Nitrogen rate and harvest date effects on energy cane yield, quality parameters, nutrient uptake and biomass chemical composition

Abstract

Enhancing energy cane biomass production can increase feedstock supply for lignocellulosic-based ethanol production. This study was conducted at the Louisiana State University Agricultural Center (LSU AgCenter) Sugar Research Station in St. Gabriel, LA, from 2013 to 2015 to evaluate yield, fiber composition, and nutrient removal rate of energy cane harvested at different dates. The treatments included two energy cane cultivars (‘Ho 02-113’ and ‘Ho 72-114’) and four N rates (0, 56, 112, and 224 kg N ha^–1), arranged in split plot in a randomized complete block design with four replications. Fifteen plants were collected once a month beginning 2 mo prior to the harvest date and then separated into leaves and stalks and analyzed for fiber and nutrient composition. Harvest date, cultivar, and N rate significantly affected energy cane yield (stalk and leaves), quality components, lignocellulosic composition, and nutrient removal rate (p < .05). On average, harvesting energy cane 1 and 2 mo earlier than the scheduled harvest date reduced stalk yield by 4% and 20% and leaf yield by 3% and 15%. Minimal changes on stalk fiber content and lignocellulosic composition in response to the treatments were documented but no specific and consistent trend was established. With leaf included, energy cane biomass production was increased by 32%; however, this can result in significant removal of plant essential nutrients, particularly N (105 kg ha^–1), P (16 kg ha^–1), and K (116 kg ha^–1). The long-term impact of complete removal of energy cane biomass from the field could result in decline in soil fertility and productivity.

Abbreviations

ICP-OES

inductively coupled plasma optical emission spectroscopy

LSU

Louisiana State University

TRS

theoretical recoverable sugars

1 INTRODUCTION

Sugarcane (Saccharum sp. Hybrid) has been cultivated in the southern parts of Louisiana since 1795 (Bischoff et al., 2008). Sugarcane is bred for large stalk diameter, ratooning ability, and disease resistance (Gravois, 2001). With the “oil shocks” of 1973 and 1979 along with price increases that led to economic disruption at international, national, and local levels greater emphasis was placed on the development of energy cane (Saccharum spp.) cultivar for bioenergy uses (Baldwin et al., 2012). The U.S. Department of Energy and U.S. Department of Agriculture perceived the potential role of biomass as a feedstock for bioenergy industry and that the U.S. is capable to produce nearly one billion dry tons of biomass annually (Perlack et al., 2005). The passing of Energy Independence & Security Act of 2007 (EISA, 2007) set the stage for a resurgence of all biomass crops, including energy cane. In this legislation, the Renewable Fuels Standard set forth goals for domestic renewable fuel production: 34 billion liters of renewable fuel production in 2008, rising to 136 billion liters by 2022, with 79 billion liters required to be produced from cellulosic ethanol and other advanced biofuels (Congressional Research Service, 2007). In 2020, the total biomass crops produced in the United States only provided 139.8 million L, which has an equivalent energy of nearly 5 quadrillion kJ, far from the set goal in 2007 (EIA, 2021). Cellulosic ethanol refers to ethanol derived from cellulose and hemicellulose from biomass. Other advanced biofuels would include ethanol derived from waste material, biomass-based diesel, biogas, and butanol and other alcohols produced through conversion of organic matter from renewable biomass (Salassi et al., 2014).

In 2007, Louisiana State University (LSU) released ‘L79-1002’, a cane used specifically as a biomass feedstock (Bischoff et al., 2008). However, as prices of the fuel decreased, the interest of using biomass feedstock crop faded away. Because energy cane was bred for high biomass and fiber content, there was a proportional reduction in sugar concentration, making it less attractive to the sugar industry.

There is an increasing uncertainty of petroleum supplies due to rising demand, decline in known reserves, and concerns over climate change and greenhouse gas emissions associated with fossil fuels usage; thus, various government programs were initiated to promote biofuels as a sustainable option to overcome these issues (Saini et al., 2014). According to Fulton et al. (2004), bioethanol can reduce greenhouse gas emissions by approximately 30–85% compared with gasoline, depending on the feedstock used. Worldwide increasing interest in the production of bioethanol is exemplified by the production of 85 billion L of bioethanol in 2011 (Avci et al., 2013; Singh & Bishnoi, 2012).

With the renewed interest in nonconventional fuel, energy cane has gained attention due to its low production cost requirement and high biomass yield potential (Kim & Dale, 2005). Energy cane has been identified as a potential and ideal feedstock source for biofuel production based on the significant energy gain it can provide considering all input–output equation while not posing any competition against food crops for prime land (Coombs, 1984; Gonzales-Hernandez et al., 2009; Hill et al., 2006; Johnson et al., 2007; Macedo, 1998; Schmer et al., 2008; Yuan et al., 2008).

Louisiana is suitable for production of energy cane for several reasons. The existing infrastructure and equipment for Louisiana's sugarcane industry can be directly used for harvesting, transport, and processing of energy cane (Baldwin et al., 2012). Also, Louisiana has a favorable climate for production of energy cane, that is, warm temperature, high annual precipitation (160 cm), and a growing season that spans from 230 to 290 d allowing for more production of biomass. Kim and Day (2011) noted all these and the fertile alluvial soils in the region led to the oldest and largest commercial sugarcane industry in the United States.

Core Ideas

• Whole-plant harvesting of energy cane increased biomass yield by 32% or 16 Mg ha⁻¹.
• Early harvesting of energy cane reduced biomass yield, sucrose content, and recovered sugar.
• Increasing N rate increased stalk and leaf yield but reduced the recovered sugar in energy cane.
• N, P, and K removal rates were 2 to 3 times higher in whole-plant than in stalk-only harvesting.

