Effects of Ecotypes and Reduced N Fertilization on Root Growth and Aboveground Development of Ratooning Sorghum × Sudangrass Hybrids

Abstract

Reduced N input while maintaining biomass production of sorghum × sudangrass hybrids (Sorghum bicolor L. × Sorghum sudanense; SSG) is essential; however, its effects on root sustainability and photosynthetic capacity during the ratooning period are not well defined in a multiple harvests system. The physiological response and root morphology of SSG were investigated under different N application levels during the ratooning period in a two-year field experiment. Treatments were all combinations of two ecotypes (late-flowering, Greenstar; early-flowering, Honeychew) and four N levels (0, 50, 100, 150 kg N ha⁻¹). The total root length, surface area, volume, tips, and dry matter (DM) were significantly influenced by both ecotype and N level, with Greenstar outperforming Honeychew. Specifically, Greenstar’s root length increased by up to three times with reduced N application (50 kg N ha⁻¹), while Honeychew showed significant root length increases only at higher N levels (100 and 150 kg N ha⁻¹). Our data support the conclusion that a low level of N (50–100 kg N ha⁻¹) was the optimal rate for ratooning root sustainability. The findings highlight the critical role of root development in sustaining biomass production and suggest that the late-flowering ecotype, Greenstar, is more suitable for a multiple harvests system with a robust root system.

1. Introduction

Sorghum × sudangrass hybrids (Sorghum bicolor L. × Sorghum sudanense) are a multipurpose crop for use as forage or dedicated cellulosic biofuels [1,2]. Sorghum × sudangrass hybrids have high biomass yields and tolerance to various environmental stresses such as high temperature, drought, and soil nutrient deficiency while requiring low agricultural input. For instance, forage sorghums generally have heat and low moisture tolerance and can be productive in areas of annual rainfall as low as 400–650 mm [3], including Southern USA and other semiarid subtropical environments [4]. Previous research indicated that hybrid forage sorghums and sorghum × sudangrass hybrids have the potential for increased biomass yields even under stressful growing conditions in temperate regions [5,6]. The physiological basis for the superior performance of sorghum hybrids has been ascribed to a sustained greater carbon exchange rate over wider environmental conditions with minimum management, but that may be contingent upon the parental genetic background more than heterosis [7].

Unlike the distinctive one-time harvest of annual row crops such as maize, wheat, rice, or soybean, multiple harvests are possible during the vegetative period (ratoon harvests) when sorghum hybrids are used for livestock feed [8,9,10]. To achieve a multiple harvests farming system, photoperiod-sensitive sorghums that remain vegetative for long periods of time are available [6,11,12,13,14]. This photosensitivity response to day-length delays heading until the day-length is from less than 13.2 h [12] to 12.2 h [13].

Even though the nutrient requirements for forage sorghum are relatively low compared with maize, adequate nutrient management is needed to sustain biomass production in multiple harvest conditions where N removal occurs. After the first harvest is conducted, the ratooning crop must distribute captured mineral nutrients to the newly growing above- and belowground parts. This will impact the subsequent pace of capturing N, or the uptake of resources from the soil. Previous research from our research group revealed that N-agronomic efficiency (NAE) was greatest at 50 kg N ha⁻¹ after the first summer harvest and then decreased with higher N application rates for a sorghum × sudangrass hybrid [6]. There is limited research on the effects of multiple harvests on bioenergy crop growth and development throughout the entire ratooning season with reduced N fertilization, as the major focus of multiple harvest research has been on perennial forage grasses [15,16,17].

While plant roots play an essential role in anchoring the plant and the acquisition of nutrients and water, the corresponding trait responses to environmental conditions are complex and not well defined. This is due to overall difficulties in root phenotype monitoring and the limitations of existing methods for evaluating the characteristics of roots and environmental conditions [18]. Studies focusing on roots and their role in nutrient uptake are essential to support the development of management strategies to increase crop production while improving nutrient-use efficiency [19]. There has been some research describing the sorghum crown root angle effect on water use efficiency and yield [20,21]; however, there is no previous research evaluating ratooning effects on existing roots for sorghum × sudangrass hybrid. There have been some efforts to achieve a better understanding of root development under field conditions. For instance, the Winrhizo system (Regents Instruments Inc., QC, Canada) has the capability of quantifying the root length, diameter, surface area, and total volume of the root, which are acquired from field conditions [22,23,24]. The current research on root morphological changes has been conducted primarily in controlled environments with seedlings and pot experiments [25,26]. However, root growth in controlled conditions is often different than that observed under field conditions due to several abiotic and biotic factors that vary widely [26,27,28,29,30]. Thus, to better understand overall plant response to summer harvest and following regrowth, it is necessary to investigate the seasonal change of sorghum × sudangrass root morphology during the ratooning period in field conditions.

