What must be considered in winter strawberry production under LEDs in Iceland?

ABSTRACT Supplementary lighting is essential to maintain year-round production in Iceland due to the extremely low natural light level in winter. In this research, the effects of high-pressure vapour sodium lamps (HPS) are compared to light emitting diodes (LED), both with similar photosynthetic photon flux density (PPFD). Strawberries (Fragaria x ananassa cv. ‘Sonata' and cv. ‘Magnum') were grown either under HPS lights or LEDs and 16°C/8°C (day/night). However, in the second winter, the day temperature was increased to 19°C under LEDs. The results showed that under the same temperature set points, the development of the flowers and the harvest was delayed by two weeks under LEDs due to a lower leaf, substrate and air temperature. However, when temperature set points were adapted, no delay under LEDs was observed. LEDs did not lead to higher yield, but to a higher energy use efficiency, while light use efficiency behaved contrary. Economic calculations clearly demonstrate that it is not justified to switch from HPS lights to LEDs. Instead, it is rather recommended to emphasise a high-yielding variety like Sonata in winter-growing of strawberries.


Introduction
Strawberries (Fragaria x ananassa) are one of the most popularly consumed berries in the world due to their nutritious properties and unique flavour (Simpson 2018).As Iceland is not self-sufficient in the production of strawberries, more than 400 t of strawberries (fobvalue: 2.322.380€, cif-price: 2.938.672€, market value: 8 €/kg; 1 € = 150 ISK) were imported in the year 2016 and more than 700 t of strawberries in the year 2022 according to Hagstofa Íslands (Statistics Iceland).In contrast, in 2016, about 100 t of strawberries (market value: 23 €/kg; 1 € = 150 ISK) in protected cultivation in Iceland were produced according to a survey by the author.The main challenges for local production of vegetables and berries in Iceland are competitive, as prices are considerably lower on imported ones (Sturludóttir et al. 2021).Despite of that the majority of consumers in Iceland support the local farmer (Halldórsdóttir and Nicholas 2016).Therefore, domestic horticulture can still be an asset for food supply in promoting sustainability, increasing food security and benefitting consumers.However, one of the main challenges being the high prices of electricity for lighting during the dark winter months (Stadler 2017;Sturludóttir et al. 2021).Nonetheless, to ensure winter production in Iceland and other high-latitude countries supplemental lighting is needed due to the extremely low natural solar radiation (Paponov et al. 2019;Appolloni et al. 2021).In lower-latitude countries like the Netherlands and Belgium, strawberries have been grown in heated greenhouses since several years (Van Delm et al. 2016).But strawberries are also cultivated in greenhouses in Norway (Verheul et al. 2007).Therefore, the use of supplemental light with a high light intensity might be a way to produce strawberries independent of the season in Iceland.Thus, the aim of the study is to establish the feasibility of growing more local food in Iceland to meet local demand.
The current primary source of supplementary lighting in Iceland is top lighting using high-pressure vapour sodium (HPS) lamps.However, light emitting diodes (LEDs) are commonly known to be more energy efficient (Singh et al. 2015).In contrast, HPS lights do not produce significant amounts of radiant heat (Singh et al. 2015;Bantis et al. 2018).As wavelengths differ between HPS and LEDs, spectral quality can have profound effects on the growth and yield of plants (Smith 1982).Among others, tomatoes (Appolloni et al. 2021), salad (Tamulaitis et al. 2005), sweet pepper (Brown et al. 1995) and strawberries (Philips 2015) were successfully grown under LEDs in other countries.Because optimal growth is among others dependent on optimal light (light intensity, light spectrum and light duration) and optimal temperature, these factors can be seen as primary determinants of the productivity of horticultural products (Gruda 2005).Therefore, it can be expected that adapted temperature settings might be necessary for growing horticultural products under LEDs (Van Delm et al. 2016).However, experience to expose plants with LEDs is missing in Iceland and as the natural light level during winter is much lower compared to more southern countries, obtained results may not directly be transferred.Since considerable variation might be expected between cultivars in their response to different light sources, two popular strawberry varieties for winter production under supplemental lighting were selected.It was tested, (1) how the yield of strawberries is influenced by different light sources with similar photosynthetic photon flux density (PPFD), (2) if temperature set points need to be adapted in dependence of the light source and (3) if growing strawberries during the darkest time is economically feasible.

