Citation
Rhizoma Peanut Proportion in Mixed-Species Pastures with Bahiagrass Affects Nutrient Cycling and Greenhouse Gas Emissions

Material Information

Title:
Rhizoma Peanut Proportion in Mixed-Species Pastures with Bahiagrass Affects Nutrient Cycling and Greenhouse Gas Emissions
Creator:
Kohmann, Marta Moura
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (145 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Agronomy
Committee Chair:
SOLLENBERGER,LYNN E
Committee Co-Chair:
DUBEUX,JOSE CARLOS
Committee Members:
DILORENZO,NICOLAS
SILVEIRA,MARIA LUCIA

Subjects

Subjects / Keywords:
cycling -- legume -- litter -- methane -- nitrous -- nutrient -- oxide
Agronomy -- Dissertations, Academic -- UF
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Agronomy thesis, Ph.D.

Notes

Abstract:
Warm-climate grasslands are N limited and rely on inorganic N inputs. Increasing fertilizer costs limit N application leading in some cases to pasture degradation. Where N fertilizer is used, emissions of greenhouse gases (GHG) occur. Integration of legumes into grasslands provides N to the system at neutral GHG emissions. The objective was to evaluate how presence and proportion of the legume rhizoma peanut (Arachis glabrata Benth.) affect nutrient cycling and GHG emissions in bahiagrass pastures (Paspalum notatum Flugge) relative to use of N fertilizer. In a litter bag study, inclusion of legumes at 33% in mixtures with grass increased decomposition of total aboveground litter relative to an unfertilized grass monoculture. Decomposition of legume started earlier and lasted longer during incubation in mixtures with grass compared with pure legume. At the end of incubation, N mineralization was greater for mixtures when as little as 33% legume was present than for N-fertilized grass, despite similar chemical characteristics of legume and fertilized bahiagrass. Findings suggest that factors other than chemical characteristics, such as microbial community diversity, play an important role in decomposition of mixed-species litter. In a litter decomposition and deposition study, greater legume proportion increased plant litter decomposition and N mineralization relative to fertilized and unfertilized bahiagrass, probably because of greater available N in litter when legumes were present. Aboveground litter deposition was unaffected by pasture composition, but existing litter was greater with less legume because of slower decomposition rates. The effect of including legume in pastures was also evaluated in terms of GHG emissions from cattle excreta. More methane was emitted from dung of animals grazing legume-grass mixtures relative to fertilized bahiagrass, probably because of greater N concentration in dung when legume was present. Emissions of nitrous oxide were greater from urine relative to dung, but were not affected by pasture composition. Fertilized bahiagrass emitted 2.5 times more GHG than legume-grass pastures due to N fertilizer use. These results indicate inclusion of legumes is a promising, sustainable alternative to inorganic fertilization in grass-based pastures, and justifies continuing investment toward developing management strategies that facilitate legume adoption. ( en )
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.
Thesis:
Thesis (Ph.D.)--University of Florida, 2017.
Local:
Adviser: SOLLENBERGER,LYNN E.
Local:
Co-adviser: DUBEUX,JOSE CARLOS.
Statement of Responsibility:
by Marta Moura Kohmann.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
LD1780 2017 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

RHIZOMA PEANUT PROPORTION IN MIXED SPECIES PASTURES WITH BAHIAGRASS AFFECTS NUTRIENT CYCLING AND GREENHOUSE GAS EMISSIONS By MARTA MOURA KOHMANN 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 2017

PAGE 2

2017 Marta Moura Kohmann

PAGE 3

To God, creator of oceans, deserts, forests and grasslands

PAGE 4

4 ACKNOWLEDGMENTS I thank all members of my graduate committee for the dedication and provision of inputs in the development of my research projects. I thank my adviser, Dr. Lynn E. Sollenberger, for the guidance, encouragement, patience, and friendship in the course of my graduate studies, and for continuously challenging me. I would like to also thank the Agronomy Department for the financial and professional support I would like to thank my colleagues for the help in field work and sample analysis, without wh ich this res earch would not have been possible: Parmeshwor Aryal Maristela Bauer, Katie Cooley Courtney Darling, Leonardo Moreno Sabrina Saraiva Daniel Schmidt Liliane Severino da Silva Valdson J. Silva and Erin Stenklyft I appreciate your help, suggestions, a nd commitment to Science. I thank Dwight Thomas and Richard Fethiere for the crucial assistance in field and laboratory work, always accompanied by laughter. I thank my parents, Carlos and Leila Kohmann, for the love, support, and encouragement to pursue a career as a scientist. I would like to thank my siblings, Andr , Laura, Tiago, and Raquel my sister in law Rebeca, and my brothers in law, Leonardo and Lucas, for the continuous love and encouragement even in the distance.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 LIST OF ABBREVIATIONS ................................ ................................ ................................ ........ 11 ABSTRACT ................................ ................................ ................................ ................................ ... 12 CHAPTER 1 OVERVIEW ................................ ................................ ................................ ........................... 14 2 LITERATURE REVIEW ................................ ................................ ................................ ....... 18 Introduction ................................ ................................ ................................ ............................. 18 Pasture Based Forage Livestock Systems ................................ ................................ .............. 19 Legume Based Forage Systems, Climate Change, and Sustainability ............................ 20 Pasture and Animal Responses on Bahiagrass Pastures ................................ .................. 21 Pasture and Animal Responses on Rhizoma Peanut Pastures ................................ ......... 22 Nutrient Cycling in Grazing Systems ................................ ................................ ..................... 26 Litter Dynamics ................................ ................................ ................................ ............... 28 Animal Excreta ................................ ................................ ................................ ................ 33 GHG Emissions from Agricultural Production ................................ ................................ ...... 34 GHG Emissions in Livestock Production ................................ ................................ ............... 35 Influence of Plants on GHG Emissions ................................ ................................ ........... 37 Emissions from Enteric Methane ................................ ................................ .................... 38 Emissions from Animal Urine and Dung ................................ ................................ ........ 39 3 LEGUME PROPORTION IN GRASSLAND LITTER AFFECTS DECOMPOSITION AND NUTRIENT MINERALIZATION ................................ ................................ ............... 44 Introduction ................................ ................................ ................................ ............................. 44 Material and Methods ................................ ................................ ................................ ............. 45 Experimental Site ................................ ................................ ................................ ............ 45 Treatments and Experimental Design ................................ ................................ ............. 45 Litter Collection, Samples and Site Preparation ................................ .............................. 46 Litter Decomposition and Nutrient Disappearance ................................ ......................... 47 Statistical Analysis ................................ ................................ ................................ .......... 49 Results and Discussion ................................ ................................ ................................ ........... 50 Initial Litter Composition ................................ ................................ ................................ 50 Total Biomass Decomposition Rate, Relative Half Life, and Extent of Decomposition ................................ ................................ ................................ ............. 50

PAGE 6

6 Disappearance of Individual Species in Mixtures ................................ ........................... 53 Remaining N and Litter C:N Ratio ................................ ................................ .................. 55 Lignin, Lignin:N, and Lignin:ADIN ................................ ................................ ............... 58 ADIN Concentration in the OM and in Total N ................................ .............................. 61 Remaining P and C:P ................................ ................................ ................................ ....... 62 Conclusions ................................ ................................ ................................ ............................. 63 4 NITROGEN FERTILIZATION AND PROPORTION OF LEGUME AFFECT LITTER DEPOSITION, DECOMPOSITION, AND NUTRIENT RETURN IN BAHIAGRASS PASTURES ................................ ................................ ................................ ............................ 72 Introduction ................................ ................................ ................................ ............................. 72 Material and Methods ................................ ................................ ................................ ............. 73 Experimental Site ................................ ................................ ................................ ............ 73 Treatments and Experimental Design ................................ ................................ ............. 74 Grazing Management ................................ ................................ ................................ ...... 74 Litter Deposition ................................ ................................ ................................ .............. 74 Litter Decomposition ................................ ................................ ................................ ....... 75 Litter a nd Nutrient Disappearance ................................ ................................ ................... 76 Chemical Composition Analysis ................................ ................................ ..................... 77 Statistical Analysis ................................ ................................ ................................ .......... 77 Results and Discussion ................................ ................................ ................................ ........... 78 Initia l Litter Composition ................................ ................................ ................................ 78 Litter Decomposition ................................ ................................ ................................ ....... 80 Change in Litter Composition During Incubation ................................ ........................... 85 Existing Litter Mass and Litter Deposition Rate ................................ ............................. 88 Existing and Deposited Litter Composition ................................ ................................ .... 90 Conclusions ................................ ................................ ................................ ............................. 90 5 NITROUS OXIDE AND METHANE EMISSIONS FROM CATTLE URINE AND DUNG IN N FERTILIZED GRASS AND LEGUME GRASS SWARDS ........................... 97 Introduction ................................ ................................ ................................ ............................. 97 Material and Met hods ................................ ................................ ................................ ............. 99 Experimental Site ................................ ................................ ................................ ............ 99 Treatments and Experimental Design ................................ ................................ ........... 100 Static Chamber Design and Installation ................................ ................................ ........ 100 Excreta Collection and Chemical Composition ................................ ............................ 101 Excreta Application ................................ ................................ ................................ ....... 102 Gas Sampling and Analysis ................................ ................................ ........................... 103 Gas Flux Calculation ................................ ................................ ................................ ..... 104 Soil Measurements ................................ ................................ ................................ ........ 105 Statistical Analysis ................................ ................................ ................................ ........ 105 Results and Discussion ................................ ................................ ................................ ......... 105 Methane Emissions ................................ ................................ ................................ ........ 105 Nitrous Ox ide Emissions ................................ ................................ ............................... 110 Effect of Pasture Management on Overall GHG Emissions ................................ ......... 115

PAGE 7

7 Conclusion ................................ ................................ ................................ ............................ 116 6 SUMMARY ................................ ................................ ................................ .......................... 122 LIST OF REFERENCES ................................ ................................ ................................ ............. 127 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 145

PAGE 8

8 LIST OF TABLES Table page 3 1 Chemical composition of plant litter at beginning of incubati on for 2 yr in Gainesv ille, FL. ................................ ................................ ................................ .................. 65 3 2 Chemical composition of plant litter at end of incubati on for 2 yr in Gainesville, FL. ..... 65 5 1 Composition of fresh dung and urine from animals consuming bahiagrass fertilized with 50 kg N ha 1 yr 1 (BGN) or bahiagras s rhizoma peanut mixed pastures (RP BG), average d over 2 yr in Gainesville, FL. ................................ ................................ ............. 118 5 2 Gas sampling frequency schedule for N 2 O and CH 4 flux measurements over 2 yr in Gainesville, FL. ................................ ................................ ................................ ................ 118

PAGE 9

9 LIST OF FIGURES Figure page 3 1 Monthly weather data at Hague, FL (18 km from experimental site) during the years of evaluation and the 30 yr average for Gainesville, FL. ................................ .................. 66 3 2 Remaining biomass and N decay of plant litter during 128 d incubation period for 2 yr in Gainesville, FL. ................................ ................................ ................................ ......... 67 3 3 Remaining rhizoma peanut biomass for mixtures of plant litter during 128 d incubation period over 2 yr in Gainesville, FL. ................................ ................................ 68 3 4 Carbon:N ratio of plant litter during 128 d incubation period for 2 yr in Gainesville, FL. ................................ ................................ ................................ ................................ ...... 68 3 5 Lignin concentration of plant litter during 128 d incubation period for 2 yr in Gainesville, FL. ................................ ................................ ................................ .................. 69 3 6 Lignin:N ratio of plant litter d uring 128 d incubation period for 2 yr in Gainesville, FL. ................................ ................................ ................................ ................................ ...... 69 3 7 Lignin:ADIN ratio of plant litter during 128 d incubation period for 2 yr in Gainesville, FL. ................................ ................................ ................................ .................. 70 3 8 Carbon:P ratio of plant litter during 128 d incubation period for 2 yr in Gainesville, FL. ................................ ................................ ................................ ................................ ...... 71 4 1 Rainfall and temperature data from the Alachua (FL) site of the Florida Automated Weather Network (FAWN). ................................ ................................ ............................... 92 4 2 Nitrogen concentration, C:N and lignin:N ratio, and acid detergent insoluble N concentration in total N for plant litter at beginning (Day 0) and end (Day 128) of incubation during 2 yr in Gainesville, FL. ................................ ................................ ......... 93 4 3 Plant litter biomass and N relative decay rates ( k ) of plant litter incubated for 128 d during 2 yr in Gainesville, FL. ................................ ................................ ........................... 94 4 4 Plant litter biomass and remaining N after 128 d of incu bation over 2 yr in Gainesville, FL. ................................ ................................ ................................ .................. 95 4 5 Litter deposition rate and existing litter mass in August and October of 2 yr in Gaine sville, FL. ................................ ................................ ................................ .................. 96 5 1 Monthly average temperature and accumulated rainfall data at Alachua, FL for the years of evaluation and the 3 0 yr average for Gainesville, FL. ................................ ....... 119

PAGE 10

10 5 2 Flux of CH 4 from dung of animals grazing bahiagrass pastures fertilized with 50 kg N ha 1 (BGN) or mixed rhizoma peanut bahiagrass (RP BG) over 2 yr in Gainesville, FL. ................................ ................................ ................................ ................................ .... 119 5 3 Percentage water filled pore space (WFPS ) in the soil with no treatment application (blank) and with application of dung or urine. ................................ ................................ 120 5 4 Cumulative CH 4 emissions from dung of animals grazing bahiagrass pastures fertilized with 50 kg N ha 1 (BGN) or mixed rhizoma peanut bahiagrass (RP BG) over 2 yr in Gainesville, FL. ................................ ................................ ............................ 120 5 5 Flux of N 2 O from dung and urine of animals grazing bahiagrass pastures fertilized with 50 kg N ha 1 (BGN) or mixed rhizoma peanut bahiagrass (RP BG) over 2 yr in Gainesville, FL. ................................ ................................ ................................ ................ 121 5 6 Cumulative N 2 O emissions from dung and urine of animal s grazing bahiagrass pastures fertilized w ith 50 kg N ha 1 (BGN) or mixed rhizoma peanut bahiagrass (RP BG) over 2 yr in Gainesville, FL. ................................ ................................ ............. 121

PAGE 11

11 LIST OF ABBREVIATIONS ADIN Acid detergent insoluble nitrogen ADF Acid detergent fiber AU Animal units CP Crude protei n DM Dry matter EF Emission factor GHG Greenhouse gas LW Live weight OM Organic matter

PAGE 12

12 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 RHIZOMA PEANUT PROPORTION IN MIXED SPECIES PASTURES WITH BAHIAGRASS AFFECTS NUTRIENT CYCLING AND GREENHOUSE GAS EMISSIONS By Marta Moura Kohmann December 20 17 Chair: Lynn E. Sollenberger Cochair: Jos C. B. Dubeux Jr. Major: Agronomy Warm climate grasslands are N limited and rely on inorganic N inputs. Increasing fertilizer costs limit N application leading in some cases to pasture degradation. Where N fe rtilizer is used, emissions of greenhouse gases (GHG) occur. Integration of legumes into grasslands provides N to the system at neutral GHG emissions. The objective was to evaluate how presence and proportion of the legume rhizoma peanut ( Arachis glabrata Benth.) affect nutrient cycling and GHG emissions in bahiagrass pastures ( Paspalum notatum Fl gge) relative to use of N fertilizer. In a litter bag study, inclusion of legumes at 33% in mixtures with grass increased decomposition of total aboveground litte r relative to an unfertilized grass monoculture. Decomposition of legume started earlier and lasted longer during incubation in mixtures with grass compared with pure legume. At the end of incubation, N mineralization was greater for mixtures when as littl e as 33% legume was present than for N fertilized grass despite similar chemical characteristics of legume and fertilized bahiagrass. Findings suggest that factors other than chemical characteristics, such as microbial community diversity, play an important role in decomposition of mixed species litter. In a l itter decomposition and deposition study, greater legume proportion increased plant litter decomposition and N mineralization relative to fertilized and unfertilized

PAGE 13

13 bahiagrass, probably because of greater available N in litter when legumes were present. A boveground litter deposition was unaffected by pasture composition, but existing litter was greater with less legume because of slower decomposition rates. The effect of including legume in pastures was also evaluated in terms of GHG emissions from cattle excreta. More methane was emitted from dung of animals grazing legume grass mixtures relative to fertilized bahiagrass, probably because of greater N concentration in dung when legume was presen t. Emissions of nitrous oxide were greater from urine relative to dung, but were not affected by pasture composition. Fertilized bahiagrass emitted 2.5 times more GHG than legume grass pastures due to N fertilizer use. These results indicate inclusion of legumes is a promising, sustainable alternative to inorganic fe rtilization in grass based pastures, and justif ies continu ing investment toward developing management strategies that facilitate legume adoption.

PAGE 14

14 CHAPTER 1 OVERVIEW The beef cattle industry in Florida is based on warm season perennial grasses which depend on nitrogen (N) fertilization for production Nitrogen is a key element in plant production, but N fertilizer costs have risen in recent years (Sollenberger et al., 2014) and producers have decreased or eliminat ed N inputs to grass based systems. Lack of maintenance fertilization, especially N, and poor grazing management have resulted in the degradation of perennial grasslands in warm climate environments where low fertility soils predominate (Braz et al., 2013) Recently in Florida there has been increasing concern regarding pasture decline, especially of bahiagrass ( Paspalum notatum Fl gge) ( Silveira et al., 2017 ) Degraded pastureland has limited potential to provide regulating, provisioning, and supporting ecosystem services (Sollenberger, 2014). Association of N fixing legumes with grasses has been referred to as one of the best alternatives for increasing sus tainability in grazing systems (Muir et al., 2011), and is an economically attractive opportunity for improving pasture nutritive value and animal production (M uir et al., 2014). G rass legume (perennial ryegrass [ Lolium perenne L.] and orchardgrass [ Dactyl is glomerata L. ] red clover [ Trifolium pr a tense L. ] and white clover [ T. repens L.] ) mixtures (50/50 to 30/70%) fertilized with 50 kg N ha 1 ha d herbage accumulation equivalent to grass monocultures fertilized with 450 kg ha 1 (Nyfeler et al., 2009), whil e total N in harvested biomass was similar to that of legume monocultures when the legume grass proportion was 40/60% (Nyfeler et al., 2011) In the southeastern USA, significant efforts have been made in developing new cultivars of rhizoma peanut ( Arachis glabrata Benth ) and promoting successful management strategies for their use in mixed swards (Ortega S. et al., 1992; Castillo et al., 2013 ; 2014 ; Mullenix et al., 2016a; 2016b ) However, there has been very limited effort to date to

PAGE 15

15 understand the impac t of legume inclusion on nutrient cycling and greenhouse gas (GHG) emissions of pastures including rhizoma peanut specifically, and warm climate legumes in general, in mixtures with grasses. An important component of nutrient cycling in grasslands is litt er decomposition. Litter decomposition can be affected by litter chemical characteristics such as carbon (C):N (Thomas and Asakawa, 1993; Rezende et al., 1999), lignin concentration (Thomas and Asakawa, 1993; Silva et al., 2012), and by environmental char acteristics such as temperature (Allison et al., 2013), rainfall (Silva et al., 2012), and ecosystem microbial abundance and community composition (Allison et al., 2013; Chapman et al., 2013). The inclusion of legumes in grasslands affect s herbage product ion and chemical composition (Nyfeler et al., 2009 ; 2011), which in turn impacts decomposition of plant litter nutrient mineralization, and nutrient cycling. Although lower C:N in grass legume mixtures than grass monocultures (~ 33 vs. 84, respectively) results in greater N mineralization from mixed litter versus that from grass monocultures (Silva et al., 2012), N con centration alone is not the only factor a ffecting litter decomposition processes in mixtures (Smith and Bradford, 2014). A n increase in dec omposition rate of multi species stands has been attributed to presence of divers e plant functional groups rather than species richness alone (Scherer Lorenzen, 2008) In addition, occurrence of non additive effects in litter decomposition of multiple spec ies was strongly associated with environmental microbial community composition (Chapman et al., 2013). These observations indicate it is not possible to predict decomposition of multi species litter based on single species decomposition studies, further su pporting the need for evaluating the effect of grass legume mixtures on nutrient cycling.

PAGE 16

16 The introduction of legumes in to previously N fertilized grass monoculture grazing systems has the potential to decrease fossil fuel use and GHG emissions related t o N fertilizer production and application (Jensen et al., 2012). In the context of climate change, it is important to recognize that ruminant livestock are both important emitters of GHG and valued produce rs of high quality human food wh ile consuming fibro us feed that is unsuitable for human consumption (Beauchemin et al., 2010). T he use of legumes in ruminant livestock diet s in warm climates has the potential to mitigate methane (CH 4 ) emissions from enteric fermentation, which are generally greater f or ani mals grazing warm season grasses relative to legumes or cool season grasses (Archimde et al., 2011). However, manipulation of animal diet s in ways that increase N output can have either positive or negative effects on enteric CH 4 emissions depending on fi ber level (Dijkstra et al., 2011). In addition to CO 2 important GHG emissions in animal production are CH 4 from dung and nitrous oxide ( N 2 O ) from urine and dung (van der Weerden et al., 2011). Increas ing dietary N may cause an increase in N output in animal excreta (Dijkstra et al., 2011 ; 2013), and N availability is thought to be the greatest driver of N 2 O emissions from animal urine and dung (Oenema et al., 1997). Low C:N in manure has also been related to an increase in CH 4 from animal dung (Jarvis et al., 1995). These relationship s ha ve led some researchers to recommend management strategies that reduce N output in animal excreta, including reduc ing dietary N and better match ing dietary N to animal pro tein requirements (Dijkstra et al., 2011 ; 2013). The complexity of tradeoffs in GHG emissions in whole production systems is enhanced by environmental factors including temperature (Bertram et al., 2010; Rafique et al., 2011; Mazzetto et al., 2014), rainf all (Mazzetto et al., 2014), soil compaction (Oenema et al., 1997), and soil water filled pore space (Rafique et al., 2011; van der Weerden et al., 2011), all of which

PAGE 17

17 influence N 2 O and CH 4 emissions from animal excreta. As N 2 O has a greater global warming potential (GWP) relative to CH 4 (265 vs. 28 times that of CO 2 respectively) (IPCC, 2014), it is possible that management strategies that increase N in animal diet s (such as the inclusion of legumes) with a focus o n reducing CH 4 emissions from enteric fe rmentation will be offset by emissions of N 2 O and CH 4 from animal dung and urine. Despite a long term emphasis on use of legumes in grasslands, limited research has been conducted to quantify the effect of legume introduction on nutrient cycling and GHG emissions of grazing systems. The overall objective of th is research is to evaluate the effect of rhizoma peanut bahiagrass mixtures vs. N fertilized bahiagrass on nutrient cycling in plant litter and GHG emissions from animal excreta Sp ecifically, the st udies conducted aim to quantify the effect of different proportions of rhizoma peanut and bahiagrass in mixed swards vs. N fertilized bahiagrass on litter deposition and decomposition rate, CH 4 emissions from dung, and N 2 O emissions from urine and dung.

PAGE 18

18 CHAPTER 2 LITERATURE REVIEW Introduction free land surface (Peters et al., 2013). In the U.S. Gulf Coast region, perennial warm season grasses are the main source of feed for livestock (Vendr amini et al., 2010), and in Florida around 4.5 million ha are grasslands (Vendramini, 2010). Planted grasslands are dominated by bahiagrass ( Paspalum notatum Fl gge ), with over one million hectares (Chambliss and Sollenberger, 1991), and bermudagrass [ Cyno don dactylon (L.) Pers.] with 128,000 ha harvested for hay (Blount et al., 2011). The beef cattle industry in Florida is based on warm season perennial grasses that depend on N fertilization for productivity and persistence A s N fertilizer costs have ri sen in recent years (Sollenberger et al., 2014), producers have decreased N inputs to grass based systems. Lack of maintenance fertilization, especially N, and poor grazing management have resulted in the degradation of perennial grasslands in warm climate environments where low fertility soils predominate (Braz et al., 2013). Recently in Florida there has been increasing concern regarding pasture decline, especially of bahiagrass (Silveira et al., 2017) Degraded pastureland has limited potential to provid e regulating, provisioning, and supporting ecosystem services (Sollenberger, 2014). Association of N fixing legumes with grasses is an economically attractive opportunity for improving pasture nutritive value and animal production (Muir et al., 2014), an d it has been referred to as one of the best alternatives for increasing sustainability in grazing systems (Muir et al., 2011). Nitrogen transfer from legumes to grasses can occur through nodule turnover, legume root exudates (Ta et al., 1986; Ta and Faris 1987), mycorrhizal fungi (Rogers et al., 2001),

PAGE 19

19 animal excreta (Ledgard, 1991), and aboveground legume litter decomposition (Rezende et al., 1999; Silva et al., 2012). Jensen et al. (2012) indicates that legume N input is greenhouse gas (GHG) neutral, wh ile production, application, and use of inorganic N fertilizers can significantly contribute GHG emissions (Lal, 2004; de Klein et al., 2006). In the southeastern USA, significant efforts have been made in developing new cultivars of rhizoma peanut ( Arach is glabrata Benth.) and promoting successful management strategies in mixed swards (Ortega S. et al., 1992; Castillo et al., 2013a; b, 2014, Mullenix et al., 2016a; b). Although rhizoma peanut is the most important perennial forage legume in Florida (Solle nberger et al., 2014), there has been very limited effort to date to understand the impact on nutrient cycling and GHG emissions of including this legume, and warm climate legumes in general, in mixtures with grasses in comparison with N fertilization. Pa sture Based Forage Livestock S ystems Livestock production occurs on around 78% of the land used in agricultural production, Wassenaar, 2007). Of this area, 2 bil lion hectares are used in extensive grazing, 1.4 in highly productive pastures, and 0.5 in feed crop production (Steinfeld and Wassenaar, 2007), with 25% of ice free land area being permanent pasture (Peters et al., 2013). In the state of Florida, around 4 .5 million hectares (ha) are covered with grasslands (Vendramini, 2010) dominated by bahiagrass ( Paspalum notatum Flgge), which covers over one million hectares (Chambliss and Sollenberger, 1991; Inyang et al., 2010). Bahiagrass, particularly Pensacola, i s extensively used for grazing and hay production (Sollenberger and Jones, 1989; Twidwell et al., 1998) because of its adaptability and performance under different management strategies, varying weather conditions (Inyang et al., 2010), and soil types (Cha mbliss and Sollenberger, 1991; Twidwell et al., 1998), performing well even under heavy grazing and low soil fertility levels (Twidwell et

PAGE 20

20 al., 1998). Bermudagrass also covers an extensive area in Florida, with 128,000 ha harvested for hay (Blount et al., 2011). Cow calf operations are mostly managed on bahiagrass pastures with continuous stocking at fixed stocking rates (Inyang et al., 2010). Legume Based Forage Systems, Climate Change, and S ustainability Some of the feasible alternative investments in grazing systems include using low cost inputs or adopting management strategies that decrease external input dependency, which increase their sustainability and resilience (Muir et al., 2011). M ixed grass legume production systems are considered one of the best alternatives to increase sustainability in grazing systems through reduced external input dependency (Muir et al., 2011). Legumes in mixtures with grasses can be an important N source (Nyf eler et al., 2009, 2011) while also providing forage with greater nutritive value and potential animal performance (Muir et al., 2014). In addition, inclusion of legumes in grasslands has the potential to decrease GHG emissions because of reduced need of N fertilizers (Jensen et al., 2012) and less emissions from ruminant enteric fermentation (Archimde et al., 2011). Legumes fix N 2 through a symbiosis with bacteria which is usually highly specific (Caetano Anolls and Gresshoff, 1991) and that occurs wit hin a context where bacteria fix N 2 for plants in exchange for photosynthetic compounds (Libault, 2014). Nodulation occurs upon the infection of plants by bacteria belonging to varying genera (Jensen et al., 2012; Libault, 2014). The nodule is considered a plant root organ, with the N 2 fixing bacteria within bacteroids (Libault, 2014). Nodule formation is the result of a complex interaction of gene induction and expression of both bacteria and plants (Caetano Anolls and Gresshoff, 1991). Nodules can be cla ssified as C sinks in plants (Libault, 2014), and their number is regulated by autoregulation of nodulation (Caetano Anolls and Gresshoff, 1991). This ability to regulat e the number of nodules is a trait that clearly differentiates this relationship from pathogenic plant bacteria

PAGE 21

21 interaction (Libault, 2014) and is a tool whereby plants avoid detrimental effects of excessive number of nodules, such as inhibition of plant growth (Sasaki et al., 2014). Pasture and Animal Responses on Bahiagrass P astures Eval uating the effect of stocking rates on bahiagrass pastures, Inyang et al. (2010) found a linear decrease in herbage mass of 5.9 to 3.2 Mg ha 1 with increasing stocking rates of 4, 8 and 12 heifers ha 1 Inyang et al. (2010) found bahiagrass herbage accumul ation rates ranging from 38 to 158 kg ha 1 d 1 with lower values of accumulation when rainfall was limiting growth or excessive to the point of restricting oxygen availability in the soil. Crude protein (CP) and in vitro digestible organic matter (IVDOM) concentrations ranged from 104 to 165 g kg 1 and 482 to 578 g kg 1 respectively. Stewart et al. (2007) evaluated animal and plant responses on continuously stocked bahiagrass grazed under varying levels of stocking rates (1.2, 2.4, and 3.6 AU ha 1 ) and N application levels (40, 120, and 360 kg N ha 1 ) to apply three contrasting management practices (low, medium and high). The authors found that herbage mass decreased from 3.42 to 2.95 Mg ha 1 with increasing stocking rate, despite the increase in N appli cation rate on higher stocking rate treatments. Animal average daily gain (ADG) under these contrasting management practices was 0.34, 0.35 and 0.28 kg for low, moderate and high management levels. These resulted in 101, 208 and 252 kg ha 1 of animal gain. Bahiagrass under high management levels had greater CP and IVDOM, achieving 140 and 505 g kg 1 in comparison with 113 and 473 g kg 1 for moderate and 99 and 459 g kg 1 respectively, for low management intensities. Sollenberger and Jones (1989) evaluate d animal and plant responses of bahiagrass under rotational stocking with variable stocking rate. They found an ADG of 0.38 kg day 1 and gain per hectare of 318 kg ha 1 across three grazing seasons for 15 to 18 month old steers grazing bahiagrass that rec eived an average of 180 kg N ha 1 yr 1 In that experiment, bahiagrass

PAGE 22

22 produced 7.1 Mg ha 1 in a period of 168 d and nutritive value of hand plucked samples averaged 116 g kg 1 for CP and 583 g kg 1 for IVDOM. Similar values of ADG were observed by Sollenberger et al. ( 1989b) on rotationally stocked bahiagrass. A decrease in bahiagrass IVDOM is usually observed in late summer and early fall (Sollenberger and Jones, 1989; Sollenberger et al., 1989b). Variations in digestibility of bahiagrass and other tropical grasses are strongly related to temperature, but the relevance of this influence differs with species, regrowth period, and soil moisture (Henderson and Robinson, 1982). Quantifying the response of bahiagrass to N applications of 0, 39, 78, 118 and 157 kg N ha 1 applied after harvest in 28 d intervals, Johnson et al. (2001) observed a quadratic response of herbage accumulation to N fertilization rate. Bahiagrass herbage accumulation was greatest with the application of 78 kg N ha 1 reaching an average of 1.5 Mg ha 1 per cut. Values of IVDOM were between 512 and 526 g kg 1 with greater values in June and August and a reduction in digestibility in July. Total N concentration and available N for ru men microbial protein synthesis increased with N fertilization (Johnson et al., 2001). Pasture and Animal Responses on Rhizoma P eanut Pastures Rhizoma peanut was first brought to the USA from Brazil in 1936 (Kerridge and Hardy, 1994; Baker et al., 1999). Around 10,000 ha of rhizoma peanut were cultivated in Alabama, Georgia and Florida 15 years ago (Baker et al., 1999), but that number has increased significantly in recent years as former row crop land in north Florida has been converted to hay production of rhizoma peanut (Dr. Ann Blount, personal communication). The cultivar new cultivars and germplasms were released in 2010 and are beginning to occupy a signifi cant land area (Quesenberry et al., 2010).

