The experiment was carried out at Florida Gulf Coast University (FGCU), located in the city of Fort Myers, Florida, United States of America. The experimental setup was in the Aquarium Room 114 of Academic Building 9 at the Water School. The Florida Gulf Coast University Work Management Center maintained the Aquarium Room 114 between 20.5˚C and 22.2˚C for the experiment duration.

The plug production experiment was conducted in a controlled environment, including a thermally insulated grow tent, LED lighting source, and environmental controls that monitored humidity, and temperature. The components of this controlled environment allowed us to maintain stable growing conditions for plug production with reduced interference from external influences. The experimental setup was similar to the arrangements used by the Yang Laboratory at the University of Connecticut [33] - [37] .

The thermally insulated grow tent (The Original Gorilla Grow Tent^® 5 × 5, Gorilla Inc., Santa Rosa, California) had dimensions of 1.52 × 1.52 × 2.12 m and weighed 33.9 kg. The artificial lighting element consisted of six light fixtures (FREELICHT 4 ft LED Grow Light 60W, Amazon Inc., Seattle, Washington) installed in the tent to provide sufficient light energy for photosynthesis. The lights were attached from the tent ceiling by support ratchets (Heavy-Duty StainlessSteel Gear Ratchets, AC Infinity, Los Angeles, California). The six linked lights collectively produce 18,000 lumens of light and 21,000 K color temperature. The lights were positioned 40 cm above the top of the plug trays for the experiment duration.

We utilized a ventilation system using a high-powered fan (CLOUDLINE T6, 6” Inline Duct Fan with Temperature Humidity Controller, AC Infinity, Los Angeles, California) with a duct opening size that is 0.15 m in diameter, an air-flow capacity of 11.38 cubic meters per minute (CMM), and a power rating of 38 watts. The fan was placed in the top opening of the grow tent in an orientation that directed air outwards. The bottom vent was left open to provide an inflow to the recirculating system of air.

We used light sensors (FUTUREHORTI Light PAR Meter PPFD Tester, Amazon Inc., Seattle, Washington) to collect data on the photosynthetic photon flux density (PPFD) and daily light integral (DLI) in moles per meter square day (mol/m²·day). To collect air moisture content and temperature data, we utilized the built-in environmental sensor of the fan (CLOUDLINE T6, 6” Inline Duct Fan with Temperature Humidity Controller, AC Infinity, Los Angeles, California). Sensors were placed in the center of the grow tent at the same elevation as the plug trays to represent the environmental conditions to which the plugs were exposed. The sensors continuously recorded humidity, temperature, and the ppm of carbon dioxide throughout the growth cycle.

We placed the plugs on a tray stand (SKU number HGC706122, Fast Fit Ltd., Hawthorne Gardening Company, Vancouver, Washington) whose dimensions were 1.22 × 1.22 m. On the stand, we placed eight smaller trays (Living Whole Foods Seed Starter Grow Trays, Amazon Inc., Seattle, Washington) with the dimensions of 0.25 × 0.51 m, which housed the growth medium (Horticube XL 104-Cell Sheets, Oasis^® Grower Solutions, Kent, Ohio). The trays were sowed with “Rex Butterhead” Lettuce Lactuca sativa seeds (Johnny’s selected seeds, Fairfield, Maine) and were subjected to different fertigation treatments.

The digestate utilized in this experiment was received from a H-WEF system setup at the Emerging Technologies Institute (ETI) at Florida Gulf Coast University. The digestate was then diluted with RO water to create the treatments, the percentage of digestate ranged from 5% to 35% with each treatment having a difference of 5%. The range of digestate concentrations from 5% to 35% was chosen to assess the impact on plant growth. The 5% concentration serves as a baseline, while incremental increases allow for the evaluation of potential response effects, and we chose the higher limit of 35% based on the pH and electrical conductivity, not being too high and allowing for viable crop growth.

The synthetic fertilizer utilized for the control tray consisted of (Jack’s Nutrients 5-12-26 Part A FeED) and (Jack’s Nutrients 15-0-0 Calcium Nitrate Part B), both produced by JR Peters Inc., Allentown, Pennsylvania. The 5-12-26 Part A fertilizer was formulated from potassium nitrate, magnesium sulfate, monopotassium phosphate, iron DTPA, iron EDTA, iron EDDHA, copper EDTA, manganese EDTA, zinc ETDA, boric acid, and ammonium molybdate. Plant growth was supported by the specific available nutrients that were 5% total nitrogen (N), 12% available phosphate (P₂O₅), 26% soluble potash (K₂O), 6.3% available magnesium (Mg), and 8.5% available sulfur (S). Additionally, the Part A fertilizer contained micronutrients such as 0.05% boron (B), 0.015% copper (C), 0.3% iron (Fe), 0.05% manganese (Mn), 0.019% molybdenum (Mo), and 0.015% zinc (Zn). The 15-0-0 Part B fertilizer (Jack’s Nutrients 15-0-0 Calcium Nitrate Part B, JR Peters, Inc., Allentown, Pennsylvania), derived from calcium nitrate (Ca(NO₃)₂), provided 15% total nitrogen (N) and 18% calcium (Ca). Together, these components facilitate the development of robust roots and leaves, which are crucial for successful plug production in vegetative plants.

