Plant growth conditions

Cultivation and growth conditions for Chrysanthemum morifolium Ramat. (“Milton Dark Pink”, “Williamsburg Purple”, and “Olympia White”) have been extensively described elsewhere (Shelp et al. 2017, 2020; Sutton et al. 2019). Briefly, the unrooted cuttings were dipped in a mixture of Stim-Root liquid (0.1% indole butyric acid) and B9 (0.25% daminozide) (Plant Products, Brampton, ON, Canada), and stored overnight in a 4 °C cooler. Each cutting was then inserted into a peat Jiffy Plug amended with 30% minerals (Model CF Hort. Plug 343040-26, Jiffy Products (N.B.) Ltd., Shippagan, NB, Canada), and maintained in the vegetative state on a misting bench in a naturally lit greenhouse. From 12 to 21 days, the misting tent was removed and the cuttings were supplied every 2 days with a fertilizer solution containing 17.8 mmol L⁻¹ N.

Rooted cuttings were dipped in a 1% solution of Safer’s Insecticidal Soap (Woodstream Canada Corporation, Brampton, ON, Canada) and individually potted in 10 cm diameter pots. The commercial soil medium was a mixture of peat moss and perlite (50:50 by volume) without a preplant fertilizer charge; the initial pH was adjusted to 5.70–6.15 with ground limestone (BM 6, Berger, Boisbriand, QC, Canada). Saturated medium extracts contained 2.9–3.4 µmol L⁻¹ Fe and <0.55 µmol L⁻¹ Mn. The potted soil was saturated with deionized water, then the pots were placed into 16 troughs in a computer-controlled ebb-and-flow subirrigation system housed in a naturally lit greenhouse with a set day/night temperature of 25 °C and humidity of 50%. Two experiments were conducted for both Mn and Fe (Mn Expt. 1: July–October 2020; Mn Expt. 2: January–March 2021; Fe Expt. 1: October–December 2020; Fe Expt. 2: February–April 2021). Greenhouse equipment failure forced the Fe Expt. 1 to be restarted when only one cultivar of chrysanthemum was available.

Plants were exposed to 1 wk of long days (12 h light:12 h dark cycle) and vegetative growth was maintained by implementing a night break with supplemental lighting (Solar Max Spectrum LED, BML Horticulture, Austin, TX; 20–50 µmol m⁻² s⁻¹ photosynthetic photon flux density at pot level) from 0030 to 0230 h. Then, the plants were pinched to equal heights to promote branching and switched to short days (10 h light:14 h dark) to induce flowering. Blackout curtains were used throughout the experiment to maintain the light:dark cycle; to compensate for any shading from the blackout curtain infrastructure, the supplemental lighting was also used under the low ambient light conditions present in the winter months. Plants were fertigated as needed every 1–3 days with the nutrient solutions described below until bud emergence, when the cultivars had a fully formed set of buds (5–6 wk). Subirrigation pumps were activated at 1100 h for 5 min to create a 2–3 cm deep irrigation head, which was allowed to drain back into the adjacent 40 L tanks for recycling. Nutrient solution was replenished as needed, every 1–2 wk. Post bud emergence, all plants were supplied with deionized water only in the same manner. Greenhouse pests were managed using an integrated pest management system (Swirskii-System, Biobest Group, Westerlo, Belgium), in combination with a Beleaf 50SG (ISK Biosciences Corporation, Concord, OH, USA) spray, once per crop cycle.

