• Open Access

Nitrogen and harvest management of Conservation Reserve Program (CRP) grassland for sustainable biomass feedstock production


Correspondence: Joseph C. Castro, tel. + 217 417 4529, fax + 217 333 5299, e-mail: castrjc@gmail.com


The Biomass Regional Feedstock Partnership has identified grasslands planted under the Conservation Reserve Program (CRP) as a potential source for herbaceous bioenergy feedstock. The goal of this project is to assess the yield potential of CRP grasslands across diverse regions. Consistent with that goal, the objective of this project was to establish yield potential and quality parameters for several different CRP grasslands, representative of different growing environments. Standard field scale agricultural practices were used as management guidelines at each location. The test locations were identified and established based on known regions containing concentrated tracts of CRP grassland and represented variable climatic parameters and production histories. Biomass production potential for CRP land dominated by either warm- or cool-season grass mixtures in each location was evaluated over the course of three growing seasons (2008, 2009, and 2010). Specifically, a mixture of warm-season perennial grasses was evaluated in North Dakota, Kansas, and Oklahoma, while a cool-season mixture was evaluated in Montana, Georgia, and Missouri. Maximum biomass yields for the three warm-season CRP sites ranged from 4.0 to 7.2 Mg ha−1 and for the three cool-season CRP sites 3.4–6.0 Mg ha−1. Our results demonstrate that CRP grassland has potential as a bioenergy feedstock resource if the appropriate management practices are followed.


The Energy Independence and Security Act (EISA) of 2007 specifies ambitious requirements for the expansion of biofuel production and use in the United States. The Renewable Fuels Standard (RFS2)—a part of the EISA of 2007—requires the production of 136 billion liters of biofuel by 2022. Within this 136 billion liter mandate, the RFS2 legislation requires over half of this fuel be derived from the production of different categories of biofuel, including cellulosic biofuel (60 GL), biomass-based diesel (3.8 GL), and other advanced biofuels (15 GL). Cellulosic biofuel is an appealing option because it is dependent on feedstock that could be sustainably supplied from dedicated perennial energy crops growing on marginal lands. This would in turn minimize direct land use changes, as well as reduce the environmental consequences caused by converting active cropland to feedstock production (Johansson & Azar, 2007; Campbell et al., 2008; Gopalakrishnan et al., 2009).

Perlack et al. (2005) proposed that up to 10 million ha of Conservation Reserve Program (CRP) grassland be dedicated to bioenergy feedstock production from which biomass production of approximately 110 million dry metric tons could be expected annually. CRP is a land retirement program that was established by the Food Security Act of 1985, which encourages farmers to convert highly erodible cropland or other environmentally sensitive land to vegetative cover. The main objectives of this program are to reduce soil erosion, enhance water supplies, and improve water quality. As of October 2011, 11.99 million ha were enrolled in the CRP program, of which 45% were composed of existing grasses and legumes, 21% was planted with improved native grasses, 10% was planted with introduced grasses and legumes, and 8% dedicated to permanent wildlife habitat (USDA-NCRS, 2011). Over the past 20 years, the CRP program has successfully maintained vegetative cover and improved soil conservation on participating farms (Gebhart et al., 1994; Gewin et al., 1999). Unfortunately, CRP land acreage continues to decrease alongside increasing commodity prices; for example, current enrollment is down 12% from the approximate 13.64 million ha enrolled in CRP in 2007.

It is increasingly being recognized that the environmental effects of bioenergy production are linked to changes in land use and/or cover (Searchinger et al., 2008). As CRP lands would require minimal land use changes for biofuel production (Walsh et al., 2003), the life cycle analysis of using these lands for biofuel production show a substantial reduction in greenhouse gas production compared to other biofuels. Therefore, these lands provide an excellent source of cellulosic feedstock without significant land-use changes, all while maintaining or improving soil carbon sequestration (Lee et al., 2007a,2007b). For example, an agricultural model predicted that switchgrass will have higher profits than conventional crops on 16.9 million ha within the United States (McLaughlin et al., 2002) and the majority of these lands will come from marginal crop production land or under CRP (McLaughlin et al., 2002; Walsh et al., 2003; Simmons et al., 2008).