PAGE 23

23 Rhizoma peanut is the most used tropical legume in the southern USA because of its high yields, nutritional value, persistence and use plasticity, and suitability for hay production, grazing, soil cover, silage, ornamental uses, and others (Kerridge and Hardy, 1994; Williams et al., 2005). It is tolerant to drought, developing well in sandy, low fertility soils (Baker et al., 1999). Rhizoma peanut has high palatability and rhizoma peanut hay costs less than alfalf a imported from the western USA, a key factor turning rhizoma peanut into a competitive species in the hay market in the Gulf Coast (Baker et al., 1999; Williams et al., 2005). Ortega S. et al. (1992) demonstrated that rhizoma peanut is also adapted to a v ariety of grazing managements, persist ing under a wide combination of grazing frequencies and residual biomass. Rhizoma peanut is a tetraploid species (Krapovickas and Gregory, 2006). It produces flowers aboveground which result in the development of the pod underground (Quesenberry et al., 2010). It is perennial, rhizomatous, with a procumbent habit and tetrafoliolate leaves with a smooth upper surface (Krapovickas and Gregory, 2006). The bi articulate fruits develop underground and have a smooth pericarp producing two to five seeds that sometimes abort into one seed only (Krapovickas and Gregory, 2006). Early selections such as Arb and Arblick were made available to producers in the 1960s but were not very successful among farmers because of their low y ield and slow establishment greater dry matter production and helped popularize the use of rhizoma peanut (Williams et al., 2005). More recently, the University o used for forage (Quesenberry et al., 2010) and germplasms Arblick and Ecoturf for forage or ornamental use (Prine et al., 2010). The ornamental aspect of these genotypes is related to the more dec umbent growth habit and bright, attractive flowers that emerge daily, however, they

PAGE 24

24 achieve comparable or slightly lesser herbage accumulation than released cultivars (Prine et al., gainst weeds, with high yields and good tolerance against vir al diseases (Quesenberry et al., 2010). Texas AgriLife climates, being resistant to late spring frost ev ents and emerging early in spring (Muir et al., 2010). Some of the challenges involved in the cultivation of rhizoma peanut include the need of using vegetative propagation and its slow establishment in the first year. Castillo et al. (2014) tested sod re moval, no till, tillage alone, and tillage after application of glyphosate as seedbed preparation methods to introduce rhizoma peanut through strip planting into pre existing bahiagrass sod. Authors found that sprout emergenc e was greater in treatments tha t included tillage, but there was also a strong trend to obtain greater longer term canopy coverage when the no till strategy was used. Controlling weeds in the planted strips is also crucial to its good establishment, and successful control has been obser ved with the use of imazapic and imazapic plus 2,4 D (Castillo et al., 2013a; b; 2014). Temperature, rainfall and daylength also drive herbage accumulation in rhizoma peanut (Valencia et al., 2001). Yield for rhizoma peanut pure stands ranges from 8310 t o 13700 Mg ha 1 (Quesenberry et al., 2010) and can be increased through irrigation (Butler et al., 2007). In a 2 yr cutting study evaluating Latitude 34, UF Tito, UF Peace, Florigraze, Arbrook, Ecoturf, and Arblick, Dubeux et al. (2017) f o und that Arblick, UF Tito, and UF Peace consistently produced greater DM accumulation than the other cultivars evaluated, reaching an average annual herbage accumulation of more than 10 Mg ha 1 yr 1 In addition, Ecoturf and Latitude 34 had superior root rhizom e biomass, i ndicating good potential for use in grazing. Herbage CP range d from 186

PAGE 25

25 to 204 g kg 1 with lower values observed with short stubble heights (Butler et al., 2007). Cathey et al. (2013) indicated that N provided through symbiotic association was not enough to maximize N accumulation and storage due to the observed increase in shoot yield with increasing doses of N fertilizer. However, in a recent study Dubeux et al. (2017) found that on average 86% of N in rhizoma peanut monocultures originated from biologic al fixation, resulting in ~200 kg N ha 1 yr 1 Williams (1994) evaluated bahiagrass receiving N fertilization rates of 0, 56, and 112 N kg ha 1 in association with rhizoma peanut. The authors observed that rhizoma peanut had a great er herbage accumulation in the spring when compared with bahiagrass, probably due to temperature during that season. The authors concluded that in mixed swards, rhizoma peanut can possibly compete with bahiagrass if rain is plentiful during spring and N fe rtilization is not applied. Nutritive value of rhizoma peanut is great er than that of other warm climate forages (Dubeux et al., 2017), and has been suggested to be responsible for high animal performance in grazing trials (Sollenberger et al., 1989a). In vitro digestible OM (IV D OM) of Arblick, Arbrook, Ecoturf, Florigraze, Latitude 34, UF Peace, and UF Tito averaged 713 g kg 1 in a 2 yr cutting study, and IV D OM w as not lower than 645 g kg 1 in that study (Dubeux et al., 2017). In a study evaluating Ecotur f, UF Tito, and UF Peace with grazing every 3 and 6 wk removing 50 and 75% of aboveground biomass, Mullenix et al. (2016a) found IV D OM and CP greater than 660 and 14 0 g kg 1 respectively. In mixed rhizoma peanut grass pastures, CP and IV D OM w ere in the r ange of 140 to 200 and 540 to 650 g kg 1 respectively, and w ere not affected by fertilization with 35 kg N ha 1 (Valencia et al., 2001). Performance of animals grazing nearly pure rhizoma peanut pastures has been evaluated for different animal types. Sol lenberger et al. ( 1989a) evaluated animal performance of 15 to 18

PAGE 26

26 month old steers grazing rhizoma peanut and found ADG of 0.97 kg day 1 This value was significantly and consistently larger than the ones found by the same animal type grazing bahiagrass pastures, which reached 0.35 kg day 1 The authors credited this difference to the greater IVDOM and CP of the rhizoma peanut relative to bahiagrass during the summer grazing season. Hernndez Garay et al. (2004) tested Arbrook and Florigraze rhizoma peanu t cultivars under grazing by Holstein replacement heifers with continuous stocking during 3 years. They found similar ADG for the cultivars except for Year 3, when Arbrook presented lower ADG due to its smaller proportion in herbage mass. Hernndez Garay e t al. (2004) also found greater CP and IVDOM for Florigraze in the third year due to less grass invasion in the pasture than in Arbrook swards. Fike et al. (2003) evaluated dairy cattle performance and intake when grazing bermudagrass or rhizoma peanut sta nds at two fixed stocking rates. The authors found herbage mass for rhizoma peanut of 3920 and 2440 kg ha 1 for pre and post graze measurements, both lower than the ones achieved by bermudagrass of 6760 and 5460 kg ha 1 Bennett et al. (1995) found ADG of 0.49 and 0.94 kg d 1 for crossbred steers grazing mixed rhizoma peanut pastures during growing and finishing phases, respectively. Valencia et al. (2001) evaluated the effect of N fertilization (0 or 35 kg N ha 1 ) in spring and stocking rate (1.5 or 2.5 b ulls ha 1 ) in summer on mixed rhizoma peanut grass pastures. The authors found no effect of treatments on herbage mass (3.0 to 3.4 Mg ha 1 ) or ADG (0.75 kg d 1 in spring and 0.53 kg d 1 in summer). Nutrient Cycling in Grazing Systems In grazing systems, t he main pools in nutrient cycling are soil, live plant biomass and plant litter, animals, animal excreta, and atmosphere (Dubeux et al., 2007; Vendramini et al., 2014). Cycling among these pools, particularly through decomposition of plant litter and anima l excreta, is crucial to the maintenance of warm season grasslands which typically are in regions where soils are low in fertility and inorganic fertilizers are expensive or inaccessible (Dubeux et

PAGE 27

27 al., 2007). In this context, the ability of legumes to fix atmospheric N 2 into plant usable compounds (Ferguson et al., 2010) can play an important role in plant mixtures and crop rotation. Nyfeler et al. (2011) evaluated the effect of monoculture and binary mixes with varying proportions of legume (red and whit e clovers) and grass (perennial ryegrass and orchardgrass) seeding rates and N fertilization on biomass total N, N 2 fixation, N transfer from legumes to grasses, N from fertilization and N from non symbiotic sources. The authors found that mixed swards had significantly greater total N when compared with pure grass or pure legume stands, with greater benefits of the presence of legumes compared with pure grass stands that received low er N fertilization rates. Treatments receiving greater N fertilization rates (450 N kg ha 1 ) had greater amounts of N from non symbiotic sources. Yields of N from symbiotic fixation and from non symbiotic sources were positively affected by the presence of grasses, indicating a stimulus in N 2 fixation and improved N assimilation into biomass in legume grass mixed swards. N from symbiotic fixation was more efficiently transferred in mix tur es with more than 70% legume. The presence of legumes also increased t he release of N during decomposition relative to pure grass litter, resulting in lesser N immobilization (Silva et al., 2012). Sigua et al. (2014) evaluated the effect of bahiagrass and bahiagrass rhizoma peanut mixed swards under grazing and no grazing on soil extractable P. The authors did not find a difference in soil P saturation (%) and soil P concentration (mg kg of soil 1 ) with the presence of grazing in comparison to the no grazing treatment. Although the authors expected to find higher concentratio n of soil extractable P under grazing, they suggest ed that plant biomass consumption and subsequently greater plant regrowth lead to greater removal of soil P relative to the ungrazed treatment. In addition, Sigua et al. (2014) argue d that animal trampling can destroy soil

PAGE 28

28 aggregates making P more susceptible to losses by leaching. Higher levels of P concentration and saturation were found in 0 15 cm soil depth relative to 15 30 cm depth. Relative to soil under bahiagrass monoculture pastures, the soil unde r the bahiagrass rhizoma peanut mixed pastures had greater P concentration reaching 25.1 vs. 19.3 mg kg 1 P hosphorus saturation was also greater in mixed bahiagrass rhizoma peanut pastures with values of 13.7% compared with 10.6% in pure bahiagrass. How ever, across the three years of the experiment, extractable soil P presented a trend to decrease in bahiagrass rhizoma peanut mixed pastures, while the bahiagrass pastures had an increasing trend for the same variable. Bahiagrass rhizoma peanut stands pres ented greater P uptake relative to pure bahiagrass, reaching values between 19.0 and 23.8 kg ha 1 compared with 6.8 and 7.0 kg ha 1 respectively. Animal trampling can also destruct soil aggregates and make P more easily lost through soil runoff. The highe r P uptake by rhizoma peanut implies that pastures cultivated with this species are likely to require more frequent P application relative to pure grass stands in order to maintain long term production (Sigua et al., 2014). Litter D ynamics Nutrient input through litter will depend on litter deposition and decomposition rates (Rezende et al., 1999; Dubeux et al., 2006a; b). Litter deposition is usually increased by senescence and animal trampling (Pal et al., 2012; Guretzky et al., 2014). High stocking rate s can reduce deposition of litter due to high herbage mass removal by grazing (Rezende et al., 1999), while fertilization can increase litter deposition due to greater herbage mass and subsequent trampling (Liu et al., 2011a; Guretzky et al., 2014). Increa ses in temperature and rainfall have also been reported to enhance deposition of litter in Brachiaria pastures (Rezende et al., 1999). Apolinrio et al. (2013) evaluated signalgrass ( Brachiaria decumbens ) litter deposition and decomposition under differen t stocking rates (SR; 2.0, 3.9, and 5.8 AU ha 1 [1 AU = 450

PAGE 29

29 kg]) and N fertilization rates (0, 150, 300 kg N ha 1 yr 1 ) over 2 yr. The authors found that N input increased plant production causing an incre ase in existing litter biomass in the second year of the experiment. Dubeux et al. (2006a; b) evaluated litter deposition and decomposition of bahiagrass under increasing management intensities (40, 120, and 360 kg N ha 1 associated with stocking rates of 1.3 2.7, 4.0 animal units [AU, one AU = 500 kg live weights] ha 1 for Low, Medium, and High management levels, respectively). At the beginning of the experiment, Low management levels had greater existing litter compared with the other treatments likely beca use at its lower SR, less biomass was r emoved from the pasture. However, e xisting litter accumulated as the grazing season advanced for all treatment s and this accumulation started earlier for High management levels relative to other treatments due to gre ater litter deposition rates in High. Litter decomposition can be affected by litter chemical characteristics such as C:N (Rezende et al., 1999), lignin concentration (Heal et al., 1997; Silva et al., 2012), and by environmental characteristics like tempe rature (Allison et al., 2013), rainfall (Silva et al., 2012), and microbial abundance and community composition (Cuteaux et al., 1995; Allison et al., 2013). Litter in warm season grasslands tend s to be of low quality in terms of chemical characteristics causing nutrient immobilization (Vendramini et al., 2014), but presence of legumes (Silva et al., 2012; Vendramini et al., 2014) and fertilization (Dubeux et al., 2006a) can improve litter quality, decomposition, and nutrient cycling. Evaluating the influ ence of stubble heights of 8, 16, and 24 cm and N fertilization rates of 50, 150 and 250 kg N ha 1 year 1 effect of stubble height on litter disappearance. However, litter decomposition rate inc reased with increasing levels of N fertilization after 8 or 16 days of deposition The authors also

PAGE 30

30 observed an immobilization of N in all treatments, which increased at greater N application rates. Apolinrio et al. (2013) assessed the effect of managemen t intensity in terms of stocking rate and N fertilization on litter decomposition of signalgrass. The authors found that increases in N fertilization resulted in increases in N concentration in litter. Litter C:N ratio was affected by treatments. Increasin g stocking rates had a negative effect on C:N when no N was applied. The combination of N fertilizer application rate and stocking rate on Pensacola bahiagrass litter dynamics was also evaluated by Dubeux et al. (2006 a ). The y observed greater N and lignin concentration in litter from high management levels, but they also reported an increase in N immobilization and mineralized N with an increase in management levels. Greater decomposition rates were also observed as management levels increased. Dubeux et al (2006a) observed that lignin:N ratio was a strong indicator of decomposition for litter in their study. This is in a greement with Heal et al. (1997) who emphasize d that although C:N ratio is a relevant indicator of litter quality and decomposition, form of C, presence of secondary compounds and other nutrients and their interactions significantly influence litter decomposition. Silva et al. (2012) evaluated litter decomposition when 0, 50, and 100% of litter was the legume calopo ( Calopogonium mucunoide s Desv.) and the remainder was signalgrass. The authors found greater decomposition rates with the addition of legumes in one of two years of evaluation. In the second year, greater lignin concentration with increasing proportion of legume was thought resp onsible for the lack of legume proportion effect on decomposition rate. Litter C:N ratio decreased with time of decomposition in the pure grass stand, while in the presence of legumes this ratio remained nearly constant during the incubation period. Rezend e et al. (1999) evaluated the effect of stocking rates on litter decomposition of pure grass and two grass legume mixes. They found that lower stocking rates resulted in a greater proportion of legume in the

PAGE 31

31 pasture, which in turn resulted in lower C:N in litter, which had a positive effect on litter decomposition. Although C:N has often been used as an indicator of litter quality, the nature, i.e., solubility of each one of these components ha s been suggested to be of greater importance in determining lit ter decomposition rate and extent (Heal et al., 1997). For example, lignin:N was found to be a good indicator of decomposition in bahiagrass litter (Dubeux et al., 2006a). As decomposition advances, C and N become more insoluble. Lignin concentration incre ases with time of decomposition because more labile C components are used by microorganisms, leaving recalcitrant materials that increase in concentration (Berg and Matzner, 1997; Berg and McClaugherty, 2008). In addition, new lignin like material is produced by decomposition microorganisms (Berg and Matzner, 1997; Berg and McClaugherty, 2008). Nitrogen concentration in plant litter also increases with decomposition (Berg and McClaugherty, 2008), and as N has been documented to inhibit lignin decomposi tion by reducing the activity of lignin degrading phenol oxidase (Carreiro et al., 2000) and inhibiting ligninase production (Berg and McClaugherty, 2008) it can further increase lignin concentration during decomposition (Berg and McClaugherty, 2008). Phe nolic groups and quinones produced during decomposition may also immobilize NH 3 increasing recalcitrant N in plant litter (Berg and McClaugherty, 2008). Dubeux et al. (2006a) reported ~50% of N immobilized in bahiagrass litter after 128 d of incubation, a nd Silva et al. ( 2012) found 50 to 68% in signalgrass calopo litter after 256 d of incubation. Rovira and Vallejo (2002) investigated change in C and N recalcitrance of plants varying in C and N quality at beginning of incubation over 2 yr. They found that C recalcitrance increased to a greater degree in plant materials with initial larger proportion of labile C, while N recalcitrance increased in all plant materials.

PAGE 32

32 In plant litter, mixtures of multiple species often resulted in positive, non additive ef fects on decomposition, such that decomposition of mixtures cannot be estimated based on decomposition of single components (Chapman et al., 2013; Zhang et al., 2013; Cuchietti et al., 2014). The increase in decomposition in mixtures relative to monocultur es was observed in a variety of plant species. In heather [ Calluna vulgaris (L.) Hull] and bracken [ Pteridium aquilium (L.) Kuhn], mixed litter decomposed to a greater extent (~48%) relative to monocultures (~37%) (Anderson and Hetherington, 1999). Chapman et al. (2013) incubated leaf litter material of four species from a mixed conifer forest (aspen [ Populus tremuloides Michx], Douglas fir [ Pseudotsuga menziesii Mirbel Franco], limber pine [ Pinus flexilis James], and ponderosa pine [ Pinus ponderos P. and C Lawson] with a wide range of chemical characteristics in one bi and all species combinations. The authors found that decomposition was greater and faster in mixtures (except those containing aspen) when compared with estimates based on single species decomposition (Chapman and Koch, 2007; Chapman et al., 2013) Although microbial activity was not related to mass losses of mixtures in their study, Chapman et al. (2013) concluded that microbial diversity was largely responsible for their greater biomass losses. Use of isotopes for determining C source I sotopic abundance and fractionation ha ve been used for many purposes. Carbon isotopes are a powerful tool for evaluating C dynamics in ecosystems (Brggemann et al., 2011), and carbon fractionation is traditionally measured using isotope ratio mass spectrometry (Werner and Brand, 2001). When evaluating C signature in soil 13 C can be used to differentiate carbon source between C3 and C4 plants (Haile et al., 2010). Isotope fractionation occurs because physical and chemical reactions are typically inclined to use one type of isotope over another, resulting in differences in reaction properties and composition of molecules (Brggemann et al., 2011). In general, the lighter isotopes are

PAGE 33

33 fav ored in biochemical reactions due to their higher kinetic energy (Glaser, 2005). The 12 C isotope represents 98.9 % of total C in the environment, followed by 13 C with 1.1% and 14 C with 10 8 % (Glaser, 2005). Isotopic fractionation in photosynthesis is affected by enzymatic fixation and CO 2 diffusion (Farquhar et al., 1982; Seibt et al., 2008), where the C fixation mechanism plays a major role in determining C fractionation in higher photosynthetic organisms (Farquhar, 1983). When stomatal conductance de creases, internal CO 2 pressure decreases and plant tissue becomes 13 C enriched (Farquhar et al., 1982; Glaser, 2005). In a single environment, C3 plant species that show large 13 C present greater water use efficiency (WUE; Farquhar et al., 1982; Hubicka et al., 1986), with greater differences found in older leaves (Rajabi et al., 2009). Animal E xcreta In beef production, around 85 to 90% of N and P consumed by animals is returned to the soil through animal urine and dung ( Whitehead, 2000). However, excreta distribution is not uniform in the pasture. Nutrients tend to accumulate under shade and near water sources at high temperature and humidity levels, where animals congregate to reduce stress caused by high temperatures (D ubeux et al., 2009). In environments where pastures comprise flat areas and hills, animal excreta tends to accumulate in the flat areas indicating that some of the nutrients deposited there have been consumed from steep areas (Schnyder et al., 2010) In grazing systems, deposition of N in urine patches ranges from 20 to 80 g m 2 while N deposition in dung patches is 50 to 200 g m 2 (Oenema et al., 1997). Grazing animals can affect nutrient cycling. Dubeux et al. (2009) evaluated management levels of low (1.4 AU ha 1 and 40 kg N ha 1 year 1 ), medium (2.8 AU ha 1 and 120 kg N ha 1 year 1 ) and high (4.2 AU ha 1 and 360 kg N ha 1 year 1 ) on bahiagrass. The study showed greater soil N, K and Mg concentration (mg

PAGE 34

34 kg 1 soil) in high relative to medium and low m anagement treatments. Greater soil P concentration was also observed in the moderate compared with the other management level treatment s Greater soil P in moderate relative to low treatments probably occurred because of greater fertilization rate in that treatment Greater herbage IVDOM, N and P concentrations at close and medium vs. large distances from water and shade were also observed at the low management level, probably a result of greater animal excreta deposition in those areas. White Leech et al. (2013a) evaluated the effect of animal excreta type and deposition frequency on bahiagrass maintained under average (60 kg N ha 1 ) or high (120 kg N ha 1 ) management levels. The study found a linear increase in dry matter harvested with increasing urin e frequency application, while increase in dung frequency application had a negative linear effect on the same variable due to sustained coverage of the grass Greater urine application frequency also resulted in greater IVDOM in the high management treatm ent. Urine application frequency also enhanced CP concentration and harvested N in bahiagrass. As observed by White Leech et al. (2013b), dry matter harvested and nutritive value (CP and IVDOM) tend to increase with higher urine frequency application in an area of 15 and 30 cm surrounding urine patches, respectively, independent of management levels. GHG Emissions from Agricultural P roduction Climate change, defined as the statistical difference in variability or mean of climate properties that continues f or more than a decade, may be triggered by natural or anthropogenic forces (IPCC, 2014). Several factors point to the intensification of the greenhouse effect, among which are the increase in air and ocean temperatures, sea level, and snow and ice melting (IPCC, 2014). Many agriculture production and management practices are sources of GHG, which are known for their effect on climate because of their capacity to trap energy in the atmosphere (IPCC, 2014). Some sources of GHG in agricultural production inclu de: methane (CH 4 ) from

PAGE 35

35 livestock enteric fermentation (Murray et al., 1976; Moss et al., 2000; Lassey, 2007; Henry et al., 2015), manure (Boadi et al., 2004; Lassey, 2007), dung (Jarvis et al., 1995; Yamulki et al., 1999), and rice production (Lassey, 2007 ); nitrous oxide (N 2 O) from urine and dung (Boadi et al., 2004; van Groenigen et al., 2005a; b; van der Weerden et al., 2011; Mazzetto et al., 2014), animal waste management (Harper et al., 2004), litter decomposition, especially if containing legumes (Pal et al., 2012, 2013), and use of N fertilizer (Bouwman, 1996; de Klein et al., 2006); and carbon dioxide (CO 2 ) from fossil fuel use, fertilizer production, and soil management (McSwiney et al., 2010). Global warming potential (GWP) refers to the radiative forcing of the emission of 1 kg of a gas relative to 1 kg of CO 2 integrated through time (Shine et al., 1990). When estimating GHG emissions from agricultural practices, GHG that not CO 2 are converted to CO 2 equivalent (CO 2 al., 2013). GHG Emissions in Livestock P roduction Previous work suggesting ruminants have low feed conversion ratios relative to monogastric animals often ignore that a large proportion of feed used in monogastric production could also be used for human co nsumption, while ~86% of feed intake by livestock does not overlap with human diet (Mottet et al., 2016). The ability of ruminants to grow and produce high quality human food when fed a fibrous feed that is unsuitable for human consumption should not be ig nored in the context of climate change and its likely detrimental consequences to agricultural practices and production (Beauchemin et al., 2010). Mottet et al. (2016) indicate that ruminants are more efficient in terms of animal protein production compare d with monogastric animals under industrial systems when feed conversion is expressed as meat produced per kg of human edible feed. Animals fed high concentrate diets may produce less methane when compared with grass fed animals (Kurihara et al., 1999; Bea uchemin and Mcginn, 2005),

PAGE 36

36 however, comparisons of pasture vs. concentrate feeding systems should also account for the emissions associated with production of the feed (Yan et al., 2010). Although the relevance of GHG sources in cattle production are depe ndent on management practices, several studies estimating emissions from dairy and beef cattle point to enteric fermentation and animal waste as the most important sources of GHG emissions. In a dairy system in Ireland, enteric CH 4 production accounts for 49% and N 2 O and CH 4 from manure management for 11% of total GHG emissions (Casey and Holden, 2005). In New Zealand, 62% of total GHG emissions from milk production come from CH 4 from enteric fermentation and 16% from direct N 2 O emissions from urine and du ng deposited on pasture; this from a system characterized by grass clover grazing throughout the year with occasional use of silage (Flysj et al., 2011). In Sweden, milk production based on roughage fodder, concentrate, and grains has 46% of emissions com ing from CH 4 from enteric fermentation and 8% from direct N 2 O from manure used in fields after storage (Flysj et al., 2011). Canadian beef cattle production presented a similar partitioning of emissions relative to dairy systems, with CH 4 from enteric fe rmentation representing 53 to 54% of emissions and manure emissions of CH 4 and N 2 O accounting for 24 to 28% of emissions (Basarab et al., 2012). In the US, beef production is characterized by grass and hay fed cow calf operations with animals finished on feedlots and consuming high concentrate diets (Pelletier et al., 2010). Analyzing different production systems in the Upper Midwest, Pelletier et al. (2010) found that enteric CH 4 represented 43% and manure CH 4 and N 2 O 21% of GHG emissions in cow calf oper ations. During the finishing phase, CH 4 from enteric fermentation represented 32 to 42% and CH 4 and N 2 O from manure 21 to 30% of GHG emissions (Pelletier et al., 2010). A study analyzing GHG emissions from animals finished on pasture in Ireland described 4 6 to 53% of

