<%BANNER%>

Harvest Management and Growth Period Effects on Perennial Bioenergy Grass Production and Composition in the Southeastern USA

MISSING IMAGE

Material Information

Title:
Harvest Management and Growth Period Effects on Perennial Bioenergy Grass Production and Composition in the Southeastern USA
Physical Description:
1 online resource (231 p.)
Language:
english
Creator:
Na, Chaein
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Agronomy
Committee Chair:
Sollenberger, Lynn E
Committee Co-Chair:
Erickson, John E
Committee Members:
Bennett, Jerry M
Vendramini, Joao Mauricio Bueno
Silveira, Maria Lucia

Subjects

Subjects / Keywords:
bioenergy -- biofuel -- elephantgrass -- energycane
Agronomy -- Dissertations, Academic -- UF
Genre:
Agronomy thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
In the southeastern USA, warm-season grasses elephantgrass (Pennisetum purpureum Schum.) and energycane (Saccharum spp. hybrid) are recognized for their biomass production and therefore are candidates for cellulosic biofuel crops. Targeted to seasonality of biomass production, the research objectives were: 1) determine the effect of harvest management (multiple vs. single and fall vs. winter) of two grasses (two elephantgrasses, ‘Merkeron’ and a breeding line UF-1, and energycane ‘L79-1002’) on biomass yield and composition; 2) assess their morphological and chemical changes throughout the growing season; and 3) investigate effects of delayed harvest after a killing freeze. Delaying a single harvest from fall to winter or harvesting twice per year increased the effective harvest period of biomass for the refinery. However, multiple harvests per year compromised long-term biomass production. Energycane biomass decreased 41% from Year 2 to 3 due to damage from sugarcane smut (Sporisorium scitamineum). Delaying harvest from fall to winter reduced the leaf percentage and increased the biomass dry matter concentration. Harvest frequency affected compositional quality of grass biomass. A single fall harvest of elephantgrass maximized the concentration of cellulose in total biomass. In contrast, energycane increased soluble sugar concentration late in the growing season resulting in a decrease in structural carbohydrate concentration. Later harvests were associated with lesser leaf percentage in total biomass and this caused N, P, and ash to decrease for single harvests. The elephantgrasses had more favorable morphological characteristics (greater height, tiller mass, and stem proportion) for biomass production than energycane. Cell wall constituents increased until late summer and then either remained relatively constant (UF-1) or increased slightly (Merkeron). In contrast, energycane cell wall constituents decreased after they peaked in summer. Delayed harvest after a freezing event increased the duration of the biomass harvesting period with varying consequences. Elephantgrass biomass yield decreased and composition changed to a greater degree than energycane when harvest was delayed, and composition was sensitive to weather conditions after the first freeze event. Decreasing leaf proportion in elephantgrass caused N and ash concentration to decrease during the delayed harvest period. Elephantgrass UF-1 showed excellent feedstock characteristics and is a candidate for cultivar release.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Chaein Na.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Sollenberger, Lynn E.
Local:
Co-adviser: Erickson, John E.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045684:00001

MISSING IMAGE

Material Information

Title:
Harvest Management and Growth Period Effects on Perennial Bioenergy Grass Production and Composition in the Southeastern USA
Physical Description:
1 online resource (231 p.)
Language:
english
Creator:
Na, Chaein
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Agronomy
Committee Chair:
Sollenberger, Lynn E
Committee Co-Chair:
Erickson, John E
Committee Members:
Bennett, Jerry M
Vendramini, Joao Mauricio Bueno
Silveira, Maria Lucia

Subjects

Subjects / Keywords:
bioenergy -- biofuel -- elephantgrass -- energycane
Agronomy -- Dissertations, Academic -- UF
Genre:
Agronomy thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
In the southeastern USA, warm-season grasses elephantgrass (Pennisetum purpureum Schum.) and energycane (Saccharum spp. hybrid) are recognized for their biomass production and therefore are candidates for cellulosic biofuel crops. Targeted to seasonality of biomass production, the research objectives were: 1) determine the effect of harvest management (multiple vs. single and fall vs. winter) of two grasses (two elephantgrasses, ‘Merkeron’ and a breeding line UF-1, and energycane ‘L79-1002’) on biomass yield and composition; 2) assess their morphological and chemical changes throughout the growing season; and 3) investigate effects of delayed harvest after a killing freeze. Delaying a single harvest from fall to winter or harvesting twice per year increased the effective harvest period of biomass for the refinery. However, multiple harvests per year compromised long-term biomass production. Energycane biomass decreased 41% from Year 2 to 3 due to damage from sugarcane smut (Sporisorium scitamineum). Delaying harvest from fall to winter reduced the leaf percentage and increased the biomass dry matter concentration. Harvest frequency affected compositional quality of grass biomass. A single fall harvest of elephantgrass maximized the concentration of cellulose in total biomass. In contrast, energycane increased soluble sugar concentration late in the growing season resulting in a decrease in structural carbohydrate concentration. Later harvests were associated with lesser leaf percentage in total biomass and this caused N, P, and ash to decrease for single harvests. The elephantgrasses had more favorable morphological characteristics (greater height, tiller mass, and stem proportion) for biomass production than energycane. Cell wall constituents increased until late summer and then either remained relatively constant (UF-1) or increased slightly (Merkeron). In contrast, energycane cell wall constituents decreased after they peaked in summer. Delayed harvest after a freezing event increased the duration of the biomass harvesting period with varying consequences. Elephantgrass biomass yield decreased and composition changed to a greater degree than energycane when harvest was delayed, and composition was sensitive to weather conditions after the first freeze event. Decreasing leaf proportion in elephantgrass caused N and ash concentration to decrease during the delayed harvest period. Elephantgrass UF-1 showed excellent feedstock characteristics and is a candidate for cultivar release.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Chaein Na.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Sollenberger, Lynn E.
Local:
Co-adviser: Erickson, John E.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045684:00001


This item has the following downloads:


Full Text

PAGE 1

1 HARVEST MANAGEMENT AND GROWTH PERIOD EFFECTS ON PERENNIAL BIOENERGY GRASS PRODUCTION AND COMPOSITION IN THE SOUTHEASTERN USA By CHAE IN NA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Chae In Na

PAGE 3

3 To my p arents and w ife

PAGE 4

4 ACKNOWLEDGMENTS Most of all, l would like to acknowledge my advisor Dr. Lynn E. Sollenberger for his unlimited passion to guide his students and offering me an opportunity to be his student. It is h e who made me an agronomist. I would like to express gratitude to the Co chair of my committee, Dr. J ohn E. Erickson. He allowed me access to all of his laboratory equipment shared his ideas with me, and showed me different point s of view from which to approach the result s of my experiments I would also like to thank Dr. Jerry M. Bennett, Dr. Jo o Vendr amini and Dr. Maria Silveira for their assistance throughout the research, answering my questions, and for serving on my committee. I thank a superior UF employee, Dwight Thomas for his consistently excellent help at the Plant Science Research and Education Unit I thank Dr. Kenneth Woodard for providing worldclass field research management expertise and for his good friendship. I thank Dr. Sollenbergers former and current students of which I am happy to be a part incl uding Kesi Liu, Hermes H. Cuervo, Daniel R. Pereira, Miguel Castillo, Kim Mullenix Nick Krueger, and Marcelo Wallau. Also, I would like to thank Richard Fethiere, Robert Querns, and Jeffrey Fedenko for their assistance with nutrient and fiber analyses, as well as the HPLC analyses. I also thank my Korean colleagues Drs. Jae Yoon Kim and Jehyeong Jung for their friendship and assistance in my off campus life. Lastly, I am grateful to my family. My wife, Hyejin Lee, has sacrificed her ongoing car e er and liv ed at great distance from her family and friends for three years and counting, coming to Florida where she had never been before. My son Gyuwhan is my treasure. Lastly, I thank m y parents Juteak Na and Kyeongja Kim, who have supported me continuously with infinite love since I was born.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 9 LIST OF FIGURES ........................................................................................................ 12 ABSTRACT ................................................................................................................... 15 CHAPTER 1 INTRODUCTION .................................................................................................... 17 2 LITERATURE REVIEW .......................................................................................... 22 Why Are Warm Season Grasses Considered for Use a s Bioenergy Feedstocks? ........................................................................................................ 22 Physiological Characteristics of Warm Season Grasses .................................. 22 Elephantgrass .................................................................................................. 23 Energycane ...................................................................................................... 24 Composition of Bioenergy Grasses and Conversion to Ethanol ............................. 26 Composition of Perennial Grasses ................................................................... 26 Fiber Analysis of Biomass ................................................................................ 27 Theoretical Ethanol Potential and Yield ............................................................ 28 Ash and Mineral Elements ................................................................................ 28 Management Effects on Perennial Bioenergy Grasses .......................................... 29 Logistical Issues and Field Management in the Biofuel Industry ...................... 29 Harvest Frequency and Seasonality Affect Biomass Yield ............................... 29 Harvest Frequency and Seasonality Affect Composition of Perennial Grasses ......................................................................................................... 31 Fertilizer Management ...................................................................................... 32 Region al Adaptation ......................................................................................... 33 Nitrogen Use Efficiency and Concentration in Biomass .......................................... 33 Importance of N in Grass based Systems ........................................................ 33 Seasonal Dynamics of N in Perennial Grasses ................................................ 34 N use Efficiency by Quantification of Isotopic N (15N) ...................................... 35 Morphology of Elephantgrass and Energycane ...................................................... 36 Importance of Morphology in Bioenergy Crops ................................................ 36 Seasonal Dynamics of Perennial Grass M orphology ....................................... 37 Carbon Sequestration of Perennial Grasses ........................................................... 38 Role of Perennial Grass Cropland for C Sequestration .................................... 38 Importance of Studying Warm season Grassland C Sequestration in Florida .. 39 3 BIOMASS HARVESTED AND PLANT PART PROPORTION RESPONSES OF PERENNIAL BIOENERGY GRASSES TO HARVEST MANAGEMENT ................. 41

PAGE 6

6 Overview of Research ............................................................................................. 41 Materials and Methods ............................................................................................ 42 Experimental Site ............................................................................................. 42 Treatments and Experimental Design .............................................................. 43 Plot Establishment and Management ............................................................... 44 Response Variables ......................................................................................... 45 Statistical Analysis ............................................................................................ 46 Results and Discussion ........................................................................................... 46 Biomass Harvested .......................................................................................... 46 Leaf:stem Ratio ................................................................................................ 49 Biomass Dry Matter Concentration ................................................................... 51 Persistence ....................................................................................................... 53 Implications of Research ......................................................................................... 54 4 BIOMASS COMPOSITION RESPONSES OF PERENNIAL BIOENERGY GRASSES TO HARVEST MANAGEMENT ............................................................ 63 Overview of Research ............................................................................................. 63 Material s and Methods ............................................................................................ 65 Experimental Site ............................................................................................. 65 Treatments and Experimental Design .............................................................. 66 Plot Establishment and Management ............................................................... 67 Biomass Fiber Analysis .................................................................................... 68 Total Nitrogen, Phosphorus, and Ash ............................................................... 69 Statistical Analysis ............................................................................................ 70 Results and Discussion ........................................................................................... 71 Van Soest Fiber Analyses ................................................................................ 71 Neutral detergent fiber ............................................................................... 71 Acid detergent fiber .................................................................................... 74 Acid det ergent lignin .................................................................................. 75 Cellulose .................................................................................................... 77 Hemicellulose ............................................................................................. 78 National Renewable Energy Laboratory Procedures ........................................ 81 Extractives ................................................................................................. 81 Total soluble sugars ................................................................................... 83 Structural hexose ....................................................................................... 84 Structural pentose ...................................................................................... 86 Lignin ......................................................................................................... 87 Mineral Composition ......................................................................................... 88 Nitrogen ..................................................................................................... 88 Phosphorus ................................................................................................ 90 Ash ............................................................................................................. 91 Harvest Management by Year Interactions for Total Biomass Composition ..... 92 Entry by Year Interactions for Total Biomass Composition ............................... 93 Implications of Research ......................................................................................... 94

PAGE 7

7 5 SEASONAL CHANGES IN MORPHOLOGICAL CHARACTERISTICS OF ELEPHANTGRASS AND ENERGYCANE ............................................................ 117 Overview of Research ........................................................................................... 117 Materials and Methods .......................................................................................... 119 Experi mental Site ........................................................................................... 119 Treatments and Experimental Design ............................................................ 119 Plot Establishment and Management ............................................................. 120 Response Variables ....................................................................................... 120 Statistical Analysis .......................................................................................... 122 Results and Discussion ......................................................................................... 122 Tiller Density ................................................................................................... 122 Tiller Mass ...................................................................................................... 124 Leaf Area Index .............................................................................................. 126 Canopy Height ................................................................................................ 128 Stem Proportio n ............................................................................................. 129 Biomass Dry Matter Concentration ................................................................. 131 Implications of Research ....................................................................................... 133 6 SEASONAL CHANGES IN CHEMICAL COMPOSITION OF ELEPHANTGRASS AND ENERGYCANE ............................................................ 148 Overview of Research ........................................................................................... 148 Materials and Methods .......................................................................................... 150 Treatments and Experimental Design ............................................................ 150 Response Variables ....................................................................................... 151 Statistical Analysis .......................................................................................... 152 Result s and Discussion ......................................................................................... 153 Neutral Detergent Fiber .................................................................................. 153 Acid Detergent Fiber ...................................................................................... 154 Acid Detergent Lignin ..................................................................................... 155 Cellulose ......................................................................................................... 157 Hemicellulose ................................................................................................. 158 Nitrogen .......................................................................................................... 160 Phosphorus .................................................................................................... 163 Ash ................................................................................................................. 163 Implications of Research ....................................................................................... 166 7 TIME AFTER A FREEZE EVENT AFFECTS PERENNIAL GRASS BIOMASS HARVESTED AND CHEMICAL COMPOSITION ................................................. 179 Overview of Research ........................................................................................... 179 Materials and Methods .......................................................................................... 180 Experimental Site ........................................................................................... 180 Plot Establishment and Management until a Freeze Event ............................ 181 Treatments and Experimental Design ............................................................ 182 Response Variables ....................................................................................... 182

PAGE 8

8 Statistical Analysis .......................................................................................... 184 Results and Discussion ......................................................................................... 185 Harvested Biomass ........................................................................................ 185 Biomass Dr y Matter Concentration ................................................................. 187 Leaf Proportion ............................................................................................... 189 Fiber Analysis ................................................................................................. 190 Nitrogen .......................................................................................................... 193 Ash ................................................................................................................. 194 Implications of Research ....................................................................................... 195 8 CONCLUSIONS ................................................................................................... 210 Effect of Harvest Management on Perennial Grasses Chapters 3 and 4 ........... 210 Morphological and Chemical Changes of Perennial Grasses During the Growing Season Chapters 5 and 6 ................................................................. 212 Effects of Delaying Harvest After a Freeze Event on Biomass Harvested and Chemical Composition Chapter 7 ................................................................... 214 Implications of the Research ................................................................................. 215 Future Research Needs ........................................................................................ 216 LIST OF REFERENCES ............................................................................................. 218 BIOGRAPHICAL SKETCH .......................................................................................... 231

PAGE 9

9 LIST OF TABLES Table page 3 1 Harvest dates for 2X, 1X Nov and 1X Dec harvest management treatments in 2010, 2011, and 2012. .................................................................................... 56 3 2 Sources of variation and levels of probability ( P ) for their effects on response variables reported in Chapter 3. ......................................................................... 56 3 3 Grass entry year interaction (P < 0.001) effect on biomass harvested. ........... 57 3 4 Harvest management year interaction ( P = 0.002) effect on biomass harvested of three perennial grasses. ................................................................ 57 3 5 Grass entry year interaction ( P < 0.001) effect on biomass leaf:stem ratio. .... 58 3 6 Harvest management year interaction ( P = 0.002) effect on leaf:stem ratio of three perennial grasses. ................................................................................. 58 3 7 Grass entry harvest management interaction ( P < 0.001) effect on biomass dry matter concentration at harvest. ................................................................... 59 3 8 Harvest management year interaction ( P < 0.001) effect on biomass dry matter concentration of three perennial grasses. ................................................ 59 3 9 Harvest management x entry interaction ( P = 0.009) effect on proportion of row without viable tillers after 3 yr defoliation. .................................................... 60 4 1 Composition and characteristics of components analyzed in Chapter 4 ............. 96 4 2 Sources of variation and levels of probability ( P ) for neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) concentrations and their effects on response variables reported in Chapter 4. .. 97 4 3 Sources of variation and levels of probability ( P ) for concentration of cell wall components (cellul ose and hemicellulose) from detergent fiber analysis and their effects on response variables reported in Chapter 4. ................................. 98 4 4 Effect of grass entry x harvest management interaction on neutral detergent fiber (NDF) concentration in leaf ( P = 0.272), stem ( P < 0.001), and total biomass ( P < 0.001). .......................................................................................... 99 4 5 Effect of grass entry x harvest management interaction on acid detergent fiber (ADF) concentration in leaf ( P = 0.090), stem ( P < 0.001), and total biomass ( P < 0.001). ........................................................................................ 100

PAGE 10

10 4 6 Effect of grass entry x harvest management interaction on acid detergent lignin (ADL) concentration in leaf ( P = 0.037), stem ( P = 0.006), and total biomass ( P < 0.001). ........................................................................................ 101 4 7 Effect of grass entry x harvest management interaction on cellulose (detergent fiber analysis) concentrati on in leaf ( P = 0.195), stem ( P < 0.001), and total biomass ( P < 0.001). .......................................................................... 102 4 8 Effect of grass entry x harvest management interaction on hemicellulose (detergent fiber analysis) concentration in leaf ( P = 0.029), stem ( P = 0.009), and total biomass ( P < 0.284). .......................................................................... 103 4 9 Sources of variation and levels of probability ( P ) for concentration of nonstructural compo nents (extractives and total soluble sugars) and their effects on response variables reported in Chapter 4. ................................................... 104 4 10 Sources of variation and levels of probability ( P ) for concentration of structural components (hexose, pentose, and total lignin) and their effects on response variables reported in Chapter 4. ........................................................ 105 4 11 Effect of grass entry x harvest management interaction on extractives in leaf ( P = 0.044), stem ( P < 0.001), and total biomass ( P < 0.001). ......................... 106 4 12 Effect of grass entry x harvest management interaction on total soluble sugars concentration in leaf ( P = 0.044), stem ( P < 0.001), and total biomass ( P < 0.001). ....................................................................................................... 107 4 13 Effect of grass entry x harvest management interaction on structural hexose concentration in leaf (P = 0.562), stem (P = 0.001), and total biomass (P = 0.001). .............................................................................................................. 108 4 14 Effect of grass entry x harvest management interact ion on structural pentose concentration in leaf ( P = 0.977), stem ( P = 0.003), and total biomass ( P = 0.139). .............................................................................................................. 109 4 15 Effect of g rass entry x harvest management interaction on lignin concentration in leaf ( P = 0.069), stem ( P = 0.002), and total biomass ( P = 0.002). .............................................................................................................. 110 4 16 Sources of variation and levels of probability ( P ) for nitrogen, phosphorus, and ash concentrations and their effects on response variables reported in Chapter 4. ......................................................................................................... 111 4 17 Effect of grass entry x harvest management interaction on nitrogen concentration in leaf ( P = 0.334), stem ( P = 0.942), and total biomass ( P = 0.955). .............................................................................................................. 112

PAGE 11

11 4 18 Effect of grass entry by harvest management interaction on phosphorus concentration in leaf ( P = 0.139), stem ( P = 0.360), and total biomass ( P = 0.348). .............................................................................................................. 113 4 19 Effect of grass entry x harvest management interaction on ash concentration in leaf ( P = 0.359), stem ( P = 0.257), and total biomass ( P = 0.853). ............... 114 4 20 Effect of harvest management x year interaction on concentrat ion of acid detergent lignin (ADL) ( P = 0.002), hemicellulose ( P < 0.001), extractives ( P = 0.011), total soluble sugars ( P < 0.001), structural hexose ( P < 0.001) and pentose ( P < 0.001), N ( P = 0.005), P ( P = 0.013), and ash ( P = 0.002) in total biomass. ................................................................................................... 115 4 21 Effect of entry by year on concentration of neutral detergent fiber (NDF) ( P = 0.009), acid detergent lignin (ADL) ( P = 0.006), hemicellulose ( P < 0.001), extractives ( P < 0.016), total soluble sugars ( P = 0.001), structural hexose ( P = 0.034) and pentose ( P < 0.001), and lignin ( P = 0.002) in total biomass. ...... 116 5 1 Energycane and elephantgrass sampling dates for responses reported in Chapter 5. ......................................................................................................... 135 6 1 Sampling dates for energycane and elephantgrass for responses reported in Chapter 6. ......................................................................................................... 168 6 2 Effect of grass entry or sampling date main effects or their interaction on leaf proportion in total biomass and proportion of total N and ash in the leaf fraction in 2010. ................................................................................................ 169 6 3 Effect of grass entry or sampling date main effects or their interaction on leaf proportion in total biomass and proportion of total N and ash in the leaf fraction in 2011. ................................................................................................ 170 7 1 Sampling and freeze event dates for delayed harves t management in 20102011, 20112012, and 20122013. ................................................................... 197

PAGE 12

12 LIST OF FIGURES Figure page 3 1 Monthly average and monthly maximum and minimum air temperatures for 2010, 2011, and 2012 for the experimental location, and the 30yr average for Gainesville, Florida. ....................................................................................... 61 3 2 Monthly rainfall for 2010, 2011, and 2012 for the experimental location and the 30yr average for Gainesville, Florida. .......................................................... 62 5 1 Seasonal changes in tiller density of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). ............................................... 136 5 2 Seasonal changes in tiller density of first and ratoon growth for three perennial grass entries in 2010 (upper) and 2011 (lower). ............................... 137 5 3 Seasonal changes in tiller mass of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). ............................................... 138 5 4 Seasonal changes in tiller mass of first and ratoongrowth for three perennial grass entries in 2010 (upper) and 2011 (lower). ............................................... 139 5 5 Seasonal changes in leaf area index of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower ). ............................................... 140 5 6 Seasonal changes in leaf area index of first and ratoongrowth for three perennial grass entries in 2010 (upper) and 2011 (lower). ............................... 141 5 7 Seasonal changes in canopy height of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). ............................................... 142 5 8 Seasonal changes in canopy height of first and ratoongrowth for three perennial grass entries in 2010 (upper) and 2011 (lower). ............................... 143 5 9 Seasonal changes in stem proportion of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). ............................................... 144 5 10 Seasonal changes in stem proportion of first and ratoongrowth for three perennial grass entries in 2010 (upper) and 2011 (lower). ............................... 145 5 11 Seasonal changes in dry matter concentration of full season growth of three per ennial grass entries in 2010 (upper) and 2011 (lower). ............................... 146 5 12 Seasonal changes in dry matter concentration of first and ratoongrowth for three perennial grass entries in 2010 (upper) and 2011 (lower). ...................... 147

PAGE 13

13 6 1 Seasonal changes in NDF concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). ............................... 171 6 2 Seasonal changes in ADF concen tration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). ............................... 172 6 3 Seasonal changes in ADL concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). ............................... 173 6 4 Seasonal c hanges in cellulose concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). ...................... 174 6 5 Seasonal changes in hemicellulose concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). .................. 175 6 6 Seasonal changes in nitrogen concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). ...................... 176 6 7 Seasonal changes in phosphorus concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). ...................... 177 6 8 Seasonal changes in ash concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). ............................... 178 7 1 Weekly average air temperature (Avg.) and average of weekly maximum (Max.) and minimum (Min.) air temperatures for 20102011, 20112012, and 201 2 2013 at the experimental location (Citra, FL) ........................................... 198 7 2 Weekly total rainfall for 20102011, 20112012, and 20122013 at the exper imental location (Citra, FL) ....................................................................... 199 7 3 Effect of days after a freeze event on biomass harvested of two perennial grass entries in three years. ............................................................................. 200 7 4 Effect of days after a freeze event on dry matter (DM) concentration of two perennial grass entries in three years. .............................................................. 201 7 5 Effect of days after a freeze event on leaf proportion of two perennial grass entries in 20102011 (upper) and 20112012 (lower). ...................................... 202 7 6 Effect of days after a freeze event on neutral detergent fiber (NDF) concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower). ..................................................................................................... 203 7 7 Effect of days after a freeze event on acid detergent fiber (ADF) concentration of two perennial grass entries in 2010201 1 (upper) and 20112012 (lower). ..................................................................................................... 204

PAGE 14

14 7 8 Effect of days after a freeze event on acid detergent lignin (ADL) concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower). ..................................................................................................... 205 7 9 Effect of days after a freeze event on cellulose concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower). ............................. 206 7 10 Effect of days after a freeze event on hemicellulose concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower). ............. 207 7 11 Effect of days after a freeze event on N concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower). ...................................... 208 7 12 Effec t of days after a freeze event on ash concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower). ............................. 209

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy HARVEST MANAGEMENT AND GROWTH PERIOD EFFECTS ON PERENNIAL BIOENERGY GRASS PRODUCTI ON AND COMPOSITION IN THE SOUTHEASTERN USA By ChaeIn Na August 2013 Chair: Lynn E. Sollenberger Cochair: John E. Erickson Major: Agronomy In the southeastern USA, warm season grasses elephantgrass ( Pennisetum purpureum Schum.) and energycane ( Saccharum spp. hybrid) are recognized for their biomass production and therefore are candidates for cellulosic biofuel crops Targeted to seasonality of biomass production, t he research objectives were : 1) determine the effect of harvest management (multip le vs. single and fall vs. winter) of two grasses (t wo elephantgrasses, Merkeron and a breeding line UF 1, and energycane L791002 ) on biomass yield and composition; 2) assess their morphological and chemical changes throughout the growing season; and 3) investigate effects of delayed harvest after a killing freeze D elaying a single harvest from fall to winter or harvesting twice per year increased the effective harvest period of biomass for the refinery. However, multiple harvests per year compromise d long term biomass production. Energycane biomass decreased 41% from Year 2 to 3 due to damage from sugarcane smut ( Sporisorium scitamineum ) Delay ing harvest from fall to winter reduced the leaf percentage and increased the biomass dry matter concentration. Harvest frequency affect ed

PAGE 16

16 compositional quality of grass biomass A single fall harvest of elephantgrass maximized the concentration of cellulose in total biomass. In contrast, energycane increas ed soluble sugar concentration late in the growing season resulting in a decrease in structural carbohydrate concentration. Later harvests were associated with lesser leaf percentage in total biomass and this caused N, P, and ash to decrease for single harvest s. T he elephantgrasses had more favorable morphological characteristics (greater height, tiller mass, and stem proportion) for biomass production than energycane. C ell wall constituents increased until late summer and then either remained relatively constant (UF 1) or increased slig htly (Merkeron) In contrast, energycane cell wall constituents decreased after they peaked in summer Delayed harvest after a freezing event increased the duration of the biomass harvest ing period with varying consequences Elephantgrass biomass yield dec reased and composition changed to a greater degree than energycane when harvest was delayed, and composition was sensitive to weather conditions after the first freeze event Decreasing l eaf proportion in e lephantgrass caused N and ash concentration to dec rease during the delayed harvest period. E lephantgrass UF 1 showed excellent feedstock characteristics and is a candidate for cultivar release.

PAGE 17

17 CHAPTER 1 INTRODUCTION Emissions of CO2 associated with burning fossil fuels contribute to increased atmospheric CO2 concentration (de la Mata et al., 2012) Greenhouse gas (GHG) emission from fossil fuel combustion is believed to be a major cause of climate change (Cheng, 2010) This association of fossil fuel use with climate change has stimulated research to identify other sources of energy. Moreover, environm ental consequences of continued largescale use of fossil fuels for energy provide impetus for research aimed at development of bioenergy technologies that mitigate GHG emissions and slow climate change (Wedin, 2004) The justification for bioenergy research is compelling for economic, national security, and environmental conservation reasons. From a national ener gy resource perspective, turmoil in the Middle East and large fluctuations in energy prices illustrate long term economic and security risks of dependence on imported oil. The high cost of petroleum imports appears here to stay as yearly average Brent Cr ude oil price has (U.S. Energy Information Administration, 2013) Bioenergy and its related products have potential for significant economic and environmental benefits to society including near zero net emissions of GHG, improved soil and water quality, and increased enterprise opportunities for a depressed rural economy (McLaughlin et al., 2002) Greater interest in dedicated bioenergy crops, however, has stimulated debate of food versus fuel. If demand for biofuel feedstock is such that crop land is converted to production of biofuels, there could be significant disruption of food supply and in creasing food cost (Mathews, 2008) Mathews (2008) referred to this devastating

PAGE 18

18 scenario as a crime against humanity, especially in developing countries. A single filling of a SUVs fuel tank with ethanol requires enough grain (450 pound of corn) to feed a person for a year (The Economist, 2007) I f du e to demand for biofuels farmers were to plant bioenergy feedstock s instead of cereal crops it would affect the cost of even rice ( Oryza sativa L.) and wheat ( Triticum aestivum L.) (Tenenbaum, 2008) For instance, since 2010 more than 40% of domestic corn ( Zea mays L.) use in the USA, up to one hundred sixty million Mg (five billion bushels) each year, has been diverted to fuel alcohol production (USDA Economic Research Service, 2012) The USDA Economic Research Service data also show that in 2010 domestic corn use for fuel alcohol exceeded animal feed and residual use. It is believed that converting corn to bio ethanol is one of reasons for recent Agflation, i.e., inflation in cost of agricultural commodities. To address both ene rgy and food security, fuel production from non food resources is rising. Breaking the Biological Barriers to Cellulosic Ethanol, the U.S. Department of Energy (DOE) publication, states that fuels derived from cellulosic biomass the fibrous, woody, and generally inedible parts of plant matter offer an alternative to conventional energy sources that support national economic growth, national energy security, and environmental goals (Department of Energy, 2006a) Cellulosic biomass is an attractive energy feedstock in the southeastern USA because supplies are abundant, it is renewable, and its use for energy does not add previously stored soil C to the at mosphere (Fargione et al., 2008) So called 2nd generation biofuel crops (cellulosic biomass) are thought to offer less competition with food crops for land use as well as improved energy efficiency and lower GHG emissions (Erisman et al.,

PAGE 19

19 2010) Non arable or marginal land (2 Gha worldwide) is suitable for production of 2nd generation energy crops if management is such that the soils are protected from erosion (Strezov et al., 2008) In addition, perennial energy crop cultivation has been associated with significantly improved soil h ealth and quality (Strezov et al., 2008) Florida is well positioned to capitalize on the national need for biofuel crops. Florida ranks first in the USA in annual growth of plant biomass because its climate conveys unique advantages for bioenergy crop production. Specifically, the state experiences relatively few frost events during mild winters. In addition, precipitation is reasonably well distributed throughout the year (Smith and Dowd, 1981) The development of production systems for high yielding energy crops that can be grown in Florida is considered essential for establishment of a sustainable biomass to energy industry. Transportation cost is high for biomass relative to energy dense materials like grain, so cellulosic biomass production must occur near the refinery to keep cost low (Brechbill et al., 2011) Thus, long term availability of sufficient amounts of reasonably priced biomass will be an important determinant of if and where new biofuel and bioenergy facilities will be built. Because of their C4 carbon fixation pathway and upright growth habit, warm season grasses like elephantgrass ( Pennisetum purpureum Schum.) and energycane ( Saccharum sp.) are widely recognized for their biomass production (Wooda rd and Prine, 1991; Woodard and Prine, 1993a) Research initiatives to identify herbaceous plants with the highest biomass yields for renewable energy purposes have consistently found elephantgrass and energycane to have the highest or near the highest DM

PAGE 20

20 product ion across the southeastern USA. Yields of 20 to 48 Mg ha1 yr1 have been reported in the region (Prine et al., 1984; Woodard and Prine, 1993a; Bouton, 2002; Woodard and Sollenberger, 2008) These species grow rapidly and es tablish canopy cover quickly so that less effort and fewer inputs are required to achieve appropriate weed control. Moreover, they can be propagated using internodes of aboveground stems, which requires relatively less labor than the rhizomebased vegetat ive propagation of Miscanthus ( Miscanthus giganteus ) and it also allows for propagation without genetic variation. Although they are regarded as important candidate species for bioenergy, only a few fieldbased experiments and morphological studies have been conducted on these species as bioenergy crops. Plant biomass production is affected by soil nutrient quantity and availability. High yielding bioenergy crops remove relatively large amounts of nutrients at harvest. Soil fertility of sandy soils in Florida is notoriously low, and replenishing nutrients removed at harvest of biomass crops is an important challenge to development of sustainable production systems. Nitrogen is poorly held by sandy soils in Florida and can be easily leached by rainfall (Obreza, 2003) Nutrient dynamics and particularly N dynamics have not yet been evaluated for many potential biofuel crops in Florida. It has been suggested that perennial crops have a large capacity for storage of nutrients, and harvest timing may impact N dynamics of bioenergy crops, particularly when harvest is delayed until after physiological maturity (Lewandowski and Heinz, 2003) There is currently no information available that describes these relationships for elephantgrass or energycane in Florida.

PAGE 21

21 For most candidate bioenergy grasses being evaluated in Flor ida, there is limited information on chemical composition as related to their potential for use as a feedstock. Chemical composition such as structural and nonstructural carbohydrates and lignin has a major effect on the utility of biomass for various pur poses, thus more information is needed on composition characteristics of various species and the effect of harvest management. A key to addressing the feedstock productivity and composition issues of the future is developing systems that combine the best plant species and cultivars with environmentally and ecologically sustainable production and harvest management practices. The proposed studies to be included in this dissertation will build on previous work conducted in the Southeast USA to provide additional information needed to identify such systems. More specifically, the research reported in this dissertation was conducted to address the following objectives: 1) determine the effect of harvest management of elephantgrass and energycane on biomass yield and composition when plants are managed as bioenergy feedstock; 2) assess morphological changes of elephantgrass and energycane throughout the growing season; and 3) determine the effect of delayed harvest after a freeze event on biomass harvested and com position.

PAGE 22

22 CHAPTER 2 LITERATURE REVIEW Why Are WarmSeason Grasses Considered for Use a s Bioenergy Feedstocks ? Physiological Characteristics of Warm Season Grasses Many warm season grasses have been considered for use as cellulosic biofuel feedstock due in part to their C4 photosynthetic pathway. In C4 plants, bundle sheath and mesophyll cells surround the vascular bundle and cause a wreathlike appearance (Kranz anatomy). This arrangement creates a physical barrier between initial carbon fixation in the mesophyll and the activity of the Calvin Cycle in the bundle sheath. In addition, there is a CO2concentrating mechanism associated with C4 plants that results in a large CO2concentration gradient between mesophyll cells and the binding site with Rubisco in the bundle sheath cell, resulting in C4 plants having up to two times greater water and N use efficiency than C3 plants (Jakob et al., 2009) Although this procedure requires two more ATPs to regenerate phosphoenolpyruvate (PEP) from pyruvate, the plant achieves a far greater advantage by minimizing or eliminating the occurrence of photorespiration. This allows these plants to synthesize sugars more efficiently than C3 plants, at least at current atm ospheric CO2 concentrations (380 ppm) (Langdale, 2011) Harvested biomass of C4 grasses in the US Gulf Coast region can total 20 to 40 Mg ha1 yr1, depending upon species, growing environment, and management (Woodard and Prine, 1993a; Anderson et al., 2008a) Among the perennial grasses that have been evaluated in the region, elephantgr ass [ Pennisetum purpureum (L.) Schum.] and energycane ( Saccharum spp. hybrid) both take advantage of the C4 pathway and have been among the most productive species in regional trials (Prine et al., 1984; Woodard and Prine, 1993a; Woodard and Sollenberger, 2008)

PAGE 23

23 Elephantgrass Elephantgrass or napiergrass is indigenous to equatorial Africa in areas of rainfall exceeding 1000 mm. It is in the tribe Paniceae of the Poaceae (Panicoideae) family (Hanna et al., 2004) Elephantgrass is a robust, creeping rhizomatous plant that perennates in the tropics and subtropics. It was introduced into the USA in 1913 (Thompson, 1919) The northern limit for survival and adequate yield under productiontype management is where the lowest temperature does not drop below 7 to 9C (Woodard and Sollenberger, 2008) In the USA, this is roughly a line across extreme southern Georgia, Alabama, and Mississippi, the southern half of Louisiana, and the southeastern portion of Texas (Woodard and Sollenberger, 2008) Plants produce many large tillers with 20 or more internodes ranging in length from 20 to 25 cm, a diameter of up to 3 cm, and a height of up to 7 m (Hanna et al., 2004) Because of its yield potential, it is considered a very promising grass for bioenergy. It has been reported that elephantgrass yield can reach 5 to 10 Mg ha1 yr1 when plants are unfertilized, 15 to 30 Mg ha1 yr1 in well fertilized pastures (Bogdan, 1977) and 70 to 85 Mg ha1 yr1 in environments with year round growing conditions and optimal management (VicenteChandler et al., 1959) In the USA, breeding efforts with Pennisetum spp. have occurred primarily in Georgia and Florida, resulting in numerous breeding lines and cultivars. The elephantgrass cultivar Merkeron is an F1 hybrid of a cross made in Georgia between a very leafy dwarf type and a tall selection (Burton, 1989) It is shortday sensitive and reaches a height of 4 to 5 m in autumn when it flowers. It is considered an excellent cultivar for bioenergy production. Merkeron has produced 172 Mg ha1 yr1 of green forage in Puerto Rico (Burton, 1989) In Florida field trials, DM yield o f 16 to 32 Mg ha1

PAGE 24

24 yr1 has been reported (Woodard et al., 1991b) Other elephantgrass breeding lines from the Georgia program include N13, N43, and N51. In Flor ida, evaluation of promising Pennisetum spp. plant introductions and breeding lines has occurred over several decades (Woodard and Prine, 1991; Woodard et al., 1991b; Spitaleri et al., 1994; Macoon et al., 2001; Macoon et al., 2002) During this period, S.C. Schank led breeding efforts in Florida with Pennisetum spp. including elephantgrass and its interspecific hybrids with pear l millet [ Pennisetum glaucum (L.) R. Br.] (Schank and Diz, 1991) During this time and prior to his death in 1997, Dr. Schank also developed a number of breeding lines of elephantgrass. In 2000, Dr. Gordon M. Prine chose several outstanding elephantgrass lines at the Green Acres Agronomy Farm in Gainesvill e, FL from what had been the breeding nursery of Dr. Schank. These breeding lines were established by Dr. Prine in a grass nursery at the University of Florida Plant Science Research and Education Unit (PSREU) at Citra, FL. The parentage of these lines has not been identified, and there is no evidence in the literature that they were evaluated under production. One of the breeding lines, now called UF 1, was observed in the nursery in 2006 by Dr. K.R. Woodard and Dr. L.E. Sollenberger, and because of its vi gorous appearance was selected for evaluation in subsequent experiments, including those reported in this dissertation. Currently, UF 1 is being considered for potential cultivar release. Energycane Modern sugarcane ( Saccharum spp.) is thought to originate from hybridization among four species; S officinarum S. sinense, S. barberi and S. spontaneum Cultivated sugarcane is predominately outcrossing, very heterozygous, and vegetatively

PAGE 25

25 propagated (Bischoff et al., 2008) It is a tall perennial tropical grass, up to 4 m or more, with thick stems up to 5 cm in diameter (James, 2003) At the USDA ARS Sugarcane Field Station in Canal Point, FL, inter specific crosses between sugarcane and other species are made with the goal of incorporating desirable traits such as ratooning ability, growth vigor, cold tolerance, and disease resistance from wild relatives into breeding lines of sugarcane as a part of a breeding program (Wang et al., 2008) These characteristics are associated with high biomass production, and when hy brids have a high concentration of cellulose instead of sucrose they are a potentially valuable feedstock resource for cellulosic ethanol production. This is why high cellulose Saccharums are called energycanes (Len et al., 2012) To better understand Saccharum spp. as a biofuel crop, it is useful to note that there are three distinctive types. These include sugarcane (primarily sugar, conventional sugarcane), Type I energycane (sug ar and fiber), and Type II energycane (primarily fiber). Cultivar L791002 is a Type II energycane developed by Louisiana State University with USDA ARS and American Sugar Cane League, Inc. (Tew and Cobill, 2008) The female parent was CP5268 and the male parent was Tainan (Bischoff et al., 2008) L79 1002 can be cropped in colder regions than the current sugarcane growing areas, particularly those areas where freezing of aboveground tissue is desi red in order to achieve desiccation prior to harvest. This cultivar has high fiber and biomass yield, 18.0 Mg ha1 of dry biomass from fiber and 6.6 Mg ha1 from extractives for a total of 24.6 Mg ha1, outstanding ratooning ability, and vigorous growth (Bischoff et al., 2008; Tew and Cobill, 2008)

PAGE 26

26 Composition of B ioenergy G rasses and C onversion to E thanol Composition of P erennial G rasses A unique feature of plant versus animal cells is the presence of a rigid cell wall. The primary constituents of plant cell walls are cellulose, hemicellulose, pectin, lignin, and proteins. Plant cell wall is composed of long crystalline cellulose microfibrils embedded in a matrix o f other polysaccharides (Vermerris, 2008a) Cellulose is regarded as the most abundant biopolymer on Earth. Within the fibril of cellulose, two glucose units are linked dimers of D glucoses called cellobiose (Perez et al., 2002; Vermerri s, 2008a) Hemicellulose in grasses contains glucuronoarabinoxylans (GAXs) that have a xylose backbone and are the predominant hemicellulosic polysaccharide in cell w alls of the grass family (Vermerris, 2008a) Lignin is an aromatic polymer synthesized from phenylpropanoid precursors (Perez et al., 2002) and lignin negatively affects the yield of fermentable sugars by shielding cellulose from degradation by providing a surface that cellulolytic enzymes adsorb to irreversibly (Akin, 2007; Vermerris, 2008a) Given the abundance of cellulose and hemicellulose in plant cell walls of perennial grasses, they represent a major source of carbohydrate for conversion to energy. Cellu lose and hemicellulose are polysaccharides that can be hydrolyzed to sugars and then fermented to ethanol, but lignin cannot be used in fermentation processes (Cherubini, 2010) For the conversion to ethanol, the cellulose and hemicellulose in biomass must be broken down into their correspondent monomers; hexoses and pentoses (Kumar et al., 2009) The cascade conversion procedures can be categorized as pretreatment, enzymatic hydrolysis (saccharification), and ethanol fermentation (Lu and Mosier, 2008) It is important to know compositional properties of

PAGE 27

27 lignocellulosic biomass because feedstock response to pretreatment, saccharification, and fermentation are controlled by properties of the feedstock. Monomers from cellulose and hemicellulose can be analyzed by the Laboratory Analytical Procedures (LAP) established by th e National Renewable Energy Laboratory (NREL). Those procedures quantify sugar monomers from extractives (nonstructural carbohydrates) and structural carbohydrates, and they measure acid soluble and acid insoluble lignin (Sluiter, 2008a; Sluiter, 2008b) Fiber A nalysis of B iomass Analyses of chemical composition and ethanol yield, such as those included in the NREL procedures, are very laborious and time consuming (Han et al., 2012) This is why cost effective methods are proposed for screening a large number of lines or seasonal changes in feedstock composition. Concentration of plant components resistant to acid detergent and neutral detergent have been used for decades to predict relative digestibility and intake of forages (Van Soest et al., 1991) These analyses have relevance to bioenergy research because biological degradation by the ruminant is somewhat similar to saccharification and fermentation in bioethanol production (Han et al., 2012) Lignin concentration is not able to be estimated correctly using the detergent f iber analysis because it measures only acidinsoluble lignin which usually underestimates actual lignin concentration (Jarchow et al., 2012) According to a recent study, 95% of actual ethanol yield can be explained by a regression model consisting of neutral detergent fiber ( NDF ) and NDF digestibility (NDFD) (Lorenz et al., 2009a) This research also showed that NDF alone is negatively correlated with actual ethanol yield. Cellulose ( acid detergent fiber [ ADF ] minus acid detergent lignin [ADL]) and

PAGE 28

28 hemicellulose (NDF ADF) concentrations are also highly correlated with theoretical ethanol potential at 91 and 51%, respectively (Lorenz et al., 2009b) Theor etical E thanol P otential and Y ield It is hard to estimate efficiency of overall fermentation procedures on a commercial scale because numerous hydrolysis methods or combinations are available, each with specific advantages and disadvantages and different r eaction yields (Lu and Mosier, 2008) To estimate ethanol yield, a reasonable formula for ethanol yield potential is needed. For t hose reasons, theoretical ethanol potential (TEP) has been proposed to estimate maximum achievable ethanol potential from a feedstock. The ethanol potential is defined as the amount of ethanol that is able to be produced from the post enzymatic hydrolysis broth, assuming 100% fermentation efficiency, as a percentage of the amount of ethanol that could theoretically be produced from the original feedstock (Department of Energy, 2006b) The TEP can be estimated by i) using NDF and ADF analysis with subsequent estimation of cellulose and hemicellulose concentration or ii) total structural carbohydrate concentration using the NREL procedure. After that, theoretical ethanol yield (TEY) can be estimated thr ough combining TEP with biomass yield data. Ash and M ineral E lements Grass maturity and soil nutrients affect not only biomass yield but N and ash concentrations (Adler et al., 2006) Desirable composition of biomass for bioenergy is dependent upon the post harvest conversion processes used. Unlike forages for livestock minimizing N and ash in biofuel feedstock is desirable (Waramit et al., 2011) This is why investigating ash and macroelemental mineral concentration is essential f or characterizing compositional quality of biomass feedstock.

