From urine to food and oxygen: effects of high and low NH4 +:NO3 - ratio on lettuce cultivated in a gas-tight hydroponic facility

In situ production of food, water and oxygen is essential for long-duration human space missions. Higher plants represent a key element in Bioregenerative Life Support Systems (BLSS), where crop cultivation can be based on water and nutrients recovered from waste and wastewater. Human urine exemplifies an important waste stream with potential to provide crops with nitrogen (N) and other nutrients. Dynamic waste composition and treatment processes may result in mineralized fractions with varying ammonium (NH4 +) to nitrate (NO3 -) ratios. In this study, lettuce was cultivated in the unique ESA MELiSSA Plant Characterization Unit, an advanced, gas-tight hydroponic research facility offering controlled environment and continuous monitoring of atmospheric gas composition. To evaluate biological and system effects of nutrient solution NH4 +:NO3 - ratio, two crop tests were run with different NH4 + to total N ratio (NH4 +:N) and elevated concentrations of Na+ and Cl- in line with a urine recycling scenario. Plants cultivated at 0.5 mol·mol-1 NH4 +:N (HiNH4 +) achieved 50% lower shoot biomass compared to those cultivated at 0.1 mol·mol-1 NH4 +:N (LoNH4 +), accompanied by higher shoot dry weight content and lower harvest index. Analyses of projected leaf area over time indicated that the reduced biomass observed at harvest could be attributed to a lower specific growth rate during the close-to-exponential growth phase. The HiNH4 + crop produced 40% less O2 over the full cultivation period. However, normalization of the results indicated a marginal increase in O2 production per time and per projected leaf area for the HiNH4 + crop during the exponential growth phase, in line with a higher shoot chlorophyll content. Mineral analysis demonstrated that the biomass content of NH4 + and NO3 - varied in line with the nutrient solution composition. The ratio of consumed NH4 + to consumed N was higher than the NH4 +:N ratio of the nutrient solution for both crop tests, resulting in decreasing NH4 +:N ratios in the nutrient solution over time. The results provide enhanced insight for design of waste processes and crop cultivation to optimize overall BLSS efficiency and hold valuable potential for improved resource utilization also in terrestrial food production systems.

In situ production of food, water and oxygen is essential for long-duration human space missions. Higher plants represent a key element in Bioregenerative Life Support Systems (BLSS), where crop cultivation can be based on water and nutrients recovered from waste and wastewater. Human urine exemplifies an important waste stream with potential to provide crops with nitrogen (N) and other nutrients. Dynamic waste composition and treatment processes may result in mineralized fractions with varying ammonium (NH 4 + ) to nitrate (NO 3 -) ratios. In this study, lettuce was cultivated in the unique ESA MELiSSA Plant Characterization Unit, an advanced, gas-tight hydroponic research facility offering controlled environment and continuous monitoring of atmospheric gas composition. To evaluate biological and system effects of nutrient solution NH 4 + :NO 3 ratio, two crop tests were run with different NH 4 + to total N ratio (NH 4 + :N) and elevated concentrations of Na + and Clin line with a urine recycling scenario. Plants cultivated at 0.5 mol·mol -1 NH 4 + :N (HiNH 4 + ) achieved 50% lower shoot biomass compared to those cultivated at 0.1 mol·mol -1 NH 4 + :N (LoNH 4 + ), accompanied by higher shoot dry weight content and lower harvest index. Analyses of projected leaf area over time indicated that the reduced biomass observed at harvest could be attributed to a lower specific growth rate during the close-to-exponential growth phase. The HiNH 4 + crop produced 40% less O 2 over the full cultivation period. However, normalization of the results indicated a marginal increase in O 2 production per time and per projected leaf area for the HiNH 4 + crop during the exponential growth phase, in line with a higher shoot chlorophyll content. Mineral analysis demonstrated that the biomass content of NH 4 + and NO 3 varied in line with the nutrient solution composition. The ratio of consumed NH 4 + to consumed N was higher than the NH 4 + :N ratio of the nutrient solution for both crop tests, resulting in decreasing NH 4 + :N ratios in the nutrient solution over time. The results provide enhanced insight for design of waste processes and crop cultivation to optimize overall BLSS efficiency and hold valuable potential for improved resource utilization also in terrestrial food production systems. KEYWORDS ammonium, nitrate, recycling, oxygen, food production, photosynthesis, life support system