The organic fertilizers were derived from the digestate from the H-WEF system. We conducted tests for the Electrical Conductivity (EC), pH, Nitrates (NO₃), Potassium (K⁺), Calcium (Ca²⁺), Ammonium, and Phosphorus Pentoxide (P₂O₅) within a temperature range of 22.6˚C to 23.3˚C. This temperature control was implemented to minimize fluctuations in EC and its potential impact on other parameters. We average three readings for NO₃, K⁺, and Ca²⁺ values to ensure accuracy. A singular test was performed to find the EC, pH, ammonium, and P₂O₅ on every sample. Table 1 presents the EC, pH, Ammonium, and Phosphorus Pentoxide (P₂O₅) values as well as the average values for Nitrates (NO₃), Potassium (K⁺), and Calcium (Ca²⁺) for the different dilutions of the digestate-based fertilizers.

The EC values varied between 190 and 1100 ppm. Among the digestate-based fertilizers, the 35D–65RO treatment had the highest EC, while the 5D–95RO treatment had the lowest. The pH levels ranged from 6.79 to 7.21 across all seven digestate treatments. The 35D–65RO treatment recorded the highest pH, whereas the 5D–95RO treatment had the lowest. NO₃ (ppm) values ranged from 13.7 to 27.3 ppm, with the 35D–65RO treatment showing the highest and the 5D–95RO treatment showing the lowest values. The K⁺ concentrations varied between 11 and 84 ppm. The 35D–65RO treatment had the highest K⁺ concentration, while the 5D–95RO fertigation had the lowest. The Ca²⁺ concentrations ranged from 6.3 to 13.3 ppm. Both the 35D–65RO and 30D–70RO fertigation treatments recorded the highest Ca²⁺ concentrations, each at 13.3 ppm, while the 5D–95RO treatment had the lowest at 6.3 ppm. Ammonium levels ranged from 45 to 232 ppm, with the 20D–80RO treatment showing the highest and the 5D–95RO treatment showing the lowest levels. The P₂O₅ values stretched from 45 to 68 ppm, with the 20D–80RO fertigation recording the highest value and the 5D–95RO fertigation recording the lowest. For the NO₃, the 20D–80RO fertigation had the second-highest concentration of nitrates. The 20D–80RO fertigation had a greater concentration of K⁺ than the 25D–75RO, 15D–85RO, 10D–90RO, and 5D–95RO treatments. The 20D–80RO fertigation treatment had the greatest value for both ammonium and P₂O₅.

The entire plug production process took 15 days, between December 8-22, 2023. We began the experiment by placing eight Horticube growth mediums in black 0.25 × 0.51 m starter trays and thoroughly saturated the Horticubes with reverse osmosis (RO) water. We then put a single pelleted “Rex Butterhead” Lettuce Lactuca sativa seed in each cell of the growth cubes, totaling 104 seeds. We repeated this process seven additional times until we had eight seed-filled Horticube trays. Next, we placed a layer of newspaper over each of the eight trays and placed the covered trays inside the grow tent, where the lighting element was turned off for 48 hours to replicate the natural seed imbibition process. After 48 hours, we used a programmable timer (Mechanical 24-Hour Programmable Dual Outlet Timer, BN Link, Santa Fe, California) to provide 16 hours of continuous light for plug production between 06:00 to 22:00 for the duration of the plug production cycle.The experimental setup allowed us to maintain ideal environmental conditions throughout the growth cycle for plug production. Using light sensors, we determined the DLI ranged from 7.76 to 13.36 mol/m²·d with an average of 11.40 mol/m²·d, which falls above the minimum recommended range of 6.5 - 9.7 mol/m²·d [38] . Utilizing the fan’s built-in environmental sensor component, we could monitor the environmental conditions over the plug production cycle. The temperature varied between 21˚C and 26.7˚C with an average value of 24.4˚C. The relative humidity over the growth period varied between 37.6% and 89.3%, averaging 58.7%.

