This is the submitted version of the article: Parada, Felipe; Ercilla Montserrat, Mireia; Arcas Pilz, Verónica; [et al.]. «Com- parison of organic substrates in urban rooftop agriculture, towards improving crop production resilience to water stress in Mediterranean cities». Journal of the Science of Food and Agriculture, (April 2021). DOI 10.1002/jsfa.11241 This version is available at https://ddd.uab.cat/record/240830 under the terms of the license 1 Comparison of organic substrates in urban rooftop agriculture, 2 towards improving crop production resilience to water stress in 3 Mediterranean cities 4 5 Short running title: Comparison of organic substrates in urban rooftop agriculture 6 resilience to water stress 7 8 Felipe Paradaa, Mireia Ercilla-Montserrata, Verónica Arcas-Pilza, Elisa Lopez-Capelc, Núria Carazod, Juan 9 Ignacio Montero, Xavier Gabarrellab, Gara Villalbaab, Joan Rieradevalla, Pere Muñozab*. 10 11 a Sostenipra Research Group 3.0 2017SGR 1683, Institute of Environmental Sciences and Technology (MDM-2015-0552), Z 12 Building, Autonomous University of Barcelona (UAB), Campus UAB, 08193 Bellaterra, Barcelona, Spain. 13 b Department of Chemical Engineering, Biological and Environmental, School of Engineering, Building Q, Universitat Autònoma de 14 Barcelona (UAB), 08193 Bellaterra, Barcelona, Spain. 15 c School of Natural and Environmental, Agriculture Building, Newcastle University, Newcastle upon Tyne, NE1 7RU, UK. 16 d Department of Agri Food Engineering and Biotechnology (DEAB), Polytechnical University of Catalonia (UPC), Campus del Baix 17 Llobregat, Castelldefels, Spain. 18 Corresponding author: Pere Muñoz (Pedro.Munoz@uab.cat). 19 20 Abstract 21 BACKGROUND: Urban agriculture contributes to meet food production demand in 22 cities. In a context of low water availability, it is important to consider alternatives that 23 are able to maintain production. This study aimed to assess the use of substrates made 24 from local materials and high water retention capacity as an alternative for urban 25 agriculture in periods with water stress. Different substrates were used for 3 consecutive 26 crop cycles of lettuce (Lactuca sativa L.) during the spring and summer periods of 2018 27 to observe these substrates performance during warmer periods of the year in an 28 integrated rooftop greenhouse near Barcelona. The substrates used were coir commercial 29 organic substrate, vegetable Compost from urban organic waste, Perlite (as control) 30 commercial standard substrate, and a Mixture of the urban Compost and Perlite (1:1). 31 Substrate crop performance was assessed under conventionally irrigation (0-5 cbar) and 32 water restricted conditions (irrigation stop until the water tension inside the perlite bags 33 reached -20 cbar). RESULTS: The results demonstrate that the Compost and Mix yields 34 were similar to those obtained from Perlite (11.5% y 3.7% of more production in a 35 restricted water condition average values). Compared to the Perlite, the organic substrates 36 increased the crops resilience to water restriction, through biomass accumulation 37 comparison, it took longer for Coir to lose water (1 and 2 test); however, when dryness 38 began, it occurred very quickly. CONCLUSION: The vegetable Compost and the 39 substrate Mixture presented tolerance to water restriction when water restriction reached 40 -20 cbar. 41 42 Keywords: Circular economy, sustainable cities, soilless system, water stress resilience, 43 water restriction, urban Agriculture. 44 45 1. Introduction 46 Currently, the increase in population within cities has created a concern due to an 47 increased demand for resources such as energy, water, and food. This situation is 48 exacerbated by the advance of climate change, causing persistent droughts, one of the 49 biggest problems to be addressed in agriculture due to the high water demand for food 50 production 1. 51 52 Urban agriculture (UA) is an alternative that originated to satisfy the higher demand, 53 contributing to food production’s sustainability by reducing different production chain 54 elements, such as energy in distribution and used packaging 2. UA can be carried out at 55 different levels: Rooftop Agriculture (RA) (3) on the ground and on building rooftops. 56 The advantage of rooftop greenhouses is access to unutilized spaces, increasing current 57 local food production, and reducing the environmental load associated with food 58 production and the buildings that sustain it 4. The evaluation of unoccupied roof spaces 59 for rooftop greenhouse (RTGs) found that these are usually small and well-ventilated and 60 have a very low relative humidity. This leads to a condition with high water consumption 61 by plants and, therefore, a propensity for crops to suffer hydric stress 5. 62 63 Barcelona is an example of a Mediterranean city area, where droughts have been repeated 64 cyclically for the past two decades. This situation has led to creating a management plan 65 that aims to prioritize water uses in cities, especially during emergencies. UA is 66 considered as a green space amenity activity rather than an agricultural activity in Spain, 67 hampered by the legal restrictions applied to these areas. As an example, tap water 68 irrigation of private gardens and city parks was forbidden during water shortages in 2008. 69 This highlights the importance to develop alternatives to alleviate drought conditions of 70 urban crop systems and maintain the food production. There is a need to study strategies 71 and technologies that allow the development of crops in water-limiting conditions, such 72 as irrigation optimization, reuse of leachates, and soilless culture systems (SCSs). SCSs, 73 is frequently used to establish crops in an artificial medium to produce food under 74 different growing conditions 6. A variety of organic and inorganic substrates could be 75 suitable for crop production under restricted water availability. 