Despite the potential for biomass feedstock production from land enrolled in CRP, data related to harvest and N management of this resource are limited. Harvest management of biomass is of special concern for feedstock producers because of possible sensitivity to frequent, or untimely defoliation. For example, in switchgrass, this sensitivity arises from the location of meristematic tissue and crown buds (Sanderson & Wolf, 1995; Mitchell et al., 1998; Sanderson et al., 1999). In addition to yield loses, changes in biomass quality were observed in high frequency harvest regimes of native grasses, which was likely due to reduced photosynthetic area (Cuomo et al., 1998). Mulkey et al. (2006) showed in switchgrass that the concentration of neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) increased between anthesis and killing-frost harvests, while total nitrogen (T-N) and ash decreased.

Many of these differences in feedstock quality can be attributed to the developmental stage which the switchgrass had reached when it was harvested. When harvested at anthesis, the switchgrass had reached sexual reproductive development, during which most nonmaintenance photosynthate was allocated to the developing seed. Conversely, when harvested after a killing-frost the switchgrass was dormant and most of the plant's storable resources had been translocated to below-ground rhizomes for storage. Similar responses are likely to be observed in other potential biomass feedstock production systems.

Perennial dedicated energy crops have also been shown to efficiently use nitrogen (Vogel et al., 2002; Mulkey et al., 2006, 2007); however, data related to which N management practices are most effective remains debatable. For example, Berg (1995) reported that for switchgrass, a yearly nitrogen application of up to 150 kg ha−1 resulted in an average yield increase of 15 kg DM kg−1 N applied. However, more recently Mulkey et al. (2006) reported that nitrogen fertilization rates could not exceed 56 kg ha−1 without affecting switchgrass persistence. In addition, changes in biomass quality, such as T-N have been reported under varying N-rates (Jung et al., 1990; Brejda et al., 1995; Mulkey et al., 2006). Altogether, limited information is available on bioenergy feedstock production on CRP land with field scale farm practices across the nation.

The Sun Grant/USDOE Regional Biomass Feedstock Partnership has identified grasslands planted under the CRP retirement program as one of four herbaceous sources with potential for use as a dedicated bioenergy feedstock. Accordingly, this program has performed replicated field trials on CRP land using field scale agricultural practices in order to assess the yield potential and suitability of CRP grassland as a bioenergy feedstock source, across logical regions of adaptation. The research presented here reports the programs assessment of yield potential, feedstock composition, and suitability of CRP grassland as a bioenergy feedstock source across the regions of national CRP land distribution.

Materials and methods

Six test locations were identified based on CRP grassland distribution in the United States. The established CRP stands were located at the following sites: Foster County, North Dakota (ND, 47.5°N 99.2°W); Ellis County, Kansas (KS, 38.8°N 99.4°W); Jackson County, Oklahoma, (OK, 34.7°N 99.3°W); Chouteau County, Montana (MT, 47.1°N 110°W); Boone County, Missouri (MO, 39°N 92.2°W); and Oconee County, Georgia (GA, 33.8°N 83.4°W). The map shown in Fig. 1 shows the six sites. The selected soil chemical properties in the top 15 cm for each location are shown in Table 1. The yearly and long-term average monthly temperature (°C) and precipitation (mm) at each location are shown in Tables 2 and 3, respectively.

Figure 1.

US Map of 2008 Conservation Reserve Program (CRP) enrollment in 2008 and research locations. One dot equals 405 ha. Triangles indicate sites with cool-season mixtures and circles indicate sites with warm-season mixtures.

Table 1. Location, conservation reserve program (CRP) enrollment year, initial species composition, and selected soil chemical properties and soil classification in top 15 cm of soil, for each of the six CRP research sites
StateLocationCRP establishedSoil pHSOCT-NSoil classificationPredominant species
  1. SOC, soil organic carbon; T-N, soil total nitrogen; BB, big bluestem; SW, switchgrass; SB, smooth bromegrass; SO, sideoats grama; LB, little bluestem; YS, yellow sweetclover; A, alfalfa; PW, pubescent wheatgrass; RC, red clover; TF, tall fescue; OR, orchardgrass.

    g kg−1   
ND47°N 99°W20017.820.82.90HaplobobollsBB, SW
KS38°N 99°W19887.624.31.90ArgiustollsSO, SW, LB, YS
OK34°N 99°W19986.65.90.64HaplustalfsSW, LB
MT47°N 110°W20087.422.03.22CalciborollA, PW
MO39°N 92°W20045.019.02.12EpiaqualfsRC, TF
GA33°N 83°W19866.2 10.40.81Kanhapludul TF, OR
Table 2. Mean monthly temperature data (°C) for each of the six CRP research sites during the study years
  1. n/a, data not available.