PAGE 37

37 emissions coming from enteric CH 4 and 13 to 15% from CH 4 and N 2 O from excreta deposited on pasture (Foley et al., 2011). Other processes responsible for GHG emissions in cattle production include pre farm operations such as chemical fertilizer lime, and agrochemicals production, transport, transfer and storage (Lal, 2004). Other on farm sources of GHG include applied N fertilizer (Casey and Holden, 2005; Foley et al., 2011) and lime (Foley et al., 2011), fossil fuels (Casey and Holden, 2005, 2 006; Foley et al., 2011) and energy used in farm operations and feed production (Casey and Holden, 2006; Basarab et al., 2012). Yearly GHG emissions from N fertilizer production and use are estimated to reach 300 Tg CO 2 (Jensen et al., 2012). Influence of P lants on GHG E missions Some studies have focused on the effect of plants on N 2 O emissions. Comparing soil N 2 O emissions at different temperatures from ryegrass or bare soil receiving urine application, Uchida et al. (2011) found that greater emissions of N 2 O from urine of dairy cattle occurred when it was applied where plants were present vs. absen t probably because of the effect of available plant sourced C on microbial population s that produces N 2 O Klumpp et al. (2010) evaluated N 2 O production in areas with low (19%) and high (35 %) densities of white clover and found no effect of clover density on emissions. Virkajrvi et al. (2010) found greater N 2 O emissions from grass clover mixes in Finland when compared with pure grass stands receiving N fertilizat ion (220 kg N ha 1 ). The authors attributed these differences to the fact that N from clover has been found to be less susceptible to losses through leaching relative to losses through gaseous forms In addition, they f ound greater emissions coming from an imal dung relative to urine, with no difference being detected if applications were made in either grass clover or pure grass pastures. Evaluating N 2 O emissions from lentil ( Lens esculenta Moench), pea ( Pisum sativum L.) and spring wheat ( Triticum aestivum L. ac. Barrie), Zhong et al. (2009) found no differences in

PAGE 38

38 emissions among treatments. In addition, some plants produce biological nitrification inhibition (BNI) molecules that are released through plant roots and decrease nitrification (Subbarao et al., 2013). The addition of BNI producing plants in production systems has the potential to decrease losses of N by N 2 O, NO, and NO 3 (Subbarao et al., 2013). Cattle grazing and trampling on pastures can also affect litter deposition, which can in turn be a r elevant source of N 2 O (Pal et al., 2012). Evaluating emissions of N 2 O from perennial ryegrass litter decomposition across 49 days, Pal et al. (2013) found higher emissions from root litter buried 2 cm deep in sieved soil, followed by emissions from root li tter on sieved soil surface, shoot litter on pasture soil, and shoot litter deposited on sieved soil. Pal et al. (2012) compared N 2 O emissions from white clover and annual ryegrass ( L olium multiflorum L.) litter on soils with 54 or 86% WFPS for 14 days. Th e authors found greater emissions coming from clover litter on lower WFPS soils, while at greater WFPS no significant differences in N 2 O emissions were found. Emissions from Enteric M ethane Ruminants are the largest contributors to anthropogenic CH 4 emissi ons in the world (Steinfeld and Wassenaar, 2007) In ruminants, enteric CH 4 is produced through anaerobic fermentation in which ruminal microorganisms hydrolyze compounds previously broken down through mastication, which results in free hydrogens. Methanog enic reactions use these hydrogens to reduce CO 2 resulting in CH 4 (rskov et al., 1968; Leek, 2004) Methane is mouth and nose Mitigation of enteric CH 4 emissions can be achieved through manipulation of diet, intake, or ruminal microflora (Johnson and Johnson, 1995) and by increasing animal production efficiency (Wall et al., 2009; Hristov et al., 2013) Several feed additives have been found to

PAGE 39

39 decrease enteric CH 4 emissions. In goats, Bhatta et al. ( 2008) found that pumpkin ( Cucurbita pepo L.) ( Ipom o ea batatas L. Lam) vine silage reduced methane emissions Tannin has also been studied and found to reduce methane emiss ions in goats (Puchala et al., 2005) and sheep (Carulla et al., 2005) Tannins had no effect on cattle where it presented a protein binding effect (Beauchemin et al., 2007) Evaluating seven legume plants for CH 4 production potential, Naumann et al. ( 2013) found a significant decrease in CH 4 production with increased condensed tannin concentration but for most legumes a concomitant reduction in volatile fatty acids (VFA) production occurred indicating a decrease in plant digestibility Archim de et al. (20 11) reviewed enteric methane emissions from livestock grazing cool and warm season legumes and C3 and C4 grasses. The authors found that when grazing warm season legumes, animals emitted 20% less CH 4 (L kg OM 1 ) compared with C4 grasses. Emissions from Animal Urine and D ung Animal urine is a significant source of N 2 O emissions (Yamulki et al., 1998; van der Weerden et al., 2011; Dijkstra et al., 2013) but not of CH 4 (Yamulki et al., 1999), while dung can produce both CH 4 and N 2 O (Yamulki et al., 1998; M azzetto et al., 2014; Mori and Hojito, 2015). Emissions of methane from dung peaks within the first 1 to 3 d after deposition due to ideal conditions for methane formation in terms of temperature, moisture (Yamulki et al., 1999), and available microbiota ( Jarvis et al., 1995). Methane production is an exclusively anaerobic process (Yamulki etal., 1999), and as most microbial mediated reaction s it is stimulated by higher temperatures (Jarvis et al., 1995; Mazzetto et al., 2014). However, high temperatures ca n create a crust on the dung pile that reduces CH 4 diffusion to the atmosphere (Yamulki et al., 1999). Rainfall can diminish available C for CH 4 formation, reducing emissions, or increase anoxic condition increasing CH 4 production (Jarvis et al., 1995). An imal diet may also influence CH 4 emissions from animal dung. Jarvis et al. (1995) found that emissions from animals grazing

PAGE 40

40 grass clover pastures had greater CH 4 emissions relative to fertilized grass, likely due to low C:N ratio. Deposition of urine and dung patches on the soil provides large amounts of N and readily available C (van Groenigen et al., 2005a). Production of N 2 O occur s through nitrification, denitrification, and nitrifier denitrification (Dijkstra et al., 2013). The organic N in dung and ur ine first goes through mineralization by heterotrophic organisms and is transformed into NH 4 + which is then used by nitrifying microorganisms and transformed into nitrite (NO 2 ) and nitrate (NO 3 ) that are used by denitrifying microorganisms to produce N 2 O and molecular nitrogen (N 2 ) (Oenema et al., 1997; van der Weerden et al., 2011). Alternatively, nitrifier denitrification occurs when NO 2 is reduced to nitric oxide (NO), N 2 O, and N 2 (Wrage et al., 2001). Part of the NH 4 + from urine is lost to the atmos phere as NH 3 (van Groenigen et al., 2005a). Studies show that more than 70% of total NH 3 volatilized to the atmosphere is lost in the first 4 days after urine is applied to the soil (Lockyer and Whitehead, 1990). Denitrification tends to be the predominant N 2 O emission process after urine is deposited on soil surface because of high soil water content, after which nitrification is the main process releasing N 2 O (de Klein and van Logtestijn, 1994). Pulses of N 2 O emissions occur immediately after urine deposi tion on soil surface, but this behavior is not observed after dung deposition (Virkajrvi et al., 2010). Dung N content can be soluble (20 to 25%), in indigestible forms (15 to 25%), or as living organisms (50 to 65%) (Oenema et al., 1997). Emissions of N 2 O in grasslands are affected by many factors such as soil compaction (Oenema et al., 1997; van Groenigen et al., 2005a), soil moisture content (van Groenigen et al., 2005a; van der Weerden et al., 2011; Giltrap et al., 2014), intensity of grazing (Rafique et al., 2011), application of N fertilizer (Klumpp et al., 2010), and presence of urine patches and dung

PAGE 41

41 (Giltrap et al., 2014). Anaerobic conditions can increase N 2 O emissions (Cardenas et al., 2007). Rainfall events tend to reduce WFPS resulting in anae robic conditions that favor N 2 O emissions (Klumpp et al., 2010; Rafique et al., 2011; van der Weerden et al., 2011), while dry soils tend to restrain emissions (Klumpp et al., 2010). Typically, greater emissions occur when WFPS is in the range of 60 and 80% when nitrification, denitrification and nitrifier denitrification are stimulated (van Groenigen et al., 2005a; b; Dijkstra et al., 2013). Rainfall events have also been associated with greater emissions of N 2 O from animal urine and dung. For example, Lessa et al. (2014) found significant emissions of N 2 O in the rainy relative to the dry season in Brazil. Temperature can also affect N 2 O production, with greater emissions at higher temperatures (Klumpp et al., 2010). Rafique et al. (2011) observed that N 2 O emissions from dairy farms in Ireland at 17C achieved values five times greater than at 5C. Uchida et al. (2011) evaluated emissions of N 2 O from bovine urine at 11, 19, and 23 C. The authors found that emissions expressed as pe rcentage of N input loss as N 2 O N increased 10 times from lower to greater temperatures (0.2 to 2.2%). Concentration of N in urine and dung is positively related to N 2 O emissions (Oenema et al., 1997), and some authors suggest that reduction of N output b y animals is an alternative for reducing N 2 O emissions from animal excreta (Dijkstra et al., 2011, 2013). Annual N deposition is one of the parameters used for estimating N 2 O emissions in the IPCC model. Concentration of N in urine is influenced by water i ntake and dietary N, varying between 1 and 20 g L 1 (Oenema et al., 1997). In an experiment using artificial urine with increasing N concentrations (119, 237, 474, and 949 N mg kg 1 dry soil), no effect on total N 2 O emissions was found. However, authors fo und an alteration in the pattern with which N 2 O is emitted with higher concentration of N in urine resulting in lower and longer peaks of N 2 O emission (van Groenigen et al., 2005a).

PAGE 42

42 Similarly, N 2 O emissions from manure (mix of dung and urine) from dairy co ws fed diets with 167 or 148 mg g 1 CP did not differ despite the significant differences found in N concentration in manure (Lee et al., 2012). The model used by IPCC (2006) to estimate N 2 O emissions from soil assumes a linear correlation between N conce ntration in animal waste or chemical or organic fertilizer, and an emission factor for N 2 O emitted per unit of N applied. The simplicity of the model used compared with the actual complexity of processes involved in N 2 O formation in the soil leads to large uncertainties in emission estimations (Flysj et al., 2011). Additionally, the default N 2 O emission factors recommended by IPCC (2006) were obtained from intensively managed temperate pastures (Oenema et al., 2005). When applied to modelling, these factor s result in uncertainty from bias, i.e., the process driving N 2 O emissions is not sufficiently understood or represented in tropical extensive grazing (Oenema et al., 2005), which can lead to systematic errors in N 2 O emission estimations (Charles et al., 1 998; Winiwarter and Rypdal, 2001; Rypdal and Winiwarter, 2001). It has also been suggested that a better understanding of factors driving N 2 O production associated with soil mapping and identification of urine patches could improve N 2 O emission estimation with models (Giltrap et al., 2014). The protocol used by IPCC also uses an emission factor for percentage of deposited N emitted as N 2 O, with one single emission factor used for both urine and dung. The IPCC (2006) encourages the production of emission factors specific for countries, especially developing and tropical countries. However, significant differences between urine and dung N 2 O emissions have been found (van Groenigen et al., 2005b; Virkajrvi et al., 2010; van der Weerden et al., 2011; Sordi e t al., 2013; Mazzetto et al., 2014). van der Weerden et al. (2011) state that obtaining emission factors separately for urine and dung would result in more accurate N 2 O emission

PAGE 43

43 estimation from urine and dung. In addition, year long studies are indicated t o provide accurate estimate s of emissions (Kebreab et al., 2006).

PAGE 44

44 CHAPTER 3 LEGUME PROPORTION IN GRASSLAND LITTER AFFECTS DECOMPOSITION AND NUTRIENT MINERALIZATION Introduction Grassland litter decomposition is an important source of nutrients in grazing systems (Dubeux et al., 2007). Increasing N fertilization increased litter decomposition rate and extent (Dubeux et al., 2006a; Liu et al., 2011), in part because of lesser litter C:N ratio which favored decomposition and N mineralization (Heal et al., 1997; Dubeux et al., 2006a; Manzoni et al., 2015). Many C4 grass pastures in warm climates receive little or no N fertilizer, resulting in greater C:N ratio and nutrient immobilization (Dubeux et al., 2006a). Alternatives to N fertilizer for increasi ng litter decomposition and N mineralization are needed. Previous research has pointed to the importance of the forms of C and N in litter not only their concentration, in determining litter decomposition rate and extent (Heal et al., 1997; Silva et al., 2012). Forms of C and N vary in different species, and increasing species richness of litter has enhanced decomposition (Anderson and Hetherington, 1999; Chapman and Koch, 2007; Chapman et al., 2013). For mixed species litter that differed in C and N quali ty and concentration (fertilized and unfertilized Scots pine [ Pinus sylvestris L.] and corn [ Zea mays L.]), Berlung et al. (2013) found that during incubation C and N were transferred between litter types through amino acids in colonizing fungal mycelia. A lthough transfers were bidirectional, net N transfer occurred toward N depleted litters, and net C transfer occurred from pine toward corn litter. These microorganism mediated interactions support the use of mixed species pastures as a means to increase li tter decomposition and nutrient mineralization. The greater N concentration of legume litter makes their inclusion an option to increase nutrient cycling in pastures based on C4 grasses. Inclusion of legumes in grasslands is an economically attractive alt ernative to increase pasture sustainability (Muir et al., 2011), while

PAGE 45

45 also increasing nutritive value and animal production potential (Muir et al., 2014). However, little information is available to evaluate effect of legume proportion in mixtures with gr asses on plant litter disappearance and nutrient dynamics, particularly in warm climates and when compared with N fertilized C4 grass pastures. The objectives of this study were to quantify decomposition rate and nutrient disappearance from litter of N fer tilized and non fertilized bahiagrass ( Paspalum notatum Fl gge ) swards compared with responses of litter varying in proportion of rhizoma peanut ( Arachis glabrata Benth.) and bahiagrass, and to identify differences in proportional disappearance of each com ponent of the rhizoma peanut bahiagrass litter mixtures during the incubation period. Material and Methods Experimen tal S ite Plant material for litter incubation was collected from pastures at the University of Florida Beef Research Unit (BRU) in Gainesv ille, FL (29.72 N, 82.35W) and the Plant Science Research and Education Unit (PSREU) at Citra, FL (29.24N, 82.10W). Soils at BRU are Sparr fine sand (loamy, siliceous, hyperthermic Grossarenic Paleudult) and Adamsville sand (hyperthermic, uncoated Aqui c Quartzipsamments). Soils at the PSREU are a Placid fine sand (sandy, siliceous, hyperthermic, Typic Humaquepts) and Tavares sand (hyperthermic, uncoated Typic Quartzipsamments). Litter incubation was on a well pasture a t BRU with the soil being an Adamsville sand (hyperthermic, uncoated Aquic Quartzipsamments). Weather data are reported from a station (Florida Automated Weather Network, FAWN) located 18 km from the incubation site (Figure 3 1). Treatments and Experiment al D esign The litter decomposition experiment was conducted in each of 2 yr (2014 and 2015), with five types of above ground litter as treatments arranged in six replicates of a randomized

PAGE 46

46 complete block design. Treatments were 100% unfertilized bahiagras s, 100% bahiagrass fertilized with 60 kg N ha 1 67% unfertilized bahiagrass plus 33% Florigraze rhizoma peanut, 33% unfertilized bahiagrass plus 67% Florigraze rhizoma peanut, and 100% Florigraze rhizoma peanut. Treatments are hereafter referred to as BG, BGN, RP33, RP67, and RP, respectively. One replicate of each treatment consisted of six bags, one bag each for incubation times of 4, 8, 16, 32, 64, and 128 d, resulting in 36 bags per treatment (six bags times six replicates). Six empty bags were incubat ed per replicate, one per incubation period, for a total of 36 blank bags across the six replicates. Blank bags were used to account for any loss of bag mass during the incubation period. Litter Collection, Samples and Site P reparation Bahiagrass herbage for separation of litter was obtained from monoculture pastures at the BRU. Both pastures were full season growth from spring until litter collection on 1 June in 2014 and 2015. The N fertilized grass received 60 kg N ha 1 14 d before herbage was harvested for litter incubation while the pasture used for BG litter received no N that growing season. For the RP33 and RP67 treatments, where bahiagrass litter was mixed with rhizoma peanut litter, bahiagrass herbage was collect ed from a mixed pasture of bahiagrass and rhizoma peanut. In all cases, bahiagrass herbage was clipped to soil level and the leaf blades from throughout the canopy were separated from the remaining herbage at the ligule. The litter fraction of bahiagrass f or this study was considered to be leaf blade, because it constituted a high proportion of forage mass above soil level. The rhizoma peanut herbage from which litter was separated was collected from monoculture pastures at the PSREU. The top 5 cm of an app roximately 20 cm tall rhizoma peanut sward canopy was removed by clipping and discarded because it included the least mature leaves. The remaining material was cut to soil level and leaflets were separated and considered to be rhizoma peanut litter. All of these older leaflets were considered to be litter

PAGE 47

47 because leaflets fall while still green, meaning that leaflet color is not a useful indicator of those that will soon fall from the plant (Fialho, 2015). In addition to all leaflets in the top 5 cm of the sward canopy, any leaflets that were not fully expanded (still folded, Fialho, 2015) were discarded and not used in the incubation. For both bahiagrass and rhizoma peanut litter, the material was dried at 60C for 72 h. Litter material was not ground prio r to incubation in order to more closely simulate conditions of actual field litter and to maintain surface integrity (Dubeux et al., 2006a). Bags used for the incubation were 15 x 20 cm, made with 100% polyester lining fabric (75 m mesh size) and closed with a heat sealer. Each bag was filled with a total of 12 g of litter, with each litter type in mixtures weighed separately to ensure the proper proportion of each species. Fifty grams of each litter material was reserved to determine litter chemical comp osition at the start of the incubation period. At initiation of incubation, bahiagrass growing in the area where bags were to be incubated was clipped to a 10 cm stubble height and bags were placed on the soil surface and subsequently covered with clipped bahiagrass to simulate litter cover. Incubation sites were protected from disturbance with wire mesh exclusion cages, and grass regrowth was trimmed biweekly to a 10 cm stubble using a hand clipper and removed from the caged area. Incubation periods were i nitiated on 17 July 2014 and 21 July 2015. Litter D ecomp osition and Nutrient D isappearance Bags were collected at each incubation time for each treatment x replicate combination, dried for 72 h at 60C, and residual soil and plant material attached to bag s were removed with a brush before weighing. Weights obtained were adjusted for any changes in bag mass based on the weight of blank (unfilled) bags. To reduce cost and because chemical composition of decomposed litter is much less variable than residual l itter mass, only litter samples from Blocks

PAGE 48

48 1 through 3 were analyzed for chemical composition. The samples were ground in a Wiley mill (Model 4 Thomas Wiley Laboratory Mill; Thomas Scientific) to pass through a 1 mm stainless steel screen and analyzed for acid detergent insoluble N (ADIN). Dry matter was determined by drying samples at 105C for 15 h and OM by ashing samples at 500C for 4 h (Moore and Mott, 1974). Phosphorus was analyzed us ing a modification of the aluminum block digestion procedure (Gallaher et al., 1975) and read by semi automated colorimetry (Hambleton, 1977). Carbon and N were analyzed through dry combustion (Vario Micro Cube; Elementar, Hanau, Germany) after samples wer e ball milled. Carbon:N and C:P ratios were obtained by dividing C concentration (g kg 1 OM) by N (g kg 1 OM) and P concentration (g kg 1 12 C and 13 C was analyzed with an isotope ratio mass spectrometer (I soPrime 100, Iso Prime, Manchester, UK) coupled with the dry combustion analyzer in order to determine relative disappearance of legume and grass in mixtures (treatments RP33 and RP67). Carbon isotope ratio was determined 1 3 1 2 C in samples to the standard Pee Dee Belemnite (PDB). Lignin was analyzed with the Daisy incubator (ANKOM Technology, 2017a). For ADIN, samples were first analyzed for acid detergent fiber (ADF) in an ANKOM fiber analyzer (ANKOM Technology, 2017b), and the resulting re sidue was analyzed for N concentration utilizing Micro Kjeldahl digestion (Gallaher et al., 1975). Lignin:N and lignin:ADIN ratios were obtained by dividing lignin concentration (g kg 1 OM) by N and ADIN concentrations (g kg 1 OM), respectively. Acid deter gent insoluble N concentration in total N was calculated by dividing ADIN concentration (g kg 1 OM) by N concentration (g kg 1 OM).

PAGE 49

49 Time dependen t responses of measure d variables were evaluated by fitting mathematical models. An inverse logistic model (Se baugh and McCray, 2003) described in Eq uation 3 1 was used to describe total and species specific biomass, N, and P decay with time ( P < 0.05) f(x) = {(A D) / [1 + (x / C) k ]} + D (3 1) in which f(x) is remaining biomass, N, or P; A is the lower limit reached at decay; k is the slope of decay during the linear phase of decay (relative decay rate, g g 1 d 1 ); D is the upper limit during decomposition; C is the day when half of observed dec omposition occurs relative to the upper and lower limits (A and D parameters), hereafter called relative half life; and x is time in days. Lignin and ADIN concentrations and lignin:N and ADIN concentration in total N followed a linear plateau model ( P < 0. 05) described in Equation 3 2 (McCartor and Rouquette, 1977; Silva et al., 2012). f(x) = A + B (x C) (3 2) in which f(x) is the concentration of lignin, ADIN, lignin:N, or ADIN in total N; A is the initial concentration or ratio; B is the rate of incr ease in concentration or ratio from beginning of incubation until plateau is reached (g kg 1 OM d 1 ); and C is the day in which concentration or ratio r each the plateau. Carbon:N, C:P and lignin:ADIN followed a single exponential model ( P < 0.05) describe d in Equation 3 3 (Wider and Lang, 1982). f(x) = A exp ( k x) (3 3) in which A is the disappearance coefficient, k is the relative decay rate (g g 1 d 1 ), and x is time in days. Statistical Analysi s Data were analyzed using the R software (R Core Team, 2016). For response variables analyzed with the inverse logistics and single exponential models, relative decay rate k was determined and analyzed. Relative half life C was also quantified and analyzed for variables fit

PAGE 50

50 to the inverse logistic equation. For response variables analyzed with the linear plateau model, days to reach plateau C and rate of increase between start of incubation and linear plateau B were analyzed. For all response variables, data pertaining to the end of incubation period (Day 128) al., 2016), where treatment was considered fixed and year was considered random. Least squares means were ana main effects ( P < 0.05). The effect of legume litter proportion (i.e., 0, 33, 67, and 100 for BG, RP33, RP67, and RP, respectively) was assessed by polynomial contrasts using mass disappearance in mixtures (RP33 and RP67 treatments) were made relative to that of the single species (BG and RP) treatments. Results and Discuss ion Initial Litter C omposition Sources of litter were not replicated, so initial litter composition data were not compared statistically. They are presented for descriptive purposes (Table 3 1). Litter treatments including legume had characteristically hig h concentrations of N, P, lignin, ADIN, and ADIN as a proportion of total N. Likewise, initial litter C:N and C:P were low for treatments including legume. Total Biomass Decomposition Rate, R elative Half Life, and Extent of D ecomposition There was no effe ct of treatment on total biomass decomposition rate during the linear phase of decomposition ( k ; P life ( P < 0.001) (Figure 3 2 .A ) was affected by litter composition. Relative half life is the day at which half of observed total litter decomposition occurred. It increased linearly from 21 to 33 d as proportion of rhizoma peanut in litter decreased, indicating that mass losses were g reater early in the incubation period as

PAGE 51

51 proportion of legume increased. Addition of legume had a quadratic effect ( P = 0.040) on remaining litter biomass at end of incubation. After 128 d, BG had greater ( P = 0.028) remaining biomass compared with RP33 (4 3 and 35%, respectively), but BG decomposition extent did not differ from the other treatments ( P > 0.145; 37, 39, and 39% for BGN, RP67, and RP, respectively). For bahiagrass monocultures, litter decomposition rate was greatest for pastures receiving mor e N fertilizer and grazed at the greatest stocking rate (Dubeux et al., 2006a). The authors reported a mass loss of 15% in the first 8 d of incubation, with decay rate being significantly less after 14 d of incubation. Similarly, Tifton 85 bermudagrass lit ter incubated for 128 d had a relative half life of 32 d, with greater mass decomposition rate at the beginning of incubation (Liu et al., 2011). Remaining litter biomass in the current study (~ 40% across treatments) was the same as the average across th ree bahiagrass management intensity treatments after 128 d of incubation (Dubeux et al., 2006a) and within the range for signalgrass [ Brachiaria decumbens (Stapf) R. D. Webster] calopo ( Calopogonium mucunoides Desv.) mixtures (30 55%) after 256 d of incuba tion ( Silva et al., 2012). When Tifton 85 bermudagrass pastures were fertilized with 50, 150, and 250 kg N ha 1 remaining aboveground litter biomass after 128 d was greater than in the current study, decreasing from 78 to 63% as N level increased in the f irst year and from 52 to 45% in the second year. The impact of legume on total litter biomass remaining at the end of incubation is not consistent in the literature. Including the legume calopo with signalgrass resulted in less litter biomass remaining in 1 yr but had no effect in the second year (Silva et al., 2012) They suggested that legume chemical composition played a role. Legume N concentration was less

PAGE 52

52 and lignin concentration greater in the year legume had no effect relative to the year when it a ffected the response. Rezende et al. (1999) evaluated the effect of stocking rates on litter decomposition of a koroniviagrass [ Urochloa humidicola (Rendle) Morrone & Zuloaga] monoculture and koroniviagrass legume mixtures with ovalifolium [ Desmodium heter ocarpon (L.) DC subsp. ovalifolium (Prain) H. Ohashi] or tropical kudzu [ Pueraria phaseoloides (Roxb.) Benth.]. They found that lower stocking rates resulted in a greater proportion of legume in the pasture, which in turn resulted in lesser litter C:N rati o and greater litter decomposition. In the current study, RP33 had less litter biomass remaining than BG, but RP67 and RP did not differ from BG. In the absence of differences in k the RP33 treatment had greater total loss of biomass in part because it h ad a greater period of linear decay (period between initial and final plateau) compared with other treatments (69 vs. ~51 d, respectively). Chemical composition of the legume may contribute to the response observed. Berg and McClaugherty (2008) indicated t hat both lignin and N concentration increase in residual litter as decomposition advances. They also stated that reduced decomposition at later incubation stages is due to an increase in lignin concentration that occurs partially because increasing N conce ntration inhibits lignin decomposition by inhibiting ligninase formation (Berg and McClaugherty, 2008) Initial litter N concentration increased from 21 to 31 g kg 1 while initial litter lignin concentration increased from 43 to 63 g kg 1 as rhizoma peanut percentage increased from 33 to 100%. Based on Berg and RP was shortened in part because of elevated initial N and lignin concentrations and interactions betwee n these components during decomposition. This finding of superior decomposition of RP33 vs. BG may have implications on how much legume is needed in pastures to enhance nutrient cycling.

