Abstract
Low natural daily light integrals (DLIs) are a major limiting factor for greenhouse production during darker months (e.g., October to February in Canada). Supplemental lighting (SL) is commonly used to maintain crop productivity and quality during these periods, particularly when the supply chain demands consistent production levels year-round. What remains to be determined are the optimum SL light intensities (LIs) for winter production of a myriad of different commodities. The present study investigated the growth and yield of sunflower (Helianthus annuus L., ‘Black oil’), kale (Brassica napus L., ‘Red Russian’), arugula (Eruca sativa L.), and mustard (Brassica juncea L., ‘Ruby Streaks’), grown as microgreens, in a greenhouse under SL light-emitting diode (LED) photosynthetic photon flux density (PPFD) levels ranging from 17.0 to 304 μmol·m−2·s−1 with a 16-hour photoperiod (i.e., supplemental DLIs from 1.0 to 17.5 mol·m−2·d−1). Crops were sown in a commercial greenhouse near Hamilton, ON, Canada (lat. 43°14′N, long. 80°07′W) on 1 Feb. 2018, and harvested after 8, 11, 12, and 12 days, resulting in average natural DLIs of 6.5, 5.9, 6.2, and 6.2 mol·m−2·d−1 for sunflower, kale, arugula, and mustard, respectively. Corresponding total light integrals (TLIs) ranged from 60 to 188 mol·m−2 for sunflower, 76 to 258 mol·m−2 for kale, 86 to 280 mol·m−2 for arugula, and 86 to 284 mol·m−2 for mustard. Fresh weight (i.e., marketable yield) increased asymptotically with increasing LI and leaf area increased linearly with increasing LI, in all genotypes. Hypocotyl length of mustard decreased and hypocotyl diameter of sunflower, arugula, and mustard increased with increasing LI. Dry weight, robust index, and relative chlorophyll content increased and specific leaf area decreased in kale, arugula, and mustard with increasing LI. Commercial microgreen greenhouse growers can use the light response models described herein to predict relevant production metrics according to the available (natural and supplemental) light levels to select the most appropriate SL LI to achieve the desired production goals as economically as possible.
Microgreens are an emerging commodity in worldwide markets. They consist of vegetable and herb seedlings that are grown in illuminated environments and harvested at or before the first true leaf stage (Kyriacou et al., 2016; Xiao et al., 2012). Microgreen production is attractive to greenhouse growers due to the increasing demand and high value, with wholesale prices ranging from $30 to $50 US per pound (Treadwell et al., 2016). Microgreens have a limited postharvest shelf life (Berba and Uchanski, 2012; Chandra et al., 2012; Kou et al., 2013); therefore, local production, particularly for fresh-cut products, represents the most promising production strategy (Kyriacou et al., 2016). However, low natural light is a major limiting factor in greenhouse production during darker months in higher-latitude regions (Dorais and Gosselin, 2002). For example, the natural outdoor daily light integrals (DLIs) in southern Canada and northern United States in October through February typically range from 5 to 15 mol·m−2·d−1 (Faust and Logan, 2018). With further 25% to 60% reductions of the outside DLIs due to transmission losses through greenhouse glazing and structural materials (Giacomelli et al., 1988; Lau and Stanley, 1989; Llewellyn et al., 2013; Ting and Giacomelli, 1987), crop-level natural DLIs may be insufficient to grow many commodities. Nonetheless, contemporary consumers expect high-quality greenhouse-grown produce year-round. Because low DLIs are generally associated with reductions in yield and quality, high-quality winter-grown produce can command a substantial price premium. Therefore, supplemental lighting (SL) may be economical for growing high-value commodities during the winter months (Dorais and Gosselin, 2002; Dorais et al., 2017).
High-pressure sodium (HPS) fixtures have been commonly used in greenhouses for SL (Hemming, 2011; Ouzounis et al., 2015). Recently, light-emitting diodes (LEDs) have been increasingly used as an alternative to HPS due to their comparatively high energy efficiency and durability (Mitchell et al., 2015; Morrow, 2008; Nelson and Bugbee, 2014; Pattison et al., 2016). In addition, the potential for spectrum and intensity control with LEDs give growers more power to manipulate their crops with light and save energy when natural lighting is sufficient (Chang et al., 2014; Llewellyn and Zheng, 2018a, 2018b; Morrow, 2008). Considerable research has been devoted to investigating different SL spectrum treatments in greenhouse environments (Bantis et al., 2018) where comparable supplemental light intensities (LIs) have yielded similar crop productivity metrics for LED vs. HPS in many commodities (Currey and Lopez, 2013; Hernández and Kubota, 2015; Llewellyn et al., 2019; Martineau et al., 2012; Poel and Runkle, 2017; Randall and Lopez, 2014).
Increasing LI has generally been associated with increases in fresh weight (FW) and percentage of dry weight (DW), decreases in height, and decreases in specific leaf area (SLA) of juvenile growth stages (including microgreens) of commodities grown in sole-source (SS) (Gerovac et al., 2016; Kitaya et al., 1998; Meng and Runkle, 2019; Park and Runkle, 2018; Samuolienė et al., 2013; Viršilė and Sirtautas, 2013; Yao et al., 2017) and greenhouse (Boivin et al., 1987; Dorais and Gosselin, 2002; Gaudreau et al., 1994; Oh et al., 2010; Rezai et al., 2018; Torres and Lopez, 2011) production environments. These relationships between LI and plant growth have borne out in seedlings of the majority of herbaceous and woody species that have been investigated worldwide (Poorter et al., 2019).
It is often not possible to infer the integrated amount of photosynthetically active radiation (PAR) a crop received [i.e., total light integral (TLI); mol·m−2] in the literature, particularly in greenhouse environments, due to incomplete descriptions of the canopy-level PPFD from natural and/or supplemental sources. Although TLI [also known as integrated photosynthetic photon flux (IPPF), Chun et al., 2001] is not yet a commonly-used metric in the scientific community, it can normalize quantifications of light “doses” from different production scenarios (e.g., SS vs. greenhouse) and lighting environments (e.g., LI, photoperiod, and lengths of production cycles), thus enabling greater relevance between different studies. When considering yield metrics on an area basis, the relationships between yield and TLI are a direct measure of light use efficiency (LUE) (e.g., g(marketable biomass)·mol−1(PAR)). A caveat of making comparisons between studies on a TLI-basis is that even if plants receive the same TLI, they may respond differently depending on the dynamics of how light was provided (e.g., 200 μmol·m−2·s−1 for 8 h vs. 100 μmol·m−2·s−1 for 16 h) and other differences in the growing environment.