PAGE 29

29 Management E ffects on P erennial B ioenergy G rasses Logistical I ssues and F ield M anagement in the B iofuel I ndustry In the biofuel industry, one barrier to utilization is cost of logistics (Rentizelas et al., 2009) Unlike petroleum, many types of biomass are characterized by seasonal variation in availability as they are harvested only at a specific optimal time to maximize yield, but the power plant must operate on a year round basis to optimize efficiency (Rentizelas et al., 2009) Although this limitation can be partially solved by using storage strategies, any storage method involves additional logistics and cost. Harves t management is one option for achieving superior seasonal distribution of biomass without large increases in cost. If the harvest window can be widened, it would improve efficiency of operation of processing plants. Delaying harvest until after aboveground shoots of m iscanthus ( Miscanthus giganteus ) were killed by freezing reduced moisture and increased leaf senescence leading to decreases in ash and N concentrations (Jrgensen and Sander, 1997; Lewandowski and Kicherer, 1997) Harvest management of reed canarygrass ( Phalaris arundinacea L. ) in Iowa and Wisconsin affected biomass composition; more mature biomass had greater concentrations of fiber, reduced digestibility, and lower protein than less mature biomass. In that research a utumn harvested biomass from a two harvests per year system had higher protein and digestibility and lower cell wall concentration th a n a single harvest in autumn. It allowed feedstock to remain in the field and improved feedstock quality by decreasing Cl, K, P, and Ca, but increasing Si. Harvest F requency and S easonality A ffect B iomass Y ield Harvest frequency and timing affect grass performance. It has been shown, f or example, that different harvest management practices can affect biomass yield and

PAGE 30

30 quality. With delayed harvest in southern Germany, miscanthus biomass yield decreased by 14 to 15% from December to February and an additional 13% decrease occurred from F ebruary to March along with a significant decrease in water concentration, ash, N, Cl, and S (Lewandowski and Heinz, 2003) Delaying harvest from October to April of the following year decreased switchgrass ( Panicum virgatum L.) yields 40% in Pennsylvania (Adler et al., 2006) ; however, only 10% of the yield reduction was caused by decreased tiller mass and 90% resulted from the inability of harvest equipment to pick up the biomass in the field. Woodard et al. (1991) investigated the effect of harvest frequency on tall perennial grasses in Florida. They observed a significant decrease in overall biomass yield of elephantgrass and energycane by increasing harvest frequency. The magnitude of yield reduction differed by number of harvests. For instance, three harvests per year decreased yield 35% compared with a single harvest, but two harves ts per year decreased yield only 12% compared with a single harvest. Twoyear averages for yield of one, two, and three harvest s per year treatments were 25.6, 22.5, and 14.2 Mg ha1, respectively (Woodard et al., 1991b) Their study indicated that elephantgrass and energycane may tolerate two harvests per year with only modest decrease in yield, so this may be a useful strategy to increase the duration of the harvest period. In addition to direct effects of defoliation on plant regrowth vigor and production, timing of defoliation can affect the degree to which perennial grasses mobilize mineral nutrients from shoots to roots or rhizomes for st orage at the end of growing season (Somerville et al., 2010) This can be important because nutrient storage affects long term survival of perennial grasses (Chaparro et al., 1995)

PAGE 31

31 Harvest F requency and S easonality A ffect C omposition of P erennial G rasses Seasonality of harvest and harvest frequency affect plant maturity and cell wall concentration and composition They affect leaf : stem ratio because stem accumulates as length of regrowth period increases and lea ves senesce at the end of growing season (Chaparro et al., 1995; Macoon et al., 2001) As a result, delaying harvest can decrease N concentration and increase dry matter concentration and may be practical i f the bioenergy crop is resistant to degradation In this case, field stored feedstock could be provided to the processing plant on a just in time basis, s aving on storage cost (Heaton et al., 2009) The effect of harvest frequency occurs in part because of its impact on leaf : stem ratio and the fact that s tructural composition of leaf and stem are very different and are affected differently by maturation. For the C3 r eed canarygrass, a single harvest in fall or winter resulted in NDF (706 mg g1 DW) and ADF (400 mg g1 DW) concent ration s that were greater than a double harvest treatment (spring + fall ; NDF of 591 mg g1 DW and ADF of 308 mg g1 DW) (Tahir et al., 2011) For sweet sorghum ( Sorghum bicolor L. Moench) in n orthern coastal China, cellulose and hemicellulose concentrations range d from 206 to 265 g kg1 DW and 159 to 191 g kg1 DW respectively, in stems at anthesis, after which they decreased until grain maturity (40 d after anthesis) (Zhao et al ., 2009) However, there was little information about the effect of harvest management on structural composition changes of warm season grasses. Seasonality of harvest of warm season grasses was investigated in Iowa (Waramit et al., 2011) In this study, the concentration of cellulose in switchgrass and big bluestem ( Andropogon gerardii Vitman) increased with increasing maturity; the average early and late season concentrations were 250 and 398 mg g1 DW

PAGE 32

32 respectively in a 2yr study There was a clear difference among species in cellulose concentration during the late season; concentrations for switchgrass were 365 mg g1 DW while those for big bluestem were 430 mg g1. Average NDF, ADF, and ADL concentrations in elephantgrass were relatively con sistent in previous research. I n a study with elephantgrass in Ethiopia, NDF, ADF, and ADL concentration s were 616, 326, and 36 mg g1 DW, respectively (Tessema and Baars, 2004) This result was similar to NDF, ADF, and ADL concentration s of separated leaf and stem of Merkeron elephantgrass where leaf concentrations were 694, 360, and 30 mg g1, respectively and stem concentrations were 742, 481, and 69 mg g1, respectively (Anderson et al., 2008a) Fertilizer M anagement Fertiliz er management of bioenergy grasses is particularly important because in order for biomass to be a viable energy source production costs must be low. In Louisiana, L791002 energycane was fertilized with N P2O5K2O levels of 0 0 0, 1790 0, and 17967134 kg ha1. Yield was affected by fertilization treatment with yield totals of 16, 25.4, and 25.2 Mg ha1, resp ectively (Bischoff et al., 2008) In Florida, when applied N rate was almost double that of the Louisiana study (N P2O5K2O; 336 85 166 kg ha1) the 2yr average yield of L791002 (20.7 Mg ha1) from a single harvest each year was similar to that observed in Louisiana. In Georgia, L791002 energycane yield was above 20 Mg ha1 during the first 3 yr without fertilizer application, but it declined to 8.1 Mg ha1 in Year 4. In contrast, Merker on elephantgrass yielded above 30 Mg ha1 during first 2 yr and then declined to 11.2 and 7.1 Mg ha1 in the third and fourth years (Knoll et al., 2012) These studies show that fertilizer inputs are an important yiel d determinant for energycane and elephantgrass. Lower input systems are possible, but yields will be

PAGE 33

33 lower in these systems and some level of fertilizer input is essential to maintain persistence and acceptable biomass yield. Regional A daptation Planting site significantly affects biomass yield of elephantgrass and energycane. Regional micro climate and soil properties may influence biomass yield and compositional quality. From Ona, FL to Auburn, AL 3 yr average yield of energycane averaged 33.9 to 24.2 Mg ha1, respectively, however, elephantgrass showed a greater yield decline when it was planted at the cooler site in Alabama ( 46.7 at Ona to 18.6 Mg ha1 at Auburn) (Woodard and Prine, 1993b) At three locations within Florida, both elephantgrass and energycane yield was above 30 Mg ha1 without any difference by site in the first year (Fedenko, 2011) However, in the same study, site differences occurred in the second growing season with highest yields in high organic matter soils in Belle Glade, F L compared with Spodosols at Ona, FL an d Entisols in Citra, FL. The author suggested that Florida is an ideal region for production of tall warm season perennial grasses. However, even if sites are within a subtropical climate, performance of perennial grasses varies due to regional differences in climate and soil fertility. NitrogenU se E fficiency and C oncentration in B iomass Importance of N in G rassbased S ystem s Nitrogen is often the essential mineral element that crops require in greatest amount, and it represents a major production cost a nd environmental concern in US agriculture. Nitrogen is a major constituent of enzymes including those involved in photosynthesis (Heaton et al., 2009) Sources of N for plant growth include N mineralization from soil organic matter, N fixation from legumes, N fertilizer, and a number of minor N inputs such as lightning to agricultural systems. Nitrogen fertilizers

PAGE 34

34 are produced from industrial N fixation, specifically the Habor Bosch process. Industrial N fixation alone accounts for almost 2% of all human energy uses (Carroll and Somerville, 2009) High N concentration is valued in feed for livestock, but in contrast, greater N concentration is associated with greater emission of environmentally harmful substances during combustion of biofuel crops (Lewandowski and Heinz, 2003) While not beneficial in fuel production, N is critical to crop growth and development (Heaton et al ., 2009) Consequently, it is critical to minimize N input in biomass production while maintaining biomass yield. Seasonal D ynamics of N in P erennial G rasses Rhizomatous perennial grasses can mobilize mineral nutrients from aboveground organs into roo ts and rhizomes at the end of growing season (Carroll and Somerville, 2009) It has been proposed that harvest management can be used as a tool to increase the degree to which N is recycled from aboveground growth to storage organs for use the next season (Lewandowski and Heinz, 2003) These authors proposed that delaying harvest to February decreased biomass harvest 14 to 15%, but the biomass had significantly less water and N than that harvested in December. In a study with switchgrass, it was observed that spring harvest of the prev ious seasons growth decreased biomass yield compared with fall; however, biomass macroelemental concentrations including N decreased with delayed harvest (Adl er et al., 2006) Another switchgrass experiment showed that N concentration peaked in June (12.4 g kg1), significantly decreased by November (3.9 g kg1), and remained relative constant until April (3 g kg1) (Wilson et al., 2013a) Changes in N concentration differ by plant part from early season to harvest period. Nitrogen is largely withdrawn from senescing leaves before abscission, and

PAGE 35

35 used for the next growing season (Van Heerwaarden et al., 2003) Cellular constituents are systematically broken down and transported out of senesc ing organs, mostly leaves, which allow > 50% of leaf N to be reallocated by the plant (Van Heerwaarden et al., 2003) For instance, it has been shown that reallocable N content is decreased in miscanthus in early winter compared to late summer by decreasing the proportion of green and dead leaves and increasing lea f litter (Heaton et al., 2009) This research suggested that N in dropped leaves is contributed and may become incorporated into the soil organic matter pool, thus potentially improving N recycling during the next season. The N concentration of m is canthus and giant reed ( Arundo donax L. ) were studied in the United Kingdom (Smith and Slater, 2011) Although miscanthus leaf N concentration decreased rapidly during November before leaf abscission, miscanthus and giant reed leaf N concentr ation was higher than in their canes. If reallocable N from leaf and stem are measured separately over the season, this information may aid in determining the ideal harvest season for optimal N use efficiency. N use E fficiency by Q uantification of I soto pic N (15N) Nitrogen use efficiency in perennial grasses may be affected by applied N rate. For instance, apparent N uptake (there was no zero N control in the experiment) in energycane shoots was 72% when 336 kg N ha1 was applied and 95% for 168 kg N ha1 (Mislevy et al., 1995) This study also showed that the 336 kg N ha1 rate did not r esult in a biomass yield increase but did increase N concentration in plant tissue compared with 168 kg N ha1. Greater N concentration is considered desirable in grasses used for forage, but in the case of biomass for bioenergy, greater N concentration ei ther has no benefit (when biomass is fermented) or it can negatively affect the conversion process (when biomass is used for combustion). Although many studies have been conducted to

PAGE 36

36 determine N recovery in forage grasses (Martha et al., 2004) there is almost no N use efficiency or N cycling information available for bioenergy grasses. Such work is needed because N inputs to biomass production must be minimized in order to keep prod uction costs low. It may be possible to investigate N fertilizer removal and carryover using isotopic N (15N) to separate applied N from various other sources such as carryover, atmospheric deposition, mineralization of soil organic matter, and biological fixation. Morphol ogy of E lephantgrass and E nergycane Importance of M orphology in B ioenergy C rops As was mentioned earlier, harvesting frequency and timing play key roles in determining biomass yield and composition. Basically, the principle of harvesting frequency and timing for biomass is similar to that for defoliation management in livestock related grassland studies. For example, Chaparro et al. (1995) reported that defoliation frequency and height affected not only biomass but also morphological development in Mott dwarf elephantgrass. This study suggested that infrequent harvests allowed the gr ass to restore leaf area and reserves, but these benefits were accompanied by lower leaf : stem ratio (Chaparro et al., 1995) Although most defoliation studies have been conducted with perennial forage grasses, it i s also important to understand physiological and morphological responses to harvest management of perennial grasses used for biofuel. Leaf area index (LAI), stalk number, canopy development, spatial distribution, and proportion of leaf are important responses to be assessed (Madakadze et al., 1998; Trcsnyi et al., 2009; Len et al., 2012) In a study with switchgrass in southern Quebec, Canada, LAI reached a maximum in August (6.1 to 8) and decreased during the remainder of the growing season (Madakadze et al., 1998) This pattern was also observed with elephantgrass

PAGE 37

37 and energycane in Florida (Woodard et al., 1993) In elephantgrass, LAI increased from 1.4 at 30 d after staging to 7.1 at 105 d after staging before gradually decreasing to 3.4 at 245 d after s taging. In energycane, LAI similarly increased to 6.9 by 161 d after mowing and then decreased to about 4. In Florida, there was a significant positive relationship between canopy height and dry biomass, with r2 values of 0.94 for elephantgrass and 0.89 for energycane (Woodard and Prine, 1993b) Seasonal D ynamics of P erennial G rass M orphology Morphological change can be used as an indicator of timing of harvest in warm season grasses. Stems are the most important organ for bioethanol production, and they generally constitute the highest proportion of total aboveground dry weight. For example, sweet sorghum stems composed 56 to 73% of aboveground biomass followed by leaf dry weight in the range of 19 to 33% (Zhao et al., 2009) and a very small proportion of inflorescence. Yield dynamics over the season are primarily associated with changes in proportion of the harvestable part (standing tiller) and nonharvestable part (residue); 21% of switchgrass biomass was left in the field after machine harvest in the fall but the amount left behind increased to 45% for a winter harvest (Adler et al., 2006) Interestingly, harvested leaf and pa nicle weight decreased but harvested stem weight increased in winter (Adler et al., 2006) Thus, leaf and stem proportions vary by season and harvest managem ent (Zhao et al., 2009) Because cost of biomass transportation is a critical factor in the bioenergy industry, harvest timing is most effective when less valuable plant organs (leaves and inflorescence) which have low bulk density and high mineral element concentration are left in the field This goal can be achieved by investigating seasonal dynamics of morphological factors.

PAGE 38

38 Carbon Sequestr ation of Perennial Grasses Role of P erennial G rass C ropland for C S equestration One of the anticipated benefits to the use of biofuel s is a reduction in GHG emissions relative to that associated with the use of petroleum. Bioenergy crops fix atmospheric CO2 which is released when crops are burned (Vermerris, 2008b) However, there is still a net emission of GHG associated with the post harvesting procedure, transportation, and field management (Vermerris, 2008b) One other component that needs to be accounted for in determining the net carbon emission of bioenergy crops is soil C sequestration. It is very important to understand that perennial grasses capture atmospheric CO2 not only in harvested aboveground plant biomass but also in the below ground and the soil C pool s. The soil C pool is a function of the dynamic equilibrium between C gains and losses from the system under a specific land use (Lemus and Lal, 2005) Increased sequestration of soil organic C (SOC) is a potentially crucial strategy for offsetting CO2 emission to the atmosphere (Lemus and Lal, 2005) The potential of perennial bioenergy grasses to offset CO2 emission through soil C sequestration is dependent on the rate of soil C additions, long term capacity of soil for C storage, and the stability of sequestered soil C overtime (McLaughlin et al., 2002) Candidate bioenergy g rasses have advantages that favor soil C sequestration including perennality high biomass production, and a deep root system which can increase soil C compared with an annual crop (Ma et al., 2000b; Lemus and Lal, 2005) Moreover, bioenergy crops can be grown on marginal and nonarable land where soil erosion and degradation occur (Lemus and Lal, 2005) For instance, if forest and prairie grassland are converted to cropland for 1st generation biofuel crop production, the time required to restore C loss associated with

PAGE 39

39 land conversion ranges from 17 to 423 yr; however, conversion of marginal and abandoned cropland to 2nd generation biofuel crop production does not result in large initial SOC loss nor require a long time to restore SO C to the pre conversion level (Fargione et al., 2008) If we choose the right location (i.e., degraded cropland or marginal land) to produce 2nd gen eration biofuel crops, the criticism could be mitigated that using cropland for biofuels sacrifices the potential C benefit due to C emissions associated with the landuse change (Searchinger et al., 2008) Importance of S tudying W armseason G rassland C S equestration in Florida Soil degradation can be associated with disruptions such as erosion and runoff, which expose SOC to oxidative processes (Follett, 2001) In perennial grasses, the large active pool of root biomass is a major source of rhizosphere deposition and fine root turnov er, sequestering C in the soil. In a switchgrass study where it was used in a Conservation Reserve Program (CRP) in South Dakota, there was no significant effect of harvest treatments (once every year vs. alternate years) on SOC but source of N fertilizer affected SOC deposition rate, with 2.4 and 4 Mg C ha1 yr1 deposited for NH4NO3N and manure, respectively (Lee et al., 2007b) This study also indicated that there were no increases in SOC without N application. In contrast, N application (0, 11 2, and 224 N kg ha1) did not affect root weight density in switchgrass in Alabama (Ma et al., 2000a) The authors indicated that the majority of root biomass was distributed in the surface soil (015 cm), with proportions of 90 and 68% of total C in the top 15 cm for samples taken from intrarow and inter row locations, respect ively. A switchgrass study in the Dakotas and Nebraska showed SOC change varied considerably, ranging from 0.6 to 4.3 Mg C ha1 yr1 with average increase across sites of 1.0 Mg C ha1 yr1 for the 30cm depth in a 5 yr study (Liebig et

PAGE 40

40 al., 2008) This st udy showed that rate of change in SOC varied by site, indicating that it is important to assess changes of SOC associated with various management systems in Florida. Moreover, most SOC studies with bioenergy grasses have been done with switchgrass as the m odel crop. It is also necessary to investigate the change of SOC associated with use of tall sub tropical perennial grasses. It has been suggested that the potential mechanism responsible for SOC accrual is an N induced increase in root mass and subsequent turnover of roots (Fornara and Tilman, 2012) Likewise, it has been argued that increases in C sequestration by perennial grass roots will be due to increased root biomass rather than increased C concentration in the root (Ma et al. 2000a) Research with grazed Tifton 85 bermudagrass ( Cynodon spp.) pastures in Florida showed that soil C concentration increased linearly as the height of post graze residual stubble increased from 8 to 24 cm over 2 yr of grazing (Liu et al., 2011) In the same pastures, greater Tifton 85 post graze stubble height and rate of N fertilizer resulted in a linear increase in particulate organic C and total soil C and N in the < 53m soil particle size fraction (Silveira et al., 2013). Thus, there are data indicating a role of management in C sequestration under forage grasses in Florida, but there are currently no published data with perennial bioenergy grasses. These data are needed to help assess the overall C budget associated with growth of perenni al grasses for biofuel.

PAGE 41

41 CHAPTER 3 BIOMASS HARVESTED AND PLANT PART PROPORTION RESPONSES OF PERENNIAL BIOENERGY GRASSES TO HARVEST MANAGEMENT Overview of Research Greenhouse gas (GHG) emissions associated with fossil fuel combustion are believed to be a major cause of climate change (Cheng, 2 010) .This association of fossil fuel use with climate change has stimulated research to investigate alternative energy sources. From a national energy resource perspective, the USA is heavily dependent on imported petroleum, and the combination of polit ical turmoil in major production regions and large fluctuations in oil prices illustrate long term economic and security risks of this dependence. To increase energy security while minimizing the impact on food supply, use of nonfood resources for energy production is rising The US Department of Energy (DOE) and the US Department of Agriculture (USDA) have estimated that 1.2 billion metric tons of biomass can be produced per year in the USA and about 30% of this production could be biomass from perennial grasses (342 million metric tons) (Perlack, 2008) Lignocellulosic biomass is believed to offer less competition with food crops for existing cropland as well as improved energy efficiency and low GHG emissions (Erisman et al., 2010) Ce llulosic biomass from perennial grasses, especially those which utilize the C4 photosynthetic pathway, is an attractive biofuel feedstock in the Gulf Coast region of the USA because supplies are abundant, they are renewable, and these grasses have high pho tosynthetic efficiency (Knoll et al., 2012) Because of their C4 carbon fixation pathway and upright growth habit, elephantgrass ( Pennisetum purpureum Schum.) and energycane ( Saccharum spp. hybrid) are widely recognized for their biomass production in the southeastern USA

PAGE 42

42 (Woodard and Prine, 1991; Woodard and Prine, 1993a) Biomass yields of 20 to 48 Mg ha1 yr1 have been reported for these species in the region (Prine et al., 1984; Woodard and Prine, 1993a; Bouton, 2002; Woodard and Sollenberger, 2008) Although regional biomass production is high, a logistical challenge facing processing plants that convert feedstock to fuel is seasonality of biomass production. This results in uneven supply of feedstock to the conversion facility and limits efficiency o f operation. Storage of biomass is one alternative, but storage increases overall cost of energy production. A preferable option would be field management practices that address the seasonality of biomass supply. Harvest management affects biomass yield (Woodard and Prine, 1991) and could possibly be used to improve distribution of biomass to the refinery. Additionally, different grass species or taxa within species may be better adapted to flexible harvest management, but there is relatively little information available describing such differences for elephantgrass and energycane. The objectives of this study were to quantify the effects of elephantgrass and energycane harvest frequency and timing on biomass yield, leaf:stem ratio, and dry matter (DM) concentration, and to compare the potential for use in biomass production systems of an elephantgrass breeding line with that of the current predominant cultivar. Materials and Methods Experimental Site The experiment was conducted during 2010, 2011, and 2012 at the Plant Science Research and Education Unit (PSREU) at Citra, FL (29.41 N, 82.17 W). The soil was a well drained Candler sand (hyperthermic, uncoated Lamellic Quartzipsamments) Initial soil characterization of topsoil (020 cm) showed an average soil pH of 7.0, and Mehlich1 extractable P, K, Mg, and Ca of 54, 20, 123, and 496 mg

PAGE 43

43 kg 1, respectively These concentrations are considered to be high for P, very low for K, and very high for Mg. Monthly average, m aximum, and minimum temperatures (Figure 3 1) and monthly precipitation (Figure 32) are shown for the experimental period. Temperature and rainfall data during the study are from the PSREU weather station that is part of the Florida Automated Weather Netw ork (http://fawn.ifas.ufl.edu/ ) T he 30yr average data are for Gainesville, Florida and were reported by the Florida Climate Center ( http://climatecenter.fsu.edu/ ) Tr eatments and Experimental Design The treatments were all factorial combinations of three grass entries and three harvest management practices. Each treatment was replicated four times in a split plot arrangement of randomized complete block design. Harves t treatment was the main plot and grass entry was the subplot. The three grass entries included two elephantgrasses, Merkeron (Burton, 1989) and a breeding line referred to as UF 1, and L791002 energycane (Bischoff et al., 2008) The two grass species were chosen because earlier work w ith biomass feedstock identified them as having the greatest potential for use in this region (Woodard et al., 1991a; Woodard and Prine, 1993a; Bouton, 2002) Merkeron elephantgrass and L79 1002 energycane are also widely available cultivars of these two species. Bre eding line UF 1 was included because preliminary research had demonstrated its potential (Sollenberger et al., 2011) and larger scale plot work was needed to compare it with existing feedstock options and to provide data to guide decisions on potential cultivar release. Three harvest management treatments were imposed that included different frequencies and timing of harvest These were i) two harvests per year (2X; one in

PAGE 44

44 summer with a ratoon harvest before first freeze in fall), ii) one harvest per year in fall ( 1X Nov ; before first freeze and at initiation of flowering of Merkeron, generally the first of the entries to flower), and iii) one harvest per year in winter ( 1X Dec ; within 1 wk after first freeze, with a freeze defined as a temperature of less than 0C at 2 m above soil level ). Harvest dates for the three treatments are shown in Table 31. The 1X Nov treatment was considered to be a control because most data reported in the literature for these species are from experiments harvested once at the end of the growing season but before a freeze event. The 2X treatment was included to evaluate plant responses to more fre quent harvest that would increase the period during which biomass could be supplied to the biorefinery. The winter (1X Dec) treatment was imposed to evaluate timing of harvest, specifically the effect of delaying harvest until after a freeze event vs. harv est prior to freezing temperature ( 1X Nov ). If 1X Dec treatment proves viable it would allow extension of the harvesting period into the winter season. Plot Establishment and Management Plots contained six rows of 6m length, with 1 m spacing between rows. Plots were established using mature aboveground stem pieces planted on 15 Dec. 2009. Thus, the 2010 data are from the establishment year of the crops, and 2011 and 2012 data are from established stands. In all 3 yr, N was applied as ammonium sulfate ((NH4)2SO4) at a rate of 150 kg N ha1 yr1, and K was applied as muriate of potash (KCl) at a rate of 90 kg K ha1 yr1. Nutrients were split applied, with applications of 50 kg N and 45 kg K ha1 in mid April and 100 kg N and 45 kg K ha1 in mid May. No P was needed based on soil test. Elephantgrass is responsive to N at rates above those used in this study (VicenteChandler et al., 1959) but the intent was to select a rate that may

PAGE 45

45 more nearly represent what will be considered practical within a production context. Limited irrigation was applied to the experiment only at sign of significant drought stress (leaf rolling). Water was applied using a traveling gun system. There were five irrigation events in 2010 totaling 60 mm, three irrigation events in 2011 totaling 50 mm, and there was no irrigation applied in 2012. Response Variables At harvest, a 4m portion of one of the two middle rows was cut using a brush cutter to a 12cm stubble height to determine biomass harvested. To minimize border effects, the 1 m portion at the end of the har vested row was not part of the yield sample. The harvested area in each plot was thus considered to be 4 m2. All material from the harvested portion of row was weighed fresh in the field and then subsampled to determine DM concentration and to calculate DM harvested. In 2010 and 2011, four additional representative tillers from each plot were handseparated into leaf (blade and sheath) and stem (including inflorescence, if present) components and dried to determine leaf : stem ratio. After the biomass sample was collected, the remaining area of the plot was clipped to the target stubble using a disk mower. Persistence was characterized after the third year of defoliation. Plots were cleared after November and December 2012 harvests, but mild early winter weat her stimulated regrowth and provided opportunity to quantify persistence on 24 Jan. 2013. After 3 yr of growth it was not always possible to distinguish individual plants, so the measure of persistence chosen was percentage of row length per plot without v iable tillers. A segment of row was included in the total length having no tillers only if it was longer than 30 cm. Areas devoid of tillers were measured within the middle four rows of each plot, i.e., a total of 24 m of row length.

PAGE 46

46 Statistical Analysis Data were analyzed using mixedmodel methods in PROC MIXED (SAS Institute, 2008) In all models, harvest treatment and grass entry were considered fixed effec ts. Year was considered a repeated measurement (fixed) for all responses except persistence because this response was quantified only at the end of the experiment. Because harvest treatment was the main plot in the randomized complete block design, block a nd block harvest treatment interaction were considered random effects. Means were compared using the pdiff test of LSMEANS. All means reported in the text are least squares means and were considered different if P If statistical analysis did not detect difference ( P > 0.05) but it showed meaningful information, it is defined as a trend. Results and Discussion Biomass Harvested There were grass entry year and harvest management year interactions for total annual biomass harvested (Table 32). T he entry year interaction occurred primarily because of poor performance of energycane in the third year. Elephantgrass UF 1 outyielded energycane by 34% in 2011 and by 70% in 2012, while in the establishment year of 2010 there was only a trend ( P = 0.15 8) toward greater biomass harvested by UF 1 (11.3%; Table 33). Merkeron yield was 28% lower than UF 1 in 2011, but it was not different in 2012 and only tended to be lower ( P = 0.158) in 2010. Removing energycane from the analysis, there was no entry year interaction for the two elephantgrasses, and across the 3yr study period UF 1 biomass harvested averaged 28.0 Mg ha1 yr1 compared with 24.3 for Merkeron ( P = 0.010). Merkeron biomass harvested was not affected by year, while both energycane and UF 1

PAGE 47

47 elephantgrass had lowest biomass in 2012 (Table 33). The thirdyear decrease in biomass harvested was more pronounced in energycane than for either of the elephantgrasses. The harvest management year interaction occurred because there was no difference in biomass harvested due to harvest management in 2010 and 2011, but by 2012 the 1X Dec treatment produced 47% more biomass than the 2X treatment (Table 3 4). Within a harvest management treatment, there was no effect of year for 1X Dec but both 2X and 1X Nov treatments produced less biomass in the third year than in either of the first 2 yr. In a previous 3yr study, Merkeron elephantgrass yielded an average of 26.8 Mg ha1 yr1 of dry biomass when harvested once yr1 (Woodard and Prine, 1991) relatively similar to the 3yr average of 24.3 Mg ha1 in the current study. Woodard and Prine (1991) reported that Merkeron biomass harvested gradually decreased from 31.9 (Year 1) to 21.4 Mg ha1 yr1 (Year 3), unlike in the current research where biomass harvested decreased only from 24.8 (2010) to 22.8 Mg ha1 yr1 (2012). In the previous work, 3yr average yields were reduced by 19 and 33% for twoand threeharvest yr1 treatments respectively, when compared with a single harvest (Woodard and Prine, 1991) There were no differences among harvest frequency treatments in plant survival in their study, however, so it seems likely that the treatment effect on biomass harvested must h ave been due to reduction in average leaf area index and canopy light interception associated with multiple harvests yr1. No known previous studies have evaluated the effect of timing of a single harvest, so the greater yield stability across years of the 1X Dec vs. 1X Nov treatment has not been reported before this experiment.

PAGE 48

48 The large quantity of biomass harvested from perennial bunchgrasses has been attributed to their ability to continue biomass accumulation over an extended growing season (Woodard and Prine, 1993b) Although they found that LAI peaked 150 d after staging and then decreased over time, elephantgrass and energycane canopies continued to intercept more than 80% of photosynthetically act ive radiation (PAR) for 250 d after the initial mowing (Woodard et al., 1993) This is possible in subtropical regions where first freeze is delayed well into the fall (NOAA, 2013) extending the biomass accumulation period. A factor that may affect yield potential of UF 1 is that unlike Merkeron and L791002, which flowered in early November, UF 1 maintained vegetative growth until the time when winter freeze killed above ground biomass in each of the 3 yr of this study. A dditionally some tillers (~10%) of Merkeron flowered in April to May 2011 and 2012. This is thought to be due to an early cessation to freezing temperatures in those years resulting in early onset of plant growth which allowed some tillers to be of suffic ient size to respond to short days and initiate flowering. No tillers of UF 1 flowered in either spring. This response of Merkeron may decrease net biomass yield by reducing tiller vegetative growth. The 10 Mg ha1 decrease in energycane biomass harvested from 2011 to 2012 was associated with severe damage from the fungal disease sugarcane smut ( Sporisorium scitamineum ). Although severity of sugarcane smut damage was not rated quantitatively, it appeared visually that there was much greater presence of s mut by the third year (2012). Previously in Louisiana, L79 1002 was rated as moderately susceptible, however, that evaluation was during the establishment year when disease infection was relatively low (Bischoff et al., 2008) An energycane selection study

PAGE 49

49 showed that L791002 had a greater number of stools (bunches) and whips (classic symptom of the disease is the production of a black whiplike structure from the central core of the mer istematic tissue) in comparison with other energycane genotypes, indicative of high level of smut susceptibility (Len et al., 2012) Affected plants may tiller profusely but with poor cane formation and small narrow leaves (Ramesh Sundar et al., 2012) thus a greater number of tillers is not an indicator of fungal resistance. It is believed that since apical meristems of tillers are infected by the disease (Ramesh Sundar et al., 2012) apical dominance is broken and the plant produc es many small but often non viable tillers. Further breeding effort for energycane smut disease resistance is needed. Leaf : stem Ratio There were entry year (P < 0.001) and harvest management year (P = 0.002) interactions for leaf:stem ratio (Table 32) Leaf:stem ratio was measured only in 2010 and 2011, but in both years UF 1 elephantgrass had the lowest ratio (0.310.41; Table 3 5). The entry year i nteraction occurred because there was no difference between Merkeron and L791002 in 2010, but in 2011, L791002 energycane had greater leaf:stem ratio than either Merkeron or UF 1 elephantgrasses. For all grass entries, leaf:stem was greater in 2011 than in 2010, but the difference between years was greatest for energycane (Table 35). Harvest management year interaction occurred because leaf:stem ratio was quite low in the July harvest of the 2X treatment in 2010 (0.37) vs. 2011 (0.63), while differences between years were less pronounced for the other harvest management treatments (Table 36). In both years the greatest leaf:stem occurred in the November harvest of the 2X treatment (0.600.70), while the least was observed with the 1X Dec

PAGE 50

50 treatment (0.240.33). This was because the growth period of the ratooning plants after summer harvest was only 100 to 107 d, so their lower leaves showed minimal senescence and stems were relatively young. In both 2010 and 2011, leaf:stem ratio decreased over the growing season. This was caused by a significant decrease in LAI during fall (Chapter 5), associated with senescence of lower leaves and resultant loss in leaf biomass, and by increasing stem DM over the season. In a Florida study in which elephantgrass was harvested at 1.2, 2.5, 3.7, and 4.9 m of growth, green leaf percentage decreased from 50% (1.2m tall) to 10% (4.9 m) (Mislevy et al., 1989) The current research showed a trend similar to that observed by Mislevy et al. (1989). Many studies in which elephantgrass leaf:stem ratio have been quantified have targeted forage production and imposed multiple harvests per year (Mislevy et al., 1989; Chaparro et al., 1995; Williams and Hanna, 1995; Hanna et al., 2004) Experiments with Mott dwarf elephantgrass showed that leaf:stem ratio was negatively correlated with plant DM production (Williams and Hanna, 1995) The same author s reported that low leaf:stem ratio for Merkeron was due to stem internode composing a greater proportion of the plant. In the current study, there was no apparent relationship between biomass yield and leaf:stem ratio (Tables 3 4, 3 6) suggesting that the pattern of morp hological development of full season bunchgrass growth does not vary greatly among entries (Chapter 5) but does differ from that of bunchgrasses cut frequently for forage. The application of forage studies to biofuel production is limited because greater leaf:stem ratio, which is desired for forage production, is not preferred for biomass feedstock for several reasons. First, the physical form of biomass has the greatest

PAGE 51

51 influence on the cost of transportation (Sokhansanj et al., 2009) and the greater bulk density of the stem fraction allows more mass of material to be transported per unit of transport volume. Higher proportion of leaf in L791002 means less bulk density, thus L791002 likely will require higher transportation cost than elephantgrass entries, if it is transported as raw material. Moreover, leaf:stem ratio affects biomass quality Leaf is reported to have greater concentration than stem of components considered to be unfavorable (e.g., ash, Cl, Si, and N) in biofuel conv ersion especially for combustion (Lewandowski and Heinz, 2003) Biomass Dry Matter Concentration Dry matter concentration was affected by harvest management entry ( P < 0.001) and harvest management year ( P < 0.001) interactions. Harvest management entry interaction occurred because there were no differences among entries in dry matter concentration for 1X Nov and 1X Dec harvest treatments, but there were differences among entr ies for the 2X treatment at both harvests (Table 37). In July for the 2X treatment, elephantgrasses had or tended to have greater dry matter concentration than energycane, while in 2X for November energycane had greater dry matter concentration than either of the elephantgrass entries (Table 37). The harvest management year interaction occurred because in 2010 the 1X Dec management resulted in greater biomass dry matter concentration than 1X Nov and 1X Nov had greater dry matter concentration than eit her of the 2X harvests (Table 3 8). In contrast, 1X Dec and 1X Nov were not different in 2011 and 2012, but in both of those years 1X Dec and 1X Nov harvest treatments had greater biomass dry matter concentration than either harvest of the 2X treatment. In 2 of 3 yr the July harvest of the 2X management had the lowest dry matter concentration (Table 38).