To understand the difference in the agronomic responses of ratooning sorghum × sudangrass hybrids, our research group previously investigated ratooning sorghum yield over 15 weeks for 3 years with early and late flowering ecotypes. We found that the practical range for additional N application rate can be reduced as much as between 50 and 100 kg ha⁻¹, and late-flowering cultivars tend to have high aboveground biomass yield and agronomic performance [6]. However, physiological response and root sustainability have not been reported for ratooning sorghum plants in response to additional N applications. Thus, the objectives of the current study were to evaluate the overall changes of root morphological traits and photosynthetic capacity in response to reduced N applications during the ratooning of early and late flowering sorghum × sudangrass hybrids.

2. Materials and Methods

2.1. Research Site and Field Management

The experiment was carried out in the Gyeongsang National University Research Farm in Jinju, Korea (35°14′39″ N 128°09′19″ E) for 2 years (2018–2019). Historically, the research site had been used for lowland rice production and was converted to upland for the current research. Soil samples were collected at a 30-cm depth using a soil auger (diameter 2.54 cm) prior to planting each year. Up to 20 samples were collected in a “W” pattern throughout the field and then composited. The chemical properties of soil are presented in

Table 1. Soil chemical properties of the research site from 2018 to 2019.

Year	pH	EC	OM	Total N	Av. P2O5	K	Ca	Mg
	(1:5)	(dS m−1)	(g kg−1)	(g kg−1)	(mg kg−1)	(cmol+ kg−1)
2018	6.3	0.20	7.0	0.37	104	0.22	2.84	0.49
2019	6.3	0.22	10.9	0.34	37	0.21	3.49	0.40

. Overall soil organic matter and mineral nutrient concentrations were fairly poor, including the total N concentration. These data indicate that the fertilizer application did not affect the overall soil N, P, and K pools, except for available P₂O₅ in 2019.

At the beginning of the growing season, the field was plowed and then rotovated multiple times each year. A pre-planting application of 100 kg ha⁻¹ of N was made using mixed fertilizer (21-17-17) that was manually broadcasted and then incorporated into the soil using a tractor-mounted rotovator. Then, 30 kg ha⁻¹ of sorghum × sudangrass hybrid seeds (1000-seed weight of 25 g) were planted using a hand-pushed two-row manual seeder (AP-2, Agritecno Yazaki Korea, Cheongju, Republic of Korea) with 15-cm inter-row and 5.5-cm intra-row spacing. To investigate ecotype differences for root development during ratooning, two distinctive ecotypes were tested in both years: the early-flowering type Honeychew and the late-flowering type Greenstar. Major field management operations and the dates of flowering for each ecotype are shown in

Table 2. Major field management operation schedule and developmental stages for two years.

Operations and Developmental Stages		2018	2019
Plowing, leveling		9 May	7 May
Preplant fertilizer application		9 May	7 May
Sowing		11 May	8 May
Summer harvest a		2 August	22 July
Additional N fertilizer application		10 August	26 July
Root sampling	Early	-	19 August (DAS b 28)
	Middle	17 October (DAS 76)	25 September (DAS 65)
	Late	23 November (DAS 113)	30 October (DAS 100)
Boot stage
(Greenstar/Honeychew)		N.D. c/4 October	21 October/20 September
Full flowering stage
(Greenstar/Honeychew)		N.D./17 October	28 October/4 October

^a The yield data can be found in a previously published paper [6]. ^b DAS, Days after summer harvest. ^c N.D., Not Detected.

. There were no significant weed issues due to the good early germination and fast-growing characteristics of sorghum × sudangrass hybrids. Irrigation was not required due to sufficient rainfall during the season.

The weather data during the growing season were collected from the Automated Weather System station, operated by the Korea Meteorological Agency (https://data.kma.go.kr/, accessed on 1 February 2020). Rainfall events during the early regrowth period were distinctively different for the 2 years, especially from the summer harvest through the first 4-week ratooning period (341 vs. 11 mm in 2018 and 2019, respectively). The experimental site received a similar amount of rainfall during the overall regrowth period; a total of 664 (2018) and 597 (2019) mm (Figure 1). Monthly mean temperature and GDD data indicate that 2018 had a warmer spring and summer, while 2019 had a warmer fall.

2.2. Summer Harvest Management and N Treatments

The experimental design was a randomized complete block (RCBD) with a split-plot limitation to randomization with four replications; the main plots were the ecotypes and the subplots were the N levels. The size of each subplot was 6 m × 7 m. The entire field was summer harvested (Table 2) using a gas-powered handheld brush cutter (KG350S, Honda, Tokyo, Japan), leaving a target stubble height of 5 cm; it is recommended to avoid picking up soil during machine harvest. Cut biomass was removed from the field immediately. To implement N level treatments during ratooning, 0 (control), 50, 100, and 150 kg ha⁻¹ of N fertilizer (urea) were broadcasted with a handheld spreader (GE-US 18 Li, Einhell, Landau an der Isar, Germany) within a week after the summer harvest in each year.

