ORIGINAL RESEARCH published: 22 March 2019 doi: 10.3389/fpls.2019.00341 Agronomic and Environmental Assessment of a Polyculture Rooftop Soilless Urban Home Garden in a Mediterranean City Anna Boneta 1† , Martí Rufí-Salís 1,2* † , Mireia Ercilla-Montserrat 1 , Xavier Gabarrell 1,2 and Joan Rieradevall 1,2 1 Sostenipra Research Group (2017 SGR 1683), Institut de Ciència i Tecnologia Ambientals (ICTA-UAB), Universitat Autònoma de Barcelona, Barcelona, Spain, 2 Department of Chemical, Biological and Environmental Engineering, Universitat Autònoma de Barcelona, Barcelona, Spain Urban planning has been focusing its attention on urban rooftop agriculture as an innovative way to produce local and reliable food in unused spaces in cities. However, there is a lack of quantitative data on soilless urban home gardens and their contribution to self-sufficiency. The aim of the present study is to provide quantitative agronomic and environmental data on an actual soilless urban garden to estimate its degree Edited by: of self-sufficiency and sustainability. For this purpose, an 18 m2 soilless polyculture Esteban José Baeza, rooftop urban home garden in the city center of Barcelona was analyzed. From 2015 Wageningen University & Research, Netherlands to 2017, 22 different crops were grown to feed 2 people in an open-air soilless system, Reviewed by: and a life cycle assessment was performed. A total productivity of 10.6 kg/m2 /year Chiara Cirillo, was achieved, meaning that 5.3 m2 would be needed to fulfill the yearly vegetable University of Naples Federico II, Italy Baptiste Jean-Paul Grard, requirements of an average citizen (in terms of weight). Considering the vegetable INRA Centre Versailles-Grignon, market basket of Catalonia, an 8.2 m2 soilless garden would be needed to cover 62% France of the market basket for one person. The top 5 most productive crops were tomato, *Correspondence: chard, lettuce, pepper and eggplant, accounting for 85.5% of the total production. The Martí Rufí-Salís Marti.Rufi@uab.cat water consumption was 3.7 L/m2 /day, and 3.3 kg/year/m2 of waste was generated. † These authors have contributed A high degree of self-sufficiency was achieved, although adjustments could be made to equally to this work adapt the production to the market basket. The environmental assessment showed that the fertilizers and their associated leachates accounted for the highest environmental Specialty section: This article was submitted to impacts in all the studied impact categories. Overall, 0.6 kg CO2 eq. was generated Crop and Product Physiology, per kg of vegetables produced. The quantitative data provided by the present study a section of the journal Frontiers in Plant Science offer a reference from which urban planners and researchers can project future Received: 14 July 2018 implementations of rooftop urban agriculture (UA) on a large scale. Accepted: 05 March 2019 Keywords: urban agriculture, soilless, polyculture, rooftop farming, home garden, life cycle assessment Published: 22 March 2019 Citation: Boneta A, Rufí-Salís M, INTRODUCTION Ercilla-Montserrat M, Gabarrell X and Rieradevall J (2019) Agronomic The continuously increasing world population is predicted to rise to almost 10 billion people and Environmental Assessment of a Polyculture Rooftop Soilless Urban by 2050 (United Nations, Department of Economic and Social Affairs, Population Division, Home Garden in a Mediterranean 2017). This situation will lead to a higher food demand and, consequently, increased pressure on City. Front. Plant Sci. 10:341. many ecosystem services (Watson et al., 2011). Moreover, the population living in urban areas doi: 10.3389/fpls.2019.00341 is also expected to increase from 54% (in 2015) to 66% by 2050 (WHO, 2016). Nevertheless, Frontiers in Plant Science | www.frontiersin.org 1 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden in the European Union, 72% of the population already lives in 2012). In terms of production, Sanyé-Mengual et al. (2013) urban areas (Eurostat, 2015). The consequence of this trend is a quantified that 150 tones of tomatoes could be produced in the progressive urban expansion that widens the distances between roof area of Barcelona. Although concerns about community production and consumption areas, increasing the dependence and industrial UA are gaining interest worldwide, home urban of cities on external resources (Scoones, 2009). In this context, gardens are overlooked and understudied (Taylor and Lovell, the necessity of rethinking our food systems is rising to achieve 2014). In addition, the existing literature concerning home urban urban sustainability and avoid intensive agricultural techniques gardens is mainly qualitative and focused on their ecosystem that have critical environmental costs. and social services provisions (Calvet-Mir et al., 2012; Cameron As a result, urban agriculture (UA) is gaining importance et al., 2012) or their contribution to food security (Kortright to facilitate access to healthy, reliable and fresh food, which and Wakefield, 2011) rather than on their agronomic and is usually difficult in cities (e.g., “food deserts”) (WHO, 2016) environmental performance. along with many other social and ecological related services. From an environmental impact perspective, urban food Specifically, urban rooftop farming, which includes gardens, production has been assessed for rooftop greenhouses (Sanjuan- greenhouses or farms placed on building rooftops, can offer Delmás et al., 2018) and community rooftop gardens (Sanyé- new landscape opportunities while restraining the burden on Mengual et al., 2015) by applying the Life Cycle Assessment agricultural land and achieving more sustainable and resilient methodology (ISO 14014). Nevertheless, there is still a gap in cities (Thomaier et al., 2015). Such is the case in which urban the literature regarding agronomic and environmental studies planners in northern global cities already include UA in their on open-air, urban, soilless and polyculture gardens. As stated agendas and policy planning (Morgan, 2009). by Specht and Sanyé-Mengual (2017), the available literature is Although soil-based agriculture is the most common urban insufficient, and new quantitative data are needed to “increase agricultural practice (Thomaier et al., 2015), soilless systems are awareness and knowledge” about urban rooftop agriculture. gaining importance as the lightest operation system. Therefore, The present research seeks to address this gap in the literature UA can be performed in unused urban spaces, such as rooftops by performing a quantitative analysis of