Evaluation of Recycled Materials as Hydroponic Growing Media

: Conventional soilless growing media, such as perlite, are mined from nonrenewable resources and can only be disposed of in landﬁlls after limited use. There is a need to investigate novel, sustainable growing media adapted from waste or engineered to be reused over multiple cycles. This study investigated waste almond shells and a recycled plastic drainage plank as hydroponic growing media alternatives. Physiochemical properties were evaluated, and a germination and greenhouse growth trial was conducted to understand the e ﬀ ect these media have on production and nutritional quality of lettuce ( Lactuca sativa L. cv. Catalogna Verde). Drought testing was carried out to understand how the media a ﬀ ected the lettuce’s response to water stress. In comparison to perlite, yields under regular irrigation were reduced by 52% in almond shells and 72% in plastic planks, although lettuce grown in almond shells still obtained commercially relevant yields. Reduced yields in almond shells were likely caused by the shell’s high salinity. Lettuce growth in plastic planks was limited by impeded root growth and low water-holding capacity. In conclusion, with minor alterations, almond shells could be used as a sustainable growing media alternative to perlite in hydroponic lettuce production. More research is needed to manufacture the planks to be conducive to plant growth.


Hydroponics
The interconnected challenges of climate change, extensive agricultural soil degradation, and dwindling freshwater resources, coupled with a rising urbanised population, threaten global food security [1][2][3][4]. Hydroponics, a system of crop cultivation without soil, offers an innovative solution for sustainable food production by preserving soil and land resources whilst reducing water usage in comparison to traditional agriculture [5][6][7]. Hydroponic systems are versatile and can range from passive, low-input systems to high-tech, fully automated systems, with system design easily adapted to local contexts [8][9][10]. Hydroponic systems can be implemented in areas that rely mainly on imported food products and are thus especially prone to food insecurity due to low availability of agricultural land or poor soil quality, such as deserts, dry coastal belts, and urban areas [11][12][13]. Hydroponics can increase food self-sufficiency in these communities, allow for local control of food resources, shorten the scale urban growing contexts [30,39,81]. These media were also tested on their ability to mitigate stress in lettuce crops during an induced drought period, in order to understand their potential to reduce irrigation requirements and buffer shocks in the hydroponic system. This study thus identifies and evaluates two recycled material options which can be used in low-input hydroponic systems and further assesses their resiliency to stresses associated with climate change, such as water shortages.

Growing Media Preparation
Almond shells (AS) and a recycled plastic drainage plank (PDP) were compared to perlite (P) as the control in this study, to evaluate their performance as growing media for lettuce production. Sinclair standard horticultural grade perlite (2.0-5.0 mm) was used (Sinclair Pro; Ellesmere Port, UK). Almond shells were provided by Borges Andalucía (Almería, Spain) as a waste product from their almond processing operations. Raw almond shells were milled and blended, respectively, to two textures, fine (<4 mm) and coarse (<4.75 mm), and these were mixed in equal volumes for use as the growing media. The PDPs are composed of 100% mixed recycled thermoplastics, arranged in a dense network of polymer noodles with approximately 25% voids. This high strength plastic material has a unique surface tension which filters and directs the flow of water.

Physical Properties of Growing Media
Physical properties were assessed on a randomly mixed composite sample of each media. The following physical properties of growing media were evaluated: compacted bulk density [82]; dry bulk density [83]; effective pore space, defined as the volumetric amount of water which saturates the media [84][85][86]; particle size distribution via dry sieving [87]; coarseness index, defined as the percentage by mass of particles with sizes >1 mm [88]; and dry matter and moisture content [89].
Water retention curves were measured for perlite and the almond shell media based on the method by Schindler and Müller [90] using the UMS Hyprop (METER Group; Munich, Germany). PDPs were not tested due to the poor contact between tensiometers and material. The following properties were obtained: water-holding capacity (water held in material at 1 kPa suction), water buffering capacity (water loss in material between 5 and 10 kPa suction), easily available water (water loss in material between 1 and 5 kPa suction), and air-filled porosity (difference between the total porosity and water content at 1 kPa suction) [54,84,91].
Since water retention curves could not be generated for the PDPs, total water-holding capacity was calculated as the volume of water held in the planks after free drainage versus the material's dry volume [84]. The air-filled porosity was calculated as the volumetric percentage of the medium filled with air at the end of free gravitational drainage, which corresponds to the volume of water lost after drainage [86].
The saturated hydraulic conductivity (Ksat) was determined for perlite and almond shell samples by the constant head method [92] using a KSAT (METER Group; Munich, Germany). PDPs were tested, but due to extremely fast drainage, Ksat could not be determined accurately.