2.3. Ratooning Chlorophyll Content and LAI

Chlorophyll content was measured on fully developed uppermost leaves, excluding the flag leaf, of 10 plants per plot using a CCM-300 (Opti-Sciences Inc., Hudson, NH, USA), which records chlorophyll content on a surface basis (mg m⁻²). Chlorophyll content data were collected at approximately 15-day intervals from days after summer harvest (DAS) 30 to 75 in 2018, and it extended from DAS 15 to 90 in 2019. Leaf area index (LAI) measurements were obtained from the center rows to minimize the border effect. The LAI of each plot was estimated using a LAI2200-C canopy analyzer (Li-Cor BioSciences, Lincoln, NE, USA) with a 90° view angle cap and six under-canopy readings in a “W” pattern. LAI data were collected at approximately 15-day intervals from DAS 30 to 75 in 2018, and from DAS 15 to 75 in 2019.

2.4. Ratooning Root Sampling

In 2018, after the different levels of N were applied, root sampling occurred at two different times (17 October and 23 November) while focused on the fully developed root morphology of the ratooning crop. After the first year of root characteristics analysis in 2018, we decided to further assess root morphological change during the 2019 growing season. Specifically, to describe root development before summer harvest, root sampling occurred at three different times (3 June, 2 July, and 22 July), and to extend our understanding of the root development of ratooning plants, regrowth plant root sampling was carried out at three different times (19 August, 25 September, and 30 October).

To obtain representative roots from each plot, we first made a circle (r = 15 cm) around the targeted plant, showing representative aboveground growth in the middle rows. Then, three shovels were inserted simultaneously at a 90-degree angle into the soil to a depth of 30 cm to collect roots in the targeted area (Figure 2). A total of 7 plant samples were collected from each plot, and the 5 most representative samples were chosen for further lab analysis. Each plant was separated into shoots and roots, and the roots were separated from the soil by adding tap water and straining the root out of the soil with a fine strainer (500 μm) [20]. The washed roots were stored in plastic zipper bags that were partially filled with water and stored at 4 °C for a maximum of 3 days. Prepared root samples were placed on a 30 cm × 40 cm acrylic tray, and enough water was added to submerge the samples to achieve clear resolution during the scanning. To obtain root images, samples were scanned (Epson Expression 12000XL, Seiko-Epson Corp. Suwa, Japan) at 400 dpi. The scanned root images were analyzed using the Winrhizo Pro ver. 2017a software program (Regent Instruments Inc., Sainte-Foy, QC, Canada), obtaining the root length, root surface area, root volume, and tips; tips are the number of root endings. The workflow of root morphological analysis is described in Figure 2.

2.5. Statistical Methods

The PROC UNIVARIATE function of SAS 9.4 software (SAS Institute Inc., Cary, NC, USA) was used to assess the normality of data distribution. In 2018 and 2019, after summer harvest measurements (ratooning period chlorophyll content, LAI, total root length, total root surface area, total root volume, tips and root DM) were analyzed with ANOVA using the PROC MIXED model. N levels and ecotype were treated as fixed effects, and DAS was treated as a repeated measure. The block and block × ecotype were treated as random effects. The year was not included in the statistical analysis because sorghum × sudangrass is an annual crop and randomization occurred in each year. N level treatments were compared using polynomial contrast [6]. The least significant difference (LSD) test (p = 0.05) was utilized for mean separation.

3. Results

3.1. Chlorophyll Content and LAI

To evaluate the effects of N fertilization on the ratooning sorghum × sudangrass hybrid, chlorophyll content and LAI were measured throughout the ratooning period. There were significant DAS × N level interactions (p = 0.001, <0.001; 2018, 2019, respectively) for both years (

Table 3. Source of variation and levels of probability (p) for chlorophyll content and LAI during the ratooning period.

Source of Variation	Chlorophyll Content		LAI
	2018	2019	2018	2019
N level (N)	<0.001 a	<0.001	<0.001	<0.001
Ecotype (E)	0.091	0.334	0.085	0.173
Days after summer harvest (DAS)	<0.001	<0.001	<0.001	<0.001
N × E	0.321	0.054	0.012	0.397
N × DAS	0.001	<0.001	<0.001	<0.001
E × DAS	0.968	<0.001	0.929	<0.001
N × E × DAS	0.460	0.807	0.865	0.075

^a Significance evaluated at 0.05 probability level and bolded.

