Bioresource Technology 100 (2009) 3526–3531

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/ locate/bior tech

Mineral concentration in selected native temperate grasses with potentialuse as biofuel feedstock

H.M. El-Nashaar a,*, S.M. Griffith a, J.J. Steiner b, G.M. Banowetz a

a United States Department of Agriculture, Agriculture Research Service, National Forage Seed Production Research Center, 3450 SW Campus Way, Corvallis, OR 97331, United Statesb USDA-ARS, National Program Staff, Beltsville, MD 20705, United States

a r t i c l e i n f o

Article history:Received 2 July 2007Received in revised form 13 January 2009Accepted 9 February 2009Available online 28 March 2009

Keywords:BiofuelBiomassFeedstockSilicaThermochemical

0960-8524/$ - see front matter � 2009 Published bydoi:10.1016/j.biortech.2009.02.051

* Corresponding author. Tel.: +1 541 738 4168; faxE-mail address: elnashah@onid.orst.edu (H.M. El-N

a b s t r a c t

Stands of native grasses along roadways, in buffer strips, riparian zones and grass prairies have poten-tial utility as feedstock for bioenergy production. The sustainability of harvesting these stands is reli-ant, in part, on knowledge of the mineral concentration of the harvested grasses because removal ofmineral nutrients such as phosphorus (P) and potassium (K) can impact subsequent biomass produc-tion and ecosystem services associated with these stands. Mineral content of biomass, particularly thatof silicon (Si), chlorine (Cl), and sulfur (S) also impacts thermochemical conversion approaches thatconvert grasses into bioproducts. This study quantified Cl, S, Si, P and K in Bromus marginatus, Elymusglaucus, Poa secunda, Pseudoroegneria, Elymus lanceolatus, Elymus trachycaulus, Leymus cinereus, Leymustriticoides, and Pseudoroegneria spicata collected at three growth developmental stages from four plantintroduction stations located in the western US. Differences (P 6 0.05) in mineral concentrations wereassociated with developmental stage, species, and location. Variability was greatest in Si concentra-tions which ranged from 1847 to 28620 mg kg�1, similar to those recorded in other grasses. Variabilityin Cl and S concentrations also occurred, but at less magnitude than that of Si. Concentrations of P andK, two mineral fertilizer components, varied approximately threefold among these grasses. Differencesin mineral concentrations among these grasses were not completely dependent upon soil mineral con-tent. Long-term evaluations of available soil mineral concentrations under contrasting managementpractices are needed to quantify how local conditions impact mineral cycling, and in turn, the sustain-ability of harvesting these stands. The data presented here establish baselines for these species in loca-tions subject to contrasting environmental and microbiological conditions that affect mineral cyclingand availability.

� 2009 Published by Elsevier Ltd.

1. Introduction

Native grasses provide forage for wildlife and livestock, and rep-resent significant components of roadside, buffer strip, and Conser-vation Reserve Program (CRP) plant communities (USDA-NRCS,2006). In some cases, for instance where these grasses are har-vested along roadways or buffer strips to control weed prolifera-tion, they may provide significant quantities of biomassfeedstock for conversion into bioenergy. In the Pacific Northwestof the US, CRP contracts affecting nearly 700,000 ha will expire be-tween 2008 and 2012 (USDA-FSA, 2008). If 90% of these lands arere-enrolled in the CRP, the remaining 10% could provide approxi-mately 160,000 Mg of biomass, assuming an average straw produc-tion similar to that of the species quantified by Banowetz et al.(2008). In the Pacific Northwest of the US alone, over 6.3 Mt of cul-tivated grass and cereal straw are produced in excess of that re-quired for conservation purposes. Such amount is sufficient to

Elsevier Ltd.

: +1 541 738 4160.ashaar).

provide biomass feedstock to supply 8% of the current electricalor transportation energy consumption for the PNW region (Bano-wetz et al., 2008). The amount of potential biomass feedstock fromstands of native grasses available has not been quantified, but it islikely that native grass biomass, combined with that provided byresidues of cultivated crops would justify a renewable bioenergyproduction infrastructure for the region.

Harvesting of cultivated and native grasses removes mineralcomponents, including phosphorus (P) and potassium (K), that va-lue as plant nutrients and must be replaced to ensure sustainableproduction of new crops for utilization and for providing the eco-system services associated with stands of these grasses (Paineet al., 1996; Brinkman et al., 2005; Henningsen and Best, 2005;Roth et al., 2005). The precise quantities of P and K, along with car-bon (C), removed with harvest are unknown. Genotypic differencesin the accumulation of minerals by some grass species have beendocumented (Christian et al., 2002, 2006; Lewandowski et al.,2003), but genotype by environment interactions that impact min-eral accumulation remain poorly understood.

