*Article* **Responses of the Lodging Resistance of** *Indica* **Rice Cultivars to Temperature and Solar Radiation under Field Conditions**

**Xiaoyun Luo 1,2, Zefang Wu <sup>3</sup> , Lu Fu 1,2, Zhiwu Dan 1,2, Weixiong Long 1,2, Zhengqing Yuan 1,2, Ting Liang 1,2 , Renshan Zhu 1,2, Zhongli Hu 1,2 and Xianting Wu 1,2,3,\***


**Abstract:** Much attention has shifted to the effects of temperature and solar radiation on rice production and grain quality due to global climate change. Meanwhile, lodging is a major cause of rice yield and quality losses. However, responses of the lodging resistance of rice to temperature and solar radiation are still unclear. To decipher the mechanisms through which the lodging resistance might be affected by temperature and solar radiation, 32 rice cultivars with different lodging resistance were grown at two eco-sites on three sowing dates over a period of three years. Based on the field observation, 12 *indica* rice cultivars which did not lodge were selected for analysis. Significant differences were found in the lodging resistance of the *indica* rice cultivars at different temperature and solar radiation treatments. The results showed that temperature was the main factor that affected the lodging resistance of *indica* rice cultivars under the conditions of this study. With the increased average daily temperature, the lodging resistance decreased rapidly, primarily due to the significant reduction in physical strength of the culm, which was attributed to the longer and thinner basal second internode. Among the 12 *indica* rice cultivars, the lodging-moderate cultivar Chuanxiang 29B was most sensitive to temperature, and the lodging-resistant cultivar Jiangan was least responsive to temperature. These results suggested that rice breeders could set the shorter and thicker basal internode as the main selection criteria to cultivate lodging-resistant *indica* cultivars to ensure a high yield at a higher ambient temperature.

**Keywords:** temperature; solar radiation; sowing date; growth duration; lodging resistance; lodging index; lodging-related traits; *indica* rice

### **1. Introduction**

Lodging severely reduces the grain yield and quality of rice [1]. Furthermore, it increases production costs by adversely affecting the harvest manipulations and heightening the grain drying demand [2,3]. According to a study by Nakajima et al. [4], rice lodging could also aggravate mycotoxin pollution that threatens animal and human health. Since the initiation of the "Green revolution" in the 1960s, semi-dwarf cultivars of rice and wheat have been developed, which have enhanced the lodging resistance significantly and increased global grain production [5–7]. However, with the large-scale cultivation of high-yielding cultivars, extensive use of fertilizers, and simplified planting techniques, such as direct-seeding, the potential risk of lodging has increased in recent years [8–10]. Thus, lodging-resistant cultivars have been developed as a genetic improvement strategy to increase the yield of rice, wheat, and other crops [11–14].

Lodging, which results from a loss of balance in plant bodies, refers to the lasting vertical stem displacement of plants [15]. In the case of rice, there are three types of lodging: culm bending, culm breaking and root lodging [3,16]. Culm breaking is generally seen at the lower internodes (including the third and fourth internodes from the plant top),

**Citation:** Luo, X.; Wu, Z.; Fu, L.; Dan, Z.; Long, W.; Yuan, Z.; Liang, T.; Zhu, R.; Hu, Z.; Wu, X. Responses of the Lodging Resistance of *Indica* Rice Cultivars to Temperature and Solar Radiation under Field Conditions. *Agronomy* **2022**, *12*, 2603. https:// doi.org/10.3390/agronomy12112603

Academic Editor: Roberto Barbato

Received: 30 August 2022 Accepted: 20 October 2022 Published: 23 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