). There were linear and quadratic responses of the chlorophyll content to the N level in 2018 and 2019, respectively, except for DAS 75 in 2019 (

Table 4. Seasonal changes of chlorophyll content (mg m⁻²) for two years (2018–2019) by N level (0, 50, 100, and 150 kg N ha⁻¹).

Year	Treatments		Ratooning Period
			DAS a 15	DAS 30	DAS 45	DAS 60	DAS 75	DAS 90
2018	N level	0 N	-	438 a b B c	419 aB	395 bB	423 aC	-
		50 N	-	527 aA	449 bcA	439 cA	466 bB	-
		100 N	-	543 aA	450 cA	432 cA	492 bAB	-
		150 N	-	537 aA	462 cA	450 cA	503 bA	-
		Polynomial contrast d		L ***Q ***	L **	L *	L ***
2019	N level	0 N	418 cB	418 cC	468 aB	462 aB	440 bB	425 bcB
		50 N	495 aA	456 bcB	467 bB	475 abB	474 bA	445 cAB
		100 N	507 aA	497 abA	480 bAB	479 bB	485 bA	459 cA
		150 N	500 aA	497 aA	495 aA	507 aA	489 aA	460 bA
		Polynomial contrast	L ***Q ***	L ***Q *	L *	L **	L **	NS

^a DAS, Days after summer harvest. ^b Means followed by different lowercase letters within a row differ significantly (LSD, p < 0.05). ^c Means followed by different uppercase letters within a column differ significantly (LSD, p < 0.05). ^d Polynomial contrast was used to compare N level (L, linear; Q, quadratic; *** p < 0.001; ** p < 0.01; * p < 0.05; NS, statistically not significant).

). In 2018, the control (0 N) consistently had the lowest chlorophyll content throughout the season compared to plots receiving the other N treatments (Table 4). At DAS 75, 50 N had a relatively lower chlorophyll content than 100 N and 150 N (466 vs. 492 and 503 mg m⁻²). To identify early and late changes in chlorophyll content, measurements at DAS 15 and 90 were added in 2019. Overall trends in 2019 were similar to the previous year, with the control having the lowest chlorophyll content throughout the season. Additionally, 50 N had a relatively lower chlorophyll content than other N treatments, and no further change occurred when the N application level reached 100 N. For both years, plots receiving N had greater chlorophyll content in the early season.

The LAI of ratooning sorghum × sudangrass hybrids showed that there were significant DAS × N level interactions (p < 0.001) for both years (Table 3). Similar to chlorophyll content in 2018, the control continuously had the lowest LAI throughout the ratooning seasons (Figure 3). DAS 30 showed the highest LAI across the N levels, then it decreased throughout the season. At DAS 30, 50 N treated had relatively lower LAI than 100 N and 150 N (4.7 vs. 5.4 and 5.7, respectively). To identify the early change of LAI, a measurement at DAS 15 was added in 2019. In 2019, except for DAS 15, the control continuously had the lowest LAI throughout the ratooning season. Additionally, 100 N and 150 N had higher LAI than 50 N at DAS 30 to 45.

3.2. Root Morphology and Dry Matter

The total root length data showed that there were three-way interactions of N × E × DAS (p = 0.011) in 2018, while there were two-way interactions of N × E (p = 0.002) and E × DAS (p < 0.001) in 2019 (

Table 5. Source of variation and levels of probability (p) for total root length, total root surface area, total root volume, tips, and root DM during the ratooning period.

Source of Variation	Total Root Length		Total Root Surface Area		Total Root Volume		Tips		Root DM
	2018	2019	2018	2019	2018	2019	2018	2019	2019
N level (N)	<0.001 a	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
Ecotype (E)	0.381	0.003	0.773	0.005	0.893	0.006	0.627	0.002	0.005
Days after summer harvest (DAS)	<0.001	<0.001	0.001	<0.001	0.019	<0.001	0.002	0.003	<0.001
N × E	0.427	0.002	0.168	0.005	0.092	0.032	0.414	<0.001	0.077
N × DAS	0.401	0.770	0.443	0.837	0.348	0.560	0.450	0.585	0.038
E × DAS	0.213	<0.001	0.119	<0.001	0.066	<0.001	0.201	0.005	<0.001
N × E × DAS	0.011	0.267	0.009	0.504	0.019	0.790	0.043	0.193	0.872

^a Significance evaluated at 0.05 probability level and bolded.

). Thus, data were presented showing the N × E and E × DAS interactions to explore the plant root responses after defoliation over the 2-year period. There were linear and quadratic effects of N level in 2018, but there was a linear response of total root length to the N level in 2019 (

Table 6. Total root length, surface area, volume, and tips affected by N level (0, 50, 100, and 150 kg N ha⁻¹), ratooning periods (early, middle, and late seasons), and ecotype (late-flowering Greenstar, early-flowering Honeychew) for two growing seasons (2018–2019).