PAGE 52

52 There are two primary factors that appeared to affect dry matter concentration in the current study, plant maturity and rainfall pattern. The effect of m aturity is supported in the literature. For example, when five C4 grasses were harvested at either 8 or 12 wk following seedling emergence (Costa and Gomide, 1991) dry matter concentration of 12wk herbage averaged 36 g kg1 greater than that of 8wk herbage. In Puerto Rico, dry matter concentration of five C4 grasses increased from ~ 200 g kg1 after 35 d of regrowth to 240 g kg1 after 45 d to 270 g kg1 after 55 d (Mendez Cruz et al., 1988) Spitaleri et al. (1995) reported that dry matter concentration of Pennisetum hybrids increased from 187 to 228 g kg1 as regrowth interval increased from 6 to 12 wk This response to maturity likely explains greater dry matter concentration of single vs. multiple harvest treatments in current experiment. Additionally, in the current study elephantgrass began growth earlier in the spring than energycane leading to earlier physiological maturation of elephantgrass plants and greater elephantgrass dry matter concentration than energycane for the summer harvest of the 2X tr eatment. This pattern of response did not carry over to the fall harvest of the 2X treatment or to 1X Nov and 1X Dec treatment s (table 37). Greatest dry matter concentration of the 1X Nov and 1X Dec treatment and the November harvest date of the 2X treatm ent occurred in 2012. These concentrations were associated with no measureable rainfall during November 2012 (Figure 31). Greatest November rainfall during the experiment occurred in 2011 (47 mm), approaching the 30yr average of 52 mm, and this was assoc iated with lowest dry matter concentration of the 1X Dec treatment in 2011. These data suggest that amount of rainfall during the fall affects dry matter concentration of standing tillers. Dry matter

PAGE 53

53 concentration is particularly important for biofuel feedstock because lesser moisture concentration reduces transportation cost and cost of additional drying that may be required during post harvest processing. Persistence Starting in 2011, portions of row devoid of tillers began to appear in some plots, so proportion of planted row without tillers was quantified after the 2012 growing season. There was harvest management x entry interaction for persistence (P = 0.009; Table 32). In the 2X harvest management, L791002 had the greatest pro portion of row without tillers present (32%; Table 39), while 1X Nov and 1X Dec harvest management of the same entry had gaps of 19 and 17%, respectively. Greater loss in the 2X treatment suggests either that more frequent harvest makes L791002 more susceptible to smut or simply that twice per year harvesting negatively affects energycane persistence. The response of elephantgrass entry UF 1 contrasted with that of energycane in that it had loss of only 9% of row with the 2X treatment but losses were 18 to 19% for the 1X Nov and 1X Dec treatments. Unlike other entries, Merkeron persistence was unaffected by harvest treatment and losses were not greater than 6%. Some elephantgrasses, including PI 300086, are susceptible to winter kill especially under multiple harvest management (Prine et al., 1988) Following 2 yr of defoliation, PI 300086 had 59, 8, and 8% stand survival for defoliation frequencies of 1, 2, and 3 harvests yr1, respectively; comparable percentages for Merke ron were 84, 75, and 81, respectively (Woodard and Prine, 1991) Thus our observation of excellent persistence of Merkeron is supported by previous results. Although UF 1 had greater gaps between tillers than Merker on for treatments harvested once per yr1, average biomass harvested of UF 1 across the 3 yr was greater than Merkeron and in 2012

PAGE 54

54 there was no difference in biomass harvested between elephantgrasses. Thus the degree of importance for UF 1 of the greater proportion of row without viable tillers than Merkeron is not totally clear at this point. It may be that UF 1 has greater ability for tiller size density compensation, allowing it to overcome the apparent disadvantage of greater gaps without tillers. Based on results of the current study, it is not possible to totally separate the impact of smut disease and defoliation frequency on energycane persistence, but it appears that both contributed significantly. Implications of Research Delaying a single harvest until after a freeze event or harvesting twice per year increased the effective harvest period of biomass and would therefore likely improve seasonal distribution of biomass to the biorefinery. Under the conditions of this experim ent, however, there was a 42% reduction in biomass harvested for the 2X treatment in the third year (averaged across entries) suggesting that increasing harvest frequency may compromise long term biomass production. This reduction was associated with a los s of stand for L791002 energycane, but stands of UF 1 and Merkeron harvested twice per year remained at greater than 90% through the third year. Energycane biomass decreased 41% from Year 2 to Year 3, due in significant measure to infestation with sugarcane smut disease. Smut resistant energycane cultivars will be required if this species is to provide a sustainable biomass supply. Delaying harvest until after a freeze event reduced the leaf percentage in harvested biomass and increased the biomass dry mat ter concentration. These factors can contribute to reduced transportation cost of biomass to the biorefinery, but it is not yet known how long after a freeze event the unharvested biomass can remain in the field without incurring major dry matter losses B reeding line UF 1 elephantgrass has excellent yield

PAGE 55

55 potential and possesses characteristics that make it an effective alternative to currently used grasses for biomass production. Evaluation under a broader range of management practices is warranted to fur ther assess its long term persistence under defoliation.

PAGE 56

56 Table 31. Harvest dates for 2X (summer and fall ratoon) 1X Nov and 1X Dec harvest management treatme nts in 2010, 2011, and 2012. Year Harvest management 2X July (summer) 1X Nov and 2X Nov (ratoon) 1X Dec 2010 30 July 10 November 9 December 2011 21 July 8 November 15 December 2012 24 July 7 November 28 November Table 32. Sources of variation and levels of probability ( P ) for their effects on response variables reported in Chapter 3. Sources of variation Total biomass harvested Leaf:stem ratio Dry matter concentration Persistence Harvest management (HM) 0.635 < 0.001 < 0.001 0.577 Entry (E) < 0.001 < 0.001 0.421 < 0.001 Year (Y) < 0.001 < 0.001 < 0.001 ---HM E 0.749 0.296 < 0.001 0.009 HM Y 0.002 0.002 < 0.001 ---E Y < 0.001 < 0.001 0.171 ---HM E Y 0.380 0.110 0.394 ---Persistence measured only at the end of the experiment, so year was not part of the model.

PAGE 57

57 Table 33. Grass entry year interaction ( P < 0.001) effect on biomass harvested. Data are means across four replicates and three harvest treatments (n = 12). Entry Year 2010 2011 2012 ---------------------------------Mg ha 1 ----------------------------------L79 1002 24.8 a A 24.2 aB 14.2 bB Merkeron 24.8 aA 25.3 aB 22.8 aA UF 1 27.6 bA 32.5 aA 24.1 cA SE 1.98 Year means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a year not followed by the same upper case letter are different ( P < 0.05) Table 34. Harvest management year interaction ( P = 0.002) effect on biomass harvested of three perennial grasses. Data are means across four replicates and three species (n = 12). Year Harvest management 2X 1X Nov 1X Dec ---------------------------------Mg ha 1 --------------------------------2010 27.0 a A 24.3 aA 25.9 aA 2011 29.0 aA 25.2 aA 27.8 aA 2012 16.9 bB 19.3 abB 24.8 aA SE 1.69 Harvest management treatments were twice per year (July and November; 2X), once per year in November ( 1X Nov ), and once per year after first freeze ( 1X Dec ). Harvest management means within a year not followed by the same lower case letter are different ( P < 0.05) Year means within a harvest management not followed by the same upper case letter are different ( P < 0.05)

PAGE 58

58 Table 35. Grass entry year interaction ( P < 0.001) effect on biomass leaf:stem ratio. Data are means across four replicates and three harvest treatments (n = 12). Entry Year 2010 2011 P value --------------------------Leaf:stem ratio ---------------------------L79 1002 0.41 a 0.70 a < 0.001 Merkeron 0.43 a 0.51 b 0.012 UF 1 0.31 b 0.41 c < 0.001 SE 0.034 Grass entry means within a year not followed by the same lower case letter are different ( P < 0.05) P value for year effect within a grass entry Table 3 6. Harvest management year interaction ( P = 0.002) effect on leaf:stem ratio of three perennial grasses. Data are means across four replicates and three species (n = 12). Year Harvest management 2X July 2X Nov 1X Nov 1X Dec ---------------------------Leaf:stem ratio ---------------------------2010 0.3 7 b 0.60 a 0.32 bc 0.24 c 2011 0.6 3 a 0.70 a 0.50 b 0.33 c P value <0.001 0.004 <0.001 0.010 SE 0.034 Harvest management treatments were twice per year (July and November; 2X), once per year in November ( 1X Nov ), and once per year after first freeze ( 1X Dec ). Harvest management means within a year not followed by the same lower case letter are different ( P < 0.05) P value for year effect within a harvest management

PAGE 59

59 Table 37. Grass entry harvest management interaction ( P < 0.001) effect on biomass dry matter concentration at harvest. Data are means across four replicates and three years (n = 12). Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec --------------------g dry matter kg 1 fresh weight --------------------L79 1002 209 b B 304 aA 318 aA 331 aA Merkeron 236 bA 260 bB 338 aA 355 aA UF 1 233 bAB 255 bB 326 aA 345 aA SE 12.8 Harvest management treatments were twice per year (July and November; 2X), once per year in November ( 1X Nov ), and once per year after first freeze ( 1X Dec ). Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05). Entry means within a harvest treat ment not followed by the same upper case letter are different ( P < 0.05) Table 38. Harvest management year interaction ( P < 0.001) effect on biomass dry matter concentration of three perennial grasses. Data are means across four replicates and three species (n = 12). Year Harvest management 2X July 2X Nov 1X Nov 1X Dec --------------------g dry matter kg 1 fresh weight --------------------2010 214 d B 252 cB 315 bB 343 aB 2011 240 bA 256 bB 298 aB 316 aC 2012 225 cAB 311 bA 368 aA 373 aA SE 8.4 Harvest management treatments were twice per year (July and November; 2X), once per year in November ( 1X Nov ), and once per year after first freeze ( 1X Dec ). Harvest management means within a year not followed by the same lower case letter are different ( P < 0.05). Year means within a harvest management not followed by the same upper case letter are different ( P < 0.05)

PAGE 60

60 Table 39. Harvest management x entry interaction ( P = 0.009) effect on proportion of row without viable tillers after 3 yr defoliation. Data are means across four replicates (n = 4) Entry Harvest management 2X 1X Nov 1X Dec --------------------------------------% --------------------------------------L79 1002 32 a A 19 bA 17 bA Merkeron 6 aB 5 aB 3 aB UF 1 9 bB 19 aA 18 aA SE 4.35 Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Year means within a harvest management not followed by the same upper case letter are different ( P < 0.05)

PAGE 61

61 Temperature (C) -100102030 Avg. Min. Max. 2010 2011 Month JanFebMarAprMayJunJulAugSepOctNovDec Temperature (C) -100102030 2012 Month JanFebMarAprMayJunJulAugSepOctNovDec 30 yr Avg. Figure 31. Monthly average and monthly maximum and minimum air temperatures for 2010, 2011 and 2012 for the experimental location, and the 30yr average for Gainesville, Florida.

PAGE 62

62 Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Rainfall (mm) 0 100 200 300 400 2010 (952 mm) 2011 (1,058 mm) 2012 (1,333 mm) 30-yr avg. (1,204 mm) Annual total rainfall Figure 32. Monthly rainfall for 2010, 2011, and 2012 for the experimental location and the 30yr average for Gainesville, Florida.

PAGE 63

63 CHAPTER 4 BIOMASS COMPOSITION RESPONSES OF PERENNIAL BIOENERGY GRASSES TO HARVEST MANAGEMENT Overview of Research Chemical composition of cellulosic biomass feedstocks plays a key role in their conversion to biofuel and the efficiency of biofuel production (Lee et al., 2007a) Gen otype x environment interaction exists and influences structural and chemical composition of lignocellulosic feedstocks (Lee et al., 2007a) Harvest management is known to be an important determinant of chemical composition of perennial grasses used for forage (Chaparro and Sollenberger, 1997), and it is reasonable to hypothesize that it affects composition of bioenergy feedstocks as well Plant biomass consists of three different types of polymers including cellulose, hemicellulose, and lignin, that are strongly bonded by noncovalent forces and by covalent cross linkages (Perez et al., 2002) Cell wall is composed of long crystalline cellulose microfibrils embedded in a matrix of other polysaccharides (Perez et al., 2002) Grasses contain glucuronoarabinoxylans (GAXs) that have a xylose backbone and are the predominant hemicellulosic polysaccharide in cell walls of the grass family (Vermerris, 2008a) Given the abundance of cellulose and hemicellulose in cell walls of perennial grasses, they represent a major source of structural carbohydrate for conversion to energ y. Lignin is a complex, amorphous, threedimensional phenolic polymer that negatively affects the yield of fermentable sugars from cellulosic biomass. This occurs as a result of shielding cellulose from microbial degradation by providing a surface that cel lulolytic enzymes adsorb to irreversibly (Akin, 2007; Vermerris, 2008a) The proportions of cellulose and lignin in biomass affect yield of biochemical conversion processes. For instance, due to high lignin concentration in wood, 36% more

PAGE 64

64 ethanol can be produced from switchgrass ( Panicum virgatum L.) than from same weight of wood biomass (McKendry, 2002) Neutral and acid detergent fiber (NDF and ADF, respectively) and acid detergent lignin (ADL) procedures (Van Soest et al., 1991) are used frequently to characterize forage cell wall composition, and those methods can be applied to investigation of the compositional characteristics of cellulosic biomass to be used for bio fuel. Limitations of the procedures include underestimation of total lignin and ADF may be contaminated by hemicellulose which causes an overestimation of cellulose and underestimation of hemicellulose (Morrison, 1980; Hatfield et al., 1994; Lowry et al., 1994; Jung et al., 1999; Jung and Lamb, 2004) Actual concentration of monomers from cellulose and hemicellulose can be analyzed by the Laboratory Analytical Procedures established by the National Re newable Energy Laboratory (NREL). Those procedures quantify sugar monomers from extractives (nonstructural carbohydrates; soluble sugars) and structural carbohydrates, and they measure acid soluble and acid insoluble lignin (Sluiter, 2008a; Sluiter, 2008b) Desirable composition of biomass for bioenergy is dependent upon the post harvest conversion process used, and unlike the situation with forage, minimizing N and ash in biofuel feedstock is preferred (Lewandowski and Heinz, 2003; Waramit et al., 2011) High levels of N and/or ash can reduce outputs of thermochemical conversion (Shahandeh et al., 2011) Warm season perennial grasses have up to two times greater N use efficiency than C3 plants (Jakob et al., 2009) and their high production of biomass with low N concentration makes them important candidate bioenergy grasses. Harvest management a ffects not only biomass yield but also its composition (Casler and Boe, 2003; Lewandowski and Heinz, 2003; Adler et al., 200 6) Miscanthus

PAGE 65

65 ( Miscanthus giganteus ) and giant reed ( Arundo donax L. ) showed a gradual decline in stem N concentration with increasing maturity in the UK (Smith and Slater, 2011) The decline in plant N with increasing maturity also has been associated with decreasing leaf proportion and much lower N concentration in stem than in leaf. Average switchgrass leaf N concentration was 13.5 mg g1 compared with 5.7 mg g1 for stem (Shahandeh et al., 2011) Similarly, giant reed and switchgrass leaf contained twofold or greater N compared to stem, and a decline in N and P concentration occurred from October to December (Kering et al., 2012) There are very limited data describing harvest management effects on chemical composition of grasses used as biofuel feedstock. The objectives of this research were to determine the effect of harvest management and grass entry on i) concentration of the Van Soest fiber fractions NDF, ADF, and acid detergent lignin (ADL), cellulose, and hemicellulose in biomass, ii) concentration of structural and nonstructural carbohydrates and lignin in biomass determined using NREL methodology, iii) mineral and ash conce ntration in biomass, and iv) the relationship between ADL using the Van Soest technique and lignin concentration using the NREL method. Materials and Methods Experimental Site The experiment was conducted during 2010 and 2011 at the Plant Science Research and Education Unit (PSREU) at Citra, FL (29.41 N, 82.17 W). The soil was a well drained Candler sand (hyperthermic, uncoated Lamellic Quartzipsamments) Initial soil characterization of topsoil (020 cm) showed an average soil pH of 7.0, and Mehlich 1 extractable P, K, Mg, and Ca of 54, 20, 123, and 496 mg kg1, respectively.

PAGE 66

66 These concentrations are considered to be high for P, very low for K, and very high for Mg. Treatments and Experimental Design Treatments were all the factorial combinations of thr ee grass entries and three harvest management practices. Each treatment was replicated four times in a split plot arrangement of a randomized complete block design. Harvest treatment was the main plot and grass entry was the subplot. The three grass entries included two elephantgrasses ( Pennisetum purpureum Schum.) Merkeron (Burton, 1989) and a breedi ng line referred to as UF 1, and L791002 energycane ( Saccharum sp p hybrid) (Bischoff et al., 2008) These species were chosen because earlier work with biomass feedstocks identified th em as having the greatest potential for use in this region. Merkeron elephantgrass and L791002 energycane were the available cultivars of these two species. Breeding line UF 1 was included because preliminary research had demonstrated its potential (Sollenberger et al., 2011) and larger scale plot work was needed to compare it with existing feedstock options and to provide data to guide potential cultivar release. Three harvest management treatments were imposed that included differences in frequency and timing of harvest. These included i) two harvests per year (2X; first harvest in July, ratoon harvest in November before fi rst freeze), ii) one harvest per year in fall ( 1X Nov ; before first freeze with timing based on flowering of Merkeron, generally the first entry to flower), and iii) one harvest per year in winter ( 1X Dec ; within 1 wk after first freeze, with a freeze defi ned as occurrence of temperature below 0C at 2 m above soil level). The 1X Nov treatment was considered a control because most data reported in the literature for these species are from experiments harvested once at the end of the

PAGE 67

67 growing season but befor e a freeze event. The 2X treatment was included to evaluate plant responses to harvest frequency. The 1X Dec treatment was imposed to evaluate timing of harvest, specifically the effect of delaying harvest until after a freeze event vs. harvest prior to fr eezing temperature ( 1X Nov ). In 2010, harvests occurred on 30 July for 2X July, 10 November for 1X Nov and 2X November and 9 December for 1X Dec In 2011, harvest dates were 21 July for 2X July, 8 November for 1X Nov and 2X Nov and 9 December for 1X De c. Plot Establishment and Management Plots contained six rows of 6m length, with 1 m spacing between rows and were established using aboveground stem pieces planted on 15 Dec. 2009. Thus, the 2010 data are from the establishment year. In both years, N was applied as ammonium sulfate [(NH4)2SO4] at a rate of 150 kg ha1, and K was applied as muriate of potash (KCl) at a rate of 90 K ha1. Nutrients were split applied, with applications of 50 kg N and 45 kg K ha1 in mid April and 100 kg N and 45 kg K ha1 in mid May. No P was needed based on soil test. Limited irrigation was applied to the experiment only at sign of drought stress (leaf rolling). Water was applied using a traveling gun system. There were five irrigation events in 2010 totaling 60 mm and three irrigation events in 2011 totaling 50 mm. At harvest, a 4m portion of one of the two middle rows was cut to 12cm stubble height using a brush cutter. Four representative tillers were selected from the harvested biomass for determination of plant part proportion and biomass composition. Those tillers were handseparated into leaf (blade and sheath) and stem (including inflorescence, if present) components. Samples were dried at 60C until constant weight. Stem samples were initially ground through a hammer mill to reduce particle

PAGE 68

68 size. Stem and leaf samples were ground to pass a 1mm stainless steel screen in a Wiley mill (Model 4 Thomas Wiley Laboratory Mill, Thomas Scientific, Swedeboro, NJ). Biomass Fiber Analysis Each component of dried biomass was analyzed for chemical composition Two different procedures were used for compositional analysis; detergent fiber analysis (Van Soest et al., 1991) and modif ied NREL procedures (Sluiter, 2008a; Sluiter, 2008b) For the detergent fiber procedure, the samples were sequentially analyzed for NDF, ADF, and ADL (Van Soest et al., 1991) with the exception that sodium sulfite and alphaamylase were excluded from the NDF analysis (Casler and Boe, 2003) The ANKOM fiber analyzer (ANKOM 2000 Fiber Analyzer, ANKOM Technology Corporation, Fairport, NY) was used for NDF and ADF determinations followed by the procedure Method for Determining Acid Detergent Lignin in Beakers (proposed by ANKOM Technology Corporation) which was used for ADL determination. Cellulose concentration was calculated as the difference between ADF and ADL and Hemicellulose concentration was calculated as the difference between NDF and ADF concentrations (Jung and Lamb, 2004; Waramit et al., 2011) In the modified NREL procedure, dried biomass was analyzed for nonstructural extractives, structural carbohydrates, and total lignin (Fedenko, 2011) For nonstructural extractives, 100 ml of deionized (D.I.) water was added to 1 g of dried sample and autoclaved in sealed pressure tubes (ACE Glass, Inc., Vineland, NJ) at 121C for 1 h. Filtered extractives were collected for high performance liquid chromatography (HPLC) analysis. The capt ured structural biomass from extraction was dried then weighed. A 0.3 g dry subsample was then use d for further two stage hydrolysis; the first stage was a 1h incubation with concentrated sulfuric acid (HPLC grade 72%, Fluka Alalytical,

PAGE 69

69 Sigma Aldrich, St. Louis, MO). In the next stage, sulfuric acid was diluted to 4% by adding D.I water then autoclaved for 1 h. Supernatant was collected and then neutralized to pH 7 for HPLC analysis. From the supernatant, acid soluble lignin determination was done at a w avelength of 240 nm using a UV Vis Spectrophotometer (StellarNet, Inc., Tampa, FL). Insoluble lignin was determined gravimetrically as total solids remaining in the crucible after vacuum filtration of hydrolyzed biomass. Total lignin was calculated as the sum of acid soluble and insoluble lignin. Extractives and hydrolyzed samples were analyzed by HPLC (Perkin Elmer Flexar system, Waltham, MA) using a refractive index detector and a Biorad Aminex HPX 87H column (300 x 7.8 mm) maintained at 50C. Sulfuric ac id ( HPLC grade, 4mM ) was used as the mobile phase at a flow rate of 0.3 mL min1 Elmers Chromera software was used to identify peaks. Linear regressions between peak areas of standard sugar concentrations were determined and used to calculate unknown sugar concentrations from the samples. A b rief explanation of each component analyzed is in Table 41 Total Nitrogen, Phosphorus, and Ash For N and P analyses, the method used was a modification of the standard Kjeldahl procedure and it was conducted at the Forage Evaluation Support Laboratory of the University of Florida. Samples were digested using a modification of the aluminum block digestion procedure (Gallaher et al., 1975) Sample weight was 0.25 g catalyst used was 1.5 g of 9:1 K2SO4:CuSO4, and digestion was conducted for at least 4 h at 375C using 6 ml of H2SO4 and 2 ml H2O2. Analysis of digestate was carried out using the Technicon Autoanalyzer and semiautomated colorimetry to determine N and

PAGE 70

70 P in the digestate (Hambleton, 1977) Samples were ashed at 500C for 6 h to determine ash concentration. Statistical Analysis Data were analyzed using mixedmodel methods in PROC MIXED (SAS Institute, 2008) In all models, harvest treatment and grass entry were considered fixed effects. Year was considered a repeated measurement (fixed). Block and all inter actions with block were considered random effects. Because there were two harvests per year of the 2X treatment, samples from both harvests were analyzed separately in the laboratory and the model effectively contains four levels of harvest management, i.e ., 2X July, 2X Nov, 1X Nov and 1X Dec Means were compared using the pdiff test of LSMEANS. All means reported in the text are least squares means and were considered different if P 0.05. Because the effects of greatest interest were harvest management entry, and their interaction, and because the harvest management entry interaction was significant in most cases, data in tables will show the means for this interaction. If the interaction was not significant, only main effect means will be compared i n the tables. If there were harvest management x year, entry x year, or harvest management x entry x year interactions for composition of leaf and stem components, interaction means are not presented nor discussed in the main body of the chapter in order t o reduce the volume of data presented and to simplify its presentation. For responses in which h arvest management x year and entry x year interactions were significant, means for measures of total biomass (not leaf or stem components) composition are prese nted briefly at the end of the chapter but are not discussed in detail. This was done because upon close inspection of these interactions the ranking of harvest management

PAGE 71

71 treatments or entry treatments was similar across years, thus the interaction was pr imarily caused by changes in magnitude and not direction of the response. Results and Discussion Van Soest Fiber Analyses Sources of variation (harvest management, entry, year, and their interactions) and levels of probability for their effects on NDF, ADF ADL, cellulose, and hemicellulose of leaf, stem, and total biomass are shown in Tables 4 2 and 43. Neutral detergent fiber Leaf NDF concentration was affected by harvest management and entry ( P < 0.001 for both, Table 4 2). Entry L791002 had greater leaf NDF concentration than elephantgrasses (Table 44). Averaged across the three grass entries, the single harvest treatments (780 and 789 mg g1 for 1X Nov and 1X Dec respectively) had greater leaf NDF concentration than the 2X harvest treatments (748 and 747 mg g1 for 2X July and 2X Nov, respectively). In stem, there was a harvest management x entry interaction ( P < 0.001, table 42). This was because L791002 had lesser NDF concentration in single harvest treatments ( 1X Nov and 1X Dec ) than in 2X harvests, but elephantgrass NDF was at least as great or greater for the single harvests compared with 2X. Within a harvest management, NDF concentration of the two elephantgrasses was greater than that of energycane for all but the November harvest of the 2X treatment. In total biomass, harvest management x entry interaction occurred ( P < 0.001). L791002 had similar or greater NDF concentration than elephantgrasses in the 2X harvest treatment but lower NDF concentration than elephantgrasses in 1X Nov and 1X Dec Within an entry, L791002 NDF concentration was least in 1X Dec and greatest

PAGE 72

72 in July and November harvests of the 2X treatment, but Merkeron and UF 1 had least NDF concentration in 2X harvests. The overall range of NDF in total elephantgrass biom ass was similar to that reported in previous research (Van Man and Wiktorsson, 2003) Most energycane fiber data reported in the liter ature were from analyses conducted after juice extraction, so viable comparisons with previous energycane research are few. In the current study, elephantgrass and energycane NDF concentration responded differently to harvest frequency. Specifically, full season growth of energycane stem and total biomass had lesser NDF concentration than stem harvested twice per year, while the pattern of response was the opposite for elephantgrass. Harvest frequency studies with reed canarygrass ( Phalaris arundinacea L. ) in Iowa showed that NDF concentration of biomass harvested twice a year (595 and 587 mg g1 for summer and ratoon, respectively) was less than that of a single fall harvest (650 mg g1) (Tahir et al., 2011) A harvest frequency and timing study in Oklahoma with switchgrass showed that a single harvest treatment had greater NDF concentration than twice a year harvests (Guretzky et al., 2011) In Easter n Canada, switchgrass NDF increased from 750 mg g1 in June to 850 mg g1 in July and remained constant for the remainder of the season (Madakadze et al., 1999) In the current research, NDF in biomass from single harvest treatments was greater than the 2X harvests for elephantgrass. This response is supported by previous forage studies with elephantgrass showing that as defoliation interval increased NDF concentration incr eased (Manyawu et al., 2003; Van Man and Wiktorsson, 2003) Observed increases

PAGE 73

73 in grass NDF concentration with increasing maturity have been related to changes in leaf:stem ratio (Madakadze et al., 1999) Although elephantgrass NDF response in the current study was similar to that observed for other grasses in previous research, energycane NDF response was different. Research with energycane has shown that it produces stem sap with relatively low Brix (range of 100 to 120 mg g1 in j uice) (Bischoff et al., 2008; Kim and Day, 2011; Te w et al., 2011) Although the Brix is low compared with commercial sugarcane, it composes a substantial portion of the biomass, and sucrose accumulation in late season is likely why energycane NDF concentration decreases as it matures. Relative to plant part NDF concentration, most research has shown that perennial grass NDF concentration in leaf was less than in stem (Griffin and Jung, 1983; Anderson et al., 2008b) For perennial C4 grasses big bluestem ( Andropogon gerardi Vitman) and switchgrass, leaf NDF varied little with increasing maturity and ranged from 642 to 663 mg g1; in contrast, stem NDF concentration of both grasses increased rapidly with maturation and averaged 8% greater than in leaves (Griffin and Jung, 1983) In the current study, leaf NDF concentration of elephantgrass was likely greater than typically observed because most previous studies involved use of the grass as forage and harvests were more frequent and did not involve the accumulation of large amounts of dead leaf as occurred in the single harvest treatments (Chapter 5). Greater NDF in leaf than stem has been observed previously in a study with elephantgrass conducted in China (Hsu et al., 1990) The difference between NDF in leaf and stem fractions was particularly pronounced for energycane, and this relates t o the

PAGE 74

74 accumulation of nonstructural sugars in energycane stem with increasing maturity, something that did not happen in leaf. Acid detergent fiber Similar to NDF, leaf ADF concentration was affected by harvest management and entry ( P < 0.001, Table 42) Energycane leaf had approximately 13% greater ADF concentration than Merkeron and UF 1 (Table 4 5). Across entries, leaf ADF was approximately 10% greater for the single harvest treatments ( 1X Nov and 1X Dec ) than either harvest of the 2X treatment. For the stem component, there was harvest management x entry interaction ( P < 0.001, Table 42). Interaction occurred because ADF concentration in L791002 stem did not differ and ranged only from 442 to 455 mg g1 across harvest management treatments, but th e range was much greater for elephantgrass and the 1X Nov and 1X Dec treatments always had greater ADF than the 2X treatments (Table 45). For total biomass there also was harvest management x entry interaction ( P < 0.001, Table 42), Interaction occurred because elephantgrass ADF concentration was greater than energycane with the exception of the 2X treatment harvested in November when there was no difference among entries (Table 45). For all three grass entries, the 2X treatment had lesser ADF concentrat ion than the single harvest treatments, and for elephantgrass the lowest ADF was observed for the 2X harvest in November. A harvest frequency study with reed canarygrass in Iowa showed that plants harvested in June had slightly greater ADF concentration (3 14 mg g1) than the ratoon harvest (302 mg g1), but ADF was greatest for a single harvest in fall (356 mg g1) (Tahir et al., 2011) In Oklahoma, a single harvest of switchgrass had greater ADF concentration than twice per year harvests (Guretzky et al., 2011) Defoliation frequency

PAGE 75

75 studies with elephantgr ass showed that increasing interval between grazing events increased ADF concentration from 360 to 398 mg g1 in Zimbabwe, 358 to 455 mg g1 in Vietnam, and 360 to 410 mg g1 in Ethiopia (Manyawu et al., 2003; Van Man and Wiktorsson, 2003; Tessema et al., 2010) In eastern Canada, switchgrass ADF concentration increased from 500 mg g1 in June to 650 mg g1 in July and remained relatively constant thereafter (Madakadze et al., 1999) This study showed that a large decrease in leaf:stem ratio occurred as plants matured, thus it was primarily changes in stem composition which affected ADF concentration in total biomass. Similar to the response o bserved for NDF in the current study, elephantgrass ADF concentration increased with single vs. 2X harvest treatments. Relative to plant part ADF concentration, in a previous elephantgrass cutting height study in Taiwan, stem had greater ADF than leaf acr oss cutting heights (Hsu et al., 1990) Similarly, mature Merkeron stem had greater ADF concentration than leaf (481 and 360 mg g1, respectively) (Anderson et al., 2008b) These results follow similar patterns to those observed for elephantgrass plant parts in the current study. Because of the substantial amount of extractable sugars that energycane stem accumulates late in the growing season, decreasing stem ADF, e nergycane leaf and stem ADF concentrations were more nearly the same than observed for elephantgrass. Acid detergent lignin Harvest management x entry interaction occurred for leaf ADL concentration ( P = 0.037, Table 42). Entry L791002 showed greater leaf ADL concentration than elephantgrass entries for all harvest management treatment s (Table 46). Interaction occurred because UF 1 leaf ADL was similar for all harvest treatments except the

PAGE 76

76 November h arvest of the 2X treatment, whereas the other entries showed greater leaf ADL for both 1X Nov and 1X Dec treatments than for either harvest of the 2X treatment. There also was harvest management x entry interaction ( P < 0.001) for stem ADL (Table 42). Unl ike leaf ADL concentration, elephantgrass showed greater stem ADL concentration than L79 1002 for each harvest treatment, and Merkeron had greater ADL than UF 1 for all but the November harvest of the 2X treatment (Table 46). Stem biomass in the 1X Nov harvest treatment had greater ADL than either of the 2X harvests for all three entries, and the November harvest of the 2X treatment had lower ADL than the July harvest for all entries except UF 1 (Table 46). For ADL in total biomass, there was harvest man agement x entry interaction ( P < 0.001), and the pattern of response was very similar to that of stem (Table 46). Specifically, ADL was greater for elephantgrass than energycane and greater for Merkeron than UF 1 elephantgrass for all but the 2X treatment in November. Reed canarygrass in Iowa had slightly lesser ADL concentration for a first harvest in June (25 mg g1) than a ratoon harvest in fall (29 mg g1), but both were less than observed for a single harvest in fall (39 mg g1) (Tahir et al., 2011) Several experiments have shown that increasing the interval between harvests is associated with increasing lignin concentration in elephantgrass (Dien et al., 2006; Tessema et al., 2010; Rengsirikul et al., 2011) Seasonal changes in lignin concentration of peren nial grasses have been studied with results showing that lignin concentration increased during summer, but the magnitude of the change varied by species (Waramit et al., 2011) In terms of plant part composition, stem was reported to have greater ADL than leaf in elephantgrass (Anderson et al., 2008b) This corresponds to results of the current

PAGE 77

77 study, where because of maturity differences single harvest treatments ( 1X Nov and 1X Dec ) had greater total ADL concentration than 2X treatments. Similar to the response of ADF, the difference between leaf and stem ADL concentration was less pronounced for energycane than for elephantgrass. Cellulose Cellulose concentration was calculated as the difference between ADF and ADL (Jung and Lamb, 2004; Waramit et al., 2011) In leaves, cellulose concentration was affected by harvest management and entry ( P < 0.001 for both, Table 43). Energycane had approximately 10% greater cellulose concentration than elephantgrass entries (Table 47). The single harvest treatments ( 1X Nov and 1X Dec ) showed greater leaf cellulose than 2X July and November harvests. This trend is very sim ilar to those observed for leaf NDF and ADF concentrations (Tables 44 and 4 5). In the stem, there was harvest management x entry for cellulose concentration ( P < 0.001). This was because L791002 had relatively constant cellulose concentration across har vest treatments, but elephantgrass cellulose concentration was generally greater with single vs. 2X harvests (Table 47). Elephantgrass stem cellulose concentration was always greater than for energycane, and the two elephantgrasses generally had similar c ellulose concentrations (Table 47). Cellulose concentration in total biomass was affected by the harvest management x entry interaction ( P < 0.001, Table 43). This response was nearly identical to that for stem, with the only effect of harvest treatment on energycane cellulose concentration observed for the 2X July treatment which was lower than 1X Nov and 1X Dec (Table 47). Merkeron showed greatest cellulose concentration for 1X Dec but there was no difference between Merkeron and UF 1 for the other harvest

PAGE 78

78 treatments, although both were greater than energycane for all but the 2X November treatment. Similar to the results for elephantgrass in the current study, reed canarygrass in Iowa had lesser cellulose concentrations in twice per year harvest treatments than single harvests with means of 273, 289, and 317 mg g1 for first harvest, ratoon, and single harvest, respectively (Tahir et al., 2011) In Iowa, cellulose concentr ation of several native, warm season perennial grasses increased with increasing maturity, with differences existing among species (Waramit et al., 2011) Less frequent grazing of elephantgrass for forage caused increased cellulose concentration (Tessema et al., 2010; Rengsirikul et al., 2011) In general, delaying a single harvest until fall is likely to maximize the concentration of cellulose in perennial grasses; however, the magnitude of this response is species dependent and the availability of that cellulose for conversion to energy is likely to be affected by increasing lignin concentratio n. Similar to the results observed in this study for elephantgrass, stem cellulose concentration (412 mg g1) was greater than leaf (357 mg g1) in mature Merkeron (Anderson et al., 2008b) In sweet sorghum ( Sorghum bicolor L. Moench) cellulose concentration was greater in leaves than stem (316 vs. 288 mg g1) (Fedenko, 2011) probably due to accumulation of nonstructural sugars in stem, simila r to the response observed for energycane in the current study. Hemicellulose Hemicellulose concentration was calculated as the difference between NDF and ADF concentrations (Jung and Lamb, 2004; Waramit et al., 2011) There was interaction of harvest management x entry for leaf hemicellulose concentration ( P =

PAGE 79

79 0.029, Table 43). The range in response was only 339 to 355 mg g1, thus the impact of treatment was minimal and the direction of the response was not consistent (Table 48). Unlike cellulose concentration, which was in the same general range for both leaf and stem (Table 47), hemicellulose concentration in leaves was greater than stem with an average of 346 and 249 mg g1, respectively, Table 48). The harvest management x entry interaction ( P = 0.009) was caused by greater hemicellulose concentration in L791002 stem vs. the elephantgrasses for the 2X July tre atment harvest but there were no differences among entries for 1X Nov and 1X Dec harvest treatments. For all three grasses, greatest stem hemicellulose concentration occurred in the 2X Nov treatment. In total biomass, hemicellulose concentration was affected by harvest management and entry ( P < 0.001 for both) but not their interaction ( P = 0.284, Table 43). In total biomass, hemicellulose concentration was greatest in L79 1002 followed by Merkeron a nd then UF 1. Similar to stem hemicellulose concentration, 1X Nov and 1X Dec harvest management showed lowest hemicellulose concentration in total biomass. Hemicellulose concentration of reed canarygrass in Iowa was relatively unaffected by harvest management with biomass from the two harvest per year treatment in June and fall and the single fall harvest treatment having concentrations of 281, 281, and 294 mg g1, respectively (Tahir et al., 2011) Other studies have shown that increasing rest periods between defoliation events resulted i n greater elephantgrass hemicellulose concentrations (Tessema et al., 2010; Rengsirikul et al., 2011) Different patterns of hemicellulose concentration were observed in several native warm season grasses in Iowa (Waramit et al., 2011) Data from the current experiment do not support the observation of greater hemicellulose concentration with increasing