76 77 One of the most used substrates is Perlite, an inorganic grow media characterized by its 78 capacity for aeration, drainage, and optimum water retention. However, a high amount of 79 energy is required for its production and transportation. Organic alternative substrates 80 widely use include Coir and Compost 7. These present desirable substrate characteristics, 81 such as high water holding, cation exchange, that are comparable to Perlite). Coir is an 82 agricultural waste and, therefore, a renewable resource. However, it must be noted that 83 Coir is a material from a tropical crop produced in minimal geographical areas far from 84 maximum horticultural use and present the same Perlite problem on transportation. 85 Compost is an alternative to Coir 8, since it is possible to obtain it locally, avoiding 86 transportation. Compost is produced from local organic waste, which is highly available 87 at city levels. This is nutrient rich, therefore reducing the need of use of nutrient solution. 88 Its use as a substrate contributes to the recovery of organic waste resources and reduction 89 of dependency on nonrenewable substrates, such as Perlite. The recent increased interest 90 in urban agricultural activities highlights the timely need to investigate low environmental 91 impact substrates. Alternative urban organic substrates growing need to easy to manage, 92 have a low environmental impact, show high moisture retention, and have nutrients that 93 are readily available for crops. 94 95 Organic substrates have been widely studied for their use in the horticultural industry 9– 11 96 but not in the UA circular economy context. There is an urge in both horticultural 97 industry and gardening (including UA) to study organic materials derived from 98 agricultural, industrial, and municipal waste streams. The disposal of such organic (also 99 referred to as biodegradable) waste materials is an environmental problem 12, and their 100 reuse as substrates might provide a suitable solution 13. Compost from municipal organic 101 waste would specifically target reduction of urban organic waste to landfill and reuse, 102 towards a short-chain circular economy and contributing to sustainable development goal 103 (SDG) 11 (sustainable cities and communities) and SDG 2 (sustainable food production) 14 104 Considering a future scenario of low water availability, which can be addressed from 105 the use of organic substrates, the need arises to study the behavior of these substrates 106 under more restrictive conditions, in such a way, to generate strategies that allow 107 maintaining food production in situations where the use of water is restricted in urban 108 communities. 109 110 We hypothesize the use of Compost as an alternative growing media to Perlite and Coir, 111 in terms of comparable crop production and lower environmental impacts. A better 112 understanding of the use of urban and agricultural residues as substrates for urban crop 113 production will contribute towards meeting SGDs 2 and 11 and urban circular economy 114 approach. The objectives are to determine the agronomic feasibility of using alternative 115 substrates for Perlite in an RTG in the context of UA and characterize the behavior of a 116 green leaf crop as an indicator of the substrates’ crop production performance under 117 conventional and restricted irrigation conditions 118 119 2. Materials and methods 120 2.1. Study site 121 The experiments were conducted in the rooftop greenhouse laboratory (i-RTG Lab), a 122 cropping system representing other UA projects developed 15. It is located in the 123 Environmental Science and Technology (ICTA-UAB) building on the campus of the 124 Universitat Autònoma de Barcelona. Protected cultivation is performed under a steel and 125 polycarbonate greenhouse structure. The weather conditions in the i-RTG Lab were 126 passively controlled. 127 128 2.2. Substrate characteristics 129 The present study focused on three growing media substrates. These consist of: Perlite as 130 control substrate, Coir, Compost from municipal vegetal wastes, and a (1:1) Mixture of 131 the vegetal Compost and Perlite. Growing media physical and chemical properties as 132 shown in Table 1. 133 134 2.3. Experimental design 135 The experiment consisted of monitoring substrates performance during 3 crop growing 136 cycles between spring and summer of 2018 in order to include warmer seasonal periods. 137 Internal and external meteorological conditions of the i-RTG were recorded (Datalogger 138 model CR3000; Campbell Scientific Inc., USA), a summary of the information in Table 139 2. 140 141 The 3 crop cycles are considered independent (Table 3). The 3 substrates selected were 142 tested under conventional irrigation (supplying all water requirements) and water 143 restricted conditions (cut the irrigation to reach -20 cbar) in triplicates during each three 144 crop cycle, except Perlite, which was tested in duplicates: two under conventional 145 conditions and the other water restricted conditions. 146 147 A 70 m2 study area within the 125 m2 i-RTG facility was used. Growing bags’ 148 dimensions were 0.4 m x1.0 m, and they had a volume of 40 L. Each plant was planted at 149 distances of 0.2 m x 0.4 m resulting in 4 plants per growing media bag, three bags per 150 rows, and 0.5 meter distance between rows, as shown in Figure 1. Oak leaf lettuces 151 (Lactuca sativa L. var. Intybacea) seedlings were planted in the 4 substrates. Crop growth 152 was monitored and growing conditions controlled, following conventional agronomic 153 guidelines for lettuce production). During the growing periods, diseases and deficiencies 154 detected in the crops were also monitored. 