Studies have been devoted to understanding nitrogen (N) utilization by crops than any other nutrient. It is the most limiting nutrient in nonlegume cropping systems and the least predictable due to its very dynamic nature. When N fertilizer is applied in the soil, it will undergo several processes and can easily be lost in the soil system. Application of N fertilizer at the optimum rate is an integral part of crop production to maximize economical return as well as to minimize environmental risks (Kanke et al., 2016; Lofton & Tubana, 2015; Raun et al., 2011; Tubana et al., 2011).

It should be noted that energy cane like sugarcane is a semiperennial that is vegetatively propagated that can be harvested annually up to 5 yr without replanting; the first harvested crop is termed plant cane and ratoon cane for each successive harvest. In Louisiana, N fertilizer is applied only once in every cropping season and usually done in early April until the beginning of May. The current LSU AgCenter N rate recommendation was established from multiple site-year response trials and then refined according to soil type and crop age (Legendre et al., 2000). Specifically, N rates between 67 to 110 kg N ha⁻¹ are recommended for plant cane and between 88 to 132 kg N ha⁻¹ for ratoon crop, with rates at the low and high end of the range recommended for sugarcane planted on light and heavy textured-soil, respectively. A study conducted by Wiedenfeld (1995) showed that sugarcane quality and yield are easily affected by N management; excess amount of N application decreased sugar yield, juice purity, as well as recoverable sucrose.

Although energy cane is considered sugarcane, information on the production of energy cane is limited. Research to date has not identified the ideal rates of fertilizer application for energy cane. For cellulosic biofuel production, the addition of extra fertilizer may be of a benefit to get more biomass. However, the effect of the added fertilizer on the composition of the stalk is not known. Mislevy et al. (1995) found only a slight benefit in biomass yield when N rate was increased from 168 to 336 kg ha⁻¹.

Sugarcane harvesting takes place during a 100-d period toward the end of the year prior to the average killing frost date. This period typically falls from October to December, with the older ratoon crops being harvested first. This is a period on which the sugar mills also operate. Supplying energy cane outside this period has agronomic and economic advantage for the biofuel industry. Although whole-plant harvesting may increase biomass yield, the complete removal of residue from the field impaired nutrient recycling. To answer these queries, this study was conducted to evaluate the effect of N rate and harvesting date on quality parameters, yield (stalk and leaves), nutrient uptake, and biomass chemical composition of two energy cane cultivars.

2 MATERIALS AND METHODS

2.1 Site location, experimental design, and layout

This study was established at the LSU AgCenter Sugar Research Station in St. Gabriel, Louisiana (30°15′47″ N′ 91°05′54″ W) on a Commerce silt loam soil (fine-silty, mixed, superactive, nonacid, thermic Fluvaquentic Endoaquepts). Before planting, composite soil samples (16 cores per quadrant) at 0–15 cm depth were collected for initial soil chemical analysis. The samples were dried, ground, and extracted with Mehlich-3 solution (Mehlich, 1984). The concentrations of selected nutrients were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES). The soil had an initial pH value of 5.5 with phosphorus (P), potassium (K), magnesium (Mg), sulfur (S), and copper (Cu) content of 34, 170, 458, 10.8, and 3.5 mg kg⁻¹, respectively.

Each plot consisted of three 1.83-m wide × 9-m long raised beds or rows. The length of the alley between plots was 3 m. The treatments included two energy cane cultivars (‘Ho 02-113’ and ‘Ho 72-114’) and four N application rates (0, 56, 112, and 224 kg ha⁻¹). Variety Ho 02-113 is a cane cultivar with high fiber and low sucrose content that can be used as a feedstock for the production of biofuels (Hale et al., 2013). The female parent of Ho 02-113 is ‘SES 234’ (Saccharum spontaneum), and the male parent is ‘LCP 85-384’, a commercial sugarcane cultivar. It has an extremely high population of small diameter stalks. The canopy is very erect, and the cultivar has excellent vigor and stubbling ability. Cultivar Ho 72-114 is also a high fiber and low sugar content cane cultivar and has been tested for high biomass production for biofuel. The female parent of Ho 72-114 is ‘CP52-068’ and the male parent is ‘US66-65-11’. All treatments were replicated four times and laid-out using split plot in randomized complete block design with cultivar as the main plots and N application rate as the subplots.

2.2 Planting, fertilization, harvesting, and plant analysis

Planting was completed on 14 Sept. 2012 using billets as planting material. Billets were produced from cutting whole stalks into pieces using a combine harvester. Each billet has an average length of 60 cm with two to three mature internodes (Hoy, 2001). Beds were opened and placed with five to six running billets, closed, and then packed with approximately 6 cm of soils with a custom roller packer. In April, urea-ammonium nitrate (UAN; 32-0-0) solution at rates of 0, 56, 112, and 224 kg N ha⁻¹ was knifed-in on both shoulders of each bed at 15 cm depth. Potassium was broadcast-applied at 60–80 kg ha⁻¹ as muriate of potash (0-0-60). Soil P level was tested high; therefore, no application of P fertilizer was made.