PAGE 80

80 maturity observed in some studies, and it seems likely that hemicellulose response is less consistent across species and treatments than are the other cell wall fractions. Similar to results in the current research, it was reported that stem hemi cellulose concentration (261 mg g1) was less than in leaf (334 mg g1) of mature Merkeron elephantgrass (Anderson et al., 2008b) In China, elephantgrass leaf had greater hemicellulose concentration than stem across several cutting height treatments (Hsu e t al., 1990) Sweet sorghum showed a similar response with hemicellulose concentration of 265 and 169 mg g1 in leaf and stem, respectively (Fedenko, 2011) Considering the Van Soest fiber components as a group, concentration responses in total biomass usually paralleled those in the stem component for the 1X Nov and 1X Dec treatments. This is because stem proportion of perennial grasses increases as they mature (Griffin and Jung, 1983; Woodard and Prine, 1993b; Woodard et al., 1993) and t his has been shown to be true for energycane and elephantgrass as well (Chapter 5). There were important differences in fiber component responses among species (elephantgrass vs. energycane). This is not thought to be due to difference in the composition o f cell wall per se but to differences in the proportion of structural (i.e., cell wall) vs. nonstructural (i.e., extractives) components, particularly as maturity increased in the 1X Nov and 1X Dec treatments. Energycane accumulated nonstructural sugars late in the growing season and as a result the proportion of most structural components was less at this time than in elephantgrass. Delayed harvest ( 1X Dec ) generally did not affect concentration of cell wall constituents vs. that o bserved in 1X Nov Thus from a compositional standpoint, it would be a viable option to extend the window of harvest to at least the occurrence of first freeze in both elephantgrass and

PAGE 81

81 energycane. Extending the harvest period by using multiple harvests has greater impact on biomass cell wall composition. In general, biomass from early harvests or subsequent ratoon harvests has lesser concentrations of cell wall constituents compared with single harvests at season end. Depending on the conversion process, t his may or may not be advantageous. More frequent harvest does result in greater leaf proportion (Chapters 3 and 5) and associated greater concentrations of N and ash (Chapter 6) which can be antiquality factors in biomass conversion. National Renewable E nergy Laboratory Procedures Extractives are the sum of all non structural components of plant tissue removed during extraction (Sluiter, 2008a) Soluble sugars and other soluble substrates compose extractives. Hexoses and pent oses are the sum of six carbon (6C ; glucose and mannose, but mannose was not detected) and fivecarbon (5C ; xylose and arabinose) structural carbohydrates, respectively. Total lignin is the sum of acid soluble and insoluble lignin. Extractives Harvest management x entry interaction occurred ( P = 0.044) in the leaf because UF 1 showed greater extractives concentration than other entries in 1X Dec but there were no differences between elephantgrass entries for the other harvest managements (Table 49 and 411 ). Single harvest ( 1X Nov and 1X Dec ) leaf extractives concentration was lower than either 2X Ju ly or 2X Nov. This was because leaf senescence occurred during late season (Woodard and Prine, 1993b) and green leaf has a relatively greater concentration of extractives than senesced leaf Harvest management x entry interaction occurred for stem extractives ( P < 0.001, Table 49). Entry L791002 stem extractives was greatest in 1X Dec (338 mg g1)

PAGE 82

82 followed by 1X Nov (313 mg g1), however, Merkeron and UF 1 had greater stem extractives in 2X Ju ly and 2X Nov treatments (275 and 268 mg g1, respectively) than 1X Nov and 1X Dec (256 and 252 mg g1, respectively). In 1X Dec L791002 (338 mg g1) showed greatest stem extractives concentration followed by UF 1 (263 mg g1) and then Merkeron (241 mg g1). In total biomass, harvest management x entry interaction occurred ( P < 0.001). This was because there was no difference among entries within the 2X Nov harvest, but L791002 had greater extractives concentration than Merkeron for all other harvest treatments and greater than UF 1 in both 1X Nov and 1X Dec Elephantgrasses generally had greater extractives concentration in the 2X than single harvest treatments, but the pattern was generally reversed for energycane. There are very limited data available for extractives concentration in grasses. Elephantgrass in Vietnam had 182 mg g1 of extractives measured using hot water extraction (Hoa et al., 2008) Stem had greater extractives concentration than leaf in sweet sorghum in Florida (502 vs. 378 mg g1) (Fedenko, 2011) This author also evaluated total extractives of several perennial grass species, and concentrations occurred over a similar range as in the current study. He also found that extractives concentration was greater in energycane than elephantgrass (273 vs. 238 mg g1) as observed in the current study. Sugar accumulation in energycane results in increased extractives concentration when using a single harvest management. Corn ( Zea mays L.) stover water soluble extractives ranged from 142 to 203 mg g1, with total sugars composing 30 to 48% of extractives (Chen et al., 2007) In Colorado, extractives concentration varied among entries including 164 mg g1 for switchgrass, 254 mg g1 for

PAGE 83

83 fescue ( Festuca arundinacea L.) and 172 mg g1 for corn stover (Thammasouk et al., 1997) Total soluble sugars Harvest management x entry interaction occurred ( P = 0.044) because there were no differences among entries within the 2X July management, but differences among entries occurred within other harvest managements and one of the eleph antgrasses always had the greatest concentration (Table 49 and 412). However, overall concentration of total soluble sugars in the leaves was many times smaller than in stem (27 vs. 164 mg total sugars g1). Thus, although differences occurred in leaf, i t contributed only marginally to differences in total biomass. In the stem, harvest management x entry interaction occurred ( P < 0.001) for concentration of total soluble sugars (Table 49). For all three entries, 2X July total soluble sugar concentration was less than 1X Dec Magnitude of the difference varied among entries (84% increase in L791002, 51% in Merkeron, and 36% in UF 1, respectively). The large difference in energycane is related to sucrose accumulation. As mentioned previously, although energycane has relatively low sugar concentration compared with sugarcane, its brix value shows measurable levels of sugar accumulation in late season (Bischoff et al., 2008; Kim and Day, 2011; Tew et al., 2011) In total biomass, there was harvest management x entry interaction ( P < 0.001). As mentioned before, bec ause the concentration of leaf soluble sugar was very small and because leaf composed a small percentage of harvested biomass, particularly for 1X Nov and 1X Dec R, total soluble sugar concentration generally followed the same patter n as in stem. Concentra tion of total soluble sugars varied most among entries for

PAGE 84

84 single vs. 2x harvest treatments. Interestingly, among elephantgrass entries, UF 1 showed 44% greater soluble sugar concentration in total biomass harvested compared with Merkeron. Further investig ation of UF 1 soluble sugar concentration is needed to determine the specific sugars present. Unlike cell wall polysaccharides, noncell wall carbohydrates are directly fermentable; however, they are susceptible to microbial degradation during storage (Dien et al., 2006) Moreover, they contain cellulosic fermentation inhibitor s including furfural (Tran and Chambers, 1986; Zhang et al., 2011) If they cannot be used for conversion to biofuel, they will decrease feedstock quality. Fedenko (2011) reported that L791002 had up to twofold greater soluble sugar concentration than Merkeron, a similar result to the current study. Within species differences in sucrose concentration were shown among four different sorghum cultivars in Texas, rang ing from 2 (forage) to 239 mg g1 (sweet) (Stefaniak et al., 2012) Thus intra species differences have occurred for grasses in addition to those observed between the two elephantgrasses in cluded in the current experiment. Structural hexose In the leaves, structural hexose concentration was affected by harvest management and entry (P < 0.001 for both) (Table 410). Entry L79 1002 showed greater hexose concentration than elephantgrass entries by 7% (Table 413). Structural hexose concentration was greater in 1X Nov and 1X Dec followed by 2X July and then 2X Nov. In stem, there was a harvest management x entry interaction ( P = 0.001). Elephantgrass entries had greater structural hexose concent ration than L791002 for all harvest treatments (Table 413). Hexose concentration in L791002 was similar for 2X

PAGE 85

85 and 1X Nov treatments, but it was least for 1X Dec Merkeron and UF 1 had lowest structural hexose concentration in 2X Nov, while UF 1 had greatest hexose concentration in 1X Nov and for Merkeron 1X Nov and 1X Dec were not different. In total biomass, harvest management x entry occurred ( P = 0.001). For 2X Nov, there was no difference among grass entries; however, entry L791002 had lower hexos e concentration in total biomass than the elephantgrasses in 2X July, 1X Nov, and 1X Dec (Table 413). Comparing harvest managements within an entry, both Merkeron and UF 1 had greatest hexose concentration in 1X Nov and 1X Dec and least in 2X Nov, while energycane hexose concentration varied relatively little and was greater in 1X Nov than in 2X Nov and 1X Dec A large amount of glucose is advantageous for ethanol production because glucose can be converted more efficiently to ethanol than most other sugars, especially, pentoses. Similar to values observed in the current study, structural hexose concentrations of 347 mg g1 have been reported with mixtures of C4 grasses in Minnesota (Gillitzer et al., 2013) Likewise the range in six carbon structural sugars of four different varieties of sorghum in Texas (Stefaniak et al., 2012) was similar to concentrations reported in the current experiment. In a switchgrass study, chemical composition of six carbon structural sugars was 300 to 337 mg g1 (Xu et al., 2011) Similar to t he results in the current study, structural glucose concentration in elephantgrass and energycane was 374 vs. 366 mg g1 in the first year and 472 vs. 416 mg g1, respectively, in the second year (Fedenko, 2011) Late harvests and increasing maturity have been observed to increase glucose and nonglucose structural sugars in reed canarygras s (233 to 286 mg g1) and switchgrass (294 to 340 mg g1) (Dien et al.,

PAGE 86

86 2006) This corresponds to what was observed for elephantgrass in the current study, but is not similar to the response of energycane. Interestingly, structural hexose concentration using the NREL procedures was similar to Van Soest cellulose concentration in the current study. Further investigation of this relationship is needed. Structural pentose In leaves, pentose concentration was affected by harvest management ( P = 0.001 Table 410). The 2X July treatment had the lowest pentose concentration compared with other treatments but there were no differences among the other three. There was harvest management x entry interaction for stem pentose concentration ( P = 0.003). The most pronounced difference was that 2X Nov had greatest pentose concentration for all entries. For both 2X treatments, L791002 had greatest pentose concentration compared with the two elephantgrasses (Table 414). There was no difference among entries in 1X Nov while in 1X Dec Merkeron had greater pentose concentration than L791002. In total biomass, structural pentose concentration was affected by harvest management and entry ( P = 0.001, P < 0.001, respectively, Table 410). Entry L791002 had greater pentose concentration than the elephantgrasses; however, the difference among entries was relatively small (45 mg of pentose g1). The 2X Nov harvest management had greater pentose concentr ation than the other defoliation treatment s. In switchgrass, concentration of five carbon structural sugars ranged from 183 to 196 mg g1 (Xu et al., 2011) Structural xylose concentration in seasonlong growth of elephantgrass in Florida was similar to energycane (135 and 145 mg g1, respectively) (Fedenko, 2011) For both reed canarygrass and switchgrass, structural five carbon

PAGE 87

87 sugars increased with increasing maturity (147 to 191 mg g1 in reed canarygrass; 210 to 253 mg g1 in switchgrass) (Dien et al., 2006) That was not observed in the current study. Pentose concentration from structural carbohydrates of 198 mg g1 has been reported for a mixture of C4 grass and 187 mg g1 for a mixture of C3 grasses in Minnesota (Gillitzer et al., 2013) Although there were statistical differences in pentose concentration in the current study, unlike hexose co ncentration, pentose was relatively constant across harvest managements and entries (ranging only from 218253 mg g1). Lignin Leaf lignin concentration was affected by harvest management ( P = 0.005), and entry ( P < 0.001) (Table 410). Entry L791002 lea f lignin concentration was greater than elephantgrass entries (Merkeron and UF 1) by 8% (218, 204, and 201 mg lignin g1, respectively, Table 4 15). Single harvest management ( 1X Nov and 1X Dec ; 214 and 212 mg lignin g1, respectively) had greater leaf lig nin concentration than 2X harvest managements (2X July and 2X Nov; 204 and 202 mg lignin g1, respectively). For stem lignin concentration, there was harvest management x entry interaction ( P = 0.002, table 410). For all entries, 1X Nov and 1X Dec biomas s had greater stem lignin concentration than either of the 2X harvest managements (Table 415). The interaction occurred because there was no difference between Merkeron and UF 1 for 1X Nov ; in contrast, within the other harvest management treatments Merkeron had greater stem lignin concentration than UF 1 which had a greater lignin concentration then L791002. In total biomass, harvest management x entry interaction occurred ( P = 0.002). Similar to stem lignin concentration, both of the 2X harvest managements generally had

PAGE 88

88 lesser lignin concentration compared with 1X Nov and 1X Dec Merkeron had greater lignin concentration than energycane for all but the 2X Nov treatment. Total lignin concentration in the current study is similar to data reported in a previous study with full season growth of elephantgrass and energycane (Fedenko, 2011) and the results of both studies agree that elephantgrass generally has greater lignin concentration than energycane. There have been concerns raised by Jung and his colleagues about the accuracy of various analyses of lignin concentration (Jung et al., 1997; Jung et al., 1999; Jung and Lamb, 2004) Similar to challenges associated with quantifying structural carbohydrates, it is difficult to determine the actual amount of lignin in the plant cell wall. Because the NREL method is substep of a series of structural c omponent analyses (Sluiter, 2008a; Sluiter, 2008b) it is very hard to find previous research that reports this approach, although there are data in the literatur e using the Klason lignin analysis procedure which does not account for acid soluble lignin In a study with three switchgrass entries, lignin concentration was determined using the NREL procedure and ranged from 214 to 230 mg g1 for full season growth (Xu et al., 2011) a range that is only slightly greater than that observed for total biomass in the current study. Mineral Composition Nitrogen Nitrogen concentration in the leaves was affected by harvest management and entry ( P < 0.001 for both, Table 416). Energycane leaf had a lower N concentration (9.9 mg g1) than elephantgrass entries (average of 11.5 mg g1) (Table 417). Leaf N concentration averaged approximately 37% lower for the single harvest treatments ( 1X Nov and 1X Dec ) compared with the 2X treatments.

PAGE 89

89 In stem, N concentration was affected by harvest management ( P = 0.002), and entry ( P = 0.011) (Table 416). Merke ron (5.1 mg g1) N concentration was greater than other entries (4.3 mg g1 in UF 1 and 4.0 mg g1 in L791002, Table 417). The 2X July treatment, which was harvested about 2 mo after the final fertilizer application, had greatest N concentration, while 1 X Nov and 1X Dec harvest management were lowest in N. Similar to stem N, total biomass N concentration was affected by harvest management ( P < 0.001) and entry ( P = 0.004) (Table 416). Merkeron had greatest N concentration. Nitrogen concentration in total biomass decreased 39% from 2X July and 2X Nov to 1X Nov and 46% to 1X Dec Similar to responses in the current study, harvest frequency of reed canarygrass in Iowa affected N concentration. When biomass was harvested twice per year (June and fall ratoon), N concentration was13.4 and 8.8 mg g1 compared with 8.3 mg g1 for a single harvest in fall (Tahir et al., 2011) Elephantgrass plant height at cutting in Japan affected N concentration, with taller plants having lesser N concentration in both leaf and stem (Hsu et al., 1990) In the same study, the difference between leaf and stem N concentration increased as height increased (from 10.8 mg g1 in leaf vs. 5.0 mg g1 in stem at a 1m cutting height; 8.6 mg g1 in leaf vs. 2.6 mg g1 in stem at a 2m cutting height). Switchgrass biomass N concentration decreased over the season (Madakadze et al., 1999) ; however, N concentration was relatively constant after September at about 5 mg g1. Miscanthus N concentration decreased until October and then remained constant for the rest of the season (Heaton et al. 2009) Similarly, delaying harvest from 1X Nov to 1X Dec in the current study did not result in a significant decrease in either

PAGE 90

90 leaf or stem N, but 1X Dec N concentration of total biomass was less than 1X Nov (Fig. 4 17) because of the decrease in leaf percentage in 1X Dec relative to 1X Nov (Chapter 3). In Georgia, Merkeron had greater overall N concentration than L791002 ( 3.8 vs. 2.7 mg g1) (Knoll et al., 2012) similar to the pattern of response in the current study. Phosphorus Phosphorus concentration in leaf, stem, and total biomass showed very similar trends. Phosphorus concentration was affected by harvest management and entry in leaf ( P = 0.002 and P = 0.00 1, respectively), stem ( P < 0.001 for both), and total biomass ( P = 0.001 and P < 0.001, respectively) (Table 416). In both part parts and in total biomass, Merkeron had greatest overall P concentration (Table 418). In leaf, stem, and total biomass, P concentration was greater for the 2X treatments compared with 1X Nov and 1X Dec For example, in total biomass, P concentration in 2X No v was approximately twice as great as the single harvest treatment. Reed canarygrass P concentration in Iowa was greater f or two harvests per year than for a single harvest (Tahir et al., 2011) similar to the response observed in the current study. Switchgrass P concentration also was affected by harvest management, and it decreased slightly from first harvest of a two harvests per year treatment to ratoon h arvest in fall to fall harvest of a single harvest per year treatment (1.3, 1.1, and 0.8 mg g1, respectively) (Guretzky et al., 2011) Giant reed and switchgrass P concentration decreased from October to December in Oklahoma (Kering et al., 2012) It is clear than plant part N and P concentration are affected differently by increasing maturity. Switchgrass l eaf N was much greater compared to stem, but leaf and stem P concentration varied only slightly (Shahandeh et al., 2011) The responses w ere similar with elephantgrass and energycane in the current study.

PAGE 91

91 Ash There were harvest management and entry effects ( P = 0.012 and P < 0.001, respectively, Table 416) on leaf ash concentration. Entry UF 1 had greatest leaf ash concentration followed by Merkeron and then L791002 (Table 419). 1X Nov and 1X Dec harvested biomass had lesser ash concentrations (48 and 44 mg g1, respectively) than 2X July and 2X Nov treatments (54 and 54 mg g1, respectively). In stem, ash concentration was affected by harvest management ( P < 0.001). The average of the single harvest treatments ( 1X Nov and 1X Dec ) was 41% less than the average of 2X Ju ly and 2X Nov (Table 419). In total biomass, ash concentration was affected by harvest management ( P < 0.001) and entry ( P < 0.012). Elephantgrass entries had greater (up to 11%) ash concentration than energycane (Table 419). Because of generally greater ash concentration in leaves than stem and greater leaf proportion in the 2X treatments (Chapter 3), total ash concentrati on in 1X Nov and 1X Dec was less than in either of the 2X treatments. The results of the current study are similar to those reported previously for Merkeron and L791002. Four year data were averaged, and Merkeron had greater ash concentration than L79 1002 (45.9 vs. 34.4 mg g1, respectively) (Knoll et al., 2012) Reed canarygrass responded differently than elephantgrass and energycane, as ash concentration increased slightly over the season (96, 106, and 107 m g g1 for two harvest per year in June and October, and for one harvest in October, respectively) (Tahir et al., 2011) In switchgrass, ash concentration in Iowa peaked in July (71 mg g1) and decreased until fall after which it remained relatively constant between 4.3 and 4.5 mg g1 (Wilson et al., 2013a) Delayed harvest after fall is not expected to affect ash

PAGE 92

92 concentration greatly. First, most biomass is stem by the end of the growing s eason and stem has lesser ash concentration than leaves. Secondly, it is much more difficult for nutrients to be leached out of stem than leaf. There are limited data for plant part ash concentration, with leaf was found to hav e greater ash than s tem because of large K removal in rice (Summers et al., 2001; Bakker and Elbersen, 2005) Harvest Management by Year Interaction s for Total Biomass Composition An indicated in the Materials and Methods, in order to present a manageable volume of data, harvest management x year interactions were not discussed in the main body of the chapter. There was harvest management x year inter action for total biomass ADL, hemicellulose, extractives, total soluble sugars, hexose, pentose, N, P, and ash concentrations, and the interaction means are presented (Table 420). For total biomass, ADL concentration in 2010 was greater than in 2011 only for the 1X Dec harvest management, and the ranking of harvest management treatments was similar in both years with lesser ADL for the 2X treatments (Table 420). Hemicellulose concentration was greater in 2011 for all harvest managements except 2X Nov, but in both years lowest hemicellulose concentrations occurred in 1X Nov and 1X Dec treatments. Extractives concentration was greater in 2010 than in 2011 for all harvest treatments, and in both years there was very little variation among harvest treatments (15 mg g1 in 2010 and 17 mg g1) in 2011. For total soluble sugars, concentrations were less in 2011 than 2010 for 1X Nov and 1X Dec only, but in both years greatest soluble sugar concentration occurred in 1X Dec Structural hexose concentration was great er in 2011 than 2010 with the exception of 1X Nov and 1X Nov had greatest hexose concentration in both years. Pentose concentration in total biomass was greater in 2011 than 2010 for all harvest management treatments and was

PAGE 93

93 greatest for 2X Nov in both years. Nitrogen concentration was greater in 2010 than 2011 for 2X Nov only, and in both years the greatest N concentration was observed for the 2X treatments. Phosphorus concentration was greater in 2010 than 2011 for all but the 2X Nov treatment, and in bo th years it was greater for the 2X treatments than for 1X Nov and 1X Dec Ash concentration was greater in 2011 than 2010 for only the 2X July treatment, but as with N and P the ranking of harvest treatments was similar in both years with greatest ash conc entration in biomass from the 2X treatments. The differences among harvest management treatments were already discussed earlier in the chapter. Although there were interactions of harvest management and year for quite a few responses, the general ranking of defoliation management treatments was similar in both years in all cases and thus the interaction was primarily due to changes in magnitude instead of direction of the response. This provides additional justification for minimizing consideration of thes e data. Entry by Year Interaction s for Total Biomass Composition There was entry x year interaction for total biomass NDF, ADL, hemicellulose, extractives, total soluble sugars, hexose, pentose, and lignin concentrations (Table 421). For total biomass NDF concentration, L791002 and Merkeron had greater NDF in the second year, but in both years NDF concentration in Merkeron was greater than for energycane. For ADL concentration, differences between years were no greater than 5 mg g1, and only UF 1 ADL concentration differed between years. Energycane had the lowest ADL in both years. Hemicellulose concentration was greater in 2011 for all entries, and in both years it was lowest in UF 1. Extractives concentration w as greater in 2010 than 2011 for all three entries, and energycane had greater concentration than Merkeron in both years. Total soluble sugars concentration was greater in 2010 than

PAGE 94

94 2011 for all entries and like for extractives Merkeron had less soluble sugars than energycane. Structural hexose co ncentration was greater in 2011 than 2010 for all three entries, and energycane had least structural hexose in both years. Structural pentose concentration was also greater in 2011 compared with 2010. Range in the response was narrow in both years, but in 2011 energycane had greater structural pentose than either of the elephantgrasses. Lignin concentration was greater in the second year for all entries, and in both years Merkeron had greater lignin concentration than energycane. Although there were interactions of entry and year for eight responses, the general ranking of entries was similar in both years in all cases and thus the interaction was primarily due to changes in magnitude instead of direction of the response. One pattern did emerge in these int eractions; most composition responses associated with greater structural constituents (e.g., NDF, structural hexose, structural pentose, and lignin) were greater in 2011, while those associated with lesser structural constituents (e.g., extractives and tot al soluble sugars) were greater in 2010 than 2011. This is consistent with the fact that 2010 was the establishment year and growth was slower to initiate while in 2011 shoot growth began earlier and more vigorously resulting in greater effective maturity at a given 2011 calendar date. This issue is discussed in more detail in Chapter 6. Implications of Research Harvest frequency (2X vs. single) significantly affects compositional quality of perennial grasses. Delaying a single harvest until fall appears to maximize the concentration of cellulose in total biomass, however, delaying harvest from 1X Nov to 1X Dec had very little impact on most responses with the exception of extractives and

PAGE 95

95 soluble sugars in energycane which increased significantly between 1X Nov and 1X Dec Relatively greater concentration of soluble sugars in energycane reduced the concentration of numerous structural components. A major factor affecting concentration differences due to harvest management was differences in leaf:stem ratio because leaves generally had greater N, P, and ash than stem. Later harvests were associated with lesser leaf percentage in total biomass and this caused N, P, and ash to decrease in 1X Nov and 1X Dec relative to 2X treatments. Total biomass N concentrati on in Merkeron decreased to a greater extent than the other entries because of greater leaf abscission. The 1X Nov and 1X Dec treatments of all entries had lesser concentrations of N and ash, component which can negatively affect some conversion processes, but they had greater lignin concentration than for 2X which may reduce accessibility of enzymes to structural hexoses.

PAGE 96

96 Table 41. Analyses conducted and c omposition al characteristics of components analyzed in Chapter 4. Procedure Analys i s Composition Van Soest NDF Cell wall structural constituents primarily composed of cellulose, hemicellulose, and lignin ADF Primarily composed of cellulose and lignin ADL Lignin (acid insoluble) Cellulose Estimated by ADF minus ADL Hemicellulose Estimated by NDF minus ADF NREL Extractives Sum of nonstructural components; nonchemically bound components (soluble sugars + organic acids) Total Soluble sugars Nonstructural carbohydrates primarily composed of glucose, fructose, and sucrose Structural hexose 6 C structural polymeric carbohydrates (glucose + mannose) Structural pentose 5 C structural polymeric carbohydrates from (xylose + arabinose) Total lignin Sum of acid soluble and insoluble lignin

PAGE 97

97 Table 42. Sources of variation and levels of probability ( P ) for neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) concentrations and their effects on response variables reported in Chapter 4. Source of variation Leaf Stem T otal NDF Harvest management (HM) < 0.001 0.716 0.332 Entry (E) < 0.001 < 0.001 < 0.001 Year (Y) 0.011 0.003 < 0.001 HM E 0.272 < 0.001 < 0.001 HM Y < 0.001 0.186 0.086 E Y 0.078 0.309 0.009 HM E Y 0.087 0.233 0.089 ADF Harvest management (HM) < 0.001 0.007 < 0.001 Entry (E) < 0.001 < 0.001 < 0.001 Year (Y) 0.591 0.056 0.141 HM E 0.090 < 0.001 < 0.001 HM Y < 0.001 0.112 0.210 E Y 0.215 0.233 0.263 HM E Y 0.317 0.330 0.458 ADL Harvest management (HM) < 0.001 0.001 < 0.001 Entry (E) < 0.001 < 0.001 < 0.001 Year (Y) 0.278 0.287 0.012 HM E 0.037 0.006 < 0.001 HM Y < 0.001 0.013 0.002 E Y 0.893 0.014 0.006 HM E Y 0.007 0.540 0.699

PAGE 98

98 Table 43. Sources of variation and levels of probability ( P ) for concentration of cell wall components (cellulose and hemicellulose) from detergent fiber analysis and their effects on response variables reported in Chapter 4. Source of variation Leaf Stem Total Cellulose Harvest management (HM) < 0.001 0.046 0.001 Entry (E) < 0.001 < 0.001 < 0.001 Year (Y) 0.375 0.006 0.008 HM E 0.195 < 0.001 < 0.001 HM Y 0.001 0.142 0.252 E Y 0.167 0.560 0.500 HM E Y 0.689 0.235 0.357 Hemicellulose Harvest management (HM) 0.026 < 0.001 < 0.001 Entry (E) 0.312 0.001 < 0.001 Year (Y) < 0.001 0.002 < 0.001 HM E 0.029 0.009 0.284 HM Y 0.001 < 0.001 < 0.001 E Y 0.038 0.026 < 0.001 HM E Y 0.354 0.629 0.198

PAGE 99

99 Table 44. Effect of grass entry x harvest management interaction on neutral detergent fiber (NDF) concentration in leaf ( P = 0.272), stem ( P < 0.001), and total biomass ( P < 0.001). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest managem ent 2X July 2X Nov 1X Nov 1X Dec Mean ---------------------------mg NDF g 1 dry matter ----------------------------Leaf L79 1002 782 784 822 823 803 A Merkeron 732 726 763 783 751 B UF 1 731 733 757 761 745 B Mean 748 b 747 b 780 a 789 a Stem L79 1002 710 abC 728 aB 694 bB 675 cC Merkeron 747 bA 752 bA 756 abA 775 aA UF 1 727 bB 734 abB 749 aA 740 abB SE 8.0 Total biomass L79 1002 738 abA 752 aA 735bC 718 cC Merkeron 743 cA 742 cB 759bA 778 aA UF 1 730 cB 734 bcC 751aB 744 abB SE 5.4

PAGE 100

100 Table 45. Effect of grass entry x harvest management interaction on acid d etergent fiber (ADF) concentration in leaf ( P = 0.090), stem ( P < 0.001), and total biomass ( P < 0.001). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec Mean ----------------------------mg ADF g 1 dry matter ---------------------------Leaf L79 1002 435 429 477 479 455 A Merkeron 388 379 417 431 404 B UF 1 392 384 413 412 400 B Mean 405 b 397 b 436 a 441 a Stem L79 1002 442 aC 445 aB 455 aB 442 aC Merkeron 501 bA 477 cA 526 aA 537 aA UF 1 484 bB 467 bA 514 aA 507 aB SE 10.0 Total biomass L79 1002 441 bB 439 bA 462 aB 453 abC Merkeron 466 cA 438 dA 498 bA 515 aA UF 1 458 bA 437 cA 495 aA 494 aB SE 7.6

PAGE 101

101 Table 46. Effect of grass entry x harvest management interaction on acid detergent lignin (ADL) concentration in leaf ( P = 0.037), stem ( P = 0.006), and total biomass ( P < 0.001). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter ar e different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec ----------------------------mg ADL g 1 dry matter -----------------------------Leaf L79 1002 57 b A 51 cA 66 aA 64 aA Merkeron 47 bB 37 cB 53 aB 54aB UF 1 48 aB 41 bB 51 aB 48 aC SE 1.5 Stem L79 1002 63 bcC 62 cB 73 aC 69 abC Merkeron 86 bA 76 cA 99 aA 98 aA UF 1 77 bB 72 bA 90 aB 87 aB SE 3.1 Total biomass L79 1002 61 bC 58 bA 71aC 68 aC Merkeron 74 bA 61 cA 87 aA 89 aA UF 1 69 bB 61 cA 82 aB 82 aB SE 2.5

PAGE 102

102 Table 47. Effect of grass entry x harvest management interaction on cellulose (detergent fiber analysis) concentration in leaf ( P = 0.195), stem ( P < 0.001), and total biomass ( P < 0.001). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec Mean --------------------------mg cellulose g 1 dry matter -------------------------Leaf L79 1002 378 378 411 414 395 A Merkeron 341 342 365 377 356 B UF 1 344 343 362 364 353 B Mean 354 b 354 b 379 a 385 a Stem L79 1002 379 aB 383 aB 383 aB 373 aC Merkeron 415 bcA 401 cA 426 abA 439 aA UF 1 407 bcA 395 cA 425 aA 420 abB SE 7.1 Total biomass L79 1002 380 bB 381 abA 391 aB 386 abC Merkeron 392 cA 378 dA 411 bA 426 aA UF 1 389 bA 377 cA 413 aA 412 aB SE 5.2

PAGE 103

103 Table 48. Effect of grass entry x harvest management interaction on hemicellulose (detergent fiber analysis) concentration in leaf ( P = 0.029), stem ( P = 0.009), and total biomass ( P < 0.284). Data are means across four re plicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec Mean ----------------------mg hemicellulose g 1 dry matter ----------------------Leaf L79 1002 347 b A 355 aA 345 bA 345 bB Merkeron 344 bAB 346 abB 346 abA 352 aA UF 1 339 bB 349 aAB 344 abA 350 aAB SE 2.6 Stem L79 1002 269 bA 283 aA 239 cA 233 cA Merkeron 246 bB 275 aAB 231 cA 238 bcA UF 1 243 bB 267 aB 235 bA 234 bA SE 4.3 Total biomass L79 1002 297 313 272 265 287 A Merkeron 277 303 261 262 276 B UF 1 271 297 256 250 269 C Mean 282 b 304 a 263 c 259 c

PAGE 104

104 Table 49. Sources of variation and levels of probability ( P ) for concentration of nonstructural components (extractives and total soluble sugars) and their effects on response variables reported in Chapter 4. Source of variation Leaf Stem Total Extractives Harvest management (HM) < 0.001 0.609 0.367 Entry (E) < 0.001 < 0.001 < 0.001 Year (Y) < 0.001 < 0.001 < 0.001 HM E 0.044 < 0.001 < 0.001 HM Y 0.200 0.046 0.011 E Y 0.207 0.210 0.016 HM E Y 0.002 0.103 0.002 Total soluble sugars Harvest management (HM) 0.004 0.001 < 0.001 Entry (E) 0.001 < 0.001 < 0.001 Year (Y) < 0.001 0.001 < 0.001 HM E 0.044 < 0.001 < 0.001 HM Y 0.003 < 0.001 < 0.001 E Y 0.259 0.031 0.001 HM E Y 0.009 0.769 0.036

PAGE 105

105 Table 410. Sources of variation and levels of probability ( P ) for concentration of structural components (hexose, pentose, and total lignin) and their effects on response variables reported in Chapter 4. Source of variation Leaf Stem Total Hexose Harvest management (HM) < 0.001 0.013 0.001 Entry (E) < 0.001 < 0.001 < 0.001 Year (Y) < 0.001 < 0.001 < 0.001 HM E 0.562 0.001 0.001 HM Y < 0.001 < 0.001 < 0.001 E Y 0.048 0.075 0.034 HM E Y 0.401 0.123 0.252 Pentose Harvest management (HM) 0.001 < 0.001 < 0.001 Entry (E) 0.141 0.132 0.017 Year (Y) < 0.001 < 0.001 < 0.001 HM E 0.977 0.003 0.139 HM Y 0.001 < 0.001 < 0.001 E Y 0.138 0.034 0.004 HM E Y 0.439 0.649 0.153 Lignin Harvest management (HM) 0.005 < 0.001 0.001 Entry (E) < 0.001 < 0.001 < 0.001 Year (Y) < 0.001 < 0.001 < 0.001 HM E 0.069 0.002 0.002 HM Y 0.021 0.388 0.303 E Y 0.054 0.197 0.002 HM E Y 0.108 0.034 0.005

PAGE 106

106 Table 411. Effect of grass entry x harvest management interaction on extractives in leaf ( P = 0.044), stem ( P < 0.001), and total biomass ( P < 0.001). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec ------------------------mg extractives g 1 dry matter ------------------------Leaf L79 1002 199 a B 196 aB 164 bB 173 bC Merkeron 221 aA 226 aA 194 bA 187 bB UF 1 216 aA 225 aA 197 bA 201 bA SE 4.0 Stem L79 1002 302 bcA 289 cA 313 bA 338 aA Merkeron 265 aC 262 aC 250 abC 241 bC UF 1 285 aB 274 abB 262 bB 263 bB SE 7.9 Total biomass L79 1002 267 bA 251 cA 269 bA 289 aA Merkeron 251 aB 248 abA 234 bcB 229 cC UF 1 262 aA 256 abA 244 bB 254 abB SE 6.0

PAGE 107

107 Table 412. Effect of grass entry x harvest management interaction on total soluble sugars concentration in leaf ( P = 0.044), stem ( P < 0.001), and total biomass ( P < 0.001). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec ------------------mg total soluble sugars g 1 dry matter ------------------Leaf L79 1002 23 bA 32 aB 14 cB 22 bB Merkeron 26 bA 40 aA 24 bA 22 bB UF 1 26 bA 35 aAB 29 abA 33 abA SE 2.9 Stem L79 1002 152 cA 165 cA 228 bA 280 aA Merkeron 97 bB 134 aB 138 aC 142 aC UF 1 138 bA 147 bB 161 bB 189 aB SE 11.2 Total biomass L79 1002 109 cA 110 cA 163 bA 205 aA Merkeron 74 cB 96 bB 106 abC 116 aC UF 1 102 cA 106 cAB 126 bB 167 aB SE 6.9

PAGE 108

108 Table 413. Effect of grass entry x harvest management interaction on structural hexose concentration in leaf ( P = 0.562), stem ( P = 0.001), and total biomass ( P = 0.001). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec Mean --------------------------mg hexose g 1 dry matter ---------------------------Leaf L79 1002 371 356 404 395 381 A Merkeron 346 334 370 369 355 B UF 1 342 333 369 357 350 B Mean 353 b 341 c 381 a 374 a Stem L79 1002 363 aB 361 aB 361 aB 342 bB Merkeron 400 bA 385 cA 415 aA 406 abA UF 1 397 bA 378 cA 413 aA 394 bA SE 7.5 Total biomass L79 1002 367 abB 360 bA 374 aB 357 bB Merkeron 382 bA 365 cA 403 aA 399 aA UF 1 381 bA 362 cA 402 aA 389 bA SE 6.2

PAGE 109

109 Table 414. Effect of grass entry x harvest management interaction on structural pentose concentration in leaf ( P = 0.977), stem ( P = 0.003), an d total biomass ( P = 0.139). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec Mean --------------------------mg pentose g 1 dry matter --------------------------Leaf L79 1002 252 266 273 272 Merkeron 251 266 271 272 UF 1 254 273 279 274 Mean 252 b 268 a 274 a 273 a Stem L79 1002 217 bA 244 aA 209 bcA 202 cB Merkeron 206 bB 233 aB 209 bA 213 bA UF 1 206 bB 230 aB 209 bA 209 bAB SE 3.1 Total biomass L79 1002 229 253 228 222 233 A Merkeron 221 246 225 225 229 B UF 1 222 246 228 218 228 B Mean 224 b 248 a 227 b 222 b

PAGE 110

110 Table 415. Effect of grass entry x harvest management interaction on lignin concentration in leaf ( P = 0.069), stem ( P = 0.002), and total biom ass ( P = 0.002). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec Mean ---------------------------mg lignin g 1 dry matter ----------------------------Leaf L79 1002 213 213 226 221 218 A Merkeron 199 197 210 211 204 B UF 1 200 197 204 202 201 C Mean 204 b 202 b 214 a 212 a Stem L79 1002 167 bC 167 bC 182 aB 177 aC Merkeron 184 bA 182 bA 203 aA 206 aA UF 1 176 bB 174 bB 199 aA 199 aB SE 3.5 Total biomass L79 1002 183 cB 186 bcAB 195 aB 190 abC Merkeron 189 bA 188 bA 204 aA 207 aA UF 1 183 bB 183 bB 201 aA 200 aB SE 2.9

PAGE 111

111 Table 416. Sources of variation and levels of probability ( P ) for nitrogen, phosphorus, and ash concentrations and their effects on response variables reported in Chapter 4. Source of variation Leaf Stem Total Nitrogen Harvest management (HM) < 0.001 0.002 < 0.001 Entry (E) < 0.001 0.011 0.004 Year (Y) < 0.001 0.002 < 0.001 HM E 0.334 0.942 0.955 HM Y < 0.001 0.048 0.005 E Y 0.395 0.646 0.381 HM E Y 0.824 0.418 0.688 Phosphorus Harvest management (HM) 0.002 < 0.001 0.001 Entry (E) 0.001 < 0.001 < 0.001 Year (Y) < 0.001 0.023 < 0.001 HM E 0.139 0.360 0.348 HM Y 0.030 < 0.001 0.013 E Y 0.205 0.496 0.365 HM E Y 0.179 0.084 0.183 Ash Harvest management (HM) 0.012 < 0.001 < 0.001 Entry (E) < 0.001 0.904 0.012 Year (Y) 0.804 0.014 0.022 HM E 0.359 0.257 0.853 HM Y 0.418 0.004 0.002 E Y 0.071 0.300 0.740 HM E Y 0.987 0.672 0.858