30 years−14.0−9.0−−3.0−11.0
30 years−−0.5
30 years2.96.110.716.
30 years−5.9−4.1−1.04.910.114.418.818.−4.0
30 years−
30 years5.77.811.916.120.624.626.625.822.616.611.57.1
Table 3. Mean monthly precipitation data (mm) for each of the six CRP research sites during the research years
  1. n/a, data not available.

30 years13.
30 years13.716.350.355.479.867.195.574.441.135.631.016.5576.6
30 years24.029.844.560.8120.3108.350.070.885.867.337.530.3729.0
30 years14.011.418.330.264.479.642.940.635.423.114.213.8388.0
30 years44.056.082.0106.0124.0102.
30 years119.1111.5126.785.198.0100.1112.096.089.788.194.294.21214.9

The predominate species varied among the six locations; C4 grasses at the ND, KS, and OK sites, C3 grasses at the MT, MO, and GA sites (Table 1). In addition to the C3 grasses, alfalfa was also a predominate species at the MT site. All locations had been managed in accordance with CRP regulations, including no N fertilization and/or aboveground biomass harvest since the start of the contract. All field sites were selected in spring 2008 and mowed at a 10–15 cm height in the spring before imposing fertilization treatments.

The experimental design was a factorial arrangement of three N rates and two harvest dates within a randomized complete block with three replicates at each location. The plot size for treatments was approximately 0.5 ha. Urea nitrogen fertilizer was annually broadcasted with the rates of 0, 56, and 112 kg N ha−1 onto each plot using a farm-scale fertilizer spreader on the dates shown in Table 4. No other fertilizer was applied as a treatment, however phosphorus was added to some of the sites early in the experiment to increase deficient levels. The fertilizer spreader was calibrated for the rate of 56 kg ha−1 and applied to that respective treatment, while for the 112 kg N ha−1 treatment plots, the spreader went over the plots twice, each at 56 kg ha−1. Fertilizer application at the OK site occurred later than at the other sites; however, this is common practice for this region, where a spring drought will often require postponed fertilizer application.

Table 4. Harvest and fertilizer application dates for each CRP research site
  1. n/a, not applicable.

  2. PSC, peak standing crop: date at which the predominant species reached anthesis; KF/EGS, date which killing frost (KF) occurred at the sites with warm-season mixtures or the end of the growing season (EGS) at the cool-season sites.

  3. a

    MO: both early and late harvest treatments at PSC were harvested again at EGS. Late harvest treatment in spring was indicated as EGS.

  4. b

    GA: EGS treatment had two cuts at both PSC and EGS.

ND7 Sep3 Sep24 Aug31 Oct23 Oct23 Oct2 Jun15 Jun21 May
KS29 Aug13 Aug22 Jul31 Oct21 Oct5 Nov1 Jul24 Mar21 Apr
OKn/a9 Sep7 Sep27 Oct4 Dec11 Nov15 Jul24 Jun28 Jun
MT8 Jul26 Jun29 Jun2 Oct23 Oct28 Oct20 Apr13 May24 Apr
MOa 16 May/15 Jul18 May/21 Jun27 May/23 Jun9 Oct18 Oct1 Nov13 Mar16 Mar19 Apr
GAb 8 Jun27 Apr24 May3 Oct20 Oct4 Octn/a16 Apr6 Apr

To monitor the effect of agronomic practices on changes in species composition, species composition at each site was estimated annually, during June and July according to the dry-weight–rank procedures described by Gillen & Smith (1986).