155 156 2.4. Irrigation management 157 The Nutrition solution was provided to the lettuces via a drip fertigation system. The 158 nutrient solution contained: HNO3 63 mg∙L-1, KPO4H2 136 mg∙L-1, KNO3 101 mg∙L-1, 159 K2SO4 174 mg∙L-1, Ca(NO3)2 164 mg∙L-1, CaCl2 111 mg∙L-1, Mg(NO3)2 148.3 mg∙L-1, 160 and microelements 0.1 mg∙L-1.) Irrigation volumes were adapted and optimized to the 161 needs of lettuces grown in the control media (perlite substrate). They ranged between 0.3- 162 0.45 L∙day-1plant-1. 163 Induced water restriction took place 20 days after transplanting to make a late temporary 164 drought when the plants were fully developing and required a higher water and nutrient 165 supply following the methodology proposed Kerbiriou et al. 16. This restriction was 166 applied by completely stopping irrigation until the perlite bags reached -20 cbar. At this 167 time, irrigation was reestablished for all the water restricted rows. Water tension in the 168 growing media was determined with an analog 12 cm tensiometer (irrometer® MLT) 169 through the hydric potential variation, with a range of 0 to -40 cbar. Also, in the third crop 170 cycle, a second hydric restriction was performed. This second restriction consisted of the 171 same irrigation stoppage, but it was only maintained until the tensiometers in the control 172 substrate perlite reached -10 cbar. 173 174 2.5. Crop system monitoring conditions; irrigation data collection and substrate 175 physicochemical analysis 176 To the water flow characterization, a daily sampling was performed in each repetition of 177 the tests, and the amount of irrigated water, the leachates drained, and its conductivity 178 was measured (Table 4). The characterization of the substrate was made by the bulk 179 density test (at the start and end of the completely essay), this was performed for each 180 growth medium by the ring method (USDA, 1999); indicating the amount of water, the 181 porosity and the bulk density. It was also shown the differential of the water content in 182 each substrate between the conventional irrigation treatment and the restriction irrigation 183 treatment ( %W- %S). 184 185 2.6. Crop sampling 186 Five crop samples were taken randomly from different repetitions of each treatment. At 187 the end of the test, when the crop was harvested, the final yield was determined (g of the 188 commercial part of lettuce, Table 5). For the commercial weight, 5 plants in each crop 189 line were selected randomly, while for the water content determination, 3 plants were also 190 selected randomly in each line. 191 192 2.7. Statistical analysis 193 The crop measurements were expressed using average values and standard deviations. 194 “R” version 3.1.2 software 17 was used to determine significant differences between the 195 different substrates and the effect of water restriction. The significance was tested using 196 a one-way analysis of variance (ANOVA). Before the statistical analysis, the assumptions 197 of ANOVA were checked by a Shapiro-Wilk and Levene. Multiple comparisons of the 198 means were determined by a post hoc Duncan test. When the data were not normally 199 distributed, a Kruskal-Wallis test was used. 200 201 3. Results and Discussion 202 The commercial production and the crop development were analyzed, and a difference 203 was detected between the lettuces in the different substrates by comparing the first tests 204 to the second test. Within the third test it was possible to appreciate a lower variability 205 between the yields of plants irrigated conventionally and with water restriction. In 206 addition, a trend towards a reduction in yield, regarding on the applied water restriction 207 (-10 and -20 cbar). 208 209 3.1. 3.1. Substrate characteristics 210 At the end of the three consecutive experiments, the bulk density (BD) for each growing 211 medium was determined and is shown in Table 4. This information allows us to describe 212 each substrate’s porosity system behavior under conventional irrigation and water 213 restriction. The Coir presented an 81.76% water content, the Perlite showed a 14.43% 214 water content, and the vegetable Compost and substrate Mixture showed 31.62% and 215 25.29% water content, respectively. The Coir showed the lowest value for the BD, with 216 0.09 g∙cm-3, followed by Perlite, Mix, and Compost, the latter with 0.23 g∙cm-3 (Table 4). 217 218 Perlite: In this study, it was not possible to see final compaction of this growing media, 219 which was possible in all the other substrates. The leachates’ conductivity in the 220 conventional irrigation ranged between 0.86 mS∙cm-1 and 1.90 mS∙cm-1 depending on the 221 percent drainage or the water consumption plants. 222 223 Coir: The amount of water at the end of the assay for the conventional irrigation Coir 224 was 71.17% (Table 4). Compared to the conventionally irrigated Perlite (57.00%), there 225 was a 14% higher WC in the Coir. Additionally, the Coir showed the smallest BD of all 226 the growing media used in this assay, with 0.1 g∙cm-3. In other word present a very low 227 weight and high water retention, characteristic desirable in a substrate. The Coir’s 228 conductivity was constant throughout the study, ranging between 1.59 mS∙cm-1 and 1.70 229 mS∙cm-1. 