Effect of Climatic Conditions Caused by Seasons on Maize Yield, Kernel Filling and Weight in Central China

Abstract

In order to evaluate the effects of climatic conditions on maize grain yield (GY), kernel weight (KW), and kernel filling and identify the optimal climatic factors for GY and KW, 2-year field experiments in three seasons, i.e., spring (S_PM), summer (S_UM), and autumn (A_UM), on maize were conducted in Central China. The results showed that S_UM had more growing degree days (GDDs) than S_PM and A_UM due to the higher mean temperature (MT), and also resulted in higher temperature stress (killing degree days (KDDs)) in maize growth duration. Meanwhile, after silking, S_PM and S_UM had more GDDs and KDDs than A_UM because of the higher MT, and the accumulated solar radiation (Ra) for S_UM was significantly higher than for S_PM and A_UM. The GY of S_PM was significantly higher than that of S_UM and A_UM, while S_UM’s GY was always the lowest, because the GDD_GD, MT_GD, and KDD_GD played significantly negative roles on GY. The final KW for S_UM was always the lowest, with GDD, MT, KDD, and Ra causing significantly negative effects, and M△T and precipitation having significant positive effects, resulting in a lower kernel filling rate during the linear kernel filling period (KFR_lkf) and a lower GDD at the maximum kernel filling rate (GDD_KFRmax). Maize KFR_lkf has significant negative linear dependences on GDD, MT, and Ra. In summary, because of the higher MT, KDD, and GDD during maize growth and kernel filling duration negatively affecting the maize kernel filling rate, the GY and KW for S_PM were the highest, and for S_UM, they were the lowest; therefore, farmers should plant S_PM first and then A_UM in Central China.

1. Introduction

Maize (Zea mays L.) is one of the most important crops in China, accounting for 20.43% and 22.42% of the world’s area and food production in 2020, respectively [1]. Central China, despite having abundant climatic resources, is a suboptimal region for maize production due to significant environmental stresses (e.g., high temperature and rainfall, and low solar radiation (Ra)). In this region, maize has a long planting period, being planted from early March to mid-July, classified as spring maize (S_PM), summer maize (S_UM), and autumn maize (A_UM) according to the planting system [2]. Therefore, it is important to understand the effects of the various climatic factors on maize yield in Central China.

The maize grain yield is closely determined by the number of kernels per unit area and final kernel weight (KW) at harvest [3,4,5], which is always reduced by 0.83 t ha⁻¹ and 0.67 t ha⁻¹ with a mean temperature (Tmean) and minimum temperature (Tmin) increased of 1 °C in different maize regions of China [6]. Under field conditions, the KW is strongly affected by episodes of abiotic stress (e.g., extreme T, water deficit or flooding, and nutrient deficiency) during the grain filling period, depending on the duration, probability, and intensity [7,8,9]. The KW depends strongly on the kernel filling rate (KFR) and linear kernel filling duration [8,10,11]. Many studies have suggested that the temperature and growing degree days (GDDs) are the main climatic factors influencing the maize phenological development duration and KW [4,6,12,13,14]. The optimum T for mid-summer maize growth and kernel filling were 21–27 °C and 27–32 °C, respectively [14,15]. The maize KW is decreased by low temperatures because of the lower leaf photosynthetic and net assimilation rate [16], with a 10 °C low temperature for 5 days being sufficient to prevent the KW from increasing [17]. However, high-temperature stress always decreased the chlorophyll index, reduced the net rate of photosynthesis [18,19], and shortened the maize kernel filling duration (KFD) [20], with a high temperature over 35 °C possibly causing kernel abortion due to the prevention of sugar–starch conversion [21,22,23]. Ra is an important climatic factor that strongly affects maize grain yield via biomass accumulation and KW [6,12,21,24,25,26]. Low light intensity results in reductions in leaf area index and photosynthesis [27], thus limiting the development of kernels and inhibiting starch formation during kernel filling, leading to lower KW [28,29]. Moreover, precipitation (Pr) is another important climatic factor affecting crop yield [25,30]. Drought stress accelerated leaf senescence, and decreased the photosynthetic electron transport of photosystem II and leaf photosynthetic capacity [20,31], as well as shortened the KFD, resulted in KW reduced [7,32]. However, waterlogging also played a negative effects on maize KW and GY [29,33]. Furthermore, it is difficult to account for the overall effect of these individual climatic factors on maize kernel filling, particularly the interaction between soil fertility, crop management, and the other factors.

In order to define the effects of climatic factors, many studies have used different planting dates, which generate different climatic conditions for maize kernel filling in fields [5,11,13,34,35]. The maize KWs were decreased by planting date delay due to the decreasing of Ra and temperature after silking [11,34], with the decreased KFR in arid regions [11,19]. However, the range of planting dates was considered to be too narrow in these studies [35,36]. Zhou et al. [35] reported that the maize KW increased initially and then declined according to the delay in planting across eight dates in the Huang-Huai-Hai Plain, China, and S_UM obtained the greatest KW due to a higher KFR and longer KFD, which were affected by a suitable temperature and Ra. However, in another previous study [2], the KW decreased initially and then increased according to eight planting dates across three seasons in Central China, with the S_UM KW being the lowest due to the higher temperature and a lower Ra and Pr, but particularly the high-temperature stress (killing degree days (KDDs)). A limitation of those studies was that the specific climatic factors affecting maize KW development were undefined. Moreover, the various climatic factors affecting the kernel filling process and KW, which determined the GY, remain unclear. In this study, six planting dates across three seasons, i.e., S_PM, S_UM, and A_UM, from the middle of March to the end of July, were set to simulate a wide range of climatic conditions for maize kernel filling with the following goals: (1) to clarify the relationship between kernel filling characteristic parameters and climatic factors under different seasons, (2) to identify the effects of climatic conditions on maize KW and GY according to seasons, and (3) to optimize the suitable maize plant season for higher GY in Central China. The results provide theoretical guidance for applying suitable date planted high-yield maize for similar areas worldwide.