Biomass yield was determined from a whole plot harvest with a farm-scale harvester at a cutting height of 10–15 cm on the dates listed in Table 4. For warm-season CRP sites, biomass was annually harvested either at the anthesis (peak standing crop, PSC) or after a killing frost (KF). For cool-season CRP sites, biomass was annually harvested either at the anthesis (peak standing crop, PSC) and/or at the end of growing season (EGS) depending on location. PSC harvest timing was determined at each location by the predominant species (as listed in Table 1) reaching anthesis. The MT site had a single cut system at either PSC or EGS. For the GA site, biomass at PSC treatment was harvested only in the spring; however, EGS treatment had two cuts and was actually a combined mass of both spring (PSC) and fall (EGS) harvests. For the MO site, biomass in both treatments was harvested in early and late in spring, respectively and both treatments were harvested again at the end of year, and early harvest was considered as PSC and late harvest was considered as EGS for data presentation. The detailed harvest timing (HT) and frequency information is described in Table 4. Above ground biomass for each plot was baled with a large round baler, weighed, and then subsampled. Subsamples were collected from the bales using a core sampler (5 cm diameter and 50 cm long) attached to an electric drill in order to determine dry matter concentration and feedstock chemical composition. Subsamples were dried at 60 °C for 48 h in a forced-air oven and ground for quality analysis in a Wiley mill (Thomas-Wiley Mill Co., Philadelphia, PA, USA) to pass a 1 mm screen and then reground to pass a 1 mm screen in a cyclone mill.

Biomass feedstock chemical composition including T-N, neutral detergent fiber (NDF), acid detergent fiber (ADF), acid detergent lignin (ADL), and ash were determined for samples collected in 2009 and 2010 using near infrared reflectance (NIR) spectroscopy. Data was unavailable in MT for both years and in OK and ND in 2010. The samples were evaluated using a Foss model 5000 scanning monochromator (Foss NIRSystems, Silver Spring, MD, USA) driven by Infrasoft International software (Port Matilda, PA, USA). The spectrophotometer was calibrated for each component by regression of chemically derived data against spectral data in a modified partial least squares regression model (Westerhaus et al., 2004). Total nitrogen for calibration samples was determined using a LECO FP-428 nitrogen analyzer (LECO Corp., St. Joseph, MI, USA). Acid detergent fiber, NDF, and ADL were analyzed with an ANKOM 200 Fiber Analyzer (ANKOM Technology, Fairport, NY, USA) (Anonymous, 2002, 2003a,2003b, respectively) using the manufacture's recommended procedures (http://www.ankom.com/media/documents/Method_5_ADF_4-13-11_A200,A200I.pdf). Ash was determined by measuring the loss in sample mass after placing them for 8 h in a muffle furnace set at 600 °C.

The effect of harvest timing and nitrogen fertilization as well as their interactive effects on yield, species composition and feedstock quality were analyzed using the PROC MIXED procedure in SAS (SAS Institute, Cary, NC, USA). Each site was analyzed independently due to differences in initial species composition, as well as distinct soil and weather characteristics at each site. Harvest timing and nitrogen fertilization were considered to be fixed variables, while year was considered random. The species composition data in the OK site was not analyzed. For the GA site, 2008 data was not included for N response data analysis because N was not applied in 2008. Differences were evaluated using a mixed-model analysis of variance, of which all the necessary assumptions were met. Contrasts were run to test differences between treatments. All comparisons were declared significant with an experiment-wise Type I error rate of α = 0.05, with the exception of feedstock quality, in which α = 0.10. A greater Type I error rate was used for feedstock quality in order to avoid making a Type II error, as these measurements had a proportionately higher variance. All comparisons were made using a Tukey adjustment, within each response variable.


Biomass yield affected by N fertility

Nitrogen fertilization had significant effects on biomass yield for all sites except MT (Table 5; Fig. 2). Maximum biomass yield typically occurred when 112 kg ha−1 of N was applied; yields were 3.5, 3.4, and 2.2 Mg ha−1 for the warm-season sites in OK, ND, and KS, respectively, and 6.2, 3.1, and 4.3 Mg ha−1 for the cool-season sites in MO, MT, and GA, respectively. Of all the cool-season grass sites, MT was the only location to show no yield increase with N fertilization, which is likely due to the grass/alfalfa mixture. Alfalfa can biologically fix N from atmosphere. The application of N fertilizer increased yield significantly at all three of the warm-season locations, with the greatest responses in OK and ND.

Figure 2.

The effect of nitrogen rate on the yield of warm- and cool-season grasses at the Conservation Reserve Program (CRP) research sites. Biomass yields were averaged across years and harvest timings.