230 231 Vegetable Compost: The conductivity on the first day of the first test was 3.80 mS∙cm- 232 1 , which decreased over time, down to 3.40 mS∙cm-1 and 2.10 mS∙cm-1 at the beginning 233 of the two other tests. At the end of the study, the Compost leachates had a conductivity 234 of 1.90 mS∙cm-1. 235 236 Mixture: The substrate Mixture indicated values ranging between Compost and Perlite, 237 for the EC`s leachates (2.27 mS∙cm-1) and the BD (0.17 g∙cm-3), indicated in Table 4. 238 During the experiment, the leachate’s EC had the same decreasing tendency reported in 239 the Compost substrate. Besides, to understand the behavior over time, the final water 240 content (WC) was evaluated together with the measures obtained daily with the 241 tensiometers placed in each substrate. 242 243 Effect of water restriction on the substrates 244 The Coir showed a 32% higher water content (  conventional irrigation treatment - 245   restricted water treatment); in this sense, the Mixture showed poor performance, at 246 18%. The vegetable Compost and Perlite had a performance of approximately 25% and 247 22%, respectively. 248 Due to each growing media’s different hydric curves, the point of restriction was not the 249 same for all of them (the minimum hydric potential reached in each substrate was 250 different) because the period of no irrigation was the same in all the substrates (Figure 2). 251 For example, during the first test, when the Perlite presented 19 cbar, the Coir and 252 Compost presented -23 and 4, respectively. The restriction period was different 253 throughout the three tests (Table 3) due to the temperature increase during the study, with 254 each test showing higher temperatures than the previous test. This induced the same water 255 stress levels in less time. 256 257 Perlite: Focusing on the tensiometers, the perlite water holding capacity (WHC) 258 remained constant through the 3 tests, with a progressive release of water content over 259 time. When water restriction was induced, the percent drainage variation occurred in 260 hours compared to the other substrates, which took approximately 2 days. After the 261 restriction period, the EC was very high at 3.45 mS∙cm-1, 2.95 mS∙cm-1, and 2.52 mS∙cm- 1 262 for each consecutive test, indicating a higher concentration of salts in the bag. Moreover, 263 it was detected that the major differences in the leachate electrical conductivities of the 264 conventional irrigated and restricted perlite bags were related to the duration of the 265 restriction periods and not just to the hydric tension of the substrate. As previously 266 explained, the temperatures increased throughout the second and third tests, reaching the 267 limiting - 20 cbar in shorter periods. Lower EC values in the second test (9 days without 268 irrigation) and the third test (7 days without irrigation) compared to the first test (14 days 269 without irrigation) once irrigation was restored. 270 271 Coir: In tests 1 and 2, the Coir showed a slow response to water restriction, but when the 272 matrix potential ranged between 5 and 8 cbar, it decreased rapidly. The results obtained 273 using the water retention curve, a high percent available water (27.60%), especially easily 274 available water (23.06%). Its water loss was more progressive than the Perlite since 275 Perlite has 19.49% available water and 8.25% easily available water. Moreover, the stress 276 response measure decreased in the last test with the same hydric demand, and the growing 277 media presented less tension in the pore system. It can explain by the collapse of coarse 278 porosity through the different processes of irrigation and drought in the essay, creating a 279 more complex porosity with a normalized pore distribution, which would explain its 280 behavior during test number 3. The treatment with restricted water present a constant EC 281 throughout the study, and no differences were detected in the conventionally irrigated 282 substrates. 283 284 Vegetable Compost: Through monitoring with tensiometers, the Compost showed a low 285 response to hydric potential in tests 1 and 2 (4 and 17.5 cbar). In the last test, the Compost 286 had similar behavior to Perlite in both water restrict treatments (-10/- 20 cbar, with 15 287 and 25 cbar for Compost and 11 and 20 cbar for Perlite, Figure 2). Compost presented a 288 similar available water content (20.67%) as Perlite, but the percent easily available water 289 was higher (17.23%). The restricted Compost’s final water content was similar to that of 290 Perlite and the water content when the Compost was conventionally irrigated. The 291 increase of BD (0.23 to 0.3 g∙cm-3) can be explained by the general irrigation management 292 of the essay was adjusted to the perlite demands. This could have meant a higher irrigation 293 input during tests 1 and 2, which could have favored particles’ arrangement, and for 294 concomitance, the increment of the bulk density. Moreover, when irrigation was stopped, 295 no leachates were detected, and after the water restriction period, the EC was the highest 296 among the substrates. 297 Specifically, in the first test, the leachates were detected 6 days after irrigation was 298 reestablished, and the conductivity was 4.70 mS∙cm-1. For the other two tests, the 299 leachates were recovered after two days, and the conductivity was 3.70 mS∙cm-1 and 2.20 300 mS∙cm-1. Nevertheless, this finding highlights that at the end of each test, the leachates’ 301 EC was the same in the restricted crops as in the conventionally irrigated crops. 