PAGE 112

112 Table 417. Effect of grass entry x harvest management interaction on nitrogen concentration in leaf ( P = 0.334), stem ( P = 0.942), and total biomass ( P = 0.955). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec Mean --------------------------mg nitrogen g 1 dry matter --------------------------Leaf L79 1002 12.5 12.9 7.4 6.6 9.9 B Merkeron 14.1 14.2 9.5 8.2 11.5 A UF 1 13.0 13.7 9.5 9.5 11.4 A Mean 13.2 a 13.6 a 8.8 b 8.1 b Stem L79 1002 5.3 4.4 3.3 3.0 4.0 B Merkeron 7.0 5.2 4.4 3.9 5.1 A UF 1 5.5 4.9 3.4 3.4 4.3 B Mean 5.9 a 4.9 b 3.7 c 3.5 c Total L79 1002 7.8 7.9 4.8 4.1 6.1 B Merkeron 9.1 8.7 5.8 4.8 7.1 A UF 1 7.6 8.0 4.6 4.2 6.1 B Mean 8.2 a 8.2 a 5.0 b 4.4 b

PAGE 113

113 Table 418. Effect of grass entry by harvest management interaction on phosphor us concentration in leaf ( P = 0.139), stem ( P = 0.360), and total biomass ( P = 0.348). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec Mean -----------------------mg phosphorus g 1 dry matter -----------------------Leaf L79 1002 1.40 1.94 0.72 0.57 1.16 B Merkeron 1.95 2.13 1.14 1.03 1.56 A UF 1 1.32 1.51 0.88 1.00 1.17 B Mean 1.56 a 1.86 a 0.91 b 0.87 b Stem L79 1002 1.30 1.44 0.70 0.63 1.02 B Merkeron 1.64 1.99 1.21 1.09 1.48 A UF 1 1.07 1.56 0.78 0.94 1.09 B Mean 1.34 b 1.66 a 0.89 c 0.89 c Total L79 1002 1.31 1.66 0.73 0.62 1.08 B Merkeron 1.71 2.04 1.19 1.07 1.50 A UF 1 1.13 1.53 0.80 0.95 1.10 B Mean 1.38 b 1.74 a 0.90 c 0.88 c

PAGE 114

114 Table 419. Effect of grass entry x harvest management interaction on ash concentration in leaf ( P = 0.359), stem ( P = 0.257), and total biomass ( P = 0.853). Data are means across four replicates and two years (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within an entry not followed by the same lower case letter are different ( P < 0.05) Entry means within a harvest management not followed by the same upper case letter are different ( P < 0.05) Entry Harvest management 2X July 2X Nov 1X Nov 1X Dec Mean -----------------------------mg ash g 1 dry matter -----------------------------Leaf L79 1002 44 45 33 32 38 C Merkeron 59 57 50 43 52 B UF 1 60 60 60 56 59 A Mean 54 a 54 ab 48 bc 44 c Stem L79 1002 44 43 28 22 Merkeron 40 43 26 25 UF 1 39 46 25 27 Mean 41 a 44 a 26 b 24 b Total L79 1002 44 44 29 24 35 B Merkeron 46 48 32 29 39 A UF 1 45 51 32 31 40 A Mean 45 a 48 a 31 b 28 b

PAGE 115

115 Table 420. Effect of harvest management x year interaction on concentration of acid detergent lignin (ADL) ( P = 0.002), hemicellulose ( P < 0.001), extractives ( P = 0.011), total soluble sugars ( P < 0.001), structural hexose ( P < 0.001) and pentose ( P < 0.001), N ( P = 0.005), P ( P = 0.013), and ash ( P = 0.002) in total biomass. Data are means across four replicates and 2 yr (n = 8). Harvest management treatments were harvested twice per year (2X July, and 2X Nov ), once per year in November (1X Nov) and once per year after first freeze (1X Dec) Harvest management means within a year not followed by the same upper case letter are different ( P < 0.05) Harvest management 2010 2011 P 2010 2011 P 2010 2011 P -mg ADL g1 dry matter ------mg hemicellulose g 1 dry matter ------------mg extractives g 1 dry matter --------2X July 69 B 67 C 0.190 270 B 294 B < 0.001 276 A 245 A < 0.001 2X No v 59 C 60 D 0.583 305 A 303 A 0.414 261 A 242 AB 0.001 1X Nov 80 A 80 A 0.869 254 C 273 C < 0.001 270 A 228 B < 0.001 1X Dec 83 A 76 B < 0.001 256 C 262 D 0.015 276 A 239 AB < 0.001 SE 2.4 3.5 5.8 ---mg total soluble sugars g1 dry matter ----mg hexose g1 dry matter mg pentose g1 dry matter 2X July 100 C 90 C 0.074 370 B 383 B 0.006 205 B 243 BC < 0.001 2X No v 105 C 103 BC 0.771 340 C 384 B < 0.001 243 A 253 A 0.001 1X Nov 147 B 116 B < 0.001 391 A 395 A 0.403 208 B 246 B < 0.001 1X Dec 185 A 140 A < 0.001 375 B 388 AB 0.008 207 B 237 C < 0.001 SE 6.3 5.8 2.6 ----mg N g1 dry matter --------mg P g1 dry matter ------mg ash g1 dry matter --2X July 8.2 A 8.1 A 0.868 1.48 A 1.28 B 0.050 41 B 49 A 0.001 2X No v 9.5 A 7.0 A <0.001 1.71 A 1.77 A 0.578 49 A 46 A 0.112 1X Nov 5.3 B 4.7 B 0.241 1.05 B 0.76 C 0.005 30 C 32 B 0.174 1X Dec 4.6 B 4.1 B 0.297 1.09 B 0.67 C 0.001 27 C 29 B 0.379 SE 0.508 0.126 1.85

PAGE 116

116 Table 421. Effect of entry by year on concentration of neutral detergent fiber (NDF) ( P = 0.009), acid detergent lignin (ADL) ( P = 0.006), hemicellulose ( P < 0.001), extractives ( P < 0.016), total soluble sugars ( P = 0.001), structural hexose ( P = 0.034) and pentose ( P < 0.001), and lignin ( P = 0.002) in total biomass. Data are means across four replicates and two years (n = 8). Entry means within a year not followed by the same upper case letter are different ( P < 0.05). Entry 2010 2011 P 2010 2011 P 2010 2011 P --mg NDF g1 dry matter ----mg ADL g1 dry matter ---------mg hemicellulose g1 dry matter -------L79 1002 722 C 749 B < 0.001 65 B 63 C 0.074 276 A 298 A < 0.001 Merkeron 748 A 762 A 0.006 77 A 78 A 0.319 273 A 279 B 0.004 UF 1 737 B 742 B 0.257 76 A 71 B 0.001 265 B 273 C 0.001 SE 4.2 2.2 3.2 ----------mg extractives g1 dry matter -------mg total soluble sugars g1 dry matter ----mg hexose g1 dry matter L79 1002 290 A 248 A < 0.001 166 A 128 A < 0.001 352 B 377 C < 0.001 Merkeron 256 C 225 B < 0.001 107 C 89 B 0.001 377 A 397 A < 0.001 UF 1 266 B 242 A < 0.001 130 B 120 A 0.038 378 A 389 B 0.013 SE 4.2 4.5 5.3 mg pentose g1 dry matter -mg lignin g1 dry matter -L79 1002 215 A 251 A < 0.001 181 B 196 B < 0.001 Merkeron 215 A 243 B < 0.001 192 A 202 A < 0.001 UF 1 217 A 240 B < 0.001 189 A 194 B 0.007 SE 1.9 2.3

PAGE 117

117 CHAPTER 5 SEASONAL CHANGES IN MORPHOLOGICAL CHARACTERISTICS OF ELEPHANTGRASS AND ENERGYCANE Overview of Research Due in part to t heir C4 carbon fixation pat hway, tropical grasses like elephantgrass ( Pennisetum purpureum Schum.) and energycane ( Saccharum spp. hybrids ) are widely recognized for their biomass production potential (Woodard and Prine, 1991; 1993a) Research efforts to identify herbaceous plants with the highest biomass yields for renewable energy purposes have consistently found elephantgrass and energycane to be attractive candidate species across the USA Gulf Coast Region. Yields of 20 to 48 Mg ha1 yr1 have been reported in the region (Prine et al., 1984; Woodard and Prine, 1991; Bouton, 2002; Woodard and Sollenberger, 2008) Morphological characteristics ar e important determinants of perennial grass biomass yield potential. Elephantgrass and energycane tillers grew vegetatively for 30 to 35 wk, and their greater yield than other grasses was associated with a longer linear growth phase (Woodard et al., 1991b) It has been reported that the long growth phase of elephantgrass was achieved in part because plants maintained a favorable light environment by altering structure of the canopy from more planophile early in the growing season to increasingly erectophile as the season progressed (Kubota et al., 1994) Morphological characteristics of candidate bioenergy species may also affect optimum harvest management. Elephantgrass leaf area index (LAI) increased more rapidly than energycane during early season growth in north Florida (Woodard et al., 1993) Elephantgrass LAI increased from 1.4 at 30 d after mowing to a maximum of 7.1 at 105 d after mowing. In contrast, energycane LAI was 0.5 at 30 d after mowing and

PAGE 118

118 reached a maximum of 6.9 at 161 d after mowing (Woodard et al., 1993) This was associated with differences in light interception, with elephantgrass reaching 90% light interception by 49 d after mowing compared with 91 d after mowing for energycane. It is reasonable to anticipate that different patterns of leaf area accumulation and light interception would be associated with differences in optimal harvest interval. Light environment may also affect tiller dynamics, as tiller senescence was observed for sugarcane ( Saccharum spp.) after canopy light interception exceeded 70% of photosynthetic active radiation (Inman Bamber, 1994) Changes in plant part proportion may also affect a range of biomass traits including dry matter concentration, nutrient removal in harvested biomass, and conversion of biomass to energy. Stems are typically considered the most important organ for bioethanol production, and they generally constitute the highest proportion of total aboveground dry weight. Elephantgrass and energycane green leaf mass reached a maximum approximately 100 d after mowing, but s tem mass and total biomass (stem and green plus dead leaves) increased linearly over 250 d resulting in much greater stem proportion with longer intervals following mowing (Woodard et al., 1993) Sweet sorghum [ Sorghum bicolor ( L. ) Moench] stems were found to compose 56 to 73% of aboveground biomass, and leaf and stem proportions varied with season and harvest management (Zhao et al., 2009) Seasonal changes in a wide range of morphological traits have not been studied comprehensively during first and ratoon growth periods and over more than one growing season for elephantgrass and ener gycane. It has been suggested that LAI, tiller number and mass, and plant part proportion are important responses to be assessed

PAGE 119

119 because of their effect on biomass accumulation and responses to defoliation (Madakadze et al., 1998; Trcsnyi et al., 2009; Len et al., 2012) The objective of this study was to quantify monthly changes in a range of morphological characteristics during first and ratoon growth of elephantgrass and energycane and relate these findings to differences in plant responses to defoliation (Chapter 3). Materials and Methods Experimental Site The experiment was conducted during 2010 and 2011 at the Plant Science Research and Education Unit (PSREU) at Citra, FL (29.41 N, 82.17 W). The soil was a well drained Candle r sand (hyperthermic, uncoated Lamellic Quartzipsamments) Initial soil characterization of topsoil (020 cm) showed an average soil pH of 7.0 and Mehlich 1 extractable P, K, Mg, and Ca of 54, 20, 123, and 496 mg kg1, respectively. Monthly average, maximu m, and minimum temperatures (Figure 31) and monthly precipitation (Figure 32) were shown previously for the experimental period. In 2010, last freeze event in spring was 7 March ( 1.4C) and first freeze event was 2 Dec ember ( 2C). In 2011, the last fre eze event in spring was 14 February ( 0.9C) but the first freeze did not occur until last sampling date, 13 Dec ember Treatments and Experimental Design The three grass entries included two elephantgrasses, Merkeron (Burton, 1989) and a breeding line referred to as UF 1, and L791002 energycane (Bischoff et al., 2008) As mentioned previously (Chapter 3), these species have potential as biofuel feedstock in this region. Merkeron elephantgrass and L791002 energycane were chosen because they are the most available cultivars of these tw o species. In addition, breeding line UF 1 elephantgrass was included because in preliminary research it

PAGE 120

120 demonstrated outstanding potential for use in bioenergy feedstock production (Sollenberger et al., 2011) and potential exists for it to be released as a cultivar. Plots used were a subset of those from the experiment described in Chapter 3. Fullseason growth sampling occurred on plots that were harvested once per year within 1 wk following the first freeze event in fall (treatment referred to as 1X Dec in Chapter 3), and first growth/ratoon growth sampling occurred on the plots that were harvested twice per year with first harvest on 30 July 2010 and 21 July 2011 (treatment referred to as 2X in Chapter 3). Data from full season growth plots were analyzed separately from first growth/ratoon growth data, so treatments were grass entries replicated four times in a randomized c omplete block design. Plot Establishment and Management Plots contained six rows of 6m length, with 1 m spacing between rows. Plots were established using aboveground stem pieces planted on 15 Dec. 2009. Thus, the 2010 data are from the establishment year and 2011 data are from well established stands. In both years, N was applied as ammonium sulfate ((NH4)2SO4) at a rate of 150 kg N ha1 yr1, and K was applied as muriate of potash (KCl) at a rate of 90 kg K ha1 yr1. Nutrients were split applied, with applications of 50 kg N and 45 kg K ha1 in mid April and 100 kg N and 45 kg K ha1 in mid May. No P was needed based on soil test. Limited irrigation was applied to the experiment only at sign of significant drought stress (leaf rolling). Water was applie d using a traveling gun system. There were five irrigation events in 2010 totaling 60 mm and three irrigation events in 2011 totaling 50 mm. Response Variables Data collection occurred at approximately 4wk intervals (Table 51). In 2011, the fully establ ished plants began growth earlier and more vigorously than the newly

PAGE 121

121 established plants in 2010, so first data collection started 4 wk earlier in 2011 than in 2010. At each collection date, LAI, canopy height, tiller density, and tiller mass were measured. The SS1 SunScan canopy analysis system (DeltaT Devices Ltd, Cambridge, United Kingdom) was used to obtain nondestructive estimates of LAI. Attenuation of photosynthetically active radiation (PPFD) was measured with a linear photosensor array probe. Ex ternal canopy beam fraction sensor (BF 3, located in a nearby plot alley) provided the simultaneous beam fraction data for SunScan system and six under canopy measurements were averaged to obtain the plot LAI estimates. The probe was placed about a 45 ang le across the four center rows (near the center of the plot) from four azimuth angles of each plot and a reading taken. To obtain consistent LAI estimates, LAI data were collected at high sun angles from 10 am to 2 pm (Eastern Standard Time) under generall y sunny or partly cloudy condition. All readings were collected following canopy closure using a default ellipsoidal leaf angle distribution parameter (ELADP) value of 1.0. When leaf senescence occurred in the late summer, the probe was placed above dead l eaves because this system could not distinguish between green vs. dead leaves. Canopy height was the average undisturbed height of plants quantified by taking 10 measurements in each plot. Height was measured to the nearest 0.1 m using a tool made from gra duated interlocking sections of PVC pipe. To measure tiller density, number of tillers was quantified in inner rows by counting in two permanently marked 0.5m sections of row per plot. To quantify tiller mass, plant part proportion, and tiller dry matter (DM) concentration, four representative

PAGE 122

122 tillers per plot were collected at each sampling event and handseparated into leaf (blade plus sheath) and stem (including inflorescence, if present) components. Fresh weight of each component was recorded. The sam ples were dried at 60 C until constant weight. Sum of leaf and stem dry weight was divided by the sum of leaf and stem fresh weight to determine DM concentration. For plant part data, only stem proportion is presented, and it was calculated as stem mass di vided by total tiller mass and multiplied by 100 to be expressed as a percentage. Statistical Analysis Data were analyzed using mixedmodel methods in PROC MIXED (SAS Institute, 2008) Data for full season growth and first growth/ratoon growth were analyzed separately because comparisons among entries was of primary interest and comparisons between full season growth and ratoon growth after midseason harvest were not deemed to be of biological importance. Years were analyzed separately because number of sampling events differed between years and sampling began earlier in 2011 than 2010 because 2010 was the year of establishment. In all models, grass e ntry was a fixed effect and sampling date (Table 51) was considered a repeated measurement (fixed). Block was considered a random effect. Means were compared using the pdiff test of LSMEANS. All means reported in the text are least squares means and were considered different if P Trends are discussed if P 0.10 > 0.05. Results and Discussion Tiller Density For the full season growth treatment, there were entry x sampling date interactions ( P < 0.001) in 2010 and 2011. In June 2010, energycane til ler density was

PAGE 123

123 greater than Merkeron and UF 1 (34, 20, and 20 tillers m2), and although this difference narrowed throughout the year, energycane had greatest tiller density at each sampling date (Fig. 5 1). In May 2011, energycane again had greater tiller density than Merkeron or UF 1 (54, 37, and 32 tillers m2). Unlike 2010, this difference lasted only through the end of June and thereafter there were no differences among treatments at any sampling date. In both years, energycane tiller number was great est early in the season and decreased later, but tiller number of elephantgrass entries did not vary widely throughout the year. In both years, tiller density of all entries changed very little after mid August. Previous research in Japan has shown that regardless of fertilizer level, elephantgrass tiller density increased from time of spring planting through July and decreased thereafter (Wadi et al., 2003) Working with sugarcane, InmanBamber (1994) reported that tiller density was a function of growing degree days, peaking about 500 degree (C) days after ratooning and then decl ining until about 1200 degree (C) days after which it stabilized. For the first growth/ratoon growth treatment, tiller density of first growth was affected by entry x sampling date interactions ( P < 0.001) in both years. Energycane had greater tiller density throughout the first growth in 2010 but only at the start of 2011 (Fig. 5 2). Tiller number was generally greater for ratoon than first growth, but variability was larger for tiller number in ratoon growth than first growth. This response is likely due to repeated cutting, as previous studies have shown that elephantgrass harvested three times per year had tiller density ranging from 45 to 122 m2, while the range for plants harvested twice per year was 37 to 42 m2 (Mukhtar et al., 2003) The latter is similar to ratoon tiller number for elephantgrass in the current study. Ener gycane had more tillers

PAGE 124

124 than elephantgrass at three of four sampling dates in 2010, but no entry differences were detected in 2011 due to the large standard errors (Fig. 52). In 2011 ratoon growth, only the sampling date effect was significant ( P < 0.001) Especially for elephantgrass there was a much larger decrease in tiller number from beginning to end of the ratoon growth compared with the first growth. This was related to poor viability of emerging tiller s in the ratoon growth. Tiller Mass For full season growth, there was entry season interaction ( P < 0.001) in 2010, but in 2011 there were entry and season effects ( P < 0.001 for both) and only a trend toward entry x season interaction ( P = 0.097). At all 2010 sampling dates except the first of the season, energycane tiller mass was less than that of the two elephantgrass (Fig. 5 3). At four sampling dates during 2010 there were differences between elephantgrasses and in each case UF 1 had greater tiller mass than Merkeron. Entry UF 1 continued to accumulate tiller mass until 23 Oct. 2010 while Merkeron and energycane tiller mass did not increase beyond 22 September. The greater period of mass accumulation for UF 1 tillers may be a function of its later f lowering date (~ 10 December vs. 10 November for Merkeron) (Sollenberger et al., 2011) Previous comparisons have shown that UF 1 had greater tiller mass than Merkeron during 2 yr in two locations in Florida, and this was associated with greater stem diameter and individual leaf area for UF 1 relative to Merkeron (Sollenberger et al., 2011) The decrease in tiller mass between November and December was associated with a freeze event on 2 Dec. 2010 ( 2C). In 2011, the elephantgrasses had greater til ler mass than energycane throughout the entire season, but the elephantgrasses did not differ

PAGE 125

125 consistently (Fig. 5 3). There was no clear difference among entries when tiller mass accumulation ceased in 2011. First growth tiller mass was affected by entry x sampling date interaction ( P < 0.001) in both 2010 and 2011, while ratoongrowth tiller mass was affected only by entry ( P = 0.009) and sampling date ( P < 0.001) main effects in 2010 and entry x sampling date interaction ( P < 0.007) in 2011. First growt h tiller mass was greater for UF 1 than Merkeron and greater for Merkeron than energycane in June and July 2010 and July 2011 (Fig. 54). Ratoon growth tiller mass was consistently greater for UF 1 than energycane in 2010 and was greater for the elephantgr asses than energycane at three of four sampling dates in 2011 (Fig. 54). In both years tiller mass was consistently greater for first growth than ratoongrowth. Previous research has shown that increasing cutting frequency may decrease average tiller mass For example, Mukhtar et al. (2003) reported that tiller mass of Merkeron elephantgrass was 8.9 g following a single harvest but only 1.8 g following a third harvest at a 60d cutting interval. Moreover, there was a negative correlation between elephantgr ass tiller density and tiller mass (Mukhtar et al., 2003) This pattern was observed in the current study as UF 1 generally had greatest tiller mass but least tiller density, and energycane often had the most tillers but their mass was least. In another study, the range in elephantgrass tiller mass was 39 to 51 g with a 90d cutting interval, but it was 11 to 16 g with a 60d cutting interval (Wadi et al., 2004) This research also showed that although rapid ratoon growth is possible with elephantgrass, a longer regrowth period may aid in maximizing crop growth rate (CGR); the range in

PAGE 126

126 CGR was 10.6 to 13.4 g m2 d1 with a 90d cutting interval vs. 6.7 to 8.6 g m2 d1 with a 60d interval. Leaf Area Index Leaf area index in full season plots was greater early in the season for Merkeron and UF 1 elephantgrasses than for energycane; however, as the season progressed, differences between species became smaller and disappeared by August in both years (Fig. 5 5) This response is similar to that observed in previous research; elephantgrass showed greater LAI than energycane in early season but by 125 d after mowing on 28 March there were no species differences (Woodard et al., 1993) Maximum LAI was between 6 and 7 for elephantgrasses each year, and this level was reached by the end of June 2010 and midJuly 2011. Similarly, maximum LAI for energycane and elephantgrass was 6.9 and 7.1, respectively (Woodard et al., 1993) In current research, energycane reached its maximum approximately 30 d after elephantgrass in each year. Energycane reached its maximum 28 to 56 d later than elephantgrass in previous Florida research (Woodard et al., 1993) Other studies have reported elephantgrass LAI of 10 or greater (Ito and Inanaga, 1988; Ishii et al., 2005) but it appears that thes e measurements of light intensity were made at soil level. This leads to an overestimation of green leaf LAI because dead leaf mass begins to accumulate approximately 100 d after the previous defoliation and increases thereafter (Woodard et al., 1993) In the current study, LAI decreased due to leaf senescence starting in September 2010 and August 2011. This is comparable to the time when dead leaf number per tiller exceeded live leaf number (~125150 d after m owing on 28 March) in a study conducted in northern Florida (Woodard et al., 1993) The decline in LAI was earlier and more greatly accentuated for elephantgrass in 2011

PAGE 127

127 due to a severe thunderstorm with locally high wind gusts that caused significant plant damage on 12 Aug. 2011. Energycane was less susceptible to storm damage, likely due to markedly lower tiller mass at the time of the storm than for the two elephantgrass entries (Fig. 53), and the decl ine in LAI for energycane began later in the 2011 season. At the end of the 2010 growing season, the first freeze was severe and killed most leaves and caused a large decline in LAI from November to December sampling dates. In southern Japan, elephantgrass live leaf number decreased from September forward and ratio of live leaf number to whole leaf number decreased from June forward at a latitude (32 N) comparable to southern Georgia (Wadi et al., 2003) In Florida, live leaf number of both elephantgrass and energycane remained nearly constant throughout the season, but dead leaf n umber increased starting in June and exceeded live leaf number by the end of July for elephantgrass and end of August for energycane. These combined effects associated with leaf senescence likely resulted in the seasonal decline in LAI observed in the curr ent study. Initial growth of the first growth/ratoon treatment followed a similar pattern as full season growth, with elephantgrass entries achieving greater LAI. This difference was sustained during the entire period leading up to the 21 July defoliation event in 2011 and to within less than 28 d of the 30 July defoliation in 2010 (Fig. 56). Defoliation reset LAI to similar levels for all entries and there were no subsequent entry effects in 2010 ( P = 0.185) (Fig.56); however, in 2011, there were differ ences among entries ( P = 0.007) with Merkeron generally having greatest LAI. It is possible that the cumulative effects of defoliation resulted in the more pronounced first growth entry effect in 2011 vs. 2010 and in the presence of an entry effect in ratoon growth in 2011 vs. no effect in 2010.

PAGE 128

128 Refoliation of Entry UF 1 in ratoongrowth of 2011 also appeared to be slowed by cumulative effects of multiple harvests in 2 yr. Canopy Height In both 2010 and 2011, there was entry sampling date interaction ( P < 0.001) for the full season treatment. In 2010, UF 1 reached a height greater than 4 m and was taller than Merkeron at all sampling dates, but it was only occasionally taller than energycane (Fig. 57). In 2011, maximum canopy height was slightly less than in 2010, but UF 1 was taller than at least one of the other entries from beginning of sampling in May through September (Fig. 58). Unlike 2010 when height continued to increase for all entries until late September to early November, in 2011 UF 1 height remained relatively constant after the 17 July sampling. This was due to the effects of the severe thunderstorm described earlier. Three out of four replicates of Merkeron and UF 1 were seriously affected by the 12 August storm that preceded sampling on 19 August. Specifically, tillers lodged due to high winds and were laid over on each other. None of the L791002 plots were damaged by the storm, likely due in part to lesser tiller mass for energycane than for the two elephantgrasses (Fig. 53). Tiller numb er of the elephantgrasses remained essentially constant through the remainder of the season following the storm (Fig. 51), so although laid over to some extent by the wind they recovered sufficiently to remain viable until the end of the growing season. In an earlier study in Florida (Woodard and Prine, 1993a) the shape of the canopy height response curve to days after mowing was similar for elephantgrass plant introduction (PI) 300086 and breeding line N51 to those observed for Merkeron and UF 1 in this study. Those authors also reported maximum elephantgrass height of between 4 and 5 m and that energycane height was generally slightly shorter than the

PAGE 129

129 elephantgrasses. Sollenberger et al. (2011) reported UF 1 was 0.46 to 0.63 m taller in September than Merkeron at two locations in Florida during 2 yr of evaluation, but energycane was not included in th at experiment For the first growth/ratoon growth treatment, first growth height of UF 1 reached 3 m or taller by harvest in July and was taller than one or both of the other entries for all sampling dates in 2010 and three of four sampling dates in 2011 (Fig. 5 8). Height of ratoon growth was greater for UF 1 across sampling dates in 2010, but in 2011 there were no difference among grass entries. Height of ratoon growth generally was 2 to 2.5 m, approximately the height of first growth in late June. Unli ke LAI, for which UF 1 and energycane showed lessened regrowth response following July harvest, ratoon canopy height response was quite similar for UF 1 and energycane in both years (Fig. 58). Stem Proportion Stem proportion of full season total tiller m ass was affected by entry and sampling date ( P < 0.001 for both) in 2010 and entry x sampling date interaction in 2011 ( P < 0.001). In 2010, stem proportion increased until September and changed relatively little thereafter; throughout the season it was gr eater for UF 1 than Merkeron or energycane (Fig. 59). Stem proportion increased throughout the entire 2011 season for all entries, and energycane stem proportion was least among entries at all sampling dates (Fig. 59). Interaction occurred because Merker on tended to have greatest stem proportion early in the year, but UF 1 showed greatest stem proportion in late season. Plant part proportion is important because it affects chemical composition of biomass. The reduction in switchgrass ( Panicum virgatum L.) ash and mineral concentration was attributed to increasing proportions of stem relative to leaf as harvest was delayed into the winter (Sokhansanj et al., 2009) With elephantgrass in Florida,

PAGE 130

130 increasing plant maturity in fall resulted not only in decreased L AI but also increased stem proportion (Woodard et al., 1993) High stem proportion can be achieved due to stem growth or it also occurs due to leaf senescence and associated decrease in leaf biomass. In Japan, elephantgrass leaf dry weight increased until September after which it decreased, but stem dry weight increased linearly throughout the season (Wadi et al., 2003) In Florida, green leaf biomass of several tall elephantgrass es was nearly constant at 5 Mg ha1 throughout the growing season while deaf leaf biomass increased from mid June to September and was relatively constant thereafter at the same mass as green leaf (Woodard et al., 1993) In the same study, stem biomass increased linearly throughout the growing season and stem accumulation was the main driver in sustained biomass accumulation of both elephantgrass and energycane (Woodard et al., 1993) This pattern of response was observed in 2011 in the current study and until October in 2010. First growth stem proportion was affected by entry and sampling dates ( P < 0.001 for both) in both years, but there was no interaction. Elephantgrass UF 1 showed the highest overall stem proportion for first growth (Fig. 5 10). Ratoongrowth was aff ected by entry and sampling date ( P < 0.001 for both) in 2010, but there was a trend toward entry x sampling interaction ( P = 0.076). In 2011, there was entry x sampling date interaction ( P < 0.001). The trend toward interaction in 2010 and the interaction in 2011 occurred because energycane was slow to refoliate following July defoliation and thus in August had greater stem proportion than the elephantgrasses. Thereafter in both years, either UF 1 or Merkeron elephantgrass generally had greater stem proportion

PAGE 131

131 than energycane, and stem proportion in ratoon growth was approximately 50% of total biomass compared with 70 to 80% for full season growth. Biomass Dry Matter Concentration Biomass DM c oncentration is an important determinant of feedstock quality because it affects transportation cost and my affect conversion processes (Adler et al., 2006; Sokhansanj et al., 2009) There was entry x sampling date interaction for DM concentration of full season growth in both 2010 ( P = 0.047) and 2011 ( P = 0.002) (Fig. 5 11). In early June 2010, energycane had greater DM concentration than the elephantgrass entries, but there were no differences during the remainder of the 2010 growing season. In 2011, there were differences in DM concentration starting in late June and continuing through the remainder of the season, with DM concentrations greater in Merkeron than energycane during this entire period and UF 1 generally was inter mediate. Regardless of entry, DM concentration increased throughout the growing period to approximately 300 to 350 mg g1 by the end of the season. In Louisiana (Legendre and Burner, 1995) energycane DM concentration averaged 368 mg g1 with a range of 248 to 447 mg g1 by December. In the current study, DM concentration did not increase to a great degree after midOctober, so delaying harvest to after first freeze in December may not improve biomass quality from the perspective of DM concentration. This response, however, is likely to be plant sp ecies and environment dependent (Chapter s 3 and 7). Greater DM concentration or a trend toward this response by energycane early in the season likely reflects delayed early season growth of energycane relative to elephantgrass during this period of time (F ig. 5 5) (Woodard et al., 1993) Greater DM concentration in elephantgrass in 2011 may be associated with the generally greater proportion of stem in elephantgrass than

PAGE 132

132 energycane in that year (Fig. 59). Diff erences in stem anatomy between elephantgrass and energycane may also affect DM concentration differences in late season. Previous work has shown that elephantgrass stem has aerenchyma tissue which makes the stem more hollow (Jennewein et al., 2012) and likely faster to dry whereas in energycane the stem is not hollow (Griffee, 2000) For first growth, there was entry x sampling date interaction for DM conc entration in both years ( P < 0.001) (Fig. 512). For ratoonregrowth in 2010 there was only a sampling date ( P < 0.001) effect, but in 2011 there was an entry x sampling date interaction ( P = 0.033). Similar to that described above for full season growth, energycane biomass had greater DM concentration at the first sampling date (June 2010 and May 2011) in both years and at the first sampling date for the ratoon in one year (August 2011). Dry matter concentration of first and ratoongrowth within the context of a two harvest per year system reached a maximum of approximately 200 to 250 mg g1, or approximately 100 mg g1 less than end of year full season growth. Previous research has identified harvest frequency as an important determinant of biomass DM concentration. When elephantgrass was harvested every 30, 40, and 60 d in Puerto Rico, DM concentration increased from 144 to 145 and then to 180 mg g1, respectively (Velez Santiago and ArroyoAguilu, 1981) In Japan, harvesting elephantgrass at 60vs. 90 d intervals resulted in DM concentrations of 143 and183 mg g1 ( Wadi et al., 2004) Hybrids between elephantgrass and pearlmillet [ Pennisetum glaucum (L.) R.Br.] were harvested every 6 and 12 wk in Florida, and DM concentration was 187 and 228 mg g1, respectively (Spitaleri et al., 1994) For several Pennisetum species harvested one, two, or three times per year in Florida, DM concentrations were

PAGE 133

133 reported to range from 327 to 372, 249 to 263, and 189 to 202 mg g1, respectively (Woodard et al., 1991a) Thus data from the current experiment fit within the context of previously reported results. Implications of Research Merkeron and UF 1 eleph antgrasses generally showed similar seasonal patterns in morphological characteristics. There were much more pronounced differences in seasonal morphological characteristics between elephantgrass and energycane. Relative to energycane, the elephantgrasses generally had 1) consistent tiller number throughout the season vs. greater tiller number early in the season for energycane, 2) greater tiller mass and earlier development of LAI than energycane, and 3) greater stem proportion and biomass dry matter conce ntration than energycane. Elephantgrass UF 1 showed desirable characteristics related to biomass yield. Specifically, tiller mass and canopy height increased until late season, likely associated with later initiation of flowering, and UF 1 generally had the greatest proportion of stem in biomass, an advantage in terms of transportation costs and also lower concentrations of N and ash that can negatively affect some conversion processes at the biorefinery. Lack of difference in biomass yield due to harvest treatment in the first 2 yr of the experiment (Chapter 3), i.e., the years when morphological characteristics were being quantified, may be due to ratoon growth quickly achieving similar LAI following harvest to that of nondefoliated, fullseason growth. This occurred because as ratoon growth was accumulating leaf area, LAI of full season growth was decreasing markedly. There was evidence of less rapid refoliation during ratoon growth in Year 2 than Year 1, and this may be an early indicator of the negativ e effects of the 2X treatment on plant vigor that were not apparent in terms of biomass harvested until Year 3 (Chapter 3). These data

PAGE 134

134 provide evidence of the value of detailed characterization of morphological responses in understanding plant biomass accumulation and responses to defoliation.

PAGE 135

135 Table 51. Energycane and elephantgrass sampling dates for responses reported in Chapter 5. Full season and ratoon growth were sampled on the same day. Year Sampling dates 2010 --6 June 30 June 28 July 23 Aug. 22 Sept. 23 Oct. 8 Nov. 8 Dec. 2011 3 May 1 June 30 June 17 July 19 Aug. 16 Sep. 11 Oct. 10 Nov. 13 Dec.

PAGE 136

136 Number of tillers m-2 010203040506070 L79-1002 Merkeron UF-1 **************** Month May Jun Jul Aug Sep Oct Nov Dec Number of tillers m-2 010203040506070 **** Figure 51. Seasonal changes in tiller density of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were < 0.001 for all in 2010; and 0.177, < 0.001, and < 0.001, respectively, in 2011. Bars show mean one standard error. Entries within a date: *, P P 01.

PAGE 137

137 Number of tillers m-2 010203040506070 L79-1002 Merkeron UF-1 ratoon L79-1002 ratoon Merkeron ratoon UF-1 *********** Month May Jun Jul Aug Sep Oct Nov Dec Number of tillers m-2 010203040506070 ** Figure 52. Seasonal changes in tiller density of first and ratoongrowth for three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values for first growth were 0.001, < 0.001, and 0.001, respectively, in 2010; and 0.017, < 0.001, and < 0.001, respectively, in 2011. P values for ratoongrowth were 0.004, < 0.001, and 0.001, respectively, in 2010; and 0.199, < 0.001, and 0.448, respectively, in 2011. Bars show mean one standard error. Entries within a date: *, P P

PAGE 138

138 Tiller mass (g) 050100150200250300 L79-1002 Merkeron UF-1 ************* Month May Jun Jul Aug Sep Oct Nov Dec Tiller mass (g) 050100150200250300 Figure 53. Seasonal changes in tiller mass of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were < 0.001 for all effects in 2010; and < 0.001, < 0.001, and 0.097, respectively, in 2011. Bars show mean one standard error. Entries within a date: *, P P

PAGE 139

139 Tiller mass (g) 050100150200250300 L79-1002 Merkeron UF-1 ratoon L79-1002 ratoon Merkeron ratoon UF-1 **** Month May Jun Jul Aug Sep Oct Nov Dec Tiller mass (g) 050100150200250300 ******** Figure 54. Seasonal changes in tiller mass of first and ratoon growth for three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values for first growth were < 0.001 for all effects in 2010; and < 0.001, < 0.001, and 0.001, respectively, in 2011. P values for ratoongrowth were 0.009, < 0.001, and 0.225, respectively, in 2010; and < 0.001, < 0.001, and 0.007, respectively, in 2011. Bars show mean one standard error. Entries within a date: *, P P

PAGE 140

140 Leaf area index 0 1 2 3 4 5 6 7 L79-1002 Merkeron UF-1 ** ** Month May Jun Jul Aug Sep Oct Nov Dec Leaf area index 0 1 2 3 4 5 6 7 ** ** ** ** ** Figure 55. Seasonal changes in leaf area index of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were 0.021, < 0.001, and < 0.001, respectively, in 2010; and < 0.001 for all effects in 2011. Bars show mean one standard error. Entries within a date: **, P

PAGE 141

141 Leaf area index 01234567 L79-1002 Merkeron UF-1 ratoon L79-1002 ratoon Merkeron ratoon UF-1 Month May Jun Jul Aug Sep Oct Nov Dec Leaf area index 01234567 Figure 56. Season al changes in leaf area index of first and ratoongrowth for three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values for first growth were < 0.001, < 0.001, and 0.325, respectively, in 2010; and < 0.001, < 0.001, and 0.929, respectively, in 2011. The P values for ratoongrowth were 0.185, < 0.001, and 0.629, respectively, in 2010; and 0.007, < 0.001, and 0.116, respectively, in 2011. Bars show mean one standard error.