Physical Properties of Growing Media
Physical properties were assessed on a randomly mixed composite sample of each media. The following physical properties of growing media were evaluated: compacted bulk density [82]; dry bulk density [83]; effective pore space, defined as the volumetric amount of water which saturates the media [84][85][86]; particle size distribution via dry sieving [87]; coarseness index, defined as the percentage by mass of particles with sizes >1 mm [88]; and dry matter and moisture content [89].
Water retention curves were measured for perlite and the almond shell media based on the method by Schindler and Müller [90] using the UMS Hyprop (METER Group; Munich, Germany). PDPs were not tested due to the poor contact between tensiometers and material. The following properties were obtained: water-holding capacity (water held in material at 1 kPa suction), water buffering capacity (water loss in material between 5 and 10 kPa suction), easily available water (water loss in material between 1 and 5 kPa suction), and air-filled porosity (difference between the total porosity and water content at 1 kPa suction) [54,84,91].
Since water retention curves could not be generated for the PDPs, total water-holding capacity was calculated as the volume of water held in the planks after free drainage versus the material's dry volume [84]. The air-filled porosity was calculated as the volumetric percentage of the medium filled with air at the end of free gravitational drainage, which corresponds to the volume of water lost after drainage [86]. The saturated hydraulic conductivity (K sat ) was determined for perlite and almond shell samples by the constant head method [92] using a KSAT (METER Group; Munich, Germany). PDPs were tested, but due to extremely fast drainage, K sat could not be determined accurately.

Chemical Properties of Growing Media
A representative, composite sample of the perlite and almond shell media were tested for chemical composition and organic matter content. These tests were not performed for PDPs due to lack of an appropriate testing method and equipment for plastic materials. Total C and N contents were determined by the Dumas technique [93]. Water-soluble nutrients were measured on a filtered extract using a weight equivalent to 60 mL of the sample and 300 mL of deionised water [82]. Cl, SO 4 -S, and NO 3 -N were determined via ion chromatography [94]. NH 4 -N was determined using colorimetric analysis [94]. P, K, Ca, Mg, Na, Fe, Mn, Cu, Zn, and B were determined using Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES) [89]. Organic matter content was determined as the loss of mass on ignition at 430 • C [89,95]. To measure pH and electrical conductivity (EC) of the growing media, an aqueous suspension was prepared in a 1:6 (v/v) ratio of growing media to water, shaken at 250 rpm for 1 h. pH was then measured potentiometrically [96], and EC was measured using an EC meter [97,98].

Lettuce Germination Trial
A germination study was carried out to evaluate the impact of the different growing media on the germination of Oakleaf lettuce (Lactuca sativa L. cv. Catalogna Verde) (Moles Seeds Ltd.; Colchester, UK). Seeds were germinated within individual cells in three germination trays, where each tray represented one experimental unit and held 10 randomly distributed cells of each growing media (P, AS, and PDP). The 13-day germination study was conducted in a growth room at Newcastle University under semi-controlled environmental conditions with temperature ranging between 20 and 30 • C and daylength of 16 h in July 2018. Trait measurements consisted of germination percentage, time of emergence, seedling height, and number of leaves, measured according to the Cornell Hydroponic Lettuce Handbook [99].

Experimental Design and Overview
A greenhouse experiment was conducted to test the effect of different growing media and an induced drought period on the growth of Oakleaf lettuce (Lactuca sativa L. cv. Catalogna Verde). The study was conducted in a greenhouse at Newcastle University farms, and the experiment ran for 34 days (July-August 2018). Seedlings at the two-leaf stage, originally germinated in Humax Original compost (Humax; Ipswich, UK), were transplanted into each growing media. Temperatures during this experiment ranged from 12 • C at night to a maximum of 41 • C during the day (mean temperature 21.9 • C), with 91% of the days having temperatures that exceeded the optimal growing range for lettuce at 15-25 • C [100]. Therefore, it should be noted that yields obtained during this trial could have been reduced due to unforeseen heat stress, which was experienced by lettuce growers across the UK [101].
Lettuce was grown in a hydroponic, drip-irrigated bag culture system, where each grow bag represented one experimental unit. A randomised complete block design was used with three growing media treatments (P, AS, PDP), two irrigation treatments (temporary induced drought [Dr] and regular irrigation [Reg]), three replications of each treatment, and three lettuce plants in each replication. Thus, a total of six treatments were evaluated (P Reg, AS Reg, PDP Reg, P Dr, AS Dr, and PDP Dr), with 9 lettuce plants grown per treatment at a planting density of 17 plants m −2 . Grow bags of almond shells and perlite were approximately 50 cm × 35 cm × 15 cm. Three stacked PDPs were also wrapped in bags of 50 cm × 21 cm × 9 cm, with holes drilled into the top of the planks for transplanting lettuce seedlings.