The harvesting for the three dates (2- and 1-mo earlier, and at scheduled harvest) was done by randomly cutting 15 plants from the middle row of each plot. Table 1 provides the harvest schedule implemented from 2013 to 2015. The whole plants were partitioned into stalks and leaves and weighed separately. The stalks were shredded and analyzed for sugar quality parameters using SpectraCane Near Infrared System (Bruker Corporation, Billerica, Massachusetts) to determine theoretical recoverable sugars (TRS), sucrose content, total soluble solids (Brix), and fiber content. Following this analysis, grab samples of shredded stalk were collected for each plot. Leaves and shredded stalks samples were dried at 60 °C for 48 h and ground to pass a 1-mm sieve and analyzed for total N using CN 91 analyzer (Vario el cube; Elementar), and elemental (P, K, S, calcium [Ca], and Mg) composition by nitric acid-hydrogen peroxide digestion procedure followed by ICP-OES. The nutrient uptake was computed as nutrient concentration × stalk dry weight. Lignocellulosic composition was determined using ANKOM²⁰⁰⁰ Filter Bag method. Ground samples weighing 0.5 g was placed in ANKOM F57 filter bags and heat sealed and underwent a series of extractions for neutral detergent fiber, acid detergent fiber, and acid detergent lignin. The residue after neutral detergent fiber extraction is predominantly composed of hemicellulose, cellulose, and lignin, whereas the acid detergent fiber extraction is composed of cellulose and lignin and acid detergent lignin residue represents lignin. The different lignocellulosic composition was computed as:

$$\begin{equation*}{\rm{\% Hemicellulose = \% NDF}} - {\rm{\% ADF}}\end{equation*}$$

$$\begin{equation*}{\rm{\% Cellulose = \% ADF}} - {\rm{\% ADL}}\end{equation*}$$

and

$$\begin{equation*}{\rm{\% Lignin = \% ADL}}\end{equation*}$$

where NDF is neural deterget fiber, ADF is acid detergent fiber, and ADL is acid detergent lignin. After collecting the 15 whole plants at the scheduled harvest (i.e., December, November, and October for 2013, 2014, and 2015, respectively), all three rows of each plot were harvested using a Case IH 8800 Series single row chopper (Case IH Agriculture) and loaded to wagon with load cell to determine the plot weight.

TABLE 1. Harvest schedule of the energy cane in St. Gabriel, LA, from 2013 to 2015 cropping season

Harvest dates	2013 plant cane	2014 first ratoon	2015 second ratoon
2 mo earlier	Oct.	Sept.	Aug.
1 mo earlier	Nov.	Oct.	Sept.
Scheduled harvest	Dec.	Nov.	Oct.

2.3 Data analysis

Statistical analysis was done using SAS 9.4 software (SAS Institute, 2012). The year that was synonymous to crop age was initially included as a factor. The results came back with a significant four-way interaction effect for almost all the variables measured, thus the analysis of variance (ANOVA) was performed for each year or crop age. The ANOVA was performed using PROC MIXED to evaluate the effects of cultivar, N rate, harvest dates, and their interactions on stalk yield, nutrient concentration and uptake, and fiber composition. Mean separation was done by Tukey–Kramer post-hoc test for any significant effect at p < .05.

3 RESULTS

3.1 Climatic condition

The monthly average precipitation and temperature from 2013 to 2015, and the average from 1991 to 2020 (30 yr) are presented in Figure 1. The average monthly precipitations were 145, 120, and 149 mm for 2013, 2014, and 2015, respectively, with 2013 and 2015 receiving more than 1,700 mm total precipitation and only 1,430 mm for 2014. The highest monthly precipitation was received in May 2014 and November 2015. Many months across these years recorded cumulative rainfall above the 30-yr average; however, there was no extreme water-related stress observed nor interference on field operations except with the harvesting season in 2015 being relatively wet. The average monthly temperature was very similar across these years with the highest temperature (25 °C) recorded in June, July, and August. The temperature in April and May in 2013 and 2014 was lower compared with the 30-yr average, other than this the 2013 to 2015 monthly average temperature pattern did not deviate substantially from the 30-yr average.

FIGURE 1

Monthly cumulative precipitation (a) and average temperature (b) in St. Gabriel, LA for 2013, 2014, 2014, and from 1991 to 2020 (30 yr)

3.2 Stalk and leaf yield and stalk fiber content

The year or crop age was initially treated as a factor when the ANOVA was performed. The results returned with a significant four-way interaction effect for almost all the variables. Thus, the ANOVA was done for each year with harvest date, cultivar, and N rate as factors. Nevertheless, the yearly means and standard error values of these variables were presented in Table 2. The average stalk yield and fiber content tended to decline with year or crop age, a pattern not observed for leaf yield. The average stalk and leaf yield for 2013, 2014, and 2015 was 28.5, 19.6, and 19.7 Mg ha⁻¹, and 13.2, 14.0, and 21.4 Mg ha⁻¹, respectively. The stalk fiber content showed a steady decline with year at the rate of 3.2%. There was no significant three-way interaction effect recorded on these variables across years. For some parameters, a significant two-way interaction effect was detected, which mostly occurred in 2015. For example, cultivar effect was not consistent across harvest date for leaf yield and stalk fiber content, which was not the case for stalk yield. Based on general pattern, the early-harvested and unfertilized energy cane had a significant reduction in stalk yield whereas the Ho 02-113 cultivar was more productive than Ho 72-114. In 2014 and 2015 (both ratoon crops), stalk yield was reduced by 5.1 and 10.7 Mg ha⁻¹, respectively, when harvested 2 mo earlier. Similar response of leaf yield was observed in 2014 and 2015 with reduction at 5 and 6.9 Mg ha⁻¹, respectively. The results also indicated that the optimal production of stalk and leaf yield was attained at 112 N kg ha⁻¹ application rate. Although stalk fiber content was reduced at higher N rate, the magnitude of reduction varied with cultivar and harvest date.