Growing conditions
A greenhouse experiment with strawberries (Fragaria x ananassa cv.'Sonata' and cv.'Magnum') was carried out in the research greenhouse of the Agricultural University of Iceland at Reykir, Hveragerði (21°12'W, 64°0'N), South Iceland in two winter seasons (winter 2017/2018, winter 2018/2019).The experiment was set up in two greenhouse compartments (60 m 2 ) with different light sources: one of the compartments had HPS lights (18 HPS lamps (Philips,600 W)) and the other had LED lights (36 Green power LEDs: DR/B LB HO (Type1); the manufacturer Signify indicated no detailed data about the light spectrum).The position of the lights was adjusted to provide a similar PPFD within the two light treatments.
The PPFD at plant canopy height was measured in each light treatment by a quantum sensor (LI-190SA; LI-COR, Lincoln, NE) and by a light sensor logger (LI-250A; LI-COR, Lincoln, NE).PPFD amounted for the HPS treatment in average 277 µmol m −2 s −1 in the first winter season and 269 µmol m −2 s −1 in the second winter season, respectively, and for the LED treatment 279 µmol m −2 s −1 in the first winter season and 278 µmol m −2 s −1 in the second winter season, respectively.White plastic on all surrounding walls helped to get a higher light level at the edges of the growing area.The light was applied for 16 h (from 07:00 to 23:00).Solar irradiation stayed below 5 kWh m −2 week −1 during both winter seasons and can therefore be neglected.
Five hundred heavy tray plants of each variety were planted in 16 hanging gutters with a plant density of twelve plants m −2 .Planting took place on 7 December 2017 in the first winter experiment and on 1 October 2018 in the second winter experiment.Each hanging gutter was equipped with 16 five-litre pots filled with Klasmann substrate TS 4 (Klasmann-Deilmann GmbH, Geeste, Germany) and each pot contained four strawberry plants.The growth period ended on 5 April 2018 in the first winter experiment and on 14 January 2019 in the second winter experiment, respectively.
In the first winter experiment, the temperature set points under both light sources were the same, 16°C during the day and 8°C during the night.However, in the second winter, the temperature set point during the day was increased to 19°C in the LED treatment to compensate for the additional radiation heating under HPS lights ( = adapted temperature set points), while the temperature set point during the night was under both light sources 8°C.Plants received standard nutrition through drip irrigation (100 ml drip −1 ; watering from 09:00 to 21:00) containing 8,4 kg/7,3 kg calcium nitrate (until 10 white fruits/harvest); 0,2 l/0,25 l DTPA (6% Fe) (until 10 white fruits/harvest); 50 g EDDHA (6% Fe); 1,3 kg potassium sulphate; 3,6 kg magnesium sulphate; 1,7 kg monopotassium; 2,9 kg/7,3 kg potassium nitrate (until 10 white fruits/harvest); 51 g manganese sulphate (32,5% Mn) per 100 l irrigation water.The pH and electrical conductivity (EC) values of the watering system were monitored during the plant growth and adjusted to an EC of 1,5-1,7 mS cm −1 and a pH of 5,8 in the applied water.Runoff was arranged at the growing stage to 10-20% at sunny days and 0-5% on cloudy days and at flowering and carrying green fruits to 25-30% on sunny days and 10-15% on cloudy days.In both winter growing seasons, 800 ppm CO 2 was applied.Bumblebees were used for pollination of flowers.

