Impact of low-temperature, overcast and rainy weather during the reproductive growth stage on lodging resistance of rice

The objectives of this study were to explore the mechanism by which the lodging resistance of the rice population during the late growth period responds to low-temperature, overcast and rainy weather during the reproductive growth stage. Field experiments were conducted using indica rice Yliangyou2 (lodging-resistance variety), IIyou084 (lodging-susceptible variety) and japonica rice Wuyunjing23 (lodging-resistance variety) and W3668 (lodging- susceptible variety) in 2013 (high temperature and strong radiation during the rice reproductive growth stage), 2012 and 2014 (low temperature and weak radiation during rice reproductive growth stage). The results showed that the length of the basal internodes and the height of the gravitational centres were greater in plants grown in 2014. Dry weight of basal culms, culm diameter, lignin content and total content of structural carbohydrates (lignin and cellulose) in basal internodes were reduced in these plants, causing a significant reduction in the bending stress and lodging resistance of the rice stems. Low-temperature, overcast and rainy weather had a greater effect on lodging resistance in indica rice compared with japonica rice. This was reflected in a greater reduction in the lignin content of the indica rice stems, which yielded a significantly lower breaking strength and bending stress.

promotes internode elongation and reduces culm wall thickness. This reduction results in fragile supporting tissue and low stem strength, thus aggravating lodging 16 . However, no study has systematically investigated how low-temperature, overcast and rainy weather affects the reproductive growth stage and rice stem strength and the associated differences between varieties.
A rare period of long low-temperature, overcast and rainy weather was encountered during the reproductive growth stage of rice in East China in 2014. In this study, we conducted experiments related to lodging on two indica hybrid rice varieties and two japonica rice varieties under the same field management conditions. The lodging resistance capacities of rice showed significant differences in 2014. Therefore, we analysed the differences in lodging resistance of rice stems from 2012 to 2014 in terms of physical properties, morphological characteristics, and chemical composition to provide a reference for research into lodging resistance in rice.

Materials and Methods
Comparison of climatic conditions. Long-term means of meteorological data from 2011 to 2015 are listed, which were analysed using growth periods of japonica rice varieties (Supplemental Table S1). Since the reproductive stage (from panicle initiation [PI] to the heading stage [HS]) is the key period for internode development, weather conditions in 2012 and 2014 were extreme, characterised by low temperature, low PAR and more rainfall, especially in 2014. As a common weather condition, 2013 was compared with 2014 to determine how extreme weather could affect rice lodging. Climatic conditions were significantly different between the two years ( Fig. 1). The climatic differences are illustrated below with japonica rice as an example because the growth periods between japonica and indica rice are similar.   (3) Greater rainfall was observed before the rice HS in 2014. From TS to PI in 2013, the annual rainfall was 142.8 mm, and 7 rainy days were recorded. The annual rainfall was 192.0 mm, and 12 rainy days were recorded from TS to PI in 2012. During this period in 2014, the annual rainfall was 338.6 mm, and 20 rainy days were observed. From PI to HS, the annual rainfall was 171.7 mm (with 5 rainy days recorded) in 2012 and 72.9 mm (with 6 rainy days recorded) in 2013. During this period in 2014, the annual rainfall was 185.9 mm, with 20 rainy days observed. From HS to MS, the annual rainfall and the number of rainy days in 2014 were only slightly less than their respective measurements in 2013, while the annual rainfall and the number of rainy days in 2012 was close to zero.