Some studies have investigated sufficiently high LI to show distinct maxima in various morphological and physiological parameters, similar to the saturation phenomena of photosynthesis responses to high LI (van Iersel, 2017). He et al. (2019) grew cherry tomato seedlings for 25 d under SS LED LI ranging from 50 to 550 μmol·m−2·s−1 with a 12-h photoperiod (TLI ranged from 54 to 594 mol·m−2). They showed that FW, DW, and health index [i.e., stem diameter / stem height × DW, Fan et al. (2013)] all plateaued at 300 μmol·m−2·s−1 (i.e., TLI of 324 mol·m−2) and that photoinhibition occurred at higher LI. Rezai et al. (2018) and Soustani et al. (2014) investigated the effects of different levels of neutral-density shade on growth and morphology. Leaf number, FW, DW, and oil yield of 5-month-old sage (Salvia officinalis L.) were all maximized when grown outdoors under 30% shade vs. deeper shade treatments or full sunlight (Rezai et al., 2018). Soustani et al. (2014) showed that shade level had greater morphological influences than seed size in Quercus castaneifolia seedlings, with seedlings grown at moderate levels of shade having the greatest biomass. However, insufficient information was provided, in both of these studies, to ascribe absolute LI levels to their treatments or calculate TLI. Viršilė and Sirtautas (2013) grew borage (Borago officinalis L.) as microgreens (no seeding density was provided) for either 10 or 12 d (production times are conflicting) under SS LED LI ranging from 110 to 545 μmol·m−2·s−1 with a 16-h photoperiod (TLI ranged from 63 to 314 mol·m−2, assuming 10-d production). They found that borage HL decreased and FW increased as PPFD increased from 110 to 440 μmol·m−2·s−1, followed by the reverse trends at 545 μmol·m−2·s−1. Our previous study (Jones-Baumgardt et al., 2019) investigated the SS production of four Brassicaceae microgreens grown under LED LI ranging from 100 to 600 μmol·m−2·s−1 with a 16-h photoperiod (TLI ranged from 58 to 380 mol·m−2) and found linear reductions in HL with increasing LI, but saturating increases in FW and DW with increasing LI. Gerovac et al. (2016) and Samuolienė et al. (2013) also both investigated the influences of SS LED LI on several Brassicaceae microgreens. Gerovac et al. (2016) reported that there were LI treatment effects in FW in only one of three genotypes and that the maximum LI was insufficient to saturate biomass production in that genotype. They also reported biomass accumulation on a per plant basis, which is not truly reflective of how microgreens are grown and sold commercially (i.e., in very dense canopies). Samuolienė et al. (2013) may have used sufficiently high LI (545 μmol·m−2·s−1) to saturate biomass accumulation but they did not provide biomass accumulation data in their results. Fu et al. (2012a, 2012b) investigated the effects of SS LI, which ranged from 100 to 800 μmol·m−2·s−1 of combined HPS and sunlight dysprosium [i.e., metal halide (MH)] light (with a 14-h photoperiod), on 10-d-old lettuce seedlings that were then harvested 27 d after sowing (TLI ranged from 86 to 685 mol·m−2). They showed that FW plateaued at LI between 400 and 600 μmol·m−2·s−1 (i.e., TLI between 343 and 514 mol·m−2). Notably, they also attempted to relate results from indoor production (i.e., plant factories) to greenhouses, but they did not account for the substantial contributions of natural lighting in the overall incident PAR in most greenhouse environments. Erwin and Gesick (2017) related leaf-level photosynthesis of chard, kale, and spinach to the anticipated optimum LI for production of these commodities based on their light saturation points (LSP). However, they also provided a detailed discussion of the myriad limitations of relating leaf-level measurements to growth and yield in production scenarios.
The goal of SL in a greenhouse production scenario is to balance the increases in yield with the additional costs of inputs (i.e., lighting infrastructure and electricity). Since the efficiency of the conversion of electricity into biomass is extremely sensitive to LI at moderate DLIs (Bugbee and Monje, 1992), it is likely that there is an economic optimum for the SL LI to use in any greenhouse commodity production scenario. This optimum may depend on many other factors, including supply chain demands, geographic location, time of year (i.e., availability of natural light), and many integral aspects of the growing environment (e.g., temperature, VPD, CO2 concentration, nutrient availability, length of production cycle, etc.). The objective of this study was to quantify the effects of supplemental LED LI during the winter months in a higher-latitude region (Ontario, Canada) on the growth and yield of four economically important greenhouse-grown microgreens: sunflower, kale, arugula, and mustard. By reporting the results in terms of both DLI and TLI, they are hopefully more broadly applicable to disparate production environments. TLI may be a particularly useful metric for comparing the results of different LI trials because it accounts for differences in photoperiod, natural DLI (if natural light is present), supplemental LI, and production time. This is particularly important for greenhouse production systems, where natural DLIs can range by an order of magnitude, depending on weather, season, and geographic region (Faust and Logan, 2018; Poorter et al., 2019).
Materials and Methods
An SL gradient experiment was performed at a commercial greenhouse near Hamilton, ON, Canada (lat. 43°14′N, long. 80°07′W). This involved intentionally creating heterogeneity in canopy-level SL PPFD (through careful fixture placement) within a common growing environment, rigorously assessing the spatial variability in canopy-level SL PPFD at high resolution (performed at night), and assigning selected trays of microgreens to very specific fixed coordinates (i.e., PPFD treatments). The application of gradient-type experimental designs for LI investigations is described further in the Discussion section. The greenhouse was a modern facility comprising 28 gutter-connected (i.e., Venlo-style) bays, each measuring 6.4 m × 70.7 m, with the gutters running in the east–west direction. The whole greenhouse was outfitted with mobile sub-irrigation benches, each measuring 5.46 m × 1.61 m, that could be moved throughout the facility on rails. The roof of the entire production area was recently (Spring 2017) refitted with UVA-transmissive acrylic cladding (16mm Acrylite® Alltop ultraviolet transmitting acrylic double-skin sheet; Evonik Cyro Canada Inc., Toronto, ON, Canada). Typical spectral photon flux distribution (300 to 800 nm) at bench level from natural and SL sources are provided in Fig. 1. The experimental growing environment consisted of four adjacent benches that were fixed in place on the rails, with the near side of the experimental plot positioned ≈3 m from the west end of a centrally located greenhouse bay.
Relative spectral photon flux distribution, at crop level, of natural light (midday) and supplemental light-emitting diode (LED) light.
Citation: HortScience horts 55, 2; 10.21273/HORTSCI14478-19
Lighting setup.
Ten LED fixtures (LX502G; Heliospectra AB, Gothenburg, Sweden) were hung in pairs 1.4 m above the benches near each of the four corners and in the geometric center of the four-bench growing area. The Euclidian coordinates (in meters; from the northeast corner of the experimental area) of the center of each fixture are provided in Table 1. These fixtures had three individually controllable spectrum channels for blue (B), white (W), and red (R) LEDs. The dominant wavelength and full width at half maximum were 445 nm and 20 nm for B and 660 nm and 20 nm for R channels, respectively. The broad-spectrum photon flux from the W channel was described as having a 5700 K correlated color temperature (CCT). All channels were operated on full power, providing an SL photosynthetic photon flux ratio (average ± sd, n = 200) of B (400–500 nm), green (G, 500–600 nm), and R (600–700 nm) of B (15.5 ± 1.2): G (11.6 ± 0.6): R (72.9 ± 1.7) (Fig. 1).
The Euclidian coordinates (in meters) of the center of each fixture relative to the upper left corner of the growing area.