Irrigation and Nutrient Solution
Nonrecirculating drip irrigation (0.9 L h −1 flow rate) was used to control water and nutrient supply to the lettuce, with one drip emitter per plant. Perlite irrigation conditions were used as the baseline for all treatments. The irrigation schedule was formulated based on producing 30% leachate by volume in the perlite grow bags, which was equivalent to 7 events of 5 min per day. All seedlings were first irrigated for 14 days. After this period, irrigation was stopped in three replicates of each growing media to simulate a drought treatment. This lasted until the matric potential in the control growing media (perlite) reached −20 kPa (~7 days), followed by deployment of regular irrigation for all replicates for another week until harvest.
All crops were irrigated from the same 200 L reservoir containing nutrient solution made with potable water. This was prepared from stock solutions (Appendix A Table A1) based on formulations for leafy greens [102]. Each stock solution was added in equal proportions to the reservoir to maintain an EC that was 1.15-1.25 mS cm −1 above potable water EC, which resulted in a final EC of 1.5-1.6 [99]. The pH of irrigation water was maintained between 5.6 and 6.0 [99]. Leachate from each grow bag was collected and tested periodically throughout the experiment for pH and EC.

Measurements
During the induced drought period and thereafter until harvest, 30-min matric potential readings were taken using T5 tensiometers (METER Group; Munich, Germany). Stomatal conductance readings and infrared thermal images were taken on lettuce before, during, and after the drought period as indicators of stress, using a Porometer AP4 (Delta-T Devices Ltd.; Burwell, UK) and FLIR Thermal Camera (FLIR Systems, Inc.; West Malling, UK), respectively. Temperature readings from thermal images were normalised using a lettuce plant potted in compost as a temperature reference, which was incorporated in each picture during the imaging process [103].
At time of harvest, plant shoot fresh weight, root fresh weight, and number of leaves were determined. Errors in root weight measurements could have occurred as roots were broken in the media during extraction, especially for lettuce grown in PDPs. To obtain dry shoot and root weights, lettuce samples were freeze-dried for 7 days at −20 • C and then weighed.
Tissue analysis was also performed on homogenised samples of the lettuce leaves in triplicates for each treatment to obtain nutrient content. Samples were oven-dried and ground to pass through a 0.5 mm screen. N content was determined by the Dumas technique [93]. P, K, Mg, Ca, S, Mn, Zn, B, and Cu contents were determined by ICP-OES [89].

Statistical Analysis
All statistical analyses were carried out using the R commander package [104] in R [105]. Data were subjected to a two-way general linear model of analysis of variance (ANOVA) to determine treatment effects. Where significant effects were determined, Tukey's test was used to separate differences between treatment means at the 95% (p < 0.05) level of confidence. For data sets with repeated-measures, a linear mixed-effects model was first applied using the nlme package in R [106], and a two-way, repeated-measures ANOVA was performed. Normality of residuals was tested with qqnorm and where necessary, data were cube root transformed. Table 1 displays physical and chemical properties evaluated for the growing media used in this study compared to values considered ideal for soilless media. An additional chemical analysis was also performed for perlite and ground almond shells (Table A2). Values from material testing (n = 1). a [84]; b [107]; c [108]; d [109]; e [91]; f [40]; g [110]. * Ideal values apply only to organic media.

Physical and Chemical Properties of Growing Media
In general, perlite had suitable physical and chemical properties as expected, except for its generally low water-holding properties, due to its fast drainage. The main disadvantages of the almond shell media were its high bulk density, low air-filled porosity, and high EC, consistent with conclusions obtained by Valverde et al. [51]. The almond shells did have a higher water-holding capacity and easily available water than perlite, seen as a major advantage of this medium. This is in contrast to a previous study by Urrestarazu et al. [49], which reported low water-holding capacity and easily available water as the main problems of using almond shells as soilless media. Since the same source of almond shells (Borges Andalucía) was used in both studies, the differences likely arise from the fact that the almond shells were processed to a much finer particle size in this study (coarseness index of 59% compared to 85%) which drastically increased the water-holding capacity by 169% and easily available water by 1708% compared to values found by Urrestarazu et al. [49]. However, the use of finer almond shell particles in this study also raised the bulk density and lowered the air-filled porosity outside the range of ideal values. Therefore, an ideal balance of fine and coarse particles must be found to improve the use of almond shells as a soilless growing media in the future. This could potentially be achieved by processing the almond shells solely from blending (rather than from a mix of blending and milling as in this study), which would give a coarseness index of 79% while still having nearly 50% of particles within the ideal range of 0.25-2.0 mm ( Figure A1).
A major disadvantage of the PDP was its low effective pore space, which in both prior studies by the supplier and in this study was found to prevent root penetration and expansion. Additionally, since water can only be held in the PDPs via surface tension in the macropores created by the polymer noodles, this material has a very limited water-holding capacity. Thus, the PDPs could be improved as a hydroponic substrate by increasing total effective porosity, which could be adjusted during manufacturing. Additionally, creating a heterogenous pore size distribution within the media would likely improve its water-holding capacity, making this substrate more amenable to plant growth [85]. However, as the porosity of this material increases, it loses structural stability; thus, more research is needed to identify the ideal porosity needed in this medium to provide adequate conditions for plant growth. Figure 2 displays the particle size distribution (a) and water retention curves (b) for perlite and ground almond shells. The 0.25-2.0 mm particle size range is considered as the ideal for growing media [107], as coarser materials will generally have excessive aeration and low water retention, and finer materials may prevent root penetration and decrease plant available water and aeration [25,111,112]. As seen in Figure 2a, the almond shell media used in this study had 74.6% of Agronomy 2020, 10, 1092 8 of 26 particles (by weight) within the range of 0.25-2.0 mm, while perlite had only 36.2% of particles within this range. Further, the almond shell media had 17.4% of particles within the range of 0.1-0.5 mm compared to only 7.77% for perlite, which is the particle size range predicted to be responsible for easily available water and water-holding capacity in soilless media [113,114].
Agronomy 2020, 10, x FOR PEER REVIEW 8 of 28 The almond shells also had a more heterogenous particle size distribution than perlite, owed to the different processing methods used (blending and milling), which created a heterogenous pore size distribution, allowing water to be held at various tensions in the media [85,86]. This is evidenced in Figure 2b, with the almond shell water retention curve exhibiting a low, gentle slope compared to perlite, as water drains slower from different pore sizes. Indeed, the drying of the saturated almond shells took 39 days compared to 13 days for perlite. In contrast, the rapid desaturation of perlite at −0.9 kPa indicates that this material has a more homogenous pore size distribution with a higher proportion of larger pores which drained quickly; this also contributes to higher air-filled porosity in comparison to the almond shells [115].