Experimental materials and design. Two super hybrid indica rice varieties (lodging-resistant
Yliangyou2 and lodging-susceptible IIyou084) and two japonica rice varieties (lodging-resistant Wuyunjing23 and lodging-susceptible W3668) were used as the experimental materials. Field experiments were conducted at an experimental base at Nanjing Agricultural University (Nanjing, China) on the Baolin Farm in Yanling Town, Danyang City, Jiangsu Province, China. The area is located in a subtropical monsoon climate, and the soil at the site is classified as alluvial loamy. The plough layer of the experimental soil contained the following: 21.02 g/kg organic matter, 1.12 g/kg total nitrogen, 70.60 mg/kg available nitrogen, 13.23 mg/kg available phosphorus, and 119.41 mg/kg available potassium, at a pH of 6.85. Seedlings were raised in a dry nursery using nutrition seedling trays. The seedlings were transplanted into the field, and the row spacing used for rice transplantation was 13.3 cm × 30.0 cm, with two seedlings per hole. Growth periods of japonica and indica rice from 2012 to 2014 are presented in Table 1. Table S2 shows the nitrogen (N) application treatments used in the field experiments. The total N application was 270 kg/ha, which included a base-tiller fertilizer of 135 kg/ha (basal: tillering = 50%: 50%) and a panicle fertilizer of 135 kg/ha (PI: spikelet differentiation = 60%: 40%). Basal, tillering, PI, and spikelet differentiation fertilizers were applied 1 day before transplantation, 7 days after transplantation, at the young panicle differentiation stage, and at the second leaf (penultimate leaf) stage, respectively. The experiments were arranged in a randomized complete block design with three replications. Each plot is 24 m 2 (4 m × 6 m) and all plots were covered with plastic films on 30 cm-wide ridges to prevent fertilizer contamination. Phosphorous was applied for basal fertilizer one day before transplantation with super-phosphate at a rate of 90 kg/ha P 2 O 5 . Potassium was applied with potassium chloride at a rate of 120 kg/ha K 2 O, 50% as basal fertilizer and 50% as panicle initiation fertilizer. Other management practices followed conventional cultivation methods.
Field sampling and calculation methods. Thirty days after heading, 10 main stems with consistent growth were selected to measure the HGC and the length of the first and second basal internodes. The HGC was measured as follows: the fresh plant was placed on the tip of a pencil, and the fulcrum point was adjusted to achieve balance. The distance from the base to the pencil tip was defined as the HGC (cm). Next, an AIKON RX-5 digital force gauge (Company: AIKON, Japan) was used to break the midpoint of the second basal internode and measure the bending load at a distance of 8 cm between two fulcrum points. The second basal internode was then cut off from the midpoint, and a vernier calliper was used to measure the inner and outer diameters of the major and minor axes in a hollow oval stem after the removal of the leaf sheath. Physical parameters on the second basal internode were calculated using the following formulas (Ookawa, 1992   Chemical composition analysis of rice stems. Non-structural carbohydrate (NSC) content was determined according to the method of Yoshida et al. 17 . For structural carbohydrate (SC) content analysis, cellulose content was determined using the method of Updegraff et al. 18 and Zhang et al. 15 . Lignin content was determined in accordance with Ishimaru et al. 13 .
Statistical analysis. Data were processed using Microsoft Excel 2003. Analysis of variance was performed using SPSS 17.0 for windows based on the significant difference test (Duncan) at 0.05 or 0.01 probability level. Plots were generated using OriginPro 8.0. Table 2 summarises the yield and yield components of different varieties. However, the interaction between years and varieties was only observed in panicles and 1000-grain weight. The lodging-resistant varieties Yliangyou2 and Wuyunjing23 had 12% and 31% higher yields, respectively, than the lodging-susceptible varieties IIyou084 and W3668. Although no significant difference was discovered among the years, effects still existed by comparing the two most extreme weather conditions in 2013 and 2014. For indica rice varieties, interannual comparisons showed that the grain yields of Yliangyou2 and IIyou084 were reduced by 10.9% and 10.1%, respectively, in 2014 compared with 2013. The lower yields in 2014 were due to reductions in the numbers of productive panicles and spikelets per panicle, which resulted in significantly fewer spikelets (15.0% and 18.5% fewer spikelets were observed in Yliangyou2 and IIyou084, respectively). For japonica rice varieties, Wuyunjing23 and W3668 yields increased by 3.4% and 5.4%, respectively, in 2014 compared with 2013. Although a reduction in productive panicles resulted in fewer spikelets, the seed-setting rate and the 1000-grain weight both increased in the japonica rice varieties (the seed setting rate of Wuyunjing23 and W3668increased by 5.4% and 3.4%, and the 1000-grain weight increased by 3.1% and 5.8%, respectively), which were likely the main reasons for the observed yield increases. Moreover, due to the higher seed-setting rate and 1000-grain weight in 2012 than in 2013, the yields were even higher in IIyou084 and Wuyunjing23, which might have been the result of good weather conditions during the HS to the MS in 2012.