Sampling and measurements
Each hanging gutter was split into two blocks and varieties were randomly distributed on either side of the gutter in six repetitions in each compartment.Each treatment consists therefore of 32 plants per repetition.
To be able to determine the duration until harvest, one unpollinated flower was marked once a week ( = each week 6 flowers per variety and light treatment).These marked flowers were daily observed until their pollination and at harvest of the fruits that were marked at the flowering stage, the harvest date was registered.
To be able to characterise the development of flowers and fruits and their harvest, the number of open flowers and fruits on each plant of one pot of each variety and repetition in each light treatment was weekly counted (1 pot × 4 plants × 2 varieties × 6 repetitions × 2 light treatments) and fruits were regularly harvested and classified into extra-class (>25 mm), first-class (>18 mm) and into not marketable fruits (too little fruits (<18 mm), unshaped fruits and mouldy fruits) by weight and number.Following this classification, fruits with more than 15 g were classified into extra-class, fruits with more than 5 g into first-class and fruits with less than 5 g were not marketable.Flower thinning practise (reducing the number of flowers to 10 or fewer per inflorescence) to be able to increase marketable yield was not conducted.
Twenty four-hour air temperature and air temperature during the day were regularly recorded by dataloggers.Substrate temperature was measured weekly in 1-2 cm depth by a portable thermometer (testo 926, Testo SE & Co. KGaA, Titisee-Neustadt, Germany) and leaf temperature weekly by a portable infrared contact thermometer (BEAM infrared thermometer, TFA Dostmann GmbH & Co. KG, Wertheim-Reicholzheim, Germany) in one pot of each variety and repetition in each greenhouse compartment (1 pot × 2 varieties × 6 repetitions × 2 light treatments).As in the second winter experiment, the temperature was increased in the LED treatment, the adapted temperature was measured.
The daily energy use (kWh) by supplemental lighting was measured by dataloggers and the total used energy during the growth period was determined.The energy use efficiency (total cumulative marketable yield in weight per kWh), light use efficiency (total cumulative marketable yield in weight per mol −1 of incident light) and profit margin were calculated.

Statistical analysis
The experiments were performed on the basis of randomised blocks for the different varieties with six repetitions, each with four plants (one pot) in each treatment (a total of 32 plants per light treatment and variety).SAS Version 9.4 was used for statistical evaluations.Substrate temperature, leaf temperature, yield parameters, energy use efficiency and light use efficiency were subjected to one-way analyses of variance with the significance of the means tested with the Tukey/Kramer HSD-test at p = 0.05.

Flowering
One to two days after flowering fruits were pollinated by bumblebees.The number of flowers and fruits increased up to a maximum of 45-65 flowers and fruits per plant (4-6 inflorescences with each 8-14 fruits), independent of the light source (Figure 1).After the maximum was reached, the amount decreased due to the start of the harvest.Sonata had a higher number of flowers and fruits compared to Magnum.
In the first winter season, where the temperature set points were the same, the development of flowers and fruits was delayed by two weeks under LEDs compared to strawberries that were grown under HPS lights (Figure 1(a)).Then, fruits ripened two weeks earlier under HPS lights and consequently, the number of flowers and fruits decreased two weeks earlier.
However, when temperature set points were adapted in the second winter season, the development of the flowers and fruits under HPS lights was one week behind of strawberry plants that received LEDs and the higher temperature set point during the day.Therefore, the development of the strawberries could be speeded up when temperature set points were adapted.