PAGE 53

53 Disappearance of I nd ividual Species in M ixtures Treatment affected d isappearance of the rhizoma peanut component of litter during incubation ( P < 0.001) and its relative half life ( P = 0.048) (Figure 3). Rhizoma peanut litter mass remaining at the end of incubation as a percentage of that present at the start increased linearly as rhizoma peanut percentage increased from 33 to 100% ( P = 0.007), with values of 27, 35, and 39% for RP33, RP67, and RP, respectively. Relative half life was achieved later in RP33 compared with other rhizoma peanut treatments (27 d compared with 20 d in RP67 and RP), probably because losses of rhizoma peanut biomass in RP33 were more extensive and occurred fo r a longer time compared with that of the other treatments containing legume. There were no differences among treatments in proportion of bahiagrass litter mass remaining at the end of incubation as a percentage of that present at the start. In the mixed legume grass treatments, the period of linear decline in rhizoma peanut biomass started earlier (7 and 9 d for RP33 and RP67, respectively) relative to that of the bahiagrass component of the mixture (19 and 16 d for RP33 and RP67, respectively). Thus, during the incubation period the proportion of bahiagrass in the mixed litter was increasing. At the end of incubation, the RP33 treatment contained only 25% rhizoma peanut. The legume grass litter proportion in RP67 changed relatively little over the incubation period, from 67% rhizoma peanut at the beginning to 63% at the end of incub ation. Decomposition of rhizoma peanut litter started at Day 10 of incubation in monoculture (compared with Day 7 9 in mixtures) and bahiagrass started at Day 14 in monoculture (compared with Day 16 19 in mixture). Thus, it seems that including the grass s lightly shortened the lag phase of legume decomposition, while the legume slightly increased the lag phase of the grass. For both components, the linear decomposition phase was longer in RP33 (55 and 51 d for rhizoma peanut and bahiagrass components, respe ctively) compared with RP67 (39 and 41 d for rhizoma peanut and bahiagrass components, respectively).

PAGE 54

54 Mixing additional species to monoculture plant litter may result in synergistic effects on decomposition rates, and these effects can be non additive rel ative to the responses observed for the monocultures (Zhang et al., 2013; Cuchietti et al., 2014). Incubation of rapidly (greater N and lesser fiber concentrations) and slowly decomposing materials (lesser N and greater fiber concentrations) in mixtures s howed that decomposition of rapidly decomposing species increased in 73 and 70% of mixture combinations incubated for 2 or 9 mo, respectively (Cuchietti et al., 2014). They found that after 270 d of incubation mixing species with larger differences in ini tial N concentration showed a greater synergistic effect on decomposition than those with smaller differences. Incubating plant litter from native species in Inner Mongolia, including a perennial bunchgrass, a semi shrub, and a perennial forb in one and t wo species combinations, Liu et al. (2007) found a positive non additive effect of mixtures on decomposition rate compared with single species. They attributed this to differences in initial N and P concentrations in incubated material. Unlike in the curre nt study, incubation of a legume ( Melissitus ruthenica ) with a grass ( Leymus chinensis or Stipa baicalensis ) in Inner Mongolia accelerated biomass losses from the grasses with no effect on legume decomposition (Zhang et al., 2013). The authors suggest ed th at the soil microbial community present at the incubation site was likely better adapted to decompose grasses, since the area encompasses a grass dominated ecosystem. Incubation of single and two species mixtures of fertilized and unfertilized maize and p ine needles for 190 d showed a transfer of C in the direction of pine to maize litter in all mixture combinations, while N was always transferred from greater to lesser N materials (Berglund et al., 2013). The authors argue that this transfer of C and N m ay have occurred as amino acids were transported by fungi, further complicating identification of individual species decomposition during incubation of mixtures.

PAGE 55

55 Factors in addition to litter chemical composition have been proposed to explain the synergis tic effects of mixed species on litter decomposition. Chapman et al. (2013) incubated leaf litter material of four species from a mixed conifer forest with a wide range of C:N and lignin:N ratios (71, 45, 47, 71, and 12, 14, 18, and 26, respectively) in on e bi and all species combinations for 3 10 and 27 mo periods. They found that total, bacterial, and fungal phospholipid fatty acids (an indicator of microbial biomass) were 70 and 20% greater for mixed compared with single species material after 1 0 and 27 mo of incubation, respectively, indicating a strong positive effect of mixing species on the microbial community (Chapman et al., 2013). Gram positive bacteria biomass was greater in mixed relative to single species plant litter, and four species mixtures had greater fungal and bacterial biomass after 10 mo compared with two species mixtures (Chapman et al., 2013). Although microbial population was not correlated to mass loss in mixed species, they suggested that diversity of microbial population in mixed species litter may be responsible for decomposition rate differences in mixed relative to single species litter. Remaining N and Litter C:N R atio Litter treatment had no effect on k for N decay ( P = 0.373) or relative half life of N ( P = 0.137), but N remaining at the end of incubation was affected ( P < 0.001) by proportion of legume in litter (linear [ P < 0.001] and quadratic [ P = 0.0015] effects). At 128 d of incubation, remaining N was greater for the BG treatment (85%) than for any treatment t hat included legume ( P < 0.002), which averaged 56, 56, and 55% for RP33, RP67, and RP. Remaining N of the BGN treatment (78%) did not differ ( P = 0.673) from that of BG. Similar to our findings, Silva et al. (2012) reported greater remaining N for signalg rass (73%) relative to signalgrass calopo mixture and pure calopo (61 and 50%, respectively).

PAGE 56

56 There was an increase in remaining N at 4 and 8 d of incubation relati ve to initial composition at Day 0 driving predicted initial remaining N above 100% (Figur e 3 2 .B ). Greater plant litter N remaining for observed versus model predicted values early in the incubation period were also observed for white clover ( Trifolium repens L.) and attributed to immobilization of environmental N (McCurdy et al., 2013). In Ti fton 85 bermudagrass, N immobilization increased as stubble heights (8 to 24 cm) and N fertilization (50, 150, or 250 kg N ha 1 yr 1 ) increased, with maximum immobilization starting progressively earlier during incubation as N fertilization increased (Liu et al., 2011). Nitrogen mineralization in that experiment was not observed before 64 d of incubation for the greater N fertilizer level treatments and N immobilization increased throughout the whole incubation period for the 50 kg N ha 1 yr 1 level. Similar to the response of remaining biomass in the current study, differences were observed in N remaining at the end of incubation even though k was similar across treatments. The period during which a linear decrease occurred in remaining N was only 14 d for BG, compared with 119, 74, and 30 d for RP33, RP 67, and RP (Figure 3 2 .B ). The greater time period of linear decline in remaining N in the RP33 treatment is probably related to the greater decomposition rate of the rhizoma peanut component in that t reatment. The effect of legume proportion was linear ( P < 0.001) for k of litter C:N ratio (Figure 3 4), with k values becoming more negative as legume proportion decreased ( 0.0019 for RP to 0.0055 g g 1 d 1 for BG). The rate of decrease in remaining li tter C:N ratio during incubation of BG was similar to that of BGN ( P = 0.530; k = 0.0055 and 0.0047 g g 1 d 1 respectively) but differed from all legume containing treatments ( P < 0.026). The decrease in C:N ratio during decomposition indicates greater disappearance of C relative to N compounds during incubation,

PAGE 57

57 suggesting that a larger proportion of total litter C was available for microbial decomposition relative to that of N. Carbon:N ratio in 50 50 signalgrass calopo and 100% calopo litter was near ly constant during a 256 d incubation, remaining between 20 and 30, while C:N in pure signalgrass litter decreased from 100 to a range of 77 to 93 (Silva et al., 2012). The decrease in remaining N values indicated N mineralization, which was observed in pu re signalgrass litter despite its high (above 70) C:N ratio at the end of the experiment. Liu et al. (2011) observed N immobilization of Tifton 85 bermudagrass with initial C:N ratios as low as 40, which are considerable lower than those reported by Silva et al. (2012). These data indicate that both immobilization and mineralization of N occur at contrasting C:N ratios, and support the concept that the forms of C and N are also of major importance in nutrient turnover from plant litter (Heal et al., 1997). There were linear ( P < 0.001) and quadratic ( P = 0.022) effects of legume proportion on remaining litter C:N at the end of incubation. After 128 d, C:N ratio was 20 for BG, which was greater than for any other treatment ( P < 0.005). The quadratic effect o ccurred because C:N ratio for the treatments containing legume did not differ (16, 15, and 13 for RP33, RP67, and RP, respectively), although C:N of RP33 and RP approached significance ( P = 0.06). The endpoint C:N ratio of BGN was 16, which was less than B G and greater than RP only ( P = 0.035) (Figure 3 4). At the end of incubation, all treatments reached C:N ratios less than 25 indicating favorable composition for mineralization (Heal et al., 1997). The similar C:N ratio in BGN and mixtures of bahiagrass a nd rhizoma peanut suggests that the integration of legumes into grass swards can achieve litter C:N ratios that favor N mineralization, similar to those achieved through N fertilizer application. This, associated with the greater initial N concentration (1 3, 17, 21, 26, and 31 g kg OM 1 for BG, BGN, RP33, RP67, and RP, respectively) and lower remaining N when

PAGE 58

58 legumes are added to grass litter indicate greater potential N return to the environment in legume containing treatments compared with N fertilized BG However, the magnitude of the effect of N return will depend on litter deposition mass. Lignin, Lignin:N, and L ignin:ADIN Proportion of legume affected litter lignin concentration at the end of the incubation period ( P = 0.034), rate of lignin concentration increase during incubation ( P < 0.001), and days to reach a linear plateau in lignin concentration ( P < 0.001). Rate of increase in lignin concentration before it reached the linear plateau was greater for RP than al l other treatments ( P < 0.022). Number of days for lignin concentration to stabilize were 44, 73, 73, 77, and 76 for RP, RP67, RP33, BGN, and BG, respectively. There was a quadratic effect ( P = 0.0136) of legume proportion on lignin concentration at the en d of incubation, with lignin concentration increasing from 220 g kg 1 for BG to 259 and 269 g kg 1 for RP33 and RP67 and then decreasing to 232 g kg 1 for RP. Although N is considered to be the most common compound limiting decomposition and C:N ratio is often used as a measurement of litter quality, nature of C, secondary compounds, other nutrients, and their interactions may significantly impact decomposition (Heal et al., 1997). During litter decomposition, recalcitrant materials tend to increase in con centration over time because more labile materials disappear and new, lignin like components are produced by microorganisms (Berg and Matzner, 1997; Gijsman et al., 1997, Berg and McClaugherty, 2008). In our study, N and lignin concentration prior to incub ation were greater at larger proportions of rhizoma peanut. During decomposition, rate of increase in lignin concentration was greater in RP compared with RP33 and RP67 (5.8 compared with 3.4 and 3.0 g kg 1 OM d 1 respectively), but this increase ceased e arlier (44 d) compared with RP33 and RP67 (73 d). Soluble N has been documented to inhibit lignin decomposition (Carreiro et al., 2000). The faster accumulation of

PAGE 59

59 lignin in RP probably occurred because the greater N concentration in this treatment inhibit ed lignin degradation, causing it to increase in concentration The lower lignin concentration found at 128 d in RP relative to the other rhizoma peanut treatments probably occurred due to its linear plateau being reached sooner compared with RP33 and RP67 Increasing lignin concentration with the inclusion of legume in grass litter was observed for signalgrass and calopo mixtures (Silva et al., 2012). The authors found that grass and legume grass mixtures reached a linear plateau in lignin concentration a t approximately 30 d of incubation, with the plateau occurring at concentrations of 160, 200, and 220 g lignin kg 1 OM for grass, the 50 50 mixture, and legume litter, respectively. Although the lignin concentrations reported by Silva et al. (2012) are sim ilar to what we observed, a longer period was necessary to achieve a linear plateau in the present study (Figure 3 5). In elephantgrass root decomposition, lignin concentration also increased during incubation reaching a linear plateau at concentrations of 220 or 200 g kg 1 OM at 240 or 118 d (Silva et al., 2015). There was a quadratic effect ( P = 0.016) of legume proportion on days for lignin:N ratio to reach linear plateau, with values of 32, 46, 40, and 33 d for BG, RP33, RP67, and RP treatments, respec tively (Figure 3 6). The plateau was reached at Day 38 for BGN. These values are at the lower part of the range of 40 to 60 d reported by Dubeux et al. (2006 a ) when evaluating bahiagrass decomposition. Lignin:N ratio at the end of the incubation period dec reased linearly ( P < 0.001) from 8.5 to 5.2 as legume proportion in litter increased (Table 3 2). Similar to the results of the current study, lignin:N ratio of litter at the end of the incubation period decreased with addition of calopo to signalgrass, an d it was at the lower end of the range (8 25) reported ( Silva et al., 2012). Dubeux et al. (2006a) observed an increase in lignin:N ratio in bahiagrass decomposition over time, with this ratio being lower as stocking rate and N

PAGE 60

60 fertilization rate increased The authors suggest that lignin:N is a better indicator than C:N of bahiagrass litter quality and decomposition rate because of its indication of C availability to microbial decomposition. In our study, lower lignin:N ratio was not associated with greate r decomposition rate. One possible explanation for the different relationship of lignin:N to decomposition in our study relative to that of Dubeux et al. (2006 a ) is that the form of N in RP33 was more soluble than in the treatments with greater proportions of rhizoma peanut. At 128 d ays of incubation N bound to fiber relative to total N concentration in litter material was 430 g ADIN kg 1 N, while in RP67 and RP these values were 637 and 560 g ADIN kg 1 N, respectively. These results emphasize that the ava ilability of N during decomposition, and not only that of C, is important to determine nutrient cycling in the environment. Increasing the proportion of legume in litter resulted in a linear decline ( P = 0.006) in lignin:ADIN ratio at the end of the incubation period (Table 3 2). Proportion of legume in litter affected lignin:ADIN ratio decay rate ( P < 0.001), with the rate decreasing linearly ( P < 0.001) as legume proportion increased ( 0.0083, 0.0054 0.0025, and 0.0018 g g 1 d 1 for RP, RP67, RP33, and BG, respectively; Figure 3 7). This difference indicates that although N is becoming more unavailable in all treatments relative to C, more N became unavailable in legume litter with lignin:ADIN rat io being 1.5 to almost 5 times greater than in other treatments. This is in agreement with Rovira and Vallejo (2002), who found that N recalcitrance increases to a greater degree in plants with contrasting chemical characteristics during decomposition, whi le increasing C recalcitrance during incubation occurs primarily in plant material with initially more labile C composition. Silva et al. (2012) also observed greater decomposition rate of lignin:ADIN in signalgrass calopo mixture and pure calopo relative to pure signalgrass litter. Since lignin concentration stabilized in that study at around 32 d, the reduction in lignin:ADIN was attributed

PAGE 61

61 to increasing ADIN concentration due to formations of complexes between N and lignin. Carbohydrates and lignin pheno lic groups and quinones may react with NH 3 during decomposition, making it unavailable for microbial utilization (Berg and McClaugherty, 2008). A similar response was observed in our study in regards to lignin:ADIN, since ADIN concentration stabilized at 8 2 d and lignin within a range of 43 to 77 d. As lignin concentration reached a plateau earlier and faster in the RP treatment, lignin:ADIN decay rate was greater for RP compared with other treatments (Figure 3 7). ADIN Concentration in the OM and in T otal N Legume proportion in litter had no effect on number of days to reach the linear plateau for ADIN concentration in OM or total N, but both responses increased with time of incubation. At the end of the incubation period, ADIN concentration in OM increased linearly with increasing legume proportion ( P < 0.001), while the linear effect on ADIN concentration in total N approached significance ( P = 0.107) (Table 3 2). Acid detergent insoluble N concentration in the OM plateaued at 82 d, while ADIN concentratio n in N stabilized at 65 d. Bahiagrass litter ADIN concentration in total N for pastures managed at a range of intensities, reached a plateau of approximately 500 g kg 1 meaning that around 50% of total N was likely unavailable for microbial utilization (D ubeux et al., 2006a). In pure signalgrass, around 68% of N was bound to fiber after 256 d of incubation, while less (54%) N was unavailable when legume proportion was 50 and 100% (Silva et al., 2012). In elephantgrass roots, ADIN also increased with incuba tion time reaching a linear plateau after 128 d in 2 yr (Silva et al., 2015). Lesser ADIN concentration in total N in the second year of their experiment lead to greater root decomposition rates in that year relative to the first year. These results and th ose of our study support the premise that nature of N and C are important indicators of decomposition and nutrient cycling.

PAGE 62

62 Remaining P and C:P There was no effect of plant litter composition on remaining P decay rate ( P = 0.5461) or relative half life ( P = 0.190), but increasing litter legume proportion linearly increased ( P < 0.001) remaining P at the end of incubation from 39% for BG to 45, 58, and 66% for RP33, RP67, and RP, respectively. Treatments BG and BGN were also different, with P remaining proportions of 39 and 61%, respectively ( P = 0.0148). Remaining P observed in this study was similar to that reported by Dubeux et al. (2006), who found ~40% remaining P for bahiagrass litter incubated for 128 d. Phosphorus may be an important limiting factor in decomposition of plant litter in tropical environments, where its avail ability in the soil is typically low (Gijsman et al., 1997). In bahiagrass pastures, litter decomposition contributed little to P inputs to the soil, with estimated P returns of only ~ 3 kg P ha 1 (Dubeux et al., 2006a). Legume proportion in plant litter h ad a linear effect on decay rate of C:P ( P < 0.001) and linear ( P < 0.001) and quadratic ( P = 0.02) effects on final C:P ratio. The C:P ratio increased ( k = 0.0019 g g 1 d 1 ) for the BG treatment during incubation, indicating P immobilization (Figure 3 8). This was also observed in bahiagrass litter incubated for 128 d (Dubeux et al., 2006a). All other treatments decreased in C:P ratio during incubation and decay rates were not different from each other ( P > 0.135). Final C:P ratio was greater for BG (249) compared with all other treatments ( P < 0.001; 141, 138, 118, and 85 for BGN, RP33, RP67, and RP treatments, respectively; Table 3 2). The classification of plant litter material by nutrient richness for both C:P and C:N followed the same order, with BG > BGN > RP33 > RP67 > RP throughout incubation, except for initial C:P ratio in BG being slightly less than that of BGN (Table 3 1). This agrees with Manzoni et al. (2015), who found that plant litter C:N and C:P at beginning of incubation are

PAGE 63

63 strongly rela ted, i.e., nutrient richness of N tends to indicate the same for P and the reverse is also true. Conclusions Losses of total biomass were greater for litter mixtures of rhizoma peanut bahiagrass consisting of 33% rhizoma peanut compared with an unfertilized bahiagrass monoculture (35 relative to 43% remaining biomass, respectively). Mixtures with greater proporti ons of rhizoma peanut and an N fertilized bahiagrass monoculture had intermediate values of remaining total biomass after 128 of decomposition (~38%) and did not differ from BG and RP33 treatments. The differences in total biomass losses occurred even thou gh decomposition rate k was not different across treatments. The difference was due to the greater period during which linear decay (period between initial and final plateau) in decomposition occurred in RP33 compared with other treatments (69 and ~51 d, r espectively). In addition, decomposition of the rhizoma peanut component was greater in RP33 relative to RP67 and RP, but decomposition of the bahiagrass component was not affected by presence of legumes vs. in monocultures. Greater decomposition of rhizom a peanut in RP33 was associated with a longer period of linear decline (55 d) compared with other treatments (39 and 31 d for RP67 and RP, respectively). Relative to incubated monocultures, the linear decrease in incubated biomass in mixtures started earli er in the rhizoma peanut and later in the bahiagrass component. Remaining N at end of incubation was less in rhizoma peanut relative to bahiagrass treatments. Similarly to biomass decomposition, differences in the extent of remaining N losses were probabl y due to the longer period of linear decline in rhizoma peanut treatments (~74 d; as along as 119 d in RP33) relative to pure bahiagrass (~21 d). These results indicate greater N mineralization when legumes were present in litter even though C:N of BGN was similar to that of RP33 and RP67 by the end of incubation.

PAGE 64

64 Nature of C and N changed during incubation, with both becoming more unavailable as decomposition advanced. Lignin:ADIN ratio decreased as incubation advanced, implying that N became unavailable at a faster rate than C during decomposition. This change was on average almost four times greater in RP when compared with the other treatments. Between 43 and 64% of N present at the end of incubation was unavailable for microbial utilization. Lignin:N, lignin:ADIN, and ADIN concentration in total N were similar between RP33 and BG at the end of incubation, indicating that other factors such as microbial activity may have been responsible for greater biomass losses observed in RP33. Phosphorus immobilizat ion was observed in BG litter but not in any other treatment. This experiment allowed us to simultaneously compare fertilizer to legume sourced N input in decomposition and nutrient cycling of grass based systems. Our results indicate that in terms of nut rient return to the environment, legume grass mixtures are superior to unfertilized grass monoculture, and are equal or superior to fertilized grass or legume monoculture. This response is primarily related to the greater N and P mineralization in the pres ence of legumes and to greater decomposition of the legume component in mixtures in grasses. However, part of the increase in decomposition cannot be attributed solely to differences in chemical composition across treatments and is likely related to greate r diversity in microbial populations in the presence of mixtures relative to monocultures.

PAGE 65

65 Table 3 1. Chemical composition of plant litter at beginning of incubation for 2 yr in Gainesville, FL. Treatments were: bahiagrass fertilized with 60 kg N ha 1 (BGN), bahiagrass (BG), mixtures of bahiagrass and rhizoma peanut in 67 33% (RP33) and 33 67% (RP67), and pure rhizoma peanut (RP) Treat N P Lignin ADIN g ADIN kg 1 N Lignin:ADIN Lignin:N C:N C:P _____ ________________ ___ g kg 1 OM _____ ________________ ___ ______ ____________ ______ dimensionless _________ ____________ ___ BGN 17 2.6 29 1.6 92 18 1.7 28 188 BG 13 2.5 27 1.1 88 27 2.2 38 193 RP33 21 3.3 43 2.5 119 19 2.0 23 151 RP67 26 3.7 53 2.9 110 19 2.0 19 136 RP 31 4.1 63 3.4 105 20 2.1 16 124 Table 3 2 Chemical composition of plant litter at end of incubation for 2 yr in Gainesville, FL. Treatments were: bahiagrass fertilized with 60 kg N ha 1 (BGN), bahiagrass (BG), mixtures of bahiagrass and rhizoma peanut in 67 33% (RP33) and 33 67% (RP67), and pure rhizoma peanut (RP). Treat. N P Lignin ADIN g ADIN kg 1 N Lignin:ADIN Lignin:N C:N C:P ________________ ________ g kg 1 OM ________________ ________ ________ ____________ ____ dimensionless _____ ____________ _______ BGN 34 3.9 229 19 542 14 7 16 141 BG 26 2.4 220 14 515 18 9 20 249 RP33 35 4.4 259 15 430 18 8 16 139 RP67 39 5.7 269 25 637 11 7 15 119 RP 45 6.9 232 25 560 11 5 13 86 SE 1.6 1.0 14 2 7 3 0.4 0.7 35 Lin effect P < 0.00 1 P < 0.0 01 NS P < 0.00 1 P = 0.107 P = 0.006 P < 0.001 P < 0.001 P < 0.001 Quad. effect NS NS P = 0.014 NS NS NS NS P = 0.022 P = 0.02 Polynomial contrast of effect of legume inclusion level Contrasts were calculated using BG, RP33, RP67, and RP only.

PAGE 66

66 Figure 3 1. Monthly weather data at Hague, FL (18 km from experimental site) during the years of evaluation and the 30 yr average for Gainesville, FL. Total annual rainfall was 1546 mm in 2014 and 1344 mm in 2015, compared with the 30 yr average of 1203 mm.

PAGE 67

67 Figure 3 2. Remaining biomass and N decay of plant litter during 128 d incubation period for 2 yr in Gainesville, FL A ) Remaining biomass and B) R emaining N T reatments a re: bahiagrass (BG), bahiagrass fertilized with 60 kg N ha 1 (BGN), mixtures of bahiagrass and rhizoma peanut in 67 33% (RP33) and 33 67% (RP67), and pure rhizoma peanut (RP) Points represent measured response variables (mean SE) and line s represents model estimates. For remaining biomass, e quations were: BG = {(39 99)/[1 + (x/31) 1.9 ]} + 99; BGN = {(32 101)/[1 + (x/33) 1.9 ]} + 101; RP33 = {(29 97)/[1 + (x/34) 1.7 ]} + 97; RP67 = {(36 96)/[1 + (x/25) 2.0 ]} + 96; RP = {(34 9 9 )/[1 + (x/23) 1.8 ]} + 99 For remaining N, equations were: BG = {(83 108 )/[1 + (x/ 19 ) 4.3 ]} + 108 ; BGN = {( 75 106 )/[1 + (x/ 23 ) 2.8 ]} + 106 ; RP33 = {(47 102)/[1 + (x/28) 1.0 ]} + 102; RP67 = {(51 98)/[1 + (x/26) 1.3 ]} + 98; RP = {(53 99)/[1 + (x/20) 2.2 ]} + 99 ; x is time in days. A) B )

PAGE 68

68 Figure 3 3. Remaining rhizoma peanut biomass for mixtures of plant litter during 128 d incubation period over 2 yr in Gainesville, FL. Treatments a re: bahiagrass and rhizoma peanut in 67 33% (RP33) and 33 67% (RP67) and pure rhizoma peanut (RP) Points represent measured response variables (mean SE) and line s represents model estimates. E quations were: RP33 = {(13 100)/[1 + (x/26) 1.1 ]} + 100; RP67 = {(30 98)/[1 + (x/25) 1.5 ]} + 98; RP = {(36 99)/[1 + (x/23) 1.8 ]} + 99. Number of days is referred to as x in equations; x is time in days. Figure 3 4. Carbon:N ratio of plant litter during 128 d incubation period for 2 yr in Gainesville, FL. Treatments are: bahiagrass (BG), bahiagrass fertilized with 60 kg N ha 1 (BGN), mixtures of bahiagrass and rhizoma peanut in 67 33% (RP33) and 33 67% (RP67), and pure rhizoma peanut (RP). Points represent measured response variables (mean SE) and line s represents model estimates. Equations were: BG = 36 e ( 0.0053 x) ; BGN = 27 e ( 0.0046 x) ; RP33 = 25 e ( 0.0041 x) ; RP67 = 20 e ( 0.0027 x) ; RP = 17 e ( 0.0020 x) ; x is time in days.

PAGE 69

69 Figure 3 5. Lignin concentration of plant litter during 128 d incubation period for 2 yr in Gainesville, FL. Treatments are: ba hiagrass (BG), bahiagrass fertilized with 60 kg N ha 1 (BGN), mixtures of bahiagrass and rhizoma peanut in 67 33% (RP33) and 33 67% (RP67) and pure rhizoma peanut (RP). Points represent measured response variables (mean SE) and line s represents model e stimates. Equations were: BG = 220 + 2.6 (x 74); BGN = 228 + 2.5 (x 77); RP33 = 258 + 3.2 (x 70); RP67 = 269 + 3.0 (x 71); RP = 238 + 4.7 (x 38); x is time in days. Figure 3 6. Lignin:N ratio of plant litter during 128 d incubation p eriod for 2 yr in Gainesville, FL. Treatments are: bahiagrass (BG), bahiagrass fertilized with 60 kg N ha 1 (BGN), mixtures of bahiagrass and rhizoma peanut in 67 33% (RP33) and 33 67% (RP67) and pure rhizoma peanut (RP ) Points represent measured respon se variables (mean SE) and line s represents model estimates. Equations were: BG = 5.9 + 0.1 6 ( x 26 ) ; BGN = 6.9 + 0.15 (x 35 ) ; RP33 = 7.5 + 0.14 (x 42 ) ; RP67 = 6 .9 + 0.12 (x 4 1) ; RP = 5.9 + 0.16 (x 26 ); x is time in days.

PAGE 70

70 Figure 3 7. Lignin:ADIN ratio of plant litter during 128 d incubation period for 2 yr in Gainesville, FL. Treatments are: bahiagrass (BG), bahiagrass fertilized with 60 kg N ha 1 (BGN), mixtures of bahiagrass and rhizoma peanut in 67 33% (RP33) and 33 67% (RP67) and pure rhizoma peanut (RP ) Points represent measured response variable s (mean SE) and line s represents model estimates. Equations were: BG = 21 e ( 0.0011 x) ; BGN = 16 e ( 0.00 09 x) ; RP33 = 23 e ( 0.0029 x) ; RP67 = 19 e ( 0.0049 x) ; RP = 2 4 e ( 0.0062 x) ; x is time in days.