Introduction
Human spaceflight missions to remote locations call for on-site food production and resource recirculation. In contrast to Low Earth Orbit, complete resupply of resources from Earth is challenged by increased mission duration and travel distances. Higher plants represent a key element for regeneration of air, water and food for astronauts (Wheeler et al., 2002;Pannico et al., 2022), either as part of stand-alone Bioregenerative Life Support Syste ms (BLSS), o r in a combination with physicochemical methods. The MELiSSA project (Micro Ecological Life Support System Alternative) is an international collaboration led by the European Space Agency (ESA), emphasizing the development of a BLSS to support long-term space missions. In an integrated loop, waste streams from a crew compartment are to be processed by thermophilic, heterotrophic and nitrifying bacteria to provide input to photosynthetic compartments with higher plants and algae to regenerate mineralized nutrients and carbon dioxide (CO 2 ), and produce food, pure water, and oxygen (O 2 ) for the crew (Godia et al., 2002). An important objective of the MELiSSA project is to develop mathematical models that can predict critical processes such as plant growth, photosynthesis, and transpiration, based on parameters such as plant species, cultivation conditions, and the processed waste stream fed to the plants.
In a scenario of crop cultivation based on mineralized human waste, urine is a valuable resource due to its high nitrogen (N) content. N is an essential element and a critical input factor for crop cultivation, with profound effects on plant growth and development as a constituent of numerous vital compounds such as nucleic acids, chlorophyll, amino acids, proteins, ATP, phytohormones, auxin and cytokinins (Marschner and Marschner, 2012;Taiz, 2015;De Bang et al., 2021). The use of recycled organic waste for crop cultivation typically requires mineralization of organic compounds by physicochemical and/or microbial processes. N is primarily taken up by plant roots as inorganic ammonium (NH 4 + ) and nitrate (NO 3 -) (Marschner and Marschner, 2012;Taiz, 2015), that may be derived from urine. While fresh urine is rich in urea-N (up to 12 g L -1 ), this compound is rapidly hydrolyzed upon storage in a non-sterile environment to NH 4 + /NH 3 that may be further converted to NO 3 by physicochemical and/or microbial processes (Udert et al., 2006;Larsen et al., 2021). As the result of this upstream treatment depends on the mineralization strategy and the process conditions, the resulting nutrient solution provided to the plants may vary in the absolute concentrations of NH 4 + and NO 3 and the ratio between them. This may in turn affect not only the plant uptake rates of these inorganic N species, but also other physiological and metabolic processes such as uptake of other nutrients, enzyme activity, photosynthesis and respiration, water balance, signaling pathways, leaf expansion, and root architectureeventually influencing the overall plant growth and crop yield (Guo et al., 2007a;Guo et al., 2007c;Andrews et al., 2013;Liu and Von Wiren, 2017). A nutrient solution containing both NH 4 + and NO 3 is typically preferred to optimize plant growth and development rather than using either NO 3 or NH 4 + as the sole source of N, although the use of NH 4 + as the dominant N source should be avoided to reduce risk of ammonium toxicity and reduced plant growth (Roosta and Schjoerring, 2008;Taiz, 2015;Song et al., 2021;Weil et al., 2021;Hameed et al., 2022). Beyond N, urine contains other plant macronutrients such as potassium, phosphorus, sulfur, and lower levels of calcium and magnesium (Udert et al., 2006;Larsen et al., 2021). On the other hand, feeding crops with processed urine introduces non-essential elements such as sodium (Na + ) and chloride (Cl -). Plants naturally accumulate salts, and the presence of Na + and Clin the nutrient solution may affect plant physiology and growth. At some concentrations, NaCl has been demonstrated to act as a eustressor with beneficial effects on crop quality Rouphael et al., 2018). However, high concentrations of salt is among the most limiting factors for plant growth and may cause injurious abiotic stress, altering morphological and physiological plant traits and ultimately reducing crop yield (Chinnusamy and Zhu, 2006;Acosta-Motos et al., 2017). To make possible a new generation of scientific studies related to the higher plant compartment, one of the most complex compartments in the MELiSSA loop, a Plant Characterization Unit (PCU) was recently designed and assembled on the premises of Department of Agricultural Sciences of University of Naples Federico II in Italy as part of the ESA MELiSSA PaCMan (PlAnt Characterization unit for closed life support systemengineering, MANufacturing & testing) activity (Pannico et al., 2022). The PCU is a gas-tight cultivation chamber for characterization of higher plants, offering extensive monitoring and control of the cultivation environment, including its separated hydroponic and atmospheric loops.
Lettuce is a highly relevant species for bioregenerative life support systems with high harvest index, efficiency (per area, time, and volume), and potential yield (edible biomass, O 2 and water), combined with relatively modest horticultural requirements (Berkovich et al., 2004). Lettuce has been successfully cultivated onboard the International Space Station (Khodadad et al., 2020) and represents a healthy and nutritious addition to the human diet, rich in vitamin C, antioxidants, polyphenols and dietary fiber (Khodadad et al., 2020). Furthermore, it is one of the most used vegetables in terrestrial hydroponic cultivation systems (Zhu et al., 2020), illustrating the benefits of a deeper understanding of plant responses to nutrient solutions derived from organic waste in commercial crop production also on Earth, with the potential to improve resource utilization and reduce environmental impact of industrial food production systems.
The underlying hypotheses and scenario of this study was that strategies and operating conditions of upstream waste processing will affect the composition of the mineralized nutrients fed to the plants in a life support system. As these strategies and conditions may be designed and controlled, but also give rise to fluctuations and dynamics, it is critical to understand and predict consequences of nutrient solution composition on crop cultivation. More specifically, urine ammonification and nitrification may result in nutrient solutions with varying concentrations of NH 4 + and NO 3 -, including their relative ratio, which will impact downstream crop cultivation. Furthermore, beyond the incoming nutrient streams, nutrient utilization rates inside the plant compartment will dictate the evolution of the nutrient solution in a closed-loop scenario with complete recycling of water, nutrients and non-nutrients. This study aimed at investigating biological and system effects of high and low NH 4 + :N ratios on lettuce cultivated in a urine recycling scenario, exploiting the recently established PCU facility offering controlled hydroponic cultivation conditions and state of the art monitoring of critical parameters such as O 2 production.
2 Materials and methods 2.1 Lettuce seedling production About 100 lettuce seeds (Lactuca sativa L. cultivar 'Grand Rapids') were disinfected in 5% sodium hypochlorite solution for 15 minutes followed by three rinsing cycles in ultrapure water. The disinfected seeds were dispersed onto absorbent paper, moist with ultrapure water and incubated at room temperature under indirect lighting. After 48 hours, germinated seeds with about 1 cm radicle were transferred to Styrofoam trays with vermiculite soaked in nutrient solution prepared according to Peiro et al. (2020) at 0.5X strength, and incubated at 24°C in a dedicated nursery with 65% relative humidity (RH), a light/dark regime of 16/8 hours and a light intensity of 400 ± 50 mmol·m -2 ·s -1 (PPFD). Ten days after sowing, 18 equally developed seedlings (evaluated by visual inspection) were selected from the pool of 100 germinated seeds. Each seedling was installed into a bottomless 50 mL test tube (CLS430829, Corning, NY, USA) filled with rockwool (Delta 4G 42/40, Grodan, Roermond, The Netherlands) cut into a cylindrical shape and vertically split in two. The seedling's root system was sandwiched between the two halves of the rockwool column and inserted into the 50 mL tube. The top surface of the rockwool was then covered with grafting mastic (Cortimax, Cisa Adriatica S.r.l., Pescara, Italy) to minimize gas exchange between the root and shoot zones of the cultivation chamber. PPFD and spectral composition were recorded by eighteen spectral scans (one per plant position, measured at a height of 3 cm above the medium surface) using a spectral radiometer (MSC15, Gigahertz-Optik, Türkenfeld, Germany). The largest deviation between this average light intensity and the intensity at any of the plant positions was 68 μmol·m -2 ·s -1 (PPFD), corresponding to 16% of the average intensity. The air temperature and RH were automatically controlled as described by Pannico et al. (2022), at 26°C and 50% during the day and 20°C and 70% during the night, respectively. The atmospheric CO 2 concentration was controlled at 1000 ppmv (parts per million by volume) by automatic injection of pure CO 2 during the day phase, while it was allowed to increase during the night phase due to respiration. Atmospheric O 2 concentration started at ambient levels and increased based on plant photosynthesis. The atmospheric O 2 concentration was not controlled beyond one chamber venting per crop test (Day 19) to avoid O 2 levels rising above 25%.

