2022 Volume 91 Issue 3 Pages 247-266
The fresh fruit yield of tomatoes (Solanum lycopersicum) is determined by yield components and related traits. In low-truss cultivation in Japan, light use efficiency in the same cultivar is not significantly affected by plant density or seedling stage at transplanting. Total dry-matter production equals light use efficiency × light intercepted by plants. Light interception is determined by leaf area index and the light extinction coefficient in the plant canopy. Light use efficiency is determined by leaf photosynthetic rate and light extinction coefficient. Light use efficiency is expressed as a function of daytime CO2 concentration. The high yield of modern tomato cultivars in the Netherlands is due to an increase in total dry matter production of plants, not to an increase in dry matter partitioning of fruits. An increase in the photosynthetic rate and a decrease in the light extinction coefficient may have increased light use efficiency among Dutch cultivars. Although modern Japanese cultivars have a high content of both water and soluble solids of fruits, yield has not increased. Since the yield of greenhouse tomatoes in Japan has increased little since the 1980s, Japanese researchers have attempted to improve the yield using current Japanese cultivars. Crop growth models can help management decisions on cultural practices, control of greenhouse climate, fertilization, and irrigation. Several models of tomato growth including TOMSIM and TOMGRO have been developed. A growth model based on the relationships among yield components and related traits has been developed to predict dry matter production and yield of greenhouse tomatoes. The model recommended a leaf area index with the highest dry matter production at a certain solar radiation. To improve yield of greenhouse tomatoes by using the model, a one-year experiment was carried out in a greenhouse. The annual yield of tomatoes with Brix at least 5° was at least 50 kg·m−2 for a popular Japanese cultivar, ‘CF Momotaro York’. For the near future, the growth models show promise for growers in helping yield improvement and crop management in greenhouses.
A greenhouse provides suitable temperature, humidity, CO2 concentration and light for crop photosynthesis and growth. Yields of crops such as tomato, cucumber, and sweet pepper are clearly higher in greenhouses than in fields or orchards. In the Netherlands, the yield of greenhouse tomatoes was ca. 30 kg·m−2·year−1 in the 1980s, but more than 60 kg·m−2·year−1 in the 2000s (Kwantitatieve Informatie voor de Glastuinbouw, 2017). In Japan, the average tomato yield has increased only slightly since the 1980s, with 10.2 kg·m−2 produced in winter and spring and 4.1 kg·m−2 produced in summer and autumn (Ministry of Agriculture, Forestry and Fisheries of Japan, 2020). In the prefecture with the highest tomato yield, the yield is currently 15.9 kg·m−2 in winter and spring and 7.1 kg·m−2 in summer and autumn.
The reasons for the high yield in the Netherlands include improvements in greenhouse transmissivity and cultivation technology, sophisticated environmental control, and breeding of high-yielding cultivars. The purpose of environmental control in a greenhouse is to regulate the growth and maturation rates, mainly by controlling temperature, and to increase dry matter production, mainly by controlling light, carbon dioxide and crop canopy. This review was originally published in Japanese (Higashide, 2018) and is now updated with the current knowledge of growth and yield prediction. This paper focuses on factors important for dry matter production, and therefore yield improvement, by comparing tomato production in the Netherlands and Japan, and providing an update on growth modelling in Japanese greenhouse tomato cultivars.
The hierarchy of tomato-yield components and elements associated with tomato yield is shown in Figure 1 (Higashide, 2013). Fresh fruit yield is determined by dry fruit yield and fruit dry matter content: increasing the former or decreasing the latter increases fresh fruit yield. However, a decrease in fruit dry matter content has almost the same meaning as a decrease in soluble solids (Brix) and is not usually welcomed. Dry fruit yield can be increased by increasing either total dry matter production or the fraction of dry matter distributed to fruits. The latter is defined as the ratio of dry matter distributed to fruits to total dry matter, and is called a harvest index. The distribution of dry matter to fruits depends on the strength of single-fruit sinks and the number of fruits per plant (Heuvelink and Dorais, 2005; Tanaka and Fujita, 1972). Sink strength is the ability to attract photosynthetic products, which are produced mainly by leaves; the magnitude of sink strength determines the distribution ratio of photosynthates in a plant. In tomatoes, fruit is the strongest sink organ, and photosynthates are distributed to fruits at a high rate.
Hierarchy of yield components and related traits in tomatoes. Solid arrows: positive associations; dashed arrows: negative associations. The figure is reproduced from Higashide (2013) with permission from Nova Science Publishers.
Total dry matter production is determined by light use efficiency and the amount of light intercepted by plants. The latter increases with an increase in leaf area per area of ground, but the rate of this increase peaks when the leaf area index (LAI) becomes high. Thus, if the LAI is high, total dry matter production is increased mainly by an increase in light use efficiency. Light use efficiency is positively affected by the photosynthetic rate of individual leaves and negatively affected by the extinction coefficient of plant canopy. These two factors are often determined by the characteristics of cultivars rather than by temperature and nutrient supply in a greenhouse. In the same cultivar, increasing CO2 concentration increases leaf photosynthetic rate at the same light intensity and therefore increases light use efficiency (De Gelder et al., 2005; Fierro et al., 1994; Hicklenton and Jolliffe, 1978; Nederhoff, 1994; Tremblay and Gosselin, 1998; Tripp et al., 1991).
Such a hierarchical structure is used in crops other than tomatoes. Analysis of yield-related factors for yield improvement was reported in cucumber and sweet pepper (Higashide et al., 2012a; Maeda and Ahn, 2021; Watabe et al., 2021).
In tomato plants, photosynthesis takes place mainly in leaves. Light at wavelengths between 380 nm and 710 nm is effective in photosynthesis. It is called photosynthetically active radiation (PAR) and accounts for about 50% of total solar radiation (Ohtani, 1997). Tomatoes in commercial greenhouses form canopies, and light interception by the canopy can be calculated by subtracting light intensity under the canopy from that above the canopy (Fig. 2). For example, if the relative light intensity is 100% above the canopy and 10% under the canopy, 90% of light is intercepted by the canopy. This value multiplied by PAR during the day is daily light interception (MJ·m−2·d−1). The cumulative amount of daily light interception for the growing period is the cumulative light interception. The ratio of intercepted light (90% in the above example) can be measured, but also calculated from LAI and light extinction coefficient, as described in detail below.
Estimation of light interception by the plant canopy.
If crops grow without pests, physiological disorders, or equipment failures, their growth and yield are determined mainly by the amount of light intercepted by the plants, and total dry matter production can be expressed as a function of the amount of intercepted light. Kaneko et al. (2015) investigated the relationship between intercepted light and total dry matter production in three-truss cultivated tomatoes, using the cultivar ‘Momotaro York’. Plants were grown at two different densities (3.9 and 2.6 plants·m−2) using a nutrient film technique system. Seedlings were transplanted at different stages—the 2–3 leaf stage (very early), before anthesis (early), first flower anthesis (standard), and at the third flower anthesis (late)—and the first to third fruit trusses were harvested. Fruit yield per area was significantly higher at high plant density than at low density, and was increased by transplanting younger seedlings (Fig. 3). The relationship between cumulative light interception by tomato plants and total aboveground dry matter production is shown in Figure 4. Since seedling stage at transplanting and plant density vary, both cumulative light interception and total dry matter production are distributed over a wide range. Total aboveground dry matter production is linearly positively correlated with cumulative light interception (Fig. 4). The slope of the regression line (Fig. 4) is the light use efficiency, and its value of 1.93 means that 1 MJ of PAR produces 1.93 g of dry matter. Plant density and the seedling stage at transplanting have no significant effects on light use efficiency. The difference in total dry matter production and yield may be caused by the difference in cumulative light interception.
