• Open Access

Seasonal nitrogen dynamics of Miscanthus×giganteus and Panicum virgatum

Authors

  • EMILY A. HEATON,

    1. Department of Crop Sciences, University of Illinois, 1201 W. Gregory Drive, 379 Madigan Lab., Urbana, IL 61801, USA,
    Search for more papers by this author
    • 1Current address: Department of Agronomy, Iowa State University, 1403 Agronomy Hall, Ames, IA, USA, e-mail: heaton@iastate.edu

  • FRANK G. DOHLEMAN,

    1. Department of Plant Biology, University of Illinois, 1201 W. Gregory Drive, 379 Madigan Lab., Urbana, IL 61801, USA,
    Search for more papers by this author
  • STEPHEN P. LONG

    1. Department of Plant Biology, University of Illinois, 1201 W. Gregory Drive, 379 Madigan Lab., Urbana, IL 61801, USA,
    2. Institute for Genomic Biology, University of Illinois, 1206 W. Gregory Drive, Room 126 IGB. Urbana, IL 61801, USA
    Search for more papers by this author

Stephen P. Long, tel. +1 217 333 2487, fax +1 217 244 7563, e-mail: slong@uiuc.edu

Abstract

There is a tradeoff to consider when harvesting perennial biomass crops: harvest too late and yield declines, harvest too early and risk higher mineral contents, particularly nitrogen (N). Allowing the crop to completely senesce and recycle nutrients has several advantages, including improved feedstock quality and reduced fertilizer requirements, but it comes at a risk, particularly in temperate climates where snow and ice can reduce or destroy harvestable biomass. The effect of harvest time on the N concentration ([N]) and biomass of Panicum virgatum and Miscanthus×giganteus was evaluated at three sites in Illinois over two years. In both species [N] of standing biomass significantly declined with time (P<0.0001). Interestingly, there was no significant interaction effect of species and sample date on [N] (P=0.2888), but there was a highly significant interaction effect on the total N in standing biomass (P<0.0001). The amount of standing N was directly related to biomass yield. Seasonal changes in standing N differed among locations, suggesting that harvest time recommendations for N management depend on location. P. virgatum would have potentially removed as much as 187 kg N ha−1 if harvested green, and as little as 5 kg N ha−1 if harvested in late winter. Because of higher biomass yields, M. ×giganteus standing N ranged from 379 kg N ha−1 in June to <17 kg N ha−1 in February. Importantly, there was little overall change in [N] between an early winter (December) harvest and a late winter (February/March) harvest, indicating the benefits of N cycling in the system can be realized by end of the growing season and thus, at least from an N economy perspective, there is no reason to risk yield losses by delaying harvest over the winter.

Introduction

Perennial grasses that can translocate nutrients from aboveground to belowground organs for overwinter storage before harvest may provide cleaner biomass feedstock than annual crops that cannot directly cycle nutrients. Deciduous trees can recover nutrients from leaves in the fall, however, most are stored in the wood parenchyma and so will be removed with a harvest of wood (Kauter et al., 2003). Minimizing mineral contents and understanding what controls them is important in a fuel crop, as some plant nutrients such as nitrogen (N) and sulfur become atmospheric pollutants upon combustion of the fuel or feedstock and others (e.g. chlorine and silica) can disrupt boiler function (Lewandowski & Kicherer, 1997; McKendry, 2002). While detrimental in fuel feedstock, N is critical to crop growth and is the nutrient needed by the plant in the greatest quantities (Snyder & Leep, 2007). Because N is a major component of key photosynthetic enzymes, its availability helps determine the rate at which plants turn solar energy into the stored chemical energy of plant biomass from which renewable biofuels like ethanol, biogas or bio-oil can be made.