Nutrient solution preparation, monitoring, and maintenance
The nutrient solutions were based on the recipe described by Peiro et al. (2020)  Water quality of the recirculating nutrient solution was continuously monitored, including pH (2 x Endress+Hauser Orbisint CPS11D Memosens, Germany), electrical conductivity (EC; 2 x Endress+Hauser Indumax CLS50D, Germany), dissolved O 2 (Mettler Toledo InPro 6860i, Switzerland), dissolved CO 2 (Labolytic optical CO 2 sensor 0-15 mg/L, Trondheim, Norway), and nitrate (Endress+Hauser Viomax CAS51D, Germany, installed in an at-line configuration with auto-sampling and auto-dilution of the nutrient solution). Nutrient solution temperature was continuously monitored and automatically controlled at 18 ± 0.5°C. For both crop tests, pH was adjusted to 5.9 at startup using 0.5 M H 2 SO 4 and then maintained at 5.9 by automatic addition of 0.5 M KOH throughout both crop tests (no addition of acid necessary during the cultivation period). Throughout both crop tests, the actual pH of the nutrient solution typically deviated less than 0.05 pH units from the set-point, and never more than 0.2 pH units. During both crop tests, nutrient stock solutions A and B were automatically fed at a 1:1 ratio to maintain a constant NO 3 concentration. The recirculating nutrient solution was mechanically filtered in three steps with reducing pore size from 1.6 to 0.45 μm, using filter cartridges (PP3 Sartopure -Sartobran, Sartorius, Goettingen, Germany). The filtration unit was implemented between the plant cultivation chamber and the stirred tank holding the nutrient solution that was continuously fed to the plants.

Data collection, sampling, and analysis
Atmospheric concentrations of O 2 and CO 2 inside the cultivation chamber were continuously measured by an O 2 /CO 2 gas analyzer (California Analytical Instruments, 700 LX NDIR/O 2 , Orange, CA, USA). Additionally, the O 2 and CO 2 concentrations of the pressure compensation system's gas tank were measured at crop test start and end, and before and after chamber venting. The amount of atmospheric gas inside the PCU was determined continuously based on ideal gas law, a total atmospheric volume of 4.9 m 3 , and continuously measured gas pressure and temperature. This allowed computation of daily atmospheric leak rates and plant O 2 production (Pannico et al., 2022). To keep the air humidity constant, the water produced by the plant transpiration was condensed and collected (Pannico et al., 2022). During the crop tests presented in this study, the injection rates of CO 2 and water into the atmospheric loop were inaccurate, unfortunately preventing accurate mass balances for CO 2 and water and thereby also preventing calculation of consumed amounts of CO 2 and produced amounts of water. Top-view images of the 1.2 m × 1.5 m cultivation chamber floor were acquired every hour (FLIR Blackfly S BFS-U3-200S6C-C @ 70 dpi), and projected leaf area as observed from above was calculated for images acquired every second day based on image segmentation by color. Projected Leaf Area Index (PLAI) was calculated as the measured projected leaf area of all 18 plants (m 2 ) per total cultivation area (m 2 ) (Zheng and Moskal, 2009). SPAD index for indication of chlorophyll content (average of ten measurements per plant; SPAD-502, Minolta Corp. Ltd, Osaka, Japan), dark-adapted chlorophyll fluorescence for the maximum quantum yield of photosystem II (F v /F m ) (average of two measurements per plant; Plant Stress Kit, Opti-Sciences), and leaf color index (average of three measurements per plant; Chroma Meter CR-400, Konika Minolta) were measured immediately after opening the cultivation chamber at day of harvest. Fresh weight of shoot and stem, and leaf number were determined for each plant by destructive sampling. Additionally, total leaf area of each plant was determined from top-view images of the plant's individual leaves (Canon EOS 90D + Tamron SP 35 mm @ 59 dpi) based on image segmentation by color.
The lead taproot of each plant was isolated and imaged (Canon EOS 90D + Tamron SP 35 mm @ 294 dpi for LoNH 4 + and 334 dpi for HiNH 4 + ) in a tray with water. The lead taproot was defined as the medial root with the largest diameter in the basal zone, not considering the diameter of the laterals. Root length, diameter and number of links and forks were analyzed with WinRhizo (Pro 2021a, Regent Instruments Inc.). Following counting of leaf number and determination of shoot fresh weight (FW) of each plant, shoots and roots were oven-dried at 60°C until constant weight was reached for determination of dry weight (DW) per plant. Total plant DW was calculated as shoot DW + root DW, while harvest index was calculated as shoot DW per total plant DW. Dried samples of shoots and roots were analyzed for total carbon (C) and N content via combustion analysis by a Micro Elemental Analyser UNICUBE ® (Elementar, Langenselbold Hesse, Germany) equipped with a thermal conductivity detector. Leaf soluble cations and anions were determined by liquid ion exchange chromatography (ICS-3000, Dionex, Sunnyvale, CA, USA) as described by Pannico et al. (2019).
Every two days, nutrient solution samples (25 mL) were extracted from the hydroponic loop for offline analysis. Nutrient ion concentrations were determined by ion exchange chromatography (ICS 3000 Dionex) and the results were smoothed by moving average to filter out noise from analysis uncertainty. N mass balance was calculated based on measured parameters at the start and end of each crop test, including concentrations of inorganic N in the nutrient solution (measured via optical sensor and ion chromatography methods described above), total liquid volume, total shoot and root biomass and their N content, in addition to the amount of nutrient stock solutions introduced and samples removed throughout the crop test.
Statistical analyses were performed with SPSS (IBM SPSS version 28.0.0.0), Anova with Tukey HSD post-test, tests of normality, Independent Samples t-test and Mann Whitney U test.