PAGE 142

142 Height (m) 012345 L79-1002 Merkeron UF-1 ************** Month May Jun Jul Aug Sep Oct Nov Dec Height (m) 012345 ********** Figure 57. Seasonal changes in canopy height of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were < 0.001 for all effects in 2010; and 0.005, < 0.00 1, and < 0.001, respectively, in 2011. Bars show mean one standard error. Entries within a date: *, P P

PAGE 143

143 Height (m) 012345 L79-1002 Merkeron UF-1 ratoon L79-1002 ratoon Merkeron ratoon UF-1 ***** Month May Jun Jul Aug Sep Oct Nov Dec Height (m) 012345 ****** Figure 58. Seasonal changes in canopy height of first and ratoongrowth for three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values for first growth were < 0.001 for all effects in 2010; and 0.002, < 0.001, and 0.027, respectively, in 2011. The P values for ratoongrowth were 0.001, < 0.001, and 0.418, respectively, in 2010; and 0.221, < 0.001, and 0.107, respectively, in 2011. Bars show mean one standard error. Entries within a date: *, P P

PAGE 144

144 Stem proportion (%) 0 10 20 30 40 50 60 70 80 90 100 L79-1002 Merkeron UF-1 Month May Jun Jul Aug Sep Oct Nov Dec Stem proportion (%) 0 10 20 30 40 50 60 70 80 90 100 ** ** ** ** ** ** ** ** ** Figure 59. Seasonal changes in stem proportion of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were < 0.001, < 0.001, and 0.075, respectively, in 2010; and < 0.001 for all effects in 2011. Bars show mean one standard error. Entries within a date: **, P

PAGE 145

145 Stem proportion (%) 0102030405060708090100 L79-1002 Merkeron UF-1 ratoon L79-1002 ratoon Merkeron ratoon UF-1 Month May Jun Jul Aug Sep Oct Nov Dec Stem proportion (%) 0102030405060708090100 ** Figure 510. Seasonal changes in stem proportion of first and ratoon growth for three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values for first growth were < 0.001, < 0.001, and 0.567, respectively, in 2010; and < 0.001, < 0.001, and 0.723, respectively, in 2011. The P values for ratoongrowth were < 0.001, <0.001, and 0.076, respectively, in 2010; and < 0.001 for all effects in 2011. Bars show mean one standard error. Entries within a date: **, P

PAGE 146

146 DM concentration mg g-1 0.00.10.20.30.4 L79-1002 Merkeron UF-1 Month May Jun Jul Aug Sep Oct Nov Dec DM concentration mg g-1 0.00.10.20.30.4 ************** Figure 511. Seasonal changes in dry matter concentrat ion of full season growth of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were 0.370, <0.001, and 0.047, respectively, in 2010; and < 0.001, < 0.001, and 0.002, respectively, in 2011. Bars show mean one standard error. Entries within a date: *, P P

PAGE 147

147 DM concentration mg g-1 0.00.10.20.30.4 L79-1002 Merkeron UF-1 ratoon L79-1002 ratoon Merkeron ratoon UF-1 **** Month May Jun Jul Aug Sep Oct Nov Dec DM concentration mg g-1 0.00.10.20.30.4 ******* Figure 512. Seasonal changes in dry matter concentration of first and ratoongrowth for three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values for first growth were 0.670, < 0.001, and < 0.001, respectively, in 2010; and 0.676, < 0.001, and 0.001, respectively, in 2011. The P values for ratoongrowth were 0.130, < 0.00 1, and 0.669, respectively, in 2010; and 0.015, < 0.001, and 0.033, respectively, in 2011. Bars show mean one standard error. Entries within a date: *, P 0.05; **, P

PAGE 148

148 CHAPTER 6 SEASONAL CHANGES IN CHEMICAL COMPOSITION OF ELEPHANTGRASS AND ENERGYCANE Overview of Research Dependence on imported fossil fuels has resulted in political and economic challenges for the USA and is associated with significant environmental concerns (Parrish and Fike, 2005; Kim and Day, 2011) Lignocellulose from grasses represents a potential celluloseto liquid fuel bioenergy production system (Anderson and Akin, 2008; Carroll and Somerville, 2009) Warm season perennial gras ses (C4) are a promising source of lignocellulose for conversion to biofuel in the Southeast USA (Knoll et al., 2012) Several candidate grasses have been evaluated and shown to have merit. In humid regions of the subtropics and tropics, elephant grass ( Pennisetum purpureum Schum.), or napiergrass, is known for its high biomass production (Woodard and Prine, 1993a; Morais et al., 2012) Energycane ( Saccharum spp. hybrid) is another candidate bioenergy grass that is characterized by high fiber conc entration and biomass yield, drought tolerance, and ratooning ability (Woodard and Prine, 1993b; Bischoff et al., 2008; Viator and Richard, 2012) While high levels of lignocellulose are desirable for biofuel production, high N and/or ash concentrations in biomass may reduce the efficiency of thermochemical conversion to fuel (Shahandeh et al., 2011) Thus, examination of chemical properties of biomass is important. In evaluating the compositional c haracteristics of grasses, there exists a considerable body of knowledge relative to their use as animal feed (Jung and Lamb, 2004; Anderson and Akin, 2008) For example, the detergent fiber analyses, which were originally proposed for forages (Van Soest et al., 1991) provide estimates of cell wall constituents including cellulose, hemicellulose, and lignin and can be useful

PAGE 149

149 as indicators of quality of cellulosic biom ass (Jung and Lamb, 2004; Guretzky et al., 2011) This is because plant cell wall is the primary energy source for ruminant microorganisms and in the ruminant animal these microbes face various barriers limiting access to structural carbohydrates that are similar to obstacles that occur in bioconversion to ethanol (Lorenz et al., 2009a) Further, it has been shown that cellulose and hemicellulose concentrations are correlated with theoretical ethanol yield potential (r = 0.91 and 0.51, respecti vely) (Lorenz et al., 2009b) For a low value crop such as perennial grasses for biofuel, it is essential to minimize N input and increase efficiency of N cycling (Erisman et al., 2010; Erickson et al., 2012) Previous research has shown that concentrations of N, P, and K in harvested plant material are dependent on harvest date in part because perennial grasses remobilize nutrients from above to belowground across the growing season (Adler et al., 2006; Heaton et al., 2009; Kering et al., 2012) Changes in plant part proportion as the season progresses, specifically increasing amount of stem (Chapter 5), al so contribute to changes in nutrient concentration in harvested biomass. Because N is considered an anti quality component for some biomass conversion platforms, understanding these changes in plant part composition can be quite important in terms of harvest timing or whether leaf should be included in harvested biomass. In addition, leaf may be considered of lesser value than stem in biofuel production system because its bulk density is lower than stem, increasing transportation cost (Sokhansanj et al., 2009; Brechbill et al., 2011) Most biomass composition data for perennial grasses are from experiments harvested once at the end of the growing season. Thus, there are limit ed data available

PAGE 150

150 that describe the compositional changes of perennial grasses throughout the growing season. This information, along with seasonal patterns of biomass accumulation, is valuable for identifying optimum harvest dat es and frequencies. Thus, t he objectives of this experiment were to quantify seasonal changes in i) fiber and nutrient composition of elephantgrass and energycane biomass and ii) leaf proportion and the effect of inclusion of leaf on composition of harvested biomass. Materials and Methods Data reported in this chapter came from the experiment that was described in Chapter 5. Thus, detailed information about the experimental site, weather conditions, plot establishment and management, and statistical analyses will not be reiterated h ere. In addition, composition analysis methodology was described previously in Chapter 4. Treatments and Experimental Design The experiment was conducted during 2010 and 2011 at the Plant Science Research and Education Unit (PSREU) at Citra, FL ( 29.41 N, 82.17 W). The soil was a well drained Candler sand (hyperthermic, uncoated Lamellic Quartzipsamments) Treatments were three grass entries replicated four times in randomized complete block design. The grass entries included two elephantgrasses, Merkeron (Burton, 1989) and a breeding line referred to as UF 1, and L791002 energycane (Bischoff et al., 2008) Merkeron elephantgrass and L791 002 energycane were chosen because they are the most available cultivars of these two species. In addition, breeding line UF 1 elephantgrass was included because in preliminary research it had demonstrated outstanding potential for use in bioenergy feedstock production (Sollenberger et al., 2011) and potential exists for it to be released as a cultivar.

PAGE 151

151 Plots used in this stud y were a subset of those from the experiment described in Chapter 3. Sampling occurred on plots that were harvested once per year within 1 wk following the first freeze event in fall (this treatment was referred to as 1X Dec in Chapter 3). Unlike Chapter 5, defoliation treatment (full season vs. first growth/ratoon growth) was not compared in this chapter. Response Variables Tiller sampling occurred at approximately 8wk intervals with the exception of the final collection (Table 6 1). Final sampling date varied between years because it was targeted for the end of the growing season, defined as within 1 wk of the first freeze event in fall (freeze was defined as occurrence of 0C recorded at a 2m height above soil level). The final sampling occurred 4 wk (2010) and 5 wk (2011) after the previous collection date. At each sampling date, plant material for compositional analysis was collected. Four representative tillers from each plot were cut at a 12cm stubble height and handseparated into leaf (blade plus sheath) and stem (including inflorescence, if present) components. The samples were dried at 60C until constant weight and leaf and stem data were used to calculate leaf proportion in total biomass. Dried samples were ground first in a hammer mill to re duce particle size and subsamples were ground through a Wiley mill (1 mm screen) before compositional analyses were conducted. Leaf and stem components of dried biomass were analyzed for composition. The modified method of detergent fiber analysis (Van Soest et al., 1991) was used for sequential neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) analyses. The samples were sequentially analyzed for NDF and then ADF using the ANKOM fiber analyzer (ANKOM 2000 Fiber Analyzer, ANKOM Technology

PAGE 152

152 Corporation, Fairport, NY). The ADF residue was used for analysis of ADL following the procedure Method for Determining Acid Detergent Lignin in Beakers (proposed by ANKOM Technology C orporation). Cellulose was estimated as the difference between ADF and ADL concentrations, and hemicellulose was estimated as the difference between NDF and ADF concentrations (Jung and Lamb, 2004; Waramit et al., 2011) Data reported are for total harvested biomass and were calculated using leaf and stem composition data and weighting the calculation based on the relative proportion of each plant part in total biomass. As described in Chapter 4, total N and P concentrations in the leaf and stem were determined using a standardKjeldahl method, a modification of the aluminum block digestion procedure (Gallaher et al., 1975) followed by semi automated colorimetric determination (Hambleton, 1977) Digestions were conducted at the Forage Evaluation Support Laboratory of the University of Florida. Samples were ashed using a muffle furnace for a minimum of 6 hr at 500C. Statistical Analysis Data were analyzed using mixedmodel methods (SAS Institute, 2008) In all mod els, grass entry was considered a fixed effect. Sampling date (Table 61) was considered a repeated measurement (fixed) within each year, and data were analyzed by year because 2010 was the establishment year and growth was delayed relative to 2011 resulti ng in later onset of sampling in 2010. Block was considered a random effect. Means were compared using the pdiff test of LSMEANS. All means reported in the text are least squares means and were considered different if P Trends were noted when P 0.10 and > 0.05.

PAGE 153

153 Results and Discussion Neutral Detergent Fiber There were entry x sampling date interactions for NDF concentration in both years ( P < 0.001 in 2010 and P = 0.009 in 2011; Fig. 61). The seasonal pattern of responses differed for the two g rass species in 2010. In June 2010, L791002 had greater NDF concentration than Merkeron and UF 1 (709, 656, and 659 mg g1, respectively, Fig. 61). However, NDF concentration of L791002 was lowest at the last two sampling dates of the same year. Energyc ane approached its peak NDF concentration in July and decreased through the remainder of the 2010 season. Unlike energycane, elephantgrass NDF increased throughout the growing season, although the rate of increase was slow starting in September. In 2011, energycane NDF concentration peaked later in the year than in 2010, September vs. July and then declined. Merkeron NDF concentration was greater than the others at the beginning (June) of the season, increased throughout the year, and was greatest in Decem ber as well (Fig. 6 1). Neither elephantgrass had a large increase in NDF after September. Seasonal trends of NDF concentration in switchgrass ( Panicum virgatum L.) were similar; NDF initially increased and then it remained nearly constant at around 850 mg g1 from August through the rest of the season (Madakadze et al., 1999) For Merkeron, increasing NDF concentration with increasing maturity was similar to the response observed in previous forage harvest frequency studies (Manyawu et al., 2003; Van Man and Wiktorsson, 2003) ; by increasing interval between cutting events in those studies, NDF concentration increased linearly from 704 to 785 mg g1 and from 636 to 753 mg g1, respectively. Unlike Merkeron, NDF in L791002 peaked in July 2010 and September 2011 and declined thereafter. This response has been associated with

PAGE 154

154 relatively greater nonstructural sugar concentration in energycane than in Merkeron at late sampling dates (Chapter 4). Illustrating this response, Bischoff et al. (2008) reported that L79100 2 accumulated substantial amounts of soluble solids (104 mg g1 Brix with 80 mg g1 sucrose concentration) in fresh cane at Louisiana during the harvest period. Similarly, energycane at 6 mo of growth in March in Brazil had Brix of 160 mg g1, and Brix inc reased linearly until it peaked at 220 mg g1 in September (Waclawovsky et al., 2010) In Florida at the end of growing season, L791002 energycane had 13 to 63 mg g1 greater extractives concentration than Merkeron elephantgrass (Fedenko, 2011) Water soluble carbohydrate (WSC) concentration was 2.2fold greater for L791002 than Merkeron (Woodard et al., 1991b) These results were observed even th ough energycane has less sucrose than commercial sugarcane; the amount of sucrose is sufficiently high, however, to result in relatively high concentration of extractives and low cell wall fiber concentration when compared with elephantgrass, especially late in the growing season. Acid Detergent Fiber There was entry x sampling date interaction in 2010 ( P < 0.001; Fig. 62). In 2011 there was no interaction ( P = 0.146), but there were entry and sampling date main effects ( P < 0.001 for both). In June 2010, L79 1002 had the greatest ADF; however, ADF of the two elephantgrasses was greater by September and the difference between them and energycane increased as the season progressed (Fig. 62). In 2011, Merkeron showed greatest ADF concentration followed by U F 1 and then L791002 (seasonal average ADF of 480, 464, and 442 mg g1, respectively; Fig. 6 2). All three grasses had increasing ADF concentration through the September sampling date in

PAGE 155

155 both years; thereafter elephantgrass ADF remain relatively constant or increased while energycane ADF decreased slightly. The ADF analysis was developed as a pretreatment step prior to ADL measurement (Van Soest et al., 1991) thus cel lulose and lignin (an indication of lignified fiber) are the primary components of ADF (Jung and Lamb, 2004) Similar to t he response of NDF, elephantgrass ADF increased at the fastest rate during the early part of the season. This response has been reported previously in harvest frequency studies (Manyawu et al., 2003; Van Man and Wiktorss on, 2003) whereby ADF concentration inc reased linearly with increasing interval between harvests (360 to 398 mg g1 and 358 to 450 mg g1, respectively). The seasonal pattern of response observed in the current research, i.e. increasing ADF early in the season with little change thereafter, was also observed for switchgrass in eastern Canada. Switchgrass ADF concentration increased linearly until September and then remained nearly constant during the remainder of season (Madakadze et al., 1999) The overall seasonal pattern of ADF responses in the current study follows that observed for NDF concentration. Acid Detergent Lignin There was entry x sampling date interaction for ADL in total biomass in 2010 ( P < 0.00 1), but in 2011, there were only entry and sampling date main effects ( P < 0.001 for both, Fig. 6 3). The seasonal pattern of ADL in 2010 was very similar to ADF. Energycane L791002 showed greater ADL concentration in June; however, the two elephantgrass entries had greater ADL concentration in July (Fig. 6 3). Thereafter, the magnitude of the difference between energycane and elephantgrass increased (11 mg g1 in July vs. 19 mg g1 in December, Fig. 63), and ADL of both elephantgrasses was greater than e nergycane throughout the remainder of the season. During the second

PAGE 156

156 year, the pattern of response of the three grass entries was similar across sampling dates, and Merkeron had greatest ADL followed by UF 1 and L791002 (seasonal average ADL of 81, 71, and 61 mg g1, respectively, Fig. 6 3). Averaged across the three grasses, ADL increased in 2011 from 56 to 80 mg g1 in June and November and then decreased slightly to 76 mg g1 in December. These data suggest that once structural lignin reaches a maximum i n fall, further change in lignin concentration is likely to be small and if it occurs it is likely to be a function of increasing concentration of other constituents, e.g., nonstructural sugars in energycane, rather than change in lignin content. It is kn own that ADL underestimates true lignin concentration (Jung et al., 1999; Jung and Lamb, 2004; Dien et al., 2006) however, it has been reported that ADL has a sim ilar pattern of response as Klason lignin, the analysis thought to one of best representations of true lignin concentration in legumes and grasses (Jung et al., 1997; Jung and Lamb, 2004) Using ADL for relative comparisons of lignin concentration is thus considered valid for bioenergy research. Lignin is considered to be an anti quality component in forage, and ADL is strongly negatively correlated with in vitro dry matter digestibility and dry matter digestibil ity ( r = 0.96 for both) (Jung et al., 1997) In biomass for bioenergy, the s ituation is less clear cut. Recovered lignin can provide heat during a thermochemical conversion process, but on the other hand, since it binds to structural carbohydrates in plant cell wall, it reduces enzyme accessibility to cellulose and subsequent etha nol yield (Adler et al., 2006; Guretzky et al., 2011) Klason lignin concentration in switchgrass in Pennsylvania was 10 to 33% greater in the fall compared with the spring; however, the difference narrowed in the

PAGE 157

157 second year (182 vs. 165 mg g1) compared with the first year (173 vs. 130 mg g1) (Adler et al., 2006) Similarly, there was relatively less change from June through December in Year 2 of the current study (Fig. 63), but that was likely due to the fact that growth initiated earlier in Year 2 than in Year 1 (establishment year). As a result the effective maturity of plants was greater in June of Year 2 than Year 1. An important determinant of inc reasing lignin concentration with increasing maturity is leaf:stem ratio. As plants mature, stem proportion decreases (Chapter 5) and stem typically has greater lignin concentration than leaf. Measurement of switchgrass cell wall components in Texas showed that stem lignin concentration was greater than in leaf (Shahandeh et al., 2011) Average lignin concentration in big bluestem ( Andropogon gerardi Vitman) and switchgrass plant parts (leaf and stem) w as 48 vs. 64 mg g1 and 42 vs. 56 mg g1, respectively, in West Virginia (Griffin and Jung, 1983) Cellulose Cellulose concentration was affected by entry x sampling date interaction ( P < 0.001) in 2010, but in 2011 only entry and sampling date effects were significant ( P = 0.002, and P < 0.001, respectively; Fig. 64). During 2010, interaction was caused by relatively greater cellulose concentration in energycane in June but lesser concentration than elephantgrasses from September through the rest of the season (Fig. 64). Only Merkeron continued to increase in cellulose concentration through December 2010, while cellulose in energycane decreased after September. In UF 1, the responses remained essentially unchanged from November to December. Unlike the first year, entry by sampling date interaction did not occur in 2011; however, there was a trend toward interaction ( P = 0.061). The pattern of response in 2011 was quite similar to that in 2010 whereby Merkeron cellulose concentration continued to increase until the end of

PAGE 158

158 season, but for the other entries it remained constant (UF 1) or deceased slightly (L791002) at late season (Fig. 64). Averaged across sampling dates, Merkeron and UF 1 had greater cellulose concentration than L791002 (399, 393, and 380 mg g1, respectiv ely). For elephantgrass, the observation of increasing cellulose concentration with maturity was similar to previous defoliation frequency studies (Tessema et al., 2010) in which increasing defoliation interval from 60 to 120 d increased cellulose concentration from 297 to 339 mg g1, respectively. Lesser concentration of cellulose in lateseason energycane is likely due to increasing concentrations of nonstructural sugars (Chapter 4). In comparison with previous work, greater cellulose concentration than observed in the current study was reported in full season growth of energycane (433 mg g1) in Louisiana (Kim and Day, 2011) however, it was from bagasse after nonstructural sugars were removed. Another study reported that energycane had twofold greater sugars from extractives than elephantgrass had (Fedenko, 2011) A larger amount of glucose monomers, which compose cellulose, is advantageous for ethanol production because glucose can be converted at greater efficiency to ethanol than most ot her sugars, especially compared to pentose (Dien et al., 2006) Greater cellulose concentration and high yields of biomass (Chapter 3) during late season suggest that delayed harvest results in greater cellulose yields, but the concurrent increase in lignin concentration as plants mature may interfere with conversion of the cellulose to energy. Hemicellulose There was entry x sampling date interaction for hemicellulose concentration in to tal biomass in 2010 ( P < 0.001), but in 2011 there were entry and sampling date main effects only ( P = 0.002, and P < 0.001, respectively; Fig. 6 5). In June and July 2010,

PAGE 159

159 L791002 had greater hemicellulose concentration than the two elephantgrasses, but thereafter hemicellulose did not differ from either elephantgrass (Fig. 65). In 2011, L791002 had average hemicellulose concentration of 299 mg g1 with a range of 276 (December) to 313 mg g1 (June). The energycane seasonal mean was greater than that of Merkeron and UF 1 (277 and 271 mg g1, respectively; Fig. 6 5). Unlike cellulose concentration, in both years hemicellulose concentration decreased over the season although the shapes of the curve were different between years. It is likely that part of the reason for the different shape of curves is that biomass was more mature in June 2011 than in 2010, due to earlier initiation of growth in 2011. Thus hemicellulose in this more mature material had already declined from its maximum when sampling began in June 2011. Previous research with elephantgrass had shown that hemicellulose concentration decreased from 337 to 280 mg g1 as regrowth interval increased from 60 d to 120 d, respectively (Tessema et al., 2010) There are no seasonal response data reported for energycane, but bagasse from seasonlong growth had a hemicellulose concentration of 238 mg g1 in Louisiana (Kim and Day, 2011) not greatly different than season end concentrations of 154 to 276 mg g1 in the current study. Reported trends in hemicellulose concentration with increasing plant maturity are not consistent, likely due in part to methodological differences. Hemicellulose can be determined as either sum of structural carbohydrates except glucose or as NDF minus ADF (as done in current research). Methodological differences in hemicellulose concentration have been reported in switchgrass research in Nebraska (Dien et al., 2006) There was no clear explanation in the literature why hemicellulose concentration

PAGE 160

160 determined using the two methods shows opposite trends. Hemicellulose concentration determined by the Uppsala dietary fiber procedure increased from 235 to 279 mg g1 as ma turity increased from preboot stage to a post frost sampling, whereas the Van Soest technique showed that hemicellulose decreased slightly from 318 to 311 mg g1 (Dien et al., 2006) Nitrogen Total biomass N concentration was affected by sampling date ( P < 0.001) in both years and also by entry in 2011 ( P < 0.001, Fig. 66). Nitrogen concentration in biomass declined as the season progressed in both years, although the extent of the decline was much greater in 2010 (Fig. 66). In that year (2010), N concentration averaged across entries was 20.5 mg g1 in June and it decreased to 8.2 mg g1 in July. During the remainder of the season, N concentration varied little and was range of 4.6 to 5.5 mg g1. In 2011, overall N concentration in Merkeron was greater t han L79 1002 and UF 1 (7.5, 6.6, and 6.4 mg g1, respectively, Fig. 6 6). With the exception of the June sampling date, N concentration was very similar to 2010. The difference between years in June again is likely attributed to the earlier regrowth in 2011 such that the major decrease in biomass N concentration had already occurred by the time sampling began. Similar to results in the current research, switchgrass N in Iowa decreased from 12.4 mg g1 in June to 0.39 mg g1 in November (Wilson et al., 2013a) Energycane N concentration in Florida also was a function of plant height (maturity) and decreased quadratically as plant height increased (Mislevy et al., 1995). Similar to current research, once N concentration of sweet sorghum approached a minimum threshold in Florida, biomass N remained relatively constant (Erickson et al., 2012) Once switchgrass N concentration reached a minimum in Pennsylvania, delayed harvest

PAGE 161

161 management resulted in only minor (5 to 15%) variation between fall and spring harvests (Adler et al., 2006) Because N is considered to be of no added value for conversion to energy and leaf N concentration is typically much greater than stem N, there may be value in returning leaf biomass to the production field. This may be advantageous both in terms of biomass quality and N recycling back to soil. To explore these relationships, leaf percentage in biomass and proportion of total biomass N in the leaf fraction were evaluated. Leaf percentage in total biomass was affected by entry and sampling date main effects in 2010 ( P < 0.001) and the interaction of entry and sampling date in 2012 ( P < 0.001). In 2010, leaf percentage decreased from 56% in June to 19% in December and average leaf percentage over the season was greatest in L791002 and Merkeron (34 and 33%, respectively) and least (26%) in UF 1 (Table 62). In 2011, L791002 had the greatest leaf percentage at each sampling date and leaf percentage decreased 18.5 units from June through December compared with 17 units for Merkeron and 25 units for UF 1 (Table 63). Leaf percentage in UF 1 reached 13 in Decem ber 2010 and 15 in December 2011 (Tables 62 and 63). This result indicates that reduced N concentration in total biomass in late season was a function of both an overall decrease of biomass N concentration over the season and reduced leaf proportion the plant part which has greater N concentration than stem. Proportion of total N in leaf was affected only by sampling date in 2010 ( P < 0.001) and by entry ( P < 0.001) and sampling date ( P = 0.013) in 2011. In June 2010, approximately 63% of total N was in the leaf fraction and it decreased to 32% by December. Leaf proportion in total biomass averaged 56 and 19% for those two dates

PAGE 162

162 (Table 62). By late in the season, UF 1 leaf proportion was only 19 (November) and 13% (December) but 37 (November) and 27% (D ecember) of N in harvested biomass was in leaf. Thus the proportion of N in leaf was approximately twice as great as the proportion of biomass that was leaf. This result indicates that returning elephantgrass leaf to the production field during a single fall or early winter harvest would reduce total biomass yield by 13 to 22%, but it would reduce the amount of N in harvested biomass by 28 to 41%. Similar to the previous year, in 2011 approximately 54% of total N was in leaf in June and that number decrease d only to 46% by December (Table 63) In December, proportion of total N in leaf for L791002, Merkeron, and UF 1 was 66, 104, and 147% greater than the percentage leaf in total biomass. While grasses had 15 to 40% leaf during November and December, 36 to 64% of N was in the leaf (Table 63). Previous work with energycane in Florida showed that N concentration in stem decreased 81% across growing season, while the decrease in leaf was 51% (Mislevy et al., 1995) As result, the magnitude of the difference in N concentration between leaf and stem increased over season (Mislevy et al., 1995) Relatively low N concentration in stem compared to leaf was observed in a sorghum study in Florida (12 mg g1 in leaf vs. 4.4 mg g1 in stem) (Erickson et al., 2012) Previous research with miscanthus showed significant decreases in both leaf and stem N across the growing season. Live leaf N concentration was not significantly different than stem in June but the difference between fractions increased over season (Beale and Long, 1997) This study also showed increasing stem and decreasing leaf proportion over time; however, the proportion of to tal N in leaf was relatively constant at about one third of aboveground N

PAGE 163

163 in both miscanthus and cordgrass [ Spartina cynosuroides (L.) Roth ] (Beale and Long, 1 997) This result is within the range observed in the current research. Phosphorus There was entry x sampling date interaction in both years ( P = 0.006 in 2010 and P < 0.001 in 2011). In 2010, Merkeron biomass had greatest P concentration throughout th e season until the final sampling date (Fig. 67). In 2011, Merkeron again had greatest or tended to have greatest P concentration except for the June sampl ing data. Average P concentration across the year was 1.2 mg g1 for Merkeron and 1.0 and 0.9 mg g1, for L791002 and UF 1. Extent of seasonal change in P concentration was greater in 2010 than 2011. As with other response variables, this was likely due in large part to later tiller emergence in the establishment year of 2010. As a result the rapid earl y season decline in P concentration observed in 2010 had likely already occurred by the time sampling was initiated in 2011. Overall leaf and stem P concentrations averaged 1.20 and 1.30 mg g1, respectively. Previous research with switchgrass showed an average 49% decrease in biomass P concentration over winter (Adler et al., 2006) Miscanthus P concentration decreased from June to August in both green leaf and stem (Beale and Long, 1997) In that study, there was a large difference in P concentration among plant parts; green leaf P concentration was twofold greater than stem; however, dead leaf P concentration was lower than stem and likely offset overall differences between total leaf vs. stem as dead leaf accumulated late in the season (Chapter 3). Ash There was entry x sampling date interaction for biomass ash concentration in 2010 ( P < 0.001) but only a sampling date effect in 2011 ( P < 0.001). In June 2010, ash

PAGE 164

164 concentration in elephantgrasses was greater than energycane; no differences among elephantgrass entries occurred (Fig. 68). Ash concentration was greatest at the beginning of the season (average across entries of 77 mg g1 in June), but it decreased significantly through September (29 mg g1). Thereafter it remained relatively constant in 2010 (Fig. 68). In 2011, the seasonal trend in concentration was similar to the previous year, absent the large decrease in ash early in the season that likely was not measured in 2011 because of earlier tiller emergence than in the first year. In 2011, ash concentration decreased from 54 mg g1 in June to 35 mg g1 in September (Fig. 68). Inorganic constituents, commonly referred to ash, cannot be converted to energy, and greater ash concentration is considered to negatively affect biomass quality for conversion (Wilson et al., 2013a) However, inorganic constituents such as Si, K, Ca, S, and Cl play an essential role in plant growth and development (Bakker and Elbersen, 2005) If ash can be removed from harvested biomass there may be value to both conversion and s oil nutrient composition in production fields. With this in mind, data for seasonal changes in proportion of ash in leaf biomass will be presented. In both years there was entry x sampling date interaction for proportion of ash in leaf biomass ( P = 0.012 and P < 0.001, respectively; Tables 62 and 6 3). There were no differences among entries in proportion of total ash that was in leaf biomass in June and July 2010, but UF 1 had the lowest proportion (24%) by December (Table 62). In June 2010, the proportion of leaf in total biomass ranged from 49 to 62% across entries, but leaf contained only 41 to 47% of ash. By December, much had changed as the leaf percentage ranged from 13 to 22% but the percentage of total ash in leaf biomass was 24 to 34%. Thus, by the end of the season including leaf in

PAGE 165

165 harvested biomass had a greater impact on the amount of ash harvested than on the amount of biomass harvested, and one third to one fourth of ash harvested could be returned to the field if leaf was not harv ested. In 2011, there were no differences among entries in percentage of total ash that was present in leaf biomass from June through November, but in December there was a lesser proportion of ash in elephantgrass than energycane leaf (Table 63). Early in 2011, the percentage of total biomass that was leaf and the proportion of total ash that occurred in leaf were very similar for all entries, but by December leaf percentage in total biomass ranged only from 15 to 36% while percentage of total ash in leaf ranged from 28 to 43% (Table 63). Results show that generally UF 1 leaf proportion in total biomass was lowest in both years, and proportion of ash in leaf was lower in UF 1 (2010) or as low as (2011) that of Merkeron. Data for both N and ash indicate that returning leaf to the production field is a much more attractive option for elephantgrass than for energycane because elephantgrass leaf contains significant proportions of N and ash but elephantgrass leaf contributes less to total biomass harvested than energycane leaf. Likewise, among elephantgrasses, this practice appears to have greater potential for UF 1 because leaf proportion, especially at the end of the growing season, was less than for Merkeron while the proportion of ash or N in UF 1 leaf was still in the general range of 25 to 33%. Similar to current research, Summer et al. (2001) investigated total ash concentration in different plant parts of rice straw (leaf, stem, node, and panicle). They found that leaves contained 18 to 19% total ash, w hereas, stem contained only 12% of total ash. It has been suggested that C4 grass species have lower ash concentration compared with C3 plants (Bakker and Elbersen, 2005) Further, it was suggested that

PAGE 166

166 amount of ash in herbaceous biomass can be controlled by the timing of harvest (Bakker and Elbersen, 2005) with longer harvest intervals associated with lower ash concentration due to senescence and translocation of constituents. These results support those of the current study. Implications of Research Seasonal changes in chem ical composition were characterized for three candidate bioenergy grasses in Florida by mineral analyses and quantifying cell wall constituents by Van Soest fiber analyses. Grass entries included Merkeron and UF 1 elephantgrasses and L791002 energycane. D ifferent responses were detected in composition between elephantgrass and energycane. With the exception of hemicellulose, elephantgrass cell wall constituents increased from early in the growing season until late summer and either remained relatively cons tant (UF 1) or slightly increased (Merkeron) during the remainder of the growing season. In contrast, once energycane cell wall constituents peaked in late summer, concentration of cell wall constituents tended to decrease thereafter. Greater nonstructural sugar concentration in energycane than in Merkeron at late sampling dates (Chapter 4) was associated with the decrease in concentration of fiber components in energycane. Nitrogen, P, and ash concentrations decreased with increasing maturity for all grass entries. When considering individual plant parts, leaf N and ash concentrations were much greater than in stem. Additionally, leaf percentage in total biomass decreased more rapidly over the growing season than did the percentage of total N and ash that occurred in the leaf fraction. Because N and ash are considered to be anti quality components in biomass feedstocks, there may be value in returning leaf biomass to the production field or delaying harvest into winter to allow a greater natural return of

PAGE 167

167 leaf biomass to the field by abscission. These approaches may be advantageous both in terms of biomass quality and recycling of nutrients back to soil. Data for both N and as h indicate that returning leaf to the production field is a much more attractive option for elephantgrass than for energycane. This is because elephantgrass leaf represents a lesser proportion of total biomass relative to the proportion of total N and ash contained in the leaf fraction, whereas with energycane leaf percentage and percentage of total N and ash are more nearly the same. In particular, through a large late season decrease in leaf proportion (Chapter 5), UF 1 appears to reduce leaf percentage i n biomass harvested and this may contribute to its potential value as a biomass biofuel crop.

PAGE 168

168 Table 61. Sampling dates for energycane and elephantgrass for responses reported in Chapter 6. Year Sampling dates 2010 6 June 28 July 22 September 8 Novemb er 8 December 2011 1 June 17 July 16 September 10 November 13 December

PAGE 169

169 Table 62. Effect of grass entry or sampling date main effects or their interaction on leaf proportion in total biomass and proportion of total N and ash in the leaf fraction in 2010. Data are means across four replicates (n = 4). Sampling dates were repeated measurements (6 June, 28 July, 8 Nov., and 8 Dec.). Data not available for sampling date 22 September. Sampling date means within an entry and not followed by the same lower case letter are different ( P < 0.05). Entry means within a sampling date and not followed by the same upper case letter are different ( P < 0.05). Entry Sampling date 6 June 28 July 8 Nov. 8 Dec. Mean ------------------Leaf proportion in total biomass (%) ------------------L79 1002 62 29 25 22 34 A Merkeron 57 27 26 22 33 A UF 1 49 23 19 13 26 B mean 56 a 26 b 23 b 19 c -----------Proportion of total harvested N in leaf fraction (%) ----------L79 1002 68 51 38 35 48 Merkeron 63 50 41 33 47 UF 1 7 48 37 28 42 Mean 63 a 50 b 39 c 32 d ----------Proportion of total harvested ash in leaf fraction (%) ---------L79 1002 46 aA 33 bA 31 bB 33 bA Merkeron 47 aA 37 bcA 41 bA 34 cA UF 1 41 aA 32 bA 36 abAB 24 cB SE 2.5

PAGE 170

170 Table 63. Effect of grass entry or sampling date main effects or their interaction on leaf proportion in total biomass and proportion of total N and ash in the leaf fraction in 2011. Data are means across four replicates (n = 4). Sampling dates were repeated measurements (1 June, 17 July, 16 Sep., 10 Nov., and 13 Dec.). Sampling date means within an entry and not followed by the same lower case letter are different ( P < 0.05). Entry means within a sampling date and not followed by the same upper case letter are different ( P < 0.05). Entry Sampling date 1 June 17 July 16 Sep. 10 Nov. 13 Dec. Mean ---------------------Leaf proportion in total biomass (%) -------------------L79 1002 55 aA 45 bA 41 bcA 40 cdA 36 dA Merkeron 37 aB 36 aB 29 bB 26 bB 21 cB UF 1 40 aB 36 aB 27 bB 21 cC 15 dC SE 2.2 -----------Proportion of total harvested N in leaf fraction (%) ----------L79 1002 63 62 60 64 60 63 A Merkeron 48 48 45 47 42 48 B UF 1 51 55 45 44 36 51 B mean 54 ab 55 a 50 bc 51 ab 46 c ----------Proportion of total harvested ash in leaf fraction (%) ---------L79 1002 49 aA 41 bA 45 bA 42 bA 43 bA Merkeron 44 aA 43 aA 42 aA 40 aA 30 bB UF 1 45 aA 46 aA 42 aA 36 bA 28 cB SE 2.6

PAGE 171

171 Month May Jun Jul Aug Sep Oct Nov Dec NDF mg g-1 600 650 700 750 800 ** ** NDF mg g-1 600 650 700 750 800 L79-1002 Merkeron UF-1 ** ** ** Figure 61. Seasonal changes in NDF concentration of total harvested biomass of three perennial grass entr ies in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were < 0.001 for all in 2010; and < 0.001, < 0.001, and 0.009, respectively, in 2011. Bars show mean one standard error. Entries within a date, P P 0.01.

PAGE 172

172 ADF mg g-1 300 350 400 450 500 550 600 L79-1002 Merkeron UF-1 Month May Jun Jul Aug Sep Oct Nov Dec ADF mg g-1 300 350 400 450 500 550 600 ** ** ** ** Figure 62. Seasonal changes in ADF concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were < 0.001 for all in 2010; and < 0.001, < 0.001, and 0.146, respectively, in 2011. Bars show mean one standard error. Entries within a date, ** P

PAGE 173

173 ADL mg g-1 0 10 20 30 40 50 60 70 80 90 100 L79-1002 Merkeron UF-1 ** ** ** ** ** Month May Jun Jul Aug Sep Oct Nov Dec ADL mg g-1 0 10 20 30 40 50 60 70 80 90 100 Figure 63. Seasonal changes in ADL concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were < 0.001 for all in 2010; and < 0.001, < 0.001, and 0.720, respectively, in 2011. Bars show mean one standard er

PAGE 174

174 Cellulose mg g-1 300 350 400 450 L79-1002 Merkeron UF-1 ** ** ** Month May Jun Jul Aug Sep Oct Nov Dec Cellulose mg g-1 300 350 400 450 Figure 64. Seasonal changes in cellulose concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were < 0.001 for all in 2010; and 0.002, < 0.001, and 0.061, respectively, in 2011. Bars show mean one standard error. Entries within a date, P P

PAGE 175

175 Month May Jun Jul Aug Sep Oct Nov Dec Hemicellulose mg g-1 200220240260280300320340 Hemicellulose mg g-1 200220240260280300320340 L79-1002 Merkeron UF-1 ***** Figure 65. Seasonal changes in hemicellulose concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were < 0.001, < 0.001, and 0.001, respectively, in 2010; and 0.002, < 0.001, and 0.437, respectively, in 2011. Bars show mean one standard error. Entries within a date, P ** P

PAGE 176

176 Nitrogen mg g-1 0510152025 L79-1002 Merkeron UF-1 Month May Jun Jul Aug Sep Oct Nov Dec Nitrogen mg g-1 0510152025 Figure 66. Seasonal changes in nitrogen concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were 0.176, < 0.001, and 0.923 respectively, in 2010; and < 0.001, < 0.001, and 0.654, respectively, in 2011. Bars show mean one stan dard error.