Experimental design

Each experiment was conducted on side-by-side benches with plants arranged in a split-plot randomized complete block design with 4 blocks and 10 plants per treatment. Each treatment appeared once per block. The outermost plants on each bench served as border plants and were not included in the analysis. They received a modified Sonneveld solution (7.1 mmol L⁻¹ N, 1.0 mmol L⁻¹ P, 3.5 mmol L⁻¹ K, 0.58 mmol L⁻¹ Mg, 2.6 mmol L⁻¹ Ca, 0.87 mmol L⁻¹ S, 3.5 µmol L⁻¹ Zn, 10.56 µmol L⁻¹ Fe, 5.0 µmol L⁻¹ Mn, 0.75 µmol L⁻¹ Cu, 5 µmol L⁻¹ B, and 0.5 µmol L⁻¹ Mo) (Sonneveld and Kreij 1987), prior to flower bud break (stage 2 of bloom development; Duncan Stephens et al. 2021), and then deionized water from bud break to harvest. The nutrient regimen was identical in the treatment rows, with the exceptions that Mn was replaced by 5, 2.5, and 1.25 (Mn Expt. 1) or 1.25, 0.625, and 0.3125 µmol L⁻¹ Mn (Mn Expt.2), and Fe was replaced by 10.56, 5.28, and 2.64 (Fe Expt. 1) or 2.64, 1.32, and 0.66 µmol L⁻¹ Fe (Fe Expt. 2). The concentration of Mn or Fe supplied during vegetative growth was considered as the main plot, while cultivar (“Milton Dark Pink” or “Williamsburg Purple”) was the subplot (with the exception of Fe Expt. 1, which used only one cultivar, “Olympia White”).

Beginning at bud emergence, bloom development was assessed weekly by counting the buds and rating their maturity on a scale from 1 to 6 (Duncan Stephens et al. 2021;  Fig. S1 (cjps-2021-0286_suppla.docx)). Plants were harvested at the onset of flower senescence, and shoot height, shoot fresh mass (FM), and flower FM were determined, along with a final assessment of bloom development (i.e., bud/infloresence count and stage, bloom diameter). Shoots and inflorescences were dried in a 95 °C oven for at least 3 days and dry mass (DM) of each was determined. The plants were examined for visual symptoms of nutrient deficiency throughout the experiment. In each experiment (with the exception of Mn Expt. 1), plant greenness was assessed as an estimate of chlorophyll content (Radhamani et al. 2016) in three young leaves of each plant (every other plant in Fe Expt. 1) using a SPAD 502DL Plus Chlorophyll Meter (Spectrum Technologies, Inc., Aurora, IL, USA) at bud emergence and immediately prior to harvest.

Leaf nutrient analysis

On the day the nutrient supply was removed, three recently matured, fully expanded leaves were collected from each plant and pooled according to treatment replicate (Hill Laboratories 2019). After drying in a 95 °C oven for at least 3 days, samples were individually ground into a fine powder using a Waring 7011 G blender (Waring Commercial, Torrington, CO, USA). Samples were then sent to the Agriculture and Food Laboratory at the University of Guelph for nutrient analysis. The samples were microwave acid digested, and diluted to an appropriate volume with nanopure water before measurement using an inductively coupled plasma–optical emission spectrometry method developed and validated in-house (based on USEPA Method 200.7 revision 4.4; United States Environmental Protection Agency).

Statistical analysis

All statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., NC, USA) with the PROC GLIMMIX method (α = 0.05). Normality and homogeneity of variance were confirmed before further statistical analyses were performed. Initially, cultivars were analysed together to compare responses to the main effect, and then individually for comparison of the main effect. Repeated measures analysis was used to compare data across time (i.e., inflorescence development), using a compound symmetry covariance structure. Variance was separated into fixed effects (Mn or Fe treatment and cultivar), random effects (block), and all relevant interactions within and between the fixed and random effects. Analyses of variance (ANOVAs) were performed and when effects were significant (P ≤ 0.05), the means were compared to each other using Tukey’s honest significant difference test using the slice function.

Growth of two chrysanthemum cultivars supplied with moderate levels of Mn

Morphological characteristics

In general, the shoot height, shoot DM, inflorescence DM, bloom diameter, and inflorescence/bud number at final harvest, and greenness of young leaves at bud emergence (Table 1) and final harvest ( Table S4 (cjps-2021-0286_suppla.docx)) were unaffected by the Mn treatment, regardless of cultivar. However, some differences occurred (Table 1). In Expt. 2, bloom diameter and inflorescence development were slightly affected in “Milton Dark Pink” and “Williamsburg Purple”, respectively, though the response was not linearly correlated with the Mn supply.