Results
Building on the background presented in the introduction, the results section first presents the experiment rationale with respect to nutrient solution composition and cultivation conditions. Representing the first scientific paper detailing crop cultivation results from the recently established PCU facility, plant growth and selected scientific datasets offered by this facility are then illustrated on a high level, before being explored in more detail to present effects of crop cultivation in low vs. high NH 4 + :N ratio.

Urine recycling scenario: Experiment rationale and nutrient solution design
The experiment rationale was built on the hypothesis presented in the introduction that both design and natural dynamics of upstream waste processing may result in mineralized waste streams with different NH :N) represents nitrification without increased alkalinity and thus limited by redox, as reviewed by Larsen et al. (2021). At the same time, these scenarios represent considerable different, low and high NH 4 + :N ratios for lettuce cultivation (Zhu et al., 2020;Du et al., 2022;Hameed et al., 2022) while still avoiding the use of NH 4 + as the dominant N source to limit the potential of ammonium toxicity and thereby allow acceptable growth (see Materials and methods).
To allow direct comparison of results, total N concentration was identical at startup of both crop tests (11 mM). Capitalizing on a state-of-the-art nutrient monitoring and control system, the NO 3 concentration was kept constant (9.7 mM in LoNH 4 + ; 5.6 mM in HiNH 4 + ) throughout each of the two crop tests to minimize effects of the absolute NO 3 concentration itself. This allowed the NH 4 + :N ratio to change throughout the cultivation period depending on the total consumption of NH 4 + and NO 3 in the system, demonstrating the development of the NH 4 + :N ratio over time. In line with a urine recycling scenario, the crop tests were performed with an elevated NaCl concentration of 5 mM, representing a Na:N concentration ratio of 0.46 mol·mol -1 . This ratio was designed based on typical Na:N ratios of mineralized urine. While fresh urine is typically characterized by 0.17 mol·mol -1 Na:N and Cl:N (adapted from Udert et al., 2006), several reports illustrate ratios of approximately 0.25 mol·mol -1 in urine-based fertilizer solutions (adapted from Bonvin et al., 2015;Mauerer et al., 2018;Halbert-Howard et al., 2021). However, the final composition of the mineralized nutrient solutions depends on urine storage and ammonification and nitrification processes. As urine is prone to ammonia volatilization, resulting Na:N and Cl:N ratios may be considerably elevated -exemplified by the real stored urine fraction by Udert and Wachter (2012) with Na:N and Cl:N ratios of 0.5 mol·mol -1 . Thus, the Na:N and Cl:N concentration ratios used in this study represent a urine-based nutrient solution, accounting for effects by urine storage and processing. However, effects of potential Na + accumulation over time in a closed loop system are not accounted for in this study, as such predictions would benefit from more insight in Na uptake ratios as function of nutrient solution composition and cultivation conditions.
Beyond NH 4 + /NO 3 and Na + /Cl -, the two crop tests were run with similar concentrations of macro-and micronutrients, except for SO 4 2used for charge balancing of NH 4 + /NO 3 -. The total strength of the nutrient solutions was designed so that the total electrical conductivity of the LoNH 4 + nutrient solution at start was identical to that of Peiro et al. (2020)

Crop cultivation in the ESA MELiSSA Plant Characterization Unit
The results reported in this paper were obtained in the recently established ESA MELiSSA Plant Characterization Unit (PCU), in which 18 lettuce plants were hydroponically cultivated as deep-water culture for 28 days after transplant (DAT), to a total age of 38 days, across a 1.8 m 2 plant cultivation area (Figure 1). Extensive monitoring and control systems provided stable cultivation conditions, including atmospheric temperature and RH, and nutrient solution temperature, pH and NO 3 concentration (see Materials and methods, Table 1; Figure 3). The closed nature of the atmospheric loop combined with continuous monitoring of atmospheric O 2 and CO 2 concentrations represent the basis for determination of O 2 production and CO 2 consumption by the plants. Atmospheric leak rates in percent gas volume per hour were calculated daily and demonstrated a stable and high system tightness with an average leak rate of 0.082 ± 0.017%·h -1 and 0.095 ± 0.022%·h -1 for the LoNH 4 + and HiNH 4 + crop tests, respectively. The atmospheric O 2 concentration increased during the day phase resulting from photosynthesis, while it decreased during the night phase due to respiration ( Figure 2

Root and shoot biomass and morphology at harvest
In line with the differences demonstrated by non-destructive PLAI analysis, detailed inspection of the plants after destructive sampling at 28 DAT illustrated a total leaf area of the HiNH 4 + crop which was 50% lower than that of the LoNH 4 + crop (3.5 m 2 vs. 7.0 m 2 , respectively, for the 18 plants;  Figure 5. In average, the lead taproots of HiNH 4 + were shorter (55% reduction in total length, including both mother root and lateral roots), thinner (15% reduction in average diameter) and less branched (57% less forks) than those of the LoNH 4 + plants.

Leaf chlorophyll content, color index and F v /F m at harvest
Dark-adapted chlorophyll fluorescence measurements of the maximum quantum yield of photosystem II (F v /F m ) were identical for both treatments (Table 3), while an elevated SPAD index indicated a higher chlorophyll content of the leaves of the HiNH 4 + plants. Color measurements illustrated significant differences as evaluated by the L*a*b color space, with the HiNH 4 + leaf color shifted more towards green compared to the color of the LoNH 4 + leaves.