H.M. El-Nashaar et al. / Bioresource Technology 100 (2009) 3526–3531 3527

The mineral content of grasses also limits their utilization forbioenergy production in thermochemical conversion processes be-cause certain ‘‘anti-quality” minerals cause slagging and corrosionat operational temperatures within gasification reactors (Mileset al., 1996; McKendry, 2002; Thompson et al., 2003). Gasificationof feedstock for energy production requires combusting biomass attemperatures in excess of 750 �C to produce syngas, a mixture ofcarbon monoxide and hydrogen gases suitable for catalytic conver-sion to biopower or liquid fuels. Syngas quality and yield dependsgreatly on the temperature at which the gasifier is operated. Athigher operational temperatures where carbon conversion effi-ciency is high, feedstock rich in silicon (Si), K and other alkalisvaporize and react with other mineral components in straw toform a sticky glass-like substance referred to as slag (Jenkinset al., 1998). Slag formation and corrosive alkalis reduce the life-span and utility of the gasifier hardware that further diminishesthe economic feasibility of converting straw to energy. Althoughproof of concept of straw gasification technology scalable to anon-farm production has been demonstrated (Boateng et al.,2007), little is known about differences among grasses in their suit-ability as gasification feedstock.

The purpose of this research was to evaluate differences in chlo-rine (Cl), sulfur (S), Si, P, and K concentrations in diverse grassesthat represent typical components of stands of native grasses inthe western US and to evaluate the impact of location on mineralconcentration. The concentrations of minerals in grasses and soilscollected from four locations in the western US were quantified.Knowledge of mineral concentration in these grasses will helpevaluate the long-term sustainability of their harvest.

2. Methods

A survey of mineral concentrations of nine perennial grasseswas conducted by collecting plant samples from established standsat four USDA Natural Resources Conservation Service (NRCS) PlantMaterials Centers in the western US. Briefly, aboveground plantbiomass from species representing common members of nativegrass communities in the western US was harvested at NRCS sitesin Aberdeen, ID; Pullman, WA; Corvallis, OR; and Lockeford, CA.The species included Bromus marginatus Nees ex Steud. (mountainbrome), Elymus glaucus Buckl. (blue wildrye), Elymus lanceolatus(Scribn. & J.G. Sm.) Gould (streambank wheatgrass), Elymus trac-hycaulus (Link) Gould ex Shinners (slender wheatgrass), Leymuscinereus (Scribn. & Merr.) A. Löve (basin wildrye), Leymus triticoides(buckl.) Pilger (beardless wildrye), Poa secunda J. Presl (Sandbergbluegrass), Pseudoroegneria spicata (Pursh) A. Löve (bluebunchwheatgrass), and Pseudoroegneria spicata (Pursh) A. Löve ssp. iner-mis (Scribn. & J.G. Sm.) A. Löve (beardless wheatgrass). Biomassfrom each of these grasses was harvested at 4 cm above the soilsurface. Samples from four randomly selected 30 by 30 cm quad-rants in previously established stands represented the four repli-cates for each species. Seven of the nine species were available atAberdeen and Pullman, but not all species were available at everylocation. Plant biomass was collected at three developmentalstages: vegetative (prior to stem elongation), mid-anthesis, andseed maturity. The collected plant material was dried at 80 �C for24 h and ground, utilizing a Cyclotec 1093 Sample Mill (Tecator,Hoganas, Sweden). Soil mineral analysis was conducted on three2.54 cm diameter soil cores collected to a depth of 30 cm from eachof the four quadrants where plant samples were harvested. Soilsamples were collected in late spring of the growing season afterthe stands were fertilized. Plant biomass samples were collectedfrom the same perennial stands grown in two consecutive years.

The established stands of these perennial grasses received rela-tively little management, representative of what might be ex-

pected in native stands. Specific fertilization information for theAberdeen site, a Delco loam, is not available, although the standswere managed utilizing the approach described in Cornforthet al. (2001). No K was provided to the plots. The Lockeford site,a Ramoth sandy loam, received 18 kg ha�1 N, and 22 kg ha�1 P fer-tilizer in a split November and March application. The Pullman site,a Palouse silt loam, received 89 kg ha�1 N in September, and theCorvallis site, a Woodburn silt loam, received 66 kg ha�1 N and24 kg ha�1 S in March.

Plant and soil analysis for K, P, S, and Si were performed usingmicrowave-assisted acid digestion (EPA method 3052) using anEthos D microwave station (Milestone, Monroe, CT) and subse-quent analysis on an Inductively Coupled Plasma Optical EmissionSpectrometer (ICP-OES) (Perkin–Elmer Life and Analytical Sciences,Shelton, CT). Soil analyses were conducted on dried soil utilizingthe same digestion and ICP protocols. For Cl analysis, plant and soilsamples (25 g) were extracted with 100 ml of deionized water andshaken on a New Brunswick Scientific shaker (NBS Co., Inc., Edison,NJ, USA) for 30 min at 350 rpm. After shaking, samples were fil-tered through Whatman Qualitative No. 42 filters (Florham Park,NJ) that had been washed three times with 1% H2SO4 (v/v) anddeionized water. The filtrate was analyzed colorimetrically for Cl(QuickChem method 10-117-07-1-C) on a Lachat flow injectionautoanalyzer (Hach Co., Loveland, CO). Soil organic matter wasdetermined by loss on ignition at 500 �C for 4 h and soil pH wasdetermined on a slurry consisting of 1:2 ddH2O:soil.