Bioenergy crops have an advantage over other renewable energy sources such as wind and solar cells in that they can be stored to provide a source of energy that can be drawn upon as needed. However, an annual harvest means that huge volumes of feedstock will need to be stored to supply either ethanol plants or power plants operating 365 days of the year (Fales et al., 2007). Can this feedstock be stored in the field? While the moisture content of stored biomass is typically the major determinant of dry matter losses in baled storage of feedstock (Sanderson et al., 1997; El Bassam & Huisman, 2001; Shinners & Binversie, 2007), the N concentration ([N]) may be important as well. Given that decomposition is strongly related to the amount and form of N in biomass (Lambers et al., 1998), it is reasonable to expect that in addition to making a cleaner fuel, a low [N] in biomass crops might reduce dry matter loss during storage. If plant residues with a higher ratio of carbon to nitrogen (C/N) are more resistant to decomposition in natural ecosystems (Fageria et al., 2005; Manzoni et al., 2008) then bioenergy crops with a high C/N may prove more resistant to degradation and be suitable for storage in the field, provided moisture contents can be kept acceptably low. Field-stored feedstock could then be provided to the processing plant on a ‘just-in-time’ basis, saving on the cost of large-scale covered storage.

In addition to impacting feedstock quality, N translocation can reduce the need for external fertilizer. In particular, temperate perennial grasses are known for mobilizing N from actively growing tissues to storage organs in response to winter or drought conditions (Heckathorn & Delucia, 1994; Suzuki & Stuefer, 1999). Harvesting feedstocks after senescence, the process whereby cellular constituents are systematically broken down and transported out of senescing organs as visually indicated by yellowing and death of green tissue (Lambers et al., 1998), could allow >50% of leaf N to be resorbed by the plant (van Heerwaarden et al., 2003). Further N could possibly be available from leaching of N from aboveground tissues after a rain. If the crop freezes or is harvested before senescence and resorbtion is complete, however, more N will remain in the feedstock, reducing fuel quality and increasing fertilizer demands for subsequent crops (Lewandowski & Heinz, 2003).

Two leading biomass crops, switchgrass (Panicum virgatum L.) and Giant Miscanthus (Miscanthus×giganteus Greef et. Deu. ex. Hodkinson et Renvoize) (Hodkinson & Renvoize, 2001), hereafter called M. ×giganteus, have shown great promise to produce high yields of biomass feedstock. Both of these rhizomatous, C4, perennial grasses have been studied for decades, but there is no consensus on N management for either species. In part, this lack of resolution is due to the interaction of harvest time with N management.

Both M. ×giganteus and P. virgatum have relatively lower N requirements than food crops, though their response to N fertilizer is not consistent (Parrish & Fike, 2005; Miguez et al., 2008). In some conditions, M. ×giganteus responds strongly to N fertilizer (Cosentino et al., 2007) but in other cases biomass yield has not shown a clear statistical response to N, despite nutrient removal rates that would appear to necessitate N addition (Lewandowski & Schmidt, 2006; Christian et al., 2008). P. virgatum can respond strongly to N fertilizer by producing more leaf area and more harvestable biomass (Sanderson & Reed, 2000; Ma et al., 2001), but the response is not consistent, particularly in a single harvest system after frost (Parrish & Wolf, 1992; Christian et al., 2001; Parrish & Fike, 2005).

Delaying harvest overwinter to spring has been recommended in both crops to reduce concentrations of combustion contaminants like chlorine in addition to N (Lewandowski & Heinz, 2003; Adler et al., 2006). However, there is a trade-off between fuel quality and fuel quantity, as the amount of harvestable dry matter declines with time left in the field after senescence (Lewandowski et al., 2003b), and overwinter losses of M. ×giganteus appear greater in the US Midwest (Heaton et al., 2008) than in Western Europe (Beale & Long, 1995a).

The balance between biomass loss and fuel quality is not yet clear. What is clear is that N fertilization represents major economic and environmental costs in modern agriculture. Can its application be minimized in biomass production by manipulating harvest time? The goal of the present study was to examine intra-seasonal N dynamics in M. ×giganteus and P. virgatum to identify the optimum harvest time that balances the benefits of N translocation at senescence with the risk of yield loss that comes with delaying harvest over the winter.