Year	Treatments		Total Root Length		Total Root Surface Area		Total Root Volume		Tips
			cm plant−1		cm2 plant−1		cm3 plant−1		plant−1
			Greenstar	Honeychew	Greenstar	Honeychew	Greenstar	Honeychew	Greenstar	Honeychew
2018	N level	0 N	377 a b B c	526 aB	76 aB	120 aB	1.73 aB	2.25 aC	1780 aB	2470 aB
		50 N	1063 aA	737 aAB	234 aA	165 aAB	4.22 aA	2.99 bBC	4630 aA	3380 aAB
		100 N	1079 aA	996 aA	232 aA	223 aA	4.05 aA	4.07 aAB	4490 aA	4490 aA
		150 N	1123 aA	982 aA	215 aA	226 aA	3.35 aA	4.21 aA	4690 aA	4380 aA
		Polynomial contrast d	L **Q *	L **	L **Q **	L **	L *Q ***	L **	L **Q *	L *
	Ratooning period	DAS a 76	992 aA	964 aA	201 aA	216 aA	3.41 aA	3.93 aA	4180 aA	4340 aA
		DAS 113	829 aB	657 aB	177 aA	151 aB	3.27 aA	2.83 aB	3620 aA	3020 aB
2019	N level	0 N	318 aC	226 bB	98 aC	65 bB	2.52 aC	1.52 bB	1240 aC	800 bB
		50 N	530 aB	246 bB	179 aB	80 bB	4.93 aB	2.12 bB	2110 aAB	880 bB
		100 N	512 aB	399 bA	186 aAB	136 bA	5.51 aAB	3.80 bA	1920 aB	1460 bA
		150 N	606 aA	422 bA	209 aA	141 bA	5.84 aA	3.80 bA	2370 aA	1570 bA
		Polynomial contrast	L **	L ***	L **	L ***	L **	L ***	L **	L ***
	Ratooning period	DAS 28	319 aC	198 bB	106 aC	67 bC	2.91 aC	1.86 bC	1270 aC	740 bB
		DAS 65	478 aB	377 bA	157 aB	113 bB	4.23 aB	2.70 bB	2020 aB	1510 bA
		DAS 100	677 aA	394 bA	241 aA	137 bA	6.95 aA	3.86 bA	2440 aA	1280 bA

^a DAS, Days after summer harvest. ^b Means followed by different lowercase letters within a row differ significantly (LSD, p < 0.05). ^c Means followed by different uppercase letters within a column differ significantly (LSD, p < 0.05). ^d Polynomial contrast was used to compare N level (L, linear; Q, quadratic, *** p < 0.001, ** p < 0.01, * p < 0.05).

). Table 6 shows that, in 2018, Greenstar (late-flowering ecotype) had a greater increase in total root length in response to N fertilizer compared with Honeychew (early-flowering ecotype). Specifically, in comparison with the control (0 N), Greenstar showed a 3-fold increase, while Honeychew receiving 50 N was not different than the control. Moreover, Greenstar maintained a high root length regardless of N fertilization levels, whereas the Honeychew root length increased for 100 N and 150 N compared with 50 N. A similar pattern was observed in 2019. Greenstar showed a significant increase in root length with 50 N or greater, while Honeychew demonstrated an increase in total root length only at 100 N or greater. These results indicate that the two ecotypes responded differently to N levels across the 2 years, with Greenstar generally maintaining higher root length development even at relatively low levels of N fertilization.

When the summer defoliation responses of total root length in the later season were compared in 2018, both Greenstar and Honeychew showed a decrease between DAS 76 and 113, but the degree of difference varied. For instance, Greenstar only had a 16% decrease, while Honeychew showed a 32% decrease in total root length. In 2019, three different dates were compared (DAS 28, 65, and 100) for the total root length after summer defoliation, including the earlier determination of the total root length. In 2019, for the early regrowth season (DAS 28), Greenstar had a 61% greater total root length than Honeychew, and this difference was maintained throughout the season.

Similar to the total root length, the root surface area was affected by a three-way interaction of N × E × DAS (p = 0.009) in 2018, while there were two-way interactions of N × E (p = 0.005) and E × DAS (p < 0.001) in 2019 (Table 5). There were linear and quadratic responses of the total root surface area to the N level in 2018, but there was a linear effect of the N level in 2019 (Table 6). Plots fertilized with N exhibited a greater root surface area than the control for both Greenstar and Honeychew in 2018 (Table 6); however, no differences were observed across N levels. Similarly, in 2019, Greenstar showed an increased root surface area with 50 N and above, while Honeychew demonstrated an increase in total root surface area at 100 N and above. The total root surface area in the later ratooning season was compared in 2018, and both Greenstar and Honeychew tended to decrease between DAS 76 and 113, but the magnitude of the difference differed (16.4% vs. 31.8%). In the early ratooning season of 2019 (DAS 28), Greenstar had a 58% greater total root surface area than Honeychew, a difference maintained throughout the season (21.1% and 41.8% for DAS 65 and 100, respectively).