Analysis of variance and correlation analyses were conductedutilizing the general linear model (GLM) and correlation (CORR)procedures provided by the Statistical Analysis System (SAS, Cary,NC). For the GLM procedures we considered the analysis for a com-pletely randomized block design, where the replications repre-sented the blocks and sample representative of each species werecollected from respective blocks at each location. Locations andyears were considered random and genotypes and developmentalstages were considered fixed. Main and interaction effect wereconsidered significant at P 6 0.05, unless otherwise stated. Meansof all minerals were evaluated for main effect and interaction ofthe locations, genotypes and developmental stages were comparedutilizing Tukey’ Studentized Rang (HSD) test (SAS). Mean differ-ences among species, stage of development, location effects, andcorrelations between the accumulation of Cl, S, Si, P and K weredetermined. The surveys of these established stands were repeatedtwice over two consecutive years. There were no significant differ-ences between the means of samples from year to year, so datawere combined for the analyses. All differences reported are signif-icant at P 6 0.05, unless otherwise stated. The Tukey’s test wasused for the multiple comparisons of the observed means.

3. Results and discussion

3.1. General soil characteristics

The soil pH and percent soil organic matter varied considerablyamong locations. Soil pH at Corvallis averaged 5.33, Pullman 5.68,Lockeford 6.40, and Aberdeen 7.96. Percent soil organic matter atCorvallis was 1.90, Lockeford 1.97, Aberdeen 2.27, and Pullman3.74.

3.2. Soil and plant mineral analyses

All soil and plant mineral concentrations were within theranges published (Marschner, 1986). The most abundant mineralin soil at all locations was Si, comprising approximately 20% ofthe total soil mass (Table 1). Corvallis and Aberdeen soils hadgreater concentrations of Cl than soil collected at Lockeford and

Table 1Mean chloride (Cl), sulfur (S), silicon (Si), phosphorous (P), and potassium (K)concentrations in soil from plant introduction stations in the western US at Aberdeen,ID, Corvallis, OR, Lockeford, CA, and Pullman, WA. Soil was sampled from 0 to 30 cmdepth. Data represent the mean of four replications. Values within columns followedby the same letter were not significantly different, P = 0.05.

Location Soil

Cl S Si P K

(mg kg�1) dry soil

Aberdeen 14.5 a 344 a 200,933 b 520 a 262 cCorvallis 15.0 a 284 b 221,000 a 157 c 276 bLockeford 8.8 b 231 c 174,367 c 306 b 102 dPullman 4.9 c 301 b 205,600 b 521 a 388 a

3528 H.M. El-Nashaar et al. / Bioresource Technology 100 (2009) 3526–3531

Pullman. Soil K concentrations were different (P 6 0.05) at all loca-tions. Ranking of soil P concentrations was identical to that of soilorganic content, greatest at Pullman and lowest at Corvallis.

The concentrations of Cl and S varied nearly tenfold among theseven grasses that were collected at both Aberdeen and Pullmansites (Table 2). By comparison, concentrations of Si, P and K variedthree to fivefold. In general, the Cl concentrations in plant tissues

Table 2The effect of developmental stage on aboveground biomass accumulation of chloride (Cl),from established perennial stands at Aberdeen, ID and Pullman, WA. Data represent the mlocation followed by the same letter were not significantly different, P = 0.05.