302 303 Mixture: For hydric potential, the Mixture showed an intermediate performance between 304 the Compost and the Perlite in the water restricted treatment during the first test, while 305 during the second and third tests, it showed a high response to hydric potential, with a 306 lower value (27 cbar) compared to the control (Perlite with 20 cbar) (Figure 2). This is 307 consistent with results obtained in the water retention curve as the percent available water 308 and the easily available water ranged between the values obtained in the Compost and the 309 Perlite (19.20% and 11.79%, respectively). This could be explained by the Mix having a 310 poor water content performance. The conventionally irrigated and restricted Mixture 311 substrates had the lowest WC values (31% and 49%, respectively, compared to the Perlite, 312 at 34% and 51%, respectively), confirming the relationship of low water content and low 313 hydric potential (a lower value more strongly strengthens the stress due to the fact that 314 the hydric potential is tension). The BD remained constant over time, being unaffected 315 by the irrigation treatment, and showed an average value of 0.17 g∙cm-3; the BDs of the 316 Compost and Perlite at the start of the essay were 0.23 and 0.11 g∙cm-3, respectively. In 317 the end, the Compost showed slight compaction (0.29 g∙cm-3), but this was not the case 318 in the Mixture. The EC was similar, but its behavior was closer to that of the Perlite than 319 the Compost. The Mixture had the same pattern as the Compost in the electric 320 conductivity, but the conductivity was approximately the average between the EC in the 321 Compost and the Perlite. For example, at the beginning of the three tests, the EC was 2.50 322 mS∙cm-1 (test 1), 3.30 mS∙cm-1 (test 2), and 1.80 mS∙cm-1 (test 3). The highest values after 323 the restricted period were 2.87 mS∙cm-1 (test 1), 2.36 mS∙cm-1 (test 2) and 2.20 mS∙cm-1 324 (test 3). Nevertheless, as shown, the differences between the conventionally irrigated 325 treatment and the restricted treatment are smaller than those of the Compost. 326 327 3.2. Crop production 328 The crop yields ranged from 245.7 to 490.0 g∙plant-1, and some differences were detected 329 due to the substrates, the effect of water restriction, and the meteorological conditions. 330 Studies have shown that the water content of a media has a direct influence on the fresh 331 weight gain by lettuce plants 18. The main result is that in all three alternative substrates 332 studied, commercial productions were obtained; therefore, they could be used in UA. As 333 expected, when the crops suffered under a water restriction period, production decreased, 334 but the magnitude of these losses was different among the substrates. 335 336 Conventional irrigation 337 During the first test (April), when the crops were irrigated appropriately, no significant 338 differences (p<0.05) in the yield were observed among the substrates. The yield obtained 339 ranged between 422.7 and 445.7 g∙plant-1. Comparing tests 1 and 2, different results were 340 obtained. In test 2, when appropriated irrigation was applied, the crops grown on the 341 Mixture and coir substrates obtained statistically the same production as the control 342 (which is the substrate with the highest production: 490.0 g∙plant-1), and the Compost 343 presented the lowest production (423.9 g∙plant-1, 14% less weight). In 3 test, the best 344 results in the conventional irrigated crops were obtained with Compost (408.7 g∙plant-1) 345 and Mixture (418.4 g∙plant-1). Compared to the substrate with the highest obtained weight 346 (Mixture), the Coir presented the lowest production (370.1 g∙plant-1), -11.5% less. The 347 behavior of the Compost was notably different from those of the other substrates. The 348 lettuce grown in the Compost presented successive decreasing weights with the three 349 consecutive tests (Table 5). This difference could be due to the fact that in the first test, 350 the vegetable Compost was new and could provide a large quantity of nutrients to the 351 lettuce. However, in May, a nutrient depletion was detected by measuring the leachates’ 352 electric conductivity, as noted in the previous section. The Compost leachates’ EC ranged 353 between 4.77 and 3.60 mS∙cm-1 in April, between 3.70 and 2.47 mS∙cm-1 in May, and 354 between 2.8 mS∙cm-1 and 2.2 mS∙cm-1 in July. Furthermore, a compaction of the substrate 355 was detected, which could be a further reason for the production decrease 19. 356 357 Water stress effect on the yield 358 Some differences were detected when the crops were under water restriction. Compared 359 to the control, the Mixture, and the coir substrates in the first test, the plants grown in the 360 Compost reached higher weights (322.6 g∙plant-1). These results agree with Mastouri et 361 al.19, who detected that the growth of lettuce increased as the contribution of either type 362 of Compost in the growing media increased, where the production in the soil obtained in 363 that study was 282.2 g∙plant-1, and the highest production using a Mixture of tree bark 364 Compost was 308.9 g∙plant-1. These results demonstrate that vegetable Compost from 365 urban green waste is a competitive agronomic option for use in UA. Thus, the Compost 366 was able to provide some buffering capacity to the temporary drought. The Coir did not 367 reduce stress in the lettuce as much as expected based on the material’s high water- 368 retention capacity. Previous studies have suggested that the yield decrease could have 369 been due to excessive osmotic stress from the combined effects of the drought and the 370 high salinity of the media, which would not have been reflected in the tensiometer 371 readings, as these only report matric potential (not osmotic) 20. As previously shown, the 372 leachates’ EC did not change after the water restriction periods, which could cause a 373 concentration of salts in the substrate. 