Operational Performance of a Staggered Multi-Crop Mini-Garden for Space Crop Production Using Zero-Discharge Reusable Root Modules

1. Introduction

Plant-based bioregenerative life-support systems are expected to play an important role in long-duration space missions by providing a renewable source of fresh food, contributing to O₂ production and CO₂ removal, and reducing dependence on food resupply from Earth [1]. For future missions to the Moon and Mars, crop-production technologies must be reliable, resource-efficient, and capable of supporting diet quality and crew well-being within confined and inhospitable environments [2,3,4,5]. As mission duration increases, efficient and reliable crop-production systems become increasingly important.

Space crop-production systems are constrained by limited mass, volume, power, crew time, water, nutrients, and data capacity, all of which must be carefully managed during mission operations [5,6]. Existing spaceflight plant-growth facilities range from simple biological experiments to closed-loop, feedback-controlled chambers, but they remain small relative to the scale required for substantial dietary replacement [1]. Systems such as Veggie and the Advanced Plant Habitat (APH) have provided essential platforms for space crop-production research and demonstration. APH is primarily research-oriented and provides tightly controlled environmental conditions, whereas Veggie has supported both research and crew consumption of fresh produce [7,8]. However, even APH, the largest plant-growth chamber currently used aboard the ISS, provides a limited crop-production area and root-zone volume of approximately 0.17 m² and 0.0086 m³, respectively [9].

Although current space crop-production systems remain supplemental because of their limited growing area, even small amounts of fresh produce can provide important dietary and psychological benefits for crew members [1]. This highlights the need for crew-consumption-oriented crop-production systems that can provide fresh produce repeatedly and reliably, even when they are not intended to replace the packaged diet [10,11]. For near-term ISS and early exploration-mission applications, the objective is therefore not necessarily complete food autonomy, but rather the development of small, reliable, and resource-efficient systems that can supply high-value fresh crops over time.

Future space crop-production systems must provide fresh food repeatedly rather than relying only on single crops and terminal harvests. An ISS mini-garden would need to support crop diversity, staggered planting, selective harvest, repeated replanting, and predictable water and nutrient demand within a small root-zone volume. Unlike crops emphasized in many early bioregenerative life-support studies, such as wheat, soybean, potato, peanut, and sweet potato, leafy greens, herbs, and compact fruiting crops are better suited for near-term supplemental food production because they have shorter production cycles, provide high-value fresh produce, and may improve diet quality despite contributing relatively little to total caloric intake [1,12]. However, less is known about how these crops perform when grown together, harvested selectively, and replanted repeatedly within a shared or reusable root-zone system.

Staggered planting and selective harvest may provide an operational strategy for extending fresh biomass availability while avoiding simultaneous peaks in crop water demand [13,14]. By distributing planting dates and harvest events across multiple root modules, a mini-garden system could provide smaller, more frequent harvests rather than a single large terminal harvest. This strategy may also reduce short-term demand on water delivery, humidity control, and condensate recovery systems by preventing all crops from reaching rapid growth stages at the same time. However, the success of this approach depends on crop selection, crop placement, canopy development, harvest timing, and compatibility among species with different growth rates and architectures.

The Utah Reusable Root Module (URRM) provides a useful platform for evaluating these questions because it consists of five independent root modules (RMs) that can be planted, harvested, and managed separately. This configuration makes it possible to stagger planting dates, separate crop species, and compare module-level responses over time. The URRM is also operated as a zero-discharge root-zone system, meaning that the nutrient solution is retained within the root module and there is no drainage or leaching during production [10]. Zero-discharge operation is desirable for space agriculture because it reduces waste and improves resource containment, but it also creates a nutrient-management challenge because salts and nutrients remain in the root zone unless they are taken up by plants or redistributed within the growth media [5,15]. Repeated use of the same growth media adds another operational challenge because roots, crop residues, rhizodeposits, microorganisms, and crop-derived compounds may persist across planting cycles [16,17,18]. These carryover effects can alter nutrient availability, microbial activity, media physical condition, and seedling establishment in later crops. Under these conditions, water use and bulk electrical conductivity (EC) must be interpreted together to evaluate whether repeated production is moving the root zone toward salt accumulation or nutrient depletion.

Despite increasing interest in supplemental fresh-food production for spaceflight, limited information is available on how multi-crop, repeated-harvest systems behave over several months in reusable peat-based media under zero-discharge conditions. In particular, few studies have evaluated how staggered crop calendars affect harvest distribution, module-level water demand, crop compatibility, replanting constraints, residue accumulation, microbial growth, media settling, seedling establishment, and root-zone nutrient dynamics across repeated planting cycles [11]. The consequences of repeated growth-media use remain especially underexplored for space crop-production systems, where media replacement, leaching, and extensive root-zone sampling are operationally constrained. This information is needed to develop practical crop calendars and management strategies for future ISS mini-garden systems.

The objective of this study was to evaluate the operational feasibility of a four-month staggered planting and selective-harvest strategy in a zero-discharge URRM simulator as a ground-based model for an ISS pick-and-eat mini-garden. Specifically, the study aimed to: (i) quantify harvest distribution and fresh biomass production across crops, root modules, and time; (ii) determine whether staggered planting distributed water-use peaks among modules; (iii) identify crop-placement and species-selection constraints during repeated production; and (iv) evaluate whether water-use and bulk Electrical Conductivity (EC) trends could be used to diagnose root-zone nutrient dynamics during long-term zero-discharge operation. The study also documented operational constraints associated with repeated use of peat-based media, including residue management, seedling establishment, media settling, and root-zone conditions that may contribute to carryover effects across crop cycles. By addressing these objectives, this study provides an operational framework for evaluating staggered multi-crop production in small, reusable root-zone systems for future space crop-production applications.

2. Materials and Methods

2.1. URRM Simulator and Root Module Configuration

The experiment was conducted using a five-module Utah Reusable Root Module (URRM) simulator configured for zero-discharge crop production (Figure 1). The simulator consisted of five independent RMs, each with a surface area of 759.3 cm² and a volume of 7,593 cm³. Together, the five RMs provided a total planting surface area of approximately 0.40 m² [10]. The RMs were operated as independent planting units, allowing crops to be staggered across modules and maintained at different developmental stages throughout the four-month trial.

Each RM was packed with peat-based growth media and contained two porous ceramic tubes for water and nutrient distribution [19]. Root-zone volumetric water content (θv) was monitored using two moisture sensors per module: a horizontally oriented CS650 sensor (Campbell Scientific Inc., Logan, UT, USA) and a vertically oriented TEROS ONE sensor (METER Group Inc., Pullman, WA, USA). The growth media and URRM design have been described previously [10,19] (Figure 1A).