Species Location Development stage

Bromus marginatus Aberdeen VegetativeFloweringMaturity

Pullman VegetativeFloweringMaturity

Elymus lanceolatus ssp. lanceolatus Aberdeen VegetativeFloweringMaturity

Pullman VegetativeFloweringMaturity

Elymus trachycaulus Aberdeen VegetativeFloweringMaturity

Pullman VegetativeFloweringMaturity

Leymus cinereus Aberdeen VegetativeFloweringMaturity

Pullman VegetativeFloweringMaturity

Poa secunda Aberdeen VegetativeFloweringMaturity

Pullman VegetativeFloweringMaturity

Pseudoroegneria spicata Aberdeen VegetativeFloweringMaturity

Pullman VegetativeFloweringMaturity

Pseudoroegneria spicata ssp. inermis Aberdeen VegetativeFloweringMaturity

Pullman VegetativeFloweringMaturity

were greater in plants collected from Aberdeen where soil Cl levelswere also greater. Differences in mineral concentrations among thethree developmental stages varied with each mineral. When signif-icant differences in mineral concentrations between developmen-tal stages occurred, the greatest concentrations frequentlyoccurred at the vegetative stage and declined with plant matura-tion. Previous studies showed that nitrogen and potassium concen-trations in switchgrass generally declined with maturation(Sanderson and Wolf, 1995; Madakadze et al., 1999). The declineof mineral concentration over time has provided rationalizationfor delayed harvest in other grasses (Burvall, 1997; Christianet al., 2002, 2006; Lewandowski et al., 2003; Adler et al., 2006).All nine species were not available at each location, but analysisof a balanced dataset of seven grasses collected from Pullmanand Aberdeen demonstrated significant location, plantdevelopment, and species effects on Cl, S, Si, P and K concentrations(Table 3). There was no apparent impact of different fertilizationapproaches on the accumulation of these minerals within plant tis-sues at Pullman and Aberdeen. Although total soil P was virtuallyidentical at both locations (Table 1), plant biomass collected fromAberdeen consistently contained greater concentrations of P com-pared to those measured in Pullman. Similarly, soil K was greater at

sulfur (S), silicon (Si), phosphorus (P), and potassium (K) by selected grasses collectedean (mg kg�1) of four replications. Values within columns for a species at indicated

Cl S Si P K

5029 a 1545 a 11976 a 3535 a 30925 a4376 a 680 b 11015 a 2293 b 27325 a3907 a 859 ab 14233 b 3040 c 14950 b1271 a 1976 a 11152 a 3167 a 29450 a1032 ab 627 b 9288 a 1880 b 18950 b785 b 522 b 12215 a 1171 c 14025 b

3551 a 1629 a 15850 a 3663 a 19975 a3209 a 1174 b 15713 a 2956 b 16500 b3013 a 799 c 17483 a 2195 c 13225 b1014 a 654 a 14982 a 2084 a 18525 ab1022 a 275 b 12479 a 1674 b 21825 a854 a 628 a 12685 a 1150 b 15700 b

3791 a 1866 a 14270 a 3489 a 24475 a3450 a 1060 b 14390 a 2051 b 16800 b3220 a 771 b 16127 a 1923 b 12267 c876 a 485 a 10844 a 1263 a 15325 a727 ab 366 a 10254 a 1159 b 11600 b570 b 359 a 10416 a 1146 b 9633 b

3822 a 2567 a 8728 a 4576 a 32100 a3090 a 775 b 12392 a 1950 c 25300 b3814 a 1050 b 13970 a 2639 b 17967 c1729 a 1680 a 14613 a 2614 a 34125 a744 b 703 b 9526 b 1553 b 18250 b498 b 334 b 16580 a 1255 b 16300 b

3433 a 1814 a 9447 a 3293 a 21150 a3310 a 1295 a 13807 a 2497 a 12500 b1117 b 550 b 22730 a 2334 a 16350 ab844 a 256 a 30025 a 1278 a 14333 a857 a 493 a 24120 ab 1323 a 16033 ab677 b 283 a 18103 b 1540 a 12425 b

4606 a 2283 a 20400 a 3984 a 21925 a3276 b 1698 b 16140 a 3151 b 21300 a3005 b 1266 b 10036 b 2808 c 15625 b1164 a 876 a 21385 a 2498 a 18250 a831 b 422 a 13850 b 1515 b 13550 b778 b 378 a 13740 b 856 b 10325 c

3337 a 2127 a 20285 a 3324 a 20550 a2830 a 1274 b 15323 a 2933 a 15100 b2349 a 937 c 8845 b 2756 a 10950 c1946 a 1431 a 28620 a 1974 a 21075 a974 a 577 b 18236 b 1201 b 12600 b841 a 311 c 13759 c 857 b 10975 b

Table 3Mean squares1 and degrees of freedom (df) for analysis of variance of location, plantdevelopmental stage, genotype, and interactions with chlorine (Cl), sulfur (S), silicon(Si), phosphorus (P) and potassium (K) concentrations in Bromus marginatus, Elymusglaucus, Elymus lanceolatus ssp lanceolatus, Elymus trachycaulus, Leymus cinereus, Poasecunda, Pseudoroegneria spicata, Pseudoroegneria spicata ssp inerme collected fromestablished perennial stands at Aberdeen, ID and Pullman, WA.

Source of variation df Cl S Si P K

Location (L) 1 28243* 2726* 3307* 8139* 35625*

Plant development (PD) 2 536* 1629* 3640* 2266* 140830*

Genotypes (G) 7 191* 155* 16720* 146* 28938*

Replications (R) 3 15 1.2 2398 12 798L � PD 2 82 27 8496 1.0 288L � G 5 243* 39* 12997* 93* 4440*

L � G � PD 23 101* 37* 583* 105* 4252*

* MS significant at P = 0.05.1 Mean square values are divided by 104.

H.M. El-Nashaar et al. / Bioresource Technology 100 (2009) 3526–3531 3529

Pullman, but K concentrations were consistently greater in bio-mass collected from Aberdeen. Neither site received supplementalK. The Aberdeen site was irrigated whereas the Pullman locationwas rainfed.