Nutrient ion and element composition of shoots and roots at harvest
Nutrient ion and element analyses of lettuce shoot and root biomass harvested at 28 DAT demonstrated considerable differences between plants cultivated in the two different nutrient solutions ( . This includes the cations K + (52% lower shoot content and 36% lower root content), Na + (32% lower shoot content and 22% lower root content), and Mg 2+ (47% lower shoot content). Ca 2+ illustrated the same trend (55% lower shoot content and 17% lower root content in HiNH 4 + ) although the differences were not statistically significant due to high variability in the results. Reduced nutrient ion content of HiNH 4 + plants could also be observed for the anions HPO 4 2-(44% lower shoot content) and Cl -(72% lower root content), while no statistically significant differences could be observed for SO 4 2-.
3.4 Crop performance and system dynamics at high and low NH 4 + :NO 3 ratios

O 2 production
Continuous monitoring of the atmospheric O 2 and CO 2 concentrations inside the cultivation chamber illustrated O 2 production during the day (when CO 2 was controlled at 1000 ppm by automatic CO 2 injection), and O 2 consumption and CO 2  production during the night (Figure 2). Detailed inspection of the datasets illustrates an immediate increase in atmospheric CO 2 concentration and a corresponding decrease in O 2 starting from the time at which the cultivation chamber LEDs were turned off, as illustrated in Figure 6A, detailing the first two days after the chamber venting.  SPAD index 20 ± 3 23 ± 2 p < 0.01 Color index L 53 ± 3 51 ± 3 Not sign.
Color index a -18.8 ± 0.5 -18.3 ± 0.5 p < 0.01 Color index b 34.3 ± 1.3 32.6 ± 1.8 p < 0.01 All values are tabulated as mean value ± standard deviation. Statistical significance between the two treatments is indicated as p-values. and during the second phase (both 1.1 mol·d -1 ·m -2 ). Furthermore, relating the total O 2 production to the total shoot DW at harvest also demonstrates similar values for the two crops, with 37 mol O 2 produced (over 26.6 days) per kg shoot DW for both crops. Estimated daily O 2 production throughout the crop tests was calculated based on the O 2 concentration in the cultivation chamber, the leak rate, and estimated daily gas composition of the atmospheric pressure compensation system (see Materials and methods). As the latter introduces a modest uncertainty in the calculations of daily O 2 production, the resulting datasets were smoothed by polynomial regression ( Figure 6B). Evaluating the differences in daily O 2 production at 10, 14, 18, 22 and 26 DAT based on these datasets, the HiNH 4 + plants produced 26,37,40,42 and 45% less compared to the LoNH 4 + plants, respectively.
To account for the considerable difference in biomass between the two treatments, the daily O 2 production (smoothed dataset) was evaluated per projected leaf area (PLA), measured non-destructively throughout the crop tests. These results ( Figure 6B)

Nutrient solution development
From start, the total N concentration was the same in both crop tests (11 mM; Table 1), while the NH 4 + :N ratios were different (0.1 mol·mol -1 for LoNH 4 + , 0.5 mol·mol -1 for HiNH 4 + ). Based on an advanced PCU nutrient management system, automatic feeding of nutrient stock solutions (with the same NH 4 + :N ratios as from start, thus 0.1 or 0.5 mol·mol -1 , see Materials and methods) successfully maintained a constant NO 3 concentration throughout both crop tests ( Figure 3A), at 9.7 mM for LoNH 4 + and 5.6 mM for HiNH 4 + (  Figure 2B) in the nutrient solution was most profound for K + , with a considerable increase from 3.5 to 7.4 mM during the HiNH 4 + crop test (KOH was automatically added to maintain pH at setpoint). In contrast, K + concentration decreased by 0.6 mM over the LoNH 4 + crop test.
Other ions demonstrating a concentration change of more than 0.2 mM over the course of a crop test (beyond K + and N ion species) included Ca 2+ , Na + and Cl -, demonstrating an increase during the LoNH 4 + crop test (requiring considerable feeding of nutrient stock

N mass balance
With a considerable nutrient solution volume (270 L), a limited cultivation area (1.8 m 2 ), and a crop with modest nutrient requirements (lettuce), mass balances are vulnerable to measurement inaccuracies especially with respect to ion concentrations in the nutrient solution. Nevertheless, a N mass balance was calculated (

Incorporation of C and N into edible biomass
To illustrate C and N recycling in a BLSS perspective, the amounts of C and N incorporated into edible biomass (18 plants per crop) was calculated for both crop tests (Figure 7) based on measured shoot DW at harvest (Table 2), and their composition ( In sum, the HiNH 4 + crop incorporated 48% less N (7 g) into total edible biomass compared to the LoNH 4 + crop (13 g).

Discussion
During long duration human space missions, bioregenerative life support systems may provide food, regenerated water, and O 2 for the astronauts. Given the limited availability of nutrient resources for plant production in this scenario, recycling and reuse of nutrients such as N from human urine is important. Depending on the upstream mineralization process of urine, the nutrient solution may have various NH 4 + :NO 3 ratios and elevated NaCl concentrations. To maximize production of food, O 2 and clean water, crop cultivation conditions should be optimized. However, as a BLSS consists of multiple processes, the performance of the regenerative loop including ammonification and nitrification needs to be evaluated and optimized as a whole. This illustrates a possible trade-off between the overall BLSS efficiency and crop cultivation and yield, which should be considered in a long-term perspective with continuous operation. Increased knowledge on plant responses to various optimal and sub-optimal cultivation conditions and nutrient solutions will contribute to improved computer models, BLSS efficiency and improvement in design and control of integrated processes linking waste management and crop cultivation.