the agronomic and or terraces (Nowak, 2004), which are already built spaces that environmental performance of a soilless urban home garden in are usually empty (Thomaier et al., 2015). Moreover, one of a Mediterranean city that grows a wide variety of vegetables the major risks in UA is contamination, mainly caused by (i.e., polyculture) using life cycle tools. In addition, this paper heavy metals present in soils (Ercilla-Montserrat et al., 2018). will provide useful specific indicators for policy making and In this sense, soilless practices help avoid this risk by using design planning in cities that seek to enhance UA. There is inert and non-contaminated substrates (Ercilla-Montserrat et al., little information in the literature about open-air rooftop soilless 2018). Notwithstanding that soilless systems can be perceived polyculture home urban gardens, and assessing the potential as “unnatural” or artificial (Specht and Sanyé-Mengual, 2017), yield for larger-scale urban production could be useful. This it should be considered that this practice is already highly lack of data implies uncertainty in urban management and consolidated in conventional agriculture (Savvas et al., 2013). hinders the inclusion of a food policy dimension in urban For instance, intensive greenhouse soilless food production is plans. This study seeks to shed some light by providing new performed in Almeria (Spain), the major vegetable producer quantitative data to assure the best performance of urban in southern Europe. According to Specht and Sanyé-Mengual gardens in the future. (2017), many of the vegetables for sale in the market are already The general aim of this paper is to assess the feasibility produced using soilless techniques (Specht and Sanyé-Mengual, of an open-air rooftop soilless polyculture home urban 2017). This wide use of soilless systems is due to the substantial garden in Barcelona by providing quantitative agronomic and water savings that it allows (Thomaier et al., 2015). Although environmental data. Two specific goals were designed, which are irrigation management is crucial for the performance of soilless related to the main section of the present study: systems, easy access to nutrients and water allows plants to grow faster and produce higher yields at higher densities because there • To assess the agronomic performance of the system is no competition for nutrients (Nowak, 2004). and relate it with the food demand of the citizens to Apart from community, commercial or industrial UA assess the amount of area needed to fulfill the vegetable initiatives, private home gardens have been always present consumption of the city. and still discreetly sprouting in cities (Kortright and • To assess the environmental impacts of a real case study by Wakefield, 2011). According to Calvet-Mir et al. (2012), applying the life cycle assessment methodology. there are many reasons to cultivate home gardens. The main goal is to obtain better quality and safer food, which will consequently enhance healthier diets by increasing the intake MATERIALS AND METHODS of abundant and diverse vegetables (Grewal and Grewal, 2012). Another important reason is that home gardens In this section, detailed information can be found on the system increase self-reliance and self-sufficiency, allowing certain under study as well as the experimental design and the agronomic economic independence and resilience to external dynamics and sample collection. Additionally, the stages of the life cycle (Calvet-Mir et al., 2012). Therefore, food sovereignty can assessment methodology applied for the environmental impact be seen as a form of empowerment (Grewal and Grewal, assessment are presented. Frontiers in Plant Science | www.frontiersin.org 2 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden TABLE 1 | General meteorological data in the city of Barcelona (Ajuntament de Barcelona, 2018a). 2015 2016 2017 Temperature Average (◦ C)1 18.4 18.4 18.4 Maximum (◦ C)1 34.7 32.6 34.2 Minimum (◦ C)1 2.1 3.9 1.6 Rainfall Total (mm)1 316.9 438.5 478.9 Maximum in a day (mm)2 47.0 66.0 115.1 Rainy days2 95 119 109 1 Can Bruixa Observatory. 2 Fabra Observatory. Study Area The garden under study is 18 m2 , and it is placed on a rooftop (private terrace on the first floor above the street level, south- facing) in the city center of Barcelona (41.38481N, 2.163125E). Barcelona is a very dense city, with almost 16,000 inhabitants per km2 (IDESCAT, 2018) due to its geographic location, which impedes its sprawl. The scarce land availability triggers urban rooftop agriculture as a potential solution for many urban sustainability issues related to the food supply. Moreover, FIGURE 1 | Scheme of the system under study. the Mediterranean climate allows year-round open-air farming practices. The general meteorological data for the years under assessment are presented in Table 1. As can be observed, there are TABLE 2 | Fertilizer irrigation concentration. no major temperature changes across years. However, the rainfall CaCl2 ∗ was higher and more densely concentrated in fewer days in 2017 KPO4 H2 KNO3 K2 SO4 Ca(NO3 )2 2H2 O Mg(NO3 )2 Hortilon Sequestrene than in previous years. mg · l− 1 136 101 217.5 164 111 148.3 10 10 Experimental Design nutrient solution (Table 2) was generic for the different crops The years under assessment are 2015, 2016, and 2017. Data grown (non-specific for leaf or fruit products). on the different inputs and life cycle stages are detailed in Supplementary Information 1. Moreover, to add data reliability, Agronomic Data 2014 was excluded from the analysis because it was mainly used Trials were performed for 22 different crops (Table 3), including to build and adjust the whole system. leafy and fruit vegetables. The study started with 5 pilot crops A soilless operation system was used in which plants grow in in 2014 and increased to 19 in 2015 and to 22 in 2016 and an inert substrate (perlite) and nutrients and water are supplied 2017. The inclusion criteria for the crops were their feasibility to through fertigation. As Figure 1 shows, 24 perlite bags were be grown in soilless systems and open-air conditions and their installed, and there was room for 72 plants (3 per bag). Each representability in the Mediterranean diet. Two main different bag contained 40 L of perlite and measured 1 × 0.3 m. Distance types of crops were grown: fruit and leafy crops. The main between bags (45 cm) and between crops is necessary to perform difference is that while leaf products are harvested by uprooting management tasks and to allow the cultivation of both fruit the whole