The total root volume was affected by a three-way interaction of N × E × DAS (p = 0.019) in 2018, while there were two-way interactions of N × E (p = 0.032) and E × DAS (p < 0.001) in 2019 (Table 5). Similar to the total root length and surface area, there were linear and quadratic responses of the total root volume to the N level in 2018, but there was a linear effect of the N level in 2019 (Table 6). Table 6 demonstrates that Greenstar had a more pronounced increase in total root volume in response to N levels compared to Honeychew in 2018. Specifically, compared with the control, Greenstar at 50 N exhibited a two-fold increase, while Honeychew showed no difference. Furthermore, Greenstar maintained a high total root volume with all N fertilization levels, while Honeychew had a significant increase only at 100 N and 150 N in 2018. A similar trend was observed in 2019, where the 50 N treatment had a two-fold increase compared to the control for Greenstar, but there was no difference between these treatments for Honeychew. Greenstar had greater root volume compared to Honeychew across the N levels in 2019, while the difference showed only for 50 N in 2018. In 2018, for DAS 76 vs. 113, there was no difference for Greenstar, while Honeychew showed less total root volume at DAS 113. Conversely, when the comparison was conducted earlier in the regrowth period in 2019 (DAS 28), significant differences in root development were observed for both ecotypes, while Greenstar consistently showed greater total root volume than Honeychew.

Root tips data showed that there was a three-way interaction of N × E × DAS (p = 0.043) in 2018, while there were two-way interactions of N × E (p < 0.001) and E × DAS (p = 0.005) in 2019 (Table 5). There was a difference among N level treatments, but ecotype did not affect the response in 2018 (Table 6). There were linear and quadratic effects of the N level in 2018, while there was a linear response of root tips to the N level in 2019 (Table 6). Plots treated with N had greater numbers of root tips compared with the control, but there were no differences among the N-treated plots. In 2019, Greenstar had a greater number of tips compared to the control for N applications of 50 N and above, while it was 100 N and above for Honeychew. Additionally, similar to other root-associated parameters, Greenstar had greater tip number development than that of Honeychew. When later seasons were compared in 2018, both Greenstar and Honeychew exhibited a decrease in the root tip number between DAS 76 and 113, but to different extents. For instance, Greenstar had only a 13% decrease, while Honeychew exhibited a 30% decrease in tip numbers. At DAS 28, 65, and 100 in 2019, Greenstar had 71%, 33%, and 91% greater tip numbers than Honeychew, respectively, indicating that Greenstar develops root tips faster in the early ratooning season and maintains them better into the later season compared to Honeychew.

Root dry matter (DM) data showed that there were two-way interactions of N × DAS (p = 0.038) and E × DAS (p < 0.001) in 2019 (Table 5). The polynomial contrast result showed that there was a linear effect of the N level in 2019 (

Table 7. Root dry matter (DM) affected by additional N level (0, 50, 100, and 150 kg N ha⁻¹), ratooning periods (early, middle, and late seasons), and ecotype (late-flowering Greenstar, early-flowering Honeychew) in 2019.

Treatments		Ecotype
		g plant−1
		Greenstar	Honeychew	Avg.
N level	0 N	1.22	0.72	0.97 C b
	50 N	2.45	1.03	1.74 B
	100 N	2.78	1.82	2.30 A
	150 N	2.99	1.95	2.47 A
	Avg.	2.36 a c	1.38 b
	Polynomial contrast d	L *	L ***
Ratooning period	DAS a 28	0.92 aC	0.62 aC
	DAS 65	1.98 aB	1.35 bB
	DAS 100	4.19 aA	2.18 bA
		Ratooning period
		DAS 28	DAS 65	DAS 100
N level	0 N	0.29 bB	0.74 bC	1.89 aC
	50 N	0.67 cAB	1.55 bB	3.01 aB
	100 N	1.02 cA	2.22 bA	3.66 aA
	150 N	1.11 cA	2.14 bAB	4.16 aA

^a DAS, Days after summer harvest. ^b Means followed by different uppercase letters within a column differ significantly (LSD, p < 0.05). ^c Means followed by different lowercase letters within a row differ significantly (LSD, p < 0.05). ^d Polynomial contrast was used to compare N level (L, linear; *** p < 0.001; * p < 0.05).