374 375 In the second test, in all the treatments, compared to the conventional irrigated crops, the 376 water restricted crops’ production decreased and was statistically the same between 377 treatments. In this case, the Compost results were worse than expected. First, the lettuce 378 presented the same weight as the other substrates, and the benefits detected in the previous 379 test were not detected here. Second, because the other three restricted substrates presented 380 an increase in production compared to the first test (25-30%), Compost’s production was 381 similar to that in the first test (320 g). 382 383 Compared to the previous tests, during the third test, the higher temperatures induced a 384 more rapid appearance of water stress (Figure 2). Whether the water restriction reached - 385 10 cbar or -20 cbar, the lowest production was obtained in the perlite bags. When the 386 restriction reached -20 cbar, the Mixture and the Compost substrates presented the best 387 results (295.2 and 284.9 g∙plant-1, respectively). Nevertheless, when the restriction did not 388 exceed -10 cbar, the crops grown in the Coir, and the Mixture reached the highest 389 production values (358.9 and 350.3 g∙plant-1). These results could have been perceived 390 when analyzing the water loss curves of the different substrates. As shown in the previous 391 section, in the first test, the Coir took a long time to lose water; however, when dryness 392 begins, water loss occurs very quickly and can damage crop production. 393 394 3.3. Relevance in UA 395 The consumption model within cities is characterized by being unidirectional, where 396 inputs and outputs flow prevail. 21 Firstly, diverse externalities are generated by an 397 extractives model that feeds on natural resources, and secondly, it increases new spaces 398 for agricultural production to satisfy the city’s requirements. From a circular economy 399 perspective, favor exchanges of flows between urban subsystems are part of the solution 400 to migrate to sustainable cities 22. The use of Compost and the Mixtures of substrates 401 derived from it, responds to these needs since at the city level 1; it reduces and values 402 municipal solid waste (MSW), 2; favors the recycling of nutrients, 3; as a substrate to 403 improve physical, chemical and biological properties of the culture medium. 404 (1) By using organic matter from MSW its amount will be reduced and, therefore, the 405 greenhouse gases emitted in landfill disposal 23. In this sense, the advantages can be 406 seen at the UA level and interact with more elements within the city 24. Viaene et al., 407 (2016) 25 present a study in Belgium about the opportunities and barriers of Compost 408 at the farm level, where they recommended 5 measures towards using Compost. 409 However, the study is carried out for farm conditions, the recommendations are 410 applicable in a circular economy context in the city. The third recommendation refers 411 to searching for new alternative sources of biomass from other industries to produce 412 Compost. It is possible to find different stakeholders at the city level that can regularly 413 provide biomass, such as greengrocers and coffee shops, among others. The integration 414 of agricultural production in the city would maintain a stable Compost production over 415 time due to its possible interconnections to other industries. Furthermore, research has 416 been made on Composting with common inorganic waste from the city, which has 417 shown good results, such as disposable diapers and biochar, among others. 26–28. 418 (2) By Composting the organic matter, it stabilizes, and the nutrients are available again 419 to produce new vegetables 29,30. Since nutrients are a limited resource, reincorporating 420 and reusing them in the production system is vital for the UA’s sustainable 421 development. 422 (3) The incorporation of Compost (total or partially) as a replacement for commercial 423 substrates can decrease the CO2 emissions, depending on the origin of the replaced 424 substrate. As an example, for this, in Spain, close to 80% of the Perlite used comes 425 from Turkey, South Africa, Greece, Uganda, and United Kingdom (35%, 18%, 10%, 426 8%and 8% respectively), where the reduction in the transport item could result in an 427 environmentally better process. Studies suggest that under proper Compost 428 management, environmental impacts are reduced 31. 429 430 In the present study, the vegetable Compost and the Mixture of vegetable Compost with 431 Perlite are suitable substrates in horticulture, especially in RA. Besides, it has been 432 observed that these substrates have better characteristics for preventing hydric stress in 433 summer, despite previous studies showing that Compost production could decrease due 434 to salt concentrations 19. 435 436 The yield was markedly competitive and higher than that of Perlite in April and July. 437 Lettuce is a moderately sensitive crop to salinity, similar to most of the RA crops: pepper, 438 tomato, and spinach, among others. Therefore, the results obtained in this study could be 439 directly applied to other horticultural crops 440 441 442 4. Conclusions 443 This analysis quantified lettuce’s agronomic performance grown in organic substrates, 444 including their resilience to water restriction. In the circular city context, the study of the 445 agricultural performance of environmentally friendly substrates (the recycled organic 446 municipal waste in cities) can contribute to RA implementation in Mediterranean urban 447 areas. Previous studies have addressed the environmental impacts and benefits of using 448 Coir and Compost as substrates 6,32 while considering the development of crop and 449 commercial production. Our results show that the studied organic substrates, Coir, as a 450 commercial substrate, and vegetable Compost alone or in a substrate Mixture with Perlite 451 1:1, could be used in UA, as they obtained similar or higher production than the control 452 substrate (Perlite). In summer, the best results were obtained with vegetable Compost 453 alone (408.7 g∙plant-1) and Compost mixed with Perlite (1:1) (418.4 g∙plant-1). 