PAGE 71

71 Figure 3 8. Carbon:P ratio of plant litter during 128 d incubation period for 2 yr in Gainesville, FL. Treatments are: bahiagrass (BG), bahiagrass fertilized with 60 kg N ha 1 (BGN), mixtures of bahiagrass and rhizoma peanut in 67 33% (RP33) and 33 67% (RP67) and pure rhizoma peanut (RP ). Points represent measured response variables (mean SE) and line s represents model estimates. BG = 162 e ( 0.0027 x) ; BGN = 168 e ( 0.0 0 13 x) ; RP33 = 147 e ( 0.0008 x) ; RP67 = 123 e ( 0.0008 x) ; RP = 121 e ( 0.0038 x) ; x is time in days.

PAGE 72

72 CHAPTER 4 NITROGEN FERTILIZATION AND PROPORTION OF LEGUME AFFECT LITTER DEPOSITION, DECOMPOSITION, AND NUTRIENT RETURN IN BAHIAGRASS PAST URES Introduction Cost of inorganic N fertilizer has increased in recent years, and greater cost and lack of availability may limit fertilizer applications to warm climate grasslands (Dubeux et al., 2007). This imposes challenges to maintaining adequate plant nutrition to support plant growth and sustain cattle production, especially when associated with poor grazing management (Braz et al., 2013). The outcome may be pasture degradation and severe limits to ecosystem services they provide (Sollenberger, 2 014). Introduction of legumes into warm season, perennial grass swards can provide N to the grazing system through cycling in animal excreta and plant litter decomposition. A 50 50 grass legume mixture fertilized with 50 kg N ha 1 had similar herbage accum ulation to a grass monoculture fertilized with 450 kg N ha 1 (Nyfeler et al., 2009), indicating an opportunity for some associations to provide N to the system without sacrificing production. In fertilized bahiagrass ( Paspalum notatum Flgge), Dubeux et al. (2006b) found that plant litter had the potential to contribute 57 kg N ha 1 yr 1 when the pasture received 40 kg N fertilizer ha 1 yr 1 For signalgrass [ Brachiaria decumbens (Stapf) R. D. Webster] calopo [ Calopogonium mucunoides D esv.]) plant litter mixtures with legume proportion of 0, 50, and 100%, Silva et al. (2012) found that litter decomposed to a greater extent in treatments containing legume when compared with grass alone. The differences observed were attributed to contras ting chemical characteristics of the incubated plant materials, particularly greater N concentration with the presence of legumes. Rezende et al. (1999) also observed an increase in decomposition rate of litter as proportion of legume increased in a mixed grass legume pasture.

PAGE 73

73 In addition to the effect of including legumes on litter chemical characteristics, greater microbial diversity may occur in mixed species litter and can play an important role in determining rate and extent of litter decomposition (C hapman et al., 2013). Abiotic factors also affect these responses, with greater temperature typically increasing decomposition rates (Anderson and Hetherington, 1999) and greater rainfall having either positive (Silva et al., 2012) or negative effects (Apo linrio et al., 2014) on decomposition dynamics depending on amount. Although efforts continue to increase adoption of legumes in grasslands (Castillo et al., 2014; Mullenix et al., 2016a; b), more information is needed about how proportion of legume affe cts nutrient cycling in grass swards, particularly when compared with grass monocultures under typical management including N fertilization. The objective of this study was to quantify the effects of N fertilization of bahiagrass monoculture and proportion of rhizoma peanut ( Arachis glabrata Benth.) in mixtures with bahiagrass on litter deposition, composition, decomposition rate, and nutrient disappearance. Material and Methods Experimental S ite A 2 yr experiment was conducted at the University of Florid a Beef Research Unit (BRU) in Gainesville, FL (29.72 N, 82.35W) during the summers of 2015 and 2016 on well established pastures of bahiagrass monoculture and mixtures of rhizoma peanut with bahiagrass. Soils at the site are Sparr fine sand (loamy, silic eous, hyperthermic Grossarenic Paleudult). Soil samples were taken in May 2014 to a 15 cm depth and analyzed at the University of Florida Extension Soil Testing Laboratory. Soil pH ranged from 5.9 to 6.2 and Melich 3 extractable P, K, and Mg ranged from 11 2 to 138, 52 to 78, and 83 to 122 mg kg 1 respectively; all nutrient concentrations were classified as high. Based on these results, no lime or fertilizer were applied

PAGE 74

74 except N for the N fertilized bahiagrass treatment. Weather data from the experimental years are presented in Figure 4 1. Treatments and Experimental D esign The experiment included four treatments replicated three times in a completely randomized design. Plot area was 20 x 48 m 2 Two treatments were bahiagrass monoculture, one receiving 0 kg N ha 1 (BG) and the other receiving 50 kg N ha 1 (BGN) in late spring/early summer. The other two treatments were mixtures of rhizoma peanut and bahiagrass, and the target rhizoma peanut botanical composition levels were 10 and 50%. The mixed pastures h ad approximately 10 to 20% rhizoma peanut in spring of the year before the study began. In order to achieve approximately 50% rhizoma peanut in those plots, they were sprayed with 875 mL ha 1 Fusilade (Fluazifop p butyl) on 12 May and 10 Oct. 2014. Across mixed treatments in the 2 yr of the study, botanical composition ranged from 6 to 78% and varied within treatments each year and between years for a given treatment. Because of this variation, it was deemed inappropriate to consider the mixed species treat ments as discrete levels of legume percentage; instead rhizoma peanut percentage was considered a continuous variable and analyzed using regression. Grazing M anagement Pastures were rotationally stocked with two animals per pasture. Grazing started on 3 an d 9 June in 2015 and 2016, respectively. Animals were maintained in the experimental area for residence periods of approximately 1 wk until target stubble height of 10 cm was reached for BG and BGN treatments and 15 cm for the mixtures. For BG and BGN, len gth of regrowth interval between grazing events was 3 wk. For the mixture treatments, regrowth interval was 6 wk. Litter D eposition Deposition of litter was measured during 14 d periods in the middle of the regrowth interval for each treatment following th e methodology described by Rezende et al. (1999).

PAGE 75

75 Because of the difference in regrowth intervals between grazing events for grass monocultures and mixtures, litter deposition was measured during Days 4 to 18 of the regrowth interval for BG and BGN and Days 14 to 28 for the bahiagrass rhizoma peanut mixtures. Specifically, a circular 0.25 m 2 quadrat was placed at three random locations per experimental unit, and plant material on the soil surface inside the quadrat was collected and dried at 60C for ~72 h. This was called existing litter. After 14 d, litter was collected from the same area, and it was considered deposited litter. This procedure occurred twice each year. Existing litter was collected on 18 Aug. and 13 Oct. 2015 and 2 Aug. and 11 Oct. 2016 In August 2015, deposited litter could not be collected due to excessive rainfall that flooded the experimental area. After an initial drying procedure and prior to weighing, both existing litter and deposited litter were sieved consecutively through 4 and 2 mm sieves to remove soil, transferred to a paper bag, and dried at 60C for an additional 12 h and weighed. Litter deposition rate was calculated by dividing the collected litter mass (14 d after clearing the existing litter) by 14 and adjusting for decomposition that occurred during the 14 d period (Rezende et al., 1999; Dubeux et al., 2006b; Liu et al., 2011a; Apolinrio et al., 2013). Litter D ecomposition Litter decomposition was evaluated using the litter bag technique. Plant material was collect ed by clipping herbage to soil level from at least four 0.25 m 2 quadrats per plot 3 wk after grazing. Litter from each experimental unit was kept separate and not combined across replicates for any treatment. For the purposes of the decomposition study, li tter was defined as fully expanded bahiagrass leaf blade and rhizoma peanut leaflets. The top 5 cm of the rhizoma peanut sward canopy was eliminated to avoid utilization of young, unexpanded leaflets. Separated fractions of rhizoma peanut and bahiagrass li tter were dried at 60C for 72 h. The proportion of legume and grass in this separated material was the proportion of each allocated to litter bags for

PAGE 76

76 that experimental unit In the first year, proportion of rhizoma peanut in incubated litter ranged from 7 to 44%, while in the second year this proportion was 15 to 75%. Litter was not ground to better simulate field conditions by maintaining surface integrity (Dubeux et al., 2006a). Litter bags were 100% polyester lining fabric (75 m mesh size), 15 x 20 cm in size, and closed with a heat sealer after 12 g of litter was added. Two replicate litter bags were filled per experimental unit for each of six incubation times (4, 8, 16, 32, 64, and 128 d), resulting in 12 bags per experimental unit. Two empty, heat sealed bags were placed per plot and served as blanks to account for changes in bag weight during the incubation period. Two empty bags were collected per incubation time from randomly determined plots. Bags were placed on the soil surface and protected ag ainst animal interference by the use of exclusion cages. After collection, bags were dried at 60C for 72 h, brushed to eliminate attached particles, and weighed. Blank bag weights were subtracted to account for changes in bag weight. Weight of remaining l itter was averaged across bags within a plot and collection time in order to obtain one weight per collection time per experimental unit. Incubation started on 20 September and 4 October in 2015 and 2016, respectively. Litter and Nutrient D isappearance Rem aining biomass, remaining N, remaining P, C:N, C:P, and lignin:acid detergent insoluble N (ADIN) followed a single exponential decay ( P < 0.05) model described in Equation 4 1 (Wider and Lang, 1982): f(x) = A e (k x) (4 1) where A is the disappearance coefficient, k is the relative decay rate (g g 1 d 1 ) and x is time in days. Concentration of lignin and ADIN, lignin:N ratio, and ADIN concentration in total N were analyzed using a linear plateau model ( P < 0.05; McCartor and Rouquette, 1977) as descri bed by Dubeux et al. (2006a) and Silva et al. (2012) in Equation 4 2 :

PAGE 77

77 f(x) = A + B (x C) (4 2) where A is initial chemical composition, B is the rate of increase in concentration, and C is the day in which the linear plateau was reached. Chemical Composition A nalysis Contents of the two bags collected from each experimental unit per incubation time were composited for chemical composition analysis. Samples were ground using a Wiley mill (Model 4 Thomas Wiley Laboratory Mill; Thomas Scientific) to p ass through a 1 mm stainless steel screen. Samples were analyzed for dry matter (DM) by drying at 105C for 15 h and organic matter (OM) by ashing at 500C for 4 h (Moore and Mott, 1974. Carbon and N were analyzed by dry combustion using a Flash EA 1112 NC elemental analyzer (CE Elantech, Lakewood, NJ) after samples were ball milled. Lignin was analyzed with a Daisy incubator (ANKOM Technology, 2017a). Acid detergent insoluble N was obtained by Micro Kjeldahl digestion (Gallaher et al., 1975) of samples aft er obtaining the acid detergent fiber (ADF) residue (ANKOM Technology, 2017b). All chemical composition results were expressed on an OM basis to eliminate mineral particle effects on nutrient concentration. Nutrient ratios were obtained by expressing nutri ent concentration on an OM basis (g kg OM 1 ). Statistical A nalysis Data were analyzed using the R software (R Core Team, 2016). Years were analyzed separately because weather conditions differed considerably between the first and second years of the experi ment (Figure 4 1), and rainfall and temperature affect plant litter dynamics (Dubeux et al., 2006a; b; Silva et al., 2012; Apolinrio et al., 2013). For response variables analyzed with the single exponential decay model, k (relative decay rate) was compar ed. For response variables following the linear plateau model, C (days to reach linear plateau) was analyzed. Chemical composition prior to incubation (initial composition) and at the end of the 128 d incubation

PAGE 78

78 period (final composition) were compared. Fo r comparisons of BG and BGN treatments, a linear (Lenth, 2016) and are rep orted for significant main effects ( P < 0.05). For bahiagrass rhizoma peanut mixtures, decomposition response variables were regressed on actual incubated legume proportion placed in the litter bag (based on separation of bahiagrass leaf blade and all RP l eaflets except those in the top 5 cm of the canopy) and deposited and existing litter responses were regressed on actual pasture legume proportion (from botanical composition samples take n to soil level). Results and Discussion Initial Litter C omposition At the beginning of incubation, average litter N concentration across treatments was 23 and 29 g N kg 1 OM in 2015 and 2016, respectively, and legume proportion in litter explained 91 and 97% of the variation in N concentration in the 2 yr ( P = 0.0028 and 0.0004 in 2015 and 2016, respectively; Figure 4 2). Similarly, C:N ratio at the beginning of incubation was negatively related to legume proportion in litter (R 2 = 0.89 and 0.90; P = 0.0050 and 0.0038 in 2015 and 2016, respectively). Nitrogen concentration in legume grass litter was greater than that of pure bahiagrass treatments ( P = 0.0005 and 0.0035, in 2015 and 2016, respectively), which did not differ from each other in 2015 ( P = 0.94; 16 g N kg 1 OM for both treatments) or 2016 ( P = 0.26 ; 17 and 22 g N kg 1 OM for BG and BGN, respectively). Ratio of C:N was similar between BG and BGN ( P = 0.7936 and 0.0889 in 2015 and 2016, respectively) and greater in pure bahiagrass treatments compared with legume grass mixtures ( P = 0.0005 and 0.0001 i n 2015 and 2016, respectively).

PAGE 79

79 Dubeux et al. (2006a) evaluated litter composition of bahiagrass under increasing management intensities (40, 120, and 360 kg N ha 1 associated with stocking rates of 1.3, 2.7, 4.0 animal units [AU, one AU = 500 kg live weig hts] ha 1 for Low, Medium, and High management levels, respectively). In their study, litter N concentration at beginning of incubation was 15 to 17 g kg 1 while in our study this range was 15 to 22 for the grass treatments and 18 to 35 g kg 1 OM for the mixtures. One reason for somewhat lesser N concentrations reported by Dubeux et al. (2006a) is they defined litter as senescent leaves still attached to the plant, while in the current study leaves of any age were used as long as they were fully expanded. As leaves senesce, N and other nutrients are transferred to new growing tissue. In switchgrass ( Panicum virgatum L.), between 23 and 61% of N in senescent tillers was remobilized into the plant canopy (Yang et al., 2009). In the perennial grasses Festuca r ubra and Agrostis capillaris it was demonstrated that remobilized N represented 70 to 85% of total N in new leaves at initial stages of spring regrowth compared with 35 to 45% as growing season advanced, and was sourced exclusively from senescing leaves r ather than belowground biomass (Bausenwein et al., 2001). This mechanism likely explains the difference in N concentration in current pure bahiagrass treatments compared with Dubeux et al. (2006a), but the presence of legume in the mixed treatments of the current study was primarily responsible for the much larger litter N concentrations (up to 35 g kg 1 OM) in those treatments relative to the grass monocultures At the beginning of incubation, proportion of legume had a weak relationship to lignin concentr ation (R 2 of 0.15 and 0.27; P = 0.44 and 0.37 in 2015 and 2016, respectively). Dubeux et al. (2006a) found that lignin:N rather than C:N or lignin concentration might be a better indicator of decomposition at later stages, since it indicates C availability for microorganisms. However, in our study decomposition was greater in 2015, when lignin:N was greater (4.8 in 2015

PAGE 80

80 compared to 2.8 in 2016). One reason for this may be that a larger proportion of N was unavailable for microbial utilization in the form of ADIN in that year (418 and 254 g ADIN kg 1 N in 2015 and 2016). This supports the argument of others that the form of C and N rather than C:N ratio only must be taken into account when evaluating litter decomposition dynamics (Heal et al., 1997; Silva et al., 2012). Litter D ecomposition Proportion of legume explained 93 and 75% of the variation in decomposition rate ( k ) in 2015 and 2016, with a more negative relative rate as proportion of legume increased, i.e., a more rapid rate of decomposition (Figure 4 3). As a result, after 128 d of incubation remaining biomass decreased as proportion of legume in litter increased (Figure 4 4). The observed greater initial N concentration as legume proportion increased was probably responsible for greater biomass disap pearance. In heather and bracken pure and mixed litter, addition of 50 kg N ha 1 yr 1 increased biomass losses by 11, 7, and 12% (Anderson and Hetherington, 1999). Extent of decomposition in deciduous and coniferous tree species was linearly correlated to initial N concentration, as well as that of other nutrients such as Mn and Ca (Berg e t al., 1996; Berg and Matzner, 1997). The authors attributed this effect to greater production of cellulose digesting enzymes in the greater N containing plant material, which was observed in forest floor litter enriched with soluble N (Carreiro et al., 20 00). Decay rate of the pure bahiagrass treatments were similar to each other in 2015 ( P = 0.3351, k = 0.0048 and 0.0051 g g 1 d 1 for BG and BGN, respectively) and 2016 ( P = 0.8007, k = 0.0014 g g 1 d 1 ), and were less negative (i.e., decomposition was slower) relative to treatments containing rhizoma peanut ( P = 0.0113 and 0.0013 in 2015 and 2016, respectively). Dubeux et al. (2006a) evaluated litter decomposition of bahiagrass under previously described Low, Medium, and High management levels over 128 d. The k value reported for the High

PAGE 81

81 management level was 0.0030 g g 1 d 1 which was similar to what we found for the treatments in this study as an average across years (Figure 4 3). This occurred even though the High management used by Dubeux et al. (2 006a) was associated with considerably greater N input (360 kg N ha 1 ), while legume grass mixtures in our study received no N fertilizer and BGN received only 50 kg N ha 1 yr 1 Biomass relative decay rate k has been documented to be affected by presence of legume in pastures. In pure grass ( Brachiaria humidicola ) and grass legume ( B. humidicola Desmodium ovalifolium ) pastures stocked at 2, 3, and 4 animals ha 1 decomposition constant k ranged between 0.0036 and 0.0042 g g 1 d 1 (Rezende et al., 1999). Although presence of legume was not found to have an effect on decomposition rate, the stocking rate effect on litter decomposition rate found by Rezende et al. (1999) was attributed to the greater legume proportion in the pasture under the lesser stocking rate, and the consequent lower C:N ratio in litter in that treatment. Similarly, Silva et al. (2012) performed a 256 d incubation study evaluating decomposition of signalgrass calopo mixtures with 0, 50, and 100% legume. They found greater total biomass l osses with the inclusion of the legume only in one year of the experiment. Although both lignin and N concentration increased with the inclusion of legumes relative to pure grass treatments, there was proportionally greater lignin and lesser N concentratio n in the year no statistical difference was found in biomass loss. This emphasizes that although C:N is of primary importance when evaluating litter decomposition, their form (i.e., C and N availability for decomposition) should also be considered as a key factor determining biomass and nutrient losses (Berg and McClaugherty, 2008). Relative decay rate was more negative, and therefore decomposition occurred at a greater rate in 2015 compared with 2016 across treatments (Figure 4 3). Differences in relative decay rate affected percentage biomass remaining at the end of incubation, with less than 56% of

PAGE 82

82 biomass remaining for all mixed species litter in 2015, while in 2016 the lowest amount of remaining biomass found was 68% (Figure 4 4 ). Similarly, in BG and BGN treatments annual k differences resulted in more remaining biomass at the end of the incubation period in 2016 compared with 2015 (84 and 60%, respectively), with no difference between treatments in either year ( P = 0.76 and 0.90 in 2015 and 2016, resp ectively). Although chemical characteristics at the beginning of incubation, particularly differences in N concentration, are important drivers of decomposition rate and extent, differences in decomposition may also be associated with contrasting temperat ure and rainfall in the 2 yr (Figure 4 1). Total amount of rainfall during the initial 32 d of incubation was similar between years (24 and 29 mm in 2015 and 2016, respectively), however in 2015 there were 10 d with no rainfall compared with 17 d in 2016 ( Figure 4 1). In addition, 2016 had twice as many days with minimum temperatures under 15C when compared with 2015 (14 and 7, respectively). Others have documented temperature and rainfall effects on litter decomposition. Signalgrass calopo litter disappea red to a greater extent when rainfall was greater, although greater N concentration at the beginning of incubation in the year with greater rainfall may also have influenced the observed differences (Silva et al., 2012). Excessive rainfall, however, may hi nder decomposition due to leaching of soluble carbohydrates and N from litter, as observed for signalgrass incubated for 256 d that had greater decomposition rates when annual rainfall was 847 relative to 1290 mm (Apolinrio et al., 2014). Finally, tempera ture is strongly related to microbial activity, hence affecting decomposition. In an incubation experiment with heather litter, microbial respiration increased exponentially as temperatures increased (Anderson and Hetherington, 1999). Our results indicate that although increasing legume proportion increases litter decomposition rate and extent, the magnitude of this effect depends upon moisture and temperature.

PAGE 83

83 In addition to effects associated with changes in litter chemical characteristics, litter decompo sition of species mixtures may be affected by microbial activity and diversity. Synergic, non additive effects have been observed for different species. Evaluating decomposition of pure and mixed heather and bracken, Anderson and Hetherington (1999) found that single species litter mass losses were similar 36 and 37%, respectively. However, for mixtures composed of 20, 40, 60, and 80% of one species and the remainder of the other, mass losses were more than 10 percentage units greater, reaching 48%. The aut hors speculate that synergic interaction influenced fungal colonization of litter when the two species were mixed, increasing decomposition compared with single species litter. Similarly, incubation of aspen ( Populus tremuloides Michx), Douglas fir ( Pseudo tsuga menziesii Mirbel Franco), limber pine ( Pinus flexilis James), and ponderosa pine ( Pinus ponderos P. and C. Lawson) leaves in single, two and all species mixtures showed a 70 and 20% increase in microbial biomass for the mixtures vs. single species litter after 1 and 2 yr of incubation, respectively (Chapman et al., 2013). The authors also observed that biomass losses were greater and 20 to 50% faster than expected in mixtures (except those containing aspen) based on weighted averages of individual species (Chapman and Koch, 2007; Chapman et al., 2013). Despite these differences in biomass losses and microbial biomass, total microbial biomass was correlated only to biomass disappearance in single species litter and did not explain differences in decomposition of mixed species litter. The authors indicate that diversity, rather than biomass of decompos ers per se was stimulated in mixed species litter and was responsible for the differences observed. As legume proportion increased in the current study there was also an increase in relative decomposition rate of remaining N (Figure 4 3), causing remain ing N at the end of incubation to be negatively related to legume proportion (Figure 4 4 ). In 2015, remaining N decreased faster

PAGE 84

84 (represented by the more negative k values, Figure 4 3) than in 2016 (averaged 0.0031 and 0.0023 g g 1 d 1 respectively). Th e pure bahiagrass treatments had similar remaining N relative decay rates to each other in 2015 ( P = 0.23, k = 0.0031 and k = 0.0020 g g 1 d 1 for BG and BGN, respectively) and 2016 (P = 0.32, k = 0.0013 and k = 0.0023 g g 1 d 1 for BG and BGN, respecti vely) (Figure 4 3). At the end of incubation, remaining N was similar for BG and BGN in 2015 ( P = 0.44; 75 and 82%, respectively), but in 2016 there was less remaining N for BGN compared with BG ( P = 0.04; 70 and 91%, respectively), probably due to greater initial N concentration in BGN that year. Greater decay of remaining N in 2015 compared with 2016 likely occurred due to the behavior we observed in remaining biomass, where lower temperatures and longer periods with no rainfall caused decomposition rates to be less in 2016 compared with 2015. Similar to what was observed in biomass losses, N concentration probably played a large role in determining behavior of remaining N. Other work supports this premise. Silva et al. (2012) found that signalgrass calopo litter mixtures had lesser remaining N in years when biomass decomposition was greater. Evaluating signalgrass decomposition under 0, 150, and 300 kg N ha 1 Apolinrio et al. (2014) found a linear increase in rate and extent of remaining N decomposition as fertilization increased. Considering that the initial N concentration was greater at higher legume proportions in litter, this indicates more N is cycled as legume proportion increases. Another indication of the relationship between decomposition dynami cs and weather conditions is that lignin concentration reached a linear plateau earlier in 2016 compared with 2 015 (69 vs. 91 d) Thus increasing lignin concentration, which is naturally occurring during microbial decomposition ended earlier during an incubation period when temperature and moisture were less favorable to decomposition. Similarly, lignin:N ratio stabiliz ed in almost half the time in 2016 (40 d) as 2015 (78 d).

PAGE 85

85 The potential litter N contribution to this system can be estimated based on litter deposition rate and N concentration at beginning of incubation. Considering a grazing season of 150 d, litter from pure BG treatments contributed ~25 and 26 kg N ha 1 in 2015 and 2016, while litter from legume containing treatments contribu ted ~32 and 56 kg N ha 1 These values are below those reported by Dubeux et al. (2006b) of 57 kg N ha 1 for bahiagrass fertilized with 40 kg N ha 1 under continuous stocking of 1.3 AU ha 1 However, in their study the constant presence of cattle is likely to have increased litter deposition rate because of trampling, increasing total N returned through plant biomass relative to this study. In addition, Dubeux et al. (2006b) performed their trial during the summer, when forage production and therefore litte r deposition w ere likely greater than in our experiment in which litter deposition data was collected in the fall. C hange in L itter Composition During I ncubation Concentration of N increased and litter C:N ratio decreased during incubation. Legume propo rtion explained 90 and 85% ( P = 0.0032 and 0.0083 in 2015 and 2016, respectively) of the variation in litter N concentration and 83 and 35% ( P = 0.0109 and 0.2150 in 2015 and 2016, respectively) of the variation in litter C:N (Figure 4 2). Treatments BG an d BGN did not differ in N concentration at end of 128 d of incubation in 2015 or 2016 ( P = 0.33 and 0.85; 21 and 19 g N kg 1 OM, respectively). Ratio of C:N was also similar between BG and BGN in 2015 ( P = 0.298; average C:N of 24) and 2016 ( P = 0.3913; av erage C:N of 25). Nitrogen concentration was greater in litter of legume grass mixtures compared with pure bahiagrass treatments ( P = 0.0013 and 0.0001 in 2015 and 2016, respectively), and consequently average legume grass litter C:N ratio was lower than i n BG and BGN treatments ( P = 0.0001 and 0.0013 in 2015 and 2016, respectively).

