b i o m a s s a n d b i o e n e r g y 5 8 ( 2 0 1 3 ) 2 8 6 e2 9 3
Available online at www.sciencedirect.com
http://www.elsevier.com/locate/biombioe
Nitrogen use in switchgrass grown for bioenergy
across the USA
V.N. Owens a, D.R. Viands b, H.S. Mayton b, J.H. Fike c, R. Farris d,
E. Heaton e, D.I. Bransby f, C.O. Hong g,*
a
Plant Science Department, South Dakota State University, 1110 Rotunda Lane North, Brookings, SD 57007, USA
Department of Plant Breeding and Biometry, Cornell University, Ithaca, NY 14853-1901, USA
c
Department of Crop and Soil Environmental Sciences, Virginia Tech, Blacksburg, VA 24061, USA
d
Oklahoma Agricultural Experiment Station, Oklahoma State University, Stillwater, OK 74078, USA
e
Department of Agronomy, Iowa State University, Ames, IA 50011-1010, USA
f
Department of Agronomy and Soils, Energy Crops and Forage/Livestock Management, 202 Funchess Hall, Auburn
University, Auburn, AL 36849, USA
g
Department of Life Science and Environmental Biochemistry, Pusan National University, Miryang 627-706, South
Korea
b
article info
abstract
Article history:
The effect of nitrogen (N) fertilizer on switchgrass biomass production has been evaluated
Received 15 December 2012
in a number of locations on a small-plot scale; however, field-scale information regarding
Received in revised form
switchgrass response to N and N use efficiency (NUE) in different regions of the USA is
14 July 2013
limited. Switchgrass was planted in South Dakota (SD), New York (NY), Oklahoma (OK), and
Accepted 19 July 2013
Virginia (VA) in 2008 and in Iowa (IA) in 2009. Three N levels (0, 56, and 112 kg ha1) were
Available online 17 August 2013
applied to 0.4e0.8 ha plots at each location beginning in spring the year after planting.
Biomass production, N removal, apparent N recovery (ANR), and NUE were determined at
all locations. Biomass yield response to N varied among locations and varied according to
Keywords:
Nitrogen removal
Switchgrass
Bioenergy
Nitrogen use efficiency
Biomass
initial soil N concentration. Low initial soil N concentration increased biomass yield
response to N fertilization, while high initial soil N concentration reduced this response.
High amounts of initial soil N caused fertilizer N removal to be low. Fertilizer N uptake in
switchgrass might be inhibited by competition from initial soil N. Seasonal temperature
and precipitation were not strongly correlated with biomass yield and N-use of switchgrass
at the studied locations. In this study, ANR was below 10% at all locations and years.
Nitrogen-use efficiency was significantly related to initial soil N. High NUE was observed at
Yield
locations where initial soil N was low. These data suggest that NUE depends on site-specific
N management strategies that are responsive to soil N supply and plant N status.
ª 2013 Elsevier Ltd. All rights reserved.
1.
Introduction
Switchgrass has been extensively studied for its value as a
forage, conservation, and bioenergy crop [1e5]. It offers a
number of distinct benefits including broad adaptation,
improved soil conservation and quality [2,6], reduced greenhouse gas emissions [7], and carbon sequestration [6,8e10]. In
particular, it has high yield potential on land marginal to row
* Corresponding author. Tel.: þ82 55 350 5548; fax: þ82 55 350 5549.
E-mail addresses: [email protected], [email protected] (C.O. Hong).
0961-9534/$ e see front matter ª 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.biombioe.2013.07.016
b i o m a s s a n d b i o e n e r g y 5 8 ( 2 0 1 3 ) 2 8 6 e2 9 3
crop production [11]. In previous work in South Dakota USA,
Mulkey et al. [11] found that switchgrass grown in marginal
soil was well suited for sustainable biomass energy
production.
Although switchgrass tolerates low soil fertility, optimizing
biomass and maintaining quality stands requires nitrogen (N)
fertilizer inputs and proper management. Switchgrass responds positively to N fertilization, but its response varies
with regional environment and soil fertility. Switchgrass
biomass increased with increasing N rates up to 168 kg ha1 in
low organic matter and low fertility soils in Texas USA [3], and
Vogel et al. [5] reported that 10e12 kg ha1 of N was required to
produce one tonne per hectare of switchgrass biomass in the
Midwestern USA. However, Mulkey et al. [11] reported no
benefit with N application levels above 56 kg ha1 on
switchgrass-dominated Conservation Reserve Program (CRP)
lands in South Dakota, USA. A major question regarding
switchgrass management as a bioenergy crop is optimizing N
application level. Excessive N fertilization may result in
adverse environmental and economic effects, including
accelerated N2O gas emission, NO3 leaching, and an increase
in production costs.
The amount of N removed in biomass is important in
determining fertilization needs and usefulness as a feedstock.
Matching the N application level with N removal has obvious
agronomic, economic, and environmental value. Bransby
et al. [12] fertilized switchgrass with 100 kg ha1 of N annually
four years in the Southeastern USA, and an average of
87 kg ha1 of N was removed in biomass from the field during
the last three. Stout and Jung [13] reported fertilizer N recovery
of about 31% and 23% following switchgrass fertilization at 90
and 180 kg ha1 y1 in Pennsylvania USA. However, Lemus
et al. [14] reported annualized recovery values of 10%e25% per
year with N application at 90e270 kg ha1 in Virginia USA.