The RMs were covered with a silicone-rubber containment material designed to retain the peat-based media within each RM under microgravity-relevant operating constraints, while also limiting direct evaporation from the substrate surface and allowing seedling emergence. The cover consisted of fringed silicone-rubber sheets (EMI Seals & Gaskets Ltd., Dorset, England) installed over the RM surface and held in place with aluminum strips (Figure 1B) [19]. Each module contained 20 planting positions, with each opening measuring 3.89 cm in length and 2.54 cm in width. At each opening, the silicone sheet was water-jet cut to create 11 flexible fringes on each side, for a total of 22 fringes per planting position. Individual fringes were approximately 2.3 mm wide and closed around the plant stem as seedlings emerged. The containment-cover design has been described previously [19,20].

The simulator was enclosed on four sides with removable aluminized cardboard panels that served as reflective side walls around the growing area. Ligh-emitting diode (LED) fixtures were positioned approximately 24 cm above the RM surface (Figure 1C). The system was operated without drainage or leaching; nutrient solution was replenished to maintain target root-zone θv, and system variables were monitored at 1-min intervals throughout the trial.

2.2. Environmental Conditions

The simulator was not operated inside a sealed growth chamber; therefore, air temperature and relative humidity were influenced by ambient room conditions. The photoperiod was maintained at 16 h light / 8 h dark throughout the experiment. Photosynthetic photon flux density (PPFD) was measured at the planting surface at three positions within each RM: front, middle, and back. Mean PPFD differed among RMs because of position within the simulator, ranging from 476 µmol m⁻² s⁻¹ in RM1 and RM5 to 630 µmol m⁻² s⁻¹ in RM3. Mean PPFD values were 476, 592, 630, 546, and 476 µmol m⁻² s⁻¹ for RM1, RM2, RM3, RM4, and RM5, respectively. The spatial variability in PPFD across RM positions is shown in Figure S1.

Air temperature and growth-media temperature were monitored during the trial. Growth-media temperature was measured continuously in each RM, while air temperature was monitored within the simulator. Relative humidity was measured periodically using a portable sensor, but continuous RH measurements were not available. Because PPFD and temperature differed among positions within the simulator, module position was considered when interpreting differences in crop growth, water use, and biomass production among RMs.

2.3. Crop-Production System and Staggered Planting Strategy

A four-month staggered crop-production trial was conducted to evaluate the URRM simulator as a ground-based model for an ISS pick-and-eat mini-garden. Trial day 1 corresponded to the first planting date. The crop calendar was designed to maintain multiple crop species at different developmental stages across the five independent RMs, allowing repeated planting, selective harvest, and module-level comparison over time. Representative images of the simulator on trial days 24, 38, and 50 illustrate the progressive development of the staggered multi-crop canopy across RMs (Figure S2).

The trial included seven crop species/cultivars selected to represent leafy greens, herbs, and a compact fruiting crop. These included Heartbreakers Twiggy Red Hybrid tomato (Solanum lycopersicum; Totally Tomatoes, Randolph, WI, USA), baby leaf red lettuce ‘Outredgeous’ (Lactuca sativa; Johnny’s Selected Seeds, Winslow, ME, USA), green lettuce ‘Grand Rapids’ (Lactuca sativa; Johnny’s Selected Seeds, Winslow, ME, USA), mizuna (Brassica nipposinica; Johnny’s Selected Seeds, Winslow, ME, USA), rainbow Swiss chard organic heirloom mix (Beta vulgaris subsp. cicla; Back to the Roots Inc., Oakland, CA, USA), ‘Premier’ kale (Brassica oleracea var. sabellica; Burpee Garden Products Co., Warminster, PA, USA), and bonsai basil (Ocimum basilicum; Seeds’n Such., Augusta, GA, USA).

Seeds were planted using triangular pullulan-polymer wedges inserted into the silicone-cover planting openings [13]. Two seeds were placed in each wedge. Although each RM contained 20 planting positions, a maximum of 10 positions were used during any planting cycle. Planting positions were alternated within each RM to leave unused positions available for subsequent crop cycles.

During the first planting cycle, crops were assigned to the five RMs as follows: RM1 was planted with mizuna and tomato on trial day 1; RM2 with red lettuce and basil on trial day 1; RM3 with mizuna and tomato on trial day 15; RM4 with green lettuce and basil on trial day 15; and RM5 with tomato on trial day 15 and red lettuce on trial day 29. This initial arrangement created a staggered crop calendar in which modules reached harvestable stages at different times.

Following the first harvest cycle, RMs were replanted or maintained according to crop performance, available planting positions, and the goal of extending fresh biomass availability over time. Later plantings included repeated leafy greens, and continued tomato and basil maintenance in selected modules. Because crop species differed in growth rate, canopy architecture, and harvest response, the staggered calendar also allowed evaluation of crop compatibility and planting-position effects within each RM. The complete crop calendar is summarized by RM and trial day in Figure 2.

2.4. Selective Harvest and Crop Maintenance

Harvesting was conducted as a repeated selective-harvest process rather than as a single terminal harvest. This approach was used to simulate a pick-and-eat production system, in which edible biomass was removed periodically while the production system remained active. For each harvest event, the harvest date, crop species, RM, days after sowing (DAS), harvest type, and fresh mass were recorded.

Mizuna harvests began at 21 DAS, whereas lettuce harvests began at 23 DAS. Harvest timing varied among RMs depending on crop development, canopy size, and the need to maintain continuous fresh biomass availability. Leaves or shoots were removed selectively when possible, while senescent tissue, damaged leaves, and residual debris were recorded separately.

Tomato fruits were harvested individually as they reached maturity. Red tomato harvest began on trial day 71 and continued until the end of the trial. Tomato RMs were inspected weekly, and fallen flowers, senescent tissue, and other debris were removed. Tomato plants were pruned as needed to maintain canopy structure, and limit interference with neighboring crops or planting positions. Removed non-edible tomato biomass was recorded separately from edible fruit biomass.

Basil was also managed as a continuous-harvest crop. Basil harvest began on trial day 42 and continued periodically through trial day 106. Basil RMs were cleaned weekly, and fallen or senescent material was collected, including small debris removed from the RM surface using tape. Debris biomass was recorded separately from edible harvested biomass.

Routine crop maintenance was performed throughout the trial. Cleaning was conducted after planting and after harvest events to remove loose plant material, fallen flowers, senescent leaves, and residual debris. Seedlings were not thinned; when two seedlings emerged from the same planting position, both were allowed to grow together to reduce individual stem thickness.