PAGE 86

86 The chemical form of N changed during incubation. For the legume grass mixtures, at the beginning of incubation the concentration of acid detergent insoluble N (ADIN) in tota l N averaged 68 and 88 g kg 1 in 2015 and 2016, respectively. By Day 128 of incubation, it increased to 418 and 254 g kg 1 in 2015 and 2016, respectively. Averaged across the pure bahiagrass treatments, ADIN in total N increased from 64 and 116 g kg 1 at i nitiation of incubation to 469 and 352 g kg 1 after 128 d in 2015 and 2016, respectively. At the end of incubation, g ADIN kg 1 N was similar for legume grass mixtures and pure bahiagrass treatments in 2015, but in 2016 litter from bahiagrass was greater ( P = 0.0021) meaning more N was immobilized. Carbohydrates and lignin phenolic groups and quinones released during decomposition can react with NH 3 decreasing its availability to microorganisms (Berg and McClaugherty, 2008). This change in N form was obser ved by Dubeux et al. (2006a), who found that bahiagrass after 128 d of incubation had almost half of its N unavailable for decomposition. They also found that the reduction in N availability was greater for bahiagrass pastures receiving more N fertilizatio n (360 kg N ha 1 yr 1 ) and grazed at greater stocking rates than for those managed less intensively. However, Silva et al. (2012) found that ADIN concentration was greater in litter from a signalgrass monoculture than a 50 50 mixture and 100% calopo at the end of incubation (675 for pure signalgrass and 534 g ADIN kg N 1 for signalgrass calopo and calopo litter). The C pool also changed during decomposition. Lignin concentration in mixtures increased during the incubation period from 34 to 150 and 36 to 95 g kg 1 OM in 2015 and 2016, respectively. Consequently, averaged across mixed litter treatments, lignin:N ratio increased from beginning to end of incubation, from 1.5 to 4.8 in 2015 and 1.3 to 2.8 in 2016 (Figure 4 2). Increase in lignin concentration du ring decomposition occurs through two mechanisms, the disappearance of soluble components and accumulation of recalcitrant material, and synthesis of

PAGE 87

87 new lignin like molecules particularly by fungi (Berg and Matzner, 1997; Gijsman et al., 1997; Berg and Mc Claugherty, 2008). Lignin:N ratio at the end of incubation was unaffected by legume proportion in 2015, but in 2016 it decreased linearly as proportion of legume increased ( P = 0.0086; Figure 4 2). Carreiro et al. (2000) reported a reduction in lignin degr ading enzymes with the addition of N. Dubeux et al. (2006a) found similar behavior underlying lignin accumulation during bahiagrass decomposition, in which litter from increasing levels of N fertilizer application had greater lignin concentration. The auth ors suggested that this is a product of more rapid decomposition of soluble C in high N fertilization treatments, which reduces soluble C in detriment to recalcitrant material, and also to the inhibitory effect N has on lignin decomposition. This mechanism likely played a role in our study, since no correlation between lignin and legume proportion occurred at the beginning of incubation ( P > 0.37), but after 128 d this relationship was positive in 2015 (P = 0.0037) but there was no relationship in 2016 ( P = 0.31). Lack of correlation in 2016 could be due to the slower decomposition rates in that year. In mixtures, Rovira and Vallejo (2002) evaluated change in recalcitrant C and N pools of Eucalyptus globulus Quercus ilex Pinus halepensis dead material and whole plant fresh material of Medicago sativa (all materials were dried and ground). They were incubated over 2 yr in mixtures with red earth. They found that C recalcitrance index (calculated as proportion of unhydrolyzed C relative to total organic C), i n which lignin is the main component, increased to a greater degree in plant materials that were more labile ( M. sativa ), while N recalcitrance index (estimated similarly to that of C) consistently increased in all plant materials. The authors attributed t his change to lower mineralization of recalcitrant N as well as significant incorporation of N to the unhydrolyzable pool. Additionally, Rovira and Vallejo (2002) found that the change in N recalcitrance index was strongly related to rainfall, but only wea k

PAGE 88

88 relationships were found between this index and temperature. These observations are similar to results in our study. Concentration of ADIN increased by six to seven times for mixtures and pure bahiagrass treatments in 2015 (68 to 418 and 64 to 469 g kg 1 OM, respectively) when rain events were more frequent at the beginning of incubation (Figure 4 1), but in 2016 ADIN increase was on the order of three times for both plant material types (88 to 254 and 116 to 352 g kg 1 OM for mixtures and pure bahiagrass treatments, respectively). In contrast, lignin concentration increased to a lesser degree relative to ADIN, increasing four and three times in mixtures and pure bahiagrass in 2015 (34 to 150 and 22 to 119 g kg 1 OM, respectively) and three and two times i n 2016 (36 to 95 and 31 to 60 g kg 1 OM, respectively). Hence, we observed a greater increase of ADIN (insoluble N) compared with that of lignin (insoluble C) after 128 d of incubation, and although rainfall seems to have affected both, it affected insolub le N concentration to a greater degree. Existing L itter Mass and Litter Deposition R ate Legume proportion did not explain a significant proportion of the variation in litter deposition rate ( P = 0.3818 and 0.1592 in 2015 and 2016, respectively; R 2 = 0.19 in both years; Figure 4 5). Since deposition rate was generally similar across treatments and existing litter is a function of deposition rate and litter decomposition rate (Rezende et al., 1999; Dubeux et al., 2006b), differences observed in exist ing litter can be attributed primarily to differences in decomposition rate. In August of 2015, existing litter had a very weak relationship with increasing legume proportion. However, as the season advanced, the greater decomposition rate that was observe d as legume proportion increased affected existing litter, resulting in a strong, negative correlation of pasture legume proportion and existing litter (Figure 4 5). This response indicates that at lesser legume proportions, litter decomposition was not ra pid enough to overcome greater deposition of litter, and existing litter increased at the end of the grazing

PAGE 89

89 season. These differences in existing litter carried over to August of 2016. In 2016, relative decomposition rate was less than in 2015 (a less neg ative k ) and not as strongly affected by legume proportion, which is observable by the lesser slope of the line relating legume proportion and relative decomposition rate (Figure 4 3). As a result, existing litter increased by October resulting in an overa ll greater mean existing litter relative to previous litter collection (3010 kg ha 1 in October compared with 2770 kg ha 1 in August of 2016), and was not strongly affected by legume proportion in the pasture (Figure 4 5). The BG and BGN treatments had li tter deposition rate s of 11 and 9 kg DM ha 1 d 1 in 2015 and 2016, respectively, compared with 9 and 13 kg DM ha 1 d 1 for mixtures containing rhizoma peanut (Figure 4 5). Pure bahiagrass treatments averaged greater existing litter relative to legume grass mixtures only in August 2015 (2770 and 1750 kg ha 1 respectively; P = 0.0031) and did not differ from each other in any collection time. This greater litter accumulation in treatments that do not contain legume, and in 2016 relative to 2015, are probably due to slower decomposition rates observed in pure BG and BGN treatments overall in 2016 compared with 2015. Dubeux et al. (2006a ; b ) evaluated bahiagrass litter deposition and accumulation under increasing management intensities from Low to High. They f ound that High management levels increased litter deposition and decomposition (Dubeux et al., 2006a; b). However, the greater decomposition rate was not enough to balance deposition rate, causing existing litter in bahiagrass pastures under the High manag ement intensity to increase over the grazing season, similar to what we observed in 2015 at lesser legume proportion in the pasture. Apolinrio et al. (2013) evaluated litter dynamics of signalgrass pastures managed under different stocking rates (2.0, 3. 9, and 5.8 AU ha 1 [1 AU = 450 kg]) and N fertilization rates (0, 150, 300 kg N ha 1 yr 1 ). The authors observed a carryover effect of grazing management on existing litter in the second

PAGE 90

90 year the experiment was conducted, due to cumulative effect of N fert ilizer leading to greater plant production. This was similar to what we observed in existing litter from October 2015 to August 2016 in our study. Existing and Deposited Litter C omposition Concentration of N was similar in deposited litter of mixtures rel ative to bahiagrass treatments in 2015 ( P = 0.8744, 19 g kg 1 OM), but greater in 2016 ( P = 0.0129, 24 and 19 g N kg 1 OM, respectively). In existing litter, N concentration was less in bahiagrass when compared with legume grass treatments in 2015 ( P = 0.0148; 22 and 19 g N kg 1 OM, respectively) and in 2016 ( P = 0.0007l; 26 and 20 g N kg 1 OM, respectively). These values are in the range described by Dubeux et al. (2006 a ) for bahiagrass fertilized with 40 to 360 kg N ha 1 (13 to 22 and 14 to 23 g N k g 1 OM for deposited and existing litter, respectively). Insoluble N fraction in total N was similar across treatments ( P > 0.1211) and ranged between 492 and 623 g ADIN kg 1 N for deposited and 468 and 575 g ADIN kg 1 N for existing litter. These values c orroborate our observation in the decomposition study, where decomposed materials have a large proportion of their N in an insoluble form. Conclusions Biomass losses occurred more rapidly and to a greater extent as legume proportion increased in plant lit ter; these were generally greater than those for unfertilized and fertilized bahiagrass. Losses of remaining N increased as legume proportion increased in 2015, when decomposition was more rapid than in 2016. Greater decay rate and extent of loss of biomas s and remaining N with increasing legume proportion probably occurred because of increasing concentration of N, however greater microbial diversity has been proposed by others as a mechanism for increased decomposition rates in mixed species litter compare d with monocultures. The effect of legume proportion on decomposition dynamics were more

PAGE 91

91 pronounced in 2015 relative to 2016, which was likely due to environmental conditions during incubation, including greater temperature and more frequent rainfall event s in 2015. Nature of C and N in plant litter changed over the period of decomposition, with increasing concentration of insoluble materials, likely due to accumulation and formation of new recalcitrant material as decomposition occurred. Because no differe nces among treatments were observed in litter deposition rate, greater litter relative decay rate was responsible for the lesser existing litter mass as legume proportion increased, particularly in 2015. Greater N concentration and litter decay rates with increasing legume proportion accompanied by similar litter deposition rates across treatments indicate that legume grass mixtures are a useful alternative to N fertilizer for increasing N cycling through plant litter in grasslands and that increasing legum e contribution is likely to be associated with greater levels of N release from plant litter.

PAGE 92

92 Figure 4 1. Rainfall and temperature data from the Alachua ( FL ) site of the Florida Automated Weather Network (FAWN). A) Daily rainfall and minimum, average, and maximum temperatures for the f irst 32 d of incubation in 2015. B) Daily rainfall and minimum, average, and maximum temperatures for the f irst 32 d of incubation in 2016. C) M onthly average temperature and accumulated rain fall annual data for the years of evaluation and the 30 yr average. A) B ) C )

PAGE 93

93 Figure 4 2 Nitrogen concentration, C:N and lignin:N ratio, and acid detergent insoluble N concentration in total N for plant litter at beginning (Day 0) and end (Day 128) of incubation during 2 yr in Gainesville, FL. Treatments are: unfertilized bahiagrass (BG), bahiagrass receiving 5 0 kg N ha 1 (BGN), and bahiagrass rhizoma peanut mixtures. Lines represent estimated linear correlation and points represent actual measured values. P Mixt ures = P value for differences among bahiagrass rhizoma peanut mixtures, P Grass x Mixtures = P value for comparison of the average of the two grass treatments vs. the average of the mixtures, and P BG x BGN = P value for comparison of the two grass treatments.

PAGE 94

94 Figure 4 3 Plant litter biomass and N relative decay rates ( k ) of plant litter incubated for 128 d during 2 yr in Gainesville, FL. Treatments are: unfertilized bahiagr ass (BG), bahiagrass receiving 5 0 k g N ha 1 (BGN), and bahiagrass rhizoma peanut mixtures. Lines represent estimated linear correlation and points represent actual measured values. P Mixtures = P value for differences among bahiagrass rhizoma peanut mixtures, P Grass x Mixtures = P value fo r comparison of the average of the two grass treatments vs. the average of the mixtures, and P BG x BGN = P value for comparison of the two grass treatments.

PAGE 95

95 Figure 4 4 Plant litter biomass and remaining N after 128 d of incubation over 2 yr in Gainesville, FL Treatments are: unfertilized bahiagr ass (BG), bahiagrass receiving 5 0 kg N ha 1 (BGN), and bahiagrass rhizoma peanut mixtures. Lines represent estimated linear correlation and points represent actual measured values. P Mixtures = P value for differences among bahiagrass rhizoma peanut mixtures, P Grass x Mixtures = P value for comparison of the average of the two grass treatments vs. the average of the mixtures, and P BG x BGN = P value for comparison of the two grass treatments.

PAGE 96

96 Figure 4 5 Litter deposition rate and existing litter mass in August and October of 2 yr in Gainesville, FL. Treatments are: unfertilized bahiagrass (BG), bahiagrass receiving 5 0 kg N ha 1 (BGN), and bahiagrass rhizoma peanut mixtures. P Mixtures = P valu e for differences among bahiagrass rhizoma peanut mixtures, P Grass x Mixtures = P value for comparison of the average of the two grass treatments vs. the average of the mixtures, and P BG x BGN = P value for comparison of the two grass treatments.

PAGE 97

97 CHAPTER 5 NITROUS OXIDE AND METHANE EMISSIONS FROM CATTLE URINE AND DUNG IN N FERTILIZED GRASS AND LEGUME GRASS SWARDS Introduction Livestock production systems are an important source of greenhouse gases (GHG). Emissions of GHG include methane (CH 4 ) from enteric fermentation (Moss et al., 2000; Lassey, 2007) and dung (Jarvis et al., 1995; Yamulki et al., 1999; Mori and Hojito, 2015), nitrous oxide (N 2 O) from dung, urine (Yamulki et al., 1998; Sordi et al., 2013; Lessa et al., 2014), and N fertilizer (de Kl ein et al., 2006), and carbon dioxide (CO 2 ) from fossil fuels utilized in management practices (McSwiney et al., 2010). These GHG have different capacities to trap energy and contribute to the greenhouse gas effect, quantified relative to CO 2 and referred to as global warming potential (GWP) (IPCC, 2014). Global warming potential of N 2 O is 265, which is 9.5 times greater than that of CH 4 (28) (IPCC, 2014). Therefore, changes in emissions from any source in livestock production due to management strategies c an impact their total GHG budgets significantly. Methane emissions from cattle urine are negligible, but dung can significantly contribute to emissions in livestock production systems (Jarvis et al., 1995; Yamulki et al., 1999; Cai et al., 2017). Methane e missions from dung can be affected by animal diet. Jarvis et al. (1995) evaluated CH 4 emissions from dung of animals fed fertilized grass or grass clover ( Trifolium sp.) mixed pastures. They found CH 4 emissions from dung were greater for animals grazing g rass legume mixtures, probably because of lower C:N ratio. Being a strictly anoxic process (Yamulki et al., 1999), production of CH 4 from dung can also be affected by percentage soil water filled pore space and rain events. For example, Mori and Hojito (20 15) found that CH 4 emissions increased after rain events as long as fluxes had not subsided to background levels.

PAGE 98

98 Formation of N 2 O occurs through nitrification, nitrifier denitrification, and denitrification (Dijkstra et al., 2013). Organic N in dung and u rine goes through hydrolysis and mineralization forming ammonium (NH 4 + ), which is used by nitrifying microorganisms producing nitrite (NO 2 ) and nitrate (NO 3 ). Denitrification then follows, forming molecular nitrogen (N 2 ) and N 2 O (Oenema et al., 1997; van der Weerden et al., 2011; Dijkstra et al., 2013). Alternatively, nitrifier denitrification occurs when NO 2 is reduced to nitric oxide (NO), N 2 O, and N 2 (Wrage et al., 2001). Denitrification is often the predominant N 2 O emission process in the first days after urine is deposited on the soil surface because of high soil water content, but this shifts towards nitrification as the predominant process (de Klein and van Logtestijn, 1994). Emissions of N 2 O from animal excreta are positively related to their N co ncentration (Oenema et al., 1997), and some authors suggest diet management strategies that reduce N output in animal excreta are a viable alternative to reduce N 2 O emissions from livestock (Dijkstra et al., 2011, 2013). Livestock production systems in so utheastern U.S. rely mostly on warm season perennial grasses (Vendramini et al., 2010), with N provided by chemical fertilizers. Direct N 2 O emissions from N fertilizer are estimated to be 0.01 kg N 2 O N kg 1 N input (de Klein et al., 2006). Emissions from p roduction, storage, and application of fertilizer also significantly contribute to GHG budgets (Lal, 2004). Alternatively, N from biological fixation in legumes is GHG neutral (Jensen et al., 2012), while also potentially reducing CH 4 emissions from enteri c fermentation (Archimde et al., 2011). This makes legume adoption an attractive alternative to improve sustainability of grass based livestock production. In southeastern USA, efforts have been made in developing new cultivars of rhizoma peanut ( Arachis glabrata Benth.) and promoting successful management strategies for their use in mixed swards with grass species such as bahiagrass ( Paspalum notatum Flgge) (Castillo et al., 2013a; b; Mullenix et al., 2016a; b).

PAGE 99

99 However, little is known about how this change in animal diet affects N 2 O emissions from animal excreta. Different models can be used to estimate GHG budgets. At national and regional levels, the methodology described by the Intergovernmental Panel on Climate Change (IPCC, de Klein et al., 2006 ) is often used, and indicates a direct relationship between N concentration of excreta and an emission factor representing proportion of N input emitted as N 2 O N. In this methodology, urine and dung N 2 O emissions are accounted for together, but evidence f rom field experiments shows that emissions from urine are often significantly greater than that of dung (van der Weerden et al., 2011; Sordi et al., 2013; Lessa et al., 2014; Mazzetto et al., 2014; Mori and Hojito, 2015). For CH 4 from animal manure, N conc entration is not taken into account in the emission factor despite experimental evidence that increases in N concentration can raise CH 4 emissions (Jarvis et al., 1995). The objective of this study was to quantify emissions and obtain emission factors of N 2 O from dung and urine and CH 4 from dung of animals grazing rhizoma peanut bahiagrass mixed pastures relative to N fertilized bahiagrass monocultures. Material and Methods Experimental S ite A 2 yr experiment was conducted at the University of Florida Beef Research Unit (BRU) in Gainesville, FL (29.72 N, 82.35W) during the summers of 2015 and 2016. Soils at the site are Sparr fine sand (loamy, siliceous, hyperthermic Grossarenic Paleudult). Soil samples were taken in May 2014 to a 15 cm depth and analyzed at the University of Florida Extension Soil Testing Laboratory. Soil pH ranged from 5.9 to 6.2 and Melich 3 extractable P, K, and Mg ranged from 112 to 138, 52 to 78, and 83 to 122 mg kg 1 ; all nutrient concentrations were classified as high, and no lime o r fertilizer, other than N for the grass monoculture, was applied. Climate data and weather data from the years of experiment are presented in Figure 5 1.

PAGE 100

100 zoma peanut bahiagrass pastures. Pastures were 20 x 48 m in area, and there were three replicates of each pasture type. Proportion of rhizoma peanut was 45 to 67% in 2015 and 73 to 78% in 2016. Bahiagrass pastures were fertilized with 50 kg N ha 1 in late spring/early summer of both 2015 and 2016. Immediately before grazing, values of crude protein in pastures averaged 116 and 126 g kg 1 for BGN and RP BG, respectively. Treatments and Experimental D esign For measures of GHG, treatments consisted of two fo rage systems (N fertilized grass and grass legume) and two animal excreta types arranged as a factorial experiment in four replicates of a completely randomized design. A static chamber (described below) was the experimental unit for GHG measures, and ther e were 16 chambers to accommodate the four replicates of four treatments. Two additional chambers were positioned in areas where no excreta was applied and were utilized as blanks to account for background emissions totaling 18 chambers. Chambers were dis tributed in three parallel rows of six chambers each, with chamber centers positioned 1 m apart from each other Static Chamber Design and I nstallation Gas emissions were collected from static chambers. The chambers were circular, with a radius of 30 cm (0.283 m 2 ) for urine and 15 cm (0.071 m 2 ) for dung to mimic typical area occupied by each excreta type (Haynes and Williams, 1993; White Leech et al., 2013 a ). Chamber bases and tops were made with polyvinyl chloride (PVC), and the tops were lined with an acrylic sheet to avoid any reactions of gases of interest with chamber material (Parkin and Venterea, 2010). The chamber tops were covered with reflective ta pe to provide insulation, and they were equipped with a rubber septum for sampling (Clough et al., 2012). The top was fitted with a 6 mm diameter, 10 cm length copper venting tube to ensure adequate air pressure inside

PAGE 101

101 the chamber during measurements, cons idering an average wind speed of 1.7 m s 1 (Hutchinson and Mosier, 1981; Hutchinson and Livingston, 2001). During measurements, chamber tops and bases were kept sealed by fitting bicycle tire inner tubes tightly over the area separating top and base. Base s of chambers were installed in an area excluded from grazing in a rhizoma peanut bahiagrass pasture 2 wk prior to excreta application in order to avoid any effect of soil disturbance on emissions (Rochette et al., 2012). Bases were installed to an 8 cm de pth, with 5 cm extending above ground level. Depth for installation was determined based on Clough et al. (2012), who indicated the need of > 12 cm depth h 1 of deployment considering a sampling (deployment) time of 40 min. Chamber tops were 22 cm in heigh t, which when summed with 5 cm of base extending above ground totaled 27 cm, in agreement with the indication of > 40 cm of chamber height h 1 of deployment (Clough et al., 2012). Bases were removed between the first and second year of the experiment and r einstalled at a different location in the same pasture to avoid any carryover effect of first year excreta application on second year GHG emissions. Excreta Collection and Chemical C omposition Prior to excreta collection, animals were trained to come from the treatment pastures to a corral area and stand in a stall. Two animals (Brahman x Angus crossbred yearling heifers with liveweight of ~375 kg) grazed each excreta source pasture for 1 wk before excreta collection. Excreta collection occurred once each year in 2015 and 2016. To obtain quantities of excreta required, collection occurred during 1 d for each pasture source and lasted no longer than 4 h, during which animals were shaded and had access to water. Dung and urine were collected from at least one animal per plot in each year. While animals were standing in the stall, collections were made by a person standing behind each animal using a plastic trash bag inserted into a fishing net mounted on a circular metal frame at the end of a wooden handle (Wh ite Leech et al.,

PAGE 102

102 2013 a ). After each urine or dung event, the sample was secured and the bag was replaced to avoid sample contamination. Dung and urine samples were placed in sealed plastic containers and maintained under 4C until application for a maximu m of 48 h (Yamulki et al., 1998; van der Weerden et al., 2011; White Leech et al., 2013 a ). Samples of each excreta type were composited across all animals from the three replicates of a given source pasture to obtain one sample of each excreta type per sou rce pasture treatment (White Leech et al., 2013 a ), resulting in four samples (urine and dung from each of two source pasture treatments). Three subsamples of urine and of dung per source treatment were collected at time of excreta application for chemical composition analysis. Urine samples were acidified to pH 2 using H 2 SO 4 to avoid N volatilization (White Leech et al., 2013 a ). Urine was analyzed for total N, NH 3 N, CO(NH 2 ) 2 N, P, and K, and dung was analyzed for total N, NH 3 N, P, and K. Total N in urine and dung were analyzed by a TruMac N Macro Determinator (Leco Corporation). Ammonia N in urine and dung were analyzed using a Timberline TL 2800 Analyzer (Timberline Instruments). Urea N in urine was analyzed with a Timberline TL 2800 Analyzer (Timberline Instruments) after urease solution application. Organic N in dung was obtained by calculating the difference between total N and ammonia N. Urine and dung P and K were analyzed by Inductively Coupled Plasma (ICP) using a Thermo iCAP 6300 Inductively Coupl ed Plasma Radial Spectrometer (Thermo Fisher Scientific Inc.). Urine and dung chemical composition are presented in Table 5 1. Excreta A pplication Urine and dung were deposited on soil surface inside the area determined by the base of the static chamber o nce per season (van der Weerden et al., 2011; Sordi et al., 2013). Urine was applied at a rate of 2 L and dung at 2 kg fresh weight per chamber (Haynes and Williams, 1993). Urine was applied following the methodology described in White Leech et al. (2013 a ) Briefly, 1

PAGE 103

103 L was applied in side the area de lineated by a 15 cm radius circle at the center of the 30 cm radius chamber base, and 1 L was applied in the area between the margin of the inner 15 and the outer 30 cm border of the chamber base This resulted in 2 .0 and 2.2 g of N applied per chamber (0.007 and 0.008 kg N m 2 respectively) in BGN and RP BG urine treatments respectively, and 3.6 and 5.8 g of N applied per chamber (0.051 and 0.082 kg N m 2 respectively) from dung f or BGN and RP BG treatments respectively Within a source treatment, dung was thoroughly mixed and weighed. The sample was applied uniformly across the entire chamber diameter. Excreta collection occurred on 3 and 4 Aug 2015 and 5 and 6 July in 2016. Excreta applicat ion occurred on 5 Aug 2015 and 7 July 2016. Gas Sampling and A nalysis Sampling occurred between 0900 and 1100 h, the period when temperature is closest to the daily average (Parkin and Venterea, 2010). Sampling was more frequent soon after excreta applic ation and decreased thereafter (Table 5 2). In the first year of the experiment, gas collections occurred until 230 d after treatment application. Results showed negligible emissions of N 2 O and CH 4 after one month, leading us to reduce length of the sampli ng period in the second year of the experiment. Data are presented only on results from samples collected in the 30 d following application in both years. Four subsamples were taken per deployment time per chamber, separated by 12 min intervals (t 0 t 12 t 24 and t 36 ). At t 0 the sample was collected from the area directly above soil surface (Chadwick et al., 2014) with the use of a 60 mL syringe and immediately flushed into a 30 mL glass vial equipped with a butyl rubber stopper sealed with an aluminum septa. After t 0 sample collection, chambers were closed by fitting top to base and sliding a bicycle inner tube over their juncture. Time zero samples were collected from six (two blanks and one of each excreta type and pasture source combination) of the 1 8 chambers and used to calculate gas fluxes in the other chambers (Ambus and Christensen, 1994). At times t 12

PAGE 104

104 t 24 and t 36 the syringe was inserted through the rubber septum on the top of the chamber and flushed inside the chamber two to three times befor e sampling to mix the air inside, after which a 60 mL air sample was collected and transferred into a glass vial similar to the procedure followed for t 0 sample collection. Vials were stored until analysis. Gas sample analysis was conducted using a gas chr omatographer (Agilent 7820A GC, Agilent Technologies, Palo Alto, CA) equipped with a flame ionization detector and a capillary column (Plot Fused Silica 25 m 0.32 mm, coating Molsieve 5A, Varian CP7536). Gas Flux C alculation Flux of N 2 O N and CH 4 C (mg m 2 h 1 ) was calculated as described in Equation 5 1: F = A dC/dt (5 1) w here F is flux of N 2 O ( m g m 2 h 1 ), A is the area of the chamber and dC/dt is the change in N 2 O concentration in time calculated using a linear method of integration, which is bette r able to detect differences between chambers and is indicated for use when the experimental objective is to detect emission differences among treatments (Venterea et al., 2009). Flux values were corrected for background emissions. Cumulative emissions wer e obtained by using trapezoidal integration of flux over time (Venterea et al., 2012). Emission factors of N 2 O were calculated for each treatment (type of excreta and typ e of pasture source) using E quation 5 2 (Sordi et al., 2013): EF (%) = [(N 2 O Nemited ) (N2O N control )] / N a pplied x 100 (5 2) w here EF (%) is the percentage of applied N that is emitted as N 2 O N, N 2 O N emitted is the cumulative g N 2 O N emitted from treatments in the year, N 2 O Ncontrol is the cumulative g N 2 O N emission from the control in the year, and N applied is the amount of N applied from treatments (g yr 1 ).

PAGE 105

105 Soil M easurements The excreta treatments were applied at the treatment rates in an adjacent area of the rhizoma peanut bahiagrass mixed pasture used for chamber placement. The are a was divided into four, 0.5 x 0.5 m plots in order to collect percentage water filled pore space (WFPS) data at each gas sampling time. Soil samples were collected to 15 cm, weighed immediately after collection, dried at 105C for 24 h, and weighed again to obtain volumetric water content. Percentage WFPS was obtained by dividing water volumetric content by soil porosity and multiplying the value by 100 (USDA, 1998). Soil porosity was estimated based on soil bulk density. Four samples were collected at th e experimental site to measure bulk density. Statistical A nalysis Data were analyzed using the R software (R Core Team, 2016). A mixed model was used CH 4 because only emissions from dung were quantified; and pasture source times excreta type for N 2 O) were considered fixed and year was considered random. Least squares means were cant main effects ( P < 0.05). In flux evaluation, a correlation matrix was included in the model to account for change in sampling frequency. Results and Discussion Methane E missions There was an interaction of day and source pasture on CH 4 flux ( P = 0.0 084) (Figure 5 2). Emissions were greater from dung of animals grazing RP BG on Days 2, 3, and 4 when compared with those grazing BGN pastures. In dung from animals grazing RP BG pastures, emissions were greater in the first 3 d compared with the period st arting at 6 d after treatment application ( P < 0.0078), with intermediate emissions at Day 4 and 5 (Figure 5 3) ( P > 0.0525).