The canopy-level SL PPFD and spectral quality were measured (at night) at the geometric center of 200 subplot locations (Fig. 2) using a radiometrically calibrated spectrometer (XR-Flame-S; Ocean Optics, Dunedin, Fla., USA) coupled to a 400-nm × 1.9-m patch cord with a CC3 cosine corrector. Using the Spectrasuite software package, spectral power distribution was measured and converted to photon flux using the PARspec subroutine. The SL PPFD ranged from 17 to 304 μmol·m−2·s−1 at crop level over the entire experimental area. These measurements closely matched the IES file-based lighting simulations provided, on a 5.2-cm square grid, by Heliospectra for the whole experimental area (data not shown). The SL PPFD and spectral photon flux distribution were re-measured at each location at the end of the experiment to verify steady-state SL conditions at each sample location. The LEDs were operated on a 16-h photoperiod (0600 to 2200 hr), providing SL DLIs ranging from 1.0 to 17.5 mol·m−2·d−1. A sunlight-calibrated quantum sensor (SQ-110; Apogee Instruments, Logan, UT, USA) placed just above the light fixtures in the center of the growing area and tethered to a datalogger (U-12; Onset Computer Corporation, Bourne, MA) collected natural light PPFD data on 120-s intervals throughout the trial. These data were used to calculate the natural DLI within the experimental greenhouse area for each day of the trial (Fig. 3). Although the full four-bench experimental area was covered with trays of microgreens, only the plants harvested from selected locations (for which the SL LI was known) were assessed for growth and yield.
Canopy-level supplemental photosynthetic photon flux density (PPFD; μmol·m−2·s−1) at the geometric center of each tray location (50 trays per genotype). The four genotypes (sunflower, kale, arugula, and mustard) were randomly assigned to one of four quadrants; which have different background fills and are labeled accordingly. All growth and harvest metrics were taken from plants harvested from a circle with a 5-cm radius (76.4 cm2) located in the center of each tray. The dark grey areas represent border trays and the trays located directly below the lighting fixtures, which were not assessed for growth.
Citation: HortScience horts 55, 2; 10.21273/HORTSCI14478-19
Gutter-level natural sunlight daily light integrals (DLIs; mol·m−2·d−1) from 1 to 13 Feb. 2018. The mean (± sd; n = 8 for sunflower, n = 11 for kale, and n = 12 for arugula and mustard) natural DLIs were 6.5 ± 1.7, 5.9 ± 1.7, 6.2 ± 1.7, and 6.2 ± 1.7 mol·m−2·d−1 for sunflower, kale, arugula, and mustard, respectively.
Citation: HortScience horts 55, 2; 10.21273/HORTSCI14478-19
Growing media and seeding.
Seeds of sunflower (Helianthus annuus L., ‘Black oil’, 5.99 seeds/g), kale (Brassica napus L., ‘Red Russian’, 367 seeds/g), arugula (Eruca sativa L., 641 seeds/g), and mustard (Brassica juncea L., ‘Ruby Streaks’, 612 seeds/g) were grown in fiber trays (23.5 × 48.5 × 3.5 cm) for 8 to 12 d after sowing (DAS). The growing substrate comprised (by volume) 30% peat, 30% compost, 30% coir, and 10% perlite. Soil macronutrient and micronutrient concentrations were comparable to those reported by Jones-Baumgardt et al. (2019). Trays were machine-filled (3.5 cm deep) with substrate and seeded on the day the experiment was started. Seeding rates were 94, 8, 4, and 4 g/tray, resulting in seeding densities of 0.50, 2.58, 2.25, and 2.15 seed/cm2 for sunflower, kale, arugula, and mustard, respectively. The experimental area was divided into four quadrants with each quadrant randomly assigned to one of the four genotypes (Fig. 2). Seeded trays of the respective genotypes were placed in the geometric center of each of the 54 subplot locations within each quadrant (Fig. 2). Trays of border plants were placed along the outer columns of each subplot (shaded in grey in Fig. 2). Trays of microgreens were top-irrigated until minimal drainage was observed, starting on 1 Feb. (2018) at 1400 hr and as needed thereafter (following the grower’s protocol).
Growth and morphology measurements.
Growing environment.
Due to the different lengths of production cycles and daily variations in natural lighting (Fig. 3), each crop was exposed to slightly different average natural DLIs. The average natural DLIs (mean ± sd) and TLIs were 6.50 ± 1.70 mol·m−2·d−1 and 52.0 mol·m−2 for sunflower, 5.94 ± 1.70 mol·m−2·d−1 and 65.3 mol·m−2 for kale, and 6.20 ± 1.70 mol·m−2·d−1 and 74.4 mol·m−2 for both arugula and mustard. The natural TLIs were calculated by summing the DLI of each full day of production plus half of the DLI for both the start and harvest days, over each respective genotype’s production period. The air temperature and relative humidity (RH) targets were set at constant levels of 21 °C and 40%, respectively. Air temperature and RH were recorded (every 120 s) using a greenhouse environmental control system (Damatex Inc, Montréal, QC, Canada). The average air temperature and RH were 21.0 ± 1.1 °C and 42.0 ± 4.1% (mean ± sd), respectively. No supplemental CO2 was used in the experiment.
Statistical analysis.
Data were analyzed using R statistical software (RStudio 1.1.453; Auckland, New Zealand). HL, HD, LA, SLA, RCC, and RI were analyzed using linear regressions and FW and DW were analyzed using asymptotic light response curves, with both the independent (i.e., LI) and dependent (i.e., production and harvest indices) values as continuous variables. The relationships between LI and yield (FW or DW) were determined using the asymptotic model y = a + be(cx) (Delgado et al., 1993), where y, x, a, and e represent yield, LI (i.e., DLI or TLI), estimated maximum yield, and Euler’s number, respectively. The parameters for a, b, and c were derived from nonlinear regressions. All regression analyses were evaluated at a significance level of P ≤ 0.05. The best fit model equations are only presented for parameters with significant regressions, whereas overall means (i.e., pooled data from total LI ranges) are presented when there were no treatment effects. Normality of residuals and homoscedascity of variances were confirmed using the Shapiro-Wilk and Levene tests, respectively.
Results
To make the results of this study most directly relatable to those of similar studies, the results are discussed primarily in terms of average DLI. However, both DLI and TLI models are presented when applicable. Because the growing periods were not consistent across all genotypes, growth and yield measurements modeled by TLI may provide more relevant comparisons between genotypes within the present study and also to results from other studies that have different production times and lighting conditions.
Sunflower.
There were no effects of LI on HL, DW, SLA, RI, and RCC in sunflower (Table 2). As average DLI increased from 7.5 to 23.5 mol·m−2·d−1, HD and LA increased linearly by 7.8% and 11%, respectively. FW increased asymptotically by 15% as average DLI increased from 7.48 to 23.5 mol·m−2·d−1.
Effect of daily light integrals (DLIs) and total light integrals (TLIs) ranging from 6.9 to 23.7 mol·m−2·d−1 and 59.9 to 284 mol·m−2, respectively, on growth and yield measurements of sunflower (Helianthus annuus L., ‘Black oil’), kale (Brassica napus L., ‘Red Russian’), arugula (Eruca sativa L.), and mustard (Brassica juncea L., ‘Ruby Streaks’) microgreens.