454 Nevertheless, a sequential decrease in the fresh lettuce weight grown in Compost in the 455 three tests was detected, probably due to the substrate’s loss of nutrients. 456 457 We also presented the first quantitative assessment of the agronomic response of water 458 restriction conditions in three alternative substrates of Perlite in the urban Mediterranean 459 context. We found that compared to Perlite, the organic substrates improved the 460 conditions against applied water restriction and increased the crops’ yield. Specifically, 461 the Coir tended to take a long time to lose water; however, when dryness begins, it occurs 462 very quickly, and commercial production decreases if drought induces water stress of -20 463 cbar, the Compost and the Mixture of Compost and Perlite present remarkable agronomic 464 resilience. 465 466 These results contribute to UA’s knowledge and preventive measures of droughts in 467 Mediterranean cities with quantified data. In the current climate change context, with 468 increasing droughts in summer, commercial systems that utilize Compost as a growing 469 media could reduce irrigation frequency, save water without increasing the substrate’s 470 salinity, and still produce commercially relevant yields. However, further studies should 471 determine the optimal irrigation and nutrient inputs for improving the growing conditions 472 to increase harvest weight. Additionally, this optimization could minimize the 473 environmental impacts for a lettuce unit, and kg of lettuce should be studied in a detailed 474 life-cycle assessment. More research is needed to contribute to this substrate’s 475 environmental analysis, considering the substrates’ lifespan and their waste management 476 impacts. 477 478 5. Acknowledgements 479 The authors thank the National Commission for Scientific and Technological Research 480 (Chile) for the grant awarded to F. Parada (PFCHA-CONICYT 2018 – Folio 72180248), 481 and the Spanish Ministry of Economy and Competitiveness (MINECO) for the financial 482 support of the research project “Integrated rooftop greenhouses: energy, waste and CO2 483 symbiosis with the building. Towards foods security in a circular economy” (CTM2016- 484 75772-C3-1-R) and the María de Maeztu program for Units of Excellence in R&D 485 (CEX2019-000940-M). The authors are grateful to the Secretaria d'Universitats i Recerca 486 for the award of a research scholarship (FI-DGR 2014) to Mireia Ercilla Montserrat, and 487 the Spanish Ministry of Economics, industry and competitiveness (MINECO) for 488 awarding a research scholarship under the Fertilecity project (CTM2016-75772-C3-1-R) 489 to Verónica Arcas (Investigador Predoctoral FPI 2018). 490 491 6. Conflict of Interest Statement 492 The authors do not present any type of conflict of interest. 493 494 495 7. Author contribution 496 All authors were responsible for the conception and design of the study conceived, M. 497 Ercilla-Montserrat, N. Carazo, J. I. Montero, X. Gabarrell, G. Villalba, J. Rieradevall, E. 498 Lopez-Capel, P. Muñoz, the original idea for the study. M. Ercilla-Montserrat, F. Parada, 499 V. Arcas-Pilz, and P. Muñoz set up, supervised and acquired the data for the experimental 500 tests. N. Carazo for laboratory analysis. F. Parada and M. Ercilla-Montserrat processed 501 and analyzed the data and took the lead in writing the manuscript. All authors critically 502 revised the draft for important intellectual content. All authors gave their final approval 503 to the manuscript. 504 505 8. Reference 506 507 1. OECD. Water Resources Allocation [Internet]. 2015. 144 p. Available from: 508 https://www.oecd-ilibrary.org/content/publication/9789264229631-en 509 2. Sanyé E, Oliver-Solà J, Gasol CM, Farreny R, Rieradevall J, Gabarrell X. Life cycle 510 assessment of energy flow and packaging use in food purchasing. Journal of Cleaner 511 Production. 2012;25:51–9. 512 3. Thomaier S, Specht K, Henckel D, Dierich A, Siebert R, Freisinger UB, et al. 513 Farming in and on urban buildings: Present practice and specific novelties of Zero- 514 Acreage Farming (ZFarming). Renewable Agriculture and Food Systems. 2015 515 Feb;30(01):43–54. 516 4. Nadal A, Llorach-Massana P, Cuerva E, López-Capel E, Montero JI, Josa A, et al. 517 Building-integrated rooftop greenhouses: An energy and environmental assessment 518 in the mediterranean context. Applied Energy. 2017;187:338–51. 519 5. Montero JI, Baeza E, Heuvelink EP, Rieradevall J, Muñoz P, Ercilla M, et al. 520 Productivity of a building-integrated roof top greenhouse in a Mediterranean 521 climate. Agricultural Systems. 2017 Nov 1;158:14–22. 522 6. Barrett GE, Alexander PD, Robinson JS, Bragg NC. Achieving environmentally 523 sustainable growing media for soilless plant cultivation systems – A review. Scientia 524 Horticulturae. 2016 Nov 22;212:220–34. 525 7. Savvas D, Gianquinto G, Tuzel Y, Gruda N. Soilless Culture. FAO Plant Production 526 and Protection Paper No. 217: Good Agricultural Practices for Greenhouse 527 Vegetable Crops. 