Table 1.

Morphological characteristics of two chrysanthemum cultivars supplied with varying levels of Mn up to bud emergence in Summer/Fall 2020 (Expt. 1) and Winter/Spring 2021 (Expt. 2).

Bud and inflorescence development

In both experiments, the bud/inflorescence number in each cultivar was similar across the treatments from bud emergence until final harvest, though occasionally some values were different ( Table S5 (cjps-2021-0286_suppla.docx)). The stage of bud/inflorescence development was determined weekly from bud emergence (stage 1) to harvest. Bud/inflorescence development was generally uniform across all treatments, regardless of cultivar (Fig. 1,  Table S6 (cjps-2021-0286_suppla.docx)); indeed, there were no differences among the treatments at final harvest (Table 1).

Fig. 1.

Inflorescence development of two chrysanthemum cultivars supplied with varying levels of Mn up to bud emergence. The stage of inflorescence development was determined weekly from bud emergence (stage 1) to harvest. Statistical treatment of the data are shown in  Table S6 (cjps-2021-0286_suppla.docx).

Leaf nutrient composition at bud emergence

Regardless of the experiment and cultivar, leaf tissue levels of the macronutrients, including Ca, and the micronutrients, including Fe, Zn, and Cu, were generally unaffected by the 16-fold range in Mn supply (Table 2,  Tables S7 and S8 (cjps-2021-0286_suppla.docx)). The only macronutrient exception was the slightly elevated S level in “Milton Dark Pink” with the lowest Mn supply in Expt. 2. The only micronutrient exception was the 30% decrease in Mn in “Williamsburg Purple” with the reduction in Mn supply from 5.00 to 2.5 µmol L⁻¹ in Expt. 1. Overall, the leaf Mn levels ranged from 51 to 116 mg kg⁻¹ DM across all cultivars and experiments.

Table 2.

Composition of select nutrients from diagnostic leaves of two chrysanthemum cultivars supplied with varying levels of Mn until flower bud emergence in Summer/Fall 2020 (Expt. 1) and Winter/Spring 2021 (Expt. 2).

Growth of two chrysanthemum cultivars supplied with moderate levels of Fe

Morphological characteristics at final harvest

In both experiments, the treatment did not affect the morphological characteristics (shoot height, shoot DM, inflorescence DM, bloom diameter, inflorescence development and number of inflorescences/buds at final harvest, and leaf greenness at bud emergence and final harvest), regardless of cultivar (Table 3,  Table S4 (cjps-2021-0286_suppla.docx)).

Table 3.

Morphological characteristics of three chrysanthemum cultivars supplied with varying levels of Fe up to bud emergence in Summer/Fall 2020 (Expt. 1) and Winter/Spring 2021 (Expt. 2).

Bud and inflorescence development

In Expt. 1 the bud/inflorescence number of “Olympia White” remained constant from bud emergence to harvest across all treatments. In contrast, in Expt. 2 the number increased in both “Milton Dark Pink” and “Williamsburg Purple” from emergence to harvest, though there was no treatment effect ( Table S12 (cjps-2021-0286_suppla.docx)). Nevertheless, the stage of inflorescence development increased with time in the three cultivars across both experiments, though there were no treatment effects (Fig. 2, Fig. S13).

Fig. 2.

Inflorescence development of three chrysanthemum cultivars supplied with varying levels of Fe up to bud emergence. The stage of inflorescence development was determined weekly from bud break (stage 1) to harvest. Statistical treatment of the data are shown in  Table S13 (cjps-2021-0286_suppla.docx).

