1. Introduction
With climate change, temperature increases are not only expected in the coming decades, but have already been observed [
1]. Of particular concern is that average temperature increases could compromise the world’s rice production. Masutomi et al. [
2] predicted rice yield losses for most of the rice growing regions in Asia in the near future, and Lobell et al. [
3] identified rice in Southeast Asia as one of the most important crops in need of climate change adaptation investment. Whereas the increase of average temperature and its negative effects on rice yields are widely accepted, there is no consensus about the effects of changes to the temperature range, the difference between day and night temperature on a daily basis. Despite high uncertainty [
1], a decrease in the daily temperature range is expected in the world’s rice growing regions [
4] as daily minimum temperatures increase more rapidly than daily maximum temperatures [
5]. Negative yield responses of rice to a narrowing daily temperature range due to an increase in minimum temperature have been observed in field experiments in the Philippines [
6]. However, Lobell [
4] projected impacts of a narrowing daily temperature range on rice yield as positive, even if relatively small.
As shown by the contradictory results of the two studies, the role of night temperature per se is not well understood [
7]. Experimental approaches aiming at exploring the effect of high night temperature on plant growth face the difficulty that only increasing night temperature means an increase in daily mean temperature, which also has an effect on plant growth and development, making it difficult to discriminate between both effects. To exclude the effects of a higher mean temperature, varying daily temperature ranges around the same mean temperature can be used to study plant responses. In such an experimental setup, the possibility of including high night temperatures is limited, since this would require a high mean temperature, which would lead to heat stress when using a large daily temperature range. To avoid heat stress effects during the day, but to test plant responses to a wide range of night temperatures, day and night temperature can be inverted, an approach that has been used in this study.
Another constraint of studies on high night temperature under natural conditions is the physical interaction of temperature and relative air humidity. Since minimum night temperature is usually around the dew point, a smaller daily temperature range is directly correlated with higher relative air humidity during the day. However, changes in relative air humidity have been associated to changes in photosynthetic rates [
8], dry matter and leaf area [
9], and spikelet sterility [
10] and thus, the potential effect of relative air humidity on the rice plant should not be disregarded in studies on temperature effects.
In a field study, high night temperature led to biomass reduction in one of the two rice genotypes [
11], whereas Peraudeau et al. [
7] reported no changes in total dry matter, increased leaf area for indica varieties in one out of the three experiments, but consistently lower SLA (specific leaf area) under elevated night temperature. In a meta-analysis, Jing et al. [
12] found that plant leaf growth was increased by high night temperature as well as leaf area ratio (LAR) and specific leaf area (SLA), while the effect on organ weight and total dry matter was less clear, and varied between plant functional groups. As a result, it was concluded that complexities and challenges remain when seeking general patterns of plant growth in response to high night temperature. Little information exists on the effect of temperature on partitioning in rice plants. In a climate chamber experiment, high night temperature during reproductive growth led to increased dry weight per hill, because of a higher stem weight, whereas leaf and root dry weight was not affected by temperature [
13]. Increased daily temperature in open-top chambers in the field led to higher leaf area, but the effects on partitioning of dry weight between organs was not consistent [
14]. In summary, there seems to be a larger effect of temperature on leaf area than on total dry matter, but the impact on dry matter partitioning between organs remains unclear. In all cited studies, temperature was raised during the night, in one case during day and night, with the result of a higher daily mean temperature compared to the control treatment. Hence, effects of day, night, and mean temperature cannot be distinguished.
As plants must achieve a balance between carbon assimilation (occurring during the day) and storage and growth (occurring during both day and night) [
15], there are likely differential effects of day and night temperature on carbon allocation. For example, high night temperature could enhance growth processes during the night, which in turn could induce the buildup of more photosynthetic tissue to meet the increased demand. Furthermore, high day temperature results in a higher water demand, which in turn could favor root growth. However, these processes have hardly been studied yet, even though greater insight into them could help predict the consequences of Climate Change for rice production. Therefore, the objective of this study is to disentangle effects of day, night, and mean temperature on dry matter and its partitioning between organs and leaf area, while also taking into account the effects of relative air humidity by using inverted day/night temperatures and air humidities.
2. Results
Rice plants were grown at three different temperature regimes with either “natural” (30 °C/20 °C; Tnat), constant (25/25 °C; Tcon), or inverted (20/30 °C; Tinv) day/night temperature in combination with three different relative air humidity (RH) regimes with either “natural” (40/90%; RHnat), constant (65/65%; RHcon), or inverted (90/40%; RHinv) day/night RH. Two additional treatments with either constantly low (20/20 °C; Tcon-l) or high (30/30 °C; Tcon-h) temperature, both at RHcon, were established. After two weeks of different day/night temperature and RH treatment with the same mean temperature and RH, total plant dry matter varied between 6.0 g for Tnat/RHcon and 10.4 g for Tnat/RHnat (
Table 1).