Effects of plant stage at transplanting and plant density on tomato yield in three-truss cultivation. Plant stage was very early, early, standard, or late; plant density was high or low. Fruit yield differed significantly among plant stages, and densities (P < 0.001), but no significant interaction of plant stage × density, by two-way ANOVA (n = 8–10). Figure reproduced from Kaneko et al. (2015) with permission from JSHS.
Dry matter production as a function of cumulative PAR intercepted by tomato plants grown at high or low plant density, and transplanted at different plant stages (very early, early, standard, or late). Figure reproduced from Kaneko et al. (2015) with permission from JSHS.
In low-truss tomato production, since the LAI is low over a considerable part of the cultivation period, the difference in LAI would affect total light interception and, as a result, total dry matter production and yield. Ohkubo et al. (2019) reported that differences in light interception due to leaf picking contributed to differences in total dry matter production in single-truss tomatoes.
The cumulative number of leaves and cumulative leaf area from the top of the canopy are larger in the lower part of the canopy than in the upper part, and LAI increases from the upper part to the lower part (Fig. 5). Light intensity is smaller in the lower part than in the upper part because light is intercepted by the crop. Monsi and Saeki (2005) showed that the relationship between light interception by plants and LAI is similar to the Lambert–Beer law. Therefore, light intensity within a canopy can be expressed as a function of LAI, and as the LAI increases towards the bottom of the canopy, the light intensity decreases (Fig. 5). An increase in LAI from 0.5 to 1.0 increases relative light interception from about 30% to 50%, and an increase in LAI from 1.0 to 3.0 increases relative light interception from about 50% to 90%. However, an increase in LAI from 3.0 to 4.0 increases relative light interception from about 90% to 95%, and an increase in LAI from 4 to 5 increases relative light interception only slightly. When LAI becomes higher than a certain value, the mutual shading of leaves increases, and therefore light interception does not increase much even if LAI increases.
Relative light strength and intercepted light in a plant canopy as a function of leaf area index (LAI) at a light extinction coefficient (k) of 0.75.
Two aspects have to be considered by growers in leaf area management of tomato canopy (Fig. 5). The first is avoiding excessive LAI. If LAI is too high, light intensity will be below the light-compensation point of photosynthesis and not only will not contribute to dry matter production by photosynthesis, but will also result in photosynthate consumption by respiration. Heuvelink and Buiskool (1995) showed that the yield of tomatoes increased as LAI increased, but the increase in yield was small when LAI exceeded 4.0. The second aspect is that the wasted portion of light transmitted into the greenhouse increases at low LAI. At LAI lower than 3.0, more than 10% of light is not intercepted by the canopy and illuminates the floor of the greenhouse. Many greenhouses in Japan have low eave heights, which limit crop height and LAI. Even in greenhouses with high eaves, LAI does not reach 3 when plants are small or plant density is low. Therefore, under these conditions, there may be significant potential for an increase in yield by using the available light.
Light extinction coefficient k indicates the degree of attenuation of light in the plant canopy (Fig. 5). At high attenuation (high k), light is intercepted by leaves in the upper part of the canopy and only a small amount reaches the lower part of the canopy, whereas at low attenuation (low k), a lot more reaches the lower part. On the basis of Figure 5, relative light interception can be calculated as a function of LAI and the extinction coefficient.
| [Eqn. 1.1] |
| [Eqn. 1.2] |
| [Eqn. 1.3] |
Daily light interception is obtained by multiplying relative light interception by PAR accumulated during the day. The cumulative light interception for a growing period is the cumulative amount of daily light interception.
As mentioned above, although total dry matter production is proportional to light interception, the regression slope is not affected by plant density or the seedling stage in the same cultivar (Fig. 4; Kaneko et al., 2015), indicating that total dry matter production can be expressed as a function of cumulative light interception and light use efficiency, which is a coefficient indicating the efficiency of dry matter production per unit of intercepted PAR.
| [Eqn. 2] |
In tomatoes, yield is the total weight of fruits as the fruits are the edible part. Total dry matter production includes dry weight of fruits and of other organs such as leaves and stems. If the ratio of dry matter distribution to fruits is constant, total dry matter production must be increased for yield increase. Despite differences in the ratio of dry matter distribution to fruits among cultivars, the ratio in the same cultivar is usually unchanged (Higashide and Heuvelink, 2009; Higashide et al., 2012b, 2014a, 2015; van der Ploeg et al., 2007). In other words, the target yield corresponds to a certain total dry matter production. To increase total dry matter production, light use efficiency, cumulative light interception, or both, must be increased (Higashide, 2015).
One way to increase light interception by a plant canopy is to increase the light transmittance of a greenhouse. Solar radiation and day length are impossible to control, except by blackout screen or supplementary light; therefore, to increase light transmittance the attenuation of light, by covering materials and greenhouse frames, has to be reduced. The light transmissivity of greenhouses (the percentage of light entering the greenhouse when the outdoor light is 100%) has improved (Baeza and López, 2012; Critten, 1993; von Elsner et al., 2000a, b) and reaches approximately 80% in the most modern greenhouses in the Netherlands (Hemming et al., 2011). On the other hand, the light transmissivity of Japanese greenhouses has improved little. Since solar radiation is higher in Japan than in the Netherlands, improvement in light transmissivity is less important. Indoor and outdoor solar radiation on a clear day in a greenhouse conforming to the Building Standards and Fire Defense Laws in Japan (Higashide et al., 2014c) is shown in Figure 6. Outdoor radiation on clear days shows a bell-shaped curve, but the radiation inside the greenhouse differs greatly between exposure to light and shading by frames or other structures, and the radiation fluctuates during the day. Light attenuation by covering material is 10%–15%, and 20–30% of light is blocked by the frame. The total daily light transmittance into the greenhouse is 54–60%. Structural strength is required for greenhouses in Japan to meet the standards for earthquake and wind resistance (Hasegawa, 2013). Therefore, to improve the light transmissivity of greenhouses, it is necessary to reduce the number and size of frames whilst maintaining strength.
Solar radiation on a clear day; outside and at two interior locations (P7 and P8) in a greenhouse built to meet building and fire standards (21 February 2012; NARO, Tsukuba, Japan). Figure reproduced from Higashide et al. (2014c) with permission from NARO Institute of Vegetable and Floriculture Science.