PAGE 106

106 When dung pasture source was BGN, there was no day effect on CH 4 flux during the sampling period ( P > 0.1089). Peak CH 4 emissions occurred 2 d after treatment application for both treatments (7.3 and 10.9 mg m 2 h 1 for BGN and RP BG pasture sources, respectively) and coincided with a peak in WFPS (Figure 5 3), but they decreased to background emissions after 8 d. As flux of CH 4 emis sions was greater from dung of animals grazing on RP BG pastures, cumulative emissions were also greater in this treatment relative to BGN ( P = 0.0034; 39 and 69 mg CH 4 C [dung pile] 1 ) ( Figure 5 4 ). The pattern we observed in gas flux of CH 4 was reported by others in the literature. Emission of CH 4 from soil typically increases after dung deposition with a peak occurring 1 to 3 d after dung application (Jarvis et al., 1995; Yamulki et al., 1999; Mazzetto et al., 2014; Mori and Hojito, 2015), while soils not amended with dung have non measureable CH 4 emissions (Jarvis et al., 1995). In an extensive review, Cai et al. (2017) found that on average dung deposition increased CH 4 emissions from soil by 85 g CH 4 C (dung patch) 1 relative to control s oil receiving no dung treatment deposition. Yamulki et al. (1999) found that emissions of CH 4 increased in 100 times after dung application relative to background measures. These peaks have been associated with ideal conditions for methanogenesis at time o f excreta deposition, namely presence of enteric microbiota (Yamulki et al., 1999; Mori and Hojito, 2015), temperature (Jarvis et al., 1995), and moisture (Yamulki et al., 1999) of dung piles. Yamulki et al. (1999) found that 80% of emissions occurred by 1 wk after application. Similarly, our study showed that by 5 d after treatment application between 80 and 90% of cumulative CH 4 emissions had occurred for BGN and RP BG treatments, and both treatments reached 99.8% of emissions recorded in the first month by 1 wk of treatment application. These results suggest that shorter measurement

PAGE 107

107 periods than the one utilized in this experiment may suffice to obtain site and diet specific dung CH 4 emission factors under our conditions (season, soil type, etc.) Comparing emissions of CH 4 from soils receiving dung of cattle to those of dung or soil alone, Jarvis et al. (1995) found that 69% of CH 4 came from the dung patch itself, while soils receiving no dung application had no CH 4 emissions. Jarvis et al. (1995) suggest that the emissions sourced from soil in the presence of dung were likely occurring because existing methanogenic soil microbiota were stimulated by C and moisture provided by dung associated with the formation of an anaerobic environment under the dung pile. In agreement with this premise, percentage WFPS was greater under dung relative to urine in several occasions in our study, particularly soon after treatment application (Day 2) coinciding with the peak in CH 4 emissions (Figure 5 3). Jarvis at al. (1995) found that emissions of CH 4 ceased after dung was treated with chloroform to eliminate the innate microbial population, indicating that the dung microbial population was primarily responsible for emissions (Jarvis et al., 1995). This agrees with our study in which background emissions were small (fluxes averaged 0.24 mg m 2 h 1 and did not surpass 2.9 mg m 2 h 1 ), with cumulative emissions of 0.007 g chamber 1 This relationship, however, might not be true in all environments. Chamberlain et al. (2015) performed year round CH 4 measurements from lowland pastures, where bahiagrass was planted in an area comprised by poorly drained Spodosols with depressional wetlands. Grazing occurred only during the dry season (November April) with a stocking rate of 1.6 cow ha 1 while during the wet season (May October), when water table approached the soil surface and sometimes inundated the soils, animals were removed from the pasture. They found that during the dry season presence of animals explained peaks in emissions of CH 4 However, during the wet season when animals

PAGE 108

108 were removed from the area, soil CH 4 emissions were comparable to those of dry season with cattle. Although these emissions could be related to previous C input to the environment through animal excreta, it is not possible to directly allocate CH 4 production during the wet season to presence of animals. Chamberlain et al. (2015) concluded that in lowland grazed pastures with seasonal flooding, emissions from the landscape (soil, ditches, canals, etc.) are an important source of CH 4 whereas cattle are a minor contributor to annual GHG budgets. Cumulative emissions of methane were 39 and 69 mg CH 4 C [dung pile] 1 for animals grazing BGN and RP BG, respectively. If we consider an average number of 17 defecations per day (Aland et al., 2002) for both treatments and that emissions were similar through the entire year we estimate annual emissions of 241 and 422 g CH 4 C animal 1 yr 1 for animals grazing BGN and RP BG, respectively. Therefore, CH 4 emissio ns from animals grazing legume rich pastures was 1.75 times greater than those grazing bahiagrass fertilized with a typical fertilization regime for extensive production in the southeastern U.S. This difference is likely related to the greater N concentrat ion of dung from animals grazing RP BG compared with BGN (2.9 and 1.8 g kg 1 respectively). Animal diet has been shown to affect CH 4 emissions from dung. Greater cumulative emissions were found for heifers grazing grass clover mixtures compared with N fer tilized grass pastures, with 61 and 22 mg CH 4 C (dung pile) 1 or 376 and 139 g CH 4 C animal 1 yr 1 respectively (Jarvis et al., 1995). Jarvis et al. (1995) argues that N concentration, in particular its influence on C:N ratio, was the main driver of this response with greater emissions observed from excreta with low C:N. Similarly, Mori and Hojito (2015) found that CH 4 emissions from dung tended to increase as dung C:N ratio decreased, although the relationship between C:N ratio and CH 4 was not significant The authors found emissions of 70 and 243 mg CH 4 C (dung patch) 1 when C:N was 36 and 18, respectively. Cumulative CH 4

PAGE 109

109 emissions of dung from animals grazing Brachiaria decumbens in Brazil were 53 and 155 g CH 4 C animal 1 yr 1 in winter and summer in a s ubtropical region and 175 and 442 g CH 4 C animal 1 yr 1 in winter and summer in a tropical region, respectively. These values, except for summer in the tropical region, are considerably lower than those found in this study. This difference might be related to the considerably lower N excreta application rate in Mazzetto et al. (2014) (8 kg m 2 ) compared with our study (28 kg m 2 ). Relative to CH 4 emissions from enteric fermentation, CH 4 emissions from animal dung are small. For example, cows grazing high legume (60% mix alfalfa and white clover, 40% Bromus auleticus Trin. Ex Nees) or grass based (24% birds foot trefoil [ Lotus corniculatus L.] and 76% annual ryegrass) pastures emitted 364 and 372 g animal 1 day 1 respectively (Dini et al., 2012). Henry et al. (2015) found that heifers fed high concentrate diets (85 and 36% concentrate, respectively, where concentrate was composed of corn gluten free pellets, soybean hulls pellets, liquid supplement, and mineral and vitamin supplement) emitted 45 and 130 g CH 4 day 1 respectively. Environmental factors can have both positive and negative effects on CH 4 emissions from cattle dung, but these effects are typically relevant only during the per iod immediately after dung deposition (Yamulki et al., 1999). Increase in temperature can be expected to increase emissions due to positive effect on microbial activity (Jarvis et al., 1995). However, high temperatures can stimulate the formation of a crus t on the dung patch (Jarvis et al., 1995). Although this crust forms an anaerobic environment, which could potentially increase formation of CH 4 it also reduces volatilization and gas exchange patterns with the environment, reportedly reducing dung CH 4 em issions (Jarvis et al., 1995; Yamulki et al., 1999; Nichols et al., 2016). Rain can diminish availability of C sources for the formation of CH 4 (Jarvis et al., 1995). Alternatively, it can

PAGE 110

110 enhance anoxic conditions in animal dung (Jarvis et al., 1995) or p ossibly damage any existing crust on dung patches resulting in release of previously trapped CH 4 (Yamulki et al., 1999; Mori and Hojito, 2015). These different effects of environmental conditions on CH 4 emissions from dung emphasize the relevance of seaso nality on this response. In Brazil, Mazzetto et al. (2014) found that emissions of CH 4 from dung during the summer was 2.5 and 2.9 times greater in two locations when compared with the winter season, attributed to greater temperatures and rainfall during t he summer. Similarly, Mori and Hojito (2015) found emissions of 173 and 48 mg CH 4 C ( dung patch ) 1 in summer and winter in Japan, respectively. Pulses of CH 4 emissions occurred immediately after rainfall events, but rainfall affected emissions of CH 4 only during the time prior to emissions reaching background levels. Yamulki et al. (1999) found negative and positive correlation of CH 4 emissions from dung patches with average daily temperatures and with rainfall, respectively. Negative relationships with tem perature were attributed to the formation of crust on patches, where crust formed more rapidly at higher temperatures and inhibited CH 4 volatilization. Annual emissions estimated at 241 and 422 g CH 4 C animal 1 yr 1 for BGN and RP BG based on our summer me asurements might be an overestimation of real annual emissions. Even so, these values are much lower than those indicated by EPA (2013) and IPCC (2006) for national budget estimations (1150 and 1125 g CH 4 C animal 1 yr 1 respectively). Nitrous Oxide E miss ions There was an interaction of pasture source and day ( P = 0.0049) on N 2 O N flux (Figure 5 5 ). Effect of pasture source was significant only near the end (Days 19 and 21) of the sampling period ( P < 0.0078), with greater values from animals grazing BGN relative to RP BG. However, these differences in fluxes did not translate into differences in cumulative N 2 O N emissions from source pasture, likely because negative fluxes offset positive ones. Flux of N 2 O

PAGE 111

111 N was affected by excreta type ( P = 0.0144), with greater average daily emissions from urine compared with dung (0.26 and 0.10 mg N 2 O N m 2 h 1 respectively). This resulted in different cumulative emissions for the two excreta types ( P = 0.001), with urine having greater cumulative emissions compared with dung (45 vs. 0.8 mg patch 1 Figure 5 6). Considering a N input of 2.1 and 4.7 g N (excreta patch) 1 for dung and urine, respectively, we found EF dung = 0.02% and EF urine = 2.14% N input emitted as N 2 O N. Therefore, N 2 O N emissions from urine were around 120 times greater than from dung in our study. Some authors suggest seasonality can play an important role on N 2 O emissions (Dijkstra et al., 2013), but Oenema et al. (1997) emphasizes that facto rs influencing N 2 O emissions are often site specific. Overall, WFPS has a positive effect on N 2 O emissions particularly when WFPS is in the range of 60 to 80%, where nitrification, nitrifier denitrification, and denitrification are favored (van Groenigen e t al., 2005a; b; Dijkstra et al., 2013). In Brazil, emissions of N 2 O from both cattle urine and dung was nearly zero during the dry season when no rainfall occurred (Lessa et al., 2014). Sordi et al. (2013) found a relationship of N 2 O emissions from urine with rainfall events above 20 mm day 1 but no such relationship was found with emissions from dung or with percentage WFPS for either excreta type likely. The authors argue that the weak relationship between WFPS and N 2 O emissions probably occurred becaus e of losses of water through drainage and evapotranspiration between rain events and sampling for WFPS. In the present study, N 2 O N fluxes had a small peak on Day 2 after application when WFPS was 88 and 52% for urine and dung, respectively (Figure 5 3 and Figure 5 5). However, peaks occurring after Day 2, such as the one on Day 21, did not relate to WFPS. This indicates that other unknown factors were responsible for the emissions occurring in that secondary peak.

PAGE 112

112 Concentration of N in urine can be increas ed by providing high crude protein diets to animals, and the close relationship of dietary crude protein to N 2 O emissions has led some researchers to suggest reducing dietary N as a management practice to reduce emissions (Oenema et al., 1997; Dijkstra et al., 2011; 2013). Avoiding urine and dung deposition during seasons or in areas in which WFPS is high, such as those affected by compaction, are also a strategy for reducing N 2 O emissions (van Groenigen et al., 2005b). The lack of difference in emissions o f N 2 O as related to diet might be due to the small differences found in urine N concentration from animals grazing BGN compared with RP BG (Table 5 1), which was probably related to the similar crude protein concentrations in the two diets (116 and 126 g k g 1 respectively). Concentration of N is considered one of the most relevant factors influencing N 2 O emissions particularly in urine, and availability of N is often used to estimate soil N 2 O emissions on a large scale due to site specific effects of other variables (Oenema et al., 1997). This principle is used in guidelines for emission estimates published by IPCC, which suggest the use of one emission factor of 2% for N 2 O N emissions from both urine and dung combined, multiplied by animal N output (de Kle in et al., 2006). This methodology is also used by the Environmental Protection Agency (EPA) for USA National GHG estimates (EPA, 2016). Although this emission factor of 2% is similar to that of urine found in this study (2.14%), it grossly overestimates t hat of dung (0.02%). Considering total N 2 O N 30 d cumulative emissions of 0.046 g from both excreta types (0.045 and 0.001 g N 2 O N from urine patch es and dung pile s respectively) and excretion of 6.8 g from urine and dung combined (2.1 and 4.7 g N per uri ne patch and dung pile, respectively) the emission factor would be 0.0067 g N 2 O N (g N) 1 or 0.67%, i.e., one third of that indicated by IPCC and used by EPA. Similar to our results, Mori

PAGE 113

113 and Hojito (2015) found emission factors of 0.024 and 0.684%, while van der Weerden et al. (2011) reported 0.04 and 0.29% for dung and urine, respectively, all of which are considerably less than those suggested by the IPCC (de Klein et al., 2006) and markedly different between excreta types. This implies that emissions e stimates from farm and national level GHG budgets overestimate N 2 O from animal excreta deposited on pastures when applied to production systems in the southern U.S. Our results support the work of others suggesting that emissions from urine and dung in gr azing systems should be accounted for separately (Yamulki et al., 1998; van der Weerden et al., 2011; Sordi et al., 2013; Lessa et al., 2014; Mazzetto et al., 2014; Mori and Hojito, 2015). van der Weerden et al. (2011) evaluated emissions of N 2 O from urine and dung from dairy and beef cattle grazing perennial ryegrass (90 95% by weight) white clover (5 10%) mixtures in New Zealand. Emissions from dung were consistently lower than that of urine, resulting in emission factors of 0.04 and 0.29% despite greater N deposition rates for dung (920 kg N ha 1 and 524 kg N ha 1 for dung and urine, respectively). In England, cumulative N 2 O emissions from urine and dung from dairy cows fed grass silage and concentrate diets were similar (9.5 and 9.9 mg N 2 O N [urine patch ] 1 respectively), but when transformed into emission factors as a function of N input, urine N 2 O emissions were three times greater than those of dung (Yamulki et al., 1998). Similarly, emission factors were 0.26 and 0.15% for urine and dung of cattle gr azing Paspalum paniculatum L., Axonopus compressus (Sw.) P. Beauv, and Pennisetum clandestinum Hochst. ex Chiov. in subtropical Brazil (Sordi et al., 2013). Emissions of N 2 O from dairy Friesian x Nelore cows grazing Brachiaria brizantha cv. Marandu supplem ented with corn ( Zea mays L.) and soybean [ Glycine max (L.) Merr.] meal in Brazil were measured during wet and dry seasons (Lessa et al., 2014). Emissions of N 2 O were close to zero during the dry season

PAGE 114

114 (emission factors of 0.01 and 0% from urine and dung, respectively), increasing only after water was applied to the area in which excreta was previously applied. During the wet season, N 2 O emission factors were 1.9 and 0.14% for urine and dung, respectively. This difference in emissions occurred despite grea ter N concentration and excreta N application patch 1 in the dry compared with the wet season (4.2 and 8.3 g N chamber 1 during the wet season and 4.5 and 9.5 g N chamber 1 in the dry season for dung and urine, respectively), indicating that emissions fact ors are affected by seasonality. These reports of differences between emissions of N 2 O N from urine and dung and the results of this study collectively suggest that form of N, rather than N concentration per se, is of paramount importance in determining N 2 O emissions factors from animal excreta. Sordi et al. (2013) found a significant increase in NH 4 + immediately after application of urine to the soil in field conditions, but this pattern was not observed in dung patches. Yamulki et al. (1998) also observe d that available N in the form of NH 4 + and NO 3 increased to a greater extent after urine application when compared with dung. This indicates that N in urine is more available for the formation of N 2 O through nitrification, nitrifier denitrification, and denitrification (Dijkstra et al., 2013) than N in dung. We observed negative fluxes of N 2 O in our study (Figure 5 6). Uptake of N 2 O has been reported in soils with low available N and high WFPS, when denitrifying bacteria use N 2 O as an electron receptor (Schlesinger, 2013; Mazzetto et al., 2014), but literature reports likely ignore negative fluxes of N 2 O causing N 2 O uptake to be underestimated (Schlesinger, 2013). Schlesinger (2013) estimates that up to 5% of total N 2 O emissions from soil might be offset by N 2 O uptake, with median flux of 4 g N m 2 h 1 If we consider negative fluxes of N 2 O to be equal to zero, cumulative emissions of N 2 O from dung and urine in our study would be 16 times

PAGE 115

115 and 1.5 tim es greater (13.5 and 69.4 mg N 2 O N [urine patch] 1 ), increasing emission factors to 0.29 and 3.30% for dung and urine, respectively. Uptake of N 2 O observed in our study and others after application of animal excreta (Yamulki et al., 1998; Mazzetto et al., 2014) suggest this might be a relevant process occurring in grazing systems. Effect of P asture Management on Overall GHG E missions Pasture composition did not affect cumulative emissions of N 2 O N from cattle urine or dung. Therefore, considering 17 defeca tions and 9 urinations per day (Aland et al., 2002), cumulative N 2 O N emissions on an annual basis were 5 and 148 g N 2 O N animal 1 yr 1 for dung and urine, respectively. Dung CH 4 C emissions from animals grazing RP BG were greater than from animals grazing BGN (422 and 241 g CH 4 C animal 1 yr 1 ). The GWP, i.e., the ability of a gas to trap energy when compared with CO 2 is 265 for N 2 O and 28 for CH 4 (IPCC, 2014). If we consider a stocking rate of two animals ha 1 and transforming emissions to CO 2 eq based on each gas GWP, GHG emissions from animal excreta are 127 and 116 kg CO 2 eq ha 1 yr 1 for RP BG and BGN, respectively. Direct emissions from N fertilizer can be estimated to be 0.5 kg N 2 O N yr 1 using an emission factor of 1% (de Klein et al., 2006), which is equivalent to 208 kg CO 2 eq ha 1 yr 1 Therefore, annual emissions from BGN are 2.5 times greater compared than those from RP BG (324 and 127 kg CO 2 eq ha 1 yr 1 respectively), and this difference is mostly due to emissions from N fertilizer. This differ ence could potentially be greater, since in these estimates we are not accounting for indirect emissions of N 2 O from fertilizer (de Klein et al., 2006), i.e., emissions from fertilizer production storage, and transportation, and from fuel necessary for it s application (Lal, 2004). In addition, CH 4 emissions from enteric fermentation of animals grazing tropical legumes are estimated to be 20% less than those from animals grazing C4 grasses (Archimde et al., 2011). Although these differences are partly due to presence of tannins in tropical legumes (Archimde et al., 2011), which are not present in rhizoma peanut (Naumann et

PAGE 116

116 al., 2013), these results indicate that the inclusion of legumes in grass swards is a viable alternative to improve sustainability of g razing systems in terms of GHG emissions. Conclusion There was no effect of pasture source on N 2 O N emissions from animal excreta, probably due to similar N concentration in urine. However, CH 4 C emissions from dung were greater in animals grazing RP B G relative to BGN (422 and 241 g CH 4 C animal 1 yr 1 ). This difference was probably related to greater N concentration in dung from animals grazing RP BG compared with BGN (Jarvis et al., 1995). Methane emissions from dung peaked 2 d after excreta applicat ion, coinciding with high WFPS, and achieved background levels after 8 d. Emissions from N 2 O N were greater from urine than dung, supporting existing evidence that they should be accounted for separately in model estimates. These differences are likely due to greater availability of NH 4 + in urine relative to dung (Yamulki et al., 1998; Sordi et al., 2013). Water filled pore space was related to the first peak of N 2 O N from urine and dung, but later peaks in emissions occurred despite low values of WFPS. Flu xes of N 2 O N were often negative, indicating N 2 O uptake. Nitrous oxide emission factors were 2.14% for urine and 0.02% for dung in our study; this factor for dung being much lower than the 2% indicated by IPCC (2006) and EPA (2016) for the combination of b oth excreta types. This indicates that N 2 O emission estimates from animal excreta of grazing livestock production systems may be overestimated by up to three times. When considering emissions from animal excreta and fertilizer applicatio n, RP BG emitted 12 7 compared with 324 kg CO 2 eq ha 1 yr 1 for BGN, and these differences were mostly due to annual fertilizer application in the BGN treatment. Our results suggest that inclusion of legumes in grass based systems results in lesser GHG emissions compared with typical N fertilization regimes in the southeastern U.S., indicating legumes are a more

PAGE 117

117 sustainable alternative than N fertilizer for production intensification, at least in terms of GHG emissions.

PAGE 118

118 Table 5 1. Composition of fresh (as is) dung and urine fr om animals consuming bahiagrass fertilized with 50 kg N ha 1 yr 1 (BGN) or bahiagrass rhizoma peanut mixed pastures (RP BG), average d over 2 yr in Gainesville, FL. Pasture source Excreta type Chemical component BGN RP BG _____ g kg 1 _____ Dung Total N 1.8 2.9 Ammonia N 0.1 0.1 Organic N 1.7 2.7 Total P 0.8 1.0 Total K 1.6 0.8 _____ % _____ Moisture 89.0 85.1 _____ g kg 1 _____ Urine Total N 1.0 1.1 Organic N 0.8 0.3 Ammonia N 0.0 0.0 Urea N 0.2 0.8 Total P < 0.1 < 0.1 Total K 0.7 0.8 Table 5 2. Gas sampling frequency schedule for N 2 O and CH 4 flux measurements over 2 yr in Gainesville, FL. W ee k after treatment application Frequency 1 Daily 2 3 wk 1 3 3 wk 1 4 2 wk 1 5 2 wk 1 Treatment application occurred on 5 Aug. 2015 and 7 July 2016.

PAGE 119

119 Figure 5 1. Monthly average temperature and accumulated rainfall data at Alachua, FL for the years of evaluation and the 30 yr average for Gainesville, FL. Figure 5 2. Flux of CH 4 from dung of animals grazing bahiagrass pastures fertilized with 50 kg N ha 1 (BGN) or mixed rhizoma peanut bahiagrass (RP BG) over 2 yr in Gainesville, FL. Letters in italics refer to the BGN treatment, and those not in italics refer to RP BG. Statistical differences ( P < 0.05) between treatments are indicated by uppercase letters, and differences between treatments within a sampling day are indicated by lowercase letters. Month

PAGE 120

120 Figure 5 3. Percentage water filled pore space (WFPS) in the soil with no treatment application (blank) and with application of dung or urine. Light and dark gray areas in the graph represent estimated WFPS where N 2 O emissions are expected to increase according to Van Groenigen et al. (2005) and Dijkstra et al. (2013), respectiv ely. Figure 5 4. Cumulative 30 d CH 4 emissions from dung of animals grazing bahiagrass pastures fertilized with 50 kg N ha 1 (BGN) or mixed rhizoma peanut bahiagrass (RP BG) over 2 yr in Gainesville, FL.

PAGE 121

121 Figure 5 5. Flux of N 2 O from dung and urine of animals grazing bahiagrass pastures fertilized with 50 kg N ha 1 (BGN) or mixed rhizoma peanut bahiagrass (RP BG) over 2 yr in Gainesville, FL. Letters in italics refer to the BGN treatment, and those not in italics refer to RP BG. Statistical differenc es ( P < 0.05) between treatments are indicated by uppercase letters, and differences between treatments within a day are indicated by lowercase letters. Figure 5 6. Cumulative 30 d N 2 O emissions from dung and urine of animal s grazing bahiagrass pastures fertilized with 50 kg N ha 1 (BGN) or mixed rhizoma peanut bahiagrass (RP BG) over 2 yr in Gainesville, FL.

PAGE 122

122 CHAPTER 6 SUMMARY Grasslands are an importance source of feed for livestock and provide a wide array of ecosystem services (Peters et al., 2013), but most grasslands are N limited. Increasing N fertilizer cost in recent years reduces N inputs (Dubeux et al., 2007), and combined with poor grazing management (Braz et al., 2013) may lead to pasture degradation, reducing delivery of ecosystem ser vices by grasslands (Sollenberger, 2014). Where N fertilizer is used, its production, transport, and application are associated with potentially large greenhouse gas (GHG) emissions (de Klein et al., 2006; Lal, 2004). Adoption of legumes in grassland agroe cosystems is an economic alternative to the use of inorganic fertilizers (Muir et al., 2011) that improves pasture nutritive value and animal production (Muir et al., 2014) at net zero GHG emissions (Jensen et al., 2012). Significant effort has been made t o increase legume inclusion into grasslands in warm climate regions (Castillo et al., 2014 a; b ; Mullenix et al., 2016a; b) but little is known of their effect on nutrient cycling and GHG emissions in direct comparison with N fertilized pastures. Three experiments evaluating integration of legumes into C4 grasslands are described in this dissertation. The overall objectives o f this research were to determine the impact of associating legumes with grasses on nutrient cycling and greenhouse gas emissions in warm climate grasslands and to compare these responses with those of grass monocultures. In Experiment 1 (Chapter 3), a li tter bag study was conducted to quantify decomposition rate and extent of aboveground litter from fertilize d and unfertilized bahiagrass ( Paspalum notatum Fl gge) compared with mixed litter varying in proportion of the legume rhizoma peanut ( Arachis glabra ta Benth.) and bahiagrass. Treatments were unfertilized bahiagrass ( BG), bahiagrass fertilized with 60 kg N ha 1 (BG N) mixture s of bahiagrass and rhizoma peanut in proportions of 67 33% ( RP 33) and 33 67% ( RP 67), and pure rhizoma peanut (RP). Incubation

PAGE 123

123 oc curred for 128 d in 2 yr. Decomposition of total biomass followed an inverted logistic model with three phases (initial plateau, linear decrease, and final plateau). Results indicated similar total litter decomposition rates across treatments, however addi tion of legume reduced total litter biomass remaining at the end of the incubation period for RP33 compared with BG (35 and 43%, respectively). This difference was due to a longer period of linear decline in total litter biomass during incubation for RP33 relative to BG (69 and 56 d, respectively). In mixtures, biomass losses of the rhizoma peanut component (expressed as a percentage of rhizoma peanut mass in original litter) increased linearly ( P = 0.007) with increasing bahiagrass proportion in the mixtur e (i.e., was greatest for RP33), but decomposition of the bahiagrass component was not affected by presence or proportion of rhizoma peanut. The linear decomposition phase of the rhizoma peanut component was longer in RP33 relative to RP67 and RP (55, 39, and 31 d, respectively). Remaining N at the end of incubation was similar between BG and BGN (78%; P = 0.673), but addition of rhizoma peanut reduced remaining N (~55%; linear, P < 0.001; quadratic, P = 0.0015) even though C:N ratio in BGN, RP 33 and RP 67 at the end of incubation was similar ( P > 0.68; C:N of 16) Both N and C became less available as decomposition advanced, but this occurred to a greater degree for N than C. Differences in decomposition were not explained solely by chemical characteristics, indicating that other factors such as diversity in the microbial community likely influenced decomposition, particularly in mixed species litter. Our results indicate greater mineralization of N when legume is present compared with unferti lized and fertilized bahiagrass and that relatively small percentages of legume (33%) in pasture biomass can have an important positive impact on nutrient cycling that is at least as great as N fertilizer applied at commonly used or greater rates.

PAGE 124

124 In order to quantify nutrient return in legume grass mixtures compared with fertilized and unfertilized monoculture pastures, a second experiment was carried out that included both a litter bag decomposition study and measurement of litter deposition measured in r otationally stocked pastures during 2 yr (Chapter 4). Bahiagrass monoculture treatments included fertilized (50 kg N ha 1 yr 1 ) and unfertilized bahiagrass (BGN and BG, respectively). Rhizoma peanut proportion varied from 6 to 78% in mixtures with bahiagra ss. Total litter decomposition and nutrient release from litter were greater in the first (2015) than second (2016) year of the experiment, probably because of more frequent rainfall and greater overall temperatures in 2015. Biomass decomposition was faste r ( P = 0.002 and 0.025 in 2015 and 2016, respectively) and more extensive ( P = 0.006 and 0.030 in 2015 and 2016, respectively) with increasing proportion of legume in litter, being generally greater ( P > 0.028) than values observed for BG and BGN. Nitrogen released from litter increased as legume proportion increased ( P = 0.0095). Greater rate and extent of biomass and N losses during litter incubation at greater legume proportions probably occurred because of concomitant increasing N concentration in initi al litter. However, greater microbial diversity has also been proposed as a mechanism for increased decomposition rates in mixed species litter compared with monocultures. The chemical forms of C and N in plant litter changed during decomposition, with lig nin and fiber bound N concentrations increasing likely due to both accumulation and formation of new recalcitrant material. Existing litter on the soil surface decreased as legume proportion increased. This likely occurred because in spite of similar litt er deposition rates across treatments, litter decomposition was greater at higher legume proportions in the pasture. Overall, this study showed that legume grass mixtures have greater litter N concentration than fertilized or unfertilized grass litter, gre ater litter decay rates and N mineralization, and similar litter deposition rates. In aggregate,

PAGE 125

125 these responses resulted in greater amounts of N cycling in the grassland system, decreasing the likelihood of grassland degradation, while contributing to oth er ecosystem services. Previous research has shown that legume inclusion can potentially reduce GHG emissions from grazing systems because of decreased use of N fertilizer and lesser enteric methane (CH 4 ) emissions associated with high quality legume diet s (Jensen et al., 2012). However, emissions of nitrous oxide (N 2 O) from animal urine and dung and emissions of CH 4 from dung have been documented to increase when excreta has greater N concentration, which is likely to occur with the inclusion of legumes i n animal diets. Since N 2 O has greater global warming potential (GWP) than CH 4 (298 and 25, respectively), legume adoption could cause an overall increase in GHG emissions from grazing systems. In a third experiment (Chapter 5), the effect of legume inclusi on in pastures was studied in terms of GHG emissions from cattle excreta. Emissions of N 2 O from urine and dung and emissions of CH 4 from dung were evaluated during 25 d following their deposition in 2 yr for animals grazing fertilized bahiagrass (50 kg N ha 1 yr 1 BGN) or rhizoma peanut bahiagrass mixed pastures (RP BG). Emissions of CH 4 C peaked 2 d after dung deposition, which was associated with high soil water filled pore space. Methane emissions returned to background levels 8 d after treatment application. Animals grazing RP BG emitted total greater dung CH 4 C relative to BGN (422 and 241 g CH 4 C animal 1 yr 1 ). This probably occurred because of greater N concentration in dung of animals grazing RP BG. Cumulative emissions of N 2 O N were greater ( P = 0.001) from urine than dung, with emission factors (expressed as proportion of N input lost as N 2 O N) being 2.14% for urine and 0.02% for dung. These results support existing evidence that N 2 O emissions from dung and urine should be accounted for sepa rately in model estimates, and they are considerably lower than

PAGE 126

126 indicated in the literature for national GHG budgets. Differences in N 2 O emissions relative to excreta type are likely due to greater availability of NH 4 + in urine compared with dung. Fluxes o f N 2 O N were often negative, indicating N 2 O uptake. Pasture treatment had no effect on N 2 O N cumulative emissions. When comparing BGN and RP BG overall emissions, we found that RP BG emitted 127 compared with 324 kg CO 2 eq ha 1 yr 1 for BGN. Greater emissions in BGN were associated with N fertilizer use. Our results suggest that inclusion of legumes in grass based systems results in lower GHG emissions compared with typical N fertilization regimes used in the southeastern U.S., indic ating that in terms of GHG emissions, legumes are a sustainable alternative to N fertilizer for production intensification. The overall objective of this work was to compare nutrient cycling and GHG emissions of warm season, grass based grazing systems wit h N inputs of either inorganic fertilizer or biological N fixation from legumes. Our results indicate that plant litter decomposition and N mineralization were greater when legumes were present compared with unfertilized grass, and equal to or greater than those of N fertilized grass swards. When accounting for system based GHG emissions, use of N fertilizer resulted in greater overall emissions compared with legume grass mixtures. These results indicate that inclusion of legumes is a promising and potentia lly more sustainable alternative to inorganic fertilization in grass based pastures, and these results justify continuing investments in developing management strategies that facilitate legume adoption.