which happens when the bending moment of the upper plant part is excessive [2,17]. The manner in which rice resists against culm breaking is usually assessed by the lodging index (LI) [18–21]. A decrease in the LI indicates a stronger lodging resistance capability. Several studies have investigated the correlation of LI with lodging-related traits [22–24]. However, as the LI is a ratio of the bending moment of the whole plant (BM) to the bending moment at breaking (M), it is inadequate for evaluating the lodging resistance capability under certain special conditions. For example, when the multiple differences in BM values between two cultivars are similar to the multiple differences in M values between the two cultivars, the calculated LI values of the two cultivars would exhibit no significant difference, which could be contradictory to the actual lodging resistance of the rice cultivars. As a result, an optimized parameter 4BM (equal to the value of 2M minus the value of BM), which was defined as the external force that the basal second internode could withstand, was proposed to be used along with the LI for a further accurate evaluation of the lodging resistance [25]. An increase in the 4BM indicates a stronger lodging resistance capability. Lodging generally happens at the stage of grain filling [26,27]. In the research of Ichii and Hada [28], the stem-breaking strength decreased to the minimum value and the LI increased to the maximum at 30 days after heading, which indicated that the grain-filling stage is the period when lodging often occurs. Lodging is associated with several biotic and abiotic factors: the height and weight of the plant, the length of the panicle, the plumpness of the leaf sheath, as well as the length, diameter and thickness of basal internodes influence the lodging resistance of rice [29–34]. Regarding the morphological factors, the external diameter and thickness of the cross-section from basal internodes strongly influence the breaking strength of the stem [21,35]. Additionally, the stem contents of soluble sugars, K, Si, cellulose, starch and lignin affect the basal stem-breaking strength pronouncedly [36–40]. However, some researchers have suggested that the basal stem-breaking strength is decided by structural carbohydrates (lignin and cellulose) rather than non-structural carbohydrates (soluble sugars and starch) [20,41]. Growth conditions, such as the application of different fertilizers, planting density, direct seeding methods, and sheath blight attacks, strongly affect the lodging resistance of rice plants [10,19,42–45]. Lodging is also correlated with varying environmental parameters, such as rain, wind, CO2, deep water, and resource complementarities [46–50].

With the increase in depletion of the stratospheric ozone, atmospheric levels of greenhouse gases, land-use alterations and aerosol outputs, an increase in global temperatures (global warming) and a decrease in solar radiation in Asia have been recorded in recent decades [51–54]. In the last century, the average global surface temperature recorded an elevation by 0.5 ◦C, and its estimated range of elevation is 0.3–6.4 ◦C by the end of this century [55]. An average annual reduction of 0.51 <sup>±</sup> 0.05 W m−<sup>2</sup> in solar radiation in Asia has been reported [56,57]. Many studies have shown a significant influence of the increase in global temperatures on the yield and quality of rice grains [58,59]. Additionally, the positive role and significance of solar radiation in rice grain output have also been shown [60,61]. However, few studies have considered the influences imposed on rice lodging resistance by the solar radiation and temperature variations. One study reported that an increase in the soil temperature increased the lodging risk of rice plants [18]. The low solar radiation reduced the physical strength of the stem and, thus, increased lodging susceptibility in rice [62].

More than half of the global population consumes rice as a staple food. Since lodging, high temperature, and low solar radiation have detrimental effects on rice production, understanding the effects of temperature and solar radiation on lodging is important for growing rice that is adapted to the changing global climate. In the present work, we chose two eco-sites to carry out 3-year field experimentations on three sowing dates per year, to grow rice at different temperatures and under different solar radiation treatments. A total of 32 rice cultivars with different lodging resistance capabilities were evaluated under different combinations of temperature and solar radiation. Among them, 12 *indica* rice cultivars, which did not lodge in all the sowing dates based on the field observation,

were selected for analysis. The objectives of the present study were to: (1) investigate the responses of the lodging resistance of *indica* rice cultivars to different temperature and solar radiation treatments; (2) evaluate the most and least affected cultivars among the 12 *indica* rice cultivars under different temperature and solar radiation treatments; (3) explore the relationship of the morphological, mechanical, and biochemical characteristics associated with lodging resistance with temperature and solar radiation. To cope with global climate change, a greater understanding of the climatic impact on lodging in *indica* rice will provide guidelines for rice breeders to adopt appropriate strategies for developing lodging-resistant *indica* rice cultivars in the future.

#### **2. Materials and Methods**

#### *2.1. Experimental Materials*

Thirty-two rice cultivars in total, including 18 Chinese accessions, 11 cultivars from a mini-core subset of the United States Department of Agriculture (USDA) rice gene bank [63], as well as Kasalath from India, Lemont from the USA and IR58025B from the IRRI, were selected as the experimental materials. These 32 rice cultivars were categorized into three groups (Groups 1, 2 and 3) with different lodging-resistance capabilities based on the principal component analysis and hierarchical clustering analysis of the lodging index in our previous study [25]. Group 1 (the lodging-susceptive group) comprised 3 *indica* cultivars and 4 *japonica* cultivars. There were 18 *indica* cultivars and four *japonica* cultivars in Group 2 (the lodgingmoderate group). The remaining three rice cultivars in Group 3 (the lodging-resistant group) all belong to *indica*. Table S1 lists the detailed grouping information.