The other approach to increasing light interception by plant canopy is through leaf management. The rate at which new tomato leaves appear on the plant (hereafter leaf appearance rate) depends primarily on temperature in the absence of restrictive factors such as disease or water deficiency (De Koning, 1994). Since inflorescences appear in tomato every three leaves, the appearance rate of inflorescences is 1/3 that of the number of leaves, and is also determined mainly by temperature.
Fruit number, leaf picking, and plant density affect the amount of photosynthates in the crop, but have little effect on the rate of leaf appearance (Heuvelink and Marcelis, 1996). For example, if the number of tomato fruits per truss is reduced from seven to one, leaf appearance rate increases by only 9% (Heuvelink and Marcelis, 1996). Low plant densities increase light interception and photosynthate production per plant, but have little effect on leaf appearance rate. De Koning (1994) reported the following equation for the relationship between leaf appearance rate and daily average temperature.
| [Eqn. 3] |
This equation suggests that leaf appearance rate increases almost proportionally to temperature between 17 and 23°C (Fig. 7A), with 11 leaves appearing over 30 days at 17°C, 14 leaves at 22°C, and 16.5 leaves at 27°C (Fig. 7B). This equation is derived from the Dutch cultivar ‘Counter’, but the relationships in other tomato cultivars are similar (De Koning, 1994).
Influence of temperature on leaf appearance in tomato plants. Leaf appearance rate per day (A) and leaf appearance at 17, 21, and 26°C (B). Figure produced by the author based on Eqn.3.2.2 in De Koning (1994).
An increase in leaf area increases light interception and therefore dry matter production. Because leaf appearance rate is not significantly affected by factors other than temperature, it is not necessary for growers to consider factors such as CO2 concentration, plant density, or fruit number to control leaf appearance. On the basis of the relationship between leaf appearance rate and temperature, light interception can be increased by temperature control. Increasing the temperature after transplanting accelerates increases in total leaf number and area, increasing light interception and therefore total dry matter production. The produced dry matter is used for plant growth, further increasing light interception.
If the rates of dry matter production and leaf appearance are not balanced, the area of individual leaves may be affected. At low temperature but high dry matter production owing to high light, leaf appearance rate is low, but individual leaf area may be large. At high temperature but insufficient dry matter production due to a shortage of light, leaf appearance rate is high, but the area of individual leaves may be small.
In tomatoes, a high rate of leaf appearance means that inflorescence appearance and fruit load per plant increase. If fruit load increases too much, the number of aborted fruits may increase and fruit set rate may decrease. The well-known decrease in fruit set rate during high temperature periods may be caused not only directly by a decrease in pollen fertility (Peet et al., 1998; Sato et al., 2000, 2002), but also indirectly by an increase in fruit load. In Japan, the weekly yield of tomatoes fluctuates dramatically at high temperatures in summer (Higashide, 2009b).
Leaf appearance is usually faster at 22°C than at 20°C, but in the long term, temperature only slightly affects dry matter production within the appropriate temperature range. The photosynthetic rate varies little at daily average temperatures ranging from 15°C to 25°C (De Koning, 1994). Although daily average temperature affects mainly growth and yield, short-term changes in temperature and the patterns of these changes have little effect (De Koning, 1990). Vanthoor et al. (2011) modeled the effect of temperature on growth as a relative growth rate based on the temperature suppressing (inhibition) function of Boote and Scholberg (2006). On the basis of their report, we attempted to quantify the effect of temperature in a greenhouse on dry matter production as a temperature score. Specifically, the temperature Ta (°C) measured in a greenhouse was classified as follows, and temperature scores (ScTm) obtained.
| [Eqn. 4] |
Ta, air temperature (°C);
Tbase, base temperature for instantaneous crop growth inhibition (°C);
Tmax, maximum temperature for instantaneous crop growth inhibition (°C);
Topt1, the lowest temperature within temperature range without growth inhibition (°C);
Topt2, the highest temperature within temperature range without growth inhibition (°C).
Higashide et al. (2022) compared cucumber yields in 6 short-term experiments (3–5 months) conducted over 3 years in 2 greenhouses under different environmental conditions. The temperature scores were significantly positively correlated with yields per solar radiation (r = 0.35, P < 0.01), suggesting that the effect of temperature management can be assessed using temperature score.
The improvement of light use efficiency leads to an increase in total dry matter production (Eqn. 2). In short-term tomato cultivation, Kaneko et al. (2015) found that LAI greatly influenced yield because the proportion of the low-LAI period in the whole growth period was high. In long-term tomato cultivation conducted in many countries, except for a short period of time after transplanting, LAI was sufficiently high, and further increases in LAI did not increase light interception. Accordingly, light use efficiency may greatly affect total dry matter production.
Light use efficiency differs among plant species and cultivars. Light use efficiencies of greenhouse tomatoes are approximately 2–4 g·MJ−1 PAR at the atmospheric CO2 concentration, but differ among cultivars (Heuvelink and Buiskool, 1995; Higashide and Heuvelink, 2009; Higashide et al., 2012b). Therefore, choice of cultivars with high light use efficiency is important to increase dry matter production and yield.
The photosynthetic rate in individual leaves is small at low light intensity, increases with increasing light intensity, and reaches saturation when light intensity exceeds a certain level (Fig. 8). This relationship is illustrated by the photosynthetic light–response curve, where the photosynthetic rate at saturation is the maximum photosynthetic rate (Pmax). The photosynthetic light-response curve can be expressed as a non-rectangular hyperbola, and the photosynthetic rate (P) of an individual leaf at the light intensity I′ is determined by the Pmax, respiratory rate (R), early slope (φ), and convexity (θ) (Eqn. 5; Thornley, 1976).
| [Eqn. 5] |
Photosynthetic light-response curve of an individual leaf. The maximum photosynthetic rate (Pmax), which differed in species or cultivars, determines leaf photosynthetic rate at saturation. θ determines the shape of the curve. 0 < θ < 1. Cultivar A shows higher Pmax than Cultivars B and C. Although Cultivars B and C show the same Pmax, θ in Cultivar B is higher than in Cultivar C.
Since light use efficiency indicates the efficiency of dry matter production per intercepted PAR, the photosynthetic curve and the maximum photosynthetic rate in individual leaves may affect light use efficiency. At sufficient light higher than saturation level, the photosynthetic rate of individual leaves is determined mainly by the maximum photosynthetic rate. Therefore, a cultivar with a high maximum photosynthetic rate has high light use efficiency. In general, however, in a plant canopy with sufficient LAI, not all leaves reach the maximum photosynthetic rate. Under high solar radiation, the photosynthetic rate reaches maximum in the upper leaves of the canopy, but not in the lower leaves. Exposure to more intense light increases the number of leaves reaching the maximum photosynthetic rate, but the entire canopy cannot reach this rate.
The photosynthetic rate of the entire canopy can be determined from the photosynthetic curve, the extinction coefficient, and the intensity of light within the canopy as determined from the LAI and light above the canopy. First, relative light intensity within a canopy is calculated as a function of LAI using the extinction coefficient (Fig. 5, Higashide, 2013; Monsi and Saeki, 2005). When the light intensity is determined, the photosynthetic rate is obtained from the light–photosynthetic curve. Then, the photosynthetic rate of the leaves at each location within the canopy can be calculated, and the sum of the photosynthetic rates of all leaves is the photosynthetic rate of the entire canopy.