There are two general types of switchgrass cultivars
characterized as lowland and upland. Lowland cultivars are
vigorous, tall, thick-stemmed, and adapted to wetter conditions whereas upland cultivars are short, rhizomatous,
thin-stemmed, and adapted to drier conditions [15]. The
physiological differences between the two switchgrass
types may result in different yield performance in the same
environment. Stroup et al. [16] reported that the lowland
cultivars produced greater biomass yields than upland cultivars in a test conducted in the greenhouse. Nitrogen requirements may also differ between the two cultivars.
Clyder and Porter [17] reported that lowland cultivars of
switchgrass had a lower nitrogen requirement than upland
cultivars.
So far, direct comparisons of N fertilization in replicated
studies of switchgrass across the USA are limited. This study
is one segment of the Regional Feedstock Partnership, a program funded by the US Department of Energy and administered by the Sun Grant Initiative, which was designed to
evaluate dedicated herbaceous energy crops and CRP land
across environmental gradients in the USA. Specifically, the
research reported in this paper provides more information of
switchgrass N-use to improve N management in switchgrass
grown for bioenergy across various regions of the USA. To do
this, we determined 1) switchgrass yield response to N fertilizer; 2) N removal in switchgrass biomass; 3) apparent N
287
recovery (ANR) and NUE of switchgrass grown in different
regions of the USA.
2.
Materials and methods
2.1.
Site description
This study was conducted at five locations across the USA
including South Dakota (SD), New York (NY), Iowa (IA), Oklahoma (OK), and Virginia (VA). The SD location was near Bristol, SD USA (45 160 8.27400 N; 97 50’8.969400 W) on a Nutley-Sinai
(silty clay, mixed, Chromic Hapluderts) with 2e20% slope; the
NY location was near Tompkins, NY USA (42 270 44.589600 N;
76 270 38.188200 W) on an Erie channery (fine-loamy, mixed,
mesic Aeric Fragiaquepts) with 2e8% slope; the IA location
was near Ames, IA USA (41 580 59.00100 N; 93 410 50.034600 W)
on a Clarion-Nicolette (fine-loamy, mixed, mesic Typic
Hapludolls) with 0e9% slope; the OK location was near
Muskogee, OK USA (35 440 32.999400 N; 95 380 21.1200 W) on a
Parsons-Carytown (fine, mixed, thermic Mollic AlbaqualfsAlbic Natraqualfs) with 0e3% slope; and the VA location was
near Pittsylvania, VA USA (36 550 56.265600 N; 79 110 23.884200
W) on a Mayodan (fine sandy loam, mixed, thermic Typic
Hapludults) with 2e15% slope. Seasonal temperature and
precipitation data in 2009 and 2010 were collected from
weather stations at each location (Tables 2 and 3).
2.2.
Experimental design and field management
A locally adapted switchgrass cultivar was planted at each
location. ‘Sunburst’ (SD), ‘Cave-in-Rock’ (NY and IA), ‘Blackwell’ (OK), and ‘Alamo’ (VA) were planted on 17 May 2008 (SD),
29 May 2008 (NY), 8 May 2009 (IA), 2 September 2008 (OK), and 1
July 2008 (VA). Experimental design was a randomized complete block with four replications across the landscape. Individual plot size ranged from 0.4 to 0.8 ha to allow for use of
conventional agricultural equipment. Three levels of N fertilizer (0, 56, and 112 kg ha1) were applied annually beginning
the year after establishment at all locations. Switchgrass was
harvested once annually around a killing frost the year after
establishment (Year 1) and the second year after establishment (Year 2) for SD, NY, OK, and VA. Since the IA location was
planted in 2009, only Year 1 (2010) data are included. Harvest
dates were 28 Oct. 2009 and 5 Nov. 2010 for SD, 22 Oct. 2009
and 2 Nov. 2010 for NY, 18 Nov. 2010 for IA, 13 Nov. 2009 and 28
Oct. 2010 for OK, and 10 Jan. 2010 and 10 Jan. 2011 for VA. Rain
and delayed senescence of switchgrass biomass precluded fall
harvest in VA; therefore, switchgrass was harvested in
January of the following year when soil conditions were
conducive to harvest equipment.
2.3.
Biomass yield
Yield was determined by harvesting a windrow through the
center of each plot (5.5 m 360 m) with locally available
equipment at a height of 10e15 cm. Biomass from each
windrow was baled and weighed. Subsamples (approximately
300 g) were collected with a hay probe (1.3e1.9 cm
wide 45.7e61.0 cm depth) from the center of bales for
288
b i o m a s s a n d b i o e n e r g y 5 8 ( 2 0 1 3 ) 2 8 6 e2 9 3
further analyses. The subsamples were weighed, dried at 60 C
for 48 h in a forced-air oven, reweighed to determine dry
matter yield, and ground in preparation for N analysis. All
subsamples were ground in a Wiley mill (Thomas-Wiley Mill
Co., Philadelphia, PA) to pass a 1-mm screen and reground to
uniformity in a Udy-cyclone impact mill (Udy Co., Ft. Collins,
CO) with a 1-mm screen. Collected subsamples were stored at
room temperature. Subsamples were not collected at VA in
2010 (Year 1), therefore, only Year 2 values are presented for
that site.
Where Nx ¼ N level > 0, and N0 ¼ no N applied.