Biomass production and morphology
In this research campaign, considerable differences were identified between lettuce plants cultivated in nutrient solutions containing high and low NH 4 + :N ratios. In a BLSS perspective, the results demonstrate that both NH 4 + :N ratios, representing considerably different scenarios of upstream waste treatment, allow plant growth and development, and thus fixation of C and N into edible biomass. There is, however, a clear impact on biomass productivity (produced amount per time) and therefore on the total production over a given time. Under the conditions tested during the 28-day crop tests performed in this campaign, the HiNH 4 + crop provided 40% lower edible biomass (shoot dry weight), resulting in the incorporation of 35% less C and 48% less N into edible biomass compared to the crop cultivated at low NH 4 + :N. These results are in line with previous research on lettuce, illustrating that increased ammonium concentrations strongly affect growth, resulting in decreased root and shoot biomass accumulation (Qiu et al., 2014;Wenceslau et al., 2021;Du et al., 2022;Hameed et al., 2022). Furthermore, the results are in line with general effects of ammonium toxicity on plant growth, demonstrating a biomassreducing effect of high ammonium concentrations attributed to for example oxidative stress, nutrient imbalances and reduced photosynthesis (Barickman and Kopsell, 2016;Song et al., 2021;Weil et al., 2021). In line with the biomass results, the total leaf area per plant at time of harvest was reduced by a 2-factor for the HiNH 4 + plants compared to the LoNH 4 + plants. This effect is supported by other reports on hydroponically grown lettuce and other plants, demonstrating lower leaf areas with increasing ammonium concentrations (Guo et al., 2007a;Guo et al., 2007b;Urlic et al., 2017;Wenceslau et al., 2021). Beyond the NH 4 + :N ratio, other potential stress factors include the nutrient solution's total ionic strength and the presence of Na + and Cladded to mimic a urine recycling scenario. While several studies indicate no or minor growth retardation of lettuce cultivated in the presence of seawater, elevated electrical conductivity and/or Na + and Clconcentrations similar to the cultivation conditions used in the presented work (Atzori et al., 2019 and references within), one should not rule out combined effects of salinity and NH 4 + :N ratio especially in a foreseen future closed-loop scenario with accumulating Na + and Clconcentrations and increasing total electrical conductivity. Calculation of PLAI throughout the crop tests offered a method to evaluate plant growth and PLAI-based growth rate without opening the closed cultivation chamber. While noticeable differences in PLAI (and thus biomass) between the HiNH 4 + and LoNH 4 + crops could not easily be observed until after approximately 10 days of cultivation, calculation of μ PLAI indicated a difference in specific growth rate between the two crops starting already in the early cultivation phase. This indicates that under the conditions tested, the elevated NH 4 + :N ratio introduced negative effects on plant growth also during the closeto-exponential cultivation period. At the same time, this indicates that the observed NO 2 in the nutrient solution was not the main cause of the observed differences in biomass production (and thus O 2 production, see below). While elevated NO 2 concentrations in the nutrient solution may cause reduced growth (Hoque et al., 2008), NO 2 was not observed until the late cultivation phase of the HiNH 4 + crop, at a time which μ PLAI of both crops had started to decrease and coincided. Coinciding μ PLAI towards the end of the cultivation period may be explained by reduced specific growth rates as plants mature, and the fact that the presented growth rates are based on PLAI which, by definition, cannot exceed a factor of 1 due to a limited cultivation area and overlapping leaves. Beyond the overall reduction in biomass production of the HiNH 4 + crop, these plants displayed a significantly higher root:shoot ratio than the LoNH 4 + plants, indicating a higher relative use of energy towards root growth. This is in line with previous research showing increasing root:shoot ratio with increasing NH 4 + :NO 3 ratio in several species, including lettuce (Zhu et al., 2020), tobacco (Walch- Liu et al., 2000) and cucumber (Zhou et al., 2017). More specifically, alteration of biomass partitioning between root and shoot is a known plant response to nutritional stress (Maskova and Herben, 2018) and NH 4 + :K +

FIGURE 7
Incorporation of C (black) and N (blue) into shoots (green) and roots (gray) of the LoNH 4 + (upper) and HiNH 4 + (lower) crops. All numbers represent element or biomass dry weight in gram and relate to the full crop consisting of 18 plants. The relative difference in total pie area corresponds to the relative difference in total dry weight of the two crops.
imbalance has been demonstrated to negatively influence carbohydrate accumulation, disturbing the root-to-shoot biomass partitioning (Zhao et al., 2020). In our study, NH 4 + and K + were among the nutrient ions that demonstrated the greatest differences between the biomass content of the HiNH 4 + and LoNH 4 + crops. Additionally, the two crops demonstrated differences in root size and morphology. The HiNH 4 + plants exhibited shorter primary and secondary roots, along with a significantly lower number of forks and links illustrating less branching. This is likely to be caused by the nutrient availability (especially of NH 4 + and NO 3 -) and plant nutritional status, which are known to strongly influence the development of the root system (Lima et al., 2010). NH 4 + is known to affect root architecture by inhibition of primary root growth and elongation, while promoting lateral branching (Liu and Von Wiren, 2017;Liu et al., 2013). Conversely, NO 3 has been shown to stimulate lateral root elongation (Lima et al., 2010;Marschner and Marschner, 2012). These observations represent valuable details for mathematical modelling of plant growth and development, as the root's surface area is critical for mass and energy transfer.