crop, fruits have staggered harvests during a longer and leaf crops. As different crops were growing simultaneously, period. Additionally, leaf crops have shorter cycles and can be the plant density remained constant for the whole period planted many times a year, while fruiting vegetables are planted (4 plants m−2 ). Although smaller distances would have allowed only once or twice a year. Data on when the crops grew are higher crop densities, polyculture gardens have to adapt to crop shown in Figure 2. The yield data considered the fresh weight versatility and have to be managed to satisfy the needs of all kinds of the plants. This combination and proper management results of crops (e.g., avoid shadowing between plants). in minimum variation in the seasonal yield and the efficient use The water came from two polyethylene 50 L tanks that were of space. According to Orsini et al. (2014), both are essential in filled with water from the municipal water network. Nutrients self-sufficient urban home gardens. were injected from a concentrated nutrient solution designed Data from the Catalan Administration (Gencat, 2016) on food by agronomic experts, and the irrigation flow was distributed consumption preferences were gathered to assess how the garden through drippers. The water system was open (or linear), and under study could cover the food demand of the citizens. thus, the leachates were disposed of. The leachates were not quantified for this study, but data from Sanjuan-Delmás et al. Sample Collection (2018) (based on a hydroponic tomato crop in a nearby area) Data were collected manually for each different crop every was extrapolated for their quantification and composition. The harvesting day, indicating the weight per plant (for the leafy Frontiers in Plant Science | www.frontiersin.org 3 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden TABLE 3 | List of crops grown in the present study. Inventory The Life Cycle Inventory (LCI) is the data collection portion Common name Scientific name Variety of the LCA. In this section, all the inputs and outputs are Aubergine Solanum melongena – described, indicating the data sources used to define the origin, Bean Phaseolus vulgaris Contender transport and end of life of each element or flow. In addition, Beetroot Beta vulgaris Rubra the assumptions taken into account for the analysis are presented. Broad Bean Vicia faba Histal The system boundaries are graphically described in Figure 3. Broccoli Brassica oleracea Italica Cabbage Brassica oleracea Capitata Infrastructure Cauliflower Brassica oleracea Botrytis The complete inventory of the auxiliary equipment can be found Celery Apium graveolens Dulcis in Supplementary Information Table A1. It is considered that Chard Beta vulgaris Cicla all the elements are transported with a light commercial vehicle Zucchini Cucurbita pepo – from a distributor 35 km away from Barcelona. For the waste Cucumber Cucumis sativus – scenario, it is considered that most of the equipment is recycled Endive Cichorium endivia Crispum at a recycling plant 10 km away from Barcelona and transported Green pea Pisum sativum – in a municipal waste collection lorry. The leachate tray (made Lettuce Lactuca sativa Meravella of expandable polystyrene) was not recyclable and was taken Melon Cucumis melo – to a landfill 25 km away from the system. Regarding the wood Parsley Petroselinum crispum Apium petroselinum structure, the end of life treatment was the same as that used by Green pepper Capsicum baccatum – Puy et al. (2010) in Scenario 1 (postconsumer wood from the Arugula Eruca vesicaria – recycling points). This treatment includes the transport of the Spinach Spinacia oleracea – wooden waste to the nearest recycling point and from there to a Strawberry Fragaria vesca – treatment plant where the wood is separated and chipped (24 km Thistle Cynara cardunculus away). Finally, the product is transported to a gasification plant Tomato Solanum lycopersicum Arawak in Cerdanyola del Vallès (36 km) where energy is generated from the wooden waste. Operation and maintenance crops) and the fruit yield per plant for the fruit crops. The data The water used for irrigation came from the municipal water were converted to the yield per square meter considering all of the network (tap water), and the potabilization treatment impacts garden area. Information about irrigation changes (according to were included in the assessment. Substrate bags were not available the season) and fertilizer reposition was also collected, indicating locally and therefore imported from Almeria (800 km). With the date when these activities took place. In addition, details about a lifespan of 3 years, the substrate bags are finally taken to a the overall state of the crops and the system were also acquired. landfill as an inert material. The fertilizers traveled 33 km in a No technical advice (apart from that which was needed to ensure transport van to the system. They were concentrated in a nutrient data quality) or improvement suggestions were provided by the solution and then injected into the irrigation flow, which was researchers during the study period. The management was fully approximately 2 L/h. The fertilizers leave the system as leachates performed by the garden’s owner. in the sewer network and end up in a wastewater treatment plant. Some of these treatment plants do not remove all the chemicals; Life Cycle Assessment thus, the impacts associated with the leachates were accounted The environmental impacts were quantified using a Life Cycle for as direct emissions to water (Sanjuan-Delmás et al., 2018). Assessment following ISO 14040-44 (ISO, 2006), which includes The residual biomass was deposited in a municipal organic waste four main stages. container and transported to a composting plant 13,5 km away from Barcelona. The complete inventory of the operation phase Goal and Scope can be found in Supplementary Information 1 Table A1. The LCA included the whole garden system (from raw material extraction to end of life), excluding the nursery plants (considered Life Cycle Impact Assessment negligible based on the data from Sanjuan-Delmás et al. (2018). The LCA was performed using Simapro 8.2 software by Pré No extra energy was needed except the amount included in Consultants. The method used to calculate the environmental the tap water supply from the network. The system was split impacts was ReCiPe (hierarchical) at a midpoint level. The into infrastructure (elements with a lifespan of more than Ecoinvent 3 database was used as the main source for the 5 years) and operation (elements with