The photosynthetic rate of individual leaves is affected by previous light exposure (Trouwborst et al., 2011). The maximum photosynthetic rate is high in leaves in the upper part of the canopy, which are sufficiently exposed to light, but is lower in the leaves in the lower part of the canopy, which are exposed to weaker light. Thus, even in the same plant, the maximum photosynthetic rates of individual leaves differ between upper and lower leaves. Strictly speaking, this difference should also be taken into account in calculating whole canopy photosynthetic rate. However, the photosynthetic rate of the whole canopy can be estimated from a single photosynthetic light–response curve because leaves at each level always receive the same proportion of light.
Humidity may also affect the growth and yield of tomatoes and other crops (Bakker, 1990, 1991; Jolliet and Bailey, 1992; Jolliet et al., 1993). Stanghellini and Bunce (1993) reported that extremely low humidity can lead to stomatal closure in tomato leaves, and thereby decrease transpiration and photosynthetic rate. In tomato production under low-humidity conditions, such as in the Mediterranean coastal regions, humidification by a fogging system alleviates stomatal closure and yield reduction (Romero-Aranda et al., 2002). Therefore, to control humidity, a fogging system is installed in modern greenhouses (De Gelder et al., 2012; Harel et al., 2014). Higashide et al. (2015) investigated the effects of CO2 application and humidification by a fogging system on yield and total dry matter production in three tomato cultivars. Application of CO2 in combination with humidification increased the yield and total dry matter production of all three cultivars, and at the same time increased light use efficiency, which was 1.5–1.6 times that at CO2 atmospheric concentration without humidification.
Higashide et al. (2022) reported no significant correlation of cucumber yield with solar radiation and humidity factors, such as relative humidity and vapor pressure deficit, at various CO2 concentrations. According to a review by Grange and Hand (1987), humidity has no direct effect on photosynthetic carbon assimilation of horticultural crops at a vapor pressure deficit of 0.2–1.0 kPa. In cucumber, resistance to low humidity increases after short exposure (Shibuya et al., 2017). Thus, high humidity does not directly promote dry matter production and yield of cucumbers. In cucumbers or tomatoes, the effects of humidity on light use efficiency or dry matter production are still not quantified and modelled.
The photosynthetic rate of an individual leaf plotted against CO2 concentration also shows saturation, similar to that in the photosynthetic light–response curve. With an increase in CO2 concentration, the photosynthetic rate increases, but its increment decreases. In modern greenhouses, CO2 application systems are installed to increase yield (De Gelder et al., 2005; Fierro et al., 1994; Hicklenton and Jolliffe, 1978; Nederhoff, 1994; Tremblay and Gosselin, 1998; Tripp et al., 1991). In Japan, although CO2 enrichment has not been widely used until now, it is becoming more common, and the effects and efficiencies of CO2 enrichment in large-scale greenhouses are reported (Kuroyanagi et al., 2014; Takahashi et al., 2012).
As described above, light use efficiency is closely related to the photosynthetic rate of individual leaves, which varies with CO2 concentration if the light is constant. Therefore, when the CO2 concentration is changed, light use efficiency also changes. Higashide et al. (2015) reported that CO2 application during the day at 750 ppm in combination with humidification resulted in a light use efficiency 1.5–1.6 times that at 510 ppm without humidification.
Saito et al. (2020b) compared dry matter production of three tomato cultivars at different CO2 concentrations and determined the light use efficiency at each CO2 concentration (Fig. 9). It was found that increasing CO2 concentration increased light use efficiency, and that light use efficiency and the degree to which it was affected by CO2 concentration differed among the cultivars:
| [Eqn. 6] |
LUEn, light use efficiency at n days after transplanting (DAT) (MJ·m−2·day−1);
CO2_n, daytime CO2 concentration at n DAT (μmol·mol−1);
m (coefficient): Rn, 4.9166; My, 4.5124; Mn, 4.5217;
o (coefficient): Rn, −25.434; My, −24.168; Mn, −22.74;
Rn, ‘Ringyoku’; My, ‘CF Momotaro York’; Mn, ‘Managua’.
Light use efficiency as a function of daytime CO2 concentration in a greenhouse. Rn, ‘Ringyoku’; My, ‘CF Momotaro York’; Mn, ‘Managua’. Open symbols, Exp. 2; closed symbols, Exp. 3. (R2 = 0.82 for Rn, 0.84 for My, 0.69 for Mn; P < 0.001; n = 12). Figure reproduced from Saito et al. (2020a) with permission from JSHS.
The relationship between CO2 concentration and light use efficiency is also reported in short-term cultivation. Itoh et al. (2020a) conducted similar experiments six times in a commercial greenhouse that produced tomatoes by low-truss cultivation (the plants were pinched above the third fruit truss, and the first to third truss were harvested). The time from transplanting to the end of cultivation ranged from 90 days to 140 days depending on the season. Itoh et al. (2020a) determined the light use efficiency in each experiment because CO2 concentration was not changed within one experiment and found a relationship similar to that in Figure 9: light use efficiency was low at low CO2 concentration and increased as CO2 concentration increased. A similar relationship between CO2 concentration and light use efficiency is also reported in sweet pepper (Watabe et al., 2021).
The transmission and attenuation properties of light in a plant canopy are expressed by light extinction coefficient. The light extinction coefficient: k, differs among plant species and cultivars. The k values range from 0.6 to 1.0 among tomato cultivars (Higashide, 2013; Higashide and Heuvelink, 2009; Higashide et al., 2012b).
The photosynthetic rate of an individual leaf depends on the photosynthetic characteristics and light intensity. In a canopy, light intensity depends on the position and on the degree of light transmission. Therefore, even if the photosynthetic characteristics of individual leaves (photosynthetic light–response curves) are the same, their photosynthetic rates and light use efficiencies may differ. An example is shown in Figure 10. Assuming two canopies with identical photosynthetic light–response curves and the same and sufficient LAI—in the canopy with a large light extinction coefficient k, much of the light is intercepted by leaves in the upper part of the canopy. Since light intensities are low for the leaves in the middle and lower layers of this canopy, the photosynthetic rates of these leaves are low, decreasing the photosynthetic rate of the entire canopy. Conversely, in the canopy with a small k, light reaches the leaves of the middle and lower layers. Thus, the photosynthetic rate of leaves in these layers is higher than in the canopy with a large k. As a result, the total amount of photosynthesis in the canopy with a small k is higher than in the canopy with a large k. Accordingly, light use efficiency may be affected by light extinction coefficient. LUE of the nine cultivars released over the past 50 years was significantly and negatively correlated with light extinction coefficient (Higashide and Heuvelink, 2009). As indicated by measurements and simulations, even for the same cultivar, the light extinction coefficients of canopies grown on slopes are smaller than those on flat land (Higashide, 2009a).
Light extinction coefficient affects canopy photosynthetic rate. Pg: photosynthetic rate.