Kale.
There were no effects of LI on HL and HD in kale (Table 2). LA, RI, and RCC increased linearly by 17%, 60%, and 18%, respectively, and SLA decreased linearly by 27% as average DLI increased from 6.9 to 23.5 mol·m−2·d−1. FW and DW increased asymptotically by 43% and 64%, respectively, average DLI increased from 6.9 to 23.5 mol·m−2·d−1.
Arugula.
There was no effect of LI on HL in arugula (Table 2). HD, LA, RI, and RCC increased linearly by 16%, 40%, 123%, and 21%, respectively, and SLA decreased linearly by 30% as average DLI increased from 7.2 to 23.4 mol·m−2·d−1. FW and DW increased asymptotically by 67% and 85%, respectively, as average DLI increased from 7.2 to 23.4 mol·m−2·d−1.
Mustard.
As average DLI increased from 7.2 to 23.7 mol·m−2·d−1, HD, LA, RI, and RCC increased by 12%, 20%, 102%, and 17%, respectively (Table 2). HL and SLA decreased linearly by 8.5% and 40%, respectively, as average DLI increased from 7.2 to 23.7 mol·m−2·d−1. FW and DW increased asymptotically by 61% and 78%, respectively, as average DLI increased from 7.2 to 23.7 mol·m−2·d−1.
Discussion
Gradient designs.
Greenhouse lighting trials often involve plots of different, reasonably homogeneous intra-plot LI levels (from natural and/or supplemental sources) within a common growing environment (Dansereau et al., 1998; Lopez and Runkle, 2008; Oh et al., 2010; Torres and Lopez, 2011). Treatment plots are typically separated in space to minimize light contamination from adjacent plots, which often limits the quantity of experimental plots and also can create variations in other environmental conditions (e.g., temperature, humidity, and air flow). Physical barriers, such as the opaque blinds commonly used in SS lighting research, have limited use in the greenhouse environment because they can also strongly influence the natural lighting levels at crop level. Furthermore, the requirement for proper replication of treatments inherently limits the number of LI treatment levels that can be investigated. Sometimes consecutive replication methods are used to increase the number of possible treatment levels and/or replications within the available plots (Gaudreau et al., 1994; Gent, 2014; Hoshi et al., 2011; Sublett et al., 2018; Tani et al., 2014; Zhang et al., 2019). However, because there can be considerable temporal variations in natural lighting (Hoshi et al., 2011; Lopez and Runkle, 2008; Poel and Runkle, 2017; Poorter et al., 2019), concurrent experimental designs are highly preferable for LI research trials in greenhouse environments. Furthermore, a regression analysis of data from replicated designs invariably has aggregations of data points at relatively few discreet levels of the independent variable (i.e., LI) with the need for interpolation between these levels. In contrast, gradient-type ecological approaches can outperform replicated designs when evaluating biological responses along a continuous variable (Kreyling et al., 2018), such as LI. Although technically more difficult to arrange than plots of replicated treatments, gradient designs, such as what was used in the present study, have been successfully used to establish supplemental LI effects on year-round greenhouse production (e.g., cut roses) (Bredmose, 1993, 1994).
Yield.
Microgreens are a unique commodity group in that seeding densities are very high (Murphy and Pill, 2010; Kyriacou et al., 2016), production times are extremely short (Treadwell et al., 2016), and input biomass (i.e., seed weight) and marketable yield (i.e., FW of aboveground biomass) may be of similar magnitudes (Jones-Baumgardt et al., 2019). Before activation of the photosynthetic machinery (i.e., transition to autotrophic growth), plant embryos rely solely on stored energy resources (i.e., heterotrophic growth). Therefore, during microgreens’ abbreviated lifetimes, both heterotrophic and autotrophic growth may have substantial contributions to the marketable yield. In the present study, arugula and mustard seed were approximately half the size of kale seed and two decades (100×) smaller than sunflower seed. Although the LI-mediated responses of the different genotypes cannot be directly compared due to the differences in seed sizes, seeding densities, and production periods (all of which followed commercial production protocols), it can be generalized that genotypes with larger seeds had shorter production times in conjunction with lower magnitudes of LI-mediated growth and yield responses. This suggests that genotypes with smaller seeds may present greater levels of phenotypic plasticity in response to increasing LI (Jones-Baumgardt et al., 2019), probably due to reduced reliance on stored energy resources in the seed. Therefore, the phenotypic plasticity of yield responses to increasing LI is an important production consideration when balancing the input costs of seed and lighting infrastructure among different commodities.
Because microgreens are sold on a FW basis, a major goal of commercial microgreen production is to provide growth conditions that maximize FW (Murphy and Pill, 2010) while still maintaining relatively high LUE. Once the LI responses on biomass accumulation start to level off (i.e., approach saturation) at higher LI, LUE decreases accordingly. The maximum LI in some other studies on microgreens and leafy vegetables were insufficiently high to show saturating responses in FW and DW (Gerovac et al., 2016; Lefsrud et al., 2006; Sago, 2016; Samuolienė et al., 2013). Conversely, Fu et al. (2012a, 2012b) showed saturation of FW of romaine lettuce grown in SS at LI ranging from 100 to 800 μmol·m−2·s−1. Because saturation occurred between 400 and 600 μmol·m−2·s−1, the authors recommended a target SL of 400 μmol·m−2·s−1 for winter greenhouse lettuce production, and they further recommended that shading should be provided in summer when ambient PPFD exceeded 600 μmol·m−2·s−1. However, they may not have fully accounted for the probable inter-treatment canopy temperature differences in their setup due to varying proximity to SS fixtures (HPS and MH), which may have contributed to photoinhibition and higher transpiration rates under their highest LI treatments (Berry and Bjorkman, 1980; He et al., 2019). Furthermore, their recommendations for greenhouse production environments did not account for the differences in lighting dynamics between an SS research environment and a greenhouse environment. While no clear maxima LI for yield were found in either the present study or in Jones-Baumgardt et al. (2019), the modeled asymptotic relationships between LI and yield do illustrate the trend of diminishing returns at higher LI, particularly for larger-seeded genotypes. These models should help growers pinpoint a lighting strategy that is optimum for their respective growing systems, specific commodities, and production goals.
Light use efficiency will also change quite dynamically over time, as the cotyledons unfold and become increasingly photosynthetically active. Although it is not uncommon to begin illuminating crops at the time of sowing (such as in the present study), using high levels of SL during this period could be wasted energy because SL will assuredly be used more efficiently once the seedlings begin to transition to autotrophic behavior. Therefore, if the time between germination and the transition to photosynthetic activity were known (probably the first few days after sowing, depending on genotype), then it is conceivable that the TLI levels in the yield equations presented herein could be retroactively adjusted to simulate a protocol whereby SL is provided only between cotyledon unfolding and harvest, which would increase LUE accordingly. Alternatively, because flowering is not important for microgreens and, for a given DLI, plants grown at a low LI for a long photoperiod generally accumulate more biomass than plants grown at a high LI for a short photoperiod (Kitaya et al., 1998; Oda et al., 1989; Ohyama et al., 2005; Sysoeva et al., 2010), continuous lighting may be possible for many genotypes, which could increase yield and/or shorten production time (Lefsrud et al., 2006; Sago, 2016). Supplemental lighting could also be applied only during the night, which may improve overall crop LUE and could dramatically reduce time-of-use energy costs associated with supplemental lighting during daytime hours (Sysoeva et al., 2010). Furthermore, increasing LI in a stepwise fashion, as the crop matures and canopy closes, may also save energy without dramatically affecting yield. This can (easily) be accomplished by using the dimming capabilities of most commercial LEDs, or (if logistically possible) by moving crops through various greenhouse sections that have been outfitted with increasing densities of SL fixtures. More research is needed to evaluate the efficacy of these other lighting strategies for greenhouse production of microgreens.