Some of the minerals quantified in this study, particularly Si andCl, are major contributors to slagging and corrosion within gasifica-tion reactors during thermochemical conversion processes (Jenkinset al., 1998). The P. secunda plants grown at Pullman contained thegreatest amount of Si among the seven species collected fromAberdeen and Pullman. E. glaucus plants collected at Corvallis con-tained the least amount of Si (Table 4). The concentrations of Cl, S,Si, P and K in these plant tissues demonstrated considerable vari-ability among these species. The Corvallis E. glaucus accession(No. 9056242), originally collected from Crater Lake National Parkon the rim of the volcano at approximately 2134 m elevation, wasof particular interest. The fact that this accession accumulated rel-atively small quantities of Si (Table 4) from soil that contained thegreatest amount of total Si among the four sites (Table 1) suggestedmarked differences in the quantity of available silicon between thestudy sites or the existence of physiological mechanisms that havegreat impact on Si accumulation. Plant species in general have con-siderable variability in Si composition (Datnoff, 2005), includingconcentrations of 2139 mg kg�1 in corn stalks, 623 mg kg�1 inpeach leaves, and of 14,328 mg kg�1 in Kentucky bluegrass lawnclippings (Taber et al., 2002). Wheat, oat, rye, barley, sorghum,corn, and turfgrass contain approximately 1% Si in their biomass,whereas aquatic grasses contain up to 5% Si (Epstein, 1994,

Table 4The effect of developmental stage on aboveground biomass accumulation of chloride (Cl), striticoides, collected from established perennial stands at Aberdeen, ID, Corvallis, OR, Lockefreplications. Values within columns for a species at indicated location followed by the sam

Species Location Development stage Cl

Elymus glaucus Corvallis Vegetative 2130Flowering 1646Maturity 1351

Lockeford Vegetative 2396Flowering 2290Maturity 2312

Pullman Vegetative 1526Flowering 1250Maturity 1013

Leymus triticoides Aberdeen Vegetative 2693Flowering 4189Maturity 3975

Lockeford Vegetative 2208Flowering 2586Maturity 2830

1999). Nable et al. (1990) found differences in Si concentrationsamong barley genotypes. In most cases, Cl, S, P, and K concentra-tions were more dependent on plant developmental stage thanthat of Si (Tables 2 and 4). The range of variability in the concentra-tion of Cl in the nine species included in this study was similar tothat reported for other temperate grasses including perennial rye-grass (Lolium perenne L.) and tall fescue (Schedonorus phoenix(Scop.) Holub) grown in the Pacific Northwest (Hart et al., 2005).

Grasses that are harvested for seed and other purposes are usu-ally cut while the plants are still green. As a consequence, the har-vested tissue contains a mixture of vegetative, flowering, andmature tissues. To provide an estimate of the expected mineralcontent of a mixture of tissues representing these developmentalstages, mean values of the means of individual developmentalstages were calculated for each species (Tables 5 and 6). Withfew exceptions (Si in P. secunda, P. spicata, and P. spicata ssp. iner-me), these composite mean mineral concentrations were greater inAberdeen than at Pullman. There was no apparent relationship be-tween the composite means of plant mineral concentrations andthe amounts of total soil Si and K. There were significant correla-tions between the concentrations of Cl, S, P and K at Pullman andAberdeen (Table 7). Correlations between Si concentration and thatof the other minerals were less common at Aberdeen than atPullman.

The concentration of P within plant tissues is not known to im-pact slagging during thermochemical conversion, but along with K,represents a critical mineral nutrient. Dependent upon available Pand K in local soils, both may need to be replaced to sustain contin-ued biomass production and ecosystem services provided by thesegrasses. Soil P concentrations measured at these locations were atlevels that are commonly accepted as nutritionally adequate (Wes-tern Fertilizer Handbook, 2001).

The extent of genetic relation of the isolates collected at differ-ent locations is not known and as a consequence, the extent of var-iability in mineral concentration within these species cannot bestated. Genetic variability in mineral concentrations within a spe-cies would be required to apply a conventional plant breeding ap-proach to improve mineral concentration for use as bioenergyfeedstock. Previous studies demonstrated genotypic variability inmineral concentrations among cultivars of other grasses (Bano-wetz et al., 2009a,b; El-Nashaar et al., 2009). Much effort focusedon comparisons of Si accumulation because of the abundance ofthis mineral in most soils. Although Si can negatively impactthermoconversion of biomass to bioenergy, Si has beneficial ef-fects, including increased plant resistance to lodging and drought

ulfur (S), silicon (Si), phosphorus (P), and potassium (K) by Elymus glaucus and Leymusord, CA, and Pullman, WA in the western US Data represent the mean (mg kg-1) of four

e letter were not significantly different, P = 0.05.