2. Materials and Methods

2.1. Experimental Sites

We conducted a 2-year field experiment, from March to November in 2012 and 2013, at an experimental farm of Huazhong Agricultural University in the town of Huaqiao, Hubei Province, China (30°06′ N, 115°45′ E, 30 m elevation). This is located at the middle reach of the Yangtze River and has a humid subtropical climate. The properties of the soil, determined from its 0–20 cm–deep layer, were as follows: pH of 5.73, 16.8 g kg⁻¹ organic matter, 1.6 g kg⁻¹ total N, 11.8 mg kg⁻¹ available Olsen P, and 105.8 mg kg⁻¹ exchangeable K.

2.2. Experimental Design and Cropping Management

The treatments were arranged in a split-plot design, with seasons as main plots and varieties as subplots. The three seasons had two sowing dates, each in each of the 2 years, set as main-plot treatments. Three regional maize hybrid varieties with various growth durations were designated as the subplots: Zhengdan 958 (ZD958), Denghai 9 (DH9), and Yidan 629 (YD629).

The experimental field was plowed and built into a broad bed and furrow system, with beds 200 cm wide and furrows 40 cm wide and 30 cm deep, before planting. Each subplot was 7 m long and 4.8 m wide and consisted of eight rows, with three repetitions. Maize seeds were sown manually, with three seeds per hole and 27.5 cm between holes, with wide and narrow row spacings of 80 and 40 cm, respectively. The two outside rows were borders. Plants were thinned at the three-leaf stage to two plants per hole and at the five-leaf stage to one plant per hole at a stand density of 6.0 plants m⁻². Each subplot was fertilized with 27 g m⁻² N, 15 g m⁻² P₂O₅, and 18 g m⁻² K₂O at planting, of which 27 g N, 9.8 g P₂O₅, and 14.7 g K₂O were applied by slow-release fertilizer (N:P₂O₅:K₂O, 22%:8%:12%; Kingenta Com. Ltd., Linshu, China), 5.2 g P₂O₅ was applied as superphosphate (P₂O₅, 12%), and 3.3 g K₂O was applied as potassium chloride (K₂O, 60%). All management and agronomic practices were identical for each subplot. Experiments were rainfed without appreciable waterlogging stress. Diseases, pests, and weeds were controlled by using chemicals throughout the growing seasons.

2.3. Measurements Methods

2.3.1. Kernel Weight, Kernel Filling and Grain Yield

The emergence (VE), silking (R1), and maturity (R6) stage dates of each subplot are listed in

Table 1. Maize crop phenology in three different maize seasons in 2012 and 2013.

Seasons	Varieties	Mazie Growth Stage (Day/Month)
		2012				2013
		PD	VE	R1	R6	PD	VE	R1	R6
SPM1	ZD958	14/3	3/4	29/5	18/7	18/3	2/4	3/6	22/7
	DH9			3/6	22/7			8/6	26/7
	YD629			8/6	25/7			14/6	29/7
SPM2	ZD958	28/3	8/4	5/6	22/7	26/3	8/4	7/6	26/7
	DH9			8/6	27/7			14/6	28/7
	YD629			13/6	29/7			18/6	31/7
SUM1	ZD958	4/5	10/5	27/6	17/8	5/5	12/5	27/6	13/8
	DH9			30/6	20/8			2/7	16/8
	YD629			4/7	22/8			7/7	19/8
SUM2	ZD958	23/5	28/5	11/7	4/9	20/5	25/5	9/7	26/8
	DH9			13/7	8/9			14/7	28/8
	YD629			18/7	11/9			17/7	30/8
AUM1	ZD958	7/7	11/7	24/8	22/10	7/7	12/7	23/8	24/10
	DH9			26/8	29/10			27/8	26/10
	YD629			31/8	5/11			30/8	3/11
AUM2	ZD958	22/7	25/7	9/9	22/11	21/7	25/7	5/9	20/11
	DH9			14/9	30/11			9/9	25/11
	YD629			19/9	30/11			13/9	28/11

Note: S_PM, spring maize; S_UM, summer maize; A_UM, autumn maize; PD, planting date; VE, emergence stage; R1, silking stage; R6, maturity stage.

. At least 50 plants in each subplot, representing average plants, were tagged in the R1 stage. Beginning 7 days after R1, the apical ear of two plants per subplot were sampled every 7–10 days until the kernels reached physiological maturity. After sampling, each ear was enclosed in a plastic bag and transported to our laboratory in an insulated cooler. In addition, 50 kernels after the 10th–15th of each ear were sampled [37]. We measured the dry weight of the 50 kernels for each ear with a balance that had an accuracy of 0.0001 g after drying them in a forced-air oven at 80 °C for at least 72 h.

At maturity, 20 ears from adjacent plants in the middle rows of each replication were harvested manually and threshed; we then measured the dry weight and moisture content, and the GY (g m⁻²) values were calculated at a moisture content of 14%.

2.3.2. Meteorological Data and Date Analysis

We obtained daily meteorological data (maximum temperature (T_max), minimum temperature (T_min), mean temperature (T_mean), Pr, and Ra) for 1 March to 30 November in both 2012 and 2013, from a local weather station (AWS 800, Campbell Scientific, Inc., Logan, UT, USA), at a distance of approximately 200 m from the experimental field (Figure 1).

The diurnal temperature difference (M△T, °C) was calculated as T_max – T_min.

The growing degree days (GDDs, °Cd) were defined according to McMaster and Wilhelm [38] and Yang et al. [39]:

$$GDD={\displaystyle \sum _0^{\mathrm{n}}\frac{T_{\min }+T_{\max }}{2}}-T_{base}$$

(1)

where T_base is 10 °C; the optimum temperature for maize growth (T_opt) is set to 30 °C; and if T_min < T_base, then T_min = T_base, and if T_max > T_opt, then T_max = T_opt [38,39].