TABLE 2. Means and analysis of variance for the effects of harvest date, cultivar, nitrogen rate, and their interactions on stalk yield, leaf yield, and stalk fiber content of energy cane from 2013 to 2015, St. Gabriel, LA

	Stalk yield			Leaf yield			Stalk fiber content
	2013	2014	2015	2013	2014	2015	2013	2014	2015
Effect	(28.5 ± 0.74)	(19.6 ± 0.59)	(19.7 ± 0.68)	(13.2 ± 0.49)	(14.0 ± 0.50)	(21.4 ± 0.83)	(25.0 ± 0.83)	(21.1 ± 0.17)	(18.6 ± 0.22)
	Mg ha^−1						%
Harvest date (H)
2 mo earlier	28.4	16.7 B	13.6 C	15.4 A	11.8 C	17.1	19.5 C	20.3	15.9
1 mo earlier	29.4	20.4 A	21.1 B	13.2 B	13.6 B	23.3	20.9 B	22.5	19.6
Harvest	27.6	21.8 A	24.3 A	11.0 C	16.8 A	24.0	34.5 A	20.6	20.2
p value	NS^†	<.001	<.001	<.001	<.001	<.001^†	<.001	<.001^†	<.001^†
Cultivar (C)
Ho 02-113	32.4	21.0 A	20.0	14.1 A	12.7 B	17.2	26.3 A	21.4	18.6
Ho 72-114	24.5	18.3 B	19.4	12.3 B	15.4 A	25.8	23.6 B	20.9	18.5
p value	<.001^a	<.01	NS	<.05	<.001	<.001^a	<.01	<.05^†	NS
Nitrogen rate (N)
0 kg ha^−1	25.2	16.4 C	14.9 C	14.6 A	14.5	19.0	24.9	22.1	18.3
56 kg ha^−1	26.0	19.0 B	18.6 B	10.7 B	13.1	23.0	26.0	21.1	18.7
112 kg ha^−1	30.8	21.0 A	22.4 A	15.3 A	15.2	22.7	25.4	20.5	18.9
224 kg ha^−1	32.0	22.0 A	22.8 A	12.3 B	13.4	21.1	23.6	20.8	18.4
p value	<.001^†	<.001	<.001	<.001	NS	NS	NS	<.001^†	NS
H × C	NS	NS	NS	NS	NS	<.05	NS	NS	<.05
H × N	NS	NS	NS	NS	NS	NS	NS	<.001	NS
C × N	.05	NS	NS	NS	NS	<.01	NS	<.01	<.001
H × C × N	NS	NS	NS	NS	NS	NS	NS	NS	NS

• Note. Same letter within column indicates no significant differences between treatments for each factor based on Tukey's analysis. Values in parentheses are the year (crop age) average and standard error of the corresponding variable.
• ^a Mean separation was not conducted on the given main effect due to significant two- or three-way interactions.
• ^† NS, not significant at p < .05.

3.3 Energy cane quality parameters

The energy cane TRS, Brix, and sucrose content were significantly affected by harvest date, cultivar, and N rate with significant cultivar × N rate interaction effect recorded in 2013 and 2015 (Table 3). Among the variables, only TRS incurred large disparity between years (or crop age) such that the average in 2013 was 116 kg Mg⁻¹ compared with 82 and 81 kg Mg⁻¹ in 2014 and 2015, respectively. The lowest values of TRS, Brix, and sucrose content were observed in energy cane harvested 2 mo earlier. As much as 86 kg Mg⁻¹ reduction in TRS was recorded for cane harvested 2 mo earlier than the scheduled harvest. The Ho 02-113 cane cultivar had higher TRS than Ho 72-114; a difference of 24 kg Mg⁻¹ in 2015 was recorded, whereas the lower TRS in 2013 and 2014 by Ho 72-144 was implicated by N rate effect. The N rate had a significant effect on TRS where the impact was more evident in 2014 and 2015 (both ratoon crops). The unfertilized (0 N) energy cane consistently obtained higher TRS compared with those treated with N (56, 112, and 224 kg N ha⁻¹). Similar response was observed for the Brix and sucrose content. In 2015, the application rate at 224 kg ha⁻¹, the highest N rate, reduced Brix by 1.3% and sucrose content by 2.55% in reference to the 0 N treatment (p < .001).

TABLE 3. Means and analysis of variance for the effects of harvest date, cultivar, nitrogen rate, and their interactions on theoretical recoverable sugar, Brix, and sucrose content of energy cane from 2013 to 2015, St. Gabriel, LA

	TRS			Brix			Sucrose
	2013	2014	2015	2013	2014	2015	2013	2014	2015
Effect	(116 ± 2.5)	(82 ± 3.7)	(81 ± 4.4)	(13.3 ± 0.22)	(13.8 ± 0.16)	(12.5 ± 0.21)	(9.04 ± 0.17)	(7.85 ± 0.21)	(7.33 ± 0.27)
	kg Mg^−1			%
Harvest date (H)
2 mo earlier	107 B	49 C	40 C	13.7 B	13.9	12.5 B	8.95 B	5.92	4.78 C
1 mo earlier	124 A	75 B	73 B	14.9 A	12.5	10.2 C	10.12 A	7.54	7.03 B
Harvest		122 C	126 A	11.4 C	15.1	14.8 A	8.07 B	10.08	10.17 A
p value	<.001	<.001	<.001	<.001	<.001^a	<.001	<.001	<.001^a	<.001
Cultivar (C)
Ho 02-113	127	86	92 A	14.6	14.0 A	13.0 A	9.90	8.08 A	8.06 A
Ho 72-114	105	78	68 B	12.1	13.7 B	12.0 B	8.20	7.62 B	6.60 B
p value	<.001^a	<.001^a	<.001	<.001^a	NS^†	<.001	<.001^a	<.01	<.001
Nitrogen rate (N)
0 kg ha^−1	126	106	103 A	13.4	15.0	13.1 A	9.41	9.28	8.59 A
56 kg ha^−1	123	71	90 B	13.5	14.4	12.8 A	9.42	7.49	7.93 B
112 kg ha^−1	119	84	69 C	13.4	13.3	12.2 B	9.22	7.75	6.75 C
224 kg ha^−1	95	87	57 C	13.0	12.7	11.8 C	8.16	6.88	6.04 D
p value	<.001^a	<.001^a	<.001	NS	<.001^a	<.001	<.001^a	<.001^a	<.001
H × C	NS	NS	NS	NS	NS	NS	NS	NS	NS
H × N	NS	NS	NS	NS	<.001	NS	NS	<.05	NS
C × N	<.05	<.05	NS	<.05	NS	NS	<.01	NS	NS
H × C × N	NS	NS	NS	NS	NS	NS	NS	NS	NS

• Note. Same letter within column indicates no significant differences between treatments for each factor based on Tukey's analysis at p < .05. Values in parentheses are the year (crop age) average and standard error of the corresponding variable. TRS, theoretical recoverable sugar.
• ^a Mean separation was not conducted on the given main effect due to significant two- or three-way interactions.
• ^† NS, not significant at p < .05.