). On average, the control had the lowest root DM (0.97 g plant⁻¹), while 100 N and 150 N had the highest root DM (2.30 and 2.47 g plant⁻¹, respectively, Table 7). At DAS 65 and 100, Greenstar had 47% and 92% greater root DM than Honeychew, and the difference was greatest in late season (DAS 100; 4.19 vs. 2.18 g plant⁻¹). The control (0 N) consistently recorded the lowest root DM throughout the regrowth period. Both 100 N and 150 N resulted in up to three-fold greater root DM compared with the control at DAS 28, and the trend was maintained throughout the regrowth season. At DAS 65 and 100, any level of N applied increased root DM over that of the control. The 150 N had demonstrated the highest root DM values (4.16 g plant⁻¹) at DAS 100. Among N treatments, averages of 100 N and 150 N were 30% and 41% greater than those of 50 N, respectively.

Comparing root morphological characteristics over time before summer defoliation, there was an E × DAP (days after planting) interaction for total root length, total root surface area, and root DM, while the main effects of DAP and ecotype were noted for total root volume and tips (

Table 8. Total root length, surface area, and volume in response to ecotype (late-flowering Greenstar, early-flowering Honeychew) before summer defoliation in 2019 and source of variation and levels of probability (p).

Treatments	Total Root Length		Total Root Surface Area		Total Root Volume			Tips			Root DM
	cm plant−1		cm2 plant−1		cm3 plant−1			plant−1			g plant−1
	Greenstar	Honeychew	Greenstar	Honeychew	Greenstar	Honeychew	Avg.	Greenstar	Honeychew	Avg.	Greenstar	Honeychew
DAP a 26	217 a b B c	229 aB	50 aB	53 aB	0.95	1.00	0.98 B	680	740	710 B	0.23 aC	0.27 aB
DAP 55	1209 aA	738 bA	243 aA	165 bA	3.93	2.99	3.46 A	4520	2580	3550 A	1.14 aB	1.08 aA
DAP 75	1326 aA	701 bA	302 aA	148 bA	4.74	2.51	3.63 A	4520	2780	3650 A	2.10 aA	1.10 bA
Avg.					3.21 a	2.17 b		3240 a	2030 b
Source of variation
Ecotype (E)	<0.001 d		<0.001		0.039			0.002			0.031
DAP	<0.001		<0.001		<0.001			<0.001			<0.001
E × DAP	0.020		0.036		0.137			0.064			0.041

^a DAP, Days after planting. ^b Means followed by different lowercase letters within a row differ significantly (LSD, p < 0.05). ^c Means followed by different uppercase letters within a column differ significantly (LSD, p < 0.05). ^d Significance evaluated at 0.05 probability level and bolded.

). Notable differences in the development of total root length were evident among the ecotypes. For instance, the total root length of Greenstar reached 1210 cm per plant at DAP 55, and then it plateaued (Table 8). Greenstar exhibited 64% and 89% greater root length development than Honeychew at DAP 55 and DAP 75, respectively. The trend for the total root surface area was also similar. Specifically, Greenstar exhibited 47% and 104% more total root surface area than Honeychew, and both cultivars displayed rapid root development until DAP 55, after which they plateaued. In terms of total root volume, the average for Greenstar was 3.21 cm³, while it was 2.17 cm³ for Honeychew. The tip number for Greenstar was 3240 per plant, while it was only 2030 per plant for Honeychew. Unlike other root parameters, there were no statistical differences between ecotypes at DAP 55 for root DM. At DAP 75, Greenstar had 90% greater root DM than Honeychew (2.10 vs. 1.10 g plant⁻¹, respectively).

4. Discussion

4.1. Chlorophyll Content and LAI Responses

The current study investigated seasonal changes in the above- and belowground development of two different ecotypes of sorghum × sudangrass hybrids, exploring the impact of different N application levels following summer defoliation. Aboveground development after summer defoliation in the current study indicated that N application linearly increased the leaf chlorophyll content of ratooning plants, and higher chlorophyll content was maintained until later in the growing season, which might have contributed to LAI and root development differences. Specifically, plants receiving 100 N (Avg. 7%) and 150 N (Avg. 9%) sustained higher chlorophyll content than other N applications throughout the ratooning season for both 2018 and 2019. Leaf nitrogen and photosynthesis are connected, as most of the N in leaves are associated with photosynthetic machinery [31]. In addition, it is known that leaf senescence is related to the photosynthetic capacity and chlorophyll content [32]. Previous research indicated that the supply of N increased the photosynthetic apparatus and chlorophyll content in the plant leaves, while leaf senescence was delayed in grain sorghum [33]. Similar results were reported in other C4 plants, such as sugarcane [34]. The stay-green trait delayed leaf senescence, but more importantly, it maintained leaf photosynthesis following the flowering period and increased grain yields in maize [35]. After DAS 45, there were no significant differences of LAI observed in terms of N input in 2018. This was due to measurements being taken after the plant canopy had already fully developed. Hence, the early growth period (DAS 15) in 2019 showed LAI development differences by N level began at the initiation of regrowth. A similar trend was reported in that 150 kg N ha⁻¹-applied plots had a greater LAI than 0 or 75 kg N ha⁻¹ applied plots during the heading stage of irrigated grain sorghum [36].