S Si P K

a 1063 a 1847 a 1159 a 14825 ab 890 a 3538 b 1468 b 11400 bc 614 a 2247 b 1915 c 13200 ca 1633 a 15545 a 2257 a 25250 aa 1390 b 8715 b 2152 b 16025 ba 1007 c 10931 c 2158 b 14875 ba 1592 a 17705 a 2699 a 22850 ab 811 b 10624 a 2633 a 19575 ac 599 b 12668 a 1149 b 15325 b

a 3266 a 12020 a 3444 a 27150 aa 1321 b 16050 b 2822 b 21100 ba 1787 c 21343 c 1935 c 21900 aba 583 a 7541 a 2053 a 16700 aab 528 a 8745 ab 1724 a 17925 aa 846 a 10480 b 1716 a 15950 a

Table 5Location effect on mineral concentration in aboveground biomass of native grasses collected from established perennial stands at Aberdeen, ID and Pullman, WA in western USsites. Data represent mean (n = 12) values (mg kg�1). Values for individual species within columns followed by the same letter are not significantly different, P = 0.05.

Species Location Cl S Si P K

Bromus marginatus Aberdeen 4437 a 1028 a 12408 a 2956 a 24400 aPullman 1029 b 1042 a 10885 b 2072 b 20808 b

Elymus lanceolatus ssp. lanceolatus Aberdeen 3258 a 1201 a 16348 a 2938 a 16567 aPullman 963 b 519 b 13382 b 1636 b 18683 b

Elymus trachycaulus Aberdeen 3520 a 1232 a 14929 a 2488 a 17847 aPullman 724 b 403 b 10504 b 1189 b 12186 b

Leymus cinereus Aberdeen 3575 a 1464 a 13523 a 3055 a 25122 aPullman 990 b 906 b 11363 b 1807 b 22892 a

Poa secunda Aberdeen 2620 a 1220 a 15328 b 2708 a 16667 aPullman 793 b 344 b 24083 a 1380 b 14264 b

Pseudoroegneria spicata Aberdeen 3629 a 1749 a 15525 a 3314 a 19617 aPullman 924 b 559 b 16325 a 1623 b 14042 b

Pseudoroegneria spicata ssp. inerme Aberdeen 2839 a 1446 a 14817 b 3004 a 15533 aPullman 1253 b 773 b 20205 a 1344 b 14883 a

Table 6Location effect on mineral concentration in aboveground biomass of Elymus glaucus and Leymus triticoides native grasses collected from established perennial stands at Aberdeen,ID, Corvallis, OR, Lockeford, CA, and Pullman, WA in the western US. Data represent mean (n = 12) values (mg kg�1). Values for individual species within columns followed by thesame letter are not significantly different, P = 0.05.

Species Location Cl S Si P K

Elymus glaucus Corvallis 1709 b 856 b 2911 b 1514 b 13142 bLockeford 2332 a 1343 a 11730 a 2189 a 18717 aPullman 1263 c 1000 b 13665 a 2160 a 19250 a

Leymus triticoides Aberdeen 3619 a 2125 a 16471 a 2867 a 23383 aLockeford 2541 b 652 b 8922 b 1831 b 16858 b

Table 7Correlation analysis of chloride (Cl), sulfur (S), silicon (Si), phosphorous (P) and potassium (K) concentrations in Bromus marginatus, Elymus glaucus, Elymus lanceolatus ssplanceolatus, Elymus trachycaulus, Leymus cinereus, Poa secunda, Pseudoroegneria spicata ssp inerme, collected from established perennial stands at Aberdeen, ID, and Pullman, WA,USA. n = 96.

Aberdeen Pullman

Cl S Si P K Element Cl S Si P K

0.57* �0.08 0.52* 0.54* Cl 0.07 �0.26 0.18 0.38�0.17 0.71* 0.71* S �0.41* 0.62* 0.64*

�0.31 �0.26 Si �0.48* �0.52*

0.76* P 0.46*

* K

* r value is significantly different at P = 0.05.

3530 H.M. El-Nashaar et al. / Bioresource Technology 100 (2009) 3526–3531

(Epstein, 1994), improved disease resistance (Hamel and Jeckman,1999; Brecht et al., 2004), insect and nematode resistance (Swainand Prasad, 1988), soil nutrient availability, nutrient balance with-in the plant (e.g., N, P, Zn and Mn), photosynthesis, improvedreproductive fertility, and reduced transpiration (Datnoff et al.,2001). Fertilization with Si was effective in suppressing diseasesin a number of warm and cool season turfgrasses (Datnoff, 2005).

In contrast, Si can have negative impacts on the forage quality ofgrasses. Some pathological conditions such as increased tooth wearand reduced digestibility or palatability have been attributed to sil-ica in the diets of herbivores (Brizuela et al., 1986). Shewmakeret al. (1988) demonstrated significant genotypic differences in sil-ica concentration in a variety of western range grasses, and showedthat silica reduced digestibility but apparently had no significanteffect on grazing preferences of sheep.