Apparent N recoveries (%) for each harvest were calculated
using the difference method [21]:
ANR ¼ [(mass of N removed at Nx mass of N removed at
N0)/mass of N applied at Nx] 100.
Where Nx ¼ N level > 0, and N removal was calculated by
multiplying biomass yield times N concentration in the
biomass.
2.4.
Statistical analysis was performed using SAS version 9.2 [22].
PROC ANOVA was carried out to compare the means of the
different treatments. Location and N level were considered
fixed. The least significant difference was used to separate
means among location and N level treatments when the
appropriate F test was significant (P ¼ 0.05). Year 1 and 2 refer
to the first and second year after establishment, respectively.
Year 1 is 2009 for SD, NY, OK, VA, and 2010 for IA. PROC GLM
was carried out to compare the means across locations and N
levels of Year 1 and 2. The Duncan grouping was used to
separate means across locations and N levels of Year 1 and 2.
All seasonal temperature and precipitation, initial soil N,
biomass yield, and all N removal and N use metrics were
subjected to correlation analysis using PROC CORR to identify
associations between seasonal temperature and precipitation
and biomass yield, N removal, and N-use metrics and correlations between initial soil N and biomass yield, N removal,
and N-use metrics.
Soil sampling
A hydraulic soil probe (6.6 cm internal diameter) was used to
collect soil samples at initiation of the research in May 2008
for SD, Apr. 2008 for NY, May 2009 for IA, Oct. 2008 for OK, and
Mar. 2009 for VA. Four random cores were collected from each
plot to a depth of 1 m. Each core was subdivided into depth
increments of 0e5, 5 to 15, 15 to 30, 30 to 60, and 60e100 cm,
after which soil from each of the four cores at each depth was
composited for analysis. Surface residue was removed before
sampling. Soil samples were initially sieved to pass an 8-mm
screen and dried in a forced-air oven at 40 C until consistent mass was attained. Visible plant residue and roots were
removed before drying. Dried soil samples were ground to
pass a 2-mm screen for chemical analysis.
2.5.
Chemical analysis
Concentrations of total N in switchgrass biomass were predicted for all samples using near infrared reflectance spectroscopy (NIRS) (NIRS Model 5000; Foss NIRSystems, Silver
Springs, MD) based on a calibration data set of 192 samples
representing all harvest years [18]. A set of 40 samples was
used for cross-validation. Calibration and validation statistics
were generated using WinISI (Version 1.5) system software
(Infrasoft International LLC., State College, PA). For calibration
and validation samples, total N was quantified using a Vario
Max CNS elemental analyzer (Elementar Instrument, Mt.
Laurel, NJ).
Total soil N at each depth was determined by dry combustion method in a Vario Max CNS elemental analyzer.
2.6.
Nitrogen removal
Annual nitrogen (N) removal was calculated by multiplying
the biomass yield by the nitrogen mass fraction of the dry
material. The annual Fertilizer N removal was calculated as
the difference between the total N removal from the unfertilized treatment (0 kg ha1) and total N removal from the
fertilized treatments (56 or 112 kg ha1).
2.7.
Nitrogen-use metrics
Nitrogen use in relation to dry matter yield was determined
using two different metrics: nitrogen-use efficiency (NUE) and
apparent nitrogen recovery (ANR).
Nitrogen-use efficiency (mass of biomass per mass of N) for
each harvest was calculated as [19,20]: NUE ¼ (yield at
Nx yield at N0)/mass of N applied.
2.8.
Statistical analysis
3.
Results and discussion
3.1.
Biomass yield
Location significantly affected biomass yield for Year 1 and 2
(Table 1). Averaged across N levels, biomass yield was highest
at NY and lowest at SD in Year 1 and 2 (Table 4). The order of
average biomass yield, from highest to lowest, was
NY > IA > OK > SD for Year 1 and NY > VA > OK > SD for Year
2. Seasonal temperature could be one important factor
affecting production of switchgrass. Switchgrass grows best
and responds to N fertilizer at high temperature (32.2 C day/
26.1 C night) [23]. However, average monthly temperature
was highest at VA of all locations in both years, but it was not
the highest yielding location (Table 2). Relationships between
seasonal temperature and biomass yield were not significant
(Table 2). Precipitation amount and timing has also been
shown to affect native grass yield in various regions of the
USA [24e27]. However, annual precipitation amount was
highest at VA of all locations in both years (Table 3), and there
was no significant relationship between seasonal precipitation and biomass yield (Table 3).
A locally adapted switchgrass cultivar was planted at each
location. Cultivar likely contributed to differences in biomass
yield among locations. Alamo (a lowland cultivar utilized in
VA) generally shows greater biomass production in the
Southern Great Plains compared with Sunburst, Cave-in-Rock,
and Blackwell (all upland cultivars utilized in SD, NY and IA,
and OK, respectively) in the Northern Great Plains [28].
289
b i o m a s s a n d b i o e n e r g y 5 8 ( 2 0 1 3 ) 2 8 6 e2 9 3
Table 1 e Analysis of variance (ANOVA) and probability values for biomass yield, N concentration, total N removal, fertilizer
N removal, apparent N recovery (ANR), and N use efficiency (NUE) at all locations.