a lifespan of less than background environmental data. According to Brentrup et al. 5 years) (Figure 2). (2004), the selected impact categories for the assessment were The functional unit selected for this assessment was 1 kg the following: climate change (CC), ecotoxicity (ET; which of edible fruit and leafy products cultivated in an open-air is calculated by adding terrestrial, freshwater and marine soilless polyculture rooftop home garden in a Mediterranean city. ecotoxicity environmental impacts), terrestrial acidification (TA), Moreover, the impacts per year were also used in some sections freshwater eutrophication (FE), marine eutrophication (ME), and of the discussion to add clarity. fossil fuel depletion (FDP). Frontiers in Plant Science | www.frontiersin.org 4 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden FIGURE 2 | Time data on when the crops occupied a space in the urban garden. Abbreviation meanings: L, Lettuce; T, Tomato; CH, Chard; SP, Spinach; Z, Zucchini; EN, Endive; CB, Cabbage; BP, Green pea; BT, Beetroot; ST, Strawberry; R, Arugula; P, Pepper; EG, Aubergine; BR, Broccoli; CL, Celery; ME, Melon; CC, Cucumber; BB, Broad bean; BE, Bean; TH, Thistle; and PR, Parsley. RESULTS Food production is assessed for the agronomic portion, while the environmental performance focuses on the life The agronomic and environmental performance of the study cycle impact assessment considering the data acquired system are shown and described in the following section. in the inventory. Frontiers in Plant Science | www.frontiersin.org 5 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden FIGURE 3 | Diagram of the system boundaries. Agronomic Data and Food Production were fruits and 215.7 kg were leafy vegetables. By splitting As shown in Table 4, a total of 569.7 kg of vegetables were the data in time, an average productivity of 189.9 kg/year and produced during the study period (2015–2017), of which 354 kg 10.6 kg/m2 /year was achieved. The top 5 most productive crops were tomato, lettuce, chard, pepper and eggplant (in this order), representing 85.5% of the TABLE 4 | Agronomic results from 2015, 2016, and 2017 for the present study. total production. Specifically, the average top 3 leaf vegetables were chard, lettuce and cabbage, with an average production of Total production 27, 23.2, and 15.7 kg/year, respectively. The 3 most productive Unit 2015 2016 2017 Average 2015–2017 fruit crops were tomato, pepper and eggplant, with an average of 40.0, 28.6, and 24.2 kg/year, respectively. Table 5 shows the Lettuce L Total kg 22.3 27.4 19.6 23.1 69.3 Chard L 33.5 14.6 32.9 27.0 81.0 specific results in these 3 crops from 2015 to 2016. Spinach L 10.5 5.0 7.8 15.5 The tomato summer and winter season productions resulted Tomato F 54.5 29.8 35.6 40.0 119.9 in 5.8 kg/m2 and 3.6 kg/m2 , respectively. Compared with other Zucchini F 1.0 1.0 1.0 studies, such as Orsini et al. (2014), who cultivated tomato in an Endive L 1.5 1.95 1.7 3.5 urban rooftop with a soil-based system, or Sanjuan-Delmás et al. Cabbage L 15.7 15.7 15.7 (2018), who grew tomato in an integrated rooftop greenhouse Green pea F 2.8 2.8 2.8 with a soilless system, this is a relatively low production. The Strawberry F 0.1 0.7 0.4 0.8 differences between the yield obtained in the summer and winter Arugula L 0.7 1.6 1.2 2.3 Pepper F 30.4 25.2 30.3 28.6 85.9 crops were mainly caused by meteorological conditions, as the Eggplant F 15.3 29.7 27.6 24.2 72.6 system in this study is open-air. However, the differences in yield Broccoli F 3.4 3.9 10.1 5.8 17.4 between seasons were less different than those obtained outside Celery L 4.4 15.5 1.5 7.1 21.4 urban area (Sanjuan-Delmás et al., 2018). Melon F 3.6 3.6 3.6 In contrast, the pepper crops produced 12.6 kg/m2 during Cucumber F 1 1 2.5 1.5 4.5 the summer season and 10.9 kg/m2 during the winter season. Broad bean F 5.7 8.6 7.2 14.3 In this case, the production is considerably high, even surpassing Cauliflower F 4 15.0 9.5 19.0 the industrial pepper production in soilless greenhouses (García Bean F 4.5 5.5 5.0 10.0 Thistle L 3.9 3.2 3.6 7.1 Artichokes F 2.2 2.2 2.2 TABLE 5 | Production data for 3 crops in the present study in 2015 and 2016. Total leafy production 88.6 63.4 63.8 71.9 215.7 Number of Total fruit production 115.1 100.5 138.4 118.0 354.0 Crop Season plants DAT kg plant/m2 kg/plant kg/m2 Total production 203.7 163.9 202.2 189.9 569.7 Tomato August–March 18 221 16.3 4 0.91 3.62 Total production kg/m2 11.3 9.1 11.2 10.6 March–September 12 169 17.5 1.46 5.83 Waste biomass kg/m2 3.8 2.8 3.3 3.3 Pepper May–January 6 250 18.9 3.15 12.60 Water L/(m2 · d) 3.3 3.6 4.2 3.7 July–March 6 267 16.4 2.73 10.93 L/kg 107.7 146.0 137.8 130.5 Eggplant May–March 3 310 17.1 5.70 22.80 Classification of the leafy (L) and fruit (F) products. DAT indicates days after transplanting. Frontiers in Plant Science | www.frontiersin.org 6 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden Lozano et al., 2005). Similarly, the eggplant production TABLE 6 | Environmental assessment of producing 1 kg of product. (22.8 kg/m2 ) was highly significant and exceeded the annual Infrastructure Operation Total production in Spain (4.5 kg/m2 ) (Daunay, 2008). Auxiliary Wood Emissions Waste The other crops, such as melon and green peas, had very eq. structure Fertilizers to water Substrate Water biomass low yields (<5 kg per year) and presented fungal infections. CC 9.1E-02 9.5E-03 3.9E-01 – 5.2E-03 3.9E-02 4.9E-03 5.5E-01 Therefore, they were not cultivated the following years. A few of 16.6% 1.7% 72.7% 1.0% 7.1% 0.9% the crops, such as beetroot or parsley, were not successful and TA 4.1E-04 1.2E-04 2.9E-03 – 3.8E-05 1.9E-04 2.3E-05 3.7E-03 were not considered in the analysis. 