2.5. Irrigation and Water-Use Measurements

Root-zone θv was maintained between 0.55 and 0.60 m³ m⁻³, consistent with the operational range previously used for the URRM system [10]. The irrigation solution consisted of Utah Hydroponics nutrient solution prepared with reverse-osmosis water at 1X concentration [21]. Nutrient solution was delivered independently to each RM through the valve connected to the porous-tube irrigation system. Root-zone θv was measured in each RM using CS650 sensors and logged at 1-min intervals. These real-time θv measurements were used to estimate the volume of nutrient solution required to return each RM to the target θv, based on the known volume of each RM. Although automated syringe-pump irrigation was initially planned, the pump system failed during the trial; therefore, nutrient solution was applied manually using a syringe. The volume of nutrient solution added to each RM was recorded at every irrigation event.

To standardize daily water-use calculations, all RMs were irrigated each day at 22:00 h, immediately before the dark period, to bring root-zone θv back to approximately 0.60 m³ m⁻³. This 22:00 h irrigation event was used as the cutoff point between consecutive water-use days. Therefore, each daily water-use period began at 22:00 h with the dark period and ended at 22:00 h the following day, after completion of the light period. This procedure provided a consistent starting θv among RMs and made daily water-use estimates more comparable across days. Additional irrigation was applied during the light period when needed to maintain the target θv range.

Daily water use was calculated for each RM by summing the irrigation volumes applied between two consecutive 22:00 h cutoff irrigations. System-level daily water use was calculated as the sum of water added across the five RMs during the same period. Cumulative water use was calculated separately for each RM and for the full system by summing daily water additions over the trial period. Because the system was operated without drainage, recorded nutrient-solution additions were used as the operational estimate of RM water use. These daily and cumulative values were used to compare module-level water demand and to evaluate whether the staggered crop calendar distributed water-use peaks over time.

2.6. Bulk EC Monitoring

Bulk EC was monitored in each RM to evaluate temporal changes in root-zone ionic status during continuous crop production. CS650 sensors installed within the peat-based growth media logged bulk EC at 1-min intervals throughout the trial. To reduce the effect of θv variation on bulk EC, only EC measurements collected when root-zone θv was between 0.59 and 0.60 m³ m⁻³ were used for trend analysis.

Bulk EC dynamics were then evaluated in relation to crop developmental stage, planting and harvest events, crop composition, and daily water use. In the zero-discharge URRM system, bulk EC was interpreted as more than a salinity index. Because no drainage occurred, changes in bulk EC reflected the combined effects of nutrient addition, water uptake, nutrient uptake, and ion storage or accumulation in the root zone. Therefore, bulk EC was used as a diagnostic indicator of overall root-zone ionic status and was interpreted together with water-use patterns and crop performance.

2.7. Data Processing

Harvest, irrigation, and bulk EC data were summarized descriptively to evaluate the operational performance of the staggered crop-production strategy. Harvest records were used to calculate total fresh edible biomass, biomass by crop, biomass by RM, harvest frequency, and harvest timing. Water-use records were summarized as daily and cumulative water use at both the RM and system levels. Bulk EC data were processed as described above and summarized as daily mean values using measurements collected when root-zone θv was between 0.59 and 0.60 m³ m⁻³.

No statistical comparisons among RMs were performed because the RMs were not experimental replicates. Instead, they differed intentionally in planting date, crop composition, and crop arrangement as part of the staggered mini-garden design. Therefore, RM-level differences were interpreted descriptively as operational outcomes of the staggered crop calendar rather than as replicated treatment effects.

3. Results

3.1. Staggered Planting Extended Fresh Biomass Availability over Time

The staggered planting strategy converted the URRM simulator from a single-harvest crop-production system into a continuous harvest platform. Across the 120-day trial, repeated selective harvests produced a total of 4,299 g of fresh edible biomass. Harvest began on trial day 21 and continued through the end of the experiment. During this 100-day harvest period, edible biomass was collected on 77 days, indicating that fresh biomass was available on most days after harvests began rather than only at a single terminal harvest event. This corresponded to an average production of 43.0 g fresh edible biomass per day across the harvest period, or 55.8 g per harvest day when only days with harvested biomass were considered.

In practice, fresh edible biomass was available at least every other day for nearly the entire harvest period. Harvest gaps were generally short, with most non-harvest intervals lasting only one day. Harvest diversity increased as the trial progressed. Early harvests were dominated by mizuna and red lettuce, but after approximately trial day 40, daily harvests frequently included more than one crop (Figure 3). Representative harvest sequences for mizuna, red lettuce, and green lettuce illustrate how edible biomass was removed repeatedly over time and show the approximate size of individual pick-and-eat harvests at different days after sowing (Figure S3).

The large final tomato harvest reflected the end of the trial rather than the natural end of tomato production. Tomato plants were still actively growing and producing fruit, indicating that this crop could have continued contributing to the system if the trial had been extended. Therefore, harvest peaks in leafy greens indicate opportunities to refine harvest timing, whereas the final tomato peak reflects longer production potential beyond the experimental period.

Figure 4A provides visual reference images for representative edible fresh mass portions of each crop, illustrating the approximate size of individual pick-and-eat harvests. Cumulative biomass differed among crop categories (Figure 4B). Mizuna was the largest contributor to total fresh biomass, producing 1,205 g, followed by red lettuce (928 g), green lettuce (635 g), basil (486 g), tomato (454 g), and kale (397 g). Although total harvest also varied among RMs, those module-level differences are evaluated in the following section in relation to crop composition and water use.

3.2. Independent RMs Helped Distribute Water Demand

Daily water use varied among RMs during the 4-month trial, reflecting differences in planting date, crop species, canopy development, harvest timing, and crop configuration within each module (Figure 5). Because the five RMs were operated as independent planting units, water-use peaks shifted among RMs as different crops reached periods of rapid canopy expansion. Total system water use remained below approximately 2 L d⁻¹ for most of the trial.

The highest RM-level water-use peaks occurred in different modules and crop cycles. During Cycle 1, RM3 reached the highest daily water-use peak at 1.20 L d⁻¹, followed by RM1 at 1.05 L d⁻¹ and RM2 at 0.90 L d⁻¹ (Table 1). During Cycle 2, RM2 reached the highest peak at 1.00 L d⁻¹, followed by RM3 at 0.90 L d⁻¹ and RM4 at 0.70 L d⁻¹. RM5 had the lowest water-use peaks in both cycles, reaching 0.40 L d⁻¹ in Cycle 1 and 0.45 L d⁻¹ in Cycle 2. RM1 also showed a strong reduction between cycles, decreasing from 1.05 L d⁻¹ in Cycle 1 to 0.15 L d⁻¹ in Cycle 2. These RM-level peaks occurred while leafy greens were harvested selectively throughout the trial. Repeated harvest limited canopy accumulation and helped maintain water-use peaks below the levels expected for unharvested terminal crops.

