched drip irriga- doi.org/10.1016/j.agwat.2009.03.019. tion in sandy soil of the Hetao Irrigation District. Agric. Water Manag. 179, 205–214. https://doi.org/10.1016/j.agwat.2016.07.022 Special Issue on Improving Agricultural 12