Year 1
df
Biomass yield
df
N concentration
Total N removal
df
Fertilizer N removal
ANR
NUE
Year 2
Block
Location (L)
N level (NL)
L NL
Block
Location (L)
N level (NL)
L NL
3
NSa
3
NS
NS
3
NS
NS
NS
4
<0.001
3
<0.001
<0.001
3
NS
NS
NS
2
NS
2
NS
NS
1
NS
NS
NS
8
NS
6
NS
NS
3
NS
NS
NS
3
NS
3
NS
NS
3
NS
NS
NS
3
<0.001
3
<0.001
<0.001
3
NS
NS
<0.001
2
NS
2
NS
NS
1
NS
NS
NS
6
0.004
6
NS
NS
3
NS
NS
NS
a NS: not significant.
Fuentes and Taliaferro [29] reported the mean switchgrass dry
matter yield of lowland cultivars Alamo and Kanlow to be
higher than the mean of upland cultivars Cave-In-Rock and
Blackwell every year from 1994 through 2000 in Oklahoma.
However, biomass production of Blackwell, which was planted in NY, was much greater than that of Alamo, which was
planted in VA in this study (Table 4). This result also confirms
that factors in addition to cultivar affected differences in
biomass yield among locations. Tulbure et al. [30] utilized
numerous variables to model switchgrass yields across the
USA and concluded that fertilizer application (N primarily),
genetics, precipitation, and other management practices were
the most important for explaining switchgrass yield variability despite the fact that precipitation was not significant in
this study.
Average biomass yield across locations and N levels was
always higher the second year after planting (Year 2) than the
first year (Year 1) (Table 4). Biomass yield of switchgrass
generally increases with year once it is well established, and
may take up to three years to reach its full production potential [31]. Depending on the region, it can typically produce
1/4 to 1/3 of its yield potential in the establishment year and 2/
3 of its potential the year after planting [32].
Native soil N could be an important factor affecting production of switchgrass. Stout and Jung [13] reported that
switchgrass grown in soil having higher N levels had higher
biomass accumulation. Of all locations, initial soil N at 0e5
and 5e15 cm soil depth was highest in NY (Table 5). The order
of initial soil N, from highest to lowest, was NY > IA >
OK > SD > VA. However, this order was different from that of
biomass yield (Table 4), and there was no significant relationship between initial soil N and biomass except for 0e5 cm
depth in Year 1 (Table 7). The lack of a significant correlation
was due at least in part to the fact that yield at VA was higher
than SD in Year 1 and SD and OK in Year 2 despite having less
initial soil N. There was a significant location N level
Table 2 e Seasonal temperature at all locations and correlation coefficients for relationships between seasonal temperature
and biomass yield (n [ 7 for biomass yield). Data were obtained from two harvest years in SD, NY, OK, and VA and one year
in IA USA.
Year
2009
2009
2009
2009
2010
2010
2010
2010
2010
30 y avg.a
Correlation coefficientb
Location
SD
NY
OK
VA
SD
NY
IA
OK
VA
SD
NY
IA
OK
VA
Year after
planting
1st
1st
1st
1st
2nd
2nd
1st
2nd
2nd
Averaged temperature ( C)
Annual
5.1
7.5
8.9
12.5
6.7
8.7
9.3
9.8
12.9
6.3
7.8
9.5
15.5
12.5
0.17
Jan.eSep.
Jan.eApr.
Apr.eMay
Apr.eSep.
MayeJun.
MayeJul.
7.2
8.8
10.4
14.1
9.0
10.6
11.3
11.8
14.9
8.6
9.2
11.6
17.3
14.1
0.14
5.1
0.9
1.6
5.8
3.0
0.7
0.4
2.0
5.5
3.5
0.8
0.8
8.3
5.9
0.20
9.5
10.6
10.8
15.1
11.8
12.4
14.6
12.7
16.4
10.4
9.8
13.3
17.9
14.3
0.15
15.2
15.2
15.7
19.5
17.2
17.1
19.4
18.0
21.1
16.4
15.5
18.6
23.0
19.3
0.07
15.7
14.9
17.0
20.0
16.1
17.1
18.9
18.6
21.4
16.5
15.5
18.9
22.6
19.0
0.04
16.8
16.0
17.4
20.6
18.1
18.7
20.5
20.0
22.6
18.3
17.1
20.4
24.3
20.6
0.04
a 30 y avg.: average of monthly temperatures for 30 years (1979e2008) at each location.
b Correlation coefficient: correlation coefficients for relationships between seasonal temperature and biomass yield.
290
b i o m a s s a n d b i o e n e r g y 5 8 ( 2 0 1 3 ) 2 8 6 e2 9 3
Table 3 e Seasonal precipitation at all locations and correlation coefficients for relationships between seasonal
precipitation and biomass yield (n [ 7 for biomass yield). Data were obtained from two harvest years in SD, NY, OK, and VA
and one year in IA USA.
Year
Location
Year after
planting
2009
2009
2009
2009
2010
2010
2010
2010
2010
30 y avg.a
SD
NY
OK
VA
SD
NY
IA
OK
VA
SD
NY
IA
OK
VA
Correlation coefficientb
1st
1st
1st
1st
2nd
2nd
1st
2nd
2nd
Sum of precipitation (mm)
Annual
Jan.eSep.
Jan.eApr.
Apr.emay
Apr.eSep.
MayeJun.
MayeJul.