This study used the experimental framework of M. ×giganteus and P. virgatum field trials at three locations in Illinois described in Heaton et al. (2008). The following research questions were addressed:

  • 1Does the [N] in aboveground biomass differ between M. ×giganteus and P. virgatum over the annual cropping cycle?
  • 2How much N is removed by crops of M. ×giganteus and P. virgatum harvested at different times?
  • 3Does the ratio of C to N (C/N) vary with time of year?
  • 4Is the partitioning of N between different shoot components the same in M. ×giganteus and P. virgatum?

Materials and methods

Details of planting stock, establishment and field maintenance are described in detail in Heaton et al. (2008). Field trials of M. ×giganteus and P. virgatum were established in May and June of 2002 at three University of Illinois Agricultural Research and Education Centers located in Northern, Central and Southern Illinois, spanning an approximately 5° range in latitude from 37 to 42°N. This land had been previously been planted to rotations of maize (Zea mays L.), soybean (Glycine max L. Merr.) and wheat (Triticum aestivum L.). Four 10 × 10 m plots of each species were arranged in a completely randomized design at each location with 1 m alleys between plots.

Clones of M. ×giganteus were propagated in the greenhouse from University of Illinois stock. At least 2 weeks before planting clones were acclimated outdoors, and then hand planted into the field at 1 plant m−2 following recommended practice (Bullard, 1996; Lewandowski et al., 2000). Clones had one to four tillers 10–50 cm tall at planting.

The P. virgatum cultivar Cave-in-Rock (Sharps Brothers Seed Company of Missouri, Clinton, MO, USA) is native to Illinois (USDA/ARS, 2009) and a locally adapted ecotype recommended for biomass production in the Midwest (Teel et al., 1997). Seed was broadcast at a rate of >13 kg of pure live seed ha−1 into a clean, packed seed bed following Teel et al. (1997). Plots were over seeded at the same rate during the early spring of 2003 to further ensure adequate populations.

Irrigation was used only at establishment on both crops in 2002. Weeds were controlled with a combination of herbicides and hand weeding in 2002 and 2003 and were not problematic thereafter. Plots were broadcast fertilized in 2004 with 25 kg N ha−1 (Scotts Turf Builder Lawn Fertilizer with 2% Iron, The Scotts Company, Marysville, OH, USA) and cleared by mowing each spring before emergence.

Crop stands were 3 years old when this experiment began, an age when they are generally considered mature (Lewandowski et al., 2000; Parrish & Fike, 2005). M.×giganteus had established well at all locations but stands of P. virgatum were relatively thinner at the Northern and Southern location. Although stand density was not recorded at these locations, estimations of plant population were still above the threshold guidelines of the one to two plants 0.09 m−2 needed for successful establishment (Parrish et al., 2008). At the Central location P. virgatum had 500–800 tillers m−2 when measured in 2003 (Heaton et al., 2008).

Biomass sampling

Plant biomass was measured at five dates spread across the annual crop production cycle during the 2004 and 2005 seasons using procedures modified from Beale & Long (1995b) and Roberts et al. (1993). Samples were taken at three times during the growing season in June, August and October, and two after the growing season in December and February (Table 1). Samples were not taken from within 1 m of the plot border to minimize edge effects and potential shading from adjacent plots.