PAGE 177

177 Phosphorus mg g-1 0.00.51.01.52.02.53.03.5 L79-1002 Merkeron UF-1 Month May Jun Jul Aug Sep Oct Nov Dec Phosphorus mg g-1 0.00.51.01.52.02.53.03.5 ********** Figure 67. Seasonal changes in phosphorus concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P values were 0.176, < 0.001, and 0.006 respectively, in 2010; and < 0.001, < 0.001, and 0.084, respectively, in 2011. Bars show mean one standard error. Entries within a date, ** P

PAGE 178

178 Ash mg g-1 0 10 20 30 40 50 60 70 80 90 L79-1002 Merkeron UF-1 ** Month May Jun Jul Aug Sep Oct Nov Dec Ash mg g-1 0 10 20 30 40 50 60 70 80 90 Figure 68. Seasonal changes in ash concentration of total harvested biomass of three perennial grass entries in 2010 (upper) and 2011 (lower). Entry, sampling date, and their interaction P value were 0.314, < 0.001, and < 0.001, respectively, in 2010; and 0.341, < 0.001, and 0.852, respectively, in 2011. Bars show mean one standard error. Entries within a date, ** P

PAGE 179

179 CHAPTER 7 TIME AFTER A FREEZE EVENT AFFECTS PERENNIAL GRASS BIOMASS HARVESTED AND CHEMICAL COMPOSITION Overview of Research Increasing demand for energy within the context of finite oil reserves has affected energy security of petroleum importing nation s including the USA (Sticklen, 2008) Currently, only biomass derived fuels offer renewable, liquid type alternatives to petroleum based transportation fuel s (Rollin et al., 2011) Warm season grasses have gained attention as bioenergy feedstock s for ethanol production in recent years because they can reduce reliance on import ed fossil fuel reduce greenhouse gas emission s and help rural econom ies (McLaughlin et al., 2002; Waramit et al., 2011) In previous research (Woodard and Prine, 1993a; Kering et al., 2012; Knoll et al., 2012; Viator and Richard, 2012) perennial grasses w ith the C4 photosynthetic system have demonstrated outstanding growth potential and biomass production One challenge to the use of w arm season perennial grasses as bioenergy feedstocks is the nonuniform availability of biomass throughout the year. The failure to match supply of biomass to the biorefinery with the goal of year round biofuel production may result in additional storage and handling cost s (Rentizelas et al., 2009) Harvest practices are needed that provide biomass of high quality to the biorefinery during nonpeak periods of plant growth. One such option is delaying harvest until winter. By delaying harvest, it may be possible to extend the time period when biomass is available (Finell et al., 2011) and delayed harvest may improve feedstock quality for direct combustion by reducing moisture concentration and leaching undesirable components such as Cl, K, and ash (Jrgensen and Sander, 1997; Lewandowski and Kic herer, 1997; Lewandowski and Heinz, 2003)

PAGE 180

180 D elayed harvest until winter has been reported to negatively affect perennial grass biomass harvested (Lewandowski and Heinz, 2003) For instance, switchgrass ( Panicum virgatum L. ) biomass ha rvested decreased throughout winter (10.0 Mg ha1 in Oct ober vs. 5.6 Mg ha1 in Apr il) (Wilson et al., 2013b) and the reduction was a ffected by winter weather especially snow in the Midwest USA Other authors have reported a reduction in both switchgrass biomass yiel d and moisture concentration during winter (Adler et al., 2006) These authors attributed 90% of the yield reduction due to lodging and subsequent difficult ies in gathering biomass during baling Unlike the Mid west, Florida has a mild winter (Smith and Dowd, 1981) that may affect biomass yield and chemical composition differently than in temperate regions R egional differences in severity and duration of winter may affect senescence, biomass yield and biomass quality of perennial grasses resulting in locally specific recommendations relative to delayed harvest. Due to the extensive research conducted on switchgrass, best management practices for harvest management have been developed (Mitchell and Schmer, 2012) However, little is known about the effect of delay ed harvest after first freeze on tallgrowing, warm season perennial grasses particularly in a subtropical environment where freeze event s occur only occasionally during winter. The objective of this experiment w as to quantify the effect of delayed harvest following first freeze on biomass harvested and composition of elephantgrass ( Pennisetum purpureum Schum.) and energycane ( Saccharum spp. hybrid) in Florida. Materials and Methods Experimental Site The experiment was conducted during the winter seasons of 20102011 ( Y ear 1), 20112012 ( Y ear 2), and 20122013 ( Y ear 3) at the Plant Science Research and

PAGE 181

181 Education Unit (PSREU) at Citra, FL ( 29.41 N, 82.17 W) at a site adjacent to the experiments reported in Chapter s 3 through 6. The soil was a well drained Candler sand (hyperthermic, uncoated Lamellic Quartzipsamments) Initial characterization of topsoil (020 cm) showed an average soil pH of 7.0, and Mehlich1 extractable P, K, Mg, and Ca of 54, 20, 123, and 496 mg kg1, respectively. Monthly weather conditions for the growing season (before the firs t freeze event occurred) were described previously (Chapter 3, Fig s. 3 1 and 32). For each winter season, weather data are reported for the period of delayed harvest starting with the date of first freeze event (Table 71 ) A freeze was defined as a temperature of 0C or less at 2 m above soil level that resulted in complete kill of the leaf canopy. Weather data reported are the weekly mean daily temperature and the weekly mean maximum and minimum temperatures (Fig. 7 1 ) as well as weekly total preci pitation (Figure 72) To summarize the occurrences of freezes, there were six events between the pre freeze harvest and first harvest after freeze, six events between the first and second harvest and two events between the second and third harvests 2010. In 2011, there were six freeze events between the pre freeze harvest and first harvest after freeze and two freezes between the first and second harves ts In 2012 there was only one freeze between the pre freeze harvest and first harvest after freeze and five freezes between the first and second harvests. Plot Establishment and Management until a Freeze Event Plots contained six rows of 6m length, with 1 m spacing between rows. Plots were established using aboveground stem pieces planted on 15 Dec. 2009. Thus, the Year 1 data ( 20102011 ) are from the establishment year In all 3 yr, N was applied as ammonium sulfate ((NH4)2SO4) at a rate of 150 kg N ha1 yr1, and K was applied as muriate of potash (KCl) at a rate of 90 kg K ha1 yr1. Nutrients were split applied, with

PAGE 182

182 applications of 50 kg N and 45 kg K ha1 in mid April and 100 kg N and 45 kg K ha1 in mid May. No P was needed based on soil test. Limited irrigation was applied to the experiment only at sign of significant drought st ress (leaf rolling). Ther e were five irrigation events during the growing season of 2010 totaling 60 mm, three irrigation events in 2011 totaling 50 mm and no irrigation was applied in 2012. Treatments and Experimental Design The two grass entries included Merkeron elephantgrass (Burton, 1989) and L791002 energycane (Bischoff et al., 2008) As mentioned previously (Chapter 3), these two cultivars have documented potential as biofuel feedstock s in this region. E xperimental design was completely randomized with four replications of the gr ass entr ies In Year 3, stands of L79 1002 had become sufficiently depleted that they could no longer be used in the experiment, so data are reported for Merkeron only in that year. Response Variables The initial biomass harvest occurred at times ranging from the date of the first freeze event (2010) to 15 (2011) to 25 d (2012) prior to first freeze (Table 71). In Years 2 and 3, the initial harvest occurred in advance of first freeze due to actual overnight temperatures being greater than forecast low te mperatures. For presentation of results, data from the at freeze harvest is shown as being collected on Day 0, but the counting of time after a freeze event did not start until the actual date of the freeze (Table 71). At harvest, a 2 m ( Y ear 1) or 3 m ( Years 2 and 3) length of row from a middle row of each plot was clipped. The grass was cut to 12cm stubble height using a brush cutter To minimize border effects, the 1m portion at the end of the harvested row was

PAGE 183

183 not part of the yield sample. The harvested area in each plot was thus considered to be 2 m2 ( Y ear 1) or 3 m2 ( Y ear s 2 and 3). There was one fewer harvest event in Years 2 and 3 than Year 1, allowing for the lar ger row length to be sampled in those years (Table 71). All material from the harvested portion of row was weighed fresh in the field and subsampled to determine DM concentration and to calculate DM harvested. The subsamples were dried at 60 C until constant weight. In 2010 and 2011, the dried subsamples for DM concentration were handseparated into leaf (blade and sheath) and stem (including inflorescence, if present) components to determine leaf and stem proportion in the dry matter After the biomass samples w ere collected at a given date, the remaining plot area remained undistur bed until the next sampling date. After the final harvest dat e the remaining area of the plot was clipped to the target stubble (12 cm) using a disk mower. To obtain representative wholeplant samples, stem and leaf samples were ground separately and a wh ole plant composite for each plot was made based on weighted proportions of leaf and stem Stem samples were initially ground through a hammer mill to reduce particle size. Stem and leaf samples were ground to pass a 1mm stainless steel screen in a Wiley mill (Model 4 Thomas Wiley Laboratory Mill, Thomas Scientific, Swedeboro, NJ). Each component of dried biomass from Years 1 and 2 was analyzed for biomass composition. T he m odified method of detergent fiber analysis (Van Soest et al., 1991) was used for sequential n eutral detergent fiber (N DF ) acid detergent fiber ( ADF ) and acid detergent lignin ( ADL ) analysis. The samples were analyzed for NDF and ADF using the ANKOM fiber analyzer (ANKOM 2000 Fiber Analyz er, ANKOM Technology

PAGE 184

184 Corporation, Fairport, NY) T he procedure Method for Determining Acid Detergent Lignin in Beakers (proposed by ANKOM Technology Corporation) was followed for ADL determinations. Cellulose was estimated as the difference between ADF and ADL concentration and hemicellulose was estimated as the difference between NDF and AD F concentration (Jung and Lamb, 2004; Waramit et al., 2011) As described in Chapter 4, total N and P concentrations in the leaf and stem were determined using a standardKjeldahl method, a modification of the aluminum block digestion procedure (Gallaher et al., 1975) followed by semi automated colorimetric determination (Hambleton, 1977) D igestions were conducted at the Forage E valuation S upport L aboratory of the University of Florida. Samples were ashed using a muffle furnace at a temperature of 500C for a minimum of 6 h Statistical Analysis Data were analyzed using mixedmodel methods (SAS Institute, 2008) In a ll models, grass entry was considered a fixed effect. Sampling date (Table 71) was considered a repeated measurement (fixed). Means were compared using the pdiff test of LSMEANS. All means reported in the text are least squares means and were considered different if P Only Merkeron was sampled f or biomass harvested and DM concentration in Y ear 3, so grass entry was not included in the model in that year Sampling date was considered to be fixed and a repeated measurement for Y ear 3. Data were analy zed by year because of difference in number of harvest events among years and because energycane was not part of the experiment in Year 3. A subsequent analysis was conducted with Merkeron only in which year was included as a random effect and sampling dat e was considered a repeated measure. A trend was referred to if P 0.10 and > 0.05.

PAGE 185

185 Results and Discussion During the years of study, there were large variations in weather patterns that followed the first freeze of winter. Average temperature following f irst freeze of Year 1 was much lower than for Years 2 and 3 (9.8, 15.4, and 15.4 C, respectively ) (Fig. 7 1) Although average temperature of Years 2 and 3 were similar Year 2 had additional freezing events at 2 ( 1.9 C ) and 6 wk ( 5.3 C ) after the first freeze, but there was only one additional freeze event ( 5.3 C ) in Year 3, on Day 57 after first freeze which was the day before final biomass harvest. Harvested B iomass Entry, sampling date, and thei r interaction had no e ffect on biomass yield in Y ear s 1 or 2 ( P = 0.773, 0.198, and 0.320 respectively, in Y ear 1; P = 0.153, 0.178, and 0.184, respectively, in Y ear 2, Fig. 73). In Y ear 3, there was a strong trend ( P = 0.067, Fig. 7 3) for biomass harvested to decrease from time of freeze until approximatel y 8 wk after freeze Biomass harvested was within the range reported previously for biomass crops harvested onc e per year in this region (Woodard et al., 1991b; Woodard and Prine, 1993b; Knoll et al., 2012) In Y ear 1, average harvested biomass was 26.7 and 27.6 Mg ha1 for L791002 and Merkeron, respectively. Although it was not significant ( P = 0.153), Merkeron (avg. 29.2 Mg ha1) tend ed to have greater biomass yield than L791002 (avg. 23.5 Mg ha1) in Y ear 2. In Y ear 3, it was not possible to obtain meaningful biomass harvest ed data from L79 1002 plots The y had sustained severe damage from the fungal disease sugarcane smut ( Sporisori um scitamineum ) an important pest of L791002 which was reported in previous energycane studies (Bischoff et al., 2008; Len et al., 2012; Ramesh Sundar et al., 2012; Viator and Richard, 2012) and in the other experiment at this location (Chapter 3) Although energycane L791002 was

PAGE 186

186 reported to be only moderately susceptible in the p r evious study (Bischoff et a l., 2008) at this location in the current study it was quite susceptible. The severity of smut on L791002 caused yield loss (Chapter 3), and susceptibility to smut may decrease the vigor and stand life of energycane. If L79 1002 wa s removed from the analysis and data we re analyzed with year in the model as a random effect Merkeron biomass harvested decreased (P = 0.029) from time of first freeze until the end of the delayed harvest period. The average decrease over 3 yr was from 30.7 to 18.9 Mg ha1 over a period after first freeze of 50 to 59 d. Although there was a wide range in weather conditions after the first freeze during the years of study, biomass harvested did not seem to be impacted to a greater or lesser degree by the colder Year 1 vs. th e warmer Years 2 and 3. D elayed cutting has been reported to negatively affect miscanthus biomass harvested (Lewandowski and Heinz, 2003) and switchgrass biomass decreased from 10.0 Mg ha1 in Oct ober to 5.6 Mg ha1 in Apr il in Iowa (Wilson et al., 2013b) Other authors have reported a reduction in both switchgrass biomass during winter (Adler et al., 2006) with magnitude of reduction reaching 90% due to lodging and difficult ies in gathering biomass during baling In the current study, Merkeron biomass harvested decreased during the post freeze period, but the variability was large In addition, the harvest methodology used in the current study was quite tho rough in capturing biomass, so although delayed harvest increased lodging (visual observation) the problem of gathering biomass for delayed harvest that was noted by Adler et al. (2006) did not arise. It may well be a problem in commercial production, however. Overall, yield reduction in the current study was less proportionally than the very large decreases

PAGE 187

187 reported in previous studies (Adler et al., 2006; Tahir et al., 2011; Wilson et al., 2013a) in much more severe winter climates than Florida. Biomass D ry M atter C oncentration Effect of delayed harvest on biomass dry matter (DM) concentration was evaluated in all 3 yr (Fig 7 4). There were effects of entry and sampling date ( P = 0.025, and P < 0.001, respectively) in Y ear 1 (Fig 7 4). In the second year there was entry x samp ling date interaction ( P < 0.001) and i n the third year there was an effect of sampling date ( P = 0.013). In the first year Merkeron had greater DM concentration than L791002 ( av erage of 377 vs 356 mg g1, Fig. 7 4 ). Slightly, but not significantly, greater DM concentration was observed for Merkeron vs. L791002 in a single fall harvest treatment in Florida, but the range of responses was only from 311 to 338 mg g1 (Woodard et al., 1991b) Biomass DM concentration increased in Year 1 of the current study until 34 d after freeze (386 mg g1) and then decreased slightly at final harvest on Day 50 (366 mg g1). The decrease in biomass DM concentration at the final harvest is likely due to a 33mm rainfall event that occurred the day before final harvest. In Y ear 2, entry x sampling date interaction occurred because Merkeron DM concentration increased almost two fold throughout the delayedharvest period (351, 519, and 676 mg g1 at 0, 28, and 59 d after freeze, respectively) whereas energycane increased only from 329 mg g1 at first freeze to 401 mg g1 at final harvest on Day 51. In contrast, in Year 3 Mer keron DM concentration changed relatively little and was 355, 314, and 367 mg g1 before freeze and at Day s 27 and 58 after freeze, respectively. The very large change in Year 2 DM concentration was unique for Merkeron among years and also was very different than the responses of energycane. It is proposed that weather conditions affected the response. After the killing freeze in Year

PAGE 188

188 1, the weather remained cold throughout the delayed harvest period. There were a total of 15 freeze events between first freeze and the final biomass harvest in Year 1, and the longest period between freeze events was 14 d. This type of environment apparently minimized DM loss. In both Years 2 and 3, there was a killing freeze followed by much warmer winter weather (avg. temp. in both years was 15.4C vs. 9.8 C in Year 1) Exploring ways that Years 2 and 3 were different, in Year 2 the temperature fell to 5.6 C the day aft er first freeze and during the delayed harvest period there was a 26d period without a freeze event. In Year 3, t here was a period of 29 d following first freeze without occurrence of additional freezes, and there were no temperature below 2 C until the day before the final harvest. This resulted in regrowth of new tillers in Year 3 which caused DM concentration to decline between first freeze and the first harvest date after first freeze in Year 3 (Fig. 74). Although there is not a definitive cause and effect, it may be that the very low temperature that occurred early in the delayed harvest period of Year 2 affected the integrity of elephantgrass stems, opening them up to drying and the freezefree period that followed provided excellent drying condition s. Previous research has shown differences in perennial grass DM concentration among years during a delayed harvest period in UK (Smith and Slater, 2011) There also was a large difference between entries in Y ear 2. This may be due to differ ences in stem anatomy. The surfaces of plants are covered by several layers of lipophilic material (mostly epicuticular wax ) and the primary function of ubiquitous presence of epicuticular wax is as a waterproof barrier (Purcell et al., 2005) Scanning electron micrographs showed that the outermost layer of sugarcane stem i.e., comrind, is composed of epidermal cells that contain a waxy later, moreover, its rind fiber was

PAGE 189

189 very thick (Han and Wu, 2004) The authors also indicated that the innermost layer of comrind is wrapped with pith cells. Thos e anatomical characteristics of energycane (sugarcane) stem may result in less seasonal change in stem DM concentration than in elephantgrass. Additionally, there is a w elldeveloped aerenchymalike tissue in the stem of elephantgrass (Jennewein et al., 2012) which if the outer stem layer s are disrupted may accelerate field dryi ng during periods of warm weather. Unlike elephantgrass, energycane has a relatively high concentration of water soluble carbohydrate and extractives (Woodard et al., 1991b; Fedenko, 2011) and that along with its anatomical characteristics may have reduced rate of stem dry down during delayed harvest. Leaf Proportion Leaf proportion was quantified in Years 1 and 2 only. There was a sampling date effect on leaf proportion in Y ear 1 ( P < 0.001, Fig. 7 5) and in Year 2 there were effects of entry and sampling date ( P = 0.001 and P < 0.001, respectively ). During Year 1, l eaf proportion increased to 18 d after first freeze and decreased until the final harvest (Fig. 7 5). T here was no entry effect ( P = 0.120) but elephantgrass tended to have lower leaf proportion th a n energycane in Year 1 (23.2 vs. 27.0% respectively ). In the second year, L791002 leaf proportion across sampling dates ( avg. 27.6%) was 2.6fold greater than Merkeron ( avg. 10.5% Fig. 7 5 ). L eaf proportion across ent ries decreased from 26.6 prior to first freeze to 18.2 at Day 28 and 12.4% at 59. These data show that in general delayed harvest after first freeze decreased leaf proportion of these perennial grass. Leaf is considered to be least valuable part of biomass harvested because of greater concentrations of N and ash (Adler et al., 2006; Sokhansanj et al., 2009) so delayed harvest may increase biomass quality w ithout the

PAGE 190

1 90 additional cost of remov ing leaves at harvest However, it has been shown that the amount of abscised leaves varied with post freeze weather conditions (Wilson et al., 2013a) In this research, extreme cold at first freeze in Year 2 followed by warm weather after freeze may have accelerated leaf abscission, and elephantgrass had only 5.3% leaf in total biomass by 59 d after freeze. This suggests that part of the decrease in elephantgrass biomass harvested and increase in dry matter concentration with delayed harvest is caused by leaf fall. Fiber A nalysis Biomass NDF concentration was affected by entry and sampling date ( P < 0.001 for both) in Y ear 1 (Fig. 76) and entry x sampling date interaction in Y ear 2 ( P = 0.026). There was a similar range of NDF concentration as reported in previous research (Woodard and Prine, 1991; Van Man and Wiktorsson, 2003) Across all sampling dates, average NDF concentration in elephantgrass was greater than energycane (784 and 702 mg g1, respectively, Fig 7 6). The NDF concentration increased with delayed harvest (709, 739, 754, and 771 mg g1 for 0, 18, 34, and 50 d after first freeze respectively). In the second year, M e r keron had greater NDF concentration than L791002 and interaction occurred because the magnitude of the difference between entries increased as harvest was delayed (7, 10, and 12% difference at 0, 28, and 59 d after freezing Fig. 7 6). When winter was cool in Y ear 1, the increase in NDF of Merkeron and energyc ane was similar (9% increase from pre freeze to 50 d after first freeze for both), however during the w arm er winter in Y ear 2 the difference increased with time. There were entry and sampling date effects on ADF concentration in Y ear 1 ( P < 0.001, Fig. 7 7 ) and in Y ear 2 ( P = 0.001). Similar ADF concentration was reported in previous research with elephantgrass (Van Man and Wiktorsson, 2003) During the first

PAGE 191

191 year, overall ADF concentration in Merkeron was greater than L791002 (avg. 528 vs. 451 mg g1). Across entr ies, ADF concentration increased 12% throughout the season. In Y ear 2, entry x sampling date interaction occurred because the increase in Merkeron ADF during the delayed harvest period (19%) was greater than in L791002 (10%). The t rend in ADF concentration was very similar to that of NDF Biomass ADL w as affect ed by entry and sampling date ( P < 0.001 for both, F ig. 7 8 ) in Y ear 1 and entry x sampling date interaction in Y ear 2 ( P = 0.019). The range in elephantgrass ADL concentration in the current study was similar to that reported in previous research (12.0 12.5 mg g1) (Rengsirikul et al., 2011) During the first year the last sampling date (50 d after freeze) showed greatest ADL concentration in both grasses (77 and 106 mg g1, Fig 7 8). Average ADL concentration of Merkeron (97 mg g1) was greater than for L791002 (72 mg g1) across sampling dates During the second year, entry x sampling date interaction occurred because elephantgrass ADL concentration increased to a greater degree with delayed harvest than did energycane (31 and 18% increase for Merkeron and L791002, respectively). This pattern of response was the same as described for NDF and ADF. Previous research has reported an increase in lignin concentration of warm season perennial grasses late in the season in Iowa (Warami t et al., 2011) Cellulose was affected by entry and sampling date during the first year ( P < 0.001 for both, Fig. 79 ) and entry x sampling date interaction in Year 2 ( P = 0.001). Concentration of elephantgrass cellulose in the current study is comparable to that (447462 mg g1) reported previously (Rengsirikul et al., 2011) Average cellulose concentration in Merkeron (431 mg g1) was greater than in L791002 (379 mg g1) in

PAGE 192

192 Year 1. Cellulose concentration across entries increased from 384 mg g1 in prefreeze biomass to 425 mg g1 by 50 d after first freezing (Fig. 7 9). In the second year entry x sampling date interaction occurred because the increase in Mer keron during the delayed harvest period was greater than in L791002 ( 16 and 8% respectively, from pre freeze to 59 d after first freeze). A similar overall trend for cellulose concentration during late season was reported in Iowa with perennial grasses (Waramit et al., 2011) The pattern of hemicellulose concentration response was different from the other fiber components. There was entry x sampling date interaction in the first year ( P = 0.040, Fig 7 10 ) and an effect of entry in Year 2 ( P = 0.009). The range in elephantgrass hemicellulose concentration was similar to that reported previously (232236 mg g1) (Rengsirikul et al., 2011) From pre freeze to 18 d after first freeze in Y e ar 1, Merkeron had greater hemicellulose concentration t han L791002 ( 254 vs. 243 mg g1 at prefreeze, 265 vs. 255 mg g1 at 18 d after first freeze, Fig. 7 10), but thereafter no differences were detected and the average was 253 mg g1. Unlike the first year L79 1002 showed greater hemi cellulose concentration (avg. 271 mg g1) than Merkeron (avg. 254 mg g1) throughout Year 2. Hemicellulose concentration remained relatively constant across the delayed harvest period. Similar lateseason trends of hemicellulose concentration were reported in previous research (Waramit et al., 2011) Considering the fiber components as a group and with the exception of hemicellulose, the pattern of change in concentration during the del ayed harvest period was similar. Merkeron had greater structural component concentration than energycane. When temperatures were warm during the delayed harvest period in Year 2, the magnitude of change in structural components was greater in elephantgrass than

PAGE 193

193 energycane. An overall increase in structural component concentration and a decrease in soluble carbohydrate was reported in switchgrass during the winter (Adler et al., 2006) An increase in NDF, ADF, and ADL were reported over winter in a study with reed canarygrass ( Phalaris arundinacea L.) (Tahir et al., 2011) and similar seasonal patterns in NDF, ADL, cellulose and hemicellulose was reported in switchgrass (Dien et al., 2006) In a previous study, L791002 (94.6 mg g1) had greater water soluble concentration than Merkeron (43.5 mg g1) in a single fall harvest in Florida (Woodard et al., 1991b) and a similar pattern of extractives (256 and 209 mg g1; energycane and elephantgrass, respectively ) (Fedenko, 2011) Greater concentration of extractives (mostly water soluble carbohydrates) in energycane likely caused the difference in fiber components among entries. As previously described for DM concentration, elephantgrass might be more sensitive to post freeze weather conditions because of its anatomical characteristics. If elephantgrass is exposed to warm temperature af ter a severe killing freeze, contents of the stem appear likely to lose a greater concentration of non structural constituents, perhaps by microbial activity. Nitrogen There was no effect of entry, sampling date or their interaction ( P = 0.666, 0.285, and 0.832, respectively) in the first year (Fig. 711), but in Year 2 there were entry and sampling date e ffects ( P = 0.032 and 0.011, respectively). During first year, overall average N concentration was 4.65 mg g1. In the second year, elephantgrass biomass had greater N concentration (4. 2 mg g1) than energycane (2.4 mg g1) when averaged across the delayed harvest period (Fig. 711). Average N concentration across entries

PAGE 194

194 decreased with delayed harvest and was 4.0, 3.4, and 2.5 mg g1 i n 0 28, and 59 d after the freezing event respectively, in Year 2. Nitrogen concentration was stable in Year 1 when the delayed harvest period temperature was cool, perhaps due to less structural breakdown and leaching of N from plant tissue or to great er retention of leaves (Fig. 75). W hen temperatures were warm following a very severe freeze event in Year 2 N concentration decrease d over the period. Previous research has reported mixed result s relative to change in N concentration over winter with bi omass N increasing in some experiments (Burvall, 1997; Lewandowski and Heinz, 2003; Dien et al., 2006) or remaining the same in others (Adler et al., 2006; Wilson et al., 2013a) It is believed that tem perature and rainfall may significantly affect N concentration during delayed harv est O ne difference between Y ear s 1 and 2 in the current study was the amount of leaf drop (Fig. 75) w hich was much greater in Year 2, the year in which N concentration decreased markedly Previous perennial grass studies have shown that although there were seasonal changes, l eaf is greater in N concentration than stem in general (Griffin and Jung, 1983; Mislevy et al., 1995; Beale and Long, 1997; Smith and Slater, 2011) Thus, w hen perennial grass es los e leaves N concentration in biomass decreases Ash T here were entry and sampling date effects ( P = 0.042 and 0.040, respectively) in Y ear 1 (Fig. 7 12) and a s ampling date effect in Y ear 2 ( P = 0.017). Average pr e freeze ash concentration in Year 1 was greater than at the end of delayed harvest period ( 30 vs. 25 mg g1, respectively) across entries Decreas ing ash concentration during winter w as reported in previous studies (Adler et al., 2006; Kludze et al., 2013; Wilson et al., 2013a) Merkeron had greater ash (29 mg g1) than L791002 (24 mg g1)

PAGE 195

195 throughout the period in Y e ar 1, with biomass from the l atest harvest ( 59 d after freeze) having lesser ash concentration (21 mg g1) compared to prefreeze biomass ( 27 mg g1). Ash is the inorganic residue remaining after ignition of biomass, and h erbaceous crops produce meaningful amount s of ash (Kludze et al., 2013) There are five primary components (Si, K, Ca, S and Cl) that have the greatest e ffect on ash concentration (Bakker and Elbersen, 2005) However, Si, which compose s the greatest proportion of ash, is relatively stable so remov al of other elements such as K is more likely to affect overall ash concentration (Tahir et al., 2011; Kludze et al., 2013) Further study is needed with ash, including which elements change over time and the extent to which change in ash concentration is related to leaf fall. Implications of Research Delayed harvest until after occurrence of freezing temperatures can increase the duration of the harvest period, but it may impact quality and quantity of herbaceous biomass Harvest of eleph antgrass and energycane was delayed up to 2 mo after the first killing freeze event in this experiment. Energycane L79 1002 b iomass yield was unaffected by delayed harvest but elephantgrass biomass decreased with increasing number of days in the delayed h arvest period. Leaf proportion in biomass decreased with increasing time between freeze and harvest, especially with elephantgrass. A bscission of leaf may decrease biomass yield and concentration of N and ash in biomass Biomass DM concentration generally increased with delayed harvest with large year and species variation, especially in e lephantgrass which is more sensitive to weather conditions after a freeze event. Elephantgrass generally had greater concentration of cell wall constituents than energycane. Fiber component concentration generally increased after a freeze event

PAGE 196

196 with the exception of hemicellulose. Increasing fiber component concentration was likely associated with loss of nonstructural constituents, perhaps by leaching or microbial activity. Biomass N and ash concentration tended to decrease overwinter but not to a large degree. These data suggest that in terms of biomass harvested energycane is better suited for delayed harvest following a freeze because its biomass loss is minimal. Elephantgrass appears to be more susceptible to loss of biomass and nonstructural constituents following a freeze event, particularly when warm weather follows the occurrence of freezing temperatures. Thus, post freeze weather conditions will likely determine the viability of delaying harvest, especially with elephantgrass, with colder temperatures following a freeze event likely having less negative impact on biomass harvested and composition.

PAGE 197

197 Table 7 1 Sampling and freeze event dates for delayed harvest management in 20102011, 20112012, and 20122013. Season Sampling dates and freeze events Harvest date before freeze Date of first freeze 1 st harvest after freeze 2 nd harvest after freeze 3 rd harvest after freeze 2010 2011 2 Dec. 2010 2 Dec. 2010 20 Dec. 2010 (18 DAF) 5 Jan. 2011 (34 DAF) 21 Jan. 2011 (50 DAF) 2011 2012 19 Dec. 2011 3 Jan. 2012 31 Jan. 2012 (28 DAF) 2 Mar. 2012 (59 DAF) --2012 2013 28 Nov. 2012 23 Dec. 2012 19 Jan. 2013 (27 DAF) 19 Feb. 2013 (5 8 DAF) --DAF; days after the first freeze event occurred.

PAGE 198

198 Temperature (C) -5 0 5 10 15 20 25 Avg. Min. Max. Temperature (C) -5 0 5 10 15 20 25 Weeks after first freeze 1 2 3 4 5 6 7 8 Temperature (C) -5 0 5 10 15 20 25 2010-2011 2012-2013 2011-2012 Figure 71. Weekly a verage air temperature (Avg.) and average of weekly maximum (Max.) and minimum (Min.) air temperatures for 20102011, 20112012, and 20122013 at the experimental location (Citra, FL)

PAGE 199

199 Weeks after first freeze 1 2 3 4 5 6 7 8 rainfall (mm) 0 10 20 30 40 50 2010-2011 (total 103 mm) 2011-2012 (total 69 mm) 2012-2013 (total 68 mm) Figure 7 2 Weekly total rainfall for 20102011, 20112012, and 20122013 at the experimental location (Citra, FL)

PAGE 200

200 Biomass harvested Mg ha-1 051015202530354045 L79-1002 Merkeron Biomass harvested Mg ha-1 051015202530354045 Days after freeze 0102030405060Biomass harvested Mg ha-1 051015202530354045 2010-20112011-20122012-2013 Figure 7 3 Effect of days after a freeze event on biomass harvested of two perennial grass entries in three years Entry, sampling date, and their interaction P values were 0.773, 0.198, and 0.320, respectively, in 20102011; 0. 153, 0.178, and 0. 184 respectively in 20112012; and 0.067 (sampling date) in 20122013. Bars show mean one standard error.

PAGE 201

201 DM concentration mg g -1 300 350 400 450 500 550 600 650 700 L79-1002 Merkeron DM concentration mg g -1 300 350 400 450 500 550 600 650 700 ** ** Days after freeze 0 10 20 30 40 50 60DM concentration mg g -1 300 350 400 450 500 550 600 650 700 2010-2011 2011-2012 2012-2013 Figure 7 4 Effect of days after a freeze event on dry matter (DM) concentration of two perennial grass entries in three years Entry, sa mpling date and their interaction P values were 0.025, < 0.001, and 0.928, respectively, in 20102011; < 0.001 for all in 20112012; and 0.013 (sampling date) in 20122013. Bars show mean one standard error. Entries within a date, ** P

PAGE 202

202 Leaf proportion (%) 0 5 10 15 20 25 30 35 40 L79-1002 Merkeron Days after freeze 0 10 20 30 40 50 60Leaf proportion (%) 0 5 10 15 20 25 30 35 40 Figure 7 5 Effect of days after a freeze event on leaf proportion of two perennial grass entries in 20102011 (upper) and 20112012 (lower) Entry, sampling date, and their interaction P values were 0.120, < 0.001, and 0.069, respectively, in 20102011; and 0.001, < 0.001, and 0.309, respectively in 20112012 Bars show mean one standard error.

PAGE 203

203 NDF mg g-1 600 650 700 750 800 850 L79-1002 Merkeron Days after freeze 0 10 20 30 40 50 60NDF mg g-1 600 650 700 750 800 850 ** ** ** Figure 7 6 Effect of days after a freeze event on neutral detergent fiber ( NDF ) conc entration of two perennial grass entries in 20102011 (upper) and 20112012 (lower) Entry, sampling date, and their interaction P values were < 0.001, < 0.001, and 0.714, respectively, in 20102011; and < 0.001 < 0.001, and 0. 026 respectively in 20112012. Bars show mean one standard error. Entries within a date, ** P

PAGE 204

204 ADF mg g-1 400 420 440 460 480 500 520 540 560 580 600 620 L79-1002 Merkeron Days after freeze 0 10 20 30 40 50 60ADF mg g-1 400 420 440 460 480 500 520 540 560 580 600 620 ** ** ** Figure 7 7 Effect of days after a freeze event on acid detergent fiber ( ADF ) concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower) Entry, sampling date, and their interaction P values were < 0.001, < 0.001, and 0.627, respectively, in 20102011; and < 0.001 < 0.001, and 0. 001 respectively in 20112012. Bars show mean one standard error. Entrie s within a date, ** P

PAGE 205

205 ADL mg g-1 60 70 80 90 100 110 120 130 L79-1002 Merkeron Days after freeze 0 10 20 30 40 50 60ADL mg g-1 60 70 80 90 100 110 120 130 ** ** ** Figure 7 8 Effect of days after a freeze event on acid detergent lignin ( ADL ) concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower) Entry, sampling date and their interaction P values were < 0.001, < 0.001, and 0.174, respectively, in 20102011; and < 0.001 < 0.001, and 0. 019 respectively in 20112012. Bars show mean one standard error. Entries within a date, ** P

PAGE 206

206 Cellulose mg g-1 360 380 400 420 440 460 480 L79-1002 Merkeron Days after freeze 0 10 20 30 40 50 60Cellulose mg g-1 360 380 400 420 440 460 480 ** ** ** Figure 79. Effect of days after a freeze event on cellulose concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower). Entry, sampling date, and their interaction P values were < 0.001, < 0.001, and 0.678, respectively, in 20102011; and < 0.001, < 0.001, and 0.001, respectively in 20112012. Bars show mean one standard error. Entries

PAGE 207

207 Hemicellulose mg g-1 200 210 220 230 240 250 260 270 280 290 L79-1002 Merkeron ** ** Days after freeze 0 10 20 30 40 50 60Hemicellulsoe mg g-1 200 210 220 230 240 250 260 270 280 290 Figure 7 10. Effect of days after a freeze event on hemicellulose concentration of two perennial grass entries in 20102011 (upper) and 20112012 (lower) Entry, sampling date and their interaction P values were 0.240, 0.006, and 0.040, respectively, in 20102011; and 0.009, 0.638, and 0. 272 respectively in 20112012. Bars show mean one standar d error. Entries within a date, ** P

PAGE 208

208 N mg g-1 0 1 2 3 4 5 6 7 8 L79-1002 Merkeron Days after freeze 0 10 20 30 40 50 60N mg g-1 0 1 2 3 4 5 6 7 8 Figure 7 11. Effect of days after a freeze event on N concentration of two perennial grass entries in 2010 2011 (upper) and 20112012 (lower) Entry, sampling date and their interaction P values were 0.666, 0.285, and 0.832, respectively, in 20102011; and 0.032, 0.011, and 0.622 respectively in 20112012. Bars show mean one standard error.

PAGE 209

209 Ash mg g-1 0 5 10 15 20 25 30 35 40 L79-1002 Merkeron Days after freeze 0 10 20 30 40 50 60Ash mg g-1 0 5 10 15 20 25 30 35 40 Figure 7 12. Effect of days after a freeze event on ash concentration of two perennial grass entries in 2010 2011 (upper) and 20112012 (lower) Entry, sampling date and their interaction P values were 0.042, 0.040, and 0.815, respectively, in 20102011; and 0.182, 0.017, and 0.342 respectively in 20112012. Bars show mean one standard error.

PAGE 210

210 CHAPTER 8 CONCLUSIONS Outstanding biomass yield potential of warm season perennial grasses is attractive for cellulosi c biofuel production systems In the s outheastern USA, elephantgrass ( Pennisetum purpureum Schum.) and energycane ( Saccharum spp. hybrid) are widely recognized for their biomass production and therefore are candidate species for second generation biofuel p roduction in this region. Although regional biomass production is high, a logistical challenge facing processing plants that convert feedstock to fuel is seasonality of biomass production. This results in uneven supply of feedstock to the conversion facili ty and limits efficiency of operation. Different grass species may be better adapted to flexible harvest management, but there is relatively little information available describing such differences for elephantgrass and energycane. T he research reported i n this dissertation was conducted to address the following objectives: 1) determine the effect of harvest management of elephantgrass and energycane on biomass yield and composition when plants are managed as bioenergy feedstock s ( Chapter s 3 and 4) ; 2) assess morphological and chemical changes of elephantgrass and energycane throughout the growing season ( Chapter s 5 and 6); and 3) investigate effects of delayed harvest after a killing freeze on yield and chemical composition of elephantgrass and energycane (Chapter 7). Effect of H arvest M anagement on P erennial G rasses Chapter s 3 and 4 The experiment was conducted during 2010, 2011, and 2012 at the Plant Science Research and Education Unit (PSREU) at Citra, FL ( 29.41 N, 82.17 W). The three grass e ntries included two elephantgrasses, Merkeron and a breeding line UF 1,

PAGE 211

211 and L791002 energycane. Three harvest management treatments were imposed that included different frequencies and timing of harvest. These were i) two harvests per year (2X; summer and fall ratoon), ii) one harvest per year in fall ( 1X Nov ), and iii) one harvest per year in winter ( 1X Dec ). Biomass harvested and related responses (Chapter 3) and chemical composition characteristics (Chapter 4) were measured. Delaying a single harves t until after a freeze event or harvest ing biomass twice per year increased the effective harvest period of biomass and would therefore likely improve seasonal distribution of biomass to the biorefinery. However, increasing harvest frequency to twice per y ear may compromise long term biomass production (42% biomass reduction for this treatment in the third year). Energycane biomass decreased 41% from Year 2 to Year 3 due to damage from sugarcane smut ( Sporisorium scitamineum ) disease. Delaying harvest until after a freeze event reduced the leaf percentage in harvested biomass and increased the biomass dry matter concentration. Persistence evaluation suggested that UF 1 likely has capability to adapt to loss of tiller s in the stand and achieve the same or eve n greater biomass yield than Merkeron. Harvest frequency affects compositional quality of perennial grasses. Delaying a single harvest until fall maximized the concentration of cellulose in total biomass but there was little change thereafter in elephantg rass In contrast, increasing soluble sugar concentration in energycane caused a decrease in structural carbohydrate concentration in plants not harvested until first freeze. A major factor affecting concentration differences due to harvest management was differences in leaf proportion. Later harvests were associated with lesser leaf percentage in total biomass and this caused N, P, and ash to decrease in 1X Nov and 1X Dec re lative to 2X.