To investigate the effects of the morphology of individual leaves on light interception and dry matter production, Higashide et al. (2017) compared the Japanese cultivar ‘Momotaro York’ and the Dutch cultivar ‘Gourmet’, which have different leaf shapes. For each cultivar, three types of leaf trimming were performed: (A) leaf tips were left intact (Untrimmed), (B) two leaflets were trimmed from the leaf tip when the leaf length was 5 cm or less (Young-trimmed), or (C) two leaflets were trimmed from the tip of a fully developed leaf at 71 days after transplanting (Mature-trimmed) (Fig. 11). The area of individual leaves was significantly larger in ‘Momotaro York’ than in ‘Gourmet’, whereas the photosynthetic light–response curves of both cultivars were almost the same (Higashide et al., 2017). In both cultivars, the area of individual leaves and LAIs decreased in Mature-trimmed (Higashide et al., 2017), whereas the light extinction coefficient of the canopy increased (Table 1). In Young-trimmed, the light extinction coefficient decreased in ‘Gourmet’ but increased in ‘Momotaro York’. The light use efficiency was not significantly affected by trimming in ‘Gourmet’ but was significantly decreased in Young-trimmed ‘Momotaro York’. The morphology of tomato leaves is affected by the relationships of KNOX genes and phytohormones (Ben-Gera et al., 2012; Blein et al., 2008; Burko et al., 2013; David-Schwartz et al., 2009; Jasinski et al., 2008; Kim et al., 2003; Shani et al., 2009; Shleizer-Burko et al., 2011). Based on these findings, the results of leaf tip trimming may be used for breeding to improve the light interception of canopies.
Tomato leaf shape altered by leaflet trimming, in the Japanese cultivar ‘Momotaro York’ and the Dutch cultivar ‘Gourmet’, at the end of the experiment. (A) Untrimmed; (B) Young-trimmed; (C) mature-trimmed. Arrows show the sites of trimming. First and second leaflets counting from the apex of each leaf were trimmed. Figure reproduced from Higashide et al. (2017) with permission from ASHS.
Effects of leaf tip trimming on light extinction coefficients and light use efficiency in the Japanese cultivar ‘Momotaro York’ and Dutch cultivar ‘Gourmet’.
Among Dutch tomato cultivars released in the last 50 years, modern cultivars have improved fresh and dry fruit yields (Higashide and Heuvelink, 2009). The increase in dry fruit yield is based on an increase in total dry matter production in the plant, which may be attributable to the accompanying increase in light use efficiency (Fig. 12A, B). Over this time period, there was no decrease in fruit dry matter content or increase in dry matter distribution to fruits (Higashide and Heuvelink, 2009). Therefore, these factors cannot explain the yield improvement of Dutch tomato cultivars. In the modern Dutch cultivars, the photosynthetic rate of individual leaves is increased, and the light extinction coefficient of the canopy is decreased, which seems to have contributed to the increase in light use efficiency (Fig. 12C, D).
Dry weight yield (A), light use efficiency (LUE: B), light extinction coefficient (C), and photosynthetic rate (D) of 8 tomato cultivars grown extensively in the Netherlands, i.e., ‘Moneymaker’ (◆), ‘Premier’ (■), ‘Extase’ (▲), ‘Sonatine’ (
), ‘Calypso’ (
), ‘Liberto’ (
), ‘Gourmet’ (◇), ‘Encore’ (Δ); and one Japanese tomato cultivar, ‘Momotaro Fight’ (○). Regression lines are based on data for the 8 Dutch cultivars only. Data from Higashide and Heuvelink (2009) with permission from ASHS. * Indicates significant correlation at P < 0.05.
In a review by Hay (1995), written from the perspective of breeding for yield improvement in various crops, the increase in yield of wheat (Triticum aestivum) and barley (Hordeum vulgare) over the last 100 years was attributed mainly to an increase in harvest index. Abbate et al. (1998) reported that the increase in dry weight of ears of wheat cultivars in Argentina was a result of additional distribution of dry matter to the ears. Saitoh et al. (1993) reported that the harvest index of modern rice cultivars (Oryza sativa) is higher than that of older cultivars. Morrison et al. (1999) reported that the yield of Canadian soybean (Glycine max) cultivars increased by 0.5% per year from the 1930s to the 1990s, due to an increase in harvest index, photosynthetic rate, and stomatal conductance, and a decrease in LAI. These reports indicate that yield improvement in many crops is primarily due to an increase in harvest index and not to an increase in total dry matter production.
Interestingly, Hay (1995) reported that the increase in the yield of corn (Zea mays) by breeding was due to an increase in total dry matter production, not in harvest index. The photosynthetic rate of individual leaves does not differ between modern and traditional cultivars in wheat and cotton (Abbate et al., 1998; Kuroda and Kumura, 1990a, b; Rosenthal and Gerik, 1991; Saitoh et al., 1990, 1993). These reports indicate that yield improvement in these crops was due not to an increase in the rate of individual leaf photosynthesis, but to changes in morphological factors such as the light extinction coefficient of the canopy. Tollenaar and Aguilera (1992) reported, however, that modern corn cultivars are more efficient in light use than older cultivars, and that there are no differences in light extinction coefficients or the amounts of light intercepted by the canopies between the old and modern cultivars.
Unlike wheat, cotton and rice, modern tomato cultivars released in the Netherlands have higher photosynthetic rates in individual leaves than older cultivars (Higashide and Heuvelink, 2009). According to Ho (1996), the difference in yield by tomato type, such as cherry, round, and beef tomato, is due to the difference in dry matter distribution to fruits. On the other hand, van der Ploeg et al. (2007) showed that, despite differences in the distribution of dry matter to fruits among cultivars, rates are not higher in modern than in older cultivars. They reported that the yield improvement of modern cultivars is due to an increase in total dry matter production resulting from an improvement in light use efficiency.
Tomato breeding is widespread in Japan, and many cultivars have been produced. However, the yield of greenhouse tomatoes in Japan has increased only slightly over the last 30 years (Ministry of Agriculture, Forestry and Fisheries of Japan, 2020). Tomato breeding in Japan prioritizes disease resistance and fruit quality, rather than yield improvement. In recent years, photosynthetic properties have been investigated to clarify why the yield of Japanese cultivars is lower than that of Dutch cultivars (Higashide and Heuvelink, 2009; Matsuda et al., 2011, 2013).
Higashide et al. (2012b) chose six Japanese cultivars released in the last 80 years, as reported by Aoki (1998) and Sumida et al. (2008), and compared their yield and fruit characteristics in a modern greenhouse. The cultivars used (the decade or year of release is indicated in parentheses) were ‘Sekai-ichi’ (1930s), ‘Fukuju No. 2’ (1940–50s), ‘Aichi Fast’ (1950–60s), ‘Kyoryoku-beijyu’ (1950–60s), ‘Momotaro’ (1985), and ‘Momotaro Colt’ (2003). Fruit yield was significantly higher in ‘Aichi Fast’ and ‘Kyoryoku-beijyu’ than in ‘Momotaro’ (Fig. 13). Fruit dry matter content was larger in ‘Fukuju No. 2’ than in the other cultivars, and larger in ‘Aichi Fast’ and ‘Momotaro Colt’ than in ‘Kyoryoku-beijyu’ and ‘Momotaro’ (Higashide et al., 2012b). Thus, yield and fruit dry matter content have not improved in the course of tomato breeding in Japan.