Table 1.   Dates on which P. virgatum and M. ×giganteus biomass was sampled for determination of yield, [C], and [N] at three locations in Illinois
NorthernCentralSouthern
M. ×giganteusP. virgatumM. ×giganteusP. virgatumM. ×giganteusP. virgatum
2004
 June 23June 23June 21June 21June 17June 17
 Aug. 10Aug. 10Aug. 8Aug. 8Aug. 12Aug. 12
 Oct. 5Oct. 5Oct. 6Oct. 6Oct. 7Oct. 7
 Dec. 14Dec. 14Dec. 15Dec. 15Dec. 9Dec. 9
 Feb. 10Feb. 10Feb. 8Feb. 8Feb. 7Feb. 7
2005
 June 7June 7June 8June 8June 3June 3
 Aug. 17Aug. 17Aug. 21Aug. 21Aug. 15Aug. 15
 Oct. 23Oct. 23Oct. 30Nov. 2Oct. 24Oct. 24
 Dec. 7Dec. 7Dec. 5Dec. 5Dec. 6Dec. 6
 Feb. 15Feb. 15Feb. 16Feb. 16Feb. 21Feb. 21

In this experiment, a 0.2 m−2, rectangular quadrat was used to sample biomass from two predetermined and randomly chosen 1 m2 quadrats within each plot. Although small by agronomic standards, foundational rangeland experiments evaluating the relative efficiencies of quadrat size found 0.2 m−2 quadrats to be suitable or even preferred to assess biomass production in uniform grass stands (Van Dyne et al., 1963; Bonham, 1989). The broadcast planting method used for P. virgatum provided a randomly distributed stand structure, but because M. ×giganteus rows did not equally fill the inter-row space, the 0.2 m−2 quadrat was consistently placed in one corner of the 1 m−2 quadrat, capturing both row and inter-row tiller distribution. Each subsample included 10–25 M. ×giganteus tillers and 50–125 P. virgatum tillers.

Tillers were cut by hand to a ca. 5 cm stubble height and weighed fresh in bulk. Approximately 0.5 kg subsamples were separated into green leaf, dead leaf lamina, stem and flower. Leaves were assessed visually and considered dead when they had <50% green lamina area (Morgan et al., 2006). All subsamples were then dried to a constant mass at 60–70 °C and plant dry matter calculated by multiplying the dry matter content of each component by its proportion of the fresh mass.

Elemental analysis

Following measurement of dry mass, randomly selected material from each subsample was ground in a Wiley mill with a 2 mm sieve, then a further subsample was pulverized to a talcum powder consistency using a stainless-steel pulverizer (<250 μm) (Kleco Pulverizer, Kinetic Laboratory Equipment Company, Visalia, CA, USA). A 2–6 mg sample was then weighed on an analytical balance (M2P Electric Microbalance, Sartorius AG Göttingen, Germany) and encapsulated in tin foil before [C] and [N] was determined by combustion and thermal conductivity separation in a CHN analyzer (ECS 4010, Costech Analytical Technologies, Valencia, CA, USA). Five incremental mass samples of an acetanalide standard (C8H9NO, Costech Analytical Technologies) were included in each analysis batch and used to create a standard curve for calibration of the batch. Correct analysis was verified in each batch by including two samples of apple leaves with a known [C] and [N] (Standard Reference Material 1515, US Department of Commerce, National Institute of Standards and Technology, Gaithersburg, MD, USA).

Data analysis

The mean value of the two plot subsamples was used for statistical analysis to avoid pseudo-replication (n=4). Analysis of [C] and [N] was conducted using a repeated measures mixed model analysis of variance (proc mixed, SAS Institute Inc., Cary, NC, USA). Location, species and harvest time were considered fixed variables, while year was considered random. Significance was determined using the F statistic and α=0.05.

Results

Site conditions

The soil and climate conditions for the three sites used in this trial are summarized in Table 2. Generally, the annual temperatures during the course of this experiment were similar to the long-term averages at the Northern and Central sites, but below average at the Southern location. Precipitation was lower than average at the Northern and Southern location in both years, and at the Central location in 2005. The biomass yields of M. ×giganteus seen in this experiment were considered high compared with European averages (Jones & Walsh, 2001), particularly at the Central location in 2004. Yields of P. virgatum were considered average to high at the Central location and average to low at the Northern and Southern locations compared with yields of similar ecotypes in the Midwest (Lemus et al., 2002; Heaton et al., 2004), corresponding to below-normal precipitation at these locations (Table 2) (Heaton et al., 2008).