10.9% 3.2% 79.3% 1.0% 5.0% 0.6% As a reference for local conventional production, the annual FE 2.8E-05 1.2E-06 8.4E-04 9.8E-04 5.9E-07 2.5E-05 1.6E-07 1.9E-03 1.5% 0.1% 44.8% 52.3% <0.1% 1.3% <0.1% Catalan production of tomato, pepper and eggplant is 4.10, 2.45, ME 7.1E-05 6.7E-06 2.3E-04 4.8E-03 2.4E-06 9.2E-06 1.3E-06 5.1E-03 and 2.56 kg/m2 , respectively (IDESCAT, 2018). These data show 1.4% 0.1% 4.4% 93.9% <0.1% 0.2% <0.1% that the home garden productivity of these crops was significantly ET 3.5E-03 7.5E-04 6.1E-02 3.4E-03 4.9E-05 1.7E-03 1.2E-05 7.0E-02 higher, except for the winter tomato crop. 5.0% 1.1% 86.6% 4.8% 0.1% 2.4% <0.1% The number of people that the garden could cover FDP 3.8E-02 2.9E-03 7.4E-02 - 2.9E-03 1.1E-02 1.6E-03 1.3E-01 29.3% 2.3% 56.6% 2.2% 8.4% 1.2% was calculated (Supplementary Information 2 Table B1). Units: kg CO2 eq (CC, climate change), kg SO2 eq (TA, terrestrial acidification), kg Considering that the home with the garden under study hosted P eq (FE, freshwater eutrophication), kg N eq (ME, marine eutrophication), kg oil eq two people, the consumption was adequately covered by the (FD, fossil depletion), and kg 1-4DB eq (ET, ecotoxicity). Percentages express the garden production except for zucchini, bean and cucumber relative contributions in each specific impact category. production (10%, 75–90%, and 20–55% for the demand of 2 people). For crops such as chard or eggplant, the production was largely exceeded (487–1116% and 364–707% for the demand greenhouse production and conventional agriculture are of 2 people, respectively). bound. For instance, the typical season for eggplant is from May to September (120 days) (Aktas et al., 2013). Then, the Environmental Assessment crop is uprooted, and the next crop is planted following a In this section, the Life Cycle Impact Assessment (LCIA) is monoculture production system. In contrast, the present study performed by analyzing every subcategory of the inventory. did not adhere to the conventional production seasons but The most important impacts are highlighted, and some rather extended them following the owner’s criteria and local recommendations for improvements are given. weather conditions. Therefore, continuing with the previous example, in the study system, due to the favorable weather Life Cycle Inventory conditions (mainly due to the high minimum temperature), the The inventory of the two subsystems considered (infrastructure eggplant crop was left until March (310 days). Consequently, and operation) is described in detail in Supplementary the production season was extended, resulting in a higher Information 1 Tables A1, A2. final production. These favorable local weather conditions could be affected Life Cycle Impact Assessment by the urban heat island effect, which, compared with the Table 6 shows the environmental impacts per kg of product as outskirts of the city of Barcelona, can increase temperatures well as the relative impact per impact category. The results show from 3 to 8◦ C (Ajuntament de Barcelona, 2018b). According that most of the environmental impacts are generated during the to Mimet et al. (2009), the seasonal variations that urban crops operation phase of the garden (water, substrate, waste biomass undergo can be explained by the microclimates created in composting, and fertilizers). This subsystem contributes between city centers, which, compared to rural areas, cause the earlier 68% (FDP) and 98.5% (ME) of the impact in all the studied onset of the flowering phase. This alteration in the phenology categories (Table 5). The specific analysis of the results is detailed of the plants leads to these extended seasons of production. in Section “Environmental Assessment Interpretation.” Moreover, the heat island effect could offer new opportunities for cultivating vegetables that would not grow in the surrounding rural areas (Waffle et al., 2017). These crops could find sufficient DISCUSSION climatic difference in cities to successfully grow during the winter seasons. This would be the case for tomato, eggplant In this section, the production results are discussed and and pepper, which have added value in winter considering compared with the consumption data of the region. Finally, that winter is not their typical season in the Mediterranean the environmental results shown in the results section are area. In winter, these crops are more difficult to find in interpreted and discussed. markets, more expensive and usually imported, increasing their environmental impact. Food Production Assessment The previous findings highlight the importance of garden It is remarkable that an open-air system (present study) can management and the personal criteria of the owner, whose obtain similar or higher productivity values than industrial decisions affect the final production. Garden management can, or research greenhouses and conventional agriculture. for example, concern the water use efficiency of the system This might be due to the standard seasons to which considering that water management plays a major role in Frontiers in Plant Science | www.frontiersin.org 7 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden soilless system performance compared to soil-based systems. based on higher vegetable intake represents a healthier diet, In addition, water can be one of the elements with higher decreasing the chance of suffering diet-related chronic diseases economic costs in UA systems (Sanyé-Mengual et al., 2015). (World Health Organization [WHO], 2015). Compared to other studies, such as Sanjuan-Delmás et al. However, an important portion of the harvest was shared with (2018), the water use of the present system (1350.5 L/m2 /year) other people. Having a vegetable surplus in a home garden can is relatively high This fact has multiple explanations. First, lead to the reinforcement of social relations by exchanging or it can be explained by the use of a single water circuit that giving the excess amounts to others. These practices enhance the irrigates the whole garden. The water flow was adjusted to the creation and maintenance of social relations along with social fruit crop water requirements, which turned out to be higher cohesion and inclusion. Moreover, they can create networks of than the requirements of the leaf crops, resulting in a higher knowledge transmission and turn home gardens into agronomic water consumption than that in monoculture systems where learning places (Calvet-Mir et al., 2012). the water flow is adjusted to a crop-specific water requirement. The average fruit and vegetable consumption in Catalonia Second, because the leachates were not quantified, there was for 2015 and 2016 was 160 kg per person per year, excluding no possibility that the garden owner could adjust the irrigation potatoes, which is more than the dietary intake recommended based on the drainage proportion. Moreover, the use of an by the World Health Organization (World Health Organization open water irrigation system implies that all the irrigation water [WHO], 2015). Nevertheless, of that amount, 104 kg were fruits, is consequently new water entering the system without any leaving a vegetable intake of only 56 kg per person per year. recirculation, which would immediately have an impact on new Those 56 kg were formed by 14 different vegetables, of which water consumption. 