Temperature had no statistically significant effect (
Table 2), but RHnat led to a significantly higher total dry matter per plant with an average 9.7 g versus 7.8 g and 8.1 g under RHcon and RHinv, respectively. At Tnat, RHnat led to a statistically significant higher total dry matter compared to RHcon, whereas RH had no statistically significant impact in the other temperature treatments.
Much larger variation was observed in leaf area, which ranged from 333 cm
2 under Tcon/RHinv to 890 cm
2 under Tinv/RHcon. Temperature only had a significant effect under RHcon, with the highest leaf area at Tinv (890 cm
2), and the lowest at Tnat (341 cm
2). RH had a larger effect on leaf area than temperature, with RHinv leading to a significantly lower leaf area of 356 cm
2 on average in comparison to 621 cm
2 and 596 cm
2 under RHnat and RHcon, respectively. Under constant temperature and relative humidity, both total dry matter and leaf area were significantly lower at 20 °C (2.9 g; 162 cm
2) than at 25 °C (7.9 g; 558 cm
2) and 30 °C (9.2 g; 439 cm
2) (
Table 3).
Leaf mass fraction (LMF), the leaf dry weight per plant dry weight, did not respond to temperature, but RH had a statistically significant effect (
Figure 1).
With an average of 0.20 g g
−1, the LMF was significantly lower under RHinv than under RHnat with 0.24 g g
−1 and RHcon with 0.27 g g
−1 (
Figure 1). Under the constant temperature regimes, LMF was significantly lower at 30 °C (0.19 g g
−1), than at 20 °C (0.23 g g
−1), and 25 °C (0.25 g g
−1) (
Figure 1, Insert 1). Temperature had a significant effect on the stem mass fraction (SMF) (
Figure 1, Insert 2), which is the fraction of stem dry weight per plant dry weight. Under Tinv, it was significantly lower (0.37 g g
−1, on average) than under Tnat and Tcon, which both resulted in a SMF of 0.41 g g
−1. Also RH had an effect on SMF, which was significantly lower (0.38 g g
−1) under RHcon than under RHnat (0.40 g g
−1) and RHinv (0.41 g g
−1). Constant temperature between 20 and 30 °C did not have any significant effect on SMF.
The root mass fraction (RMF), the root dry weight per plant dry weight, varied between 0.31 g g−1 for Tnat/RHnat and Tnat/RHcon and 0.38 g g−1 for Tinv/RHnat and Tinv/RHinv and was strongly influenced by temperature. Under Tnat, RMF was on average 0.31 g g−1 and significantly lower than under Tcon and Tinv with 0.34 g g−1 and 0.37 g g−1, respectively. RH as well as different constant temperatures between 20 and 30 °C had no effect on RMF. The faction of dead leaves (DLMF) varied between 0.00 g g−1 for Tinv/RHcon and 0.06 g g−1 for Tnat/RHcon and Tnat/RHinv. Nevertheless, there were no significant differences because of large variation within treatments.
The SLA varied largely and ranged from 19.7 m
2 kg
−1 (Tnat/RHinv) to 32.5 m
2 kg
−1 (Tinv/RHcon) (
Figure 2).
Both temperature and RH had large effects on SLA. On average, Tnat led to the lowest SLA with 22.6 m2 kg−1 followed by Tcon with 25.9 m2 kg−1 and Tinv with 28.8 m2 kg−1. Among the RH treatments, SLA was significantly lower under RHinv with 22.8 m2 kg−1 than under RHnat and RHcon with 26.6 and 28.0 m2 kg−1, respectively. Further, a significant interaction effect between temperature and RH was found. Under Tnat, RHnat resulted in a significantly higher SLA compared to RHinv, whereas under Tcon and Tinv, RHcon resulted in a significantly higher SLA compared to RHinv. At RHnat, temperature had no significant effect on SLA, while at RHcon, Tinv and Tcon resulted in a significantly higher SLA than Tnat and at RHinv, Tinv resulted in a significantly higher SLA than Tnat. Among the constant temperature treatments, no significant difference was found in regard to SLA.
Regressing SLA and plant organ fractions versus day and night temperature and RH of all treatments, including the different constant temperatures, resulted in a significant positive correlation between SLA and night temperature (
r = 0.65;
p = 0.030) (
Table 4).
The stem mass fraction was positively correlated with day temperature (r = 0.76; p = 0.006) and the root mass fraction was negatively correlated with day temperature (r = −0.84; p = 0.001) and positively with night temperature (r = 0.63; p = 0.037). Leaf and dead leaf mass fractions were not correlated with any of the parameters.