Table 2.   Location and environmental conditions of three sites used for evaluation of M. ×giganteus and P. virgatum in Illinois during 2004–2005
 NorthernCentralSouthern
Location (Lat., Long.)Northern Illinois Agronomy Research Center Shabbona, IL (88.15W, 41.85N)Crop Science Research Center, Savoy, IL (88.23W, 40.08N)Dixon Springs Agricultural Research Center, Simpson, IL (88.67W, 37.45N)
SoilFine-silty, mixed, superactive, mesic Typic Endoaquolls; formed from loess or silty material and the underlying till. Slope 0%Fine-silty, mixed, mesic Typic Endoaquoll; very deep and formed from loess and silt parent material deposited on the till and outwash plains. Slope 0%Fine-silty, mixed, active, mesic Oxyaquic Fragiudalfs; formed in loess and underlying weathered stone. Moderately permeable above the fragipan and very slowly permeable in the fragipan. Slope 0–2%
 Annual temperature (°C)
200410.011.613.6
20059.512.213.9
30 year average9.011.015.0
 Annual precipitation (cm)
  1. Mean annual climate data was collected daily by the Illinois Climate Network (Angel, 2007).

20047210293
2005478591
30 year average95104123

Nitrogen concentration [N]

In general, [N] declined steadily from a high of >1.5% in young shoot tissue in the spring to <0.5% in the late fall in both species (P<0.0001) (Fig. 1, Table 3). Overall, P. virgatum had a slightly, but significantly, greater [N] in standing biomass than M. ×giganteus (P<0.0001) (Table 3; Fig. 1). This was driven by a greater [N] in P. virgatum during the 2004 season, a difference that was less pronounced in 2005 (Fig. 1). When averaged across all locations, [N] declined from 1.17% to 0.33% in M. ×giganteus and from 1.65% to 0.52% in P. virgatum during the 2004 growing season. Averaged across locations in 2005, [N] dropped from 1.63% to 0.29% in M. ×giganteus and from 1.51% to 0.49% in P. virgatum. The seasonal progression of [N] reduction was generally consistent across all locations and years with the exception of the February 2004 sample date at the Southern location (Fig. 1).

Figure 1.

 Changes in nitrogen concentration (%) ([N]) of standing biomass of Miscanthus×giganteus (•) and Panicum virgatum (○) at three locations in Illinois over 2 years. Points represent least squared means ±1 SE, n=3–4.

Table 3.   Mixed model analysis of variance associated with the ratio of carbon to nitrogen (C/N ratio), the N concentration (%) ([N]) and the N in standing biomass (kg N ha−1) on five different sample dates in field trials of M. ×giganteus and P. virgatum at three locations in Illinois during the 2004 and 2005 growing seasons (n=3–4, α=0.05)
EffectDegrees of freedomF statisticProbability of >F
C/N ratio
 Sample date463.4<0.0001
 Site26.770.0014
 Sample date × site89.54<0.0001
 Species142.42<0.0001
 Sample date × species411.52<0.0001
 Site × species24.950.0078
 Sample date × site × species87.29<0.0001
[N] (%)
 Sample date4361.86<0.0001
 Site20.220.8060
 Sample date × site818.49<.0001
 Species123.46<.0001
 Sample date × species41.250.2888
 Site × species20.970.3789
 Sample date × site × species83.500.0008
Standing N (kg N ha−1)
 Sample date427.78<0.0001
 Site244.36<0.0001
 Sample date × site815.05<0.0001
 Species1280.88<0.0001
 Sample date × species410.71<0.0001
 Site × species216.83<0.0001
 Sample date × site × species87.32<0.0001

Standing N mass

The similarities in [N] between P. virgatum and M. ×giganteus across the three locations did not necessarily translate to similar amounts of N standing in the field at each sample date, largely because of differences in the amount of biomass produced by the two species. Measured biomass was higher in M. ×giganteus than P. virgatum and so, correspondingly, was standing N at most times of year (P<0.0001) (Fig. 2, Table 3). At the Northern and Central locations seasonal changes in standing N of both species followed a similar pattern in both years. At the Southern location winter losses of biomass were smaller and standing N was more variable. Elemental analysis samples were lost from the August 2004 sampling of P. virgatum and from the December and February samplings of M. ×giganteus from the 2005 season, and the missing data compounded existing variability at this location.