Drought Stress
During the drought period (25 July 2018-1 August 2018), three replicates of each growing media did not receive any irrigation, while three other replicates of each growing media remained under regular irrigation (seven irrigation events daily). From this period until harvest (7 August 2018), measures of stomatal conductance and leaf surface temperatures were taken as indicators of plant stress, with higher stress relating to lower stomatal conductance and higher leaf temperatures [103,116]. Growing media were evaluated on their ability to buffer drought stress by maintaining and holding water for the plant, therefore allowing the crop to maintain high stomatal conductance and productivity. Figure 3 shows infrared thermal images taken of lettuce crops grown in each growing media and irrigation treatment throughout the drought period. The most obvious thermal stress is evidenced by the droughted lettuce grown in the PDPs, where increased leaf surface temperatures can be observed just after the first missed irrigation. This is expected as the PDPs have very minimal water-holding capacity. Little difference is seen between regularly irrigated and droughted lettuce grown in perlite and almond shells. In some cases, stress seems to lessen in droughted crops over time, although this can be attributed to variability in daytime temperature. Still, thermal imaging can be used as a quick way to evaluate crop stress; however, to evaluate significant differences it is necessary to compare measured values of stomatal conductance and leaf surface temperatures ( Figures A2 and A3). The almond shells also had a more heterogenous particle size distribution than perlite, owed to the different processing methods used (blending and milling), which created a heterogenous pore size distribution, allowing water to be held at various tensions in the media [85,86]. This is evidenced in Figure 2b, with the almond shell water retention curve exhibiting a low, gentle slope compared to perlite, as water drains slower from different pore sizes. Indeed, the drying of the saturated almond shells took 39 days compared to 13 days for perlite. In contrast, the rapid desaturation of perlite at −0.9 kPa indicates that this material has a more homogenous pore size distribution with a higher proportion of larger pores which drained quickly; this also contributes to higher air-filled porosity in comparison to the almond shells [115].