After 48 hours had elapsed, we began the fertigation of both: 1) Control system (synthetic fertigation treatment) and seven dilutions of digestate-based fertigation treatments; 2) 5% Digestate—95% RO Water (5D–95RO); 3) 10% Digestate—90% RO Water (10D–90RO); 4) 15% Digestate—85% RO Water (15D–85RO); 5) 20% Digestate—80% RO Water (20D–80RO); 6) 25% Digestate—75% RO Water (25D–75RO); 7) 30% Digestate—70% RO Water (30D–70RO); 8) 35% Digestate—65% RO Water (35D–65RO). Each tray was fertigated by adding 200 - 400 mL of each fertigation source to their respective trays daily for 13 days; volume added depended on the rate of evaporation and uptake from the plugs. Each tray received 200 - 400 mL of their respective fertigation between the hours of 12:00 and 16:00 every day for 13 days. The volume added was based on the plug’s uptake as well as the rate of evaporation. The experimental setup is illustrated in Figure 2 .

To find the comparative growth performance of the plugs with different fertigation treatments, we collected biomass data at the end of the growth cycle. On the 15^th day of growth, we harvested 20 plugs from each treatment tray in a randomized manner for data collection.

We utilized destructive methods to obtain wet weight (g), plug head area (cm²), total leaf area (cm²), total chlorophyll content (mg/cm²), and dry weight (g) for biomass data collection. The control treatment for this experiment contains data collected from a separate experiment that was conducted under similar conditions [39] . To obtain the wet weight, we gently pulled each plug out of the Horticube tray, separated the root material from the plug, and weighed the samples. We separated the root mass since it contained fragments of the Horticube material, which, if weighed, would lead to an inconsistent analysis of the biomass. We measured these values first and immediately after harvest to ensure minimum water loss through plant evapotranspiration which would dry out the plugs and lead to inaccurate measurements.

To calculate the plug head area and the plug’s total leaf area, we installed the Leafscan app on a mobile device (iPhone 14, Apple Inc., Cupertino, California). To calculate the plug head area, we took the entire plug and placed it in the middle of a white sheet of paper with four black dots forming a 10.5 × 10.5 cm reference square, we then took a picture and uploaded it to the Leafscan application. For the plug’s total leaf area, we separated and laid out the leaves of each sample on the paper with the 10.5 × 10.5 cm reference square, taking a picture and uploading it to the Leafscan mobile application. The Leafscan application then utilizes the camera on the mobile device to photograph the leaves and, by running an algorithm that measures the green leaf area in comparison to the blank white area, generates the total leaf area value [39] . The Leafscan app calculated the area inside the contour in pixels, and by using the given reference length of 10.5 cm, it converted the leaf pixel area into the surface area [39] . We collected the plug head area and total leaf area data of 20 samples for each treatment that was later exported in a comma-separated values (csv) format. The Leafscan app measured the leaf area in square centimeters (cm²) with an accuracy of 0.01 cm² [40] .

We then utilized a chlorophyll meter (AMTAST Chlorophyll Meter for Testing Plant Chlorophyll Content Unit SPAD, Amazon Inc., Seattle, Washington) to find the total chlorophyll content (mg/cm²) of each sample. We took three chlorophyll readings per sample and then averaged the values of those readings and converted them to Soil Plant Analysis Development (SPAD) units and subsequently to total chlorophyll content (mg/cm²). The total chlorophyll content of the crop foliage is calculated by converting the atLEAF CHL values to SPAD and considering the relationship between chlorophyll content and SPAD units [41] .

To find the dry weight of the samples collected, we placed all twenty samples in individual brown paper bags and placed them in a drying oven for six days set at a temperature of 65˚C for each treatment. After six days, we returned to weigh the dried samples and divided the dry weight value obtained by 20 to find the average dry weight of each sample. We averaged the values because the weighing scale we utilized had a precision of 0.001 g and, if individually sampled, would not register on the weighing scale.

We utilized the collected data described in the previous section to derive leaf area index (LAI, cm²/cm²) and specific leaf area (SLA, cm²/g). We also converted SPAD units to total chlorophyll content (mg/cm²). The LAI (cm²/cm²) is defined as the ratio of the total leaf area (cm²) and the ground area (cm²). The SLA (cm²/g) is defined by the ratio of each plug’s total leaf area (cm²) to dry weight (g). We found the SLA (cm²/g) by dividing the average total leaf area value by the average dry weight value for each treatment. The resulting calculated value is the average SLA (cm²/g) for each treatment.

We used statistical analysis to understand and compare the effects of synthetic and digestate-based fertigation treatments on plug production. Descriptive statistics were used to demonstrate our results from the data collected. Due to near-perfect correlations between nitrogen and chlorophyll, and total leaf area and wet weight, only wet weight and chlorophyll were retained for further statistical analyses. Dry weight was also excluded as it tends to be strongly correlated with wet weight and the recorded values were averages instead of individual measurements. The remaining variables (wet weight, plug head area, and chlorophyll) did not satisfy multivariate normality, so a permutation multivariate analysis of variance (MANOVA) was performed using the vegan package [42] in R [43] . Pairwise Kruskal-Wallis tests and Dunn’s tests were conducted during post-hoc analyses via the ggstatsplot package [44] .