9 have been grown in the studied system (Table 3). Based on The selection of crops in this kind of garden is completely the vegetable market basket and the production data, the area subjective and based on garden owner perception. For example, needed to fulfill the vegetable demand was calculated for 6 out if the minimum temperature values are reached when a specific of 9 products (the data were insufficient for the other crops). crop is growing and the crop does not succeed, the garden owner’s The results show that considering a density of 4 crops per square perception of the performance of this species could be highly meter, 8.2 m2 would be necessary to satisfy 62% (by weight) of the altered, and as a consequence, the owner does not plant this crop vegetal demand for one person. again. Moreover, considering that crop selection is exclusively However, the average productivity of the present system was made by the garden owner, the crop variety and number of 10.6 kg/m2 , from which it can be deduced that 5.3 m2 are needed plants are not the same year after year. This fact makes data per person to fulfill the vegetable demand, though this area analysis more complex. In this sense, we suggest that in further does not include all the vegetables of the market basket. In this research, an accurate analysis should be performed focusing sense, improvements can be made to adapt the production to on crop performance, especially in crops that can be grown the consumption rates. For instance, more importance should in the off-season. be given to zucchini, beans and cucumber and less to eggplant and chard, whose production far exceeded the consumption Food Consumption rates. It should also be considered that diversifying the quantity Table 3 shows the agronomic data for the products that were of crops is necessary for avoiding the scenario in which only produced in the studied system and commonly consumed by 5 products represent 78% of the total production, as happened Catalan citizens (market basket) according to official Catalan in the present study. Administration sources (Gencat, 2016). They are ranked in The 5 crops present in the market basket that were not descending order by the citizen consumption rates in kg per year. cultivated in the present study are onion, carrot, mushrooms, As shown in Table 3, only 2 products were consumed and not asparagus and garlic. There are no previous data on the feasibility produced in the present system and were therefore missing from of soilless cultivation in Mediterranean areas for these crops the list (position 2 for onion and 6 for carrot). (Savvas et al., 2013). Similarly, fruit cultivation in soilless systems There were many other crops produced in the garden, has not been studied. In the present system, the only fruits that such as cauliflower, broccoli, celery, etc., though consumption were grown were melon, watermelon and strawberries, which data were not available for these crops in the administration are in positions 5, 6, and 20, respectively (per weight), in the databases. The omission of these crops could be interpreted consumption ranking of Catalonia. However, these crops were as their consumption being insignificant compared with other not successful in the present system. products. Nevertheless, the productivity of some of these A final possible consideration would be to increase the intake crops in the present study was remarkable. Taken together, of the more successful vegetables among the community near the these results suggest that the crop inclusion criteria were garden. The market basket was taken only as a reference of the based on the owner’s preference for certain vegetables or variety and the quantity of the average vegetable consumption, the preference for a mostly vegetarian diet, which would but other vegetable combinations (varieties and quantities) are include a greater diversity and quantity of vegetables than possible and healthy as well. The implementation of home urban those required by an average citizen. It could also be gardens might boost the consumption of less popular and self- said that by managing a home garden, increased vegetable produced vegetables while decreasing the quantity of other fruits intake is almost inevitable, and thus, the average market that might be imported (such as banana), thus reducing the basket does not fit the purpose anymore. This diet change overall environmental impact and increasing self-sufficiency. Frontiers in Plant Science | www.frontiersin.org 8 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden FIGURE 4 | Relative impact of specific Fertilizers within the Fertilizer item in the life cycle inventory. Environmental Assessment Among the remaining fertilizers, only Ca(NO3 )2 and K2 SO4 had Interpretation impacts above 20% of the fertilizer item impact. For instance, The greatest impact of the operation subsystem was mainly due Ca(NO3 )2 exerted the greatest impact in the CC (36.4 kg CO2 to the fertilizers and their emissions, which were the elements eq. per year; almost half of fertilizer impact) and FDP (4.48 kg oil with the greatest impact in all 6 of the impact categories eq. per year) categories. K2 SO4 had similar impacts in this impact studied. On the one hand, fertilizers accounted for between category (4.37 kg oil eq. per year). 56.6 and 86.6% of the impact in 4 out of the 6 impact Considering that the system is open (or linear), it is important categories (FDP, ET, TA, and CC). The relative impacts of the to pay special attention to keeping the balance between not types of fertilizer are shown in Figure 4. Within fertilizers stressing or overwatering the plants so that the environmental (Figure 3), Hortilon had the greatest impact for the ET impacts of the leachates are kept to a minimum and are and TA impact categories, accounting for 75.3 and 26.3% of not recirculated. the fertilizer impact, respectively. Moreover, it also exerted a The network water used for the irrigation of the garden and great impact (0.15 kg of P eq. per year) in the FE category. the substrate had negligible impacts in most of the categories However, the eutrophication impact categories were dominated [between 8.4% (FDP) and 0.2% (ME) for water and between 2.2% by the drained leachates, both the FE (0.20 kg P eq. per (FDP) and 0.03% (FE) for the substrate]. year) and ME (0.96 kg of N eq. per year) categories. These The waste biomass treatment had the lowest impact, impacts were highly significant due to the aforementioned accounting for between 0.01 and 1.2% of the impact in the assumption that most of the chemicals are not removed in the studied categories. Nevertheless, transport (municipal solid waste wastewater treatment plant and are considered