The high-temperature stress encountered during maize growth was counted for each day when T_max > 30 °C, using the KDD (°C d), as defined by Lobell et al. [40] and Butler and Huybers [41]:

$$KDD=\left\{\begin{array}{l}0\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}if\hspace{1em}{T}_{\max }\le {30}_{}\\ {\displaystyle \sum _0^n(T_{\max }-30)\hspace{1em}\hspace{1em}if\hspace{1em}{T}_{\max }>30}\end{array}\right.$$

(2)

The kernel filling dynamic was estimated by fitting the sigmoid model in Equation (3) with curve-fitting software, with the average KW plotted against GDD from 7 days after silking until maturity [42]:

$$KW=\frac{a}{1+\exp (-b\times (GDD-c))}$$

(3)

where KW is that measured (mg); GDD (°Cd) is the value after silking; and a, b, and c are estimated parameters—a is the final KW (mg), b is a parameter related to the rate of change in KW, and c is the GDD at the maximum KFR (GDD_KFRmax).

We fitted a bilinear model to estimate the KFR during the linear kernel filling stage, as in Equation (4). The model was fitted with the KW from 10% of its final value to maturity, plotted against the GDD [21,43]:

$$KW=\left\{\begin{array}{l}d+k\times GDD\hspace{1em}\hspace{1em} if\hspace{1em}GDD<f\\ d+k\times f\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}if\hspace{1em}GDD\ge f\end{array}\right.$$

(4)

where KW is that measured (mg), GDD (°Cd) is that after silking, days is the y-intercept (mg), k is the KFR during the linear kernel filling phase (KFR_lkf, mg °Cd⁻¹), and f is the total GDD during kernel filling (GDD_max°Cd).

The onset date of linear kernel filling (GDD_onset°Cd) was determined according to Equation (5), and the duration of the linear kernel filling phase (GDD_lkf°Cd) was determined according to Equation (6), which we fitted with the GDD from GDD_onset to maturity [25].

$$GD{D}_{onset}=-\frac{d}{k}$$

(5)

$$GD{D}_{lkf}=f+\frac{d}{k}$$

(6)

2.4. Statistical Analysis

A two-way analysis of variance was used to analyze the interaction effects of seasons and varieties, and a least significant difference (LSD) test at the p = 0.05 or 0.01 probability level was used to compare treatments, using SPSS software (version 16.0; SPSS Inc., Chicago, IL, USA).

3. Results

3.1. Calendar Time and Variation in Climatic Factors

There were significant differences in maize the growth duration (GD) and kernel filling duration (KFD) among different seasons across years and varieties (Table 1). The S_UM GD was always shortened by 7~16 days and 25~29 days compared to S_PM and A_UM in 2012 and 2013, respectively, while there were no significant differences according to variety. The S_PM and S_UM KFDs were almost 20 days shorter than that of A_UM (47.2 and 49.2 vs. 69.0 days), however.

Across years and varieties, the average growth degree days for total growth duration (GDD_GD) for S_UM was 142.3 and 149.8 °Cd higher than S_PM and A_UM, respectively (

Table 2. Climatic conditions during maize growth duration for different seasons in 2012 and 2013.

Seasons	Varieties	Growing Degree Days (GDD, °Cd)		Mean Daily Temperature (MT, °C)		Mean Temperature Difference (M△T, °C)		Killing Degree Days (KDD, °Cd)		Accumulated Solar Radiation (Ra,MJ m−2)		Accumulated Precipitation (Pr, mm)
		2012	2013	2012	2013	2012	2013	2012	2013	2012	2013	2012	2013
SPM1	ZD958	1404.8	1446.6	23.6	23.4	7.7	9.5	110.5	154.2	1565.8	1762.0	399.8	365.3
	DH9	1474.6	1518.9	21.0	23.6	7.7	9.4	122.4	170.6	1654.4	1855.4	401.5	398.0
	YD629	1526.6	1574.0	21.5	23.8	7.7	9.3	135.7	182.4	1707.8	1924.3	401.5	398.0
SPM2	ZD958	1476.4	1487.4	21.8	24.1	7.6	9.4	127.2	170.6	1613.6	1782.3	400.6	388.4
	DH9	1531.5	1524.1	22.0	24.2	7.6	9.3	144.3	178.4	1703.8	1828.2	400.6	388.4
	YD629	1583.4	1578.2	22.7	24.3	7.6	9.3	152.7	192.8	1764.4	1891.5	400.6	388.4
SUM1	ZD958	1593.4	1511.8	25.1	27.4	7.2	8.5	206.8	257.9	1678.9	1684.2	335.6	252.1
	DH9	1705.3	1564.8	25.3	27.4	7.2	8.6	213.8	274.7	1859.6	1748.0	394.8	252.1
	YD629	1764.5	1618.5	26.0	27.5	7.3	8.7	204.6	292.9	1916.5	1805.6	400.4	252.1
SUM2	ZD958	1664.1	1560.4	28.5	28.0	7.3	8.2	244.8	281.4	1709.8	1700.2	341.8	261.6
	DH9	1795.9	1594.9	28.5	28.0	7.3	8.2	172.5	286.2	1915.2	1734.6	442.4	273.4
	YD629	1853.5	1630.3	27.7	28.0	7.3	8.2	174.1	294.0	1950.9	1767.8	483.6	273.4
AUM1	ZD958	1488.3	1503.6	28.2	25.1	8.2	9.0	150.2	237.7	1614.1	1577.8	292.6	271.9
	DH9	1645.9	1519.0	27.8	25.0	8.3	9.1	85.6	237.7	1845.0	1602.9	453.0	271.9
	YD629	1692.5	1566.5	27.7	24.3	8.4	8.9	113.3	237.7	1913.3	1655.6	464.8	280.2
AUM2	ZD958	1385.0	1416.0	27.0	22.2	8.7	8.9	115.6	186.2	1650.6	1576.1	327.6	226.4
	DH9	1461.1	1429.7	26.2	21.8	8.6	8.9	84.3	186.2	1814.3	1605.5	340.0	226.5
	YD629	1515.7	1436.9	25.3	21.4	8.6	8.9	108.3	186.2	1804.5	1639.4	389.3	226.6
Seasons		15.4 **	7.5 *	21.1 **	21.3 **	94.3 **	44.5 **	40.2 **	34.7 **	29.2 **	332.2 **	ns	72.3 **
Varieties		6.1 *	4.6 *	ns	ns	ns	ns	ns	Ns	45.7 **	62.2 **	5.5 *	ns
Seasons * Varieties		ns	ns	ns	ns	ns	ns	ns	Ns	ns	ns	ns	ns

Note: S_PM, spring maize; S_UM, summer maize; A_UM autumn maize; * difference significant at p < 0.05; ** difference significant at p < 0.01; ns no significant difference.