Water-use responses also differed according to crop placement within each RM. At the beginning of the trial, all RMs contained tomato or basil plants in the back positions, reducing the planting area available for leafy greens. During Cycle 2, basil in RM4 and tomato in RM3 were removed, increasing the available surface area for leafy-green production. After this adjustment, RM2, RM3, and RM4 reached similar water-use peaks during subsequent crop cycles. In contrast, RMs that retained tomato plants in the back positions showed lower water-use responses after tomato establishment. In RM1, mizuna and tomato were planted on the same day during Cycle 1, and the module reached a high water-use peak. However, during Cycle 2, red lettuce was planted in the front positions while established tomato plants remained in the back. Red lettuce showed stunted growth, and RM1 reached only 0.15 L d⁻¹. In RM5, tomato was planted first and leafy greens were added 14 days later; this module showed low water-use peaks in both cycles. These responses indicate that water use was highest when leafy greens established vigorously, whereas lower peaks occurred when leafy-green growth was limited by crop configuration, previous crop effects, or competition with established tomato plants.

3.3. Crop Configuration and Species Selection Affected Productivity Under Repeated Use

Cumulative edible biomass and cumulative water use varied substantially among RMs during the 4-month trial (Figure 6). Across modules, cumulative water use ranged from 20.9 L in RM5 to 40.3 L in RM2, while cumulative edible biomass ranged from 545 g in RM5 to 1,409 g in RM2. In general, modules with greater cumulative water use also produced greater edible biomass, although this relationship differed among crop configurations. RM4 produced high edible biomass, with 998 g, despite using less water than RM2 and RM3. In contrast, RM1 and RM5, which retained tomato plants in the back positions, were among the lowest-producing modules. RM3 also produced relatively low fresh edible biomass, with 652 g, despite having the second-highest cumulative water use and being located in the highest-PPFD zone.

RM2, with the greatest total biomass, included a large contribution from red lettuce during the first cycle and basil that was continuously harvested. RM4 produced high biomass mainly from green lettuce and kale, with additional biomass from mixed greens and Swiss chard. These two modules also had the lowest water use per unit fresh edible biomass (Table 2), indicating that their higher cumulative water use was associated with greater edible biomass production. In contrast, RM1 and RM5 included larger tomato contributions but lower leafy-green biomass than RM2 and RM4, resulting in higher water use per unit fresh edible biomass. RM3 had the lowest fresh-biomass production per liter of water among the RMs, despite being located in the highest-PPFD zone. RM3 was also the only module planted with mizuna for three consecutive cycles without crop rotation, and tomato plants were present in the back positions during the first cycle before being removed at approximately 30 DAS. Across the trial, the RMs that included tomato plants at any stage were the lowest-producing modules.

RM1, RM3, and RM5 initially included tomato plants in the back planting positions; however, tomato plants in RM3 were removed at approximately 30 DAS and did not contribute fruit harvest. In RM1, tomato and mizuna were planted on the same day during the first cycle, and mizuna established successfully. In later cycles, new leafy greens were planted in front of established tomato plants. Under this configuration, red lettuce showed stunted growth, Swiss chard also performed poorly, and kale showed the greatest ability to establish, although its growth and water-use response were lower than those observed during the first-cycle mizuna crop. In RM5, tomato was planted first and red lettuce was added 14 days later. Leafy-green growth was limited from the first cycle, and RM5 remained one of the lowest-yielding and lowest-water-use modules.

Species performance changed across crop cycles. Lettuce contributed substantially to biomass production in RM2 and RM4 during Cycle 1, but red and green lettuce were less reliable in later cycles, including in modules without tomato plants in the back positions. Swiss chard was introduced in Cycle 3 and also showed limited performance under repeated-cycle conditions. In contrast, Brassica crops showed more consistent performance across repeated cycles.

Because the nutrient solution used during the trial was later found to contain no sulfur, cycle-level productivity was interpreted with caution. The effect of this nutrient omission was expected to become more evident during later cycles, after repeated cropping and continued nutrient removal from the RMs. Therefore, comparisons of crop productivity among configurations were focused primarily on the first two crop cycles, when nutrient limitation was expected to be less restrictive (Table 3).

Total fresh edible biomass decreased from 2,252.94 g in Cycle 1 to 738.70 g in Cycle 2, corresponding to a 67.2% reduction. The magnitude of this reduction differed among RMs and crop configurations In Cycle 2, tomato plants were already well established in the back positions of RM1 and RM5, and basil remained established in the back positions of RM2. RM1 and RM2 showed large reductions in biomass from Cycle 1 to Cycle 2, with decreases of 98.3% and 74.5%, respectively. In RM1, the reduction corresponded to severe stunting of red lettuce during Cycle 2. RM5 decreased by 48.8% between cycles; however, its Cycle 1 biomass was already low, producing only 172.31 g of red lettuce compared with 725.96 g in RM2, which was also planted with red lettuce during Cycle 1. In RM5, tomato had been planted 14 days before red lettuce, and leafy-green growth was limited from the first cycle. In contrast, RM3 and RM4, where tomato or basil plants were removed after Cycle 1, had smaller reductions, decreasing by 42.4% and 62.7%, respectively.

3.4. Water Use and Bulk EC Showed Module-Specific Nutrient Dynamics

Daily water use, cumulative water use, and bulk EC followed distinct trajectories among RMs during the 4-month staggered multi-crop trial (Figure 7). At the beginning of the trial, most RMs had bulk EC values between approximately 0.35 and 0.40 dS m⁻¹. The second crop cycle began at lower bulk EC than the first cycle in most RMs, with starting values approximately one-third to one-half of those measured at the beginning of the trial. In RM1, RM2, RM3, and RM4, periods of rapid crop growth were associated with large daily water-use peaks and concurrent decreases in bulk EC.

The steepest EC drawdowns occurred during mizuna-dominated periods. During Cycle 1, RM1 and RM3, both planted with mizuna, showed the greatest decreases in bulk EC, reaching 0.13 and 0.12 dS m⁻¹, respectively, by the end of the cycle. In comparison, the lettuce-dominated modules maintained higher bulk EC values at the end of Cycle 1, with RM2 at 0.15 dS m⁻¹ and RM4 at 0.17 dS m⁻¹. These patterns indicate that bulk EC declined more rapidly during mizuna growth than during lettuce-dominated growth. In the following cycles, RM2, RM3, and RM4 continued to show EC decreases during major water-use peaks, followed by partial EC recovery during lower-growth or post-harvest periods.