Morphology.
Elongated hypocotyls (≥5 cm) are an important growth attribute for many commercial microgreen growers because taller plants facilitate mechanical harvesting processes but may reduce the robustness of the harvested tissues. Although compact growth behaviors in mature plant tissues are commonly associated with increased LI (Poorter et al., 2019; Yeh and Hsu, 2004; Zervoudakis et al., 2012), juvenile plants have shown either reductions or no changes in HL with increasing LI in both greenhouses (Craver et al., 2019; Hernández and Kubota, 2014; Lopez and Runkle, 2008; Poel and Runkle, 2017; Pramuk and Runkle, 2005) and SS growing environments (Craver et al., 2018; Kitaya et al., 1998; Meng and Runkle, 2019; Potter et al., 1999; Viršilė and Sirtautas, 2013). There were only LI treatment effects in HL in mustard in the present study; while Gerovac et al. (2016), Jones-Baumgardt et al. (2019), and Samuolienė et al. (2013) all reported general reductions in HL with increasing SS LI in Brassicaceae microgreens. In mustard, the LI treatment effects on HL were only approximately half the magnitude (over the same TLI range) as was reported in Jones-Baumgardt et al. (2019). Furthermore, the absolute lengths of the hypocotyls were markedly longer in the present study than in any of these SS studies. This indicates that other aspects of the growing environment may have overshadowed the LI treatment effects on HL. Because all three of the aforementioned SS studies had the same daytime temperature (21 °C) but 4 °C (i.e., 17 °C vs. 21 °C) cooler nighttime temperatures than in the present study, it is possible that the higher nighttime temperature in the greenhouse environment conveyed a thermomorphogenic effect on hypocotyl elongation (Halliday and Davis, 2016; Heggie and Halliday, 2005). In high-density plantings, HL can also be influenced by photoreceptor-mediated neighbor sensing induced by light quality changes within the canopy (i.e., shade avoidance) (Fiorucci and Fankhauser, 2017; Park and Runkle, 2018; Roig-Villanova and Martínez-García, 2016). Shade avoidance responses related to low red to far red ratios (R:FR) should not generally occur in SS environments, unless FR wavelengths comprise a significant portion of the SS spectral photon flux distribution. Conversely, low R:FR are more likely in greenhouse environments because natural sunlight is relatively high in FR [i.e., R:FR ≈1.1 (Llewellyn et al., 2013)].
In both the present study and Jones-Baumgardt et al. (2019), the increases in DW in the Brassicaceae genotypes were disproportionate to FW with increasing LI. More specifically, the %DW (i.e., DW/FW) increased with increasing TLI; a common trend in microgreens (Gerovac et al., 2016; Jones-Baumgardt et al., 2019; Samuolienė et al., 2013) and other seedlings (Kitaya et al., 1998; Potter et al., 1999). Concurrent with increasing %DW were decreasing SLA and increasing HD (in arugula and mustard), indicating increasing thickness of aboveground tissues (i.e., stems and cotyledons). The negative linear relationships between LI and SLA in kale, arugula, and mustard were similar to the LI effects on SLA in many other commodities and growing environments (Fan et al., 2013; Gent, 2014; Sims and Pearcy, 1994; Tani et al., 2014; Torres and Lopez, 2011). Short shelf life is one of the major limiting factors for commercial microgreen production (Chandra et al., 2012). For example, Clarkson et al. (2003) found that salt-stressed baby spinach and lettuce, which has lower SLA (i.e., thicker leaves), scored 50% higher on the organoleptic scale than control treatment (no salt stress) 7 d after harvest. More research is needed to evaluate how these morphological changes may affect palatability of marketable tissues, but it is very likely that they are concomitant with potential reductions in tissue damage during harvest and processing and general increases in post-harvest shelf life (Cantwell and Suslow, 2002; Hodges and Toivonen, 2008).
With the much longer hypocotyls observed in the present study vs. microgreens grown in SS, such as in Gerovac et al. (2016), Jones-Baumgardt et al. (2019), and Samuolienė et al. (2013), a more holistic assessment of plant robustness may be appropriate to appraise how LI affects the postharvest quality of harvested tissues. RI [adapted from Fan et al. (2013)] which relates individual components of hypocotyl volume (i.e., HL and HD) to DW, may serve as a proxy for plant robustness. To contextualize the term, for a given DW, increases in HL and decreases in HD both confer reductions in robustness (i.e., tall and/or spindly plants). Despite there being no treatment effects on HD in kale and HL in kale and arugula in the present study, increasing LI still increased RI in all three Brassicaceae genotypes. Future research may be able to more directly evaluate how RI relates to aspects of postharvest quality, such as shelf life and palatability.
Chlorophyll.
CC is closely associated with human perception of the green pigmentation of leaves (Barrett et al., 2010) and is also an important indicator of how plants respond to increasing LI. Both destructive and nondestructive methods are used to measure CC (Kalaji et al., 2016). Destructive measures usually report CC in terms of per weight (FW or DW) or per LA bases (Richardson et al., 2002), depending on the study’s objectives. However, these different bases may result in quite different interpretations of the various studies’ results, depending on leaf morphology. To illustrate this, Hernández and Kubota (2014) found that CC of cucumber seedlings increased, on a per LA basis, under SL treatments (vs. no SL) at two natural DLIs. Conversely, Makus and Lester (2002) reported a 60% decrease in CC, an a per DW basis, in two greenhouse-grown cultivars of mustard under ambient light vs. 50% shade. Additionally, Kopsell et al. (2012) reported substantially lower CC, on a per FW basis, in mustard microgreens grown under 463 vs. 275 μmol·m−2·s−1 of SS fluorescent light. However, because leaf morphology characteristics (e.g., SLA) are unknown, direct comparisons of their CC results are difficult. Furthermore, RCC indices are frequently presented as an alternative to destructive CC measurements due to ease of measurement (Kalaji et al., 2016; Richardson et al., 2002). However, RCC indices also do not account for varying leaf morphology (e.g., SLA) between treatments. In the present study, when considering the magnitude of the RCC treatment effects within the context of decreasing SLA with increasing LI, it appeared that CC on a per DW basis decreased by 12% to 25% in the Brassicaceae genotypes as TLIs increased 86 to 254 mol·m−2. Tani et al. (2014) reported both increases in SLA (i.e., decreases in specific leaf mass) and decreases in RCC of lettuce grown in a greenhouse under solar panels (i.e., shading) vs. ambient lighting, regardless of time of year. At higher LI, it is conceivable that the intensity of green hues (not measured in the present study) may have increased on a per LA basis, but the total chlorophyll content decreased on a per DW basis, such as was reported by Jones-Baumgardt et al. (2019).