). Across seasons, the GDD_GD for ZD958 was always significantly lower than that for DH9 and YD629. Across years and varieties, the mean daily temperature from VE to R₆ (MT_GD) for S_UM was 4.3 and 2.1 °C higher than S_PM and A_UM, respectively (p < 0.01), resulting in the KDD_GD being enhanced by 88.5 and 81.2 °Cd (p < 0.01); however, the mean temperature difference (M△T_GD) significantly closed by 0.7~0.9 °C to 7.8 °C. Accumulate solar radiation during growth duration (Ra_GD) for A_UM significantly lower than S_PM and S_UM by 62.9 and 97.7 MJ m⁻² in two years, while the Ra_GD for ZD958 was reduced by 159.9~204.1 MJ m⁻² and 48.6~100.2 MJ m⁻² than DH9 and YD629 (p < 0.01), respectively. ZD958 obtained less accumulate precipitation (Pr_GD) than DH9 and YD 629 from VE to R6 (349.7 vs. 405.4 and 423.4 mm, p < 0.05) in 2012, while S_PM acquired more Pr_GD than S_UM and A_UM in 2013 (387.8 vs. 260.8 and 250.6 mm, p < 0.01).

The average GDD during the kernel filling duration (GDD_KF) for A_UM was 142.3 and 200.9 °Cd less than S_PM and S_UM, respectively, across years and varieties (

Table 3. Climatic conditions during maize kernel filling duration for different seasons in 2012 and 2013.

Seasons	Varieties	Growing Degree Days (GDD, °Cd)		Mean Daily Temperature (MT, °C)		Mean Temperature Difference (M△T, °C)		Killing Degree Days (KDD, °Cd)		Accumulated Solar Radiation (Ra,MJ m−2)		Accumulated Precipitation (Pr, mm)
		2012	2013	2012	2013	2012	2013	2012	2013	2012	2013	2012	2013
SPM1	ZD958	815.4	817.9	27.1	27.8	7.2	7.9	107.2	121.8	801.1	897.1	197.7	110.0
	DH9	822.2	812.7	27.7	28.1	7.3	7.8	119.0	131.8	827.5	905.5	153.0	94.2
	YD629	809.1	789.8	28.3	28.9	7.5	7.8	132.4	143.6	845.6	868.9	130.7	86.5
SPM2	ZD958	814.0	828.8	27.9	28.1	7.4	7.8	123.9	132.2	829.5	925.8	152.0	94.2
	DH9	846.4	771.5	28.4	28.9	7.5	7.8	140.9	139.6	895.7	845.8	130.7	86.5
	YD629	797.6	759.1	28.5	29.0	7.5	7.6	138.9	141.1	862.8	814.4	126.8	86.5
SUM1	ZD958	910.3	835.2	29.2	29.4	7.3	8.4	175.4	186.8	953.3	948.7	165.6	55.4
	DH9	915.2	803.4	29.3	29.7	7.5	8.6	186.6	194.6	972.6	912.6	143.1	55.4
	YD629	877.5	767.7	29.2	29.8	7.4	9.0	171.1	200.5	913.1	880.3	148.7	55.1
SUM2	ZD958	953.1	850.3	28.2	29.4	7.1	8.6	146.5	205.2	973.6	925.1	199.1	125.8
	DH9	972.7	793.9	27.8	29.2	7.2	8.6	140.0	185.4	1013.4	838.7	193.0	137.4
	YD629	933.5	775.2	27.7	29.2	7.4	8.7	140.6	183.3	984.9	821.5	167.5	137.4
AUM1	ZD958	714.7	755.5	21.8	22.0	9.2	9.1	20.0	44.1	834.1	743.3	141.2	216.2
	DH9	751.0	706.5	20.9	21.6	9.2	9.4	19.9	42.7	889.0	733.2	185.1	134.5
	YD629	686.8	700.9	19.9	20.5	9.4	9.0	12.2	31.5	835.2	736.8	195.2	142.8
AUM2	ZD958	604.7	704.1	17.2	18.8	9.5	9.2	0.2	29.1	848.1	844.6	185.0	143.2
	DH9	551.2	669.9	16.0	18.2	9.5	9.2	0.0	29.1	860.2	836.3	133.2	132.7
	YD629	499.7	631.6	15.8	17.6	9.3	9.5	0.0	29.1	763.5	845.4	133.1	90.9
Seasons		24.5 **	33.9 **	49.2 **	96.9 **	357 **	120 **	121 **	530 **	30.6 **	6.24 *	ns	ns
Varieties		ns	8.71 **	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns
Seasons * Varieties		ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns

Note: S_PM, spring maize; S_UM, summer maize; A_UM, autumn maize; * difference significant at p < 0.05; ** difference significant at p < 0.01; ns, no significant difference. The same as follows.

). Across seasons, the GDD_KF values for ZD958 were 39 °C and 62 °Cd significantly higher than for DH9 and YD629, respectively, in 2013. Across years and varieties, the MTs from R₁ to R₆ (MT_KF) for A_UM were 9.0 °C and 9.8 °C lower than S_PM and SM (p < 0.01), respectively. There were significantly differences in M△T during kernel filling (M△T_KF) among seasons, which for A_UM were significantly increased by 1.5~1.9 °C and 0.6~2.0 °C than S_PM and S_UM, respectively. Both years showed the same tendency in the average KDD during kernel filling (KDD_KF) for S_UM, in that they were significantly higher than they were for S_PM and A_UM, and A_UM’s KDD_KF values were reduced by 109.6 °C and 154.8 °Cd than S_PM and S_UM, respectively (p < 0.01). The Ra after silking (Ra_KF) for S_UM significantly increased by 125 and 130 MJ m⁻² compared to S_PM and A_UM, respectively, in 2012. There were no significant differences in Ra_KF between S_PM and S_UM ((876 and 888 MJ m⁻²), which were significantly higher than A_UM in 2013 (790 MJ m⁻²). There were no significant differences in Pr during kernel filling across seasons or varieties in two years.