Leaf nutrient composition at bud emergence

Regardless of the experiment and cultivar, the leaf macronutrient levels at bud emergence were generally unaffected by the 16-fold range in Fe supply (Table 4,  Table S14 (cjps-2021-0286_suppla.docx)). The only effects of treatment were on Mg and Ca levels in “Milton Dark Pink” in Expt. 2, which increased slightly with the reduction in Fe supply from 2.64 to 0.66 µmol L⁻¹. Leaf Fe was not different among the treatments in either experiment (Table 4,  Table S15 (cjps-2021-0286_suppla.docx)). Overall, the Fe concentrations ranged from 68.5 to 103.0 mg kg⁻¹ DM across all cultivars and treatments. Differences were evident in the concentration of some of the other micronutrients. The leaf Mn level was negatively correlated with the Fe supply in both experiments, regardless of cultivar, increasing approximately 16%–31% with the reduction in Fe. In contrast, the leaf Cu level was correlated with the Fe supply in both “Olympia White” and “Milton Dark Pink”, respectively, in experiments 1 and 2, decreasing by approximately 11%–16% with the reduction in Fe. The leaf Mo level was also affected by the Fe supply in Expt. 2; however, the response was different between the two cultivars. In “Milton Dark Pink,” the lowest Mo level was found with 0.66 µmol L⁻¹ Fe supply, whereas in “Williamsburg Purple”, the highest level was found with 0.66 µmol L⁻¹ Fe supply.

Table 4.

Composition of select nutrients from diagnostic leaves of two chrysanthemum cultivars supplied with varying levels of Fe until flower bud emergence in Summer/Fall 2020 (Expt. 1) and Winter/Spring 2021 (Expt. 2).

Strategic timing and rates of Mn and Fe fertilization improve use efficiency in subirrigated, greenhouse-grown pot chrysanthemums

The common practice in commercial greenhouses is to supply nutrients in excess to avoid deficiency symptoms which, if present, would render a plant unsaleable. These levels are higher than necessary for maximal growth, but not high enough to be toxic. Fertilizer recommendations are often based on outdated cultivars and irrigation techniques and focus on macronutrients, especially N, P, and K. For example, Green Leaf Plants (2015) recommends a complete fertilizer at 21.4–28.6 mmol L⁻¹ N with added micronutrients for greenhouse pot chrysanthemums but suggests that the rate can be reduced by 25%–50% or eliminated during the final 2–3 weeks of the crop cycle with subirrigation systems. The Sonneveld and Hoagland solutions, commonly used when growing plants in soilless media, contain 5 µmol L⁻¹ Mn and 15–25 µmol L⁻¹ Fe at 10.5–18.5 mmol L⁻¹ N (Hoagland and Arnon 1938; Sonneveld and Kreij 1987).

MacDonald et al. (2014) demonstrated that eliminating nutrient delivery altogether during the reproductive growth of a subirrigated pot chrysanthemum (“Yellow Favor”) does not affect the quality of the plants and flowers. Subsequent studies with many chrysanthemum cultivars (“Olympia White”, “Covington Yellow”, “Milton Dark Pink”, “Williamsburg Purple”, “Kingsville Yellow”, “Milton Orange”, and “Newport Bronze”), typically replicated over two growing seasons, further optimized the delivery of N, P, K, Ca, Mg, S, Zn, and (or) Cu by combining nutrient removal during reproductive growth with reduced nutrient delivery during vegetative growth, resulting in overall decreases of 75%–87.5% in nutrient delivery, without adversely affecting flower quality (Shelp et al. 2017, 2020, 2021a, 2021b; Sutton et al. 2019; Duncan Stephens et al. 2021). While the cultivars being compared often exhibited contrasting phenotypes (e.g., biomass accumulation, bloom diameter, tissue nutrient levels), the NUE always increased with decreasing nutrient delivery, and any treatment effects on morphological characteristics were minor, likely to be unnoticed by consumers.