Yield
In the first winter season with the same temperature set points both varieties ripened two weeks earlier under HPS lights (Figure 2(a)).The average ripening time from the point of view of the whole harvest period was 41 days for both varieties under HPS lights, but 45 days for Magnum and 47 days for Sonata under LEDs.At the end of the harvest period, the yield was not statistically different between light sources.However, Sonata reached about 10% more yield than Magnum.This yield difference between varieties became obvious in the middle of the harvest period (Figure 2(a)).The yield of Sonata amounted 590 g plant −1 under LEDs and 610 g plant −1 under HPS lights.The yield of Magnum under LEDs was 530 g plant −1 and 520 g plant −1 under HPS lights.Neither the total number of marketable fruits, the extra-class fruits, and the first-class fruits, nor the average weight was statistically different between light sources (Tables 1 and 2).However, the variety had an influence on the total number of marketable fruits and first-class fruits (Table 1).Sonata had a higher number of marketable fruits than Magnum.In contrast, the average fruit weight was not affected by the variety (Table 2).Extra-class fruits had a weight of on average more than 20 g and first-class fruits around 10 g, resulting in an average total marketable weight of around 13 g.But the average marketable weight decreased from 20 g fruit −1 to 10 g fruit −1 with the proceeding harvest period (results not shown).However, when temperature set points were adapted in the second winter season, strawberries under LEDs and an increased temperature set point during the day were one week earlier ripe than under HPS lights and the lower temperature set point during the day, independent of the variety.On average, Sonata was ripe after 43 days under LEDs and the higher temperature set point during the day, while under HPS and the lower temperature set point during the day fruits ripened after 50 days.In contrast, Magnum was ripe after 45 days independent of the light source and adapted temperature set points.At the end of the harvest season, the accumulated marketable yield of Sonata was independent of the light source and adapted temperature set points (Figure 2(b)).Yield under LEDs and 19°C amounted to 560 g plant −1 , under HPS and 16°C 600 g plant −1 , respectively.In contrast, Magnum under LEDs and the higher temperature set point during the day reached a significantly lower yield (430 g plant −1 ) compared to strawberries grown under HPS lights and the lower temperature set point (520 g plant −1 ).However, a higher number of unpollinated flowers was observed for Magnum under LEDs and increased temperature set point during the day compared to Magnum grown under HPS lights (results not shown).Also, in the second winter season, the    lower number of first-class fruits was responsible for the lower yield of Magnum compared to Sonata, while the average fruit weight (total extra + first-class) was not affected by the light source (Tables 3 and 4).However, Magnum had under both light sources a significant higher average marketable weight (total extra + firstclass) than Sonata.The weight of extra-class fruits was on average more than 20 g and of first-class fruits around 10 g, resulting in an average total marketable weight of around 12-14 g.Like in the previous winter season, the average marketable weight decreased from 20 g fruit −1 to 10 g fruit −1 with the proceeding harvest period (results not shown).

Growth conditions
In the first winter season with the same temperature set points, the 24-h air temperature and the air temperature during the day were on average 0,5°C higher under HPS lights than under LEDs (Table 5).Furthermore, the substrate temperature and the leaf temperature were significantly higher under HPS lights than under LEDs.However, when temperature set points were adapted under LEDs in the second winter season and the temperature set point during the day increased from 16°C to 19°C, the 24-h air temperature and the air temperature during the day was about 1,5°C higher than under HPS lights and the lower temperature set point during the day of 16°C (Table 6).Also, the substrate temperature of strawberries grown under LEDs and 19°C was significantly higher compared to strawberries grown under HPS lights and 16°C.In contrast, the leaf temperature was independent of the light source and adapted temperature set points.

Economic calculations
Strawberries grown under LEDs used daily nearly 45% lower kWh's compared to strawberries grown under HPS lights (Tables 7 and 8).Thus, energy costs for producing strawberries were lower for LEDs.In contrast, investment costs in LEDs were higher than for HPS lights.Despite of the two weeks longer growth period of strawberries grown under LEDs in the first winter season with the same temperature set points, the used energy and consequently the energy costs were about 45% lower than for strawberries grown under HPS lights (Table 7).Then, the profit margin increased by 8 € m −2 for Magnum and by 3 € m −2 for Sonata when    LEDs were used instead of HPS lights.By the selection of Sonata instead of Magnum, the profit margin could be increased by 15 € m −2 under HPS lights and by 11 € m −2 under LEDs, respectively.When strawberries were lightened with LEDs, kWh's were transferred better into yield compared to strawberries exposed to HPS lights (Table 7).In addition, Sonata had a better energy use efficiency than Magnum under LEDs.However, under HPS lights no variety effect was found.The light use efficiency provides insight into production efficiency and indicates that Sonata was significantly better using the light than Magnum under HPS lights, while no variety effect was observed under LEDs.
In the second winter season with adapted temperature set points, the profit margin was remarkable higher for Sonata compared to Magnum (Table 8).While the profit margin was independent of the light source and adapted temperature set points for Sonata, Magnum gained a 10 € m −2 higher profit margin when grown under HPS lights and 16°C compared to LEDs and 19°C.When strawberries were lightened with LEDs, kWh's were transferred better into yield compared to strawberries grown under HPS lights (Table 8).Sonata had a significantly better energy use efficiency compared to Magnum under LEDs, while this effect was not observed under HPS lights.Also, Sonata had a significantly higher light use efficiency than Magnum under both light sources.In addition, the light use efficiency was higher under HPS lights than under LEDs for the variety Magnum.