3.2. Grain Yield and Final Kernel Weight

Across seasons and varieties, the S_pM GYs were significantly increased by 200.8 and 25.4 g m⁻², and 101.7 and 32.6 g m⁻² higher than S_UM and A_UM, respectively, in 2012 and 2013 (Figure 2). The GYs for YD629 were always the lowest compared to ZD958 and DH9 in two years (p < 0.01), while DH9 obtained significantly higher GYs than ZD958 in 2013 (713.3 vs. 814.2 g m⁻²), across the three seasons.

Across the three varieties, among the three seasons, the A_UM KWs were the largest in 2012 (309.3 vs. 287.5 vs. 320.9 mg kernel⁻¹) and 2013 (275.0 vs. 263.3 vs. 346.8 mg kernel⁻¹; p < 0.01; Figure 3). The kernel dry matter accumulation dynamics for ZD958 were always the lowest in S_UM, while there were no significant differences in S_PM and S_UM for DH9 and YD629. The KWs for YD629 were significantly higher than for ZD958 and DH9 in 2012 (322.9 vs. 297.8 and 297.2 mg kernel⁻¹, respectively), while there was a significant difference between ZD958 and DH9 in 2013 (283.3 vs. 305.5 mg kernel⁻¹), across the three seasons.

3.3. Maize Kernel Filling Characteristic Parameters

Kernel filling rate (KFR) during the linear kernel filling period (KFR_lkf, mg °Cd⁻¹) varied significantly among seasons and varieties only in 2012 (

Table 4. KFR_lkf, GDD_onset, GDD_lkf, GDD_KFRmax, and GDD_max for different maize seasons in 2012 and 2013.

Years	Seasons	KFRlkf (mg °Cd−1)			GDDonset (°Cd)			GDDlkf (°Cd)			GDDKFRmax (°Cd)			GDDmax (°Cd)
		ZD958	DH9	YD629	ZD958	DH9	YD629	ZD958	DH9	YD629	ZD958	DH9	YD629	ZD958	DH9	YD629
2012	SPM1	0.547 ± 0.03 bc	0.518 ± 0.03 b	0.745 ± 0.04 a	185 ± 9 a	186 ± 9 a	259 ± 13 a	575 ± 29 a	513 ± 26 c	462 ± 23 bc	443 ± 22 a	455 ± 23 b	498 ± 25 a	720 ± 36 a	699 ± 35 b	721 ± 36 a
	SPM2	0.572 ± 0.02 b	0.518 ± 0.02 b	0.61 ± 0.02 b	190 ± 7 a	186 ± 7 a	212 ± 7 b	536 ± 19 b	513 ± 18 c	440 ± 15 c	467 ± 16 a	460 ± 16 b	465 ± 16 ab	726 ± 25 a	699 ± 24 b	653 ± 23 b
	SUM1	0.526 ± 0.01 cd	0.369 ± 0.01 d	0.604 ± 0.02 b	161 ± 4 b	116 ± 3 c	181 ± 5 c	463 ± 12 c	830 ± 21 a	474 ± 12 bc	397 ± 10 b	544 ± 14 a	424 ± 11 b	623 ± 16 b	946 ± 24 a	655 ± 16 b
	SUM2	0.509 ± 0.02 d	0.464 ± 0.02 c	0.479 ± 0.02 c	160 ± 7 b	179 ± 8 ab	158 ± 7 cd	461 ± 19 c	594 ± 25 b	582 ± 24 a	426 ± 18 ab	481 ± 20 ab	462 ± 19 ab	621 ± 26 b	773 ± 32 b	740 ± 31 a
	AUM1	0.534 ± 0.02 cd	0.612 ± 0.02 a	0.715 ± 0.03 a	115 ± 5 c	164 ± 7 b	197 ± 8 bc	588 ± 24 a	553 ± 22 bc	507 ± 20 ab	415 ± 17 ab	460 ± 18 b	487 ± 19 a	703 ± 28 a	717 ± 29 b	704 ± 28 ab
	AUM2	0.659 ± 0.03 a	0.595 ± 0.02 a	0.719 ± 0.03 a	125 ± 5 c	84 ± 3 d	145 ± 6 d	469 ± 19 c	495 ± 20 c	401 ± 16 c	369 ± 15 b	353 ± 14 c	373 ± 15 c	594 ± 24 b	579 ± 23 c	545 ± 22 c
F values
Seasons (S)		11.4 **			5.21 *			ns			ns			ns
Varieties (V)		8.75 **			ns			ns			ns			ns
S×V		ns			ns			ns			ns			ns
2013	SPM1	0.566 ± 0.03 b	0.591 ± 0.03 ab	0.457 ± 0.02 c	157 ± 8 c	174 ± 9 a	148 ± 7 a	465 ± 23 c	467 ± 23 b	610 ± 31 ab	398 ± 20 a	421 ± 21 ab	453 ± 23 a	622 ± 31 b	641 ± 32 ab	758 ± 38 a
	SPM2	0.529 ± 0.02 b	0.545 ± 0.02 b	0.392 ± 0.01 d	182 ± 6 b	146 ± 5 b	167 ± 6 b	471 ± 16 c	499 ± 17 b	710 ± 25 a	433 ± 15 a	387 ± 14 b	439 ± 15 a	653 ± 23 b	644 ± 23 ab	777 ± 27 a
	SUM1	0.563 ± 0.01 b	0.555 ± 0.01 b	0.618 ± 0.02 a	210 ± 5 a	146 ± 4 b	148 ± 4 a	420 ± 11 d	463 ± 12 b	393 ± 10 d	433 ± 11 a	375 ± 9 b	376 ± 9 b	630 ± 16 b	608 ± 15 b	541 ± 14 c
	SUM2	0.361 ± 0.02 c	0.57 ± 0.02 ab	0.514 ± 0.02 b	153 ± 6 c	138 ± 6 b	147 ± 6 a	651 ± 27 a	466 ± 20 b	482 ± 20 c	424 ± 18 a	366 ± 15 b	386 ± 16 b	804 ± 34 a	604 ± 25 b	628 ± 26 b
	AUM1	0.793 ± 0.03 a	0.584 ± 0.02 ab	0.636 ± 0.03 a	201 ± 8 a	90 ± 4 c	131 ± 5 a	425 ± 17 d	625 ± 25 a	555 ± 22 b	433 ± 17 a	444 ± 18 a	420 ± 17 ab	626 ± 25 b	715 ± 29 a	686 ± 27 ab
	AUM2	0.551 ± 0.02 b	0.655 ± 0.03 ab	0.691 ± 0.03 a	110 ± 4 d	174 ± 7 a	153 ± 6 a	551 ± 22 b	489 ± 20 b	487 ± 19 c	402 ± 16 a	450 ± 18 a	444 ± 18 a	661 ± 26 b	663 ± 27 ab	640 ± 26 b
F values
Seasons (S)		ns			ns			ns			9.91 **			ns
Varieties (V)		ns			ns			ns			ns			ns
S×V		ns			ns			ns			6.02 *			4.16 *