Figure 2.

 Changes in standing N mass (triangles) from Miscanthus×giganteus (closed symbols) and P. virgatum (open symbols) grown at three locations over 2 years shown with changes in total aboveground dry matter (circles) from Heaton et al. (2008). Values are least squared means ±1 SE, n=3–4.

In most cases, standing N stayed roughly constant or declined between June and October, even though biomass of both species nearly doubled over the same period (Fig. 2). Averaged across locations and years, M. ×giganteus standing N peaked in August at 220.6 kg N ha−1 and fell to 133.1 kg N ha−1 by February. An unexplained increase in [N] from 0.18% in December to 0.78% at the February sampling date was observed and led to an increased standing N value of 346.3 kg N ha−1 (Figs 1a and 2a). Increases in standing N between December and February were not shown elsewhere over the course of the experiment and do not have a clear biological explanation. Averaged across locations and years, standing N in P. virgatum also peaked in August at 63.0 kg N ha−1, decreasing to 22.3 kg N ha−1 by February.

Carbon/nitrogen ratio (C/N)

Averaged over all years, locations and sampling dates, the C/N ratio of M. ×giganteus biomass was 142.6, and P. virgatum biomass was 95.9 (P<0.0001) (Table 3). There were significant effects of the interaction of species, location and sampling date (Table 3), indicating that the C/N ratios of M. ×giganteus and P. virgatum biomass significantly differ depending on the growing location and time of year. Averaged over years and locations, the C/N ratio of harvestable biomass steadily increased in both species as the season progressed over winter (Fig. 3). Mean C/N ratios were considerably higher in M. ×giganteus biomass across all locations during the 2004 growing season than they were in 2005, ranging from 39.8 to 26.4 in June of 2004 and 2005, respectively, when [N] was highest, to 322.4 and 140.1 by February of 2004 and 2005 season as [N] dropped in the standing biomass (Figs 1a and 3). Changes in the C/N ratio of P. virgatum biomass were smaller and more consistent among locations and years (Fig. 3). Averaged over locations, the mean C/N ratio of P. virgatum was 28.8 and 30.4 in June of 2004 and 2005, respectively, and steadily increased to 112.0 and 113.3 by February.

Figure 3.

 The ratio of C to N in aboveground biomass of Miscanthus (•) and switchgrass (○) at three locations in Illinois over two growing seasons. Values are means±1 SE, n=3–4.

N partitioning in shoots

Partitioning of N between shoot components was analyzed at the Central location (Fig. 4). Nitrogen in live (green) leaf and stem comprised the majority of the standing N during the growing season for both P. virgatum and M. ×giganteus. Inflorescence accounted for more aboveground N in P. virgatum than it did in M. ×giganteus, reflecting a later flowering time in M. ×giganteus. Notable differences in leaf retention were apparent between M. ×giganteus and P. virgatum toward the end of the growing season. No live leaf remained in P. virgatum by October, though dead leaves were part of the harvestable biomass through the winter (Fig. 4b and d). Conversely, M. ×giganteus was still actively growing in October and had considerable N in live leaves, but these leaves dropped soon after the end of the growing season, contributing to greater N in crop litter and leaving less N in the standing crop (Fig. 4a and c).

Figure 4.

 Standing N mass in different plant organs from mature stands of (a) Miscanthus and (b) switchgrass harvested at different times in Urbana, IL during 2004. Values are means, n=4.