RM1 showed a different pattern after the first cycle. Following the large early water-use peak associated with first-cycle mizuna, daily water use remained low and relatively constant during Cycle 2. During this period, bulk EC showed little additional decline and no clear recovery, coinciding with the restricted growth observed for red lettuce in Cycle 2. RM5 also differed from the other modules. Daily water use remained comparatively low throughout the trial. Despite the lower water use, bulk EC declined steadily from the beginning of the trial and showed limited recovery during later cycles.

A late-cycle increase in bulk EC was evident in most RMs, particularly during the final weeks of the trial. This increase occurred after the main leafy-green production peaks, when leafy-green yield was lower and remaining harvests were more strongly associated with tomato and basil production. These patterns show that water use and bulk EC changed together, but the direction and magnitude of the response differed among modules, crop species, and crop configurations. Detailed RM-specific dynamics of daily water use and bulk EC in relation to planting and harvest events are shown in Figure S4–S8.

3.5. Operational Observations and Study Constraints During Repeated Planting

Several operational observations emerged during the 4-month staggered multi-crop trial. Staggered planting across independent RMs distributed both harvest events and water demand over time, with harvests occurring on most days during the harvest period and water-use peaks occurring at different times among modules. However, module productivity also depended on crop placement. Leafy greens planted in front of established tomato plants showed reduced growth. This effect was less evident in RM2, where basil remained in the back positions and leafy-green production was higher than in the tomato-containing modules. Among leafy greens, lettuce produced high biomass during the first cycle but was less reliable in later cycles. Mizuna and kale showed stronger establishment during repeated use.

Repeated use of the same peat-based growth media introduced physical constraints that became evident after harvest. Although new planting positions were available during the first four cycles, harvested stumps remained around or below the silicone rubber top cover. Stump size differed among crop species. Mizuna produced larger stumps than red or green lettuce, creating larger retained residues near the growth media and top-cover interface during repeated planting (Figure S9). Remaining stumps were covered to reduce direct contact between old residues and new crop growth and to maintain containment. However, some regrowth occurred beneath the top cover. Visual observations also showed localized mycelial growth on or near old stumps and planting positions, as well as algae growth on the surface of the growth media during later crop cycles. Mycelial growth was also observed around emerging seedlings near the planting wedges, particularly where the pullulan material had dissolved. These observations show that repeated planting in the same RM created crop-residue, containment, and sanitation challenges that differed among species and planting materials.

The trial also had several constraints that affected interpretation of later-cycle productivity. The nutrient solution used during the trial was later found to contain no sulfur, and low sulfur and iron were observed in preliminary plant and/or peat analyses near the end of the trial. In addition, initial bulk EC differed among crop cycles, with lower starting values after the first cycle. Therefore, the lower productivity observed during later cycles cannot be attributed to a single factor, but likely reflected the combined effects of crop sequence, crop placement, repeated media use, lower nutrient availability, and differences in initial root-zone EC.

The nutrient solution used in this trial was a hydroponic formulation, and no salt accumulation was observed during repeated crop cycles under zero-discharge operation. Instead, bulk EC decreased during active crop growth and remained low during subsequent cycles. Bulk EC generally remained above approximately 0.10 dS m⁻¹; however, this value reflects total ionic strength and does not confirm that all essential nutrients were present in adequate concentrations.

The study was conducted under ground-based conditions rather than in microgravity. Plant orientation, water distribution, root-zone aeration, residue behavior, and microbial conditions may differ under flight conditions. Thus, the results describe operational responses observed in a ground-based URRM mini-garden simulation.

4. Discussion

4.1. Staggered Planting Supported Continuous Harvest in a Reusable Mini-Garden

The staggered planting and selective-harvest strategy extended fresh biomass availability over most of the harvest period, converting the URRM simulator from a single-harvest production system into a repeated-harvest mini-garden. The system produced 4,299 g of fresh edible biomass, with harvests collected on 77 of the 100 days after harvest began. This pattern is important for near-term ISS crop-production systems, where the goal is not complete dietary replacement, but rather repeated access to fresh, high-value produce that can supplement the packaged diet and contribute to crew well-being [10,22]. In this context, harvest frequency is as important as total biomass, because small, repeated harvests are better aligned with a pick-and-eat mini-garden concept than a few large terminal harvests.

Early harvests were dominated by mizuna and lettuce, whereas later harvests included tomato, basil, kale, and additional leafy greens. This transition from single-crop to multi-crop harvest days increased crop diversity over time and demonstrated that independent RMs can support a dynamic crop calendar within a small total planting area. This is consistent with the operational need for supplemental space crop-production systems to provide crop diversity and repeated fresh-food availability within constrained volume and crew-time resources [5,6,22].

Selective harvest also influenced the temporal pattern of biomass availability. Several large harvest peaks occurred during the trial, especially during the final lettuce harvests and at the end of the tomato crop. However, these peaks should be interpreted partly as harvest-management outcomes rather than fixed biological limits of the system. In leafy greens, large single harvests indicate that harvest timing could be shifted earlier or divided into smaller harvests across more days. For example, instead of allowing lettuce biomass to accumulate until a large final harvest, smaller pick-and-eat harvests could be taken more frequently to improve harvest continuity and reduce the risk of quality decline. In contrast, the final tomato peak reflected the termination of the experiment, because tomato plants were still actively growing and producing fruit at the end of the trial. Therefore, the leafy-green peaks indicate opportunities to refine selective-harvest timing, whereas the tomato peak indicates that fruit production could have continued beyond the 120-day trial.

Overall, the results support staggered planting as a practical strategy for repeated fresh-biomass availability in a reusable mini-garden. However, the harvest calendar must be managed actively. Repeated production depended not only on planting dates, but also on selective-harvest timing, crop regrowth capacity, crop duration, and whether different species could be maintained simultaneously without suppressing one another. Thus, staggered planting provided the temporal framework for distributed harvest, but crop configuration and species selection determined how effectively that framework translated into usable fresh biomass.

4.2. Independent RMs Distributed Water Demand over Time

Daily water use varied among RMs because crops differed in planting date, species, canopy development, and harvest timing. Because the five RMs were operated as independent planting units, peak water demand did not occur simultaneously across all modules. Instead, water-use peaks shifted from one RM to another as different crops reached periods of rapid growth. This temporal separation of peak demand was a key operational advantage of the staggered crop calendar.