623
848
1079
1424
677
980
1180
848
1215
596
933
906
1120
1146
0.07
487
674
719
959
552
673
1081
768
991
517
709
759
833
886
0.08
98
177
291
327
119
273
139
223
449
113
255
187
300
385
0.28
48
149
234
249
102
119
120
196
177
130
166
215
244
197
0.04
411
549
549
716
467
458
978
592
586
457
538
666
630
595
0.10
104
218
174
285
169
153
366
251
194
174
180
249
274
199
0.01
163
301
219
392
233
235
495
366
281
267
270
370
338
301
0.04
a 30 y avg.: average of monthly precipitations for 30 years (1979e2008) at each location.
b Correlation coefficient: correlation coefficients for relationships between seasonal precipitation and biomass yield.
interaction for biomass yield (Table 1). Biomass yield did not
respond to N at any location in Year 1 but there was a response
in Year 2 in SD, NY, and VA (Table 4). Biomass yield increased
with N level up to 56 kg ha1 in SD, but no further yield increases were seen with additional N. This result was similar to
that of Mulkey et al. [11] who reported no benefit with N
application levels above 56 kg ha1 on switchgrass-dominated
CRP lands in SD. In NY, biomass yield decreased significantly
with N applied in Year 2. There was a lodging issue in fertilized
plots at NY in both years; therefore it was difficult to effectively harvest all biomass. This might partly account for the
negative response of biomass yield to N at NY. In VA, biomass
yield increased with each successive N level. There was no
yield response of N at IA (Year 1) and OK (Year 1 and 2). Overall
response of biomass to N might also be associated with initial
soil N. Relatively low initial soil N levels at SD and VA
Table 4 e . Biomass yield, N concentration, and total N removal as affected by N level in South Dakota (SD), New York (NY),
Iowa (IA), Oklahoma (OK), and Virginia (VA) USA in the first (Year 1) and second (Year 2) years after switchgrass
establishment.
N level (kg/ha)
Year 1
SD
Year 2
NY
IA
OK
VA
SD
NY
OK
VA
9.08a
8.96a
8.03a
8.69A
5.43a
5.26a
5.74a
5.48B
3.75a
3.94a
4.69a
4.13C
3.29a
4.48a
4.65a
4.14C
3.90a
5.42a
5.24a
4.85C
7.07A
11.49a
10.50b
9.76b
10.6A
5.52a
5.94a
5.79a
5.75C
5.21b
6.95a,b
9.09a
7.08B
4.75a
4.73a
5.04a
4.84B
2.93a
3.49a
3.31a
3.24C
8.11a
8.33a
7.99a
8.14A
1.42a
1.36a
2.20a
1.66C
2.44B
2.36a
2.76a
3.39a
2.84B
3.83a
3.87a
4.13a
3.94A
1.25a
1.35a
1.38a
1.33C
1
Biomass yield (Mg ha )
0
1.98a
56
2.69a
112
3.12a
a
Year mean
2.60D
L NL meanb
5.01B
1
N concentration (g kg )
0
3.18a
56
3.35a
112
4.59a
Year mean
3.71C
L NL mean
4.98A
1
Total N removal (kg ha )
0
6.04b
56
9.05ab
112
14.5a
Year mean
9.88D
L NL mean
23.8A
39.2a
38.5a
36.8a
38.2A
15.8a
16.1a
18.7a
16.9C
28.1a
29.5a
33.0a
30.2B
5.29a
7.43a
11.1a
7.92C
17.1B
27.1a
28.9a
33.5a
29.8A
19.2a
20.8a
21.7a
20.6B
6.56a
9.39a
14.4a
10.1C
a Year mean: mean value across N levels.
b L NL mean: mean value across locations and N levels. Upper and lower case letters are for column and row comparison. Values with same
letter within a column or row are not significantly different a P ¼ 0.05.
291
b i o m a s s a n d b i o e n e r g y 5 8 ( 2 0 1 3 ) 2 8 6 e2 9 3
Table 5 e Initial soil N concentration at various soil depth
in South Dakota (SD), New York (NY), Iowa (IA), Oklahoma
(OK), and Virginia (VA) USA.
Soil depth
increment (cm)
0e5
5e15
15e30
30e60
60e100
Total N concentration (g kg1)
SD
NY
IA
OK
VA
1.93c
1.55b
1.04b
0.50c
0.30bc
3.02a
2.29a
1.80a
1.22a
2.53b
2.26a
1.86a
1.03ab
0.68a
2.39b
1.68b
1.17b
0.86b
0.54ab
1.12d
0.55c
0.42c
0.29c
0.20c
Values with same letter within a row are not significantly different
at P ¼ 0.05.
increased the need for fertilizer N, but high initial soil N levels
at NY, IA, and OK may have reduced this need.
3.2.
Biomass N concentration and N removal
Location significantly affected N concentration in switchgrass
in both years, but N level did not (Table 1). Average switchgrass
N concentration increased with N level at all locations and in all
years, but it was not significant (Table 4). There was significant
broadleaf weed encroachment in OK in Year 1, likely leading to
elevated N in harvested biomass compared with other locations. In switchgrass production fields, many broadleaf weeds
are leguminous plants capable of fixing atmospheric nitrogen
(N2) [33]. Hong et al. [34] reported that presence of broadleaf
weeds may have led to elevated N concentration in biomass in
mixtures of warm-season grasses.