In the present study, three cultivars of chrysanthemum were supplied with an optimized macronutrient regimen, and control supplies of Zn, Cu, B, and Mo during vegetative growth. While there were cultivar differences in morphology and tissue nutrient levels, the levels of N (4.5%–6.53% DM), P (0.67%–1.02% DM), K (4.77%–6.60% DM), Ca (1.14%–1.76% DM), Mg (0.39%–0.69% DM), Zn (26.3–60.8 mg kg⁻¹ DM), Cu (2.7–9.1 mg kg⁻¹ DM), B (36.5–74.8 mg kg⁻¹ DM), and Mo (1.2–4.0 mg kg⁻¹ DM) across all treatments are in general agreement with sufficiency guidelines for chrysanthemums in the extension literature (4.0%–6.5% DM N, 0.2%–1.2% DM P, 1.0%–10.0% DM K, 0.5%–4.6% DM Ca, 0.1%–1.5% DM Mg, 5–250 mg kg⁻¹ DM Zn, 5–50 mg kg⁻¹ DM Cu, and 20–200 mg kg⁻¹ DM B) (Ontario Ministry of Agriculture, Food and Rural Affairs 2014; Plank et al. 2018; Hill Laboratories 2019). They are also in agreement with estimates of the critical deficiency concentrations in leaves of dicotyledonous plants (i.e., 1–5 mg kg⁻¹ DM Cu and 0.1–1.0 mg kg⁻¹ DM Mo, depending on the species; see Broadley et al. 2012). According to these guidelines, Cu is the only nutrient that may be low. However, common symptoms of Cu deficiency in chrysanthemum are leaf margin desiccation and suppression of flowering (Roorda van Ensinga and Smilde 1980), neither of which was seen. Furthermore, previous research from our laboratory demonstrated that leaf Cu levels as low as 3.7 mg Cu kg⁻¹ DM are not associated with visual signs of deficiency (Shelp et al. 2021a).

Our previous research generally involved the replication of experiments over two growing seasons, and the treatment responses were similar across the seasons, as well as the two culivars under consideration. Therefore, when symptoms of Mn or Fe deficiency were absent from the first experiment, it was decided to further reduce their levels in the second experiment. Overall, the Mn and Fe supplies were reduced by approximately 94% during vegetative growth, compared to the industry standards, and approximately 97% over the crop cycle. There were no signs of uniform yellowing, interveinal chlorosis, delayed flower bud development, smaller flower size with pale colour, or small necrotic spots that coalesce into large desiccating areas (Lunt et al. 1964; Roorda van Eysinga and Smilde 1980). Across all experiments, the Mn levels ranged from 44.8 to 121.8 mg kg⁻¹ and the Fe levels ranged from 68.5 to 121.8 mg kg⁻¹, which are within the ranges considered acceptable for normal plant growth (25–375 mg kg⁻¹ Mn and 20–750 mg kg⁻¹ Fe; Ontario Ministry of Agriculture, Food and Rural Affairs 2014; Planck et al. 2018; Hill Laboratories 2019). Comparison of these findings with the relatively stable morphological results leads to the conclusion that the Mn and Fe use efficiencies were improved approximately 16-fold with decreasing supply, and 32-fold over the crop cycle, in comparison to industry standards, without adversely affecting the plant and flower quality of either cultivar. This was done by removing the entire nutrient suite during the reproductive growth stage and lowering the Mn and Fe supplies during the vegetative growth stage.