direct emissions collection) far exceeded the impacts of the composting process, to the environment. contributing over 97% of the impact in all the studied categories Frontiers in Plant Science | www.frontiersin.org 9 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden (Supplementary Information 1 Table A4). Similar results were fertilizer use (applying multiple crop-specific irrigation sectors) found by Quirós et al. (2014). and leachate treatment (by promoting nutrient recirculation The infrastructure impacts were mainly focused on the FDP within the same system) should be optimized due to their (31.6%), CC (18.4%) and TA (14.1%) categories. Within the great impacts. Finally, a rainwater harvesting system should auxiliary equipment, the leachate tray (made of expandable be implemented considering that with the 2017 annual rainfall polystyrene and HDPE) was the most impacting element, (442.1 mm) (IDESCAT, 2018), 29% of the water demand for the accounting for more than 40% of the impacts in 5 out of the garden could be met. 6 categories (CC, TA, ME, ET, and FDP). For the wooden structure, the impacts accounted for between 3.2 and 0.1%, which is not very significant. Nevertheless, most AUTHOR CONTRIBUTIONS of the structure was not directly used for the agricultural practice and had mainly decorative purposes. As a possible AB acquired and gathered the data. Then, AB processed all implementation, reducing the amount of wood to the minimum the agronomic and environmental data to be included in the needed to support the stalking of the crops to minimize the present article. AB also wrote the majority of the paper. MR-S impacts caused by this structure could be considered. assessed the environmental results and their interpretation as well wrote the article and discussed the results, especially the environmental assessment section. MR-S also integrated the majority of the revisions from the journal reviewers. ME-M CONCLUSION assessed the agronomic results and their interpretation as well as wrote the article and discussed the results, especially the This investigation showed the feasibility of a real polyculture agronomic section. JR acquired and processed a portion of the soilless rooftop garden that achieved an average productivity data, assessed the global results as well as their interpretation and of 10.6 kg/m2 /year with a density of 4 plants/m2 . The top wrote the conclusions, and take-home messages. XG assessed the 5 most productive crops were tomato, chard, lettuce, pepper and paper writing quality as well as the discussion section, especially eggplant, accounting for 85.5% of the total production. Other the agronomic section. crops, such as melon and green peas, were not as successful and had very low yields (<5 kg/year). Similarly, the water use was 3.7 L/m2 /day, which is a relatively high compared to other studies. FUNDING Relative to food consumption, 9 out of the 14 vegetables that This work was supported by the Spanish Ministry of Economy, constitute the Catalan vegetable market basket were cultivated in Industry and Competitiveness (CTM2016-75772-C3-1-R) and the present study. In terms of the percentage covered, an 8.2 m2 the “María de Maeztu” program for Units of Excellence in R&D soilless garden is sufficient for covering 62% of the Catalan (MDM-2015-0552). vegetable market basket. Regarding the environmental assessment, the urban home garden produced 5.9 kg of CO2 eq/m2 /year. The majority of the ACKNOWLEDGMENTS impacts were caused during the operation phase (68.4 – 98.5% of impact among the 6 categories), particularly by the fertilizers in We are grateful to Generalitat de Catalunya for the grant 4 out of the 6 analyzed impact categories (CC, TA, ET, FDP) and awarded to ME-M (FI-DGR 2016) and the Universitat the resulting leachates for the other 2 categories (FE and ME). Autònoma de Barcelona for awarding a research scholarship Compared to operation, infrastructure had the lowest impact, (PIF-UAB 2017) to MR-S. accounting for 1.5–31.6% of the impact. The findings from this study make several contributions to the current literature on UA. Moreover, from the experience earned SUPPLEMENTARY MATERIAL from this study, we detected which future research lines regarding this type of garden could follow. First, there should be cultivation The Supplementary Material for this article can be found online trials of the crops included in the market basket but not assessed at: https://www.frontiersin.org/articles/10.3389/fpls.2019.00341/ in the present study to assess their performance. Second, the full#supplementary-material REFERENCES Aktas, H., Daler, S., Ozen, O., Gencer, K., Bay, D., and Erdal, I. (2013). The effect of some growing substrate media onyielld and fruit quality of Eggplant (Solanum Ajuntament de Barcelona Índex (2018a). Anuari Estadístic de la Ciutat melongena L) grown and irrigated by drip irrigation system in greenhouse. de Barcelona 2017. Available at: http://www.bcn.cat/estadistica/catala/dades/ J. Agric. Sci. 1, 5–12. anuari/ [accessed January 25, 2019]. Brentrup, F., Küsters, J., Kuhlmann, H., and Lammel, J. (2004). Environmental Ajuntament de Barcelona (2018b). Pla Clima - Resum Executiu - Efecte impact assessment of agricultural production systems using the life cycle assess- Illa de Calor. Available at: http://lameva.barcelona.cat/barcelona-pel-clima/ca ment methodology: I. Theoretical concept of a LCA method tailored to crop [accessed January 24, 2019]. production. Eur. J. Agron. 20, 247–264. doi: 10.1016/S1161-0301(03)00024-8 Frontiers in Plant Science | www.frontiersin.org 10 March 2019 | Volume 10 | Article 341 Boneta et al. Assessment of Urban Home Garden Calvet-Mir, L., Gómez-Baggethun, E., and Reyes-García, V. (2012). Beyond food Sanjuan-Delmás, D., Llorach-Massana, P., Nadal, A., Ercilla-Montserrat, M., production: ecosystem services provided by home gardens. a case study in Muñoz, P., Montero, J. I., et al. (2018). Environmental assessment of an Vall Fosca, Catalan Pyrenees, Northeastern Spain. Ecol. Econ. 74, 153–160. integrated rooftop greenhouse for food production in cities. J. Clean. Prod. 177, doi: 10.1016/j.ecolecon.2011.12.011 326–337. doi: 10.1016/j.jclepro.2017.12.147 Cameron, R. W. F., Blanuša, T., Taylor, J. E., Salisbury, A., Halstead, A. J., Sanyé-Mengual, E., Cerón-Palma, I., Oliver-Solá, J., Montero, J. I., and Henricot, B., et al. (2012). The domestic garden - Its contribution to urban green Rieradevall, J. (2013). Environmental analysis of the logistics of agricultural infrastructure. Urban For. Urban Green. 