Note: S_PM, spring maize, S_UM, summer maize, A_UM, autumn maize, KFR_lkf, kernel filling rate during linear kernel filling period; GDD_onset, growing degree days for onset of linear kernel filling; GDD_lkf, growing degree days for duration of linear kernel filling; GDD_KFRmax, growing degree days at maximum kernel filling rate; GDD_max, total growing degree days for kernel filling. Different letters after values indicate significant differences at p < 0.05 levels, and * = p < 0.05, ** = p < 0.01, and ns = p >0.05, respectively.

). We saw the same tendency in KFR_lkf among three varieties: S_UM’s KFR_lkf was significantly lower than that of S_PM and A_UM in 2012 (p < 0.05); meanwhile, in each season, the KFR_lkf values for YD629 were higher than ZD958 and DH9 (p < 0.05). In 2013, A_UM obtained the highest KFR_lkf among three varieties (p < 0.05), while the KFR_lkf for ZD958 in S_PM was higher than that of S_UM (0.548 vs. 0.462 mg °Cd⁻¹), and that for YD629 was lower in S_PM than S_UM (0.425 vs. 0.566 mg °Cd⁻¹; p < 0.05). In S_PM, YD629’s KFR_lkf was lower than that of ZD958 and DH9 (0.425 vs. 0.548 and 0.568 m g °Cd⁻¹; p < 0.05), while the KFR_lkf in S_UM for ZD958 was the lowest (0.460 vs. 0.563 and 0.566 mg °Cd⁻¹; p < 0.05).

Across varieties, there were significant differences in the GDD for the onset of the linear kernel filling (GDD_onset, °Cd) among seasons only in 2012, with no significant differences in 2013 (Table 4). All varieties showed the same tendency, in that S_PM had the highest GDD_onset compared to S_UM and A_UMin in 2012 (p < 0.05), while DH9 showed the same tendency in 2013. The GDD_onset was lower in A_UM for ZD958 and S_PM forYD629 than in the other two seasons in 2013 (p < 0.05). In 2012, the GDD_onset for YD629 in S_UM was lower than in S_PM and A_UM, and in S_UM, the GDD_onset values for ZD958 and YD629 were higher than that for DH9 (161 °C and 170 °C vs. 148 °Cd). However, in the S_PM, the GDD_onset for ZD958 and DH9 was higher than for YD629 (170 °C and 160 °C vs. 108 °Cd; p < 0.05), while ZD958’s GDD_onset in S_UM and A_UM was higher than for the S_PM in 2013 (p < 0.05).

There were no significant differences in GDD during linear kernel filling (GDD_lkf, °Cd) across the seasons, varieties, and years (Table 4). The variation in the GDD_lkf for different varieties among seasons was different in two years. The GDD_lkf for ZD958 was lowest during S_UM than S_PM and A_UM in 2012 (462 vs. 555 and 529 °Cd), but it was the highest in 2013 (535 vs. 468 and 488 °Cd). On the contrary, the GDD_lkf in S_UM was always higher than in S_PM and A_UM for DH9 (712 vs. 513 and 524 °Cd) and YD629 (528 vs. 451 and 454 °Cd) in 2012, despite being the lowest for DH9 (464 vs. 483 and 557 °Cd) and YD629 (437 vs. 660 and 521 °Cd) in 2013. In S_PM and A_UM, the GDD_lkf values for ZD958 and DH9 were higher than for YD629, while the GDD_lkf for ZD958 was lowest, and for DH9, it was the highest in S_UM in 2012. In contrast, in 2013, the GDD_lkf for ZD958 and DH9 in S_PM was lower than YD629, and the GDD_lkf for ZD958 was highest in S_UM and lowest in A_UM.

The GDD at the maximum KFR (GDD_KFRmax, °Cd) did not significantly differ among seasons and varieties in 2012, while it was lower in S_UM than in S_PM and A_UM across varieties in 2013 (p < 0.05; Table 4). All varieties showed the same variation in 2012, in that the GDD_KFRmax was lower in A_UM than in S_PM and S_UM (p < 0.05), while the GDD_KFRmax values in S_UM for DH9 and YD629 were lower than in both S_PM and A_UM in 2013 (p < 0.05). The variation among different varieties during one season showed that the GDD_KFRmax values for YD629 in S_PM and A_UM were higher than ZD958 and DH9 in 2012 (p < 0.05), which had similar values to the S_PM in 2013. During S_UM, DH9 had the highest GDD_KFRmax in 2012; however, this was highest for ZD958 in 2013 (p < 0.05).