While the present study was performed during early February, when natural light levels were quite low, the TLI models presented herein may also be relevant to other times of the year and production environments. For example, in Southern Ontario, Canada, the typical “supplemental lighting season” extends from early October through March, when the natural lighting levels can vary by a factor of 4 (i.e., from 5 to 20 mol·m−2·d−1) (Faust and Logan, 2018). With an understanding of the ambient lighting dynamics within their production facility, a grower can use the presented TLI models to predict their marketable yield for a given SL intensity and, therefore, determine other aspects of their production protocols (e.g., the growing area, length of production, temperature, CO2 concentration, etc.) required to achieve their production goals. Furthermore, although saturation was not observed during the trial, it is conceivable that TLIs in excess of ≈300 mol·m−2 may be excessive for producing microgreens. Locally, this TLI level is achievable (assuming 10-d production cycle and 75% light transmission) with natural light alone during the summer months (i.e., May through August) when the natural DLI may be >40 mol·m−2·d−1 (Faust and Logan, 2018). Furthermore, shading (e.g., retractable curtains or whitewash) may be required to reduce canopy-level LI (also greenhouse temperatures) during periods of excessive LI to reduce photoinhibition.
In most light-limiting greenhouse crop production scenarios (e.g., winter production), SL will increase yield and quality. How much SL to provide is largely related to the commodities’ LUE. Although the crops in the present study all showed LI treatment effects on yield, LUE are quite low relative to many other commodities due to relatively high seeding densities, short production times, and the heterotrophic/autotrophic paradox. While this may be a compelling argument against using high intensities of SL for microgreens, their high value and year-round demand may offset the relatively low LUEs, compared with other crops, particularly for smaller-seeded crops.
Conclusion
This study investigated the responses of four microgreen genotypes to various levels of supplemental LED lighting. The relationships between LI (expressed both in terms of average DLIs and TLIs) and key morphological parameters are presented and discussed. Commercial greenhouse growers can use the light response models described herein to make predictions of relevant microgreen production metrics according to the available light levels during winter in high-latitude regions; enabling growers to optimize the choice of the most appropriate SL levels to achieve their desired production goals, as economically as possible.
Literature Cited
Bantis, P., Smirnakou, S., Ouzounis, T., Koukounaras, A., Ntagkas, N. & Radoglou, K. 2018 Current status and recent achievements in the field of horticulture with the use of light-emitting diodes (LEDs) Scientia Hort. 235 437 451
Barrett, D.M., Beaulieu, J.C. & Shewfelt, R. 2010 Color, flavor, texture and nutritional quality of fresh-cut fruits and vegetables: Desirable levels, instrumental and sensory measurement, and the effects of processing Crit. Rev. Food Sci. Nutr. 50 369 389
Berba, K.J. & Uchanski, M.E. 2012 Postharvest physiology of microgreens J. Young Investig. 24 1 5
Berry, J. & Bjorkman, O. 1980 Photosynthetic response and adaptation to temperature in higher plants Annu. Rev. Plant Physiol. 31 491 543
Boivin, C., Gosselin, A. & Trudel, M.J. 1987 Effect of supplementary lighting on transplant growth and yield of greenhouse tomato HortScience 22 1266 1268
Bredmose, N. 1993 Effects of year-round supplementary lighting on shoot development, flowering and quality of two glasshouse rose cultivars Scientia Hort. 54 69 85
Bredmose, N. 1994 Biological efficiency of supplementary lighting on cut roses the year round Scientia Hort. 59 75 82
Bugbee, B. & Monje, O. 1992 The limits of crop productivity Bioscience 42 494 502
Cantwell, M. & Suslow, T. 2002 Postharvest handling systems: Minimally processed fruits and vegetables, p. 445–463. In: A.A. Kader (ed.). Postharvest technology of horticultural crops. University of California, Division of Agriculture and Natural Resources Publication 3311
Chandra, D., Kim, J.G. & Kim, Y.P. 2012 Changes in microbial population and quality of microgreens treated with different sanitizers and packaging films Hort. Environ. Biotechnol. 53 32 40
Chang, C.L., Hong, G.F. & Li, Y.L. 2014 A supplementary lighting and regulatory scheme using a multi-wavelength light emitting diode module for greenhouse application Light. Res. Technol. 46 548 566
Chun, C., Tominaga, M. & Kozai, T. 2001 Floral development and bolting of spinach as affected by photoperiod and integrated photosynthetic photon flux during transplant production HortScience 36 889 892
Clarkson, G.J.J., O’Byrne, E.E., Rothwell, S.D. & Taylor, G. 2003 Identifying traits to improve postharvest processability in baby leaf salad Postharvest Biol. Technol. 30 287 298
Craver, J.K., Boldt, J.K. & Lopez, R.G. 2018 Radiation intensity and quality from sole-source light-emitting diodes affect seedling quality and subsequent flowering of long-day bedding plant species HortScience 53 1407 1415
Craver, J.K., Boldt, J.K. & Lopez, R.G. 2019 Comparison of supplemental lighting provided by high-pressure sodium lamps or light-emitting diodes for the propagation and finishing of bedding plants in a commercial greenhouse HortScience 54 52 59
Currey, C.J. & Lopez, R.G. 2013 Cuttings of impatiens, pelargonium, and petunia propagated under light-emitting diodes and high-pressure sodium lamps have comparable growth, morphology, gas exchange, and post-transplant performance HortScience 48 428 434
Dansereau, B., Zhang, Y., Gagnon, S. & Xu, H.-L. 1998 Stock and snapdragon as influenced by greenhouse covering materials and supplemental light HortScience 33 668 671
Delgado, E., Parry, M.A.J., Lawlor, D.W., Keys, A.J. & Medrano, H. 1993 Photosynthesis, ribulose-l,5-bisphosphate carboxylase and leaf characteristics of Nicotiana tabacum L. genotypes selected by survival at low CO2 concentrations J. Expt. Bot. 44 1 7
Dorais, M. & Gosselin, A. 2002 Physiological response of greenhouse vegetable crops to supplemental lighting Acta Hort. 580 59 67