Fruit yield of 6 Japanese tomato cultivars (S, ‘Sekai-ichi’; F, ‘Fukuju No. 2’; K, ‘Kyoryoku-beijyu’; A, ‘Aichi Fast’; M, ‘Momotaro’; MC, ‘Momotaro Colt’) released over the past 80 years. The same letters indicate no significant difference (n ≥ 12; P < 0.05 by Tukey’s multiple-comparison test). Figure reproduced from Higashide et al. (2012b) with permission from ASHS.
The positive correlation between dry matter content and soluble solids in tomato fruits is well known (Stevens and Rudich, 1978). However, Higashide et al. (2012b) found no such correlation in the six old and modern Japanese cultivars listed in the previous section. Fruit water content, the index opposite to fruit dry matter content, was also not significantly correlated with soluble solids in all six cultivars. But a strong negative correlation was found when ‘Momotaro’ and ‘Momotaro Colt’ data were excluded (Fig. 14). The fruit soluble solid contents of ‘Momotaro’ and ‘Momotaro Colt’ were considerably higher than those of the other four cultivars, indicating that the composition and deliciousness of tomato fruit improved with the ‘Momotaro’ release.
Association between fruit water and soluble solid contents of six Japanese tomato cultivars (S, ‘Sekai-ichi’; F, ‘Fukuju No. 2’; K, ‘Kyoryoku-beijyu’; A, ‘Aichi Fast’; M, ‘Momotaro’; MC, ‘Momotaro Colt’). Linear regression excludes M and MC. Figure reproduced from Higashide et al. (2012b) with permission from ASHS.
The high water and soluble solids contents in ‘Momotaro’ and ‘Momotaro Colt’ indicate that the fruits are juicy and sweet. Although water and soluble solids contents in fruits are generally negatively correlated, Kawabata et al. (2002) reported that changes in sugar concentrations were inconsistent with changes in dry matter content under conditions of restricted fruit growth mechanically by acrylic sleeves, and that sugars in fruits may accumulate as water-soluble forms. However, the fruit growth of ‘Momotaro’ and ‘Momotaro Colt’ was not restricted in Higashide et al. (2012b). By these cultivar characteristics, the contents of soluble carbohydrates and/or water in the fruits of ‘Momotaro’ and ‘Momotaro Colt’ may be higher than in the other cultivars without growth restriction. This might suggest that the nutrient composition and palatability of tomato fruits changed after the development of ‘Momotaro’.
Despite an increase in light use efficiency (Fig. 9; Itoh et al., 2020a; Saito et al., 2020b) and in total dry matter production with increasing CO2 concentrations, the increase in fruit yield is lower in ‘Momotaro York’ than in the other cultivars (Higashide et al., 2015). Yasuba et al. (2011) also detected differences in the rate of increase of tomato yield among cultivars under a combination of CO2 enrichment and fogging system. In corn, Rogers and Dahlman (1993) reported an increase in total dry matter production and dry matter distribution to fruits by CO2 enrichment. Tripp et al. (1991) reported an increase in dry matter distribution to tomato fruits by CO2 enrichment in seven cultivars. Nederhoff (1994) reported an increase in dry matter distribution to fruits by CO2 enrichment in cucumbers and sweet peppers, but not in tomatoes. Similarly, Higashide et al. (2015) found no effect of CO2 application with humidification on the percentage of dry matter distributed to fruits in ‘Asabiyori 10’ and ‘Aichi Fast’, and a decrease in ‘Momotaro York’ (Fig. 15A). The distribution of dry matter to fruits is determined mainly by sink strength, which is strongly affected by the number of fruits per truss (Heuvelink, 1996; Heuvelink and Marcelis, 1989). Higashide et al. (2015) reported that CO2 enrichment had no effect on the number of fruits per truss but increased the number of trusses with immature fruits per plant from 6 to 7 in ‘Momotaro York’, whereas the latter effect was not observed in ‘Asabiyori 10’ or ‘Aichi Fast’ (Fig. 15B). It was assumed that the assimilation products due to the increases in leaves and inflorescence were insufficient to allocate to the increased number of fruit, and thus reduced the distribution of dry matter to fruits in ‘Momotaro York’.
Fraction of dry matter allocated to fruits (A), and number of trusses with immature fruits per plant (B) of three Japanese cultivars (Ab, ‘Asabiyori 10’; Af, ‘Aichi Fast’; My, ‘Momotaro York’) grown under ambient CO2 without fogging (AMC) or elevated CO2 with fogging (ECF) at 171 or 165 days after transplanting, respectively. The results of two-way ANOVA (n = 12) with environmental conditions (Env) and cultivar (Cv) as independent variables and their interaction (Env × Cv) for each dependent variable are shown in each panel. NS: not significant (P ≥ 0.05). The same letters indicate no significant difference (P < 0.05; by Tukey’s multiple-comparison test). Figure reproduced from Higashide et al. (2015) with permission from JSHS.
Grafting is widely used in Japan to prevent soil-borne diseases, but it has not been frequently used in hydroponics because of the low risk of such diseases. However, in recent years, rootstock cultivars from the Netherlands have been used in Japan. The type of rootstock reportedly affects not only disease avoidance, but also yield (Khah et al., 2006; Mohammed et al., 2009; Oda et al., 1996; Sada, 1981; Turkmen et al., 2010). Masuda and Furusawa (1991) reported that the quality of tomatoes was improved by Japanese rootstock cultivars, but yield did not change. Flores et al. (2010) and Savvas et al. (2011) reported that the yield and soluble solids contents of tomatoes were increased by using a salt-resistant rootstock under salt stress. Fan et al. (2011) compared rootstock cultivars in China and the United States under salt stress conditions and found no effect on yield. Despite these studies, the mechanism for the effect of rootstock on yield has not been clarified.
To improve yields of Japanese cultivars, Higashide et al. (2014a) investigated the effects of a Dutch interspecific rootstock (S. lycopersicum × S. habrochaites) on dry matter production and yield. The Dutch cultivar ‘Gourmet’ (G) and the Japanese cultivar ‘Momotaro York’ (My) were self-grafted (G/G; My/My: scion/rootstock), and grafted on the Dutch rootstock cultivar ‘Maxifort’ (Mx; De Ruiter Vegetable Seeds, the Netherlands) (G/Mx; My/Mx), or the on the Japanese rootstock cultivar ‘Spike 23’ (S; S. lycopersicum; AiSan Seed Co., Ltd., Aichi, Japan) (G/S; My/S). The fruit yield of My/Mx was significantly higher than that of My/My, whereas fruit dry matter content was not affected (Figs. 16 and 17). For both scions ‘(Gourmet’ and ‘Momotaro York’), the difference in yield among rootstocks was mainly due to the difference in dry fruit yield. The rootstocks did not affect the maximum photosynthetic rate of the scions or dry matter distribution to fruits. Total dry matter production was higher in the combinations of scion/rootstocks with higher light use efficiency (Fig. 18). Therefore, the difference in total dry matter production was possibly due to the difference in light use efficiency, which also resulted in the difference in yield.