O 2 production
In total, the HiNH 4 + crop produced 41% less O 2 than the LoNH 4 + . However, as the growth of the HiNH 4 + plants was impaired, the O 2 production should be evaluated also in terms of specific production. When O 2 production was normalized by shoot dry weight at harvest, the amount of O 2 produced was 37 mol·kg -1 for both the LoNH 4 + and the HiNH 4 + crops. On a high level, O 2 production following biomass production is in line with the general mechanism of photosynthesis, converting light energy to chemical energy utilized for the synthesis and accumulation of organic matter (Evert and Eichhorn, 2013;Taiz, 2015), and hence, the photosynthetic rate is determining for biomass productivity. Furthermore, measurements of chlorophyll fluorescence on darkadapted leaves demonstrated that the F v /F m of both crops (0.83) was within a range indicating a healthy and well-functioning photosystem II (0.79 -0.84) (Bjorkman and Demmig, 1987 (Blanke et al., 1996;Qiu et al., 2014;Hu et al., 2015). However, the scientific literature on chlorophyll content as a response to NH 4 + concentration and NH 4 + :NO 3 ratio demonstrate complex relationships and varying results. Song et al. (2021) demonstrated that lettuce and cabbage chlorophyll content was higher in plants exposed to an NH 4 + :NO 3 ratio of 50:50, than both 0:100 and 100:0, while Zhu et al. (2020) reported a higher chlorophyll concentration in lettuce seedlings exposed to an NH 4 + :NO 3 ratio 25:75 than both 100:0, 50:50 and 0:100. Studies of other species also show varying results, for example, an increase in chlorophyll content with increasing NH 4 + :NO 3 ratios was not discovered in kale (Assimakopoulou et al., 2019). Different theories are proposed to explain an increased chlorophyll content as a response to high ammonium. The higher chlorophyll concentration may be caused by ammonium-stress inhibiting leaf expansion, resulting in a denser chlorophyll content rather than an actual total increase (Sanchez-Zabala et al., 2015). Furthermore, carbon skeletons in the leaves favor NH 4 + assimilation. Based on this, it has been proposed that a higher chlorophyll content might be a strategy to increase photosynthetic CO 2 assimilation to produce more carbon skeletons and thereby mitigate NH 4 + accumulation (Sanchez-Zabala et al., 2015).

Biomass composition
The 3 :N ratio in the nutrient solution was reported for example for rocket salad (Kim et al., 2006). The average nitrate content of the LoNH 4 + and HiNH 4 + edible biomass (shoots) was 6.0 and 0.8 g NO 3 -N per kg dry weight, respectively (Table 4). Considering the shoot dry weight content ( :NO 3 ratio in the nutrient solution Savvas et al., 2006;Roosta and Schjoerring, 2008;Fallovo et al., 2009). For example, Weil et al. (2021) reported that the concentration of both K + and Ca 2+ in lettuce shoots declined significantly with increasing nutrient solution NH 4 + :N ratio, concluding that the uptake of cationic nutrients and the plant growth were reduced when the NH 4 + :N ratio exceeded 0.50 mol·mol -1 . The observed effects on nutrient ion content have been explained by antagonism (Lasa et al., 2001;Marschner and Marschner, 2012); that the NH 4 + ion resembles the K + ion in radius and hydration shell size and may therefore move through K + channels (Marschner and Marschner, 2012), and that uptake of NO 3 occurs simultaneously with the uptake of Ca 2+ or K + so that an increasing proportion of NO 3 will increase the content of Ca 2+ and K + in leaves (Du et al., 2022). The content of NaCl in the nutrient solutions, which was designed to mimic a urine recycling scenario, may have impacted plant nutrient uptake in both the high and low NH 4 + -scenarios. In general, excess Na + uptake lowers the uptake of essential ions and disturbs the root-shoot transport. It has been widely observed that Na + , even at low levels, inhibits the transport systems for K + uptake due to the chemical similarities between the two ions (Amini and Ehsanpour, 2005;Kronzucker et al., 2013). For some functions, Na + may also replace potassium, such as for example for maintenance of vacuole osmotic potential (Wakeel et al., 2011). Na + may cause osmotic stress, which decreases the passive uptake of calcium, resulting in a lower concentration of calcium in the plant. Another possible effect of NaCl in plant nutrition, is that Clmay antagonize the uptake of nitrate, while it assists uptake of ammonium (Song et al., 2021 (Raab and Terry, 1994). Additionally, plants exposed to abiotic stress often accumulate reactive oxygen species (ROS). Accumulation of soluble sugars can mitigate this toxic effect caused by plant stress by contributing to ROS scavenging (Keunen et al., 2013;Peshev and Van Den Ende, 2013). Sugars can also protect chloroplasts and thereby stabilize photosynthesis under stress conditions, and they may serve as osmoprotectants . Additionally, the uptake of NO 3 against the electrochemical gradient requires more use of fixed carbon compared to acquisition of NH 4 + .

Nutrient consumption and nutrient solution evolution in a long-term perspective
Maximizing resource utilization limits the possibilities of flushing the hydroponic loop and restarting the system with a fresh and balanced nutrient solution. Thus, long term development of the nutrient solution needs to be understood, modelled, and ultimately designed and controlled in a trade off with the productivity of the plants and the design and performance of upstream processes.  (Glass et al., 2002). N preference may be linked to energy considerations as uptake of NO 3 requires energy while NH 4 + can be directly absorbed, hence many plants prefer to take up NH 4 + if both are available (Mengel and Kirkby, 2001;Marschner and Marschner, 2012;Du et al., 2022). Additionally, NO 3 uptake is inhibited by ammonium, and lettuce appears to absorb ammonium faster than NO 3 when the source contains both N-forms (Savvas et al., 2006;Britto and Kronzucker, 2013). However, plant preference for NH 4 + or NO 3 is known to vary across plant species, physiological phase and environmental conditions , and thus observed N preference may only be valid for a given species, developmental stage, and cultivation conditions. An improved understanding of the NH 4 + to NO 3 preference, for example as function of cultivation conditions, should be further pursued, as this may hold potential to reduce ammonium toxicity even at conditions of high NH 4 + :N ratios. During the LoNH 4 + crop test which required a considerable amount of nutrient stock solution feeding, accumulation of Na + , Ca 2+ and to a certain extent Clwas observed together with a marginal reduction of K + . This illustrates imbalance between the system's feed rates and consumption rates. Nutrient ion levels in the nutrient stock solutions may be tuned relative to for example total N as far as charge balancing allows in order to achieve a balanced nutrient solution. Accumulation of Na + and Cl -, however, represents a potential general challenge for the utilization of mineralized human waste. Accumulation of Na + in the nutrient solution may to a certain extent be mitigated by increased plant Na + uptake rates with increasing Na + concentrations in the nutrient solution (Neocleous and Savvas, 2017;Breśet al., 2022). Nevertheless, the extent and the implications of long-term accumulation of non-nutrients in crop cultivation based on human waste are important aspects that would benefit from further attention in a BLSS perspective. The ratio between consumed NH 4 + and consumed NO 3 affects nutrient solution pH. NH 4 + uptake causes release of protons and thereby reduced pH in the root zone, which again may impact nutrient uptake (Lasa et al., 2001;Marschner and Marschner, 2012;Weil et al., 2021). This effect was clearly demonstrated during the HiNH 4 + crop test with a substantial amount of base added to maintain a constant pH. The HiNH 4 + scenario, with an NH 4 + :N ratio of 0.5 mol·mol -1 is based on urine ammonification and partly nitrification without addition of alkalinity (Larsen et al., 2021). In a BLSS scenario, the results presented illustrate that no or low alkalinity addition during the nitrification step results in increased alkalinity requirement during the crop cultivation step instead. In this case, addition of OHduring crop cultivation should be carefully balanced with appropriate anions to avoid nutrient imbalance over time. In this context, it is interesting to notice that addition of macronutrient ions prone to be present at low concentrations in mineralized urine due to precipitation (such as Ca 2+ ) could possibly serve a dual purpose of being a counter ion for OHcharge balancing and being a required nutrient solution supplement. Such feeding strategies must, however, be carefully tuned and consider factors such as salt availability and solubility.