The impact of straw removal as bioenergy feedstock or for otherpurposes will be associated with removal of minerals, including K,P, and other plant nutrients, from the soil. The location effectsdemonstrated in this study suggest that the impact of mineral re-moval on soil fertility and ecosystem services will be highly depen-dent upon location-specific characteristics not defined by thisstudy.

4. Conclusions

Variability in the concentration of minerals that impact thesuitability of grasses for thermochemical conversion to bioenergyexists among the nine native grass species evaluated in this study.There was a significant location effect in mineral concentrationsamong plants grown at Pullman and Aberdeen, and factors unde-fined by this study affected the quantities of minerals accumulatedby these grasses at different locations. Mineral concentration with-in plant tissues was affected by plant developmental stage and inmany cases, declined with plant maturation. Changes in mineralconcentration with plant development may impact the numberof cuttings of these grasses as well as their utilization for forageor bioenergy feedstock. Location-specific factors will determinethe impact of straw removal on soil fertility and associated ecosys-tem services that depend on plant productivity.

Acknowledgements

The authors thank Machelle Nelson and Don Streeter for techni-cal assistance. We are most grateful to Ralph Fisher, Larry and Da-

H.M. El-Nashaar et al. / Bioresource Technology 100 (2009) 3526–3531 3531

vid Gady, Don Wirth, George Pugh, and Rominger Farm for theircooperation and assistance with on-farm research. We are alsograteful to personnel at the USDA Natural Resources ConservationService Plant Materials Centers in Aberdeen, ID, Corvallis, OR, Lock-eford, CA, and Pullman, WA for their cooperative spirit and time inthe collection of soil and plant materials from their location.

References

Adler, P.R., Sanderson, M.A., Boateng, A., Weimer, P.J., Jung, H.-J.G., 2006. Biomassyield and biofuel quality of switchgrass harvested in fall or spring. Agron. J. 98,1518–1525.

Banowetz, G.M., Boateng, A.A., Steiner, J.J., Griffith, S.M., Sethi, V., El-Nashaar, H.M.,2008. Assessment of straw biomass feedstock resources in the PacificNorthwest. Biomass Bioenergy 32, 629–634.

Banowetz, G.M., Griffith, S.M., El-Nashaar, H.M., 2009a. Mineral content of grassesgrown for seed in low rainfall areas of the Pacific Northwest and analysis of ashfrom gasification of bluegrass (Poa pratensis L.) straw. Energy and Fuels.doi:10.1021/ef800490w.

Banowetz, G.M., Griffith, S.M., Steiner, J.J., El-Nashaar, H.M., 2009b. Mineralaccumulation by perennial grasses in a high-rainfall environment. Energy andFuels. doi:10.1021/ef800488.

Boateng, A.A., Banowetz, G.M., Steiner, J.J., Barton, T.F., Taylor, D.G., Hicks, K.B., El-Nashaar, H.M., Sethi, V.K., 2007. Gasification of Kentucky bluegrass (Poapratensis L.) straw in a farm-scale reactor. Biomass Bioenergy 31, 153–161.

Brecht, M.O., Datnoff, L.E., Kucharek, T.A., Nagata, R.T., 2004. Influence of silicon andchlorothalonil on the suppression of gray leaf spot and increase plant growth inSt. Augustine grass. Plant Dis. 88, 338–344.

Brinkman, T.J., Deperno, C.S., Jenks, J.A., Haroldson, B.S., Osborn, R.G., 2005.Movement of female white-tailed deer: effects of climate and intensive row-crop agriculture. J. Wildlife Manage. 69, 1099–1111.

Brizuela, M.A., Detling, J.K., Cid, M.S., 1986. Silicon concentration of grasses growingin sites with different grazing histories. Ecology 67, 1098–1101.

Burvall, J., 1997. Influence of harvest time and soil type on fuel quality in reedcanary grass (Phalaris arundinacea L.). Biomass Bioenergy 12, 149–154.

Christian, D.G., Riche, A.B., Yates, N.E., 2002. The yield and composition ofswitchgrass and coastal panic grass grown as a biofuel in Southern England.Bioresour. Technol. 83, 115–124.

Christian, D.G., Yates, N.E., Riche, A.B., 2006. The effect of harvest date on the yieldand mineral content of Phalaris arundinacea L. (reed canary grass) genotypesscreened for their potential as energy crops in southern England. J. Sci. FoodAgric. 86, 1181–1188.

Cornforth, B., St. John, L., Ogle, D., 2001. Seed production standards for conservationplants in the Intermountain west, NRCS Technical Notes, Technical Notes –Plant Materials No. 14, Boise, ID:USDA,NRCS, p. 14.

Datnoff, L.E., 2005. Silicon in the life and performance of turfgrass. Online App.Turfgrass Sci. doi:10.1094/ATS-2005-0914-01-RV.