Fresh fruit yield of plants from reciprocal grafting between Dutch and Japanese cultivars at 126 days after transplanting. Dutch cultivars: G, ‘Gourmet’; Mx, ‘Maxifort’; Japanese cultivars: My, ‘Momotaro York’; S, ‘Spike 23. The same letters indicate no significant difference within the same scion cultivar (P < 0.05; n = 12; by Bonferroni post-hoc test). Figure reproduced from Higashide et al. (2014a) with permission from JSHS.
Total fruit dry weight (bars), and dry matter content of mature fresh fruits (points and lines), of tomato plants from reciprocal grafting between Dutch and Japanese cultivars, measured at 126 days after transplanting. Dutch cultivars: G, ‘Gourmet’; Mx, ‘Maxifort’; Japanese cultivars: My, ‘Momotaro York’; S, ‘Spike 23’. Both immature and mature fruits were included in the measurement of fruit dry weight. The same letters indicate no significant difference within the same scion cultivar (P < 0.05; n = 12; by Bonferroni post-hoc test). Figure reproduced from Higashide et al. (2014a) with permission from JSHS.
Association between total aboveground dry matter and light-use efficiency. Dutch cultivars: G, ‘Gourmet’, Mx, ‘Maxifort’; Japanese cultivars: My, ‘Momotaro York’, S, ‘Spike 23’. Slashes show grafting: scion/rootstock. *** P < 0.001. Figure reproduced from Higashide et al. (2014a) with permission from JSHS.
Crop growth models can help in pre-season and in-season management decisions on cultural practices, control of greenhouse climate, fertilization, and irrigation (Boote and Scholberg, 2006; Marcelis et al., 2009; Spitters et al., 1989). The models are also expected to be used to improve cultivation management technology, labor management, and sales planning. In growth models for greenhouse tomatoes such as TOMGRO and TOMSIM (Bertin and Heuvelink, 1993; Heuvelink, 1995; Jones et al., 1991), observed LAI values are used to predict canopy assimilation and dry matter production. Using such relational expressions, dry matter predictions can be used to extrapolate fruit fresh weight. In addition to TOMSIM and TOMGRO, several other models of tomato growth have been reported and compared (Boote and Scholberg, 2006; Kuijpers et al., 2019). Higashide et al. (2011) reported that a dry matter production model can be used to optimize nitrate fertilizer application. Higashide et al. (2014b) analyzed the effects of auxin biosynthesis inhibitors on tomato seedlings by using a growth model that estimated leaf elongation and dry matter production. Mochizuki et al. (2019) developed a model of elongation growth and dry matter production to clarify the effect of ultrafine bubbles in nutrient solution on the growth of tomato seedlings. Using the model, they succeeded in predicting plant growth under an ultrafine bubbles solution with high accuracy.
Environmental control in a greenhouse is important to improve crop yield. Only 1.6% of total greenhouse area in Japan was managed by advanced environmental control systems in 2014, and this increased to just 2.7% in 2018 (Ministry of Agriculture, Forestry and Fisheries, 2018). Although the total area with installed advanced environmental control systems is steadily increasing, not many greenhouses have these systems yet. The issues in installing environmental control systems are the initial cost, the lack of knowledge and experience in using the equipment, and the difficulties in choosing the correct settings and managing the system. The environmental control equipment in a greenhouse includes ventilation windows, a heating system, heat pumps, circulation fans, a fogging system, light shielding curtains, heat insulation curtains, a CO2 applicator, and irrigation and hydroponics equipment, all of which operate on the basis of data measured by sensors. Although growers may expect yield improvement and labor-saving by installation of environmental control equipment, more expensive equipment is more difficult to use and requires more technical knowledge to operate. The number of settings to be controlled is as large as ca. 300, and the optimum value of each setting depends on the environment. Therefore, it is not easy to perform appropriate control only on the basis of experience and intuition and this is one of the reasons preventing the installation of more environmental control systems in Japan.
Saito et al. (2020b) reported predictions of growth and yield of greenhouse tomatoes from non-destructive measurements. A flow diagram of growth and yield prediction is shown in Figure 19. First, leaf appearance rate and the resulting number of leaves are determined on the basis of daily average temperature. Leaf area per plant is obtained from leaf number and individual leaf area, and LAI is obtained by multiplying leaf area per plant-by-plant density. Individual leaf area is calculated as leaf length × leaf width × coefficient. The differences in individual leaf area among cultivars may be expressed by the differences in the length, width, and coefficients. Individual leaf area increases as the leaves grow, but leaf expansion stops at a certain size. The individual leaf area at this stage and the time needed to reach it are largely determined by cultivar. Since individual leaf areas and LAIs cannot be measured by destructive sampling in commercial production, they must be estimated in a non-destructive manner from the mean individual leaf area of each cultivar, and used as inputs for the prediction (Saito et al., 2020b). At present, the data on leaf number and leaf area cannot be obtained from sensors or images and have to be obtained manually. In the future, acquisition of these data is likely to be automated for practical use.
Flow diagram for calculation of yield and dry matter production. Temp., daily average air temperature; LAI, leaf area index; PAR, photosynthetically active radiation; LUE, light use efficiency; DW, dry weight; FW, fresh weight. Figure reproduced from Saito et al. (2020a) with permission from JSHS.
Saito et al. (2020b) determined the amount of light intercepted by plant canopy from LAI and measured PAR. Light use efficiency is obtained from a function formula prepared in advance for each species and cultivar. As mentioned above, light use efficiency is a function of daytime CO2 level. Daily dry matter production is calculated by multiplying the amount of intercepted light by light use efficiency. Total dry matter production is obtained by integrating daily dry matter production. Dry fruit yield is obtained by multiplying total dry matter production by the ratio of dry matter distribution to fruits, and fresh fruit yield is obtained by dividing dry fruit yield by fruit dry matter content.
The ratio of dry matter distribution to fruits and fruit dry matter content are largely determined by the species and cultivar. These values measured in advance are used as coefficients. From environmental data such as temperature, solar radiation, and CO2 concentration, leaf appearance rate, amount of intercepted light, and light use efficiency are calculated by functional expressions. The coefficients of these functional expressions differ depending on the species and cultivar, and the characteristics of each crop and cultivar may be expressed by these coefficients. By directly inputting crop information such as the number of leaves and leaf area, corrections can be made at various time points, and dry matter production and yield of the observed plants can be predicted as outputs.