Table 5. Analysis of variance (ANOVA) and probability values for yield, for 3 years at each of the CRP research sites
Source of variationMOKSNDOKMTGA
  1. HT, harvest timing; ns, non significant.

Year × N-ratensnsns0.0028nsns
Year × HTns<.00010.03030.0017<.00010.0073
N-rate × HTnsnsns0.037nsns
Year × N-rate × HTnsnsnsnsnsns

Biomass yield affected by harvest timing

A significant interaction between HT and year was present at all locations except MO (Table 5). The effects of HT on yield, at each of the locations, during the 3 years of the study are shown in Fig. 3. No consistent patterns of HT effects on biomass yields were observed across locations and years. The locations which had warm-season mixtures (ND, KS, and OK) had a variable response to HT over the 3 years of the study. In 2009, harvesting at EGS tended to produce greater biomass in all sites, but there is no consistent response to harvest timing in any other year at these locations. The OK location was established late during the first year of the study and thus, data was not available for a measurement of PSC harvest yield at this location. The locations with cool-season mixtures (MT, MO, and GA) had more consistent responses to HT than the warm-season locations. In MT, yields were equivalent when harvested at either EGS or PSC in both 2008 and 2010. At the same site in 2009, yield was about 60% greater when harvested at PSC rather than at EGS. In GA, yields were greater in both 2008 and 2009, by 160% and 43%, respectively, when harvested at EGS rather than at PSC. Yield was slightly greater at PSC in 2010; however, this difference was insignificant. When considered as a group, yield from the cool-season sites were greater when carried out at EGS; this trend was consistent across all 3 years, except in MT in 2009.

Figure 3.

The effect of harvest timing on yield of warm- and cool-season grasses at the Conservation Reserve Program (CRP) research sites. Biomass yields were averaged across N rates. Asterisks indicate significant difference between harvest treatments for a given location with P < 0.05.δData not available for OK in 2009 due to late establishment date.

Species composition affected by harvest timing and N application

The species composition results of the sites with predominant cool-season grass composition are shown in Table 6. In MT, alfalfa and pubescent wheatgrass competition were both significantly affected by both HT and N application. When harvested at EGS instead of PSC, the composition favored pubescent wheatgrass rather than alfalfa. The losses in alfalfa were more than compensated by the increase in pubescent wheatgrass yields. The effect of added N on species composition in MT was not observed until the third year of the study, when added N favored pubescent wheatgrass significantly, while decreasing the proportion of alfalfa. In MO, there were no significant differences in species composition changes observed between PSC and EGS; however, harvesting at EGS favored tall fescue, rather than yellow sweetclover or white clover. Nitrogen application significantly increased tall fescue during all 3 years in MO; however, adding more than 56 kg ha−1 only resulted in greater composition during the first year. In GA, species composition was not affected by HT or N application.

Table 6. Species composition (%) of sites with predominant cool-season grass mixtures
StateSpeciesYearHarvest timingN-rate (kg ha−1)
  1. Asterisk and letters indicate significant differences between treatments during a given year and location (harvest timing or N-rate) at α = 0.05.

  2. PSC, peak standing crop; EGS, end of growing season.

Pubescent wheatgrass200858.081.1*
MORed clover200818.915.824.3a18.0b9.7c
Yellow sweetclover20084.
White clover20085.
Tall fescue200862.771.153.7a66.8b80.2c
GATall fescue200847.150.646.949.350.4
Orchard grass2008

The species composition results of the sites with predominant warm-season grass composition are shown in Table 7. There was a limited response to both HT and N application on species composition in ND and KS sites. Only in KS during 2010 was there a significant effect of N. At that location in 2010, there was a significant reduction in yellow sweetclover observed as N rate increased.

Table 7. Species composition (%) of sites with predominant warm-season grass mixtures
StateSpeciesYearHarvest timingN-rate (kg ha−1)
  1. Letters indicate significant differences between treatments (harvest timing or N-rate) at α = 0.05.