Discussion
In Iceland under low natural light conditions, winter production of strawberries was possible with a high supplemental light level of either HPS lights or LEDs.However, the development of flowers and fruits as well as the ripening was influenced by the light source and delayed by two weeks under LEDs in the first winter season with the same temperature set points (Figure 1 (a)).The lower substrate temperature, leaf temperature, 24-h air temperature and air temperature during the day under LEDs could have caused this negative influence on plant growth and beginning of ripening, whereas strawberries grown under HPS lights profited from radiant heating (Table 5).Indeed, according to Hernández and Kubota (2015) the higher canopy air temperature under HPS lights compared to the blue and red LEDs was responsible for the 28% greater shoot dry mass of cucumbers, the 28-32% higher shoot fresh weight and the 9-12% higher leaf number.Also, Särkkä et al. (2017) reported that the growth of cucumbers was decreased under LEDs (top and interlighting) and the lower temperature compared to HPS lights (HPS top and HPS interlights) through reduced cell growth and indirectly through sink strength.In contrast, Stadler (2021) reported that the tomato harvest was delayed by half a week when plants were grown under HPS lights compared to LEDs in young plant production.Also, in the study of Dysko and Kaniszewski (2021) tomatoes that got LEDs fruited earlier than plants that got HPS lights, thus increasing the early yield.However, it was found that the development and the beginning of the harvest under LEDs could be speeded up when temperature set points were adapted in the second winter season (Figures 1(b) and 2(b)).Therefore, the regulation of temperature in accordance with the used light source seems to be important to control the development.This was in accordance with Van Delm et al. (2016) who see the regulation of temperature and lighting strategy as important for plant balance between earliness and total yield.All experiments that compare HPS and LED lights need to assess the differences in plant temperature to ensure that any effect of temperature can be separated from the effects of light (Davis and Burns 2016).In Iceland, adapting the temperature under LEDs does not involve additional production costs due to geothermal energy.However, heating costs must be added to horticultural production in particular under LEDs in other countries.Especially for strawberries further research is needed as  according to a recent study, only a few studies on the effects of light and temperature factors on strawberry production in greenhouses have been conducted so far (Tang et al. 2020).
In contrast to the fact that the development was influenced by the light source, the amount of the marketable yield was not affected by the light source, neither in the first winter season with the same temperature set points nor in the second winter season with adapted temperature set points (Figures 2(a) and 2(b)).However, contrasting results regarding the effect of light sources on yield are reported in the literature.For example, for tomatoes, Dueck et al. (2012a) described a lower production under LEDs than under HPS lights, while Wacker et al. (2022) reported a 10% higher tomato yield under LEDs compared to HPS lights and concluded that switching from HPS lamps to LEDs enables increasing productivity.The significantly lower yield of Magnum under LEDs in the second winter (Figure 2(b)) was possibly caused by a lower number of pollinated flowers.Assuming the number of unpollinated flowers under LEDs and increased temperature would have been lower, the same yield value under HPS lights and LEDs might have been expected, as it was the case for the variety Sonata.
Yield differences regarding which variety was grown were observed in both winter seasons.The marketable yield of Magnum was about 10% lower compared to Sonata (Figure 2).This was related to a lower number of marketable fruits (Tables 1 and 3), more precise the number of first-class fruits due to significantly more fruits with deformation.Sonata was a better variety under supplemental lighting with a high PPFD during winter than Magnum.Therefore, the selection of the variety is gaining increased importance in view of economic aspects.
However, as the optimal photosynthetically active radiation that improved the overall quality of strawberries was shown to be 339.6-452.8μmol m −2 s −1 (Zhong et al. 2011), a higher yield might have been possible by applying instead of less than 300 µmol m −2 s −1 a higher PPFD value as in the present study.Among that, by selecting a different light spectrum of the LEDs, it is possible to influence yield (Choi et al. 2015).It has been pointed out that adding more red and blue light led to a significant increase in the amount and production of harvested strawberry fruit as well as a shortening of the fruit's maturity period compared to nonlightened strawberries (Hidaka et al. 2016).Indeed, Guiamba et al. (2022) concluded that strawberry plants require prolonged and high light intensities with a high red-light component for maximum performance and biomass production.Folta and Childers (2008) demonstrated that the growth of strawberry plants cultivated for 40 days in a growth chamber was higher when illuminated with combined blue and red LED light compared to red or blue one alone.Therefore, a suitable temperature, light intensity and light spectrum can increase the yield and quality of strawberries.
The energy use and consequently the energy costs could be reduced by 45% with LEDs compared to HPS lights (Tables 7 and 8).This value was comparable to 40% less electricity of LEDs in tomato production in the study by Wacker et al. (2022).However, the energy use efficiency increased significantly with the use of LEDs compared to HPS lights, while the light use efficiency acted the opposite way (Tables 7 and 8).Furthermore, Sonata can convert incident light into marketable yield more efficiently than Magnum.Also, Särkkä et al. (2017) found that the electrical use efficiency (kg yield J −1 ) increased when HPS light was replaced with LEDs in cucumbers, but concluded that the high capital cost is still an important aspect delaying the LED technology in horticultural lighting.Indeed, Tables 7 and 8 indicate that the profit margin does not justify the use of LEDs.At the current stage of LED technology, the best lighting solution for high-latitude winter growing appears to be HPS top lights combined with LED interlights according to Särkkä et al. (2017).However, the authors concluded also that a solution for the near future could be a combination of LED and HPS top lights in order to maintain a suitable temperature but reduce energy use.Indeed, also in the study of Rakutko et al. (2020), a Hybrid lighting system is seen as the best available technique as it significantly increased the efficiency of light energy use by cultivated plants leading to shorter pre-fruiting period, higher plant productivity and improved quality of fruits.In Norway, HPS top lighting was more efficient for tomato production than LED interlighting (Verheul et al. 2022).All lighting systems at high-latitudes have in common that temperature and supplemental light control need to be optimised for each supplemental light system (Wacker et al. 2022).Also, in lower-latitudes (Netherlands), a combination of HPS and LEDs as top lights is suggested by Dueck et al. (2012b) as the most promising alternative (for tomato production) when considering production parameters and costs for lighting and heating.In contrast, Kuijpers et al. (2021) found that the operational return of a tomato greenhouse can be improved by 9% when switching from HPS lights to LEDs.Furthermore, protocols for the correct management of plant irrigation and growth need the be developed before HPS lamps can be switched with LEDs (Davis and Burns 2016).Overall, it can therefore be concluded that more detailed scientific studies are necessary to understand the effect of different spectra using LEDs on plant physiology and to investigate the response to supplemental light quality of economically important greenhouse crops and validate the appropriate and ideal wavelength combinations for important plant species (Bantis et al. 2018;Gómez et al. 2013;Hernández and Kubota 2015;Singh et al. 2015).Moreover, a more successful way to increase the profit margin than with the light source seems to be with the selection of a good yielding variety (Tables 7 and 8).