2013; 528 8. Ulm F, Avelar D, Hobson P, Penha-Lopes G, Dias T, Máguas C, et al. Sustainable 529 urban agriculture using compost and an open-pollinated maize variety. Journal of 530 Cleaner Production. 2019 Mar 1;212:622–9. 531 9. Roehrdanz M, Greve T, de Jager M, Buchwald R, Wark M. Co-composted 532 hydrochar substrates as growing media for horticultural crops. Scientia 533 Horticulturae. 2019 Jun 27;252:96–103. 534 10. Urrestarazu M, Salas MC, Padilla MI, Moreno J, Elorrieta MA, Carrasco GA. 535 Evaluation of different composts from horticultural crop residues and their uses in 536 greenhouse soilless cropping. Acta Hortic. 2001 Mar;(549):147–52. 537 11. Verdonck O, Penninck R, De Boodt M. The physical properties of different 538 horticultural substrates. Acta Hortic. 1984 Jun;(150):155–60. 539 12. European Commission. Communication from the commission to the European 540 parliament, the council, the European economic and social committee and the 541 committee of the regions - A new Circular Economy Action Plan For a cleaner and 542 more competitive Europe. [Internet]. 2020 [cited 2020 Nov 17]. Available from: 543 https://eur-lex.europa.eu/legal- 544 content/EN/TXT/?uri=COM%3A2020%3A98%3AFIN 545 13. European Commission. Environment- Biodegradable waste [Internet]. [cited 2020 546 Nov 17]. Available from: 547 https://ec.europa.eu/environment/waste/compost/index.htm 548 14. UN. Sustainable Development Goals. Zero Hunger(Goal 2). [Internet]. United 549 Nations - Sustainable Development Goals. 2015 [cited 2020 Nov 17]. Available 550 from: https://www.un.org/sustainabledevelopment/hunger/ 551 15. Ercilla-Montserrat M, Muñoz P, Montero JI, Gabarrell X, Rieradevall J. A study on 552 air quality and heavy metals content of urban food produced in a Mediterranean city 553 (Barcelona). Journal of Cleaner Production. 2018 May; 554 16. Kerbiriou PJ, Stomph TJ, Van Der Putten PEL, Lammerts Van Bueren ET, Struik 555 PC. Shoot growth, root growth and resource capture under limiting water and N 556 supply for two cultivars of lettuce (Lactuca sativa L.). Plant and Soil. 2013 Oct 557 24;371(1–2):281–97. 558 17. R Development Core Team. R: A language and environment for statistical 559 computing. R Foundation for Statistical Computing, Vienna, Austria.; 2014. 560 18. Valença D da C, Carvalho DF de, Reinert F, Azevedo RA, Pinho CF de, Medici LO, 561 et al. Automatically controlled deficit irrigation of lettuce in “organic potponics.” 562 Scientia Agricola. 2018 Jan;75(1):52–9. 563 19. Mastouri F, Hassandokht MR, Padasht Dehkaei MN. The effect of application of 564 agricultural waste compost on growing media and greenhouse lettuce yield. Acta 565 Horticulturae. 2005 Dec;(697):153–8. 566 20. Wallach Rony. Physical Characteristics of Soilless Media. In: M. Raviv JHL, editor. 567 Soilless Culture: Theory and Practice. 2008. p. 41-116. 568 21. World Economic Forum. Circular Economy in Cities Evolving the model for a 569 sustainable urban future [Internet]. Switzerland; 2018 Mar. Available from: 570 http://www3.weforum.org/docs/White_paper_Circular_Economy_in_Cities_report 571 _2018.pdf 572 22. Lucertini G, Musco F. Circular Urban Metabolism Framework. One Earth. 2020 Feb 573 21;2(2):138–42. 574 23. Lou XF, Nair J. The impact of landfilling and composting on greenhouse gas 575 emissions – A review. Bioresource Technology. 2009 Aug 1;100(16):3792–8. 576 24. De Corato U. Agricultural waste recycling in horticultural intensive farming systems 577 by on-farm composting and compost-based tea application improves soil quality and 578 plant health: A review under the perspective of a circular economy. Science of The 579 Total Environment. 2020 Oct 10;738:139840. 580 25. Viaene J, Van Lancker J, Vandecasteele B, Willekens K, Bijttebier J, Ruysschaert 581 G, et al. Opportunities and barriers to on-farm composting and compost application: 582 A case study from northwestern Europe. Waste Management. 2016 Feb 1;48:181– 583 92. 584 26. Colón J, Mestre-Montserrat M, Puig-Ventosa I, Sánchez A. Performance of 585 compostable baby used diapers in the composting process with the organic fraction 586 of municipal solid waste. Waste Management. 2013 May 1;33(5):1097–103. 587 27. Espinosa-Valdemar RM, Sotelo-Navarro PX, Quecholac-Piña X, García-Rivera 588 MA, Beltrán-Villavicencio M, Ojeda-Benítez S, et al. Biological recycling of used 589 baby diapers in a small-scale composting system. Resources, Conservation and 590 Recycling. 2014 Jun 1;87:153–7. 591 28. Guo X, Liu H, Zhang J. The role of biochar in organic waste composting and soil 592 improvement: A review. Waste Management. 2020 Feb 1;102:884–99. 593 29. Jack A, Thies J. Compost and Vermicompost as Amendments Promoting Soil 594 Health. In: Uphoff N, editor. Biological Approaches to Sustainable Soil Systems 595 [Internet]. CRC Press; 2006 [cited 2020 Aug 14]. p. 453–66. (Books in Soils, Plants, 596 and the Environment; vol. 20060863). Available from: 597 http://www.crcnetbase.com/doi/abs/10.1201/9781420017113.ch31 598 30. Harrison RB. Composting and Formation of Humic Substances. In: Jørgensen SE, 599 Fath BD, editors. Encyclopedia of Ecology [Internet]. Oxford: Academic Press; 600 2008 [cited 2020 Aug 14]. p. 713–9. Available from: 601 http://www.sciencedirect.com/science/article/pii/B9780080454054002627 602 31. Martínez-Blanco J, Colón J, Gabarrell X, Font X, Sánchez A, Artola A, et al. The 603 use of life cycle assessment for the comparison of biowaste composting at home and 604 full scale. Waste Management. 2010 Jun 1;30(6):983–94. 605 32. Martínez-Blanco J, Muñoz P, Antón A, Rieradevall J. Life cycle assessment of the 606 use of compost from municipal organic waste for fertilization of tomato crops. 607 Resources, Conservation and Recycling. 2009 Apr;53(6):340–51. 608