11, 129–137. doi: 10.1016/j.ufug.2012. products from roof top greenhouses in mediterranean urban areas. J. Sci. Food 01.002 Agric. 93, 100–109. doi: 10.1002/jsfa.5736 Daunay, M.-C. (2008). “Eggplant,” in Handbook of Plant Breeding: Vegetables II, eds Sanyé-Mengual, E., Orsini, F., Oliver-Solà, J., Rieradevall, J., Montero, J. I., and J. Prohens and F. Nuez (New York, NY: Springer), 163–220. doi: 10.1007/978- Gianquinto, G. (2015). Environmental and economic assessment of multiple 0-387-74110-9_5 cultivation techniques and crops in open-air community rooftop farming in Ercilla-Montserrat, M., Muñoz, P., Montero, J.-I., Gabarrell, X., and Villalba, G. Bologna (Italy). Agron. Sustain. Dev. 35, 1477–1488. doi: 10.1016/B978-0- (2018). A study on air quality and heavy metals content of urban food produced 88415-758-8.50009-0 in a Mediterranean city (Barcelona). J. Clean. Prod. 195, 385–395. doi: 10.1016/ Savvas, D., Gianquinto, G., Tuzel, Y., and Gruda, N. (2013). Soilless culture. Good j.jclepro.2018.05.183 Agricultural Practices for Greenhouse Vegetable Crops. Rome: FAO. Eurostat (2015). Urban Europe – Statistics on Cities, Towns and Suburbs – Scoones, I. (2009). The politics of global assessments: the case of the Executive Summary – Statistics Explained. Available at: https://ec. International Assessment Of Agricultural Knowledge, Science And Technology europa.eu/eurostat/statistics-explained/index.php/Urban_Europe_-_statistics_ For Development (IAASTD). J. Peasant Stud. 36, 547–571. doi: 10.1080/ on_cities,_towns_and_suburbs_-_executive_summary [accessed March 20, 03066150903155008 2019]. Specht, K., and Sanyé-Mengual, E. (2017). Risks in urban rooftop agriculture: García Lozano, M., Escobar, I., and Berenguer, J. J. (2005). Green-pepper fertigation assessing stakeholders’ perceptions to ensure efficient policymaking. Environ. in soilless culture. Acta Hortic. 697, 543–547. doi: 10.17660/ActaHortic.2005. Sci. Policy 69, 13–21. doi: 10.1016/j.envsci.2016.12.001 697.71 Taylor, J. R., and Lovell, S. T. (2014). Urban home gardens in the global north: Gencat (2016). Consum Alimentari. Departament d’Agricultura, Ramaderia, a mixed methods study of ethnic and migrant home gardens in Chicago, IL Pesca i Alimentació. Available at: http://agricultura.gencat.cat/ca/departament/ urban home gardens in the global north: a mixed methods study of ethnic estadistiques/alimentacio/consum-alimentari [accessed January 25, 2019]. and migrant home gardens in Chicago, IL. Renew. Agric. Food Syst. 30, 22–32. Grewal, S. S., and Grewal, P. S. (2012). Can cities become self-reliant in food? Cities doi: 10.1017/S1742170514000180 29, 1–11. doi: 10.1016/j.cities.2011.06.003 Thomaier, S., Specht, K., Henckel, D., Dierich, A., Siebert, R., Freisinger, U. B., IDESCAT (2018). Idescat. Anuari estadístic de Catalunya. Pluviometria. et al. (2015). Farming in and on urban buildings: present practice and specific Precipitació mensual. Comarques i Aran. Available at: https://www.idescat.cat/ novelties of Zero-Acreage Farming (ZFarming). Renew. Agric. Food Syst. 30, pub/?id=aec&n=217 [accessed January 24, 2019]. 43–54. doi: 10.1017/S1742170514000143 ISO (2006). ISO 14040: Environmental Management – Life Cycle Assessment – United Nations, Department of Economic and Social Affairs, Population Division Principles and Framework. Geneva: International Organization for (2017). World Population Prospects: The 2017 Revision, Methodology of the Standardization. United Nations Population Estimates and Projections. Working Paper No. Kortright, R., and Wakefield, S. (2011). Edible backyards: a qualitative study of ESA/P/WP.250. New York: United Nations. household food growing and its contributions to food security. Agric. Hum. Waffle, A. D., Corry, R. C., Gillespie, T. J., and Brown, R. D. (2017). Urban heat Values 28, 39–53. doi: 10.1007/s10460-009-9254-1 islands as agricultural opportunities: an innovative approach. Landsc. Urban Mimet, A., Pellissier, V., Quénol, H., Aguejdad, R., Dubreuil, V., and Rozé, F. Plan. 161, 103–114. doi: 10.1016/j.landurbplan.2017.01.010 (2009). Urbanisation induces early flowering: evidence from platanus acerifolia Watson, R., Albon, S., Aspinall, R., Austen, M., Bardgett, B., Bateman, I., et al. and prunus cerasus. Int. J. Biometeorol. 53, 287–298. doi: 10.1007/s00484-009- (2011). UK National Ecosystem Assessment:understanding Nature’s Value to 0214-7 Society. Synthesis of Key Findings. Available at: http://eprints.lancs.ac.uk/49673/ Morgan, K. (2009). Feeding the city: the challenge of urban food planning. Int. [accessed January 24, 2019]. Plan. Stud. 14, 341–348. doi: 10.1080/13563471003642852 WHO (2016). Global Report on Urban Health. Available at: https://apps.who.int/ Nowak, M. (2004). Urban Agriculture on the Rooftop. Ph.D. thesis, Cornell iris/handle/10665/204715. University, Ithaca, NY. World Health Organization [WHO] (2015). Population nutrient intake goals for Orsini, F., Gasperi, D., Marchetti, L., Piovene, C., Draghetti, S., Ramazzotti, S., et al. preventing diet-related chronic diseases. World Health Organ. 346, 871–876. (2014). Exploring the production capacity of rooftop gardens (RTGs) in urban agriculture: the potential impact on food and nutrition security, biodiversity Conflict of Interest Statement: The authors declare that the research was and other ecosystem services in the city of Bologna. Food Secur. 6, 781–792. conducted in the absence of any commercial or financial relationships that could doi: 10.1007/s12571-014-0389-6 be construed as a potential conflict of interest. Puy, N., Rieradevall, J., and Bartrolí, J. (2010). Environmental assessment of post- consumer wood and forest residues gasification: the case study of Barcelona Copyright © 2019 Boneta, Rufí-Salís, Ercilla-Montserrat, Gabarrell and Rieradevall. metropolitan area. Biomass Bioenergy 34, 1457–1465. doi: 10.1016/j.biombioe. This is an open-access article distributed under the terms of the Creative Commons 2010.04.009 Attribution License (CC BY). The use, distribution or reproduction in other forums Quirós, R., Villalba, G., Muñoz, P., Font, X., and Gabarrell, X. (2014). is permitted, provided the original author(s) and the copyright owner(s) are credited Environmental and agronomical assessment of three fertilization treatments and that the original publication in this journal is cited, in accordance with accepted applied in horticultural open field crops. J. Clean. Prod. 67, 147–158. doi: academic practice. No use, distribution or reproduction is permitted which does not 10.1016/J.JCLEPRO.2013.12.039 comply with these terms. Frontiers in Plant Science | www.frontiersin.org 11 March 2019 | Volume 10 | Article 341