Location significantly affected total N removal for both
years, but N level did not (Table 1). Average total N removal
across locations and N levels was higher in Year 1 than in Year
2 (Table 4) despite the fact that yield was greater in Year 2. This
was mainly due to higher N concentration in harvested
biomass in Year 1. Perennial C4 grasses, such as switchgrass,
translocate up to 30% of shoot N to rhizomes and roots during
senescence [35]. Parrish and Wolf [36] observed significant
redistribution of N into belowground biomass of switchgrass
at the end of the growing season. Heggenstaller et al. [37] also
observed that N level affected switchgrass root biomass and
nutrient partitioning, with 140 kg ha1 maximizing root
biomass. Switchgrass in Year 2 may not have required as
much applied N as in Year 1 because some of the N from Year
1 was retained in belowground biomass or may have
remained in the soil. This might partially account for lower N
concentrations in switchgrass grown in Year 2 compared with
Year 1. Also, a dilution effect from increased biomass in Year
2, driven largely by a greater proportion of biomass from internodes, could also be a reason for lower N concentrations in
switchgrass grown in Year 2. Average total N removal across N
levels was different among locations, and was highest in NY
and lowest in SD in both years (Table 4). High total N removal
at NY compared to other locations resulted from high biomass
production while the opposite was true for SD. Relatively high
total N removal at OK resulted from high levels of N in harvested biomass in Year 1 as described previously.
The average amount of calculated fertilizer N removed with
biomass did not significantly increase with N level (Table 6).
Fertilizer N removal was negatively correlated with initial soil N
concentration at the all soil depths (Table 7), with the highest
correlations occurring at 0e5, 30e60, and 60e100 cm soil
depths in Year 1. The high initial soil N concentration at NY
along with a negative response to fertilizer N caused fertilizer N
removal to be lowest among locations, even though differences
were not significant among locations (Table 6). Based on these
results, fertilizer N uptake in switchgrass might be inhibited by
high levels of initial soil N. Our results were similar to that of
Stout and Jung [13] who reported high initial soil N resulted in a
lower fertilizer N accumulation rate in switchgrass.
3.3.
Apparent N recovery and N use efficiency
Apparent nitrogen recoveries were below 10% at all locations
and years (Table 6). Maximum ANR was 7.59% with 112 kg ha1
Table 6 e . Nitrogen level effect on fertilizer N removal, apparent N recovery (ANR), and N use efficiency (NUE) in switchgrass
biomass harvested at South Dakota (SD), New York (NY), Iowa (IA), Oklahoma (OK), and Virginia (VA) USA the first (Year 1)
and second (Year 2) years after establishment.
N level (kg/ha)
Year 1
SD
Fertilizer N removal (kg ha1)
56
3.02a
112
8.50a
a
Year mean
5.76A
ANR (%)
56
5.39a
112
7.59a
Year mean
6.49A
NUE (kg kg1)
56
12.75a
112
10.22a
Year mean
11.49A
NY
Year 2
IA
OK
SD
NY
OK
VA
0.66a
2.35a
1.50A
0.31a
2.87a
1.59A
1.46a
4.88a
3.17A
2.13a
5.76a
3.95A
1.82a
1.34a
1.58A
1.63a
2.47a
2.05A
2.83a
7.81a
5.32A
1.18a
2.10a
1.64A
6.90a
6.40a
6.65A
2.60a
4.36a
3.48A
3.81a
5.14a
4.48A
3.25a
5.66a
4.46A
2.92a
2.20a
2.56A
5.06a
6.97a
6.01A
2.01a
8.54a
5.27A
3.10a
2.73a
0.19A
3.04a
7.62a
5.33A
27.02a
11.96a
19.49AB
17.65a
15.44a
16.55C
6.79a
2.19a
4.49B
30.94a
34.64a
32.79A
a Year mean: mean value across N levels. Upper and lower case letters are for column and row comparison. Values with same letter within a
column or row are not significantly different a P ¼ 0.05.
292
b i o m a s s a n d b i o e n e r g y 5 8 ( 2 0 1 3 ) 2 8 6 e2 9 3
Table 7 e Correlation coefficients for relationships between initial soil N concentration and biomass yield, N concentration
of switchgrass biomass, total N removal, fertilizer N removal, apparent N recovery (ANR), and N use efficiency (NUE) (n [ 2
for 0e5, 5e15, 15e30, and 30e60 cm soil depth and n [ 1 for 60e100 cm soil depth) in switchgrass harvested the first (Year
1) and second (Year 2) years after establishment. Data were obtained from five soil depths in SD, IA, OK, and VA and four
soil depths in NY USA.
Native soil N
in depth (cm)
Biomass yield
Year 1
0e5
5e15
15e30
30e60
60e100
Year 2
0e5
5e15
15e30
30e60
60e100
0.982*
0.868
0.820
0.940
0.993
0.536
0.448
0.583
0.667
0.384
N concentration
Total N removal
Fertilizer N removal
ANR
NUE
0.083
0.365
0.407
0.058
0.064
0.843
0.446
0.374
0.772
0.478
0.996*
0.892
0.850
0.978*
1.000***
0.814
0.435
0.352
0.687
0.106
0.980*
0.940
0.909
0.989*
0.993
0.872
0.773
0.840
0.956*
0.903
0.544
0.524
0.423
0.476
0.989
0.609
0.590
0.493
0.540
0.985
0.984*
0.943
0.974*
0.997**
0.979
0.708
0.622
0.585
0.724
0.986
*, **, and *** denote significance at 0.05, 0.01 and 0.001 levels of probability, respectively.