Mechanisms for improved Mn and Fe use efficiencies

Increasing NUE as a function of decreasing nutrient supply can result from improvements in ENA and (or) ENU. Nutrient budgets have been used to demonstrate that increased ENA is the primary mechanism to obtain sufficient N, P, and S for the growth of chrysanthemums on decreasing supplies of the respective macronutrient (MacDonald et al. 2014; Shelp et al. 2017, 2020; Sutton et al. 2019). The NRAMP and IRT protein families in dicotyledonous plants have members that are involved in the transport of Mn and Fe from the rhizosphere into root cells. For example, AtNRAMP1 is a plasma membrane-localized, metal/H⁺ symporter that mediates Mn uptake by epidermal and cortical cells in roots (Cailliatte et al. 2010; Castaings et al. 2016). Under Mn deficiency, AtNRAMP1 expression is moderately upregulated, and the atnramp1 mutant grows poorly and accumulates less Mn than the WT, suggesting that this protein is a high-affinity Mn uptake transporter (Cailliatte et al. 2010). Active Mn uptake may also be mediated by the high-affinity Fe transporter AtIRT1 (Korshunova et al. 1999; Castaings et al. 2016). In response to Fe deficiency, the nuclear AtFIT gene is activated, inducing the upregulation of AtFRO2 and AtIRT1 (Connolly et al. 2002; Varotto et al. 2002). Better knowledge of the affinities and kinetics of these Mn and Fe transporters would be useful in assessing their importance with decreasing nutrient delivery, especially when the reduction is insufficient to elicit visible signs of deficiency (Griffiths and York 2020).

The possibility of enhanced ENU with decreasing Mn or Fe supply cannot be excluded. Temporary stores of Mn and (or) Fe in the roots and source leaves could be remobilized by the upregulation of tonoplastic and Golgi network efflux proteins AtZIP1 and AtNRAMP2,3,4 (Thomine et al. 2000; Lanquar et al. 2005, 2010; Milner et al. 2013; Bastow et al. 2018; Gao et al. 2018). Remobilization could also be increased by upregulating the plasma membrane-localized proteins AtYSL1 and AtYSL3, which transport Mn and Fe between translocation streams (Waters et al. 2006). Notably, remobilization would be especially important during the reproductive growth of chrysanthemums as Mn and Fe can be removed at flower bud emergence with no adverse effects on plant and flower quality (this study; Shelp et al. 2017, 2020, 2021a, 2021b; Sutton et al. 2019; Duncan Stephens et al. 2021).

The methods employed here did not allow for a nutrient balance to estimate the relative contribution of the nutrient solution vs. cuttings, pot mix and Jiffy plug; or, to assess the relative importance of ENA and ENU in sustaining the growth of flowering chrysanthemums when the micronutrient supply (Mn, Fe, Zn, Cu) was reduced during vegetative growth and then the entire nutrient supply was removed during reproductive growth (c.f. Shelp et al. 2020 vs. Shelp et al. 2021a). Furthermore, a nutrient balance would be useful for studying remobilization only if the contents of various plant parts change in response to the treatments. To overcome this limitation, isotopes of Mn and Fe could be supplied to the root systems or a leaf (via the leaf flap feeding method) at various times throughout the vegetative and reproductive growth stages to monitor the simultaneous movement of labelled and unlabelled Mn or Fe (Biddulph et al. 1959; Tsukamoto et al. 2006; Valentinuzzi et al. 2020). This would allow the uptake, storage, and translocation of Mn and Fe to be tracked as a function of plant development. Gene expression analysis via quantitative polymerase chain reaction of potential Mn and Fe transporters in both roots and leaves would give greater insight into mechanisms responsible for Mn and Fe uptake and remobilization when the supplies are lowered, albeit not excessively lowered so that visible and morphological symptoms of deficiency become apparent. Other factors such as root architecture and release of root storage pools, could have also contributed to improved NUE, but these possibilities are beyond the scope of this study.

Potential interactions among Mn, Fe, Zn, Cu, and Ca

Transporters such as NRAMP1, IRT1, YSL1, and ECA3 have broad specificity for essential divalent cations (Jeong et al. 2017; Alejandro et al. 2020; He et al. 2021; Rai et al. 2021). This non-specificity and the related antagonism could explain the inverse correlation, observed in some cases here, of the leaf Mn, Zn, and Ca levels with the decreasing Fe supply (Rai et al. 2021). They could also support the existence of crosstalk between interconnected pathways for the regulation of metal homeostasis (Rai et al. 2021). Nonetheless, these findings indicate that the balance of divalent cations in the nutrient solution might be an important consideration when nutrient delivery during the vegetative growth stage approaches the low sufficiency range.