Discussion

Previous work has shown M. ×giganteus grown in northern and central Illinois loses more harvestable biomass when left to stand though the winter than was expected from experiences in Europe (Heaton et al., 2008). This result led to the recommendation that the crop might be harvested as soon as possible after senescence, provided the mineral nutrient content of feedstock was low enough to (1) meet fuel quality specifications, and (2) avoid excessive removal of nutrients from the field. This study examined N dynamics in crops of M. ×giganteus and P. virgatum in relation to biomass production to address four questions:

(1) Does the [N] in above-ground biomass differ between M. ×giganteus and P. virgatum?

On average, the [N] of P. virgatum biomass was significantly greater than that of M. ×giganteus (P<0.0001) (Table 3, Fig. 1). Both species showed a significant decline in [N] throughout the season at all locations that was most pronounced between the period of active growth in June and the onset and completion of senescence from August to October (Fig. 1). A decline in [N] is consistent with nutrient cycling in perennial rhizomatous grasses and other studies of M. ×giganteus and P. virgatum (Bransby et al., 1998; Long & Beale, 2001) and is a recognized survival strategy in clonal plants like M. ×giganteus (Suzuki & Stuefer, 1999). Although [N] continued to decline between October and December, the change was only ca. 0.1% (Fig. 1), indicating there would be little benefit to [N] by harvesting in winter over late fall. An October harvest is consistent with recommendations for biomass production of P. virgatum (Fike et al., 2006; Parrish et al., 2008), but represents a significant deviation from some European findings on M. ×giganteus that showed significant reductions in [N] from delaying harvest over winter (Lewandowski & Heinz, 2003). It may be that winters are harsher in the Midwest than in Western Europe and cold temperatures come earlier in the season, forcing senescence or killing any remaining live material earlier in the year.

How much N is too much in biofuel feedstock? Guide values for acceptable amounts of N in biomass feedstock for combustion suggest 0.6% N as an upper allowable threshold (Kauter et al., 2003). Importantly, the present study demonstrates that neither M. ×giganteus nor P. virgatum would be acceptable at that level if harvested in summer, even though biomass yields were often highest then. Typically [N] declined below 0.6% by December and coincided with some loss in harvestable yield, the loss varying with location and year. A part of this loss was likely translocation of soluble carbohydrates belowground, which may be critical to early vigor during the subsequent growing season (Beale & Long, 1997; Madakadze et al., 1999b; Clifton-Brown & Lewandowski, 2000; Monti et al., 2008a).

(2) How much N might be removed by crops of M. ×giganteus and P. virgatum harvested at different times?

Very little fertilizer (25 kg N ha−1 in 2004) was applied during the course of this study. Results clearly show any summer harvest would necessitate a large fertilizer input especially in M. ×giganteus, an August harvest of which would have removed more than 300 kg N ha−1 at the Central location in both years of this study (Fig. 4). Even though the amount of N translocated to belowground tissues was not measured here, the decline of N in the standing biomass to ca. 40 kg N ha−1 by December demonstrates how efficient M. ×giganteus may be in recycling N. Contrary to some European experience (Long & Beale, 2001; Lewandowski & Heinz, 2003) there appears to be little advantage in delaying harvest to late vs. early winter in terms of absolute N removal from M. ×giganteus (Figs 1 and 2). While P. virgatum and M. ×giganteus are both capable of extracting considerable N from soil reserves and are very efficient at recycling N (Bransby et al., 1998; Christian & Haase, 2001), N was still taken off the field in this experiment and even minimal removal will eventually necessitate that maintenance fertilizer be added in the long-term unless atmospheric deposition and free-living N fixation can keep pace (Christian et al., 2008).

(3) Does the C/N ratio vary with time of year?