3.4 Lignocellulosic composition

The lignocellulosic components of energy cane stalk (unpressed) and leaves were not affected by cultivar and N rate but were influenced by harvest date (Figure 2). Across the years, the total lignocellulosic content of the leaves and stalk was similar with ranges of 64–87% and 68–87%, respectively. The impact of harvest date on the content and distribution of lignocellulosic component was not consistent across years. Generally, the total lignocellulosic content of leaves was increased by early harvesting in 2014 and 2015, whereas for stalk, this pattern was observed in 2013 and 2015. The lignin content of both leaves and stalk was relatively stable across years and harvest dates recording an average of 20% with a range of 17–24%. This was not the case for cellulose and hemicellulose content. Cellulose and hemicellulose in both leaves and stalk tended to increase with year or crop age. With the inclusion of harvest date, the response became more complex. Overall, the observation was that the cellulose in leaves was higher in energy cane that was harvested 2 mo and 1 mo earlier than the scheduled harvest. This was the same pattern obtained for stalk except in 2013. Here, the stalk recorded higher amount of cellulose (34%) compared with the earlier harvest date. With respect to the hemicellulose content of the leaves and stalk, the scheduled harvest had a clear positive impact but only in 2015.

FIGURE 2

Harvest date effects on lignocellulosic composition (%) of energy cane stalk and leaves in 2013 (plant cane), 2014 (1st ratoon), and 2015 (2nd ratoon) in St. Gabriel, LA. For each lignocellulosic component (hemicellulose, cellulose, and lignin) of each plant part (leaves and stalks), values with the same uppercase letter are not significantly different at p < .05

3.5 Leaves and stalk nutrient removal rate

The amount of N, P, K, Ca, Mg, and S removed by energy cane stalk and leaves from 2013 to 2015 are presented in Tables 4 to 6. Significant effects of harvest date, cultivar, and N rate were observed on leaf nutrient removal rates across years with a few significant two-way interaction effects. More significant two-way interaction effects were recorded for stalk nutrient removal rates across years. Considering the general patterns, K was taken up in the largest amount by both leaves and stalks followed by N in 2013 and 2014; however, a shift between the removal rate of these nutrients occurred in 2015. The N and K removal rate by stalk was almost twice as much as the leaves and this was more evident in 2013 (plant cane). The general pattern with harvest date differed across years. In 2013 (plant cane), the pattern indicated that earlier harvesting resulted in higher nutrient removal rate by energy cane than the scheduled harvesting. In 2014 and 2015, which were both ratoon crops, the energy cane that was harvested at 1 and 2 mo earlier generally removed lesser amounts of nutrients than those harvested at scheduled date. At scheduled harvest, leaves' nutrient removal rates were 65–130 kg N, 48–192 kg K, 6–30 kg P, and 9–17 kg S per ha. Substantial amounts of Ca (10-67 kg ha⁻¹) and Mg (17-36 kg ha⁻¹) were removed by leaves, and these were even higher than amount of P removed. The amount of N and bases (K, Ca, Mg) removed by leaves was generally higher than what was removed by stalk. The nutrient removal rate was computed as the product of nutrient concentration and yield. Yield was more responsive to the treatments and had a higher impact than the nutrient concentration (data not shown) on nutrient removal rate. For example, the N rate consistently impacted most of the nutrients (not only N) removed by leaves and stalk across years, mainly due to the stalk and leaf yield increased with increasing N rate (Table 2). Besides, the level of nutrients in the soil (except for N) was ensured as nonlimiting, and, when recommended, a uniform application of fertilizer was made to attain sufficient level in the soil.

TABLE 4. Means and analysis of variance for the effects of harvest date, cultivar, nitrogen rate, and their interactions on leaves and stalks macronutrients removal rate of energy cane in 2013, St. Gabriel, LA

	Leaves						Stalks
Effect	N	P	K	Ca	Mg	S	N	P	K	Ca	Mg	S
	kg ha^−1
Harvest date (H)
2 mo earlier	108 A	11.6 A	117 A	16 A	56 A	17 A	54	38	257 A	27	15	32
1 mo earlier	85 B	8.4 B	81 B	13 B	45 B	13 B	63	40	126 B	36	21	22
Harvest	65 C	6.1 C	48 C	10 C	36 B	9 C	53	26	160 B	26	10	21
p value	<.001	<.05	<.001	<.001	<.001	<.001	NS	<.001^a	<0.001^a	NS	<.001^a	<.0001^a
Cultivar (C)
Ho 02-113	98 A	9.8 A	90 A	14 A	50 A	13	46	33	170	21	12	23
Ho 72-114	73 B	7.6 B	74 B	12 B	41 B	13	67	37	192	38	18	27
p value	<.001	<.001	<.01	<.01	<.01	NS	<.001^a	NS^†	NS	<.05^a	<.01^a	<.05^a
Nitrogen rate (N)
0 kg ha^−1	87 AB	9.5 A	84 A	12 B	42 B	14	38	31	162	28	12 B	23
56 kg ha^−1	84 BC	8.7 A	85 A	12 B	46 B	13	48	35	167	24	13 B	24
112 kg ha^−1	73 C	7.2 B	66 B	11 B	41 B	11	58	38	203	39	20 A	29
224 kg ha^−1	99 A	9.6 A	93 A	16 A	53 A	13	83	35	192	27	15 B	25
p value	<.05	<.01	<.01	<.05	<.001	NS	<.001^a	NS	NS	NS	<.01	NS
H × C	NS	NS	NS	NS	NS	NS	<.01	<.001	<.001	<.001	<.001	<.001
H × N	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS
C × N	NS	NS	NS	NS	NS	NS	<.05	NS	NS	NS	NS	NS
H × C × N	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS

• Note. Values with same letter within a column for each factor indicate no significant differences based on the Tukey‘s post-hoc analysis at p < .05.
• ^a Mean separation was not conducted on the given main effect due to significant two- or three-way interactions.
• ^† NS, not significant at p < .05.