#### *2.2. Experimental Design*

For phenotypic information acquisition of the tested cultivars at varying temperatures and solar radiations, we chose two eco-sites to carry out 3-year field experiments on three sowing dates (SDs) per year, which totaled three experiments. The location of the first experiment was Xindu in China's Sichuan province (30◦49051.5400 N, 104◦0603.4400 E; 547.7 m altitude), and the experiment was conducted during the rice-growing season from April to September in 2015. The location of the other two experiments was Ezhou in China's Hubei province (30◦22020.7500 N, 114◦4507.7800 E; 23.6 m altitude), and the experiments were conducted during the rice-growing seasons from May to October in 2017 and 2018, respectively. Every experiment was arranged in a split plot design with sowing dates (SDs) as main plots and cultivars as subplots. The 32 cultivars were all sown on April 11 (SD1), April 20 (SD2) and April 28 (SD3) in 2015 at the Xindu site. Following growth to about the fourth leaf stage, we transplanted seedlings from each SD to paddies in an experimental block roughly sizing 16 m <sup>×</sup> 6 m (96 m<sup>2</sup> ). Three blocks were set up for the three SDs. Each block was randomly arranged with 32 plots (each 2 m <sup>×</sup> 0.6 m in size; 1.2 m<sup>2</sup> ) for the 32 cultivars. Each cultivar was planted in triplicate rows (10 plants per row) at each experimental plot, where the hill spacing was 20 cm × 20 cm. Isolation of consecutive plots was accomplished at one-row spacing of 20 cm so that the growth impact on adjacent plants could be minimized and the plant cultivar could be clarified. In the case of the Ezhou site, the 32 cultivars were all sown on May 8 (SD4), May 23 (SD5) and June 7 (SD6) in 2017 for the second experiment, and on May 10 (SD7), May 25 (SD8) and June 9 (SD9) in 2018 for the third experiment. The seedling transplantation and plot designs for each SD were identical to those for Xindu.

In each experiment, 375 kg ha−<sup>1</sup> of basal fertilizer in total, which was the compound fertilizer containing each 15% *w*/*w* N, P and K, was applied to each block one day before the seedling transplantation. As an N source, urea was split-applied during the tillering stage at 150 kg ha−<sup>1</sup> and during the panicle initiation stage at 75 kg ha−<sup>1</sup> . About a 5-cm water depth was maintained in the experimental field post-transplant and remained flooded until 7 days before maturity. Intensive control of diseases, insects and weeds was implemented, in order to avoid biomass or yield loss.

#### *2.3. Meteorological Data Collection 2.3. Meteorological Data Collection*

The sources of the sunshine hours and temperature data, including the daily maximum, average and minimum temperatures, were the China Meteorological Data Service Centre [64] and the relevant local meteorological stations near the experimental fields (Figure 1). Due to the lack of equipment for measuring solar radiation, the solar radiation data could not be directly recorded at the meteorological stations. Therefore, an Angstrom empirical model was employed to simulate the 2015 solar radiation data for Xindu, and the 2017 and 2018 solar radiation data for Ezhou based on the sunshine hours [65,66]: The sources of the sunshine hours and temperature data, including the daily maximum, average and minimum temperatures, were the China Meteorological Data Service Centre [64] and the relevant local meteorological stations near the experimental fields (Figure 1). Due to the lack of equipment for measuring solar radiation, the solar radiation data could not be directly recorded at the meteorological stations. Therefore, an Angstrom empirical model was employed to simulate the 2015 solar radiation data for Xindu, and the 2017 and 2018 solar radiation data for Ezhou based on the sunshine hours [65,66]:

the seedling transplantation. As an N source, urea was split-applied during the tillering stage at 150 kg ha–1 and during the panicle initiation stage at 75 kg ha–1. About a 5-cm water depth was maintained in the experimental field post-transplant and remained flooded until 7 days before maturity. Intensive control of diseases, insects and weeds was