The total GDD values for kernel filling duration (GDD_max, °Cd) differed among years, seasons, and varieties. There were no significant differences in the GDD_max among the varieties for S_PM and A_UM in 2012 or for A_UM in 2013. ZD958’s GDD_max was higher in S_PM than S_UM and A_UM in 2012 (723 vs. 622 and 648 °Cd), but that it was highest in S_UM in 2013 (717 vs. 637 and 643 °Cd; p < 0.05; Table 4). The variation in the GDD_max during S_UM changed for DH9, with it being the highest across three seasons in 2012 (860 vs. 699 and 648 °Cd) but the lowest in 2013 (606 vs. 643 and 689 °Cd). The GDD_max for YD629 in S_PM and S_UM was higher than in A_UM in 2012 (687 vs. 698 and 625 °Cd), whereas in S_PM and A_UM in 2013, it was higher than in S_UM (768 vs. 663 and 584 °Cd). Among varieties, the GDD_max for YD629 was higher than for ZD958 and DH9 in S_PM in 2013 (768 vs. 637 and 643 °Cd), while during the S_UM season, the GDD_max for DH9 was the highest in 2012 (860 vs. 622 and 698 °Cd), and for ZD958, it was the highest in 2013 (717 vs. 606 and 684 °Cd).

3.4. Relationships among GY, KW, Kernel Filling Characteristic Parameters and Climatic Factors

A correlation analysis was performed to explore how climatic conditions affect the GY, KW, and kernel filling characteristic parameters. The temperature index during maize growth duration, including the GDD_GD, MT_GD, and KDD_GD, always played significantly negative roles on GY, as well as Ra_GD, while there were no significant correlations between the GY, M△T_GD, and Pr_GD (

Table 5. Correlation coefficients of GY, KW, KFR_lkf, GDD_onset, GDD_lkf, GDD_KFRmax, GDD_max, and climatic factors.

	GDD	MT	M△T	KDD	Ra	Pr
GY	−0.384 *	−0.528 **	−0.152	−0.315 *	−0.336 *	0.257
KW	−0.421 *	−0.592 **	0.400 *	−0.673 **	−0.589 **	0.397 *
KFRlkf	−0.565 **	−0.563 **	0.492 **	−0.573 **	−0.523 **	0.170
GDDonset	0.345 *	0.290	−0.343 *	0.230	0.210	0.110
GDDlkf	0.240	0.110	−0.180	0.070	0.110	0.110
GDDKFRmax	0.429 **	0.190	−0.409 *	0.080	0.210	0.260
GDDmax	0.440 **	0.270	−0.374 *	0.190	0.230	0.180

Note: Correlation coefficients significant at * = p < 0.05, and ** = p < 0.01. GDDs, growing degree days; MT, mean daily temperature; M△T, mean temperature difference; KDDs, killing degree days; Ra, accumulated solar radiation; Pr, accumulated precipitation; GY, grain yield; KW, kernel weight; KFR_lkf, kernel filling rate during linear kernel filling period; GDD_onset, growing degree days for onset of linear kernel filling; GDD_lkf, growing degree days for duration of linear kernel filling; GDD_KFRmax, growing degree days at maximum kernel filling rate; GDD_max, total growing degree days for kernel filling.

). The regression analysis showed that maize’s GY linearly decreased by 0.314 g m⁻² and 0.634 g m⁻², respectively, when the GDD_GD and KDD_GD increased by 1 °C d within the studied range (Figure 4A,C). If the MT_GD was lower than 25.8 °C, the GY was initially reduced by 42.4 g m⁻² for each 1 °C increase, and then the GY was stabilized at 703.5 g m⁻² when the MT_GD exceeded 25.8 °C (Figure 4 B). The same tendency occurred in GY and Ra_GD: the GY was higher than 722.5 g m⁻² if the Ra_GD was less than 1805 MJ m⁻², and the GY decreased by 0.499 g m⁻² (MJ m⁻²)⁻¹ with Ra_GD increased (Figure 4D).

During the maize kernel filling stage, the GDD_KF, MT_KF, KDD_KF, and Ra_KF had significant negative correlations with the KW, whereas the M△T_KF and Pr_KF played significant positive roles (Table 5). The further regression analysis showed a quadratic response of KW to GDD_KF and MT_KF, with the KW increasing pristinely and then decreasing when the GDD_KF and MT_KF exceeded 538.4 °Cd and 27 °C, respectively (Figure 5A,B). On the contrary, the KW initially decreased when the M△T_KF increased, and then increased steeply when the M△T_KF exceeded 8.1 °C (Figure 5C). The KWs were reduced linearly by 0.441 mg kernel⁻¹ for each 1 °Cd increase in KDD_KF beyond 44.3 °Cd, especially in S_PM and S_UM (Figure 5D). The final KW was greater than 273.9 mg kernel⁻¹ if the Ra_KF was less than 925.3 MJ m⁻², and the KW increased by 0.399 mg kernel⁻¹ (MJ m⁻²)⁻¹ as the Ra_KF decreased (Figure 5E). The KW exhibited a positive linear relationship with Pr_KF, with it increasing by 0.338 mg kernel⁻¹ for each 1 mm increase within the studied range (Figure 5F).

The final KW was principally determined by the KFR; the KFR_lkf exhibited significant negative correlations with GDD_KF, MT_KF, KDD_KF, and Ra_KF; and the M△T_KF had a significant positive correlation with KFR_lkf (Table 5). The regression analysis showed that KFR_lkf had significant, linear, negative responses to GDD_KF, MT_KF, and Ra_KF. The KFR_lkf decreased sharply by 0.0523, 0.012, and 0.0741 mg °C d⁻¹ when the GDD_KF, MT_KF, and Ra_KF increased by 100 °Cd, 1 °C, and 100 mm, respectively (Figure 6A,B,E). There were bilinear relationships between the M△T_KF, KDD_KF, and KFR_lkf; the KFR_lkf increased by 0.169 mg °Cd⁻¹ for each 1 °C increase in the M△T_KF if the M△T_KF exceeded 8.63 °Cd (Figure 6C), but when the KDD_KF was greater than 44.0 °Cd, the KFR_lkf decreased by 0.0965 mg °C d⁻¹ for each 100 °Cd increase in the KDD_KF (Figure 6D).