This result is important for spaceflight crop-production systems because water delivery, humidity control, condensate recovery, and overall water balance have finite capacities [5,6,14,19]. In previous URRM simulator and EDU trials conducted without staggered planting and without repeated selective harvest, daily water use reached approximately 5 L d⁻¹ during periods of full canopy development [13,14,19]. Under those conditions, high transpiration rates can increase the risk of relative humidity approaching saturation within the growth chamber, particularly below dense canopies where air movement is restricted. In contrast, the staggered crop calendar used in the present study maintained total system water use below approximately 2 L d⁻¹ for most of the trial. This indicates that staggered planting can reduce synchronized system-level transpiration peaks and help avoid short periods of excessive atmospheric water load [14].

The independent RM architecture contributed to this effect because each module could be maintained at a different crop stage. When one RM had a large canopy and high transpiration demand, neighboring modules often had newly planted crops, recently harvested crops, or lower canopy cover. This spatial separation likely maintained a more open canopy structure across the full growing area, which can improve air movement around and below the canopy.

This is especially relevant because turning off the lights does not eliminate water loss in this system. Previous URRM measurements showed that nighttime evapotranspiration (ET) remained substantial, with nighttime ET approximately 0.46 of daytime ET [14,19]. Therefore, if the system reaches an excessive whole-canopy water-use load, the humidity burden may persist during the dark period. Reducing photoperiod or light intensity could lower transpiration demand, but this would also reduce daily light integral and potentially slow crop growth. Staggered planting provides a more operationally favorable strategy because it moderates whole-system water demand without requiring major changes to the lighting schedule.

Selective harvest also contributed to water-demand management. By repeatedly removing edible leaf biomass, the canopy was prevented from accumulating to the same extent as in an unharvested terminal crop. Therefore, the combination of independent RMs, staggered planting, and selective harvest distributed transpiration demand over time while maintaining repeated fresh biomass production helped maintain repeated fresh biomass production while avoiding synchronized whole-system water demand.

4.3. Crop Configuration and Repeated Planting Affected Productivity

Crop configuration strongly affected whether the available RM volume was effectively converted into edible biomass. A key finding of this study was that maintaining root-zone θv near 0.60 m³ m⁻³ and supplying nutrient solution throughout the trial did not eliminate competition among crops sharing the same RM. At first, competition within the RMs might have been expected to be limited because θv was maintained at a high level and nutrient solution was supplied regularly. Under this assumption, tomato plants could be placed in the back positions while leafy greens occupied the front positions, allowing multiple crops to share the same RM. However, the results showed the opposite pattern. Leafy greens planted in front of established tomato plants showed poor establishment, low water use, and reduced biomass production. Therefore, the main limitation was not simply the amount of water or nutrient solution supplied, but the interaction among crop architecture, crop age, canopy position, rooting volume, nutrient capture, and repeated use of the same media.

Intercropping can improve resource-use efficiency, biodiversity, and overall productivity when compatible species are combined, but these benefits depend on reducing excessive competition for light, water, nutrients, and rooting space [23]. Belowground interactions are especially important in this context because root systems compete for limited resources, and plant performance can be restricted when neighboring root systems overlap within confined rooting volumes [23,24]. These interactions are likely to be amplified in spaceflight crop-production systems, where root-zone volume is severely constrained and long-duration crops may occupy the substrate while subsequent crops are established. In the present study, the negative response of leafy greens planted after tomato establishment suggests that shared RM volume created a belowground competition problem that could not be resolved simply by maintaining high θv or regular nutrient-solution supply.

These results indicate that mixing long-duration crops, especially tomato, with short-cycle leafy greens in the same RM is not an appropriate crop configuration for this system. Although the original crop layout placed tomato plants in the back positions and leafy greens in the front positions, this arrangement reduced the performance of later leafy-green cycles. In contrast, basil appeared to be more compatible with leafy greens, likely because its canopy architecture, harvest pattern, and growth habit created less sustained competition than tomato. Thus, separating long-duration fruiting crops and short-cycle leafy greens among different RMs appears to be a better strategy than mixed permanent–short-cycle cropping within the same RM.

The response of RM1 and RM5 illustrates the importance of crop age and establishment sequence. In RM1, tomato and mizuna were planted on the same day, and mizuna competed well during the first cycle. However, in later cycles, new leafy greens were planted in front of already established tomato plants, and crop performance declined. Red lettuce failed, Swiss chard also performed poorly, and kale showed the best ability to grow under this configuration, although its growth and water-use response remained lower than first-cycle mizuna. In RM5, tomato was planted first and red lettuce was added 14 days later; in this case, leafy-green growth was limited from the first cycle, and RM5 remained one of the lowest-yielding and lowest-water-use modules. These patterns indicate that leafy greens were more successful when established at the same time as tomato than when introduced after tomato had already occupied the root zone. Therefore, the main limitation of this configuration appeared to be belowground competition within the shared RM volume rather than canopy interference among modules.

The lower productivity of tomato-containing modules was likely driven by more than one mechanism, but aboveground light competition was probably not the dominant factor. The tomato cultivar used in this study had a compact dwarf canopy, and visual observations showed that the tomato canopy did not substantially shade newly planted leafy greens or invade neighboring RM areas (Figure S10). Instead, the main limitation appeared to occur within the shared root-zone volume. Established tomato plants may have occupied a large portion of the available rooting volume and competed with newly planted leafy greens for nutrients and localized water uptake zones, even though the overall RM θv was maintained near the target level. The CS650-based irrigation strategy maintained average RM θv, but it did not prevent localized belowground competition within the shared media volume. In addition, tomato-mediated rhizosphere effects cannot be ruled out. Tomato roots can release α-tomatine into the rhizosphere, and this compound has been shown to influence soil bacterial communities in a concentration-dependent manner [25]. Therefore, established tomato plants may have altered the local root-zone environment through both resource competition and rhizosphere-mediated interactions. Additional experiments would be required to separate these mechanisms.

Repeated planting in the same peat-based media also likely contributed to productivity differences among RMs. RM3 did not retain tomato after the first cycle, but it was planted with mizuna across three cycles and produced less fresh edible biomass than expected given its relatively high cumulative water use and high-PPFD position. This response should not be interpreted only as low water-use efficiency, because mizuna has a higher dry matter fraction than lettuce and therefore produces less fresh mass per unit dry matter. However, repeated mizuna planting without crop rotation may also have negative carryover effects in the reused media. Plant–soil legacy effects can persist across crop cycles through residual roots, rhizodeposits, microbial shifts, and the buildup of crop-specific biological constraints[16,17,18,25]. Such legacy effects can influence nutrient cycling, microbial community composition, pathogen pressure, and ultimately the establishment and growth of subsequent crops. Crop rotation can reduce negative plant–soil legacy effects and limit the buildup of soil-borne pathogens or crop-specific microbial constraints [17,26]. Thus, the lower productivity of RM3 probably reflected a combination of repeated-cropping constraints, crop-establishment failure, crop-specific legacy effects, and crop physiology rather than tomato competition alone.