$$R\_G/R\_A = a + b \times n/N \tag{1}$$

where *R<sup>G</sup>* and *RA*, respectively, denote the global and extraterrestrial solar radiations (both MJ m−<sup>2</sup> day−<sup>1</sup> ), whereas *n* and *N,* respectively, represent the actual and potential daily sunshine hours. Following the method in a study by Chen et al. [67], "*a*" and "*b*", the empirical factors, were found to be 0.15 and 0.55 at Xindu, and 0.13 and 0.52 at Ezhou, respectively. The Nash–Sutcliffe equation (NSE) values were 0.81 and 0.84 for Xindu and Ezhou, respectively, which indicated that the model ran well [68]. MJ m−2 day−1), whereas *n* and *N,* respectively, represent the actual and potential daily sunshine hours. Following the method in a study by Chen et al. [67], "*a*" and "*b*", the empirical factors, were found to be 0.15 and 0.55 at Xindu, and 0.13 and 0.52 at Ezhou, respectively. The Nash–Sutcliffe equation (NSE) values were 0.81 and 0.84 for Xindu and Ezhou, respectively, which indicated that the model ran well [68].

where *RG* and *RA*, respectively, denote the global and extraterrestrial solar radiations (both

*Agronomy* **2022**, *12*, 2603 4 of 19

implemented, in order to avoid biomass or yield loss.

**Figure 1.** The daily average, minimum, and maximum temperatures and solar radiation during the whole rice growth season in 2015 at Xindu (**A**), and in 2017 (**B**) and 2018 (**C**) at Ezhou. **Figure 1.** The daily average, minimum, and maximum temperatures and solar radiation during the whole rice growth season in 2015 at Xindu (**A**), and in 2017 (**B**) and 2018 (**C**) at Ezhou.

The effective accumulated temperature (EAT) in the determined growth duration was calculated as: The effective accumulated temperature (EAT) in the determined growth duration was calculated as:

$$\text{EAT (}^{\diamond}\text{C)}=\sum(\text{T}-T\_{0})\times\text{Growth duration (d)}\tag{2}$$

EAT (°C) =*∑*(T − *T0*) × Growth duration (d) (2) where T and *T0* (10 °C for rice) are the daily average temperature and the biological zero where T and *T<sup>0</sup>* (10 ◦C for rice) are the daily average temperature and the biological zero temperature, respectively [58,69].

temperature, respectively [58,69]. The cumulative solar radiation (CSR) in the determined growth duration was calcu-The cumulative solar radiation (CSR) in the determined growth duration was calculated as:

$$\text{CSR (M) m}^{-2}\text{)}=\sum \text{R} \times \text{Growth duration (d)}\tag{3}$$

CSR (MJ m−2) =∑R × Growth duration (d) (3) where R (MJ m–2 d–1) is the daily solar radiation, which is calculated as the *RG* from the where R (MJ m−<sup>2</sup> d −1 ) is the daily solar radiation, which is calculated as the *R<sup>G</sup>* from the formula (1).

#### formula (1). *2.4. Plant Sampling and Measurements*

lated as:

*2.4. Plant Sampling and Measurements*  The growth period, which spanned from the sowing to the maturation stages, varied from 90 to 146 days depending on the different cultivars and environmental factors. According to the cultivar growth period in each plot, we recorded the full-heading date, which was defined as the date on which 80% emergence of all panicles occurred from the The growth period, which spanned from the sowing to the maturation stages, varied from 90 to 146 days depending on the different cultivars and environmental factors. According to the cultivar growth period in each plot, we recorded the full-heading date, which was defined as the date on which 80% emergence of all panicles occurred from the flag leaf sheath. At both experimental sites, we assessed the lodging-related traits of the plants in each plot 25 days following the full-heading date, i.e., the grain filling stage. The days of the determined growth duration from the sowing date to the measuring date were also recorded for each cultivar plot (Table 1). For the phenotypic determination, eight plants were picked from the middle of each plot to avoid the marginal effects. These harvested plants, which had all tillers and intact roots, were transferred to the laboratory in plastic buckets filled with water for subsequent analyses. For lodging-related traits determination, the representative samples of plant main tillers were collected. We examined only the traits associated with the second elongated internode (starting from the main tiller root) in view