Moreover, the GDD_KF had significant positive effects on the GDD_onset, GDD_KFRmax, and GDD_max, which were significantly and negatively affected by the MT_KF. The GDD_lkf had no relationships with any climatic factors (Table 5). When the GDD_KF increased by 1 °Cd, the GDD_onset, GDD_KFRmax, and GDD_max increased linearly by 0.188 °C, 0.171 °C, and 0.328 °Cd, respectively, and the GDD_onset values were stabilized if the GDD_KF values were higher than 825 °Cd, (Figure 7A,C,E). The GDD_onset stabilized at 169 °Cd if the M△T_KF was less than 8.4 °C, and then it decreased by 33.3 °Cd when the M△T_KF increased by 1 °C (Figure 7B). However, the GDD_KFRmax decreased by 33.4 °Cd for each 1 °Cd increase in the M△T_KF if the M△T_KF was lower than 8.6 °C (Figure 7D), while with each 1 °Cd increase in the M△T_KF GDD_max always decreased linearly by 33.8 °Cd (Figure 7F).

4. Discussion

With global climate change, the mean temperature has increased by 1.2 °C since 1960, [44] and the severity of drought stress has increased [45]; together, the climate change and drought stress cause negative effects on maize due to more frequent extreme climatic conditions [46,47], which have a significant influence on maize kernel filling and yield [47]. In Hou et al.’s studies, the maize yield was reduced by 0.83 t ha⁻¹°C⁻¹ with the mean temperature increase in different maize regions of China [6]. Therefore, clarifying the impacts of climatic conditions on maize yield and KW after pollination is vital, which will aid the expansion of breeding and targeted management strategies. Maize growth duration is determined by the variation in weather conditions and each variety’s growth degree days requirements [48,49]. Our results show that the GD and KFD lengthen when the planting date is delayed due to the lower MT_KF and GDD_KF requirements; these results confirmed the findings of studies on the North China Plain and in Central China that maize growth duration is shortened by increased temperature [2,35,50]. Moreover, in our study, we found that high-temperature stress (KDD_KF) occurs more frequently in S_UM kernel filling duration possibly because of the higher Ra_KF, while M△T_KF and Pr_KF are increased when the planting date is delayed.

The maize grain yield is controlled by the kernel number per unit area and individual KW [51,52]; the kernel number decreased by 42–77% under drought stress from tassel emergence to 6 days after silking [53], and the maize yield decreased by 0.85 t ha⁻¹ with accumulated photosynthetically active radiation reduced by 100 MJ [6]. The different planting date is an effective approach to examine the climatic conditions’ effects on maize growth and KW, without variation in the soil environment [5,13,34]. Our results showed that S_PM always obtained a higher GY than S_UM and A_UM, and S_UM GY always obtained the lowest in Central China, because the higher GDD_GD, MT_GD, and KDD_GD during S_UM growth duration had a significantly negative impact on GY, especially the higher temperature stress result in lowest KW for S_UM. The results were confirmed by that, the temperature increased with the plant date delayed, which resulted in maize GY redcued New Mexico State, US [5], and the S_UM‘s GY was highest due to the suitable climatic conditions in the North China Plain [35].

Under drought stress and higher temperature stress, maize’s KW was reduced by 5.0–11.0% and 18.0–37.6% in the field and the laboratory, respectively [20]; the kernel number decreased by 42–77% under drought stress after silking [53]. The kernel numbers were constant for different planting dates, indicating that KW is the dominant factor for GY on the North China Plain [35]; however, in a previous report, we found that the kernel number and KW are equally important for GY in Central China [2]. Maize KW is determined by the combination of KFR and linear KFD [8,10,11], which always vary according to the weather conditions [5,54]. In the present study, the KW was lowest for S_UM and highest for A_UM. The maize linear kernel filling period (KFR_lkf) and KFD initially declined and then increased with the planting date; in contrast, GDD_KF exhibited the opposite tendency. These findings are not consistent with previous studies, which found that KW decreases linearly with the planting date in temperate regions [11,13] and initially increased and then decreased in cool regions and on the North China Plain [35,55]. These findings might be explained by three factors: firstly, S_UM has a shorter KFD than A_UM; secondly, thermal parameters, such as GDD_KF, MT_KF, and KDD_KF, and Ra_KF affect KW negatively; and finally, KW exhibits positive relationships with M△T_KF and Pr_KF. All of these effects affirm that the maize KW is closely related to the KFR during the linear kernel filling period [8,11,56], and the KW decreases when the KFD is shortened by drought stress [32,54]; plant time delays always especially reduce negative drought effects on the KFR and KW in temperate Argentina [57], and it has been confirmed that there is no relationship with Pr in irrigated systems [35].

The maize kernel filling progress was determined by the dry matter supplied from the source [58] and was expressed by different kernel filling parameters [59], which always relate to the climatic conditions during KFD, especially the effective KFD [7,9,60]. The kernel growth rate correlates with temperature and temperature range, and the effective grain-filling period correlates with accumulated temperature and Ra on the North China Plain [35]. According to our regression analyses, KFR_lkf exhibits significant negative correlations with GDD_KF, MT_KF, KDD_KF, and Ra_KF. In addition, KFR_lkf increases by 0.169 mg °Cd⁻¹ for each 1 °C increase when M△T_KF is larger than 8.63 °Cd. This finding confirms that KFR is the main factor affecting KW, thus leading us to conclude that maize KW was determined by KFR and the duration of the phases [61]. Moreover, GDD_KF has positive linear relationships with GDD_onset, GDD_KFRmax, and GDD_max, while MT_KF has a negative relationship. This is because high-temperature stress may be caused by higher Ra in S_UM after silking, thus reducing the amount of radiation that can be used for photosynthesis, and therefore a shorter KFD [11,20], while less Pr causes drought stress in Central China [7,32], thereby decreasing KW. These findings are consistent with those of previous studies [13,35].

5. Conclusions

Our study demonstrated that, during S_UM growth and the kernel filling duration, there were higher temperatures, which resulted in more GDDs than for S_PM and A_UM, and there were a greater number of KDDs, while after silking, S_PM and S_UM obtained higher GDDs and seriously higher KDDs than A_UM. GDD, MT, and Ra played significantly negative roles on KFR_lkf, resulted in S_UM KW were the lowest and A_UM KW were the highest. Therefore, the GY linearly decreased, while GDD, KDD, MT, and Ra increased, respectively. We conclude that the different seasons of maize GY manifestations of the S_PM were highest, and for S_UM, they were the lowest; therefore, farmers should plant S_PM first and then A_UM in an attempt to minimize plant S_UM in Central China.