Species performance also changed across cycles, showing that crop selection for a reusable mini-garden should be based on repeated-cycle performance rather than first-cycle yield alone. Lettuce performed well during the first cycle, but red and green lettuce were less reliable in later cycles, even in RMs without tomato plants in the back positions. However, observations from a separate URRM EDU study indicate that lettuce performance after repeated RM use may improve when a recovery period is included between crop cycles. In that study, red lettuce was planted after approximately six months of continuous RM use that included tomato production, but the RMs were left fallow for approximately two months before replanting. Red lettuce did not perform as well as in a first-cycle planting, but little to no stunting was observed in most RMs [11]. This contrasts with the severe stunting observed in the present trial when lettuce was replanted without a fallow period. Although the EDU study was not designed to isolate the mechanism responsible for improved lettuce establishment, the comparison indicates that a fallow period may allow partial recovery of the root-zone environment before replanting a sensitive crop such as lettuce. Swiss chard also showed limited performance under repeated-cycle conditions, including in configurations without tomato competition. In contrast, Brassica crops performed better across repeated cycles. These responses indicate that species suitability for reusable RMs depends not only on initial productivity, but also on establishment reliability after previous crops have been harvested and on whether sufficient recovery time is provided between cycles.

Repeated use of the same peat-based media introduced physical and sanitation-related constraints that were not apparent in single-cycle evaluations. Although new planting positions were available for the first four cycles, harvested stumps remained around or below the silicone rubber top cover. In the original harvest approach, stumps and roots were intentionally retained in the peat-based growth media to degrade while maintaining containment. These constraints were already evident after the first cycle and differed among crop species (Figure S9). Mizuna produced larger stumps than red or green lettuce, leaving larger residues near the top-cover interface.

To reduce direct contact between old stumps and new crop growth, remaining stumps were covered. However, covering the stumps also allowed some continued growth beneath the top cover, where conditions were likely wetter, darker, and less aerated than above the cover. These conditions may have favored localized decomposition and microbial growth near new planting positions. In addition, the pullulan-based seed wedges may have provided an additional carbon source for microbial activity during seedling establishment. This interpretation was consistent with visual observations of mycelial growth on or near old stumps and planting positions, particularly where the pullulan material had dissolved.

Overall, the results show that crop configuration, crop sequence, and repeated media use must be considered together when designing reusable RMs. A suitable strategy would be to separate crop types among RMs, with long-duration crops grown independently from high-turnover leafy greens. For leafy-green RMs, crop rotation and residue management will likely be necessary to reduce negative legacy effects across cycles. Crop selection should consider not only edible yield, but also stump size, ease of residue management, regrowth potential, sanitation risk, and ability to establish after previous crops. One possible strategy would be to use lettuce in the first cycle, when clean planting positions support vigorous establishment and stump residues are easier to manage, followed by more reliable crops in reused media, such as mizuna or kale, in subsequent cycles. Another option would be to harvest vigorous first-cycle crops earlier, before large stumps develop.

4.4. Bulk EC Showed Depletion Rather than Salt Accumulation Under Zero Discharge

One concern with zero-discharge crop production is that repeated irrigation without leaching could lead to progressive salt accumulation in the root zone. This is especially important for reusable growth media, where salts, root residues, and crop-derived compounds may accumulate across successive crop cycles. In the present trial, however, bulk EC did not increase progressively during repeated RM use. Instead, bulk EC declined during periods of active crop growth and remained low during later crop cycles. This indicates that, under the nutrient concentration and irrigation strategy used in this study, the zero-discharge system did not produce a salt-buildup problem. Rather, the dominant root-zone response was depletion of soluble ions from the peat-based media.

This finding is important because it shows that the absence of leaching does not necessarily result in salinity accumulation when nutrient inputs are closely matched to plant water use and uptake. In conventional container or hydroponic production, leaching or solution replacement is often used to prevent undesirable salt accumulation. In contrast, the URRM system was operated without drainage, so unused ions could only remain in the root zone, be taken up by plants, become immobilized, precipitate, or interact with the media. The observed decline in bulk EC therefore indicates that plant uptake and other removal or immobilization processes exceeded ion accumulation during active growth. From an operational perspective, this is encouraging because it suggests that reusable zero-discharge RMs can avoid salinity buildup, at least under the tested conditions. However, it also shows that nutrient depletion can become a greater limitation than salt accumulation.

The EC response also helps explain why later-cycle productivity declined. The second crop cycle began with lower bulk EC than the first cycle in most RMs, and periods of rapid growth were associated with further EC drawdown. This means that later crops were often established in media with lower initial soluble-ion availability than earlier crops. In a reusable RM, this can affect seedling establishment and early vegetative growth, especially for crops that are sensitive to nutrient availability during early development. Therefore, reduced productivity in later cycles should not be interpreted only as an effect of crop competition or plant legacy effects. It also likely reflected a changing root-zone nutrient environment, where repeated cropping reduced the pool of soluble nutrients available to newly planted crops.

However, bulk EC has an important limitation: it measures total ionic strength, not individual nutrient sufficiency. A low bulk EC indicates that the concentration of soluble ions is low, but it does not identify which nutrients are depleted. Likewise, a bulk EC value that remains above a minimum threshold does not confirm that all essential nutrients are present in adequate amounts. For example, the nutrient solution used in this trial was later found to contain no sulfur, and preliminary plant and/or peat analyses near the end of the study indicated low sulfur and iron. These nutrient-specific limitations would not be detected from bulk EC alone. A root zone could maintain measurable EC because other ions are present while still being deficient in a specific essential element such as sulfur or iron.

This distinction is especially important for interpreting the late-cycle decline in lettuce and chard performance. Bulk EC trends showed that the root-zone solution was becoming depleted, but they could not determine whether poor establishment was associated with nitrate depletion, sulfur deficiency, iron limitation, crop-specific biological effects, or a combination of these factors. Sulfur omission is particularly relevant because sulfur is required for amino acid and protein synthesis and may become more limiting as plant biomass accumulates over repeated cropping. Iron limitation may also affect chlorophyll formation and photosynthetic function, especially in rapidly growing leafy greens. Therefore, the observed EC decline should be interpreted as evidence of overall nutrient removal rather than as evidence of depletion of any specific ion. The sulfur and iron results demonstrate the importance of ground-based ion-specific analysis for interpreting bulk EC responses and identifying the actual nutritional constraints before flight.