of the common occurrence of culm-breaking-lodging type at the lower internodes [2,70]. Measurement of the culm length (CL) was accomplished between the lower node of the basal second internode and the panicle tip. Measurement of the stem length (SL) was accomplished between the lower nodes of the basal second and third internodes. Panicle length (PL) measurement was accomplished from the bottom node to the tip of a panicle. Determination of the fresh weight (FW) was accomplished for the zone between the lower node of the basal second internode and the panicle tip. A YYD-1 plant lodging tester (TOP Instrument, Zhejiang China) was utilized for determining the breaking resistance (BR) at the middle point of the basal second internode with a leaf sheath as per a priorly reported procedure [29]. Each sampled basal second internode with a leaf sheath was placed on the groove of supporting pillars that were 5 cm apart. The plant lodging tester, which was arranged perpendicularly to the middle internode, was then progressively loaded. BR measurement was accomplished when the internode was pushed to break, and the tester readout was recorded as the BR value (kg). Measurements of the culm diameter (CD), culm wall thickness (CT), as well as the external and internal diameters of the minor and major axes in an oval cross-section at the basal second internode center were accomplished using a digital vernier caliper. The physical parameters were calculated following the studies of Ookawa and Ishihara [71] and Ookawa et al. [72], and our previous study [25] as follows:

Bending moment of the whole plant (BM, g cm) = CL (cm) × FW (g) (4)

where CL is the culm length from the lower node of the basal second internode to the panicle tip, FW is the fresh weight from the lower node of the basal second internode to the panicle tip.

Bending moment at breaking (M, g cm) = 1/4 <sup>×</sup> BR (kg) <sup>×</sup> Spacing between supporting pillars (cm) <sup>×</sup> <sup>10</sup><sup>3</sup> (5)

> where BR is the breaking resistance of the basal second internode with leaf sheath; spacing between supporting pillars is 5 cm.

$$\text{Lodging index (LL\\_\%)}=\text{BM/M}\times100\%\tag{6}$$

where BM and M are calculated from formulas (4) and (5), respectively.

The external force that the basal second internode could withstand (4BM, g cm) = 2M − BM (7)

where BM and M are calculated from formulas (4) and (5), respectively.

$$\text{Section modulus (SM, mm}^3\text{)} = \pi/32 \times (\mathbf{a}\_1^3 \mathbf{b}\_1 - \mathbf{a}\_2^3 \mathbf{b}\_2)/\mathbf{a}\_1\tag{8}$$

$$\text{Culm diameter (CD, mm)} = (\mathbf{a}\_1 + \mathbf{b}\_1)/2 \tag{9}$$

$$\text{Culm wall thickness (CT, mm)} = (\mathbf{a}\_1 \cdot \mathbf{a}\_2 + \mathbf{b}\_1 \cdot \mathbf{b}\_2)/4 \tag{10}$$

where a<sup>1</sup> and b<sup>1</sup> respectively denote the outer diameters of the minor and major axes in an oval cross-section, whereas a<sup>2</sup> and b<sup>2</sup> respectively represent the inner diameters of the minor and major axes in an oval cross-section.

$$\text{Bending stress (BS, g mm}^{-2}) = \text{M/SM} \times 10 \tag{11}$$

where M and SM are calculated from formulas (5) and (8), respectively.

Following the determination of all lodging-related parameters, the basal second internode with the leaf sheath was subjected to 105 ◦C oven-drying for 30 min, then dried at 70 ◦C until the weight was constant. The dried basal second internodes were then ground to a fine powder for determining structural carbohydrates. The cellulose content (CC) was measured using a commercial cellulose content assay kit (Boxbio Science, Beijing, China), which was modified following the method of Updegraff [73]. Since the acetyl bromide method for the determination of lignin appears to have earned the most widespread accep-

tance [74], we used a commercial lignin content assay kit (Boxbio Science, Beijing, China), which was modified following the method of Johnson et al. [75], to measure the lignin content (LC) of the samples. The CD, CT, LC and CC were only measured in 2018 at Ezhou.

**Table 1.** The determined growth durations (days) of twelve *indica* rice cultivars on all the sowing dates across two locations and three years.


<sup>a</sup> SD1, 11 April 2015; SD2, 20 April 2015; SD3, 28 April 2015; SD4, 8 May 2017; SD5, 23 May 2017; SD6, 7 June 2017; SD7, 10 May 2018; SD8, 25 May 2018; SD9, 9 June 2018. All abbreviations imply the same below as well.