at SD in Year 1. This recovery rate is much lower than found by
other researchers who reported N recovery of 31% following
switchgrass fertilization at 90 kg ha1 y1 in Pennsylvania USA
[13], but similar to annualized values of about 10% with
90 kg ha1 in Virginia USA [14]. At this recovery rate, more
than 90% of total fertilizer N is unaccounted for, and may be
susceptible to loss through leaching, denitrification, and
volatilization, or some portion of it may have become
sequestered in belowground pools, i.e., roots, microbial
biomass, and/or soil organic matter. In addition to wasting
money on unnecessary fertilizer, excessive N application
beyond plant requirements could cause adverse environmental effects through N2O gas emission and NO3 leaching.
Location significantly affected NUE in Year 2, but N level
did not in either year (Table 1). There was a strong negative
correlation of NUE with initial soil N concentration at all soil
depths in both years (Table 7). The order of NUE from highest
to lowest was VA > SD > OK > NY in Year 2 (Table 6). This was
opposite of the order of initial soil N concentration
(NY > OK > SD > VA) (Table 5). Compared to other locations SD
had a relatively high NUE. Low initial soil N at SD resulted in
high fertilizer N accumulation in switchgrass and an increase
in biomass yield response to N fertilization. These data suggest that N-use efficiency depends on site-specific N management strategies that are responsive to soil N supply and
plant N status.
4.
Conclusion
Based on the results in this study, biomass yield response of
switchgrass to N fertilizer varied among locations across the
USA. Biomass yield response to N fertilization depended on
initial soil N concentration. Low initial soil N increased
biomass yield response to N fertilization while high initial soil
N reduced the overall response. High amounts of initial soil N
caused fertilizer N removal to be low. Seasonal temperature
and precipitation were not significantly correlated to biomass
yield and N-use of switchgrass at the studied locations. In this
study, ANR was below 10% at all locations and years, suggesting further work is necessary to assess the fate of unrecovered N fertilizer. Nitrogen-use efficiency was significantly
related to initial soil N concentration. Higher NUE values were
observed at locations where initial soil N level was low.
Acknowledgments
This research was supported by funding from the North
Central Regional Sun Grant Center at South Dakota State
University through a grant provided by the US Department of
Energy Office of Biomass Programs under award number DEFC36-05GO85041.
references
[1] Caslers MD, Boe AR. Cultivar environment interactions in
switchgrass. Crop Sci 2003;43(6):2226e33.
[2] McLaughlin SB, Walsh ME. Evaluating environmental
consequences of producing herbaceous crops for bioenergy.
Biomass Bioenergy 1998;14(4):317e24.
[3] Muir JP, Sanderson MA, Ocumpaugh WR, Jones RM, Reed RL.
Biomass production of ‘Alamo’ switchgrass in response to
nitrogen, phosphorus, and row spacing. Agron J
2001;93(4):896e901.
[4] Sanderson MA, Read JC, Reed RL. Harvest management of
switchgrass for biomass feedstock and forage production.
Agron J 1999;91(1):5e10.
[5] Vogel KP, Brejda JJ, Walters DT, Buxton DR. Switchgrass
biomass production in the midwest USA: harvest and
nitrogen management. Agron J 2002;94(3):413e20.
[6] Lee DK, Owens VN, Doolittle JJ. Switchgrass and soil carbon
sequestration response to ammonium nitrate, manure, and
harvest frequency on Conservation Reserve Program land.
Agron J 2007;99(2):462e8.
[7] Mclaughlin SB, Samson R, Bransby D, Wiselogel A. Evaluating
physical, chemical, and energetic properties of perennial
b i o m a s s a n d b i o e n e r g y 5 8 ( 2 0 1 3 ) 2 8 6 e2 9 3
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
grasses as biofuels. In: Abstracts of BioEnergy ’96-The
Seventh National Bioenergy Conference: Partnerships to
Develop and Apply Biomass Technologies, Nashville, TN
USA. September 15-20, 1996. p. 1e8.
Garten CT, Wullschleger SD. Soil carbon dynamics beneath
switchgrass as indicated by stable isotope analysis. J Environ
Qual 2000;29(2):645e53.
McLaughlin SB, DeLaTorreUgarte DG, Garten CT, Lynd LR,
Sanderson MA, Tolbert VR, et al. High-value renewable energy
from prairie grasses. Environ Sci Technol 2002;36(10):2122e9.
Liebig MA, Kronberg SL, Gross JR. Effects of normal and
altered cattle urine on short-term greenhouse gas flux from
mixed-grass prairie in the northern great plains. Agr Ecosyst
Environ 2008;125(1e4):57e64.
Mulkey VR, Owens VN, Lee DK. Management of switchgrassdominated Conservation Reserve Program lands for biomass
production in South Dakota. Crop Sci 2006;46(2):712e20.
Bransby DI, Ward CY, Rose PA, Sladden SE, Kee DD. Biomass
production from selected herbaceous species in the
Southeastern USA. Biomass 1989;20(3e4):187e97.
Stout WL, Jung GA. Biomass and nitrogen accumulation in
switchgrass: effects of soil and environment. Agron J
1995;87(4):663e9.