Rice yield.
Physical properties. Significant (Table 3). With regard to the source of the LI increase, WP and M were significantly reduced for both of the indica varieties, while the reduction in M was greater. For the two japonica varieties, the increase in LI was mainly due to a reduction in M, as WP even increased in 2012 and 2014. The indica varieties exhibited no obvious differences in SL among the years, whereas FW was significantly reduced, resulting in a lower WP. However, due to a significant increase in SL, the japonica varieties had no reduction in WP. Compared with 2013, the M values of Yliangyou2, IIyou084, Wuyunjing23 and W3668 were reduced by 21.0%, 13.6%, 3.1%, and 6.6% in 2012 and by 33.1%, 34.8%, 19.4%, and 22.2% in 2014, respectively. Among the different varieties, indica rice exhibited a more obvious reduction in M. This was mainly due to a significant reduction in BS. In 2014, BS was reduced by 20.4% and 39.8% in Yliangyou2 and IIyou084, but by only 4.3% and 13.7% in Wuyunjing23 and W3668, respectively. However, in 2012, BS was reduced by 12.9%, 29.0% for indica varieties and by 17.1%, 31.3% for japonica varieties, which did not show much difference. The Z values showed inconsistent changes between varieties. A significant difference among years was observed in the japonica varieties, but not in the indica varieties. The results showed that low-temperature, overcast and rainy weather could increase the risks of rice lodging and had a greater effect on indica varieties than on japonica varieties in terms of two stem strength parameters (BS and M). Figure 2 presents the configuration of the internodes and HGC in different varieties. On average, indica rice was taller than japonica rice. The HGC and the lengths of internodes VI, V and IV significantly differed among the years (Table 4). Compared with 2013, the lengths of the first, second and third basal internodes in Yliangyou2 increased by 38.5%, 70.8% and 42.3 in 2014, respectively. While in IIyou084, these increases were 31.5%, 71.5% and 33.4%. The HGC values for Yliangyou2 and IIyou084 were 3.3% and 7.1% greater in 2014 compared with 2013, respectively. For japonica rice, the lengths of the first and second basal internodes increased significantly. Compared to 2013, the first and second basal internodes increased by 413.0% and 108.8%, respectively, in Wuyunjing23 in 2014. The first and second basal internodes of W3668 increased by 288.2% and 98.7%, respectively, over the same period. These differences were extreme and significant. The HGC values for Wuyunjing23 and W3668 were 20.3% and 10.3% greater in 2014 compared to 2013. However, the increase in internode lengths of all varieties in 2012 was less than in 2014. Moreover, the HGC in 2012 was even lower than that in 2014, which was also reflected by an interaction effect between years and varieties on HGC (Table 4). Table 5, there was no interaction between varieties and years. It also shows that compared with 2013, the dry weight of basal culms of Yliangyou2, IIyou084, Wuyunjing23 and W3668 was significantly reduced by 21  Carbohydrates in basal culms and leaf sheaths. In addition, to analyse whether the carbohydrates in the basal culms and leaf sheaths were altered under extreme weather conditions, NSC, cellulose, and lignin contents were measured (Table 6)    content decreased by 2.3% and 19.1%, and the total content of SC (lignin and cellulose) decreased by 12.8% and 23.4 %in Wuyunjing23 and W3668, respectively. Moreover, the interaction between years and varieties was observed in the effect of lignin contents in the basal culms. The lignin content in the basal stems showed greater interannual differences in indica rice varieties compared with japonica rice varieties. Interestingly, leaf sheaths had relatively high contents of NSC, cellulose and lignin, especially in the indica varieties.

Discussion
Low-temperature, overcast and rainy weather decreased the rice lodging resistance. In this study, the climate condition in 2013 was normal when compared with the average value of multi-year. However, the daily mean temperature and PAR were lower before the rice HS in 2012 and 2014 compared with 2013, especially for 2014. Moreover, a longer period of low-temperature, overcast and rainy weather was encountered in 2012 and 2014. There was no significant difference in PAR, temperature and rainfall after the HS between 2013 and 2014, but there was nearly no precipitation in 2012 ( Fig. 1 and Supplemental Table S1). The results showed that low-temperature, overcast and rainy weather exerted a profound impact on lodging resistance in rice stems, with a significant increase in the LI (Table 3). TS to PI occurs during the vegetative growth period of rice plants, while PI to HS is critical for elongation and the plumpness of stem internodes. We inferred that the LI increase in 2012 and 2014 was mainly due to prolonged low-temperature, overcast and rainy weather from PI to HS, the reproductive growth period of rice.