#### *2.5. Data Analysis*

Acording to the field observation, 12 *indica* rice cultivars, including the nine cultivars (Chuan 106B, 345B, Huanghuazhan, Jinlongsimiao, Chuanxiang 29B, Chenghui 3203, Guichao 2, II-32B and Teqing) in the lodging-moderate group and the three cultivars (R379, 9311 and Jiangan) in the lodging-resistant group, which did not lodge on all the sowing dates, were selected for the following data analysis. The measured data were analyzed via the Statistical Product and Service Solutions software for Windows (SPSS, Ver. 20.0; IBM, NY, USA). For each rice cultivar, the mean of eight main tillers was determined as the lodging-related parameter for that cultivar plot. The average daily temperature (*Tmean*) and solar radiation (*Rmean*), effective accumulated temperature (EAT) and cumulative solar radiation (CSR) for each SD were calculated based on the mean values of all the 12 *indica* rice cultivars in the determined growth durations. A Tukey test was employed to accomplish the significance analysis of the *Tmean*, *Rmean*, EAT, CSR, LI and 4BM for the different sowing dates in each year, where *p* = 0.05 was set as the statistical significance level. To determine the relationship of the temperature and solar radiation parameters with the lodging-related parameters, we minimized the inter-cultivar differences in traits using the standardized data according to the method described in a study by Fan and Liu [76]. The standardized data, which were referred to as "relative data", were calculated according to the formula below:

The relative lodging index = the lodging index of a given cultivar on one sowing date/average lodging index of this cultivar on all the sowing dates.

The other relative values of lodging-related parameters were calculated using the formula for the relative lodging index described above. The stepwise regression analysis and path analysis were performed to determine the relationships of temperature and solar radiation parameters to the lodging resistance. The correlation analysis was conducted using Pearson's *r* coefficient, and the correlations were regarded as statistically significant when *p* < 0.05. The correlation coefficient was decomposed into direct effects and indirect effects by path analysis, with the result that the relative importance of the temperature and solar radiation for the lodging resistance were revealed. The values of direct and indirect effects were calculated according to the study of Luo et al. [77]. The GraphPad Prism software for Windows (Ver. 5.0; San Diego, CA, USA) was utilized for depicting the scatter diagrams and histograms.

#### **3. Results**

### *3.1. Effects of Sowing Date on Determined Growth Durations and Temperature and Solar Radiation Conditions*

The determined growth durations of all the 12 *indica* rice cultivars were shortened as the sowing dates were delayed in each year (Table 1). Compared with the first sowing date, the number of days of the determined growth durations in the last sowing date were reduced by 5−14 d, 2−11 d, and 1−12 d in 2015, 2017, and 2018, respectively. The average determined growth durations of the three sowing dates at Ezhou were reduced by 6−30 d compared with those at Xindu.

The average daily temperature (*Tmean*) of the experiment during the determined growth durations increased significantly with the delay of the sowing dates in each year (Figure 2A). *Tmean* on SD1 was 0.4 and 0.8 ◦C lower than that on SD2 and SD3, respectively. *Tmean* in SD4 was 0.2 and 0.3 ◦C lower than that on SD5 and SD6, respectively. Additionally, *Tmean* on SD7 was 0.5 and 0.8 ◦C lower than that on SD8 and SD9, respectively. Furthermore, the *Tmean* showed significant differences between locations and years (Table S2). The *Tmean* from SD1 to SD3 in 2015 at Xindu was significantly lower than that from SD4 to SD6 in 2017 and from SD7 to SD9 in 2018 at Ezhou by 4.3 and 5.1 ◦C, respectively. However, the effective accumulated temperature (EAT) showed no significant variations among the three sowing dates in each year. Similar to the *Tmean*, the EAT at Xindu was significantly lower than that at Ezhou (Figure 2B, Table S2). *Agronomy* **2022**, *12*, 2603 8 of 19