Drought Stress
During the drought period (25 July 2018-1 August 2018), three replicates of each growing media did not receive any irrigation, while three other replicates of each growing media remained under regular irrigation (seven irrigation events daily). From this period until harvest (7 August 2018), measures of stomatal conductance and leaf surface temperatures were taken as indicators of plant stress, with higher stress relating to lower stomatal conductance and higher leaf temperatures [103,116]. Growing media were evaluated on their ability to buffer drought stress by maintaining and holding water for the plant, therefore allowing the crop to maintain high stomatal conductance and productivity. Figure 3 shows infrared thermal images taken of lettuce crops grown in each growing media and irrigation treatment throughout the drought period. The most obvious thermal stress is evidenced by the droughted lettuce grown in the PDPs, where increased leaf surface temperatures can be observed just after the first missed irrigation. This is expected as the PDPs have very minimal water-holding capacity. Little difference is seen between regularly irrigated and droughted lettuce grown in perlite and almond shells. In some cases, stress seems to lessen in droughted crops over time, although this can be attributed to variability in daytime temperature. Still, thermal imaging can be used as a quick way to evaluate crop stress; however, to evaluate significant differences it is necessary to compare measured values of stomatal conductance and leaf surface temperatures (Figures A2 and A3).  Table 2 displays values of stomatal conductance and normalised leaf surface temperatures averaged across media and irrigation treatments from several readings taken throughout the drought period. It is important to note that these measurements are highly dependent on other environmental conditions, such as varying air/media temperature, wind speed, humidity, irradiance, and complexities of the canopy structure (i.e., leaf angles and layering) [117]; in order to compensate for that, a normalisation technique was utilised as described in Prashar and Jones [103,118].
Growing media, irrigation, and time all significantly affected stomatal conductance and leaf surface temperatures (p < 0.001). Average values of stomatal conductance were lower for droughted lettuce compared to regularly irrigated lettuce grown in the same media (p < 0.05), thus demonstrating that lettuce undergoing drought in each media experienced signs of stress. However, the significant interaction between growing media and irrigation treatment for both stomatal conductance and leaf surface temperatures (p < 0.01) demonstrates that the induced drought affected the lettuce grown in each media differently. Droughted lettuce grown in perlite had significantly higher stomatal conductance than those grown in PDPs (p < 0.001), with those grown in almond shells not being significantly different from either (p > 0.05). The only significant difference in normalised leaf surface temperatures averaged over the drought period was for droughted lettuce grown in PDPs, which showed higher values compared to all other treatments (p < 0.001). There were no significant differences in stomatal conductance or leaf surface temperatures between regularly irrigated crops over the drought period (p > 0.1).   Table 2 displays values of stomatal conductance and normalised leaf surface temperatures averaged across media and irrigation treatments from several readings taken throughout the drought period. It is important to note that these measurements are highly dependent on other environmental conditions, such as varying air/media temperature, wind speed, humidity, irradiance, and complexities of the canopy structure (i.e., leaf angles and layering) [117]; in order to compensate for that, a normalisation technique was utilised as described in Prashar and Jones [103,118].
Growing media, irrigation, and time all significantly affected stomatal conductance and leaf surface temperatures (p < 0.001). Average values of stomatal conductance were lower for droughted lettuce compared to regularly irrigated lettuce grown in the same media (p < 0.05), thus demonstrating that lettuce undergoing drought in each media experienced signs of stress. However, the significant interaction between growing media and irrigation treatment for both stomatal conductance and leaf surface temperatures (p < 0.01) demonstrates that the induced drought affected the lettuce grown in each media differently. Droughted lettuce grown in perlite had significantly higher stomatal conductance than those grown in PDPs (p < 0.001), with those grown in almond shells not being significantly different from either (p > 0.05). The only significant difference in normalised leaf surface temperatures averaged over the drought period was for droughted lettuce grown in PDPs, which showed higher values compared to all other treatments (p < 0.001). There were no significant differences in stomatal conductance or leaf surface temperatures between regularly irrigated crops over the drought period (p > 0.1). Figures A2 and A3 show the individual readings of stomatal conductance and leaf surface temperature measurements taken on different days. Results show that after the first missed irrigation, droughted lettuce growing in the PDPs already showed lower stomatal conductance (p < 0.1) and higher leaf surface temperatures (p > 0.05) than the regularly irrigated lettuce. By the third missed irrigation, droughted lettuce grown in almond shells showed a lower stomatal conductance than the regularly irrigated lettuce grown in perlite and almond shells (p < 0.05), and by the 15th missed irrigation, droughted lettuce grown in perlite also had a lower stomatal conductance than regularly irrigated lettuce grown in the same media (p < 0.001). Leaf surface temperatures of lettuce grown in both perlite and almond shells did not significantly differ from regularly irrigated crops throughout the drought period. Values are mean ± standard error (n = 9 for irrigation main effect means; n = 6 for substrate main effect means; n = 3 for interaction means). Means followed by a different letter in the same column indicate significant differences by Tukey's test at a significance level p < 0.05. ANOVA p-values were generated from cube root transformed data, with boldface used for significance (p < 0.05). a Means are averaged from readings taken at time points relating to the following numbers of missed Lettuce grown in perlite showed the lowest signs of stress, maintaining the highest stomatal conductance and lowest leaf surface temperatures throughout the drought period compared to other media, although values for lettuce grown in almond shells were not significantly different when averaged over the entire drought period (Table 2). Almond shells were able to provide some buffering capacity to drought but did not reduce stress in the lettuce as much as expected based on high water retention characteristics, which were also evidenced through tensiometer readings taken throughout the drought period, showing that almond shells retained moisture at the lowest suction potential throughout the drought ( Figure A4). The enhanced stress seen in the lettuce grown in almond shells compared to perlite under drought may have been due to exaggerated osmotic stress from the combined effects of the drought and the high salinity of the almond shell media, which would not have been reflected in tensiometer readings of matric (not osmotic) potential [86].
Lettuce grown in the PDPs were extremely responsive to drought due to its almost negligible water-holding capacity. The effects of drought stress were likely compounded by the high heat retention of the material. Thermal conductivities of thermoplastics are an order of magnitude higher than for common horticultural substrates, such as perlite, clay, peat, and bark, so the PDPs are more receptive to ambient temperature changes [119,120]. These results show that PDPs cannot provide any buffering capacity to decrease plant stress in times of drought and should therefore only be used in hydroponic systems that utilise a constant supply of water, such as deep water culture, floating raft systems, and nutrient film technique [10]. Table 3 displays results from the germination trial. While there was no significant difference in final germination percentage between media (p > 0.05), seedlings grown in perlite had a higher final height and number of real leaves than those grown in the PDPs, which were higher than those grown in almond shells (p < 0.001). A faster emergence time was observed when germinated in perlite compared to PDPs (p < 0.05), with the emergence time for seeds sown in almond shells not being significantly different from the other media (p > 0.05). This demonstrates that seeds could germinate effectively in PDPs, which would improve the material's use in plant production as transplanting seedlings into its dense structure would not be necessary. Values are mean ± standard error (n = 3). Means followed by a different letter in the same column indicate significant differences by Tukey's test at a significance level p < 0.05. Table 4 provides an overview of lettuce yield characteristics at the time of harvest based on growing media and irrigation treatments. Lettuce grown in perlite had a higher average fresh and dry shoot weight than lettuce grown in almond shells, followed by lettuce grown in PDPs (p < 0.001). Root fresh and dry weight were higher in lettuce grown in almond shells than in perlite, which had a higher root fresh and dry weight than lettuce grown in PDPs (p < 0.001). The number of leaves per plant was not significantly different for lettuce grown in perlite and almond shells, but was higher in these media than in PDPs (p < 0.05). Thus, the higher shoot weights for lettuce grown in perlite compared to almond shells can be attributed to either larger leaf areas or thicker leaves rather than the number of leaves. The induced drought period lowered fresh shoot weights for all treatments when compared to regularly irrigated lettuce grown in the same media (p < 0.01). Values are mean ± standard error (n = 3). Means followed by a different letter in the same column indicate significant differences by Tukey's test at a significance level p < 0.05.