Indica and japonica rice responded differently to extreme weather. Genotype is a key factor in determining rice lodging, and variations are often observed between rice varieties 6,19 . The reason for the LI increase varied in the indica and japonica varieties. The results showed that FW was markedly reduced in indica rice, but not in japonica rice. The number of spikelets per panicle also showed the same trend, decreased significantly in indica rice, but a little in japonica rice ( Table 2). For japonica rice, FW did not decrease and SL increased, which resulted in a significant increase in WP. This phenomenon explained the remarkable increase in the LI of the two japonica varieties. Although the WP of the indica rice decreased while that of japonica rice increased, the M values of both types of rice were reduced. In addition, this reduction in M as well as in BS was more obvious in indica rice than in japonica rice (Table 3). These results lead us to infer that the yield and lodging resistance of indica rice are more significantly affected by low-temperature, overcast and rainy weather. In addition, elongated basal internodes and increased plant height were much more obviously observed in japonica rice, while the lignin content in the basal internodes showed a greater interannual difference in the indica rice varieties compared with japonica rice. This phenomenon might explain why greater reductions in M and BS were observed in the two indica rice varieties but not in the two japonica rice varieties.
Light may be the main cause of rice lodging under low-temperature, overcast and rainy weather. Low light and shading can cause excessive growth in plants, with elongated basal internodes, higher HGC, and lower lodging resistance [20][21][22] . Such reduction of lodging resistance has also been observed in maize 23 and soybean 24 . A similar phenomenon was observed in this study. Weather changes had an obvious impact on many morphological characteristics, such as HGC and length of the elongated internodes of basal stems, which may be induced by an increased gibberellin content according to previous studies 25, 26 . In addition, the stem strength of rice stems is also closely related to their plumpness. The plump components of rice stems mainly include NSC (e.g., soluble sugar and starch) and SC (e.g., cellulose and lignin). Scholars have suggested that stems contain a high content of starch. During the grain filling stage after heading, the starch retained in stems is used to improve lodging resistance in plants 27 . However, Zhang et al. 28 reported that starch and soluble sugar contents in stems have little relationship with bending load. Instead, bending load is mainly associated with cellulose and lignin contents in basal stems, and greater cellulose and lignin contents result in greater stem strength 15 . Former research has indicated that light intensity can promote the expression of genes related to Lignin biosynthesis, thus increasing the lignin content 29,30 . Here, an obvious reduction in lignin content was also observed, and such a decrease was possibly due to relatively low PAR in 2012 or 2014. More in-depth, light quantity (PAR) was associated with stem traits associated with lodging, while light quality (F:FR) played a role in determining rooting traits that affect the anchorage capacity of wheat plants 31,32 . However, in our present study, the comprehensive effect of temperature, light and rainfall was emphasised. Further investigation is necessary to determine the separated effect of single factors on rice lodging and to compare whether it is the same.
Potential value of understanding the effect of low-temperature, overcast and rainy weather on rice lodging resistance. Frequent extreme weather will bring great risks to rice production, so it is meaningful to determine the potential effect of such weather on rice growth and development. Recently, multiple studies have focused on whether cultivation measures have an impact on rice lodging, such as planting density 33 or nitrogen application rates [34][35][36] . In addition, plant growth regulators can be involved in the formation of stem morphology and substance accumulation, thus affecting the mechanical strength and lodging resistance of stems. Gibberellins and paclobutrazol or uniconazole have shown significant and rapid effects on rice lodging 37,38 . Our results clarified that low-temperature, overcast and rainy weather reduced the lodging resistance of rice and have a negative effect on yields, and the response of rice to this extreme weather was different between japonica and indica varieties. This is helpful for us to choose the appropriate measures to avoid the rice lodging and ensure the rice high yield when the extreme weather is encountered.