Strategies to maximize food and O 2 production in a BLSS perspective
In a Lunar greenhouse perspective, cultivation area will be limited and should be utilized in an efficient way. In a scenario of constant plant density throughout the cultivation period, plants should be cultivated to an adult stage, as the plant growth rate (g per day) and the O 2 production rate (mol per day) are highest during the end of the cultivation period, thus increasing also the average growth rate and the average O 2 production rate as evaluated over the full cultivation period. However, such a strategy comes at the cost of a low utilization rate of the available cultivation area (low PLAI), especially during the first part of the cultivation period. In a more optimized scenario, initial plant density can be high to achieve a high PLAI even with small plants. As the plants grow, the plant density can be reduced to give room for the expanding plants. In such an optimized scenario, specific production per time and PLA should be considered when aiming at optimizing food (biomass) and O 2 production. Under the conditions tested, the specific O 2 production per time and PLA reached a maximum around 13 DAT ( Figure 6B), indicating that adolescent lettuce plants were more effective in O 2 production per time and PLA than seedlings and adult plants. With O 2 production linked to biomass production, this period with the highest specific O 2 production per time and PLA coincides with the period of highest specific growth rate (here estimated as μ PLAI ; Figure 4). Thus, in an optimized scenario with dynamic plant density to continuously operate at a high PLAI, the results obtained under the conditions tested suggest that plants should be cultivated up to approximately 16 -20 days after transplant (26 -30 days after germination) to maintain a maximized specific biomass (food) production per time and PLA and a maximized specific O 2 production per time and PLA. Furthermore, it is interesting to notice the indication of a higher O 2 production per time and PLA of the HiNH 4 + crop compared to that of the LoNH 4 + crop ( Figure 6B), illustrating that as long as PLAI is kept high, O 2 productivity may be high even under cultivation conditions suboptimal for growth. Such an effect, that may be attributed to the plant's complex physiological response to mitigate suboptimal cultivation conditions, is especially interesting considering the possible need to accept suboptimal conditions for a given BLSS step in order to maximize the total efficiency of the whole loop.

Conclusion
The results of this study demonstrate the importance of understanding and considering direct and indirect effects of upstream waste processing on crop cultivation in BLSS. The urine utilization scenarios of 0.1 and 0.5 mol·mol -1 NH 4 + :N demonstrated significant effects on plant development, plant nutrient composition and O 2 production. Under the conditions tested, plants cultivated at the high NH 4 + :N ratio demonstrated a high preference for ammonium combined with a reduced specific growth rate and thus a reduced total biomass production. O 2 production per time and projected leaf area reached a maximum during the adolescent plant phase, coinciding with exponential growth and indicating a strong relationship between biomass and O 2 production. Interestingly, the specific O 2 production per time and projected leaf area was marginally higher for the plants cultivated at high NH 4 + :N ratio, in line with a higher chlorophyll content. In a life support perspective, the results illustrate design concepts for crop cultivation strategies depending on the performance of upstream processes (e.g. urine mineralization and high/low NH 4 + :N) and downstream needs (food or O 2 ). In addition, parameters such as plant density and cultivation duration may be adjusted or balanced against each other to optimize performance. Ultimately, the results may guide the design and control of both crop cultivation and other processes in a regenerative BLSS loop towards a common trade-off and to achieve a greater good rather than optimizing individual processes. The unique features of the ESA MELiSSA Plant Cultivation Unit (PCU) with its gas-tight atmospheric and hydroponic loops with extensive monitoring and control enabled collection of comprehensive and detailed data on crop O 2 production, plant growth via image-based analyses, and nutrient solution dynamics throughout the entire cultivation period. Together with future expanded capabilities of the PCU, such as analyses of CO 2 consumption and water production, this will contribute to improved understanding of plant responses and further advancement of computer models for predictions of plant growth and production in various BLSS and cultivation scenarios.
This study represents a further advancement towards increased waste-and resource recycling in plant-based food production systems. A deeper understanding of these processes and their impacts is crucial for the production of food and for the regeneration of water and O 2 in future human space exploration. Additionally, considering relevant terrestrial challenges such as depletion of mineral resources and pollution of ground waters due to nutrient runoff, nutrient recycling fuels sustainable agriculture also on Earth. Exploitation of waste streams for plant production illustrates synergies between space exploration and terrestrial food production, and knowledge on plant responses to different resource utilization scenarios improves the development and design of systems and processes for both Earth and space.

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions
ØJ, MS and A-IJ designed the experiment. MS, ØJ, AP and CQ executed the crop tests and performed analyses and calculations. All authors contributed to interpretation of results. MS and ØJ prepared the draft manuscript. All authors contributed to the article and submitted and approved the submitted section.

Funding
This research was funded by the European Space Agency (ESA) through the MELiSSA PaCMan 2 project (Plant characterization unit for closed life support system -engineering, manufacturing and testing), contract number 4000115852/15/NL/AT.