  2. PSC, peak standing crop; KF, killing frost.

NDBig bluestem200829.233.435.833.424.7
Smooth bromegrass20086.
KSSideoats grama200820.122.120.918.224.2
Little bluestem200819.419.822.818.717.3
Yellow sweetclover200827.219.823.423.923.1

Feedstock quality affected by harvest timing and N application

A significant interaction between year and both HT and N rate occurred for all feedstock quality variables, except with T-N and NDF, in MO. These results are shown in Tables 8 and 9, respectively. Ash contents were inconsistent with harvest timing and N application across the locations during the study period. Even though no consistent patterns were observed for harvest timing, ash content of cool-season grasses had lower variability (60–72 g kg−1) than warm season grasses (52–92 g kg−1) across both years and N rates. In most years, adding fertilizer decreased ash content, with the exception of GA and KS in 2010, when adding nitrogen fertilizer resulted in increased ash content.

Table 8. Effect of harvest timing (HT) on biomass feedstock quality (g kg−1)
  1. Different letters indicate significant differences within each CRP research site for each component (α = 0.10).

  2. T-N, total nitrogen; NDF, neutral detergent fiber; ADF, acid detergent fiber; ADL, acid detergent lignin; PSC, peak standing crop; KF/EGS, killing frost (KF) or end of growing season (EGS).

Table 9. Effect of N-Rate on biomass feedstock quality (g kg−1)
ComponentN-rate (kg ha−1)MOGAKSNDOK
  1. Different letters indicate significant differences within each CRP research site for each component (α = 0.10).

  2. T-N, total nitrogen; NDF, neutral detergent fiber; ADF, acid detergent fiber; ADL, acid detergent lignin.

Ash072.3a66.8ab72.2a/b62.0c88.9a66.9 d101.356.0
NDF0613.6630.4705.3a632.4b645.5 d722.0a671.8703.4
ADF0391.5a387.6a395.5a377.3a/c361.6 d487.1a362.8435.7
56369.3b380.3a/b389.5a360.0b366.3 d445.6b374.8425.5
112364.7b374.9a/b379.6a/b359.0c376.9 d409.0c380.2418.7

Total nitrogen (T-N) content was greater when harvested at PSC compared to at EGS, with the exception of MO and GA in 2010. The effect of HT on T-N was inconsistent for the cool-season CRP locations; for warm-season locations however, a delay in harvest consistently resulted in a lower T-N. The response of T-N content to nitrogen fertilizer application was variable depending on site. There was no response in MO, while T-N increased as N-rate increased in GA and OK. In KS, however, T-N decreased as N-rate increased.

The structural carbohydrate composition of feedstock determined by NDF, ADF, and ADL increased with delayed HT (KF/EGS). However, this trend had some inconsistencies. For example, in 2009 and 2010 in GA, as well as in 2010 in KS, NDF decreased with delayed HT. Nitrogen application did not have a significant effect on NDF at any of the sites, except KS, where NDF increased with N rate in 2009, but decreased with N rate in 2010. At most of the sites, in most years, adding fertilizer decreased the concentration of ADF and ADL.


Biomass yield responses to N application were site specific and species dependent. At sites which were composed of no more than 10% legume species, biomass yield responded favorably to N application, while also maintaining the predominant species over time. However, at sites where legume species were more prevalent, responses to N application were limited. For example, in MT, which was initially composed of 40% alfalfa, yield did not respond to N application until the third year of study. In addition, at this site, N application increased pubescent wheatgrass composition and decreased alfalfa composition. These results agree with those of Mallarino & Wedin (1990), which demonstrated that N application markedly reduced the legume species composition. Similar results were observed at the KS site, which initially had a significant proportion (23%) of yellow sweetclover. However, yellow sweetclover declined with fertilization and composition moved in favor of sideoats grama, rather than yellow sweetclover. Similarly, legume composition at the MO site decreased with N application.

Biomass had no consistent responses to harvest timing, with the exception of MO and OK, which produced higher biomass with delayed harvest. Previous reports have demonstrated that switchgrass monocultures show significant response to harvest timing (Mulkey et al., 2006), while mixed warm-season grass stands are not sensitive to harvest timing (Mulkey et al., 2007). In addition, Mulkey et al. (2007) reported that an early harvest reduced switchgrass and indiangrass yields, while big bluestem yields remained consistent when harvested at anthesis. In our study, the warm-season mixtures had a variable response to HT for both yield and species composition at all three sites, while cool-season mixtures tended to produce more biomass when harvested at EGS. Accordingly, producers who are interested in having a flexible harvesting schedule may be more interested in mixed stands rather than monocultures. One benefit of cool-season mixtures is that they have earlier biomass accumulation, which can be harvested during the early season when warm-season grasses are not ready for biomass harvest (Florine et al., 2006). Therefore, combining both warm- and cool-season grasses at a regional scale for feedstock production systems could result in a consistent supply of biomass.