Effect of Growing Media on Yields
Shoot fresh and dry weights of lettuce crops grown in regularly irrigated perlite and almond shells, as well as perlite under drought, were consistent with hydroponic lettuce yields reported in the literature, which range from 160 to 268 g plant −1 for fresh weight and 6.4 to 13.3 g plant −1 for dry weight, depending on system type and lettuce variety [121][122][123][124]. This therefore indicates production relevant to commercial standards for lettuce crops grown in these treatments, showing that almond shells could still be used as an environmentally conscious media replacement, even though fresh weight yields were 52% lower than in perlite. Lettuce grown in PDPs were not able to reach a commercially relevant yield.
A possible reason for the lower yields of lettuce grown in almond shells compared to perlite is the high salinity of the almond shell media. Indeed, Chong [125] identified high salt content as one of the main limitations of using organic-derived wastes as growing media alternatives. A study by Andriolo et al. [121] showed that lettuce exposure to salinity levels above 2.0 mS cm −1 in nutrient film technique hydroponic systems led to reductions in fresh shoot weight, with salinity levels adjusted via applied nutrient concentrations. In this experiment, the nutrient solution was kept within an optimal range of 1.5-1.6 mS cm −1 , and the salinity of the almond shell media was measured as 0.574 mS cm −1 ; this could have therefore brought the salinity level above the 2.0 mS cm −1 threshold. The EC of leachate from the almond shell media during the first week of growth was found to be exceptionally high, at an average of 4.07 mS cm −1 , compared with perlite and PDPs at 1.06 and 1.30 mS cm −1 , respectively. High salinity levels cause decreases in leaf expansion and total leaf area, thus limiting plant growth and lowering final yields [126][127][128][129]. Since the EC of the almond shell leachate lowered by the end of the growth trial (at an average EC of 1.2 mS cm −1 by the third week), and lettuce grown in almond shells and perlite did not significantly differ in the number of leaves, it is possible that the initial exposure of the lettuce to high saline levels in the almond shells after transplantation caused reduced leaf areas, thereby preventing future growth by lowering water and mineral uptake and rates of photosynthesis. Thus, in accordance with recommendations by Chong [125] for organic waste materials with a high initial soluble salt content, it is possible that yields of lettuce grown in almond shells could be increased if the media is washed or leached by running one or several irrigation cycles to lower its EC prior to use.
Additional causes of the lower yields in the almond shells could be low air-filled porosity, high water retention capacity, and slow desaturation rate, which combined could have led to waterlogging with regular irrigation over time, reducing oxygen supply to plant roots [26,130,131]. This could be amended by changing the irrigation schedule or by adjusting the particle size distribution of the media to be coarser, encouraging aeration and drainage. Finally, as the fresh and dry root weight was significantly higher in lettuce plants grown in almond shells compared to perlite, this indicates that plants grown in the almond shells may have been experiencing nutrient deficiencies. Indeed, the high C:N ratio (48) of the almond shell media in comparison to ideal levels for soilless media (20-40) could have led to N immobilisation in the substrate [91,110,132], and N deficiencies have been shown to lead to decreases in shoot/root ratios in plants as N and biomass becomes allocated to the roots [133,134]. However, the degree to which this influenced yields is not certain. A study from Broadley et al. [134] showed that stomatal conductance, an indicator of plant stress and C assimilation, was significantly lower in N-limited lettuce plants than in control plants; however, in this study, while the stomatal conductance for lettuce grown in almond shells was lower compared those grown in perlite, it was not significantly different.
For lettuce grown in PDPs, a likely cause for lower yields could be restricted root growth, as it was observed that roots formed a confined ball between the first and second stacked planks. Indeed, roots penetrated only through the first plank in which the seedlings were set. Poor root growth, caused by high mechanical impedance of the substrate, has been identified as a major factor contributing to reduced vegetative growth for plants grown in both soils and soilless media [135][136][137]. The low water-holding capacity of the material was also a likely cause of reduced plant growth, as plant roots were only able to take up water during irrigation events, with excess water being immediately leached from the material. This issue was compounded by the low and confined root growth, which further reduced the surface area available for water uptake [132].
Yields for lettuce grown in PDPs would likely be improved if manufactured with increased porosity and a more variable pore size distribution, which would allow for better root development and potentially increase water retention capabilities. PDPs would be better used as a hydroponic substrate in hydroponic systems featuring continuous water supply. In general, it was shown that plants could grow in the material, so PDPs could be used in green roofs even if not specifically for crop production [79,80]. Since germination was effective in PDPs (Table 3), it would be easy to continuously seed and grow in this medium. This would allow recycled plastic planks to be incorporated as a long-life material for sustainable building design, providing dual purposes of drainage as well as a medium for plant growth.