Experimental and analytical modeling similar to that of Saito et al. (2020a) has been carried out in low-truss tomato cultivation. With the same approach, Itoh et al. (2020a) conducted six low-truss tomato cultivation experiments and used one of them to verify the models produced in the other five experiments. Those authors found that the predicted total dry matter production was almost equal to that observed when CO2 concentration ranged from about 400 μmol·mol−1 to 650 μmol·mol−1. Thus, yield prediction by modeling is also useful in low-truss tomato production. Similarly, Watabe et al. (2021) modeled dry matter production in sweet pepper in a greenhouse. On the basis of the analysis of yield components, growth, and photosynthetic characteristics (Higashide and Heuvelink, 2009), they pointed out that improvement of light use efficiency is important for yield improvement. The authors suggest their model may be useful for growth simulation and yield improvement in sweet peppers, similar to the use of the model for yield prediction and improvement in tomatoes (Saito et al., 2020a).
The analysis and modeling of dry matter production are also used in high-Brix tomato production. These tomatoes, which are peculiar to Japan and have a high content of soluble solids, are produced commercially and sold at a high price. Several cultivars have intrinsically high Brix, but many other cultivars are stressed by decreasing irrigation or adding salt to a nutrient solution to increase the content of fruit Brix. However, commonly, the yield decreases drastically, to 1/3 or less that of standard production without stress.
Itoh et al. (2020b) reported an evaluation of this type of salinized low-truss cultivation. The electrical conductivity of the applied nutrient solution was gradually increased from 0.6 to 3.5 dS·m−1 in the salt-stressed group and from 0.6 to 2.0 dS·m−1 in the non-stressed group; the nutrient solutions were otherwise managed in the normal manner. Fresh fruit yield was significantly lower in the salinized group than in the non-stressed group, whereas fruit Brix was significantly higher in the salinized group than in the non-stressed group, and individual fruit weight tended to be lower in the salinized group (Table 2). In all three experiments, there was no significant difference in light use efficiency between the two groups (Fig. 20). In other words, there was no difference in the efficiency of dry matter production between the salt-stressed and non-stressed groups, but the fruit weight per fruit was lower in the salt-stressed group, and fruit dry matter content was increased by the high soluble solids. Since an increase in fruit dry matter content reduces fresh weight yield (Fig. 1), it can be assumed that yield decrease in the salt-stressed group resulted from the increase in fruit dry matter content.
Effect of salinity treatment on fruit Brix and fruit fresh weight.
Total aboveground dry matter production as a function of cumulative PAR intercepted by tomato plants subjected to salinized or non-salinized treatment. Light use efficiency values (95% confidence intervals; treatment) were 2.07 (1.99–2.15; non-salinized) and 2.13 (2.02–2.26; salinized) in Experiment 1 (A); 3.17 (3.02–3.32; non-salinized) and 3.27 (3.09–3.44; salinized) in Experiment 2 (B); and 3.27 (3.02–3.53; non-salinized) and 3.51 (3.30–3.72; salinized) in Experiment 3 (C). Regression lines with the same letter are not significantly different at 95% confidence interval (n = 16–22). ***Significant correlation at P < 0.001. Figure reproduced from Itoh et al. (2020b) with permission from JSHS.
Dry matter production increases with an increase in leaf area, but it is inefficient when leaf area is too large or too small. Saito et al. (2020b) recommended LAIs based on the photosynthetic properties of the cultivars (Fig. 21), which correspond to the highest dry matter production at a certain solar radiation. The recommended LAIs are small at low solar radiation and large at high solar radiation (Fig. 21). The strength of solar radiation varies daily depending on the weather and also differs greatly by season. Saito et al. (2020a) estimated target LAIs from average solar radiation and corresponding recommended LAI for each season. During their experiment, leaf area was managed to adjust it to the target LAIs by planned leaf removal, which depended on the leaf appearance rate. Changes in the target and observed LAIs in each cultivar during the experiment are shown in Figure 22.
Light-response curves of photosynthesis as a function of photosynthetic photon flux density (PPFD) (A), and recommended leaf area index (LAI) as a function of daily cumulative solar radiation (B) from June 2016 to March 2017. Rn, ‘Ringyoku’; My, ‘CF Momotaro York’; Mn, ‘Managua’. Figure reproduced from Saito et al. (2020b) with permission from JSHS.
Target, predicted, and observed leaf area index (LAI) for three tomato cultivars: Rn, ‘Ringyoku’; My, ‘CF Momotaro York’; Mn, ‘Managua’. Average values ± standard error (bars) shown for observed LAIs (38 and 206 days after transplanting [DAT]: n = 4; 325 DAT: n = 6). Figure reproduced from Saito et al. (2020a) with permission from JSHS.
Since light use efficiency increases with CO2 concentration, it is preferable to always maintain high CO2 concentration in the greenhouse for yield improvement. However, in summer, the ventilation windows of the greenhouse are always fully opened to reduce air temperature and even if CO2 is applied, a considerable part of it escapes outside. Therefore, to apply CO2 efficiently, Saito et al. (2020a) changed the CO2 concentration target in the greenhouse according to the season.
To improve yield of greenhouse tomatoes by managing LAI and CO2, Saito et al. (2020a) used a model for prediction of growth and dry matter production and carried out an experiment for one year using three cultivars: ‘Ringyoku’ (National Agriculture and Food Research Organization, Tsukuba, Japan), the popular Japanese cultivar ‘CF Momotaro York’ (Takii & Co., Ltd., Kyoto, Japan), and the Dutch cultivar ‘Managua’ (Rijk Zwaan, the Netherlands (Fig. 23). At 38 days after transplanting, predicted total dry matter production was within the standard error range of the observed values in all cultivars. At 206 and 325 days after transplanting, it was within the standard error range of the observed values in ‘CF Momotaro York’ and ‘Managua’, but was larger than the observed values in ‘Ringyoku’. Annual fresh fruit yield was predicted to be 59.7 kg·m−2 in ‘Ringyoku’, 53.8 kg·m−2 in ‘CF Momotaro York’, and 59.2 kg·m−2 in ‘Managua’, whereas observed yield was 55.6 ± 2.8 kg·m−2 (mean ± standard deviation) in ‘Ringyoku’, 50.6 ± 4.3 kg·m−2 in ‘CF Momotaro York’, and 60.7 ± 5.5 kg·m−2 in ‘Managua’. Thus, the observed yield of ‘Ringyoku’ was lower than predicted yield, but the predicted yields of ‘CF Momotaro York’ and ‘Managua’ were within the standard deviation of the observed yield. Thus, using this model, Saito et al. (2020a) succeeded in predicting the yields of these two cultivars, which allowed them to dramatically increase their yield while meeting the taste requirements of Japanese consumers. The annual yield of tomatoes with Brix of at least 5° was over 50 kg·m−2 for ‘CF Momotaro York’ (Fig. 24; Higashide, 2020). In the near future, the growth models show promise for growers to facilitate yield improvement and crop management in greenhouses.
Predicted and observed total dry matter production in three tomato cultivars: Rn, ‘Ringyoku’; My, ‘CF Momotaro York’; Mn, ‘Managua’. Average values ± standard error (bars) shown for observed total dry matter production (38 and 206 DAT, n = 4; 325 DAT, n = 6). Figure reproduced from Saito et al. (2020a) with permission from JSHS.
Average yield of Japanese tomato cultivars and yield of ‘CF Momotaro York’ obtained using model-based management. Figure reproduced from Higashide (2020) with permission from JSHS.