Datnoff, L.E., Snyder, G.H., Korndorfer, G.H., 2001. Silicon in Agriculture. ElsevierScience, The Netherlands.

El-Nashaar, H.M., Banowetz, G.M., Griffith, S.M., Casler, M.D., Vogel, K.P., 2009.Genotypic variability in mineral composition of switchgrass. Bioresour. Technol.100, 1809–1814.

Epstein, E., 1994. The anomaly of silicon in plant biology. Proc. Natl. Acad. Sci. USA91, 11–17.

Epstein, E., 1999. Silicon. Ann. Rev. Plant Physiol. Plant Mol. Biol. 50, 641–664.Hamel, S.C., Jeckman, J.R., 1999. Impact of mineral silicon products on powdery

mildew in greenhouse grown turf. In: Rutgers Turfgrass Proc., vol. 31, Rutgers,NJ, p. 5.

Hart, J.M., Mellbye, M.E., Gingrich, G.A., Young III, W.C., Silberstein, T.B., Christensen,N.W., 2005. Spring potassium and chloride application for grass seedproduction in the Willamette Valley. In: Young III, W.C. (Ed.), Seed ProductionResearch at Oregon State University, USDA-ARS Cooperating.

Henningsen, J.C., Best, L.B., 2005. Grassland bird use of riparian filter strips insoutheast Iowa. J. Wildlife Manage. 69, 198–210.

Jenkins, B.M., Baxter, L.L., Miles Jr., T.R., Miles, T.R., 1998. Combustion properties ofbiomass. Fuel Process. Technol. 54, 17–46.

Lewandowski, I., Clifton-Brown, J.C., Andersson, B., Basch, G., Christian, D.G.,Jørgensen, U., Jones, M.B., Riche, A.B., Schwarz, K.U., Tayebi, K., Teixeira, F.,2003. Environment and harvest time affects the combustion qualities ofMischanthus genotypes. Agron. J. 95, 1274–1280.

Madakadze, I.C., Stewart, K., Peterson, P.R., Coulman, B.E., Smith, D.L., 1999.Switchgrass biomass and chemical composition for biofuel in eastern Canada.Agron. J. 91, 696–701.

Marschner, H., 1986. Mineral Nutrition of Higher Plants. Academic Press, New York.McKendry, P., 2002. Energy production from biomass (part 2): conversion

technologies. Bioresour. Technol. 83, 47–54.Miles, T.R., Miles Jr., T.R., Baxter, L.L., Bryers, R.W., Jenkins, B.M., 1996. Boiler

deposits from firing biomass fuels. Biomass Bioenergy 10, 125–138.Nable, R.O., Lance, R.C.M., Cartwright, B., 1990. Uptake of boron and silicon by barley

genotypes with differing susceptibilities to boron toxicity. Ann. Bot. 66, 83–90.Paine, L.K., Peerson, T.L., Undersander, D.J., Rineer, K.I.C., Bartlelt, G.A., Temple, S.A.,

Sample, D.W., Klemme, R.M., 1996. Some ecological and socio-economicconsiderations for biomass energy crop production. Biomass Bioenergy 10,231–242.

Roth, A.M., Sample, D.W., Ribic, C.A., Paine, L., Undersander, D.J., Bartelt, G.A., 2005.Grassland bird response to harvesting switchgrass as a biomass energy crop.Biomass Bioenergy 28, 490–498.

Sanderson, M.A., Wolf, D.D., 1995. Switchgrass biomass composition duringmorphological development in diverse environments. Crop Sci. 35, 1432–1438.

Shewmaker, G.E., Mayland, H.F., Rosenau, R.C., Assay, K.H., 1988. Silicon in C-3grasses: effects on forage quality and sheep preference. J. Range Man. 42, 122–127.

Swain, B.N., Prasad, J.S., 1988. Influence of silica content in the roots of rice varietieson the resistance to root-knot nematodes. Ind. J. Nematol. 18, 360–361.

Taber, H.G., Shogren, D., Lu, G., 2002. Extraction of silicon from plant tissue withdilute HCl and HF and measurement by modified inductive coupled argonplasma procedures. Commun. Soil Sci. Plant Anal. 33, 1661–1670.

Thompson, D.N., Shaw, P.G., Lacey, J.A., 2003. Post-harvest processing methods forreduction of silicon and alkali metals in wheat straw. Appl. Biochem.Biotechnol. 105, 205–218.

USDA-FSA, 2008. Conservation Reserve Program–Monthly CRP Acreage Report.<http://content.fsa.usda.gov/crpstorpt/rmepegg/MEPEGGR1.HTM>. Websiteconfirmed March 2008.

USDA-NRCS, 2006. Resource effects of biomass energy production. <http://soils.usda.gov/sqi/management/files/Biomass_Conservation_Issue_Brief.pdf>(Website verified March 2008).

Western Fertilizer Handbook, 2001. Interstate Printers and Publishers. Inc. Danville,IL.