Overall, these results show that managing a zero-discharge mini-garden is not only about reducing water use. A module with low water use can still move toward nutrient depletion, whereas modules with vigorous leafy-green growth may show EC drawdown during rapid growth followed by partial recovery after harvest. Therefore, water use and bulk EC should be evaluated together as a diagnostic approach for interpreting root-zone status, crop scheduling, and repeated use of peat-based growth media in reusable RMs. However, bulk EC represents the combined ionic status of the root zone and does not identify which nutrients are accumulating or being depleted. This limitation is important in closed-loop soilless systems because nutrient ratios can shift over time even when total EC appears stable [27]. For ISS operation, bulk EC should therefore be interpreted as a flight-compatible indicator of overall root-zone ionic status rather than as a direct measure of specific nutrient availability. Because routine ion-specific analysis is not available during flight, the value of bulk EC depends on ground-based calibration. For the URRM system, this includes developing crop-specific relationships among bulk EC, pore-water EC, and nitrate-N depletion for crops intended for flight. These relationships can help identify crops that rapidly deplete nitrate-N and support adjustment of nutrient-solution concentrations before launch. In this study, the main EC-related risk was nutrient depletion rather than salinity buildup, highlighting the need to adapt nutrient formulations and replenishment strategies for repeated cropping in zero-discharge media.

4.5. Operational Constraints and Lessons for Future Crop Calendars

The 4-month trial showed that repeated use of the same RM is not limited only by water delivery, EC control, or crop productivity. Several operational constraints emerged during harvest, replanting, residue management, and seedling establishment. These constraints are especially important for the URRM because growth-media containment and food safety must be maintained during repeated planting and harvest events. Therefore, future crop calendars must integrate biological performance with operational handling, including harvest method, residue management, planting-material design, media condition, and crew procedures.

One important lesson was that the original harvest approach, which left stumps and roots in the peat-based growth media to degrade, may not be optimal for repeated crop cycles. Although this approach minimized disturbance and supported media containment, retained stump residues became localized sites of rapid microbial growth. Covering the stumps reduced direct contact with new crops, but likely created humid, poorly aerated microsites around decomposing tissue. Based on these observations, new harvest protocols are being developed to remove more stump material at harvest while preserving growth-media containment.

Planting materials and media condition also affected seedling establishment. The pullulan-based wedges were useful for seed placement and containment, but mycelial growth was observed near emerging seedlings, especially where the pullulan material had dissolved. This suggests that the wedges may have provided readily available carbon for microorganisms during germination and early seedling establishment. In response, wedge size has been reduced from approximately 1 cm on each side to about half that size to maintain seed-placement function while reducing the amount of available carbon introduced into the RM.

The physical condition of the peat-based media changed with repeated crop cycles. After approximately two cycles, peat settling reduced contact between the planting wedges and the growth medium in some positions. This can limit water transfer to the wedge and delay seedling emergence. In some cases, seedlings germinated below the silicone rubber fringes and began opening true leaves before fully emerging above the cover. Therefore, future crew procedures should include checking and, when necessary, gently adjusting the peat surface before planting to ensure good wedge–media contact, as well as monitoring seedling emergence past the fringes during early establishment. In response to these observations, peat packing density is being increased from 0.12 to 0.14 g cm⁻³ to reduce settling and improve contact between the wedge and the growth medium.

Moisture conditions near the top cover also require further optimization. Although previous studies indicated that θv from 0.45 to 0.70 m³ m⁻³ supported similar plant growth in the URRM simulator [28], the present system introduced an impermeable top cover that changed the near-surface microenvironment. At the target θv of approximately 0.60 m³ m⁻³, high moisture near the cover likely contributed to algae development and mycelial growth around wedges and decomposing residues. Therefore, current tests are evaluating a lower target θv of approximately 0.50 m³ m⁻³ to reduce excessive surface moisture while maintaining germination, seedling emergence, crop productivity, and growth-media containment.

The trial also showed that mixed cropping should be avoided when the goal is production rather than competition testing. Long-duration tomato and short-cycle leafy greens were not compatible within the same RM because competition occurred in both directions: established tomato reduced later leafy-green performance, while vigorous mizuna appeared to reduce tomato growth when both crops were established together.

These observations indicate that tomato yield potential was likely underestimated in the mixed-crop configuration. A tomato-only RM would provide a better estimate of tomato yield potential, crop duration, water use, and nutrient demand under zero-discharge conditions. Similarly, leafy-green productivity is best evaluated in RMs without long-duration fruiting crops occupying the back positions. For production, tomato and leafy greens should therefore be grown in separate RMs, allowing each crop type to be managed according to its own growth pattern, harvest frequency, residue characteristics, and nutrient demand.

Overall, the main operational lesson from this trial is that reusable RMs require crop calendars designed around both biological performance and operational handling. Leafy-green calendars should consider not only fresh biomass yield, but also stump size, ease of harvest, residue decomposition, microbial risk, compatibility with media containment, and reliability of establishment after previous crops. Mizuna and kale appeared more reliable than lettuce and chard during repeated use, but crop selection should balance productivity with operational cleanliness and ease of repeated management. These lessons support the use of independent RMs to separate crop types, reduce unwanted crop competition, and manage crop-specific nutrient demand. They also show that harvest protocols, residue handling, planting materials, water-content targets, and media packing are as important as irrigation and EC control for successful repeated cropping in a zero-discharge ISS mini-garden.

5. Conclusions

The 4-month URRM trial demonstrated that a staggered, selective-harvest crop calendar can support repeated fresh biomass production in zero-discharge reusable root modules, producing 4,299 g of edible fresh biomass with harvests available on most days after the first harvest. Independent RMs helped distribute crop development and water demand over time, keeping total system water use below approximately 2 L d⁻¹ for most of the trial and avoiding synchronized peak demand across modules. However, productivity and operational reliability depended strongly on crop configuration, species selection, and repeated-use constraints. Mixed cropping of long-duration tomato with short-cycle leafy greens was not optimal because competition occurred in both directions, indicating that tomato and leafy greens should be managed in separate RMs to better match their growth patterns, harvest frequency, residue characteristics, and nutrient demand. Bulk EC monitoring provided useful information on overall root-zone ionic status, but water use alone was not sufficient to diagnose nutrient balance, and the main EC-related risk was nutrient depletion rather than salinity accumulation. Overall, an ISS mini-garden should be designed as an operational crop calendar in which species selection, planting position, harvest timing, water demand, residue management, media condition, and nutrient balance are coordinated across independent RMs to maintain productivity, containment, and food safety during repeated crop cycles.