Dorais, M., Mitchell, C.A. & Kubota, C. 2017 Lighting greenhouse fruiting vegetables, p. 159–169. In: R. Lopez and E.S. Runkle (eds.). Light management in controlled environments. Meister Media, Middletown, DE
Erwin, J. & Gesick, E. 2017 Photosynthetic responses of swiss chard, kale, and spinach cultivars to irradiance and carbon dioxide concentration HortScience 52 706 712
Fan, X.-X., Xu, Z.-G., Liu, X.-Y., Tang, C.-M., Wang, L.-W. & Han, X.-L. 2013 Effects of light intensity on the growth and leaf development of young tomato plants grown under a combination of red and blue light Scientia Hort. 153 50 55
Faust, J.E. & Logan, J. 2018 Daily light integral: A research review and high-resolution maps of the United States HortScience 53 1250 1257
Fiorucci, A.-S. & Fankhauser, C. 2017 Plant strategies for enhancing access to sunlight Curr. Biol. 27 931 940
Fu, W., Li, P. & Wu, Y. 2012a Effects of different light intensities on chlorophyll fluorescence characteristics and yield in lettuce Scientia Hort. 35 45 51
Fu, W., Li, P., Wu, Y. & Tang, J. 2012b Effects of different light intensities on anti-oxidative enzyme activity, quality and biomass in lettuce HortScience 39 129 134
Gaudreau, L., Charbonneau, J., Vézina, L.P. & Gosselin, A. 1994 Photoperiod and photosynthetic photon flux influence growth and quality of greenhouse-grown lettuce HortScience 29 1285 1289
Gent, P.N. 2014 Effect of daily light integral on composition of hydroponic lettuce HortScience 49 173 179
Gerovac, J.R., Craver, J.K., Boldt, J.K. & Lopez, R.G. 2016 Light intensity and quality from sole-source light-emitting diodes impact growth, morphology, and nutrient content of Brassica microgreens HortScience 51 497 503
Giacomelli, G.A., Ting, K.C. & Panigrahi, S. 1988 Solar PAR vs. solar total radiation transmission in a greenhouse Trans. ASAE 31 1540 1543
Halliday, K. & Davis, S. 2016 Light-sensing phytochromes feel the heat Science 354 832 833
He, W., Huang, Z.-W., Li, J.-P., Su, W.-X., Gan, L. & Xu, Z.-G. 2019 Effect of different light intensities on the photosynthate distribution in cherry tomato seedlings J. Hort. Sci. Biotechnol. 94 611 619
Heggie, L. & Halliday, K. 2005 The highs and lows of plant life: Temperature and light interactions in development Intl. J. Dev. Biol. 49 675 687
Hemming, S. 2011 Use of natural and artificial light in horticulture-interaction of plant and technology Acta Hort. 907 25 35
Hernández, R. & Kubota, C. 2014 Growth and morphological response of cucumber seedlings to supplemental red and blue photon flux ratios under varied solar daily light integrals Scientia Hort. 173 92 99
Hernández, R. & Kubota, C. 2015 Physiological, morphological, and energy-use efficiency comparisons of LED and HPS supplemental lighting for cucumber transplant production HortScience 50 351 357
Hodges, D.M. & Toivonen, P.M.A. 2008 Quality of fresh-cut fruits and vegetables as affected by exposure to abiotic stress Postharvest Biol. Technol. 48 155 162
Hoshi, T., Higa, H., Goto, K. & Niibori, K. 2011 Effects of supplemental lighting on the quality of tomato seedlings raised in greenhouses Acta Hort. 907 117 123
Jones-Baumgardt, C., Llewellyn, D., Ying, Q. & Zheng, Y. 2019 Intensity of sole-source light-emitting diodes affects growth, yield, and quality of Brassicaceae microgreens HortScience 54 1168 1174
Kalaji, H., Dąbrowski, P., Cetner, M., Samborska, I., Łukasik, I., Brestic, M. & Zivcak, M. 2016 A comparison between different chlorophyll content meters under nutrients deficiency conditions J. Plant Nutr. 40 1024 1034
Kitaya, Y., Niu, G., Kozai, T. & Ohashi, M. 1998 Photosynthetic photon flux, photoperiod, and CO2 concentration affect growth and morphology of lettuce plug transplants HortScience 33 988 991
Kopsell, D., Pantanizopoulos, N., Sams, C. & Kopsell, D.E. 2012 Shoot tissue pigment levels increase in ‘Florida Broadleaf’ mustard (Brassica juncea L.) microgreens following high light treatment Scientia Hort. 140 96 99
Kou, L.P., Luo, Y.G., Yang, T.B., Xiao, Z.L., Turner, E.R., Lester, G.E., Wang, Q. & Camp, M.J. 2013 Postharvest biology, quality and shelf life of buckwheat microgreens Food Sci. Technol. 51 73 78
Kreyling, J., Schweiger, A.H., Bahn, M., Ineson, P., Migliavacca, M., Morel-Journel, T., Christiansen, J.R., Schtickzelle, N. & Larsen, K.S. 2018 To replicate or not to replicate – that is the question: How to tackle nonlinear responses in ecological experiments Ecol. Lett. 21 1629 1638
Kyriacou, M., Rouphael, Y., Di Gioia, F., Kyratzis, A., Serio, F., Renna, M., De Pascale, S. & Santamaria, P. 2016 Micro-scale vegetable production and the rise of microgreens Trends Food Sci. Technol. 57 103 115
Lau, A.K. & Stanley, L.M. 1989 Solar radiation transmission and capture in greenhouses Can. Agr. Eng. 31 205 214
Lefsrud, M.G., Kopsell, D.A., Augé, R.M. & Both, A.J. 2006 Biomass production and pigment accumulation in kale grown under increasing photoperiods HortScience 41 603 606
Llewellyn, D. & Zheng, Y. 2018a Finetuning LEDs for a better light: How light spectrum makes a difference. 15 June 2018. <https://www.greenhousecanada.com/structures-equipment/lighting/finetuning-leds-for-a-better-light-how-light-spectrum-makes-a-difference-32450>
Llewellyn, D. & Zheng, Y. 2018b Using feedback control strategies to more efficiently manage LEDs. 15 May 2018. <https://www.greenhousecanada.com/structures-equipment/lighting/using-feedback-control-strategies-to-more-efficiently-manage-leds-32408>
Llewellyn, D., Schiestel, K. & Zheng, Y. 2019 Light-emitting diodes can replace high-pressure sodium lighting for cut gerbera production HortScience 54 95 99
Llewellyn, D., Zheng, Y. & Dixon, M. 2013 Survey of how hanging baskets influence the light environment at lower crop level in ornamental greenhouses in Ontario, Canada HortTechnology 23 823 829
Lopez, R.G. & Runkle, E.S. 2008 Photosynthetic daily light integral during propagation influences rooting and growth of cuttings and subsequent development of New Guinea impatiens and petunia HortScience 43 2052 2059
Makus, D.J. & Lester, G. 2002 Effect of soil type, light intensity, and cultivar on leaf nutrients in mustard greens Yaredai Zhiwu Kexue 54 23 28
Martineau, V., Lefsrud, M.G., Tahera-Naznin, M. & Kopsell, D.A. 2012 Comparison of supplemental greenhouse lighting from light emitting diode and high pressure sodium light treatments for hydroponic growth of Boston lettuce HortScience 47 477 482