4.2. Root Morphology and Dry Matter Responses

Throughout this study, N fertilization after summer defoliation linearly increased total root length during ratooning. However, two ecotypes exhibited different degrees of belowground development after summer defoliation in both 2018 and 2019. Greenstar continued to increase its total root length as the regrowth season progressed, while Honeychew maintained its total root length after DAS 65. These results indicated that late-flowering ecotypes such as Greenstar tend to increase total root length until the later season and maintain this growth until the end of the season. In contrast, early-flowering ecotype Honeychew appeared to cease total root length development earlier in the season and exhibited a greater decrease later in the season. Thus, the two ecotypes exhibited distinct responses in total root length under varying N levels across the two-year regrowth period, with Greenstar displaying more consistent growth and maintenance compared to Honeychew. Thus, the late-flowering ecotype is a better option for multiple-harvest management as it has the superior ability to extend its root system, which is essential for sustaining water and nutrient supply throughout the ratooning season. It might also help to maintain a well-developed plant canopy during ratooning with a relatively low level of N fertilization.

To the best of our knowledge, there has been no previous research testing the effect of N fertilization on ratooning root development for sorghum. In the ratooning sugarcane study, there were genotype differences in root length density after 90 and 270 days of ratooning growth [37]. In the study, sugarcane genotypes under drought stress had high root length density in lower soil layers. However, under moderate drought conditions, plants tend to increase root systems in order to take up more water from inadequate soil moisture environments [38]. Previous research also suggested that plants are believed to be capable of modifying root growth to meet water demands during droughts [39]. In terms of the total root surface area and volume for both Greenstar and Honeychew, values continued to increase linearly as the regrowth season progressed, while Greenstar continued to have a greater root surface area and volume than Honeychew. Chung et al. (2020) proposed that an increase in root tip number represents the accumulation of lateral roots or an increase in roots emerging from the lateral root [40]. Thus, the greater tip number for Greenstar in the current study suggests that the late-flowering ecotype has the capability to increase new roots and branching of roots. Additionally, Greenstar is able to maintain existing roots for longer periods during ratooning. Nitrogen fertilization is required as it increases both above- and belowground plant development. However, requirements vary, as Greenstar showed higher growth under low levels of N fertilization (50 N).

Unexpectedly, although differences in aboveground growth were evident in the early stages of ratooning, the variation in growth due to treatment was not observed after the mid-growth phase. This implies that in the later stages of regeneration, there are no significant differences attributable to ecotype or N application. However, this inference is contradicted when the results of both aboveground and belowground growth are considered. The root development of the ratooning crop exhibited differences until the late growth phase, influenced by N application and ecotype. Therefore, N supplementation following the summer harvest not only directly enhances the growth of the ratooning crop but also facilitates nutrient uptake from the soil by promoting root development. The results obtained from this study demonstrate a similarity to the aboveground biomass yield of our research [6]. Especially even under unfavorable weather conditions in 2019, Greenstar exhibited greater development of root DM compared to Honeychew. When examining the cultivars before the summer harvest was executed, rapid development was observed up to DAP 55. Greenstar exhibited greater development in terms of root length, surface area, volume, and tips, indicating differing root development potential between the two cultivars. This might affect differences in plant regrowth after defoliation by ecotypes. Therefore, well-developed existing roots might be crucial for vigorous above- and belowground regrowth in SSG. Thus, under unfavorable weather conditions, the late-flowering ecotype Greenstar displayed a greater development and maintenance of the root system. As SSG is often grown in marginal fields where nutrients are scarce, choosing a late-flowering ecotype with a low N fertilization level is a practical option for growers.

5. Conclusions

We investigated changes in aboveground and belowground growth development among sorghum × sudangrass hybrid cultivars following post-cut fertilization. The results indicated no significant difference in chlorophyll content and LAI between the two ecotypes. While chlorophyll content varied with N application levels, LAI did not show significant variation. Different from aboveground parts, belowground development showed variation between ecotypes, particularly under the relatively favorable growing conditions of 2018, where differences were less pronounced compared to the unfavorable environment of 2019. In the latter years, differences in root growth among cultivars were evident from the early growth stages of ratooning, with Greenstar exhibiting rapid and vigorous root development. As overall root development did not increase at 100 kg N ha⁻¹ or above, reduced post-cut N application rates of 50–100 kg ha⁻¹ are recommended with a late-flowering ecotype, as excessive N input for a low-value crop does not align with economic goals and sustainability.