PAGE 212

212 D ecrease in N and ash wa s associated with amount of leaf abscission For single harvest treatments because of decreased leaf proportion, compositional characteristics of total biomass usually were very similar to those of stem 1X Nov and 1X Dec treatments of all entries had lesser concentrations of the components N and ash, but they had greater lignin concentration than for 2X. Morphological and C hemical C hanges of P erennial G rasses During the G rowing S eason Chapter s 5 and 6 Plots used were a subset of those from the experiment described in Chapter 3. Fullseason growth sampling occurred on plots that were harvested once per year and first growth/ratoon growth sampling occurred on the plots that were harvested twice per year with first har vest on 30 July 2010 and 21 July 2011. Morphological characteristics (Chapter 5) and chemical constituents (Chapter 6) were measured throughout the growing season. Merkeron and UF 1 elephantgrasses generally demonstrated similar seasonal patterns in morphological characteristics. Differences in seasonal morphological characteristics between elephantgrass and energycane were detected. Relative to energycane, the elephantgrasses generally had 1) a more consistent number of tillers throughout the season, 2) g reater tiller mass and earlier development of LAI than energycane, and 3) greater stem proportion and biomass dry matter concentration than energycane. Elephantgrass UF 1 showed desirable characteristics for biomass production including increasing tiller m ass and canopy height until late in the growing season and a greater proportion of stem. During ratoon growth grasses rapidly achieved the same LAI as fullseason growth which explains why there were no yield differences between one and two harvests per year treatments in the first 2 yr of the

PAGE 213

213 study reported in Chapter 3. These data provide evidence of the value of detailed characterization of morphological responses in understanding plant biomass accumulation and response to defoliation. Seasonal changes in chemical composition were quantified by assessing cell wall constituents using the Van Soest fiber analyses and through miner al analyses Different patterns of response were observed between elephantgrass and energycane. C ell wall constituent s of the t wo elephantgrasses responded similar ly Cell wall constituents except hemicellulose increased until late summer and either remained relatively constant (UF 1) or slightly increased (Merkeron) during the rema inder of the growing season. In contrast, once en ergycane cell wall constituents peaked in late summer, they decreased thereafter because of increasing concentrations of nonstructural components. Nitrogen and ash are considered to be anti quality factors in some biomass conversion processes. Because le af N and ash concentrations are much greater than those in stem, N and ash can be return ed to the field efficiently by leaf abscission as leaves mature. T here may be value in returning remaining leaf biomass to the production field at time of harvest Data for both N and ash indicate that returning leaf to the production field is a more attractive option for elephantgrass than for energycane. In elephantgrass, leaf biomass as a percentage of total biomass is considerably smaller than leaf N and ash content as a percentage of total N and ash. Thus, there is an option to reduce N and ash in harvested biomass to a greater degree than biomass yield if leaf were to be returned at harvest. Currently, it is not certain the degree to which the loss of

PAGE 214

214 biomass is com pensated for by greater biomass quality or by return of nutrients for use in subsequent growing seasons. Effects of Delaying Harvest After a Freeze Event on Biomass Harvested and C hemical C omposition Chapter 7 The experiment was conducted during the winter seasons of 20102011, 20112012, and 20122013 at the Plant Science Research and Education Unit (PSREU) adjacent to the site where the experiments were conducted that are reported in Chapter s 3 through 6. Harvest of two grass entries (Merkeron elephantgrass and L791002 energycane) was delayed incrementally for up to 2 mo after the first killing freeze event and change in biomass harvested and biomass composition were quantified. Delayed harvest after a freezing event can increase the duration of the biomass harvest period. E lephantgrass biomass yield and composition were more sensitive to weather conditions after the freeze event than energycane. Energycane b iomass yield was unaffected by delayed harvest but elep hantgrass biomass decreased with increasing number of days in the delayed harvest period. Leaf proportion in biomass decreased with increasing time between freeze and harvest, especially for elephantgrass, and this caused N and ash concentration to decreas e during the delayed harvest period. Fiber components of elephantgrass generally increased after the freeze because stem of this species appear ed to lose a greater amount of nonstructural constituents. These data suggest that energycane may be better suit ed for delayed harvest following a freeze because its biomass loss is minimal. Elephantgrass appears to be more susceptible to loss of biomass and nonstructural constituents following a freeze event, particularly when warm weather follows the occurrence o f freezing temperatures. Thus, post freeze weather conditions will likely determine the viability of

PAGE 215

215 delaying harvest, especially with elephantgrass, with colder temperatures following a freeze event likely having less negative impact on biomass harvested and composition. Implications of the Research It is possible to increase the period when biomass can be provided to the biorefinery by delaying harvest until first freeze without compromising quality and biomass yield of elephantgrass and energycane. In co ntrast, multiple harvests during the growing season, at least in the colder subtropics like northern Florida, may sacrifice long term biomass yield and persistence as well as increase concentrations of minerals and ash in the biomass that can interfere wit h some conversion processes Leaf fall from full season growth, especially when harvest is delayed until after a freeze event will likely increase nutrient recycling for the next growing season, transportation efficiency, and feedstock quality for the bior efinery. The final decision regarding harvest frequency and timing must take into account the species being used and the regional circumstances related to availability of feedstock to the conversion plant. It may well be most efficient if large production areas are divided into management units such that a given unit is harvested twice annually only one year out of three or four. In other years it can be harvested once per year some time during the fall through winter period. Because energycane loses less biomass than elephantgrass during the winter, it may be an excellent candidate to supply biomass later in the winter season. Seasonlong evaluation of morphological characteristics provides information for determining optimal harvest dates. A large decrease in mineral nutrient and ash concentrations over the season, associated with leaf abscission, indicated that nutrient recycling and feedstock quality may be improved if harvest occurs when grass es are mature. Elephantgrass UF 1 showed excellent feedstock c haracteristics including

PAGE 216

216 extended growth, high biomass yield, and late flowering, and it likely merits release as a biomass biofuel crop. If energycane cultivars are to have a future for use as sources of feedstock for bioenergy in this region, smut resist ance will be a necessary plant characteristic. In the absence of such resistance, long term yield is compromised and persistence for more than three growing seasons appears unlikely. Changes in biomass yield and composition of perennial grasses are depende nt on species and weather following the initial freeze. Elephantgrass is more sensitive to weather conditions than energycane, and if warm weather follows the first freeze elephantgrass should be harvested in a relatively narrow time window to avoid loss o f biomass. Future Research Needs Current experiments have answered important questions, but a number of newly emerging challenges and questions will need to be addressed by future research. A lthough efficiency of nutrient recycling is an important consideration for biomass production, the amount of nutrient that remains on the field and the degree to which it contributes to production the following year have not been established. Measurement of N allocation and translocation for the next season will provide further information regarding nutrient use efficiency of perennial grasses. It was not possible to fully assess the performance of energycane b ecause the impact of sugarcane smut was so devastating E nergycane cultivar s selected for smut resistanc e should be evaluated and compared with smut susceptible types and elephantgrass. To more fully assess the benefits of cultivation of perennial grass es it is important to evaluate impacts on ecosystem service s. For instance, the degree of C sequestration associated with long term imposition of various harvest management practices has not yet been investigated. Field observations suggest that bioenergy grass plots provide habitat for many wildlife

PAGE 217

217 species Assessment of species richness and population sizes of various wildlife species are needed to more clearly establish the value of production field s as wildlife habitat. Especially for elephantgrass, there were large border effects such that over time plants in border rows became more robust than in inner rows and likely competed aggressively with plants in inner rows for light and nutrients. Greater row spacing may be needed to address this constraint in future research.

PAGE 218

218 LIST OF REFERENCES Adler, P.R., Sanderson, M.A., Boateng, A.A., Weimer, P.J., Jung, H. J.G., 2006. Biomass Yield and Biofuel Quality of Switchgrass Harvested in Fall or Spring. Agron. J. 98, 15181525. Akin, D.E., 2007. Grass lignocellulose: strategies to overcome recalcitrance. Appl. Biochem. Biotechnol. 137140, 315. Anderson, W., Akin, D., 2008. Structural and chemical properties of grass lignocelluloses related to conversion for biofuels. J. Ind. Microbiol. Biotechnol. 35, 355366. Anderson, W.F., Casler, M.D., Baldwin, B.S., 2008a. Improvement of perennial forag e species as feedstock for bioenergy. In: Vermerris, W. (Ed.), Genetic improvement of bioenergy crops. Springer New York, pp. 347376. Anderson, W.F., Dien, B.S., Brandon, S.K., Peterson, J.D., 2008b. Assessment of bermudagrass and bunch grasses as feedstock for conversion to ethanol. Appl. Biochem. Biotechnol. 145, 1321. Bakker, R.R., Elbersen, H.W., 2005. Managing ash content and quality in herbaceous biomass : an analysis from plant to product. 14th European Biomass Conference and Exhibition ETA Ren ewable Energies, Paris, France, p. 4. Beale, C.V., Long, S.P., 1997. Seasonal dynamics of nutrient accumulation and partitioning in the perennial C4grasses Miscanthus giganteus and Spartina cynosuroides Biomass and Bioenergy 12, 419428. Bischoff, K.P. Gravois, K.A., Reagan, T.E., Hoy, J.W., Kimbeng, C.A., LaBorde, C.M., Hawkins, G.L., 2008. Registration of L 791002 Sugarcane. J. Plant Reg. 2, 211217. Bogdan, A.V., 1977. Tropical pasture and fodder plants. p. 475. Bouton, J.H., 2002. Bioenergy crop breeding and production research in the Southeast, Final Report for 1996 to 2001. Brechbill, S., Tyner, W., Ileleji, K., 2011. The economics of biomass collection and transportation and its supply to indiana cellulosic and electric utility facilities. Bio energy Res. 4, 141152. Burton, G.W., 1989. Registration of Merkeron Napiergrass. Crop Sci. 29, 13271327. Burvall, J., 1997. Influence of harvest time and soil type on fuel quality in reed canary grass ( Phalaris arundinacea L.). Biomass and Bioenergy 12, 149 154.

PAGE 219

219 Carroll, A., Somerville, C., 2009. Cellulosic biofuels. Annu. Rev. Plant Biol. 60, 165182. Casler, M.D., Boe, A.R., 2003. Cultivar X environment interactions in switchgrass. Crop Sci. 43, 22262233. Chaparro, C.J., Sollenberger, L.E., Jones, C.S., 1995. Defoliation effects on Mott elephantgrass productivity and leaf percentage. Agron. J. 87, 981985. Chen, S. F., Mowery, R.A., Scarlata, C.J., Chambliss, C.K., 2007. Compositional analysis of water soluble materials in corn stover. J. Agric. F ood. Chem. 55, 59125918. Cheng, J., 2010. Biomass to renewable energy processes. CRC Press, Boca Raton. Cherubini, F., 2010. The biorefinery concept: Using biomass instead of oil for producing energy and chemicals. Energy Convers. Manage. 51, 14121421. C osta, J.L., Gomide, J.A., 1991. Drying rates of tropical grasses. Trop. Grasslands 25, 325332. de la Mata, L., Cabello, P., de la Haba, P., Agera, E., 2012. Growth under elevated atmospheric CO2 concentration accelerates leaf senescence in sunflower ( Helianthus annuus L.) plants. J. Plant Physiol. 169, 13921400. Department of Energy, 2006a. Breaking the biological barriers to cellulosic ethanol: a joint research agenda available at http://genomicscience.energy.gov/biofuels/ (verfied 15th May 2013). In: Department of Energy (Ed.). DOE EERE. Department of Energy, 2006b. Theoretical Ethanol Yield Calculator available at http://www1.eere.energy.gov/biomass/ethanol_yield_calculator.html (verified 14th Nov. 2012). DOE EERE. Dien, B.S., Jung, H.J.G., Vogel, K.P., Casler, M.D., Lamb, J.F.S., Iten, L., Mitchell, R.B., Sarath, G., 2006. Chemical composition and response to dilute acid pretreatment and enzymatic saccharification of alfalfa, reed canarygrass, and switchgrass. Biomass and Bioenergy 30, 880 891. The Economist, 2007. The end of cheap food, Dec 8th 2007 news article. The Economist. Erickson, J.E., Wood ard, K.R., Sollenberger, L.E., 2012. Optimizing sweet sorghum production for biofuel in the southeastern USA through nitrogen fertilization and top removal. Bioenergy Res. 5, 8694. Erisman, J., Grinsven, H., Leip, A., Mosier, A., Bleeker, A., 2010. Nitrog en and biofuels; an overview of the current state of knowledge. Nutr. Cycling Agroecosyst. 86, 211223.

PAGE 220

220 Fargione, J., Hill, J., Tilman, D., Polasky, S., Hawthorne, P., 2008. Land clearing and the biofuel carbon debt. Science 319, 12351238. Fedenko, J.R., 2011. Biomass yield and composition of potential bioenergy feedstocks. University of Florida, Gainesville, FL. Finell, M., Arshadi, M., Gref, R., 2011. Carbohydrate composition in delayed harvested reed canary grass. Biomass and Bioenergy 35, 10971102. Fo llett, R.F., 2001. Soil management concepts and carbon sequestration in cropland soils. Soil Tillage Res. 61, 77 92. Fornara, D.A., Tilman, D., 2012. Soil carbon sequestration in prairie grasslands increased by chronic nitrogen addition. Ecology 93, 20302 036. Gallaher, R.N., Weldon, C.O., Futral, J.G., 1975. An Aluminum block digester for plant and soil analysis. Soil Sci. Soc. Am. J. 39, 803806. Gillitzer, P.A., Wyse, D.L., Sheaffer, C.C., Taff, S.J., Lehman, C.C., 2013. Biomass production potential of g rasslands in the oak savanna region of Minnesota, USA. Bioenergy Res. 6, 131141. Griffee, P., 2000. Saccharum officinarum Food and Agricultue Organization, Rome, Italy. Griffin, J.L., Jung, G.A., 1983. Leaf and stem forage quality of big bluestem and swi tchgrass. Agron. J. 75, 723. Guretzky, J.A., Biermacher, J.T., Cook, B.J., Kering, M.K., Mosali, J., 2011. Switchgrass for forage and bioenergy: Harvest and nitrogen rate effects on biomass yields and nutrient composition. Plant and Soil 339, 6981. Hamble ton, L.G., 1977. Semiautomated method for simultaneous determination of phosphorus, calcium and crude protein in animal feeds. J. Assoc. Off. Anal. Chem. 60, 845852. Han, G.P., Wu, Q.L., 2004. Comparative properties of sugarcane rind and wood strands for structural composite manufacturing. Forest. Prod. J. 54, 283288. Han, K. J., Pitman, W.D., Kim, M., Day, D.F., Alison, M.W., McCormick, M.E., Aita, G., 2012. Ethanol production potential of sweet sorghum assessed using forage fiber analysis procedures. GC B Bioenergy, n/an/a. Hanna, W.W., Chaparro, C.J., Mathews, B.W., Burns, J.C., Sollenberger, L.E., Carpenter, J.R., 2004. Perennial Pennisetums In: Moser, L.E., Burson, B.L., Sollenberger, L.E. (Eds.), Warm Season (C4) Grasses. ASA CSSA SSSA, Madison, WI, pp. 503 535.

PAGE 221

221 Hatfield, R.D., Jung, H. J.G., Ralph, J., Buxton, D.R., Weimer, P.J., 1994. A comparison of the insoluble residues produced by the Klason lignin and acid detergent lignin procedures. J. Sci. Food Agric. 65, 5158. Heaton, E.A., Dohleman, F.G. Long, S.P., 2009. Seasonal nitrogen dynamics of Miscanthusgiganteus and Panicum virgatum GCB Bioenergy 1, 297307. Hoa, D.T., Man, T.D., Hau, N.G., 2008. Pretreatment of lignocellulosic biomass for enzymatic hydrolysis. ASEAN 25, 341346. Hsu, F., Hong, K., Lee, M., Lee, K., 1990. Effects of cutting height on forage yield, forage and silage quality of napier grasses (written in Chinese with abstract in English). J. Agr. Assoc. China 151, 7788. InmanBamber, N.G., 1994. Temperature and seasonal effects on canopy development and light interception of sugarcane. Field Crops Res. 36, 4151. Ishii, Y., Yamaguchi, N., Idota, S., 2005. Dry matter production and in vitro dry matter digestibility of tillers among napiergrass ( Pennisetum purpureum Schumach) v arieties. Grassland Sci. 51, 153163. Ito, K., Inanaga, S., 1988. Studies on dry matter production of napiergrass : I. Comparison of dry matter productivities and growth parameters between Tokyo and Miyazaki (written in Japanese with abstract in English). Jpn. J. Crop Sci. 57, 9096. Jakob, K., Zhou, F., Paterson, A., 2009. Genetic improvement of C4 grasses as cellulosic biofuel feedstocks. In Vitro Cell. Dev. Biol. Plant 45, 291305. James, G., 2003. Sugarcane / Glyn James. Blackwell Science, Oxford. Jarchow, M.E., Liebman, M., Rawat, V., Anex, R.P., 2012. Functional group and fertilization affect the composition and bioenergy yields of prairie plants. GCB Bioenergy 4, 671679. Jennewein, S., Gilbert, R., Wright, A., Glaz, B., Rowland, D., Bennett, J., Schnell, R., 2012. Physiological and morphological effects of high water tables on giant reed ( Arundo donax ), elephant grass ( Pennisetum purpureum ), energycane and sugarcane ( Saccharum spp.). ASA, CSSA and SSSA Annual Meetings, Cincinnati, OH. Jrgensen, U., Sander, B., 1997. Biomass requirements for power production: how to optimise the quality by agricultural management. Biomass and Bioenergy 12, 145147. Jung, H. J.G., Lamb, J.F.S., 2004. Prediction of cell wall polysaccharide and lignin concentrations of alfalfa stems from detergent fiber analysis. Biomass and Bioenergy 27, 365373.

PAGE 222

222 Jung, H. J.G., Mertens, D.R., Payne, A.J., 1997. Correlation of acid detergent lignin and Klason lignin with digestibility of forage dry matter and neutral detergent fiber. J. Dairy Sci. 80, 16221628. Jung, H. J.G., Varel, V.H., Weimer, P.J., Ralph, J., 1999. Accuracy of Klason lignin and acid detergent lignin methods as assessed by bomb calorimetry. J. Agric. Food. Chem. 47, 20052008. Kering, M.K., Butler, T.J., Biermacher J.T., Guretzky, J.A., 2012. Biomass yield and nutrient removal rates of perennial grasses under nitrogen fertilization. Bioenergy Res. 5, 6170. Kim, M., Day, D., 2011. Composition of sugar cane, energy cane, and sweet sorghum suitable for ethanol produc tion at Louisiana sugar mills. J. Ind. Microbiol. Biotechnol. 38, 803807. Kludze, H., Deen, B., Dutta, A., 2013. Impact of agronomic treatments on fuel characteristics of herbaceous biomass for combustion. Fuel Process. Technol. 109, 96102. Knoll, J.E., Anderson, W.F., Strickland, T.C., Hubbard, R.K., Malik, R., 2012. Low input production of biomass from perennial grasses in the coastal plain of Georgia, USA. Bioenergy Res. 5, 206214. Kubota, F., Matsuda, Y., Agata, W., Nada, K., 1994. The relationship between canopy structure and high productivity in napier grass, Pennisetum purpureum Schumach. Field Crops Res. 38, 105110. Kumar, P., Barrett, D.M., Delwiche, M.J., Stroeve, P., 2009. Methods for pretreatment of lignocellulosic biomass for efficient hydrolysis and biofuel production. Ind. Eng. Chem. Res. 48, 37133729. Langdale, J.A., 2011. C4 cycles: past, present, and future research on C4 photosynthesis. The Plant Cell 23, 38793892. Lee, D.K., Owens, V.N., Boe, A., Jeranyama, P., 2007a. Composition of herbaceous biomass feedstocks. North Central Sun Grant Center, South Dakota State University, Brookings, SD. Lee, D.K., Owens, V.N., Doolittle, J.J., 2007b. Switchgrass and soil carbon sequestration response to ammonium nitrate, manure, and harvest frequency on conservation reserve program land. AGRONOMY JOURNAL 99, 462468. Legendre, B.L., Burner, D.M., 1995. Biomass production of sugarcane cultivars and early generation hybrids. Biomass and Bioenergy 8, 5561. Lemus, R., Lal, R., 2005. Bioenergy crops and carbon sequestration. Crit. Rev. Plant Sci. 24, 121.

PAGE 223

223 Len, R.G., Gilbert, R.A., Korndorfer, P.H., Comstock, J.C., 2012. Selection Criteria and Performance of Energycane Clones ( Saccharum spp. S. spontaneum ) for Biomass Production Under Tropical and Subtropical Conditions. CEIBA 51, 11. Lewandowski, I., Heinz, A., 2003. Delayed harvest of miscanthus influences on biomass quantity and quality and environmental impacts of energy production. Eur. J. Agron. 19, 45 63. Lewandowski, I., Kicherer, A., 1997. Combustion quality of biomass: practical relevance and experiments to modify the biomass quality of Miscanthus x giganteus Eur. J. Agron. 6, 163177. Liebig, M.A., Schmer, M.R., Vogel, K.P., Mitchell, R.B., 2008. Soil carbon storage by switchgrass grown for bioenergy. Bioenergy Res. 1, 215222. Lorenz, A.J., Anex, R.P., Isci, A., Coors, J.G., de Leon, N., Weimer, P.J., 2009a. Forage quality and composition measurements as predictors of ethanol yield from maize ( Zea mays L.) stover. Biotechnol. Biofuels, England, p. 5. Lorenz, A.J., Coors, J.G., de Leon, N., Wolfrum, E.J., Hames, B.R., Sluiter, A.D., Weimer, P.J., 2009b. Characterization, genetic variation, and combining ability of Maize traits relevant to the production of cellulosic ethanol. Crop Sci. 49, 8598. Lowry, J.B., Conlan, L.L., Schlink, A.C., McSweeney, C.S., 1994. Acid detergent dispersible lignin in tropical grasses. J. Sci. Food Agric. 65, 4149. Lu, Y., Mosier, N.S., 2008. Current technologies for fuel ethanol production from lignocellulosic plant biomass. In: Vermerris, W. (Ed.), Genetic improvement of bioenergy crops. Springer New York, pp. 161 182. Ma, Z., Wood, C.W., Bransby, D.I., 2000a. Impacts of soil management on root characteristics of switchgrass. Biomass and Bioenergy 18, 105112. Ma, Z., Wood, C.W., Bransby, D.I., 2000b. Soil management impacts on soil carbon sequestration by switchgrass. Biomass and Bioenergy 18, 469477. Macoon, B., Sollenberger, L.E., Moore, J.E., 2001. Defoliation effects on leaf blade proportion and nutritive value of four Pennisetum genotypes. Soil Crop Sci. Soc. Fla. Proc. 60, 114119. Macoon, B., Sollenberger, L.E., Moore, J.E., 2002. Defoliation effects on persistence and productivity of four Pennisetum spp. genotypes. A gron. J. 94, 541548. Madakadze, I.C., Coulman, B.E., Peterson, P., Stewart, K.A., Samson, R., Smith, D.L., 1998. Leaf area development, light interception, and yield among switchgrass populations in a short season area. Crop Sci. 38, 827834.

PAGE 224

224 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, 696701. Manyawu, G.J., Chakoma, C., Sibanda, S., Mutisi, C., Chakoma, I.C., 2003. The effect of harvesting interval on herbage yield and nutritive value of Napier grass and hybrid Pennisetums Asian Austral. J. Animal Sci. 16, 9961002. Martha, G.B., Corsi, M., Trivelin, P.C.O., Alves, M.C., 2004. Nitrogen recovery and loss in a fertilized elephant gr ass pasture. Grass Forage Sci. 59, 8090. Mathews, J.A., 2008. Opinion: is growing biofuel crops a crime against humanity? Biofuels, Bioprod. Biorefin. 2, 9799. McKendry, P., 2002. Energy production from biomass (part 1): overview of biomass. Bioresour. T echnol. 83, 3746. McLaughlin, S.B., de la Torre Ugarte, D.G., Garten, C.T., Lynd, L.R., Sanderson, M.A., Tolbert, V.R., Wolf, D.D., 2002. High value renewable energy from prairie grasses. Environ. Sci. Technol. 36, 21222129. Mendez Cruz, A.V., Siberio To rres, V., Fernandez van Cleve J., Fontanet E., Ramirez Oliveras, G., 1988. Yield and nutritive value of hay from five tropical grasses at three harvesting intervals. J. Agric. Univ. Puerto Rico 72, 109118. Mislevy, P., F.G., M., B., A., 1989. Changes in elephant grass plant components with maturity 1 percentage dry yield. In: R., D. (Ed.), Proc. Int. Grassl. Congr.16th. The French Grassl. Soc., Nice, France, pp. 951952. Mislevy, P., Martin, F.G., Adjei, M.B., Miller, J.D., 1995. Agronomic characteristic s of US 721153 energycane for biomass. Biomass and Bioenergy 9, 449457. Mitchell, R.B., Schmer, M.R., 2012. Switchgrass harvest and storage. In: Monti, A. (Ed.), Switchgrass: A valuable biomass crop for energy. Springer, London, UK, pp. 113127. Morais, R., Quesada, D., Reis, V., Urquiaga, S., Alves, B.R., Boddey, R., 2012. Contribution of biological nitrogen fixation to Elephant grass ( Pennisetum purpureum Schum.). Plant and Soil 356, 2334. Morrison, I.M., 1980. Hemicellulosic contamination of acid detergent residues and their replacement by cellulose residues in cell wall analysis. J. Sci. Food Agric. 31, 639645. Mukhtar, M., Ishii, Y., Tudsri, S., Idota, S., Sonoda, T., 2003. Dry matter productivity and overwintering ability of the dwarf and normal napiergrasses as affected by the planting density and cutting frequency. Plant Prod. Sci. 6, 6573.

PAGE 225

225 NOAA, 2013. Fall Freeze Occurrence 10% Probability 28 F ( 2 C) or less on an Earlier Date http://www.ncdc.noaa.gov/oa/climate/freezefrost/frostfreemaps.html verified 3/4/2013. National Climatic Data Center. Obreza, T.A., 2003. Prioritizing citrus nutrient management decisions. Soil and Water Science Department, Florida Cooperativ e Extension Service, UF/IFAS, Gainesville, FL. Parrish, D.J., Fike, J.H., 2005. The biology and agronomy of switchgrass for biofuels. Crit. Rev. Plant Sci. 24, 423 459. Perez, J., Munoz Dorado, J., de la Rubia, T., Martinez, J., 2002. Biodegradation and bi ological treatments of cellulose, hemicellulose and lignin. Int. Microbiol. 5, 5363. Perlack, R.D., 2008. Biomass as feedstock for a bioenergy and bioproducts industry: the technical feasability of a billionton annual supply. Oak Ridge National Laborator y, Oak Ridge, TN. Prine, G.M., Dunavin, L.S., Brecke, B.J., Stanley, R.L., Mislevy, P., Kalmbacher, R.S., Hensel, D.R., 1988. Model crop systems: Sorghum, napiergrass. In: Smith, W.H., Frank, J.R. (Eds.), Methane from biomass: A systems approach. Elsevier Applied Science, New York, NY and London, England., pp. 83102. Prine, G.M., Mislevy, P., Shiralipour, A., Smith, P.H., 1984. Elephantgrass, energy crop for humid subtropics., In 4th Annual Solar and Biomass Energy Workshop, Atlanta, Georgia, pp. 124127. Purcell, D.E., Leonard, G.J., O'Shea, M.G., Kokot, S., 2005. A chemometrics investigation of sugarcane plant properties based on the molecular composition of epicuticular wax. Chemometr. Intell. Lab. 76, 135147. Ramesh Sundar, A., Leonard Barnabas, E., M alathi, P., R., V., 2012. A mini review on smut disease of sugarcane caused by Sporisorium scitamineum In: Mworia, J. (Ed.), Botany. Rengsirikul, K., Ishii, Y., Kangvansaichol, K., Pripanapong, P., Sripichitt, P., Punsuvon, V., Vaithanomsat, P., Nakamanee, G., Tudsri, S., 2011. Effects of inter cutting interval on biomass yield, growth components and chemical composition of napiergrass ( Pennisetum purpureum Schumach) cultivars as bioenergy crops in Thailand. Grassland Sci. 57, 135141. Rentizelas, A.A., Tolis, A.J., Tatsiopoulos, I.P., 2009. Logistics issues of biomass: The storage problem and the multi biomass supply chain. Renewable Sustainable Energy Rev. 13, 887894.

PAGE 226

226 Rollin, J.A., Zhu, Z., Sathitsuksanoh, N., Zhang, Y.H.P., 2011. Increasing cellulose accessibility is more important than removing lignin: A comparison of cellulose solvent based lignocellulose fractionation and soaking in aqueous ammonia. Biotechnol. Bioeng. 108, 2230. SAS Institute, 2008. SAS Online DOC 9.2. SAS Inst, Cary, NC. Schank, S.C., Diz, D.A., 1991. A seeded type of hybrid hexaploid elephantgrass with potential for livestock production in the tropics and subtropics. Conf. on Livestock in the Tropics. University of Florida, Gainesville, FL, pp. 713. Searchinger, T., Heimlich, R ., Houghton, R.A., Dong, F., Elobeid, A., Fabiosa, J., Tokgoz, S., Hayes, D., Yu, T.H., 2008. Use of U.S. croplands for biofuels increases greenhouse gases through emissions. Science 319, 12381240. Shahandeh, H., Chou, C.Y., Hons, F.M., Hussey, M.A., 2011 Nutrient partitioning and carbon and nitrogen mineralization of switchgrass plant parts. Commun. Soil Sci. Plant Anal. 42, 599615. Sluiter, A., 2008a. Determination of extractives in biomass; laboratory analytical procedure (LAP). National Renewable Energy Laboratory, Golden, Colorado. Sluiter, A., 2008b. Determination of structural carbohydrates and lignin in biomass;laboratory analytical procedure (LAP). National Renewable Energy Laboratory, Golden, Colorado. Smith, R., Slater, F.M., 2011. Mobilization of minerals and moisture loss during senescence of the energy crops Miscanthusgiganteus Arundo donax and Phalaris arundinacea in Wales, UK. GCB Bioenergy 3, 148157. Smith, W.H., Dowd, M.L., 1981. Biomass Production in Florida. J. Forest. 79, 508515. S okhansanj, S., Mani, S., Turhollow, A., Kumar, A., Bransby, D., Lynd, L., Laser, M., 2009. Largescale production, harvest and logistics of switchgrass ( Panicum virgatum L.) current technology and envisioning a mature technology. Biofuel. Bioprod. Bior. 3, 124141. Sollenberger, L.E., Woodard, K.R., Vendramini, J.M.B., Chase, C.D., Y., L., M., G., Seib, J., K.A., L., Gerardo Cuervo, H., 2011. Are all elephantgrasses invasive? Characterization of natural populations and cultivated types of a bioenergy gras s in Florida., ASA CSSASSSA meeting. ASA/CSSA/SSSA, San Antonio, TX. Somerville, C., Youngs, H., Taylor, C., Davis, S.C., Long, S.P., 2010. Feedstocks for lignocellulosic biofuels. Science 329, 790792. Spitaleri, R.F., Sollenberger, L.E., Schank, S.C., S taples, C.R., 1994. Defoliation effects on agronomic performance of seeded Pennisetum hexaploid hybrids. Agron. J. 86, 695698.

PAGE 227

227 Spitaleri, R.F., Sollenberger, L.E., Staples, C.R., Schank, S.C., 1995. Harvest management effects on ensiling characteristics and silage nutritive value of seeded Pennisetum hexaploid hybrids. Postharvest Biol. Technol. 5, 353362. Stefaniak, T.R., Dahlberg, J.A., Bean, B.W., Dighe, N., Wolfrum, E.J., Rooney, W.L., 2012. Variation in biomass composition components among forage, bi omass, sorghum sudangrass, and aweet sorghum types. Crop Sci. 52, 19491954. Sticklen, M.B., 2008. Plant genetic engineering for biofuel production: towards affordable cellulosic ethanol. Nature Reviews: Genetics 9, 433443. Strezov, V., Evans, T.J., Hayma n, C., 2008. Thermal conversion of elephant grass ( Pennisetum Purpureum Schum) to biogas, biooil and charcoal. Bioresour. Technol. 99, 83948399. Summers, M.D., Jenkins, B.M., Hyde, P.R., Williams, J.F., Scardacci, S.C., 2001. Properties of rice straw as influenced by variety, season and location. ASAE Annual International Meeting, Sacramento, California. Tahir, M.H.N., Casler, M.D., Moore, K.J., Brummer, E.C., 2011. Biomass yield and quality of reed canarygrass under five harvest management systems for b ioenergy production. Bioenergy Res. 4, 111119. Tenenbaum, D.J., 2008. Food vs. fuel: diversion of crops could cause more hunger. Environ Health Perspect 116, A254257. Tessema, Z., Baars, R.M.T., 2004. Chemical composition, in vitro dry matter digestibili ty and ruminal degradation of Napier grass ( Pennisetum purpureum (L.) Schumach.) mixed with different levels of Sesbania sesban (L.) Merr. Anim. Feed Sci. Technol. 117, 2941. Tessema, Z.K., Mihret, J., Solomon, M., 2010. Effect of defoliation frequency and cutting height on growth, dry matter yield and nutritive value of Napier grass ( Pennisetum purpureum (L.) Schumach). Grass Forage Sci. 65, 421 430. Tew, T.L., Cobill, R.M., 2008. Genetic improvement of sugarcane ( Saccharum spp.) as an energy crop. In: Vermerris, W. (Ed.), Genetic improvement of bioenergy crops. Springer New York, pp. 273294. Tew, T.L., Miller, J.D., Dufrene, E.O., Cobill, R.M., Garrison, D.D., White, W.H., Grisham, M.P., Pan, Y.B., Legendre, B.L., Richard, E.P., 2011. Registration of 'Ho CP 91 552' sugarcane. J. Plant Reg. 5, 181190. Thammasouk, K., Tandjo, D., Penner, M.H., 1997. Influence of extractives on the analysis of herbaceous biomass. J. Agric. Food. Chem. 45, 437443. Thompson, J.B., 1919. Napier and merker grasses. Univ. Fla. A gric. Exp. Stn. Bull., Gainesville, FL.

PAGE 228

228 Tran, A., Chambers, R., 1986. Lignin and extractives derived inhibitors in the 2,3butanediol fermentation of mannoserich prehydrolysates. Appl. Microbiol. Biotechnol. 23, 191197. Trcsnyi, Z.K., Fieldsend, A.F., Wolf, D.D., 2009. Yield and canopy characteristics of switchgrass ( Panicum virgatum L.) as influenced by cutting management. Biomass and Bioenergy 33, 442448. U.S. Energy Information Administration, 2013. Europe Brent spot price, available at http://www.eia.gov/petroleum/ (verified 15th May 2013). USDA Economic Research Service, 2012. U.S. domestic corn use. Van Heerwaarden, L.M., Toet, S., Aerts, R., 2003. Current measures of nutrient resorption efficiency lead to a substantial underestimation of real resorption efficiency: facts and solutions. Oikos 101, 664 669. Van Man, N., Wiktorsson, H., 2003. Forage yield, nutritive value, feed intake and digestibility of three grass species as affected by harvest freq uency. Trop. Grasslands 37, 101110. Van Soest, P.J., Robertson, J.B., Lewis, B.A., 1991. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci., United States, pp. 35833597. Velez Santiago, J., ArroyoAguilu, J.A., 1981. Effect of three harvest intervals on yield and nutritive value of seven napiergrass cultivars. J. Agric. Univ. Puerto Rico Agric. Exp. Stn. 65, 129 137. Vermerris, W., 2008a. Composition and biosynthesis of lignocel lulosic biomass. In: Vermerris, W. (Ed.), Genetic improvement of bioenergy crops. Springer New York, pp. 89142. Vermerris, W., 2008b. Why Bioenergy Makes Sense. In: Vermerris, W. (Ed.), Genetic i mprovement of b ioenergy Crops. Springer New York, pp. 342. V iator, R.P., Richard, E.P., 2012. Sugar and energy cane date of planting effects on cane, sucrose, and fiber yields. Biomass and Bioenergy 40, 8285. VicenteChandler, J., Silva, S., Figarella, J., 1959. The Effect of Nitrogen Fertilization and Frequency o f Cutting on the Yield and Composition of Three Tropical Grasses. Agron. J. 51, 202206. Waclawovsky, A.J., Sato, P.M., Lembke, C.G., Moore, P.H., Souza, G.M., 2010. Sugarcane for bioenergy production: an assessment of yield and regulation of sucrose conte nt. Plant Biotechnol. J., England, pp. 263276.

PAGE 229

229 Wadi, A., Ishii, Y., Idota, S., 2003. Effects of the level of fertilizer input on tiller and leaf development in relation with dry matter accumulation of napiergreass and kinggrass. J. Jap. Soc. Grassland. Sci. 49, 311323. Wadi, A., Ishii, Y., Idota, S., 2004. Effects of cutting interval and cutting height on dry matter yield and overwintering ability at the established year in Pennisetum species. Plant Prod. Sci. 7, 8896. Wang, L. P., Jackson, P.A., Lu, X., Fan, Y. H., Foreman, J.W., Chen, X. K., Deng, H.H., Fu, C., Ma, L., Aitken, K.S., 2008. Evaluation of Sugarcane Saccharum spontaneum Progeny for Biomass Composition and Yield. Crop Sci. 48, 951961. Waramit, N., Moore, K.J., Heggenstaller, A.H., 2011. Composition of native warm season grasses for bioenergy production in response to nitrogen fertilization rate and harvest date. Agron. J. 103, 655662. Wedin, D.A., 2004. C4 grasses: resource use, ecology, and global change. In: Moser, L.E., Burson, B.L., Sollenberger, L.E. (Eds.), Warm Season (C4) Grasses. ASA CSSASSSA, Madison, WI, pp. 1550. Williams, M.J., Hanna, W.W., 1995. Performance and nutritive quality of dwarf and semi dwarf elephantgrass genotypes in the southeastern USA. Tropical Grasslands 29, 122127. Wilson, D.M., Dalluge, D.L., Rover, M., Heaton, E.A., Brown, R.C., 2013a. Crop management impacts biofuel quality: Influence of switchgrass harvest time on yield, nitrogen and ash of fast pyrolysis products. Bioenergy Res. 6, 103113. Wilson, D .M., Heaton, E.A., Liebman, M., Moore, K.J., 2013b. Intraseasonal changes in switchgrass nitrogen distribution compared with corn. Agron. J. 105, 285294. Woodard, K.R., Prine, G.M., 1991. Forage yield and nutritive value of elephantgrass as affected by harvest frequency and genotype. Agron. J. 83, 541546. Woodard, K.R., Prine, G.M., 1993a. Dry matter accumulation of elephantgrass, energycane, and elephantmillet in a subtropical climate. Crop Sci. 33, 818824. Woodard, K.R., Prine, G.M., 1993b. Regional performance of tall tropical bunchgrasses in the Southeastern U.S.A. Biomass and Bioenergy 5, 321. Woodard, K.R., Prine, G.M., Bachrein, S., 1993. Solar energy recovery by elephantgrass, energycane, and elephantmillet canopies. Crop Sci. 33, 824830. Woodar d, K.R., Prine, G.M., Bates, D.B., 1991a. Silage characteristics of elephantgrass as affected by harvest frequency and genotype. Agron. J. 83, 547551.

PAGE 230

230 Woodard, K.R., Prine, G.M., Bates, D.B., Chynoweth, D.P., 1991b. Preserving elephantgrass and energycane biomass as silage for energy. Bioresour. Technol. 36, 253259. Woodard, K.R., Sollenberger, L.E., 2008. Production of biofuels crops in Florida: Elephantgrass. Agronomy Department, Florida Cooperative Extension Service, UF/IFAS, Gainesville, FL. Xu, J.L., Chen, Y., Cheng, J.J., Sharma Shivappa, R.R., Burns, J.C., 2011. Delignification of switchgrass cultivars for bioethanol production. Bioresour. Technol. 6, 707720. Zhang, K., Agrawal, M., Harper, J., Chen, R., Koros, W.J., 2011. Removal of the fermentation inhibitor, furfural, using activated carbon in cellulosic ethanol production. Ind. Eng. Chem. Res. 50, 1405514060. Zhao, Y.L., Dolat, A., Steinberger, Y., Wang, X., Osman, A., Xie, G.H., 2009. Biomass yield and changes in chemical composition of sweet sorghum cultivars grown for biofuel. Field Crops Res. 111, 5564.

PAGE 231

231 BIOGRAPHICAL SKETCH ChaeIn Na was born in 1982 in the small town of Goryeong, Gyeongbuk province, South Korea. As a descendent of more than three centuries of diligent farmer s, he loved to work with his parents on a rice and oriental melon farm. He received a Bachelor of Science in A gricultural B iology (20002006) and Master of Science in Agronomy (20062008) from Kyungpook N ational U niversity, South Korea. When he became a sophomore in college he served his country as a member of the Korean A uxiliary P olice (20022003). After he completed his Master of Science degree, h e continued his interest in agriculture and joined the Ph.D program in the Agronomy D epartment of the University of Florida as a Research Assistant in fall 2009 under the guidance of Dr. Lynn E. Sollenberger. He graduated in summer 2013.