TABLE 5. Means and analysis of variance for the effects of harvest date, cultivar, nitrogen rate, and their interactions on leaves and stalks macronutrients removal rate of energy cane in 2014, St. Gabriel, LA

	Leaves						Stalks
Effect	N	P	K	Ca	Mg	S	N	P	K	Ca	Mg	S
	kg ha^−1
Harvest date (H)
2 mo earlier	106 B	10	113	34 C	12 B	9 B	51 B	20 B	117 B	12	7 B	8 C
1 mo earlier	106 B	11	115	45 B	14 B	11 B	60 B	26 A	137 AB	15	8 AB	11 B
Harvest	121 A	11	108	56 A	17 A	15 A	70 A	29 A	148 A	16	10 A	13 A
p value	<.05	NS^†	NS	<.001	<.001	<.001	<.001	<.001	<.05	<.05^a	<.001	<.001
Cultivar (C)
Ho 02-113	116 A	11 A	121 A	48	14	12	64	25	143	20	10	11 A
Ho 72-114	105 B	10 B	103 B	42	14	11	58	25	126	9	7	10 B
p value	<.05	<.05	<.01	NS	NS	NS	NS	NS	<.05	<.001^a	<.001^a	<.01
Nitrogen rate (N)
0 kg ha^−1	93 C	10	98 B	35 B	11 C	11	44 C	25	118	11	6	13 A
56 kg ha^−1	98 C	10	103 B	37 B	11 C	10	50 C	25	131	11	7	11 B
112 kg ha^−1	119 B	11	119 A	51 A	16 B	12	62 B	26	140	15	9	10 C
224 kg ha^−1	133 A	11	127 A	57 A	18 A	13	86 A	25	149	20	11	8 D
p value	<.001	NS	<.01	<.001	<.001	NS	<.001	NS	<.05	<.001^a	<.001^a	<.001
H × C	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS
H × N	NS	NS	NS	NS	NS	NS	NS	NS	NS	<.05	NS	NS
C × N	NS	NS	NS	NS	NS	NS	NS	NS	NS	<.05	<.05	NS
H × C × N	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS

• Note. Values with same letter within a column for each factor indicate no significant differences based on the Tukey‘s post-hoc analysis at p < .05.
• ^a Mean separation was not conducted on the given main effect due to significant two- or three-way interactions.
• ^† NS, not significant at p < .05.

TABLE 6. Means and analysis of variance for the effects of harvest date, cultivar, nitrogen rate, and their interactions on leaves and stalks macronutrients removal rate of energy cane in 2015, St. Gabriel, LA

	Leaves						Stalks
Effect	N	P	K	Ca	Mg	S	N	P	K	Ca	Mg	S
	kg ha^−1
Harvest date (H)
2 mo earlier	140 AB	21 B	158 B	39 B	21 B	13	46 B	23 B	116 B	12 B	10 B	7 B
1 mo earlier	158 A	32 A	199 A	69 A	28 A	18	73 A	34 A	155 A	21 A	16 A	12 A
Harvest	130 B	30 A	192 A	67 A	26 A	17	73 A	34 A	151 A	21 A	15 A	12 A
p value	<.01^a	<.001	<.05	<.001	<.01	<.001^a	<.001	<.001	<.001	<.05	<.05^a	<.001
Cultivar (C)
Ho 02-113	143	29	188	57	24	17	65	32	146	22	15	11 A
Ho 72-114	143	27	178	60	27	15	63	29	136	14	13	9 B
p value	NS^†	NS	NS	NS	NS	<.05^a	NS	<.05	NS	NS	<.01^a	<.05
Nitrogen rate (N)
0 kg ha^−1	85	21	128	32 C	13	14 B	40 D	27	113	11	8	13 A
56 kg ha^−1	122	25	171	48 B	19	15 B	48 C	29	137	15	11	10 B
112 kg ha^−1	172	32	212	75 A	33	17 A	71 B	33	151	23	17	9 C
224 kg ha^−1	190	33	221	79 A	36	18 A	98 A	33	162	24	19	8 C
p value	<.001^a	<.001^a	<.001^a	<.001	<.001^a	<.001	<.001	<.001	<.001	NS	<.001^a	<.001
H × C	NS	NS	NS	NS	NS	<.001	NS	NS	NS	NS	NS	NS
H × N	<.001	NS	NS	NS	NS	NS	NS	NS	NS	NS	.05	NS
C × N	<.001	<.01	<.001	NS	<.01	NS	NS	<.05	NS	NS	.05	NS
H × C × N	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS

• Note. Values with same letter within a column for each factor indicate no significant differences based on the Tukey's post-hoc analysis at p < .05.
• ^a Mean separation was not conducted on the given main effect due to significant 2- or 3-way interactions.
• ^† NS, not significant at p < .05.

4 DISCUSSION

There was no water- nor heat-related stress observed on energy cane during the three cropping years. The highest rainfall event in 2013 and 2014 occurred during the early growth stage of energy cane, whereas in 2015, it was close to the maturity stage. According to Richard and Anderson (2014), dry conditions and sunlight promote tillering at the beginning of each growing season, whereas frequent rainfall and cloudy conditions discourage tillering. The grand growth period is where the cane biomass accumulates and, during this period, sufficient soil moisture is critical to sustain growth (Woodard & Prine, 1993).