The C/N ratio of both species significantly increased over the season (P<0.0001) and the difference in C/N ratios of M. ×giganteus and P. virgatum that developed over the duration of this experiment is a novel finding (Table 1, Fig. 3). Interestingly, C/N ratios in M. ×giganteus were similar to P. virgatum in the summer, but then increased more through the autumn and winter (Fig. 3). This may be because more leaves dropped from M. ×giganteus than P. virgatum (Fig. 4), thus effectively raising the C/N ratio of standing M. ×giganteus biomass since leaves typically have a lower C/N ratio than stems (Jenkins et al., 1998; Monti et al., 2008b). The high values of C/N found during the winter for both species are similar to those seen after a winter harvest in Italy (Monti et al., 2008b) and in the range typically seen in woody plants, e.g. Lambert et al. (1980). Since a high C/N ratio is frequently associated with slow decomposition in natural systems (Manzoni et al., 2008), harvesting M. ×giganteus and P. virgatum over the winter when C/N ratios are high could be a strategy to improve feedstock stability in storage.

A major unknown surrounding the suitability of M. ×giganteus for use in the Midwest is the capacity of this tropical grass to senesce and translocate soluble compounds before a killing freeze, given that previous work has typically been in milder climates, e.g., (Beale & Long, 1997; Clifton-Brown & Lewandowski, 2000). Although senescence was not measured directly, a notable finding of this study is that M. ×giganteus was able to achieve similar C/N ratios to P. virgatum by October, suggesting senescence was underway. Green leaves were still predominant in M. ×giganteus at the first killing frost with only lower canopy leaves completely yellow or brown. By contrast, P. virgatum was predominantly brown at the same time, consistent with the known ability of Cave-in-Rock to successfully translocate nutrients even at high latitudes, before being killed by frost, making it a recommended cultivar for short season areas (Madakadze et al., 1998). Concentrations of N in P. virgatum at the Illinois sites were consistent with trials in Canada (Madakadze et al., 1999a, b).

(4) Is partitioning of N within shoots the same in M.×giganteus and switchgrass?

Analysis of the proportion of N within different aboveground tissues suggests the C/N ratio is lower in P. virgatum because dead leaves with residual N remained attached to the stem over the winter (Fig. 4). Retention of leaves by biomass feedstocks has been linked to higher ash contents and residual mineral content with negative impacts on biofuel quality (Lewandowski et al., 2003a; Turn et al., 2003). We show here that the dramatic increase in the C/N ratio of M. ×giganteus by the end of winter results largely from the loss of the attached dead leaves. Importantly, the N in dropped leaves is contributed to the leaf litter and may become incorporated into the soil organic matter pool, thus potentially improving nutrient retention within M. ×giganteus crops.

This study demonstrated that harvest time critically influences N dynamics in mature stands of M. ×giganteus and P. virgatum in the temperate climate of Illinois. This is the first time intraseasonal variability in N status of these leading biomass crops has been studied in side-by-side comparisons over multiple locations with important implications for harvest time and N management. Overall, M. ×giganteus and P. virgatum showed similar patterns of seasonal changes in [N], as would be expected from two perennial rhizomatous grasses. However, when efficiency of N translocation is coupled with high biomass production in M. ×giganteus, this crop is capable of removing significantly more N from the field if harvested too early. Results indicate that delaying harvest can reduce (1) the [N] in biomass feedstock, and (2) replacement N fertilizer requirements, but the majority of reductions are realized by late autumn and the crop should be harvested before risk of weather-related losses overwinter.

Acknowledgements

This research was funded by the Illinois Council on Food and Agricultural Research (C-FAR) and the Department of Crop Sciences at The University of Illinois. We thank the Illinois Agricultural Experiment Station of the University of Illinois for providing land and facilities for these trials. We also thank Mike Kleiss, Robert Dunker, Steve Ebelhar, Carl Hart, Lyle Paul, Joseph Crawford, Rebecca Arundale, Janel Woods, Allison Luzader, Meagan Wells, Katie Ciccodiccola, and especially Emily Doherty for their help in trial maintenance and sample preparation; and Dr. Jim Darling for providing elemental analysis equipment.

Ancillary