Lemus R, Brummer EC, Burras CL, Moore KJ, Barker MF,
Molstad NE. Effects of nitrogen fertilization on biomass yield
and quality of large fields established switchgrass in
southern Iowa, USA. Biomass Bioenergy 2008;32(12):1187e94.
Moser LE, Vogel KP. Switchgrass, big bluestem, and
indiangrass. In: Barnes RF, Miller DA, Nelson CJ, editors.
Forages. An introduction to grassland agriculture, vol. 1.
Ames, IA: Iowa State Universtiy; 1995. p. 409e20.
Stroup JA, Sanderson MA, Muir JP, McFarland MJ, Reed RL.
Comparison of growth and performance in upland and
lowland switchgrass types to water and nitrogen stress.
Bioresource Technol 2003;86(1):65e72.
Clyder L, Porter JR. An analysis of variation between upland
and lowland switchgrass, Panicum virgatum L., in central
Oklahoma. Ecology 1996;47(6):980e92.
Garcia-Ciudad A, Garcia-Criado B, Perez-Corona ME, Vazquez
de Aldana BR, Ruano-Ramos AM. Application of nearinfrared reflectance spectroscopy to chemical analysis of
heterogeneous and botanically complex grassland samples. J
Sci Food Agric 1993;63(4):419e26.
Zemenchick RA, Albrecht KA. Nitrogen use efficiency and
apparent nitrogen recovery of Kentucky bluegrass, smooth
bromegrass, and orchard grass. Agron J 2002;94(3):421e8.
Novoa R, Loomis RS. Nitrogen and plant production. Plant
Soil 1981;58(1e3):177e204.
Crasswell ET, Godwin DC. The efficiency of nitrogen
fertilizers applied to cereals in different climates. In:
Thinker PB, Lauchli A, editors. Advances in plant nutrition.
New York: Praeger; 1984. p. 1e55.
SAS Institute Inc. SAS/STAT 9.2 User’s guide. Cary, NC: SAS
Institute Inc.; 2008.
Barnes RF, Nelson CJ, Collins M, Moore KJ. Forages: an
introduction to grassland agriculture. In: Volence JJ,
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
293
Nelson CJ, editors. Environmental aspects of forage
management. Ames, IA: Blackwell Publishing; 2003.
p. 99e124.
Briggs JM, Knapp AK. Interannual variability in primary
production in tallgrass prairie: climate, soil moisture,
topographic position, and fire as determinants of
aboveground biomass. Am J Bot 1995;82(8):1024e30.
Lee DK, Boe A. Biomass production of switchgrass in central
South Dakota. Crop Sci 2005;45(6):2583e90.
Shiflet TN, Dietz HE. Relationship between precipitation and
annual rangeland herbage production in Southeastern
Kansas. J Range Manage 1973;27(4):272e6.
Smart AJ, Dunn BH, Jobnson PS, Xu L, Gates RN. Using
weather data to explain herbage yield on three great plains
plant communities. Rangeland Ecol Manag
2007;60(2):146e53.
McLaughlin SB, Kszos LA. Development of switchgrass
(Panicum virgatum) as a bioenergy feedstock in the United
States. Biomass Bioenerg 2005;28(6):515e35.
Fuentes RG, Taliaferro CM. Biomass yield stability of
switchgrass cultivars. In: Janick J, Whipkey A, editors. Trends
in new crops and new uses. Alexandria, VA: ASHS Press;
2002. p. 276e82.
Tulbure MG, Wimberly MC, Boe A, Owens VN. Climatic and
genetic controls of yields of switchgrass, a model bioenergy
species. Ag Ecosyst Eviron 2012;146(1):121e9.
Samson R. Switchgrass production in Ontario: a
management guide (PDF) [monograph on the internet].
Resource efficient agricultural production (REAP) e Canada.
Available from: http://www.reapcanada.com/library/
Bioenergy/2007%20SG%20production%20guide-FINAL.pdf;
2007.
Bransby DI. Switchgrass profile. Bioenergy feedstock
information network (BFIN) [monograph on the internet].
Oak Ridge National Laboratory; 2005. Available from: http://
bioenergy.ornl.gov/papers/misc/switchgrass-profile.html.
Newton WE, Orme-Johnson WH. Nitrogen fixation volume II:
symbiotic associations and cyanobacteria. In: Boland MJ,
Farnden KJF, Robertson JG, editors. Ammonia assimilation in
nitrogen-fixing legume nodules. Baltimore: University Park
Press; 1980. p. 33e52.
Hong CO, Owens VN, Lee DK, Boe A. Switchgrass, big
bluestem, and indiangrass monocultures and their two- and
three-way mixtures for bioenergy in the northern great
plains. Bioenerg Res 2013;(6):229e39.
Heckathorn SA, Delucia H. Retranslocation of shoot nitrogen
to rhizomes and roots in prairie grasses may limit loss of N to
grazing and fire during drought. Funct Ecol
1996;10(3):396e400.
Parrish DJ, Wolf DD. Managing switchgrass for sustainable
biomass production. In: Proc. symp. liquid fuels from
renewable resources, Nashville, TN, 13e14 December 1992.
St. Joseph, MI: ASAE; 1992. p. 34e9.
Heggenstaller AH, Moore KJ, Liebman M, Anex RP. Nitrogen
influences biomass and nutrient partitioning by perennial,
warm-season grasses. Agron J 2009;101(6):1363e71.