The processing of plants into leaves and stalk was done in this study to estimate the amount of potential residue (leaves) that may be produced from stalk-only harvesting of energy cane. The reduction in stalk yield by almost 9 Mg ha⁻¹ occurred after the first year of harvest in 2013. Stalk yield normally declines with cane crop age. Viator et al. (2010) reported that the yield decline with each yearly harvest of a crop cycle can be attributed both to crop age and injury incurred during mechanical harvesting, especially with the cold and sometimes freezing temperature in the temperate regions. The results from the present study showed that the residue produced by energy cane was substantially higher than sugarcane. Franco et al. (2013) documented the amount of residue produced from sugarcane harvesting averaging at 10.7 Mg ha⁻¹. This is significantly lower compared with the 16.2 Mg ha⁻¹ residue produced from energy cane in this study (Table 2).

In this study, the energy cane at scheduled harvest consistently obtained the highest stalk yield across years with an average yield of 25 Mg ha⁻¹. Mislevy et al. (1995) reported similar response of energy cane and Erianthus to harvest date. Their findings showed that (a) the highest total dry biomass yield was attained at the maturity stage and (b) the green leaf yield decreased by 75% when harvested in December as opposed to harvesting in October. According to Richard et al. (1995), the length of the growing season dictates the crop biomass yield. Unlike the year-round warm temperature in the tropical region, the cool and short growing season in temperate climate regions limit cane biomass accumulation. Such growing condition prevails in Louisiana on average years. In addition, several studies showed that tall perennial grasses like energy cane cannot tolerate continuous harvesting at an immature stage without reducing the yield potential of the subsequent ratoon crops (Mislevy & Fluck, 1992; Mislevy et al., 1992; Mislevy et al., 1995; Mislevy et al., 1997; Viator et al., 2010; Woodard & Prine, 1993). All these and the present study suggest that earlier harvesting of energy cane in Louisiana may compromise biomass yield.

Mislevy et al. (1995, 1997) reported that harvest date impacted both the quantity and the quality components of millable stalk of energy cane. Viator et al. (2010) also reported that early harvest resulted in higher stalk moisture contents and reduced ratoon longevity. With the work of Legendre (1975) showing that sucrose content of sugarcane increases as the season progresses, December harvesting was established as ideal for plant cane to maximize sugar yield. This response is quite similar as to how the sucrose content of energy cane in this study responded to November and December harvesting. At these harvest dates, the average sucrose content was 10% compared with the 5–6% obtained when energy cane was harvested in August and September (Table 3). Crop age along with N status, soil moisture, and temperature are factors that influence maturation and sucrose accumulation in sugarcane stalk (Bull, 2000; Tubana et al., 2007). However, the low sucrose content and low levels of TRS and Brix are the main characteristics that makes energy cane an ideal source of feedstock. Therefore, this is considered a positive aspect of early harvesting of energy cane.

Besides harvest date, the significant role of cultivar and N rate on TRS, Brix, and sucrose content was demonstrated in this study (Table 3). The lower levels of TRS, Brix, and sucrose content make Ho 72-114 a better energy cane cultivar than Ho 02-113. These parameters were also reduced in N-fertilized energy cane especially at high N application rates. For example, the average amount of TRS across years decreased from 112 kg Mg⁻¹(0 N) to 73 kg Mg⁻¹ with the application of 224 kg N ha⁻¹. Several studies in sugarcane demonstrated similar trend between TRS and N rate (Lofton & Tubana, 2015; Muchow et al., 1996). According to Muchow et al. (1996), the significant decrease in recoverable sugars with increasing N rate was associated with the decrease in sucrose content in stalk and higher biomass produced by N-fertilized cane than the unfertilized ones. The accumulation of biomass in well N-fertilized cane increased both canopy interception and utilization of solar radiation. Orgeron (2012) also evaluated the response of stalk weight, percent fiber, cane yield, TRS, and sugar yield to increasing N rate application with the outcome indicating that the application rate of 67 kg N ha⁻¹ was as effective as the 157 kg N ha⁻¹.

The stalk fiber content is another quality component that can separate sugarcane from energy cane. In the present study, higher differences in fiber content were observed between years (or crop age) than between cultivars and N rates. A notable impact of harvest date on stalk fiber content was observed but it was not consistent across years. With early harvesting, stalk fiber content was reduced by 15% in 2013, whereas in 2014 and 2015, differences between harvest dates ranged from 0.3 to 4.3% only. The preliminary results of the introgression program conducted by Cana Vialis (a Monsanto group of Company) showed that the selected F1 clones (cross between a commercial hybrid and S. spontaneum) had fiber productivity that ranged from 30.6 to 40.2 Mg ha⁻¹(Matsuoka et al., 2012). This was estimated based on the millable stalks production of these clones per linear meter (35-40) and their fiber content (15.4 to 19.9%). Using this as standard, the average fiber content of the two cultivars evaluated in this study was above average. With proper scheduling of harvest date and optimal application of N fertilizer, further increases in fiber content can be achieved. Even so, the stalk yield is a major factor that determines the (final) fiber productivity. Considering the highest stalk yield and fiber content attained in the present study from the best combinations of harvest date, cultivar, and N rate, the highest estimated fiber productivity was only 11 Mg ha⁻¹.