Meng, Q. & Runkle, E.S. 2019 Far-red radiation interacts with relative and absolute blue and red photon flux densities to regulate growth, morphology, and pigmentation of lettuce and basil seedling Scientia Hort. 255 269 280
Mitchell, C.A., Dzakovich, M.P., Gomez, C., Lopez, R., Burr, J.F., Hernández, R., Kubota, C., Currey, C.J., Meng, Q., Runkle, E.S., Bourget, C.M., Morrow, R.C. & Both, A.J. 2015 Light-emitting diodes in horticulture, p. 1–87. In: J. Janick (ed.). Horticultural Reviews Vol. 43. Wiley, New York
Morrow, R.C. 2008 LED lighting in horticulture HortScience 43 1947 1950
Murphy, C.J. & Pill, W.C. 2010 Cultural practices to speed the growth of microgreen arugula (roquette; Erica vesicaria subsp. Sativa) J. Hort. Sci. Biotechnol. 85 171 176
Nelson, J.A. & Bugbee, B. 2014 Economic analysis of greenhouse lighting: Light emitting diodes vs. high intensity discharge fixtures PLoS One 9 6 e99010
Oda, M., Aoki, S., Nagaka, M. & Tsuji, K. 1989 Nutrient solution culture of lettuce under artificial light. II. Growth promotion induced by continuous illumination with low light intensity Environ. Control Biol. 27 75 82
Oh, W., Runkle, E.S. & Warner, R.M. 2010 Timing and duration of supplemental lighting during the seedling stage influence quality and flowering in petunia and pansy HortScience 45 1332 1337
Ohyama, K., Manabe, K., Omura, Y., Kozai, T. & Kubota, C. 2005 Potential use of a 24-Hour photoperiod (continuous light) with alternating air temperature for production of tomato plug transplants in a closed system HortScience 40 374 377
Ouzounis, T., Rosenqvist, E. & Ottosen, C.O. 2015 Spectral effects of artificial light on plant physiology and secondary metabolism: A review HortScience 50 1128 1135
Park, Y. & Runkle, E.S. 2018 Far-red radiation and photosynthetic photon flux density independently regulate seedling growth but interactively regulate flowering Environ. Exp. Bot. 155 206 216
Pattison, P.M., Tsao, J.Y. & Krames, M.R. 2016 Light-emitting diode technology status and directions: Opportunities for horticultural lighting Acta Hort. 1134 413 426
Poel, B.R. & Runkle, E.S. 2017 Seedling growth is similar under supplemental greenhouse lighting from high-pressure sodium lamps or light-emitting diodes HortScience 52 388 394
Poorter, H., Niinemets, U., Ntagkas, N., Siebenkäs, A., Mäenpää, M., Matsubara, S. & Pons, T.L. 2019 A meta-analysis of plant responses to light intensity for 70 traits ranging from molecules to whole plant performance New Phytol. 223 1073 1105
Potter, T.I., Rood, S.B. & Zanewich, K.P. 1999 Light intensity, gibberellin content and the resolution of shoot growth in Brassica Planta 207 505 511
Pramuk, L.A. & Runkle, E.S. 2005 Photosynthetic daily light integral during the seedling stage influences subsequent growth and flowering of Celosia, Impatiens, Salvia, Tagetes, and Viola HortScience 40 1336 1339
Randall, W.C. & Lopez, R.G. 2014 Comparison of supplemental lighting from high-pressure sodium lamps and light-emitting diodes during bedding plant seedling production HortScience 49 589 595
Rezai, S., Etemadi, N., Nikbakht, A., Yousefi, M. & Majidi, M.M. 2018 Effect of light intensity on leaf morphology, photosynthetic capacity, and chlorophyll content in sage (Salvia officinalis L.) Korean J. Hort. Sci. Technol. 36 46 57
Richardson, A.P., Duigan, S. & Berlyn, G. 2002 An evaluation of noninvasive methods to estimate foliar chlorophyll content New Phytol. 153 185 194
Roig-Villanova, I. & Martínez-García, J. 2016 Plant responses to vegetation proximity: A whole life avoiding shade Front. Plant Sci. 7 236 246
Sago, Y. 2016 Effects of light intensity and growth rate on tipburn development and leaf calcium concentration in butterhead lettuce HortScience 51 1087 1091
Samuolienė, G., Brazaitytė, A., Jankauskienė, J., Viršilė, A., Sirtautas, R., Novičkovas, A., Sakalauskienė, S., Sakalauskaitė, J. & Duchovskis, P. 2013 LED irradiance level affects growth and nutritional quality of Brassica microgreens Central European J. Biol. 8 1241 1249
Sims, D.A. & Pearcy, R.W. 1994 Scaling sun and shade photosynthetic acclimation of Alocasia macrorrhiza (Araceae) to a transfer from low to high light Amer. J. Bot. 79 449 455
Soustani, B.F., Jalali, S.G., Sohrabi, H. & Shirvany, A. 2014 Growth responses to irradiance regime along an ecological gradient of Quercus castaneifolia seedlings of different provenance Ecol. Res. 29 245 255
Sublett, W.L., Barickman, T.C. & Sams, C.E. 2018 The effect of environment and nutrients on hydroponic lettuce yield, quality, and phytonutrients Horticulturae 4 48 63
Sysoeva, M., Markovskaya, E. & Shibaeva, T. 2010 Plants under continuous light Review Plant Stress 4 5 17
Tani, A., Shiina, S., Nakashima, K. & Hayashi, M. 2014 Improvement in lettuce growth by light diffusion under solar panels J. Agr. Meteorol. 70 139 149
Ting, K.C. & Giacomelli, G.A. 1987 Solar photosynthetically active radiation transmission through greenhouse glazings Energ. Agr. 6 121 132
Torres, A.P. & Lopez, R.G. 2011 Photosynthetic daily light integral during propagation of Tecoma stans influences seedling rooting and growth HortScience 46 282 286
Treadwell, D.D., Hochmuth, R., Landrum, L. & Laughlin, W. 2016 Microgreens: A new specialty crop. Univ. Florida IFAS Ext. Bul. HS1164
Viršilė, A. & Sirtautas, R. 2013 Light irradiance level for optimal growth and nutrient contents in borage microgreens Rural Dev. 6 272 275
van Iersel, M. 2017 Light, photosynthesis, and plant growth, p. 21–28. In: R. Lopez and E.S. Runkle (eds.). Light management in controlled environments. Meister Media, Middletown, DE
Xiao, Z., Lester, G., Luo, Y. & Wang, Q. 2012 Assessment of vitamin and carotenoid concentrations of emerging food products: Edible microgreens J. Agr. Food Chem. 60 7644 7651
Yao, X.-Y., Liu, X., Xu, Z.-G. & Jiao, X.-L. 2017 Effects of light intensity on leaf microstructure and growth of rape seedlings cultivated under a combination of red and blue LEDs J. Integr. Agr. 16 97 105
Yeh, D.M. & Hsu, P.Y. 2004 Differential growth and photosynthetic response of selected cultivars of English ivy to irradiance J. Hort. Sci. Biotechnol. 79 633 637
Zervoudakis, G., Salahas, G., Kaspiris, G. & Konstantopoulou, E. 2012 Influence of light intensity on growth and physiological characteristics of common sage (Salvia officinalis L.) Braz. Arch. Biol. Technol. 55 89 95
Zhang, M., Catherine, C.M. & Runkle, E.S. 2019 Manipulating growth, color, and taste attributes of fresh cut lettuce by greenhouse supplemental lighting Scientia Hort. 252 274 282