Figure 1 .
Figure 1.Open flowers and fruits per plant with different light sources and same temperature set points in the first winter season (a) and adapted temperature set points in the second winter season (b).Values are the means and ± SE of 24 plants in each light treatment.

Figure 2 .
Figure 2. Accumulated marketable yield with different light sources at same (a) and adapted (b) temperature set points.Values are the means and ± SE of 24 plants in each light treatment.Letters indicate significant differences at the end of the experiment (HSD, p ≤ 0.05).

Table 1 .
Number of harvested fruits with different light sources and same temperature set points.

Table 2 .
Average weight with different light sources and same temperature set points.

Table 3 .
Number of harvested fruits with different light sources and adapted temperature set points.

Table 4 .
Average weight with different light sources and adapted temperature set points.

Table 5 .
24-h air temperature, air temperature during the day, substrate temperature and leaf temperature with different light sources at same temperature set points.

Table 6 .
24-h air temperature, air temperature during the day, substrate temperature and leaf temperature with different light sources at adapted temperature set points.
Letters indicate significant differences at the end of the experiment (HSD, p ≤ 0.05).

Table 7 .
Used energy, energy use efficiency, light use efficiency and profit margin with different light sources at same temperature set points.

Table 8 .
Used energy, energy use efficiency, light use efficiency and profit margin with different light sources at adapted temperature set points.Investment in lights (€ m −2 ): 7,45 for LEDs, 4,39 for HPS. *