**Figure 2.** Differences in the average daily temperature (*Tmean*) (**A**) and solar radiation (*Rmean*) (**B**), effective accumulated temperature (EAT) (**C**) and cumulative solar radiation (CSR) (**D**) of the twelve *indica* rice cultivars on the three sowing dates of each year during the determined growth durations. SD1, 11 April 2015; SD2, 20 April 2015; SD3, 28 April 2015; SD4, 8 May 2017; SD5, 23 May 2017; SD6, 7 June 2017; SD7, 10 May 2018; SD8, 25 May 2018; SD9, 9 June 2018. All abbreviations imply the same below as well. Vertical bars indicate standard errors (±), *n* = 12 for each SD. Different lowercase letters on the bars of the three SDs in each year indicate significant differences determined by the Tukey test at 5% probability level. **Figure 2.** Differences in the average daily temperature (*Tmean*) (**A**) and solar radiation (*Rmean*) (**B**), effective accumulated temperature (EAT) (**C**) and cumulative solar radiation (CSR) (**D**) of the twelve *indica* rice cultivars on the three sowing dates of each year during the determined growth durations. SD1, 11 April 2015; SD2, 20 April 2015; SD3, 28 April 2015; SD4, 8 May 2017; SD5, 23 May 2017; SD6, 7 June 2017; SD7, 10 May 2018; SD8, 25 May 2018; SD9, 9 June 2018. All abbreviations imply the same below as well. Vertical bars indicate standard errors (±), *n* = 12 for each SD. Different lowercase letters on the bars of the three SDs in each year indicate significant differences determined by the Tukey test at 5% probability level.

*3.2. Effects of Sowing Date on LI and* △*BM*  The average solar radiation (*Rmean*) of the experiment during the determined growth durations showed an inconsistent tendency with the delayed sowing dates in each year

ing dates in each year and it was greater at Xindu than that at Ezhou.

The LI and △BM, which were used to evaluate the lodging resistance of rice, were significantly affected by the sowing dates (Figure 3). With the delayed sowing dates, the LI increased significantly in each year (Figure 3A). The LI on SD3 was 22.65% higher than

higher than SD7. Although the tendency of LI was not always significant in some cultivars, the overall increasing tendencies of LI of the 12 *indica* cultivars were identical (Table 2). In addition, the LI was also significantly affected by locations and years (Table S2). The average LI of the three SDs (SD1, SD2, and SD3) in 2015 at Xindu was significantly lower than that of the three SDs (SD4, SD5, and SD6) in 2017 and the three SDs (SD7, SD8, and SD9) in 2018 at Ezhou by 41.03% and 36.97%, respectively. Contrary to the LI, the △BM decreased significantly with the delay in the sowing dates in each year (Figure 3B). Compared to the first sowing date, the △BM of the last sowing date was lower by 35.52%, 57.62% and 47.93% in 2015, 2017 and 2018, respectively. The decreasing trends of △BM of all the 12 *indica* cultivars were significant (Table 3). Furthermore, the average △BM of the three SDs in 2015 at Xindu was significantly higher than that of the three SDs in 2017 and 2018 at Ezhou by 255.88% and 179.51%, respectively. However, there was no significant difference in the △BM between 2017 and 2018 at Ezhou (Table S2). These results indicated that the lodging resistance of 12 *indica* rice cultivars weakened with the delay in the sow(Figure 2C). From SD1 to SD3, the *Rmean* was relatively constant. However, the *Rmean* on SD4 was 0.7 and 1.1 MJ m−<sup>2</sup> day−<sup>1</sup> higher than that on SD5 and SD6, respectively. Compared to the declining trend from SD4 to SD6, the *Rmean* displayed the opposite trend from SD7 to SD9, where the *Rmean* on SD7 was lower than that on SD8 and SD9 by 0.3 and 0.4 MJ m−<sup>2</sup> day−<sup>1</sup> . In addition, the *Rmean* was also significantly affected by locations and years (Table S2). The *Rmean* from SD1 to SD3 in 2015 at Xindu was significantly lower than that from SD4 to SD6 in 2017 and from SD7 to SD9 in 2018 at Ezhou by 1.8 and 3.9 MJ m−<sup>2</sup> day−<sup>1</sup> , respectively. As a result of delayed sowing dates, the cumulative solar radiation (CSR) decreased significantly from SD1 to SD3 and from SD4 to SD6 (Figure 2D). In addition, the CSR displayed an insignificant declining trend from SD7 to SD9. On the other hand, the CSR in 2015 at Xindu and in 2017 at Ezhou were significantly lower than those in 2018 at Ezhou (Figure 2D, Table S2).

Thus, a total of nine temperature and solar radiation treatments with significant differences were established for the twelve *indica* rice cultivars by setting nine sowing dates across two eco-sites and three years.