Effect of Drought on Yields
Overall, the effect of drought decreased fresh shoot yields of lettuce grown in perlite, almond shells, and PDPs by 41%, 53%, and 82%, respectively, compared to yields of regularly irrigated lettuce in the same media. Reductions in shoot dry weights were less profound, with reductions of 22.7%, 15.9%, and 80% for droughted lettuce grown in perlite, almond shells, and PDPs, respectively. For perlite and almond shells, this is consistent with the 25% reduction in shoot dry weight seen in a study by Kerbiriou et al. [138], where two cultivars of butterhead lettuce potted in soil and river sand were placed under a drought period 21 days after transplanting and then allowed a recovery period of one week, which was similar to the timeline of drought in this study. Only lettuce crops grown in perlite were able to produce commercially relevant yields after drought. This indicates that commercial systems which utilise perlite as a growing media could potentially reduce irrigation frequency, conserving water, and still produce commercially relevant yields.
Since the almond shell media was able to maintain the lowest water tension over the drought period ( Figure A4), this could allow for improved water use efficiency in plants if other stresses, such as salt stress from the media, were removed and overall yields were increased [139].

Nutritional Quality
Micronutrient deficiencies are a major problem in both the developed and developing world, and these issues are expected to increase with declining soil fertility, increased drought, and high CO 2 levels associated with climate change [140][141][142][143]. Thus, there is a demand for vegetables of high nutritional quality, and, therefore, a need to identify media which can appropriately transfer nutrients to plant roots, especially in cases of water-or nutrient-limited conditions [22,26,144,145]. Table 5 displays mean nutrient concentrations of lettuce tissue evaluated after harvest for each treatment individually, as well as optimal nutrient concentrations for lettuce defined by literature. Lettuce grown in almond shells showed N, Mg, and Fe deficiencies, while lettuce grown in perlite showed Mg, Fe, and B deficiencies. Lettuce grown in PDPs showed no signs of nutrient deficiencies. In addition, all treatments except for perlite under drought showed excess levels of Zn in the leaf tissue; however, these values of high Zn were consistent with ranges of Zn (105-238 mg kg −1 ) found for 18 different cultivars of lettuce grown hydroponically in both inorganic and organic nutrient solutions [146]. Values are mean ± standard error (n = 3). Means followed by a different letter in the same column indicate significant differences by Tukey's test at a significance level p < 0.05. Boldface is used to indicate means lying outside the optimal values. a [147]; b [148]; c [149]; d [150].

Effect of Growing Media on Nutritional Quality
Despite massive reductions in yields, lettuce grown in PDPs consistently had the highest (or not significantly different from the highest) concentrations of all mineral nutrients. This can possibly be explained by the confined root ball structure which formed in between the planks in the grow bag. In times of irrigation, the nutrient solution was applied from the drip stakes directly onto the root ball, which may have allowed for immediate nutrient uptake, compared to the solution being absorbed by the media and then gradually diffusing to the roots such as in perlite and almond shells.
Lettuce grown in almond shells had significantly higher levels of micronutrients, including Mn, Zn, and B, compared to lettuce grown in perlite. This could be due to the higher presence of these minerals in the substrate itself (Table A2), as well as the low pH (5.2) of the almond shell media, as these micronutrients are all preferentially absorbed at pH values below 6.0 [151]. However, lettuce crops grown in PDPs did not have significantly different concentrations of Mn and Zn than those grown in almond shells, despite the high pH of this media. Still, it is possible that the pH of the PDPs did not affect nutrient uptake in lettuce crops, since the nutrient solution had little contact with the substrate before falling directly on the root ball of crops during irrigation, and little nutrient solution was held in the material after irrigation events. Additionally, a study by Neocleous et al. [152] showed that higher salinity levels generally increased the concentrations of Mn, Zn, and B in baby lettuce leaves; this could have further influenced the higher concentrations of these nutrients present in lettuce grown in almond shells due to the relatively high EC of this growing media.
The lower levels of macronutrients such as N, P, K, Mg, and S found in lettuce crops grown in almond shells versus either perlite or PDPs could also be due to the low pH of the almond shells (5.2), as these nutrients are preferentially absorbed at higher pH levels between 6.5 and 8.0 [151]. This issue may have been further exaggerated throughout the crop cycle, as a previous study by Valverde et al. [51] observed a 0.5 decrease in the pH of almond shells during the growth of sweet peppers, likely caused by the production of organic acids from degrading carbon in the media, which occurs during aerobic composting [153]. The fact that the N content of lettuce grown in almond shells was below optimal levels and was significantly lower than N levels for crops grown in perlite indicates that decreased N uptake could have been a limiting factor in lettuce growth, likely caused by N immobilisation in the substrate because of the high C concentration [91,110,132].