Cell walls are the major component of plant biomass and consist mainly of three organic compounds: cellulose, hemicellulose, and lignin or collectively termed as lignocellulosic (Yang, 2001). The lignocellulosic composition of cell walls varies widely among species (Popper et al., 2011) and may vary within species (Knox, 2008). The average cellulose, hemicellulose, and lignin content of the stalk in the present study were 29, 27, and 21%, respectively which were very similar to leaves at 29, 29, and 19%, respectively (Figure 2). These % cellulose levels were lower than the % cellulose (44%) in energy cane stalk reported by Ogata (2013) but with comparable results on hemicellulose (22%) and lignin (24%) content. It is interesting to note that the sugarcane bagasse (42%, 25%, and 20%) and energy cane (43%, 24%, and 22%) lignocellulosic composition based on Kim and Day (2011) were very similar and consistent with the report of Ogata (2013). An earlier study conducted by Legendre and Burner (1995) and Baoder and Barrier (1998) also showed the higher cellulose content (38%) of the bagasse fraction of several commercial sugarcane cultivars. Perhaps, the difference in cellulose content can be partly attributed to the differences in the method of analysis implemented in these studies. In the present study, there were small changes in lignocellulosic content of energy cane stalk in response to the treatments. Thus, the stalk yield remains a strong contributor to increase the production of lignocellulosic materials more so if whole-plant harvesting of energy cane is implemented raising biomass yield further with the addition of leaves. According to Burner et al. (2009) one third of the biomass yield of the energy crops they evaluated was leaves with cellulose and lignin content at 48.2% and 16.7%, respectively. Burner et al. (2009) suggested that delaying harvest beyond a freeze in more temperate regions could improve feedstock quality for cellulosic conversion by reducing water concentrations, but this could also reduce the yield of leaves, lignin, ash, and cellulose.

The results from this study indicated the potential of leaf biomass contribution in doubling the production of energy cane biomass yield. However, the long-term impact of complete removal of energy cane biomass from the field can deteriorate soil fertility and subsequently, the sustainability of the entire production system. The annual nutrients removed by energy cane leaves and stalks were estimated in this study (Tables 4, 5, 6). Leite et al. (2016) reported that the N, P, and K removal rates of sugarcane stalks were estimated at 32–168, 5–57, and 26–713 kg ha⁻¹, respectively with lower values for leaves at 19–77, 0.6-4.9, and 2–96 kg ha⁻¹, respectively. The present study confirmed that energy cane, just like sugarcane, removed a considerable amount of these primary nutrients along with Ca, Mg, and S (Tables 4, 5, 6). Conversely, there were apparent differences in leaves nutrient removal rate between energy cane and sugarcane, with energy cane removing two to three times more N, P, and K than sugarcane. The continuous removal of leaves may result in steady decline in soil nutrient content specifically K, Mg, and Ca. Sufficient levels of K in the soil is essential for sugarcane longevity therefore in restoring the productivity of ratoon crops (Schultz et al., 2010; Weber et al., 2002). Similar to the findings of our study, Monti et al. (2008) reported that Ca was mostly concentrated in leaves, whereas K was equally distributed between leaves and stalk of miscanthus. Reumerman and Van de Berg (2002) also reported that a high Ca/K ratio in miscanthus leaves tended to lower slag occurrence. Overall, the value of almost doubling the biomass production in energy cane by implementing whole-plant harvesting can potentially result in increased investment on fertilizers and agronomic practices. The purpose is to match the nutrient removal rate of the whole-plant harvesting system thus preventing yield losses due to poor soil fertility.

5 CONCLUSIONS

This study documented the potential impact of harvest date, cultivar, and N application rate on energy cane biomass yield. The early harvesting of energy cane incurred an average of 12% reduction in stalk yield and 9% reduction in leaf yield. Across harvest dates, the inclusion of leaf in energy cane harvesting raised the total biomass yield by 32% with value equivalent to 16 Mg ha⁻¹. Clearly, the harvesting system (whole-plant vs. stalk) has a greater impact than switching harvest dates on energy cane biomass production. It is important to note that the early harvesting has an added benefits in the form of reduced sucrose content and low levels of TRS and Brix. These are quality components that makes energy cane an ideal source of feedstock. A future research endeavor is to determine whether these benefits can offset the biomass yield decline associated with early harvesting. The role of cultivar and N rate on all measured parameters was mostly implicated with harvest dates. However, mean values and general patterns suggest that Ho 02-113 is a better energy cane cultivar than Ho 72-114. With respect to N rate, 56 and 112 kg N ha⁻¹ were considered optimal because at these rates both the maximum total (stalk and leaf) biomass yield and maximum reduction in TRS, Brix, and sucrose content were attained. Small changes on stalk fiber content and lignocellulosic composition in response to the treatments were documented but no specific and consistent trend was established.

The increase in biomass yield with proper choice of cultivar and N rate was lower than the additional biomass yield from including leaves during harvesting. The caveat of whole-plant harvesting in energy cane is a potential increase in investment for fertilizer and agronomic practices to regain soil fertility. Nutrient removal rates were increased by two to three times when leaf was included in the harvesting. With whole-plant harvesting system, energy cane can remove N, P, and K at rates of 105, 16, and 116 kg ha⁻¹, respectively. The value of leaves if retained in the field as residue versus collecting them for additional feedstock should be carefully considered. The long-term impact of complete removal of energy cane biomass from the field can deteriorate soil fertility and cease organic matter accumulation that can subsequently reduce soil quality and productivity.

ACKNOWLEDGMENTS

Published with the approval of the Director of the Louisiana Agricultural Experiment Station with the number 2020-306-34890. This research paper was partially funded by USDA-NIFA.

AUTHOR CONTRIBUTIONS

Marilyn Sebial Dalen: Formal analysis; Investigation; Methodology; Writing – original draft. Brenda S. Tubana: Conceptualization; Funding acquisition; Investigation; Writing – review & editing. Samuel Kwakye: Methodology; Writing – review & editing. Kun-Jun Han: Methodology; Writing – review & editing.

CONFLICT OF INTEREST

The authors declare no conflict of interest.