Biomass feedstock quality illustrated varying results to both harvest timing and N application. Previous research has reported that delayed harvest decreases ash concentration (Lewandowski et al., 2003; Mulkey et al., 2007) and that N application increases ash concentration (Mulkey et al., 2006). Our results did not agree with these observations; however, the range of ash concentration reported here are similar to those reported in other research (Lewandowski et al., 2003; Florine et al., 2006; Mulkey et al., 2006, 2007). One possible reason for the high variation in ash content might be due to low biomass yield along with the diverse species composition, which accumulated biomass at different times of the growing season.

The concentration of T-N has been previously shown to decrease with delayed harvest timing (Clark, 1977; Griffin & Jung, 1983; Hayes, 1985; Twidwell et al., 1988). The results reported here show a similar response, as T-N was shown to decrease with delayed harvest, with the exception of GA. Previous reports have suggested a seasonal decline in T-N (Griffin & Jung, 1983; Twidwell et al., 1988) is possibly due to remobilization of N within the plant to the roots and other storage organs for winter survival (Owensby et al., 1970; Vogel et al., 2002). The concentration of T-N was also shown to respond to N-application; however, this response was site dependent. At most sites, N application resulted in greater T-N, while in KS, T-N decreased with N application.

In most instances, structural carbohydrates determined by NDF, ADF, and ADL tended to increase with delayed harvest and warm-season mixtures accumulated more structural carbohydrates than cool-season mixtures. These results agree with previous studies (Florine et al., 2006; Mulkey et al., 2006, 2007); however, our results showed a greater variability in feedstock composition associated with location and year.

Conservation Reserve Program (CRP) land is a potentially important land resource for sustainable biomass feedstock production. Based on the 2005 Billion Ton study (Perlack et al., 2005), 10.3 million ha of CRP land could annually provide about 50 million tons of dry biomass with an annual yield of 4.25 Mg ha−1, which accounts for approximately 10% of total biomass coming from agricultural land resources. Our study indicates that the maximum biomass yields for the three warm-season CRP sites ranged from 4.0 to 7.2 Mg ha−1 and for the three cool-season CRP sites 3.4–6.0 Mg ha−1. These yield ranges were similar to other studies evaluating potential biomass feedstock production in CRP lands or naturalized grass lands. For example, the 3 year study in South Dakota showed biomass yield in two warm-season grasslands managed for wildlife habitat ranged from 2.4 to 2.9 Mg ha−1 and from 2.0 to 3.6 Mg ha−1, respectively and these yields increased to 3.9–8.2 Mg ha−1 and to 2.2–4.6 Mg ha−1 with N fertilization of 224 kg N ha−1, respectively (Mulkey et al., 2007). Similarly, Florine et al. (2006) reported that the potential biomass yield in naturalized grassland in southern Iowa ranged from 0.8 to 8.2 Mg ha−1. Accordingly, adding N is essential to maximizing yields and maintaining consistency of desirable species, unless the predominant species are legumes.

The Conservation Reserve Program was originally established for soil and water conservation, not biomass production. Accordingly, in order for CRP to be a reliable source of sustainable biofuel feedstock, the appropriate management considerations must be taken into account that can produce sustainable yields. This study evaluated grasslands planted under the CRP retirement program as a herbaceous biomass source with potential use as a dedicated bioenergy feedstock. The results presented here demonstrate, using field scale agricultural practices, that CRP land is a potential resource for bioenergy feedstock production if the appropriate management practices are followed. Biomass from CRP lands could increase biomass production through renovating fields to high yielding species and/or cultivars recently developed for biomass feedstock production. Regardless of which species are grown, the proper management practices must be followed to ensure that the grass stand can provide dependable yields and maintain its conservation properties. The results presented here indicate that desirable yields are attainable with proper harvest timing and N management practices. As most CRP stands are mixtures, harvest timing is flexible; however this flexibility may be limited by species composition. Adequate N application is crucial to obtaining the yields outlined by the Billion-Ton study, especially in systems where nonleguminous species are prevalent.


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 DE-FC36-05GO85041.