Effect of Drought on Nutritional Quality
In general, drought reduces both nutrient uptake by roots and transport of nutrients from the roots to the shoots because of decreased nutrient diffusion in the growing media due to a lack of moisture and lower transpiration rates in the crops from drought stress [154][155][156][157]. These effects, therefore, contributed to the significantly lower levels of N, P, and K found in all lettuce crops which underwent a drought period (p < 0.01). Because plants require much lower levels of micronutrients in general, the effects of drought stress do not influence micronutrient concentrations as severely as macronutrients [158]; this is likely why reductions in the concentrations of most micronutrients were not observed in the droughted lettuce.
Because of the similar or higher micronutrient levels of lettuce grown in droughted almond shells compared to droughted perlite, as well as the almond shell media's remarkable ability to retain water at low potentials over time, this shows that the almond shell media could be used to maintain the nutritional quality of hydroponically-produced crops in cases of water shortages, provided yield can be increased after amending EC levels of the substrate. While lettuce grown in PDPs under drought had consistently high levels of all nutrients, this media could not be used to maintain production under drought conditions due to the material's low water-holding capacity, resulting in extremely reduced yields.

Conclusions
Hydroponics offers an innovative technological solution for sustainable agriculture, preserving soil resources whilst building food self-sufficiency in areas with limited available agricultural land, such as urban areas and arid regions. There are opportunities to increase the sustainability of hydroponic systems by identifying locally-available waste sources that can be repurposed into valuable products and used as feedstocks for soilless growing media. This provides an alternative waste management strategy which supports circular economies and local food production. Soilless growing media can be used to increase the resiliency of hydroponic systems by holding water and nutrients for plants in cases of water or energy shortages. This study showed that almond shells, an agricultural waste product common in the Mediterranean, western United States, and Middle East, could provide a sustainable alternative to perlite. While yields of lettuce grown in almond shells were lower than in perlite, they were still commercially relevant. Leaching the almond shells prior to use could improve yields by lowering the EC of the growing media and thus reducing possible osmotic stress in crops. Adjusting the irrigation schedule and nutrient solution to maintain an ideal balance of water and air in the media and to eliminate any macronutrient deficiencies could also improve the yields achieved with the almond shell media. The high easily available water capacity of the almond shell media, as well as its ability to maintain low water tensions throughout drought, shows that this material would be ideal for use in areas that have limited freshwater sources. The plastic drainage plank, an innovative material engineered from recycled thermoplastics, was able to grow highly nutritious lettuce crops, although not to a commercial yield. This substrate could likely be improved by manufacturing with a higher porosity and wider pore size distribution. It is also recommended that this material be used in hydroponic systems which provide a constant water supply to plants, such as floating raft systems or nutrient film technique. If improved, the PDPs could offer an important growing media alternative, since they can be cleaned and reused almost indefinitely, which is not possible for almond shells. More research is needed to identify soilless growing media options through reuse of waste sources in different local contexts, as well as to identify ways to improve overall hydroponic system sustainability within a circular economy. In this way, hydroponics can be used to produce food with lower use of resources and build equitable, self-sufficient food systems worldwide.

Appendix B
Perlite and almond shells were tested for chemical composition, with results displayed in Table A2.

Appendix C
The particle size distribution of the milled (fine texture) and blended (coarse texture) almond shells are provided in Figure A1. The coarseness index was 45.2% for milled and 79.1% for blended almond shells. These were mixed in equal volumes to make the final almond shell media used in this study.

Appendix C
The particle size distribution of the milled (fine texture) and blended (coarse texture) almond shells are provided in Figure A1. The coarseness index was 45.2% for milled and 79.1% for blended almond shells. These were mixed in equal volumes to make the final almond shell media used in this study.

Appendix D
Additional data taken during and after the induced drought period are provided in Appendix D, showing stomatal conductance ( Figure A2), leaf surface temperatures ( Figure A3), and matric potential readings ( Figure A4).

Appendix D
Additional data taken during and after the induced drought period are provided in Appendix D, showing stomatal conductance ( Figure A2), leaf surface temperatures ( Figure A3), and matric potential readings ( Figure A4). Figure A2. Stomatal conductance readings taken during induced drought period. Figure A2. Stomatal conductance readings taken during induced drought period. Figure A2. Stomatal conductance readings taken during induced drought period.