The dilution of commercial fertilizers to achieve optimal macronutrient levels can result in micronutrient levels that are inadequate to sustain plant growth and quality, or exacerbate the adverse effects of an unsuitable micronutrient balance. Interestingly, there is quite a range in the approximate ratios of Zn:Cu (from 2:1 to 10:1), Fe: Mn (from 1:1 to 5:1), and B: Mo (15:1 to 40:1) among common commercial (Fusion Plant Products 15-5-17; Peters Professional 17-3-7 Peat-Lite Neutral Cal-Mag, ICL Fertilizers, Dublin, OH, USA)) and the more historical Sonneveld and Hoagland solutions. To date, the levels of Zn, Cu, Mn, and Fe have all been successfully reduced in a research setting to below the levels obtained in the commercial setting with the dilution of Fusion Plant Products 15-5-7 (this study; Shelp et al. 2021a, 2021b). Thus, it may be possible to further reduce in a research setting the levels of Zn, Cu, Mn, and Fe, and to proceed with the optimization of the two remaining essential micronutrients, B and Mo, which have been shown to accumulate in greenhouse-influenced waterways in southwestern Ontario (Ministry of Environment 2012; Maguire et al. 2018). It may however, become necessary to fine-tune the micronutrient balance. For example, Fe-deficient plants accumulate additional Cu in leaf tissues (Waters et al. 2012), and higher Cu levels in the nutrient solution, relative to Zn, reduce the availability of Zn to plants, and vice versa (Imtiaz et al. 2003). Once completed, the delivery strategy can be validated in a commercial setting using modern cultivars. Together, this research should indicate whether the current commercial fertilizers need to be reformulated to meet the needs of the modified subirrigation strategy for chrysanthemums.

Prospects for the floricultural industry

Subirrigation systems are popular environmentally friendly means of floricultural greenhouse production (MacDonald et al. 2013; Ferrarezi et al. 2015; Semananda et al. 2018). Because fertilizer recommendations are generally based on overhead irrigation systems and outdated cultivars, an opportunity exists to optimize nutrient delivery and potentially lower nutrient inputs. To date, our modified delivery strategy for modern cultivars of chrysanthemum has been optimized with macronutrients in a research setting using a closed subirrigation system (Shelp et al. 2017, 2020; Sutton et al. 2019), and then validated in a commercial setting using an open subirrigation system where the spent nutrient solution is released directly into the municipal sewage system in compliance with the Nutrient Management Act (Shelp et al. 2021b). In both cases, macronutrient delivery to modern chrysanthemum cultivars can be reduced in the subirrigation solution by approximately 75% during vegetative growth, compared to common fertilizer formulations, and then removed entirely during reproductive growth. Therefore, reducing the fertilizer rate results in proportional cost savings (Shelp et al. 2021b), and decreases the risk of environmental contamination.

It would also be of interest to study the timing and optimization of nutrient delivery using drip irrigation, another common environmentally friendly method for irrigating indoor floricultural crops (Lévesque et al. 2009). Unlike subirrigation, drip irrigation is scalable to any size, and reducing nutrient delivery would lower the need for overirrigation. This research would validate the new fertilization rates and the modified delivery strategy with a different irrigation system, enabling our research to reach more growers.

In Canada, there were almost five million indoor potted chrysanthemums produced in 2020 (Statistics Canada 2021). Our chrysanthemum research could encourage growers to adjust the timing and application of fertilizers, leading to input and cost reductions without sacrificing plant quality. Ultimately, our innovative strategy could be adopted for all potted ornamental crops grown indoors across Canada, which totalled in excess of 80 million plants in 2020, thereby improving the overall sustainability of the floricultural industry.