PAGE 127

127 LIST OF REFERENCES Aland, A., L. Lidfors, and I. Ekesbo. 2002. Diurnal distribution of dairy cow defecation and urination. Appl. Anim. Behav. Sci. 78: 43 54. Allison, S.D., Y. Lu, C. Weihe, M.L. Goulden, A.C. Martiny, K.K. Treseder, and J.B.H. Martiny. 2013. Micr obial abundance and composition influence litter decomposition response to environmental change. Ecology 94:714 725. Ambus, P., and S. Christensen. 1994. Measurement of N 2 O emission from a fertilized grassland: An analysis of spatial variability. J. Geoph ys. Res. 99(D8): 16549 16555. Anderson, J., and S. Hetherington. 1999. Temperature, nitrogen availability and mixture effects on the decomposition of heather [ Calluna vulgaris (L.) Hull] and bracken [ Pteridium aquilinum (L.) Kuhn] litters. Funct. Ecol. 13 (1): 116 124. Angers, D.A., and S. Recous. 1997. Decomposition of wheat straw and rye residues as affected by particle size. Plant Soil 189: 197 203. ANKOM Technology. 2017a. Determining acid detergent lignin in Daisy incubator. Method 9 https://www.ank om.com/sites/default/files/document files/Method_9_Lignin_in_Daisy.pdf ( accessed 1 Jan 2017). ANKOM Technology. 2017b. Acid detergent fiber in feeds Filter bag technique (for A2000 and A2000I). Acid Deter g. Fiber Method 12. https://www.ankom.com/sites/default/files/document files/Method_13_NDF_Method_A2000_RevE_4_10_15.pdf ( accessed 1 Jan 2017). Apolinrio, V.X.O., J.C.B. Dubeux, A.C.L. Mello, J.M.B. Vendramini, M.A. Lira, M.V.F. Santos, and J.P. Muir. 2013. Deposition and decomposition of signal grass pasture litter under varying nitrogen fertilizer and stocking rates. Crop Ecol. Physiol. 105(4): 999 1004. Apolinrio, V.X.O., J.C.B. Dubeux, A.C.L. Mello, J.M.B. Vendramini, M.A. Lira, M.V.F. Santos, and J.P. Muir. 2014. Li tter decomposition of signal grass grazed with different stocking rates and nitrogen fertilizer levels. Agron. J. 106(2): 622 627. Archimde, H., M. Eugne, C. Marie Magdeleine, M. Boval, C. Martin, D.P. Morgavi, P. Lecomte, and M. Doreau. 2011. Comparison of methane production between C3 and C4 grasses and legumes. Anim. Feed Sci. Technol. 166 167: 59 64. Baker, C., A. Blount, and K. Quesenberry. 1999. Peanut Stunt Virus infecting perennial peanuts in Florida and Georgia. Florida Department of A griculture & Consumer Services, Division of Plant Industry.

PAGE 128

128 Basarab, J., V. Baron, ". Lpez Campos, J. Aalhus, K. Haugen Kozyra, and E. Okine. 2012. Greenhouse gas emissions from calf and yearling fed beef production systems, with and without the use of growth promotants. Animals 2: 195 220. Bausenwein, U., P. Millard, and J.A. Raven. 2001. Remobilized old leaf nitrogen predomi nates for spring growth in two temperate grasses. New Phytol. 152(2): 283 290. Beauchemin, K.A., J.H. Henry, S.M. Little, T.A. M cAllister, and S.M. McGinn. 2010. Life cycle assessment of greenhouse gas emissions from beef production in western Canada: A case study. Agric. Syst. 103:371 379. Beauchemin, K.A., and S.M. McG inn. 2005. Methane emissions from feedlot cattle fed barley o r corn diets. J. Anim. Sci. 83: 653 661. Beauchemin, K.A., and S.M. McG inn. 2006. Met hane emissions from beef cattle: Effects of fumaric acid essential oil, and canola oil. J. Anim. Sci. 84: 1489 1496. Beauchemin, K.A., S.M. McGinn, T.F. Martinez, and T .A. McAllister. 2007. Use of condensed tannin extract from quebracho trees to reduce methane emissions from cattle. J. Anim. Sci. 85(8): 1990 1999. Bennett, L.L., A.C. Harnmond, M.J. Williams, W.E. Kunkle, D.D. Johnson, R.L. Preston, and M.F. Miller. 1995. Performance, carcass yield, and carcass quality characteristics of steers finished on rhizoma peanut ( Arachis glabrata ) tropical gras s pasture or concentrate. J. Anim. Sci. 73: 1881 1887. Berg, B., G. Ekbohm, M. B. Johansson, C. McClaugherty, F. Ru tigliano, and A. V. de Santo. 1996. Maximum decomposition limits of forest litter types: A synthesis. Can. J. Bot. 74(5): 659 672. Berg, B., and E. Matzner. 1997. Effect of N deposition on decomposition of plant litter and soil organic matter in forest sy stems. Environ.Rev. 5(March): 1 25. Berg, B., an d C. McClaugherty. 2008. Plant l itter: Decomposition, humus format ion, carbon sequestration. 2nd Ed. Springer. Berglund, S.L., G.I. Agren, and A. Ekblad. 2013. Carbon and nitrogen transfer in leaf litter mixtures. Soil Biol. Biochem. 57: 341 348. acid effect on N 2 O emissions from cow urine patches at a range of soil temperatures. p. 106 108. In 19th World Congress of Soil Science. Soil Solutions for a Changing World. Brisbane, Australia.

PAGE 129

129 Bhatta, R., O. Enishi, N. Takusari, K. Higuchi, I. Nonaka, and M. Kurihara. 2008. Diet effects on methane production by goats and a comparison between measurement methodologies. J. A gric. Sci. 146: 705 715. Bhatta, R., K. Tajima, N. Takusari, K. Higuchi, O. Enishi, and M. Kurihara. 2006. Comparison of sulfur hexafluoride tracer technique, rumen simulation technique and in vitro gas production techniques for methane production from ruminant feeds. Int. Congr. Ser. 1293: 58 61. Blount, A. R. T. Wilson, J. Ferrell, R. Mizell, and J. Bosques. 2011. Bermudagrass stem maggot a new pest in Florida. Publ. SS AGR 379 Inst. Of Food Ag. Sci., Univ. of Florida, Gainesville, FL. http://edis.ifas.ufl.edu/ag384 (accessed 1 Jan 2017). Boadi, D.A., K.M. Wittenberg, S.L. Scott, D. Burton, K. Buckley, J. a. Small, and K.H. Ominski. 2004. Effect of low and high forage diet on enteric and manure pack greenhouse gas emissions from a feedlot. Can. J. Anim. Sci. 84: 445 453. Bouw man, A.F. 1996. Direct emission of nitrous oxide from agricultural soils. Nutr. Cycl. Agroecosystems 46: 53 70. Braz, S.P., S. Urquiaga, B.J.R. Alves, C.P. Jantalia, A.P. Guimares, C. a. dos Santos, S.C. dos Santos, .F. Machado Pinheiro, and R.M. Boddey 2013. Soil carbon stocks under productive and degraded pastures in the Brazilian Cerrado. Soil Sci. Soc. Am. J. 77:914 928. Brienen, R.J.W., W. Wanek, and P. Hietz. 2011. Stable carbon isotopes in tree rings indicate improved water use efficiency and dr ought responses of a tropical dry forest tree species. Trees 25: 103 113. Brggemann, N., A. Gessler, Z. Kayler, S.G. Keel, F. Badeck, M. Barthel, P. Boeckx, N. Buchmann, E. Brugnoli, J. Esperschtz, O. Gavrichkova, J. Ghashghaie, N. Gomez Casanovas, C. K eitel, A. Knohl, D. Kuptz, S. Palacio, Y. Salmon, Y. Uchida, and M. Bahn. 2011. Carbon allocation and carbon isotope fluxes in the pl ant soil atmosphere continuum: A review. Biogeosciences 8: 3457 3489. Butler, T.J., J.P. Muir, M.A. Islam, and J.R. Bow. 2 007. Rhizoma peanut yield and nutritive value are influenced by harvest technique and timing. Agron. J. 99: 1559 1563. Caetano Anolls, G., and P.M. Gresshoff. 1991. Plant genetic control of nodulation. Annu. Rev. Microbiol. 45: 345 382. Cai, Y., S.X. Chang, and Y. Cheng. 2017. Greenhouse gas emissions from excreta patches of grazing animals and their mitigation strategies. Earth Science Rev. 171: 44 57.

PAGE 130

130 Cardenas, L.M., D. Chadwick, D. Scholefield, R. Fychan, C.L. Marley, R. Jones, R. Bol, R. Well, and A. Vallejo. 2007. The effect of diet manipulation on nitrous oxide and methane emissions from manure application to incubated grassland soils. Atmos. Environ. 41:7096 7107. Carreiro, M.M., R.L. Sinsabaugh, D.A. Repert, and D.F. Parkhurst. 2000. Microb ial enzyme shifts explain litter decay responses to simulated nitrogen deposition. Ecology 81(9): 2359 2365. Carulla, J.E., M. Kreuzer, A. Machmller, and H.D. Hess. 2005. Supplementation of Acacia mearnsii tannins decreases methanogenesis and urinary nit rogen in forage fed sheep. Aust. J. Agric. Res. 56: 961 970. Casey, J.W., and N.M. Holden. 2005. Analysis of greenhouse gas emissions from the average Irish milk production system. Agric. Syst. 86: 97 114. Casey, J.W., and N.M. Holden. 2006. Quantificat ion of GHG emissions from sucker beef production in Ireland. Agric. Syst. 90: 79 98. Castillo, M. S. L. E. Sollenberger, A. R. Blount, J. Ferrell, C. I. Na, M. Williams, and C. Mackowiak. 2014. Seedbed preparation techniques and weed control strategies for strip planting rhizoma peanut into warm season grass pastures. Crop Sci. 54: 1868 1875. Castillo, M.S., L.E. Sollenberger, A.R. Blount, J Ferrell, M.J. Williams, and C.L. Mackowiak. 2013a. Strip planting a legume into warm season grass pasture: D efoliati on effects during the year of establishment. Crop Sci. 53(2): 724 731 Castillo, M.S., L.E. Sollenberger, J.A. Ferrell, A.R. Blount, M.J. Williams, and C.L. Mackowiak. 2013b. Strategies to control competition to strip planted legume in a warm season grass pasture. Crop Sci. 53: 2255 2268. Cathey, S.E., T.R. Sinclair, and C.L. Mackowiak. 2013. Nitrogen limitation of rhizoma peanut growth. J. Plant Nutr. 36(2): 311 328. Chadwick, D.R., L. Cardenas, T.H. Misselbrook, K. a. Smith, R.M. Rees, C.J. Watson, K.L Mcgeough, J.R. Williams, J.M. Cloy, R.E. Thorman, and M.S. Dhanoa. 2014. Optimizing chamber methods for measuring nitrous oxide emissions from plot based agricultural experiments. Eur. J. Soil Sci. 65: 295 307. Chamberlain, S.D., E.H. Boughton, and J.P. Sparks. 2015. Underlying ecosystem emissions exceed cattle emitted methane from subtropical lowland pastures. Ecosystems 18(6): 933 945. Chambliss, C., and L. Sollenberger. 1991. Bahiagrass: The foundation of cow calf nutritio n in Florida. p. 74 80. In Florida beef cattle short course. Gainesville.

PAGE 131

131 Chapman, S.K., and G.W. Koch. 2007. What type of diversity yields synergy during mixed litter decomposition in a natural forest ecosystem? Plant Soil 299: 153 162. Chapman, S.K., G.S. Newman, S.C. Hart, J.A Schweitzer, and G.W. Koch. 2013. Leaf litter mixtures alter microbial community development: mechanisms for non additive effects in litter decomposition. PLoS One 8(4): e62671. Charles, D., S. Jones, B.M.R, A.G. Alway, H.S. Eggleston, an d R. Milne. 1998 Treatment of uncertainties for n ational e stimates of greenhouse g as e missions. AEA Technology, Culham, UK. Clough, T.J., P. Rochette, S.M. Thomas, M. Pihlatie, J.R. Christiansen, and R.E. Thorman. 2012. Chamber design. p. 19 33. In de Klein, C., Harvey, M. (eds.), Nitrous o xide chamber methodology g uidelines. 1st ed. Ministry for Primary Industries, Wellington, New Zealand. Cuteaux, M. M., P. Bottner, and B. Berg. 1995. Litter decomposition, climate and litter quality. Tree 10(2): 63 66. Cuchietti, A., E. Marcotti, D.E. Gurvich, A.M. Cingolani, and N. Prez Harguindeguy. 2014. Leaf litter mixtures and neighbor effects: Low nitrogen and high lignin species increase decomposition rate of high nitrogen and low lignin neighbors. Appl. Soil Ecol. 82: 44 5 1. de Klein, C. A.M. R.S.A. Novoa, S. Ogle, K.A. Smith, B.G. McConkey, A. Mosier, and K. Rypdal. 2006. N 2 O Emissions from managed soils, and CO 2 emissions from lime and urea application. p. 11.1 11.54. In 2006 IPCC Guidelines for National Greenhouse Gas I nventories. de Klein, C.A.M., and R.S.P. van Logtestijn. 1994. Denitrification and N 2 O emission from urine affected grassland soil. Plant Soil 163: 235 242. Dijkstra, J., O. Oenema, and A. Bannink. 2011. Dietary strategies to red ucing N excretion from ca ttle: I mplications for methane emissions. Curr. Opin. Environ. Sustain. 3(5): 414 422. Dijkstra, J., O. Oenema, J.W. van Groenigen, J.W. Spek, A.M. van Vuuren, and A. Bannink. 2013. Diet effects on urine composition of cattle and N 2 O emissions. Animal 7(s2):292 302. Dini, Y., J. Gere, C. Briano, M. Manetti, P. Juliarena, V. Picasso, R. Gratton, and L. Astigarraga. 2012. Methane emission and milk production of dairy cows grazing pastures rich in legumes or rich in grasses in Uruguay. Animals 2: 288 300. Dubeux, J.C.B. 2005. Management strategies to improve nutrient cycling in grazed Pens bahiagrass pastures. Ph.D. diss Univ. of Florida, Gainesville.

PAGE 132

132 Dubeux, J.C.B., A.R. Blount, C. Mackowiak, E.R .S. Santos, J.D. Pereira Neto, U. Riveros, L. Garcia, D.M. Jaramillo, and M. Ruiz Moreno. 2017. Biological N 2 fixation, belowground responses, and forage potential of rhizoma peanut cultivars. Crop Sci. 57: 1027 1038. Dubeux, J.C.B., L.E. Sollenberger, L. A. Gaston, J.M.B. Vendramini, S.M. Interrante, and R.L. Stewart. 2009. Animal behavior and soil nutrient redistribution in continuously stocked 1510. Dubeux, J.C.B., L.E. Sollenberger, S.M. Interrante, J.M.B. Vendramini, and R.L. Stewart. 2006a. Litter decomposition and mineralization in bahiagrass pastures managed at different intensities. Crop Sci. 46(3):1305 1310. Dubeux, J.C.B., L.E. Sollenberger, B.W. Mathews, J.M. S cholberg, and H.Q. Santos. 2007. Nutrient cycling in warm climate grasslands. Crop Sci. 47:915 928. Dubeux, J.C.B., L.E. Sollenberger, J.M.B. Vendramini, R.L. Stewart, and S.M. Interrante. 2006b. Litter mass, deposition rate, and chemical composition in b ahiagrass pastures managed at different intensities. Crop Sci. 46(3):1299 1304. EPA. 2016. Inventory of U. S greenhouse gas emissions and sinks: 1990 2014. Washington, DC. Farquhar, G.D., M.H.O. Learyb, and J.A. Berry. 1982. On the relationship betwee n carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Aust. J. Plant Physiol. 9: 121 137. Ferguson, B.J., A. Indrasumunar, S. Hayashi, M. H. Lin, Y. H. Lin, D.E. Reid, and P.M. Gresshoff. 2010. Molecular analysis of legume nodule development and autoregulation. J. Integr. Plant Biol. 52(1): 61 76. Fike, J.H., C.R. Staples, L.E. Sollenberger, B. Macoon, and J.E. Moore. 2003. Pasture forages, supplementation rate, and stocking rate effects on dairy cow performance. J. Dairy Sci. 86: 1268 1281. Fialho, C.A. 2015. Morphogenetic and structural characteristics of forage peanut ( Arachis pintoi Krapovickas & Gregory cv. Belmonte) subjected to grazing intensities under continuous stocking. http://www.teses.usp.br/teses/dispo niveis/11/11139/tde 04052015 110506/ ( accessed 2 Aug. 2017). Flysj, A., M. Henriksson, C. Cederberg, S. Ledgard, and J. Englund. 2011. The impact of various parameters on the carbon footprint of milk production in New Zealand and Sweden. Agric. Syst. 10 4: 459 469.

PAGE 133

133 farm systems modelling of greenhouse gas emissions from pastoral suckler beef cow production systems. Agric. Ecosyst. Environ. 142: 222 230. Gallaher, R.N., C.O. Weldon, and J.G. Futral. 1975. An aluminum block digester for plant and soil analysis. Soil Sci. Soc. Am. J. 39: 803 806. Gijsman, A.J., H.F. Alarcn, and R.J. Thomas. 1997. Root decomposition in tropical grasses and legumes, as affected by so il texture and season. Soil Biol. Biochem. 29(9 10): 1443 1450. Giltrap, D.L., P. Berben, T. Palmada, and S. Saggar. 2014. Understanding and analysing spatial variability of nitrous oxide emissions from a grazed pasture. Agric. Ecosyst. Environ. 186(3): 1 10. Glaser, B. 2005. Compound specific stable Soil Sci. 168: 633 648. Guretzky, J.A., W.H. Schacht, A. Wingeyer, T.J. Klopfenstein, and A. Watson. 2014. Litter deposition and nitrogen return in rot ationally stocked smooth bromegrass pastures. Agron. J. 106(1): 175. Haile, S., V. Nair, and P.K.R. Nair. 2010. Contribution of trees to carbon storage in soils of silvopastoral systems in Florida, USA. Glob. Chang. Biol. 16: 427 438. Hambleton, L.G. 197 7. Semiautomated method for simultaneous determination of phosphorus, calcium and crude protein in animal feeds. J.A.O.A.C 60: 845 852. Haynes, R.J., and P.H. Williams. 1993. Nutrient cycling and soil fertility in the grazed pasture ecosystem. Adv. Agron. 49: 119 199. Harper, L.A., O.T. Denmead, and T.K. Flesch. 2011. Micrometeorological techniques for measurement of enteric greenhouse gas emissions. Anim. Feed Sci. Technol. 166 167: 227 239. Harper, L.A., R.R. Sharpe, T.B. Parkin, A. De Visscher, O. van Cleemput, and F.M. Byers. 2004. N itrogen cycling through swine production systems: A mmonia, dinitrogen, and nitrous oxide emissions. J. Environ. Qual. 33: 1189 1201. Heal, O.W., J.M. Anderson, and M.J. Swift. 1997. Plant litter quality and decomposition: An historical overview. p. 3 30. In Cadisch, G., Giller, K.E. (eds.), Driven by n ature: Plant litter quality and decomposition. CAB Internationl. Henderson, M.S., and D.L. Robinson. 1982. Environmental influences on yield and in vitro true digestibility of warm season perennial grasses and the relationships to fiber components. Agron. J. 74: 943 946.

PAGE 134

134 Henry, D.D., M. Ruiz Moreno, F.M. Ciriaco, M. Kohmann, V.R.G. Mercada nte, G.C. Lamb, and N. DiLorenzo. 2015. Effects of chitosan on nutrient digestibility, methane emissions, and in vitro fermentation in beef cattle. J. Anim. Sci. 93(7): 3539 3550. Hernndez Garay, A., L.E. Sollenberger, C.R. Staple s, and C.G.S. Pedreira. Florigraze 1355 1360. Hristov, A.N., T. Ott, J. Tricarico, A. Rotz, G. Waghorn, A. Adesogan, J. Dijskstra, F. Montes, J. Oh, E. Kebreab, S.J. Oosting, P.J. Gerber, B. Henderson, H.P.S. Makkar, and L. Firkins. 2013. SPECIAL TOPICS Mitigation of methane and nitrous oxide emissions from animal ope rations: III A review of animal management mitigation options. J Anim. Sci. 91: 5095 5113. Hubicka, K.T., G.D. Farquhar and R. Shorter. 1986. Correlation between water use efficiency and carbon isotope discrimination in diverse peanut (Arachis) germplasm. Aust. J. Plant Physiol. 13: 803 816. Hutchinson, G.L., and G.P. Livingston. 2001. Vents and seals in non steady state chambers used for measuring gas exchange between soil and the atmosphere. Eur. J. Soil Sci. 52(4): 675 682. Hutchinson, G.L., and A.R. Mosier. 1981. Improved soil cover method for field measurement of nitrous oxide fluxes. Soil Sci. Soc. Am. J. 45: 311 3 16 Inyang, U., J.M.B. Vendramini, L.E. Sollenberger, B. Sellers, A. Adesogan, L. Paiva, and A. Lunpha. 2010. Forage species and stocking rate effects on animal performance and 10 85. IPCC. 2006. 2006 IPCC Guidelines for nati onal greenhouse gas inventories ( Eggleston H.S., Buendia L., Miwa K., Ngara T. and Tanabe K., e ds.). IGES, Japan. IPCC. 2014. Climate Change 2014: Synthesis report. Contribution of working g roups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (R K Pachaur and L A Meyer, e ds.). Geneva, Switzerland. Jarvis, S.C., R.D. Lovell, and R. Panayides. 1995. Patterns of methane emission from excreta of grazing animals. Soi l Biol. Biochem. 27:1581 1588. Jensen, E.S., M.B. Peoples, R.M. Boddey, P.M. Gresshoff, H. Hauggaard Nielsen, B. Alves, and M.J. Morrison. 2012. Legumes for mitigation of climate change and the provision of feedstock for biofuels and biorefineries. A review. Agron. Sustain. Dev. 32: 329 364. Johnson, C.R., B.A. Reiling, P. Mislevy, and M.B. Hall. 2001. Effects of nitrogen fertilization and harvest date on yield, digestibility, fiber, and protein fractions of tropical grasses. J. Anim. Sci. 79: 2439 24 48.

PAGE 135

135 Johnson, K. A. M. Huyler, H. Westberg, B. Lamb, and P. Zimmerman. 1994. Measurement of methane emissions from r uminant l ivestock using a SF6 tracer t echnique. Environ. Sci. Technol. 28: 359 362. Johnson, K.A., and D.E. Johnson. 1995. Methane emissions from cattle. J Anim. Sci. 73: 2483 2492. Johnson, K.A., H. H. Westberg, J.J. Michal, and M. W. Cossalman. 2007. The SF6 tracer technique: M ethane measurement from ruminants. p. 33 67. In Makkar, H.P.S., Vercoe, P.E. (eds.), Measuring methane production fr om r uminants. Springer Netherlands. Kebreab, E., K. Clark, C. Wagner Riddle, and J. France. 2006. Methane and nitrous oxide emissions from Canadian animal agriculture: A review. Can. J. Anim. Sci. 86: 135 157. Kerr idge, Peter, C., and B. Hardy (e ds). 199 4. Biology and agronomy of forage Arachis. International Center for Tropical Agriculture. Klumpp, K., J.M.G. Bloor, P. Ambus, and J. F. Soussana. 2010. Effects of clover density on N 2 O emissions and plant soil N transfers in a fertilised upland pasture. Plant Soil 343: 97 107. Krapovickas, A., and W.C. Gregory. 2006. Taxonomia del genero Arachis (Leguminosae). Bonplandia 8(1 4): 1 206. Kurihara, M., T. Magner, R. A. Hunter, and G.J. McCrabb. 1999. Methane production and energy partition of cattle in the tropics. Br. J. Nutr. 81(3): 227 34. Lal, R. 2004. Carbon emission from farm operations. Environ. Int. 30(7): 981 90. Lassey, K.R. 2007. Livestock methane emission: From the individual grazing animal through national inventories to the global methane cy cle. Agric. For. Meteorol. 142: 120 132. L ee, C., A.N. Hristov, C.J. Dell, G.W. Feyereisen, J. Kaye, and D. Beegle. 2012. Effect of dietary protein concentration on ammonia and greenhouse gas emitting potential of dairy manure. J. Dairy Sci. 95: 1930 1941. Leek, B.F. 2004. Digestion in the Ruminant Stomac h. p. 438 474. In Reece, W.O. (E Physiology of Domestic Animals. 12th ed. Cornell University Press, New York. Le nth, R. V. 2016. Least squares means: The R p ackage lsmeans. J. Stat. S oftw. 69(1):1 33. Lessa, A.C.R., B.E. Madari, D.S. Paredes, R.M. Boddey, S. Urquiaga, C.P. Jantalia, and B.J.R. Alves. 2014. Bovine urine and dung deposited on Brazilian savannah pastures contribute differently to direct and indirect soil nitrous oxide em issions. Agric. Ecosyst. Environ. 190: 104 111.

PAGE 136

136 Libault, M. 2014. The carbon nitrogen balance of the nodule and its regulation. Biomed Res. Int. 2014: 1 7. Liu, K., L.E. Sollenberger, M.L. Silveira, Y.C. Newman, and J.M.B. Vendramini. 2011 a Grazing int bermudagrass pastures: I. M ass, deposition rate, and chemical composition. Agron. J. 103(1):156 162. Liu, K., L.E. Sollenberger, M.L. Silveira, J.M.B. Vendramini, and Y.C. Newman. 2011 b bermudagrass pastures: II. D ecomposition and nitrogen mineralization. Agron. J. 103: 163 168. Liu, P., O.J. Sun, J.H. Huang, L.H. Li, and X.G. Han. 2007. Nonadditiv e effects of litter mixtures on decomposition and correlation with initial litter N and P concentrations in grassland plant species of northern China. Biol. Fertil. Soils 44: 211 216. Lockyer, D.R., and S.C. Jarvis. 1995. The measurement of methane losses from grazing animals. Environ. Pollut. 90(3): 383 90. Lockyer, D.R., and D.C. Whitehead. 1990. Volatilization of ammonia from cattle urine applied to grassland. Soil Biol. Biochem. 22(8): 1137 1142. Manzoni, S., J.A. Trofymow, R.B. Jackson, and A. Porpo rato. 2015. Stoichiometric controls on carbon, nitrogen, and phosphorus dynamics in decomposing litter. Ecol. Monogr. 80(1): 89 106. Mazzetto, A.M., A.S. Barneze, B.J. Feigl, J.W. Van Groenigen, O. Oenema, and C.C. Cerri. 2014. Temperature and moisture af fect methane and nitrous oxide emission from bovine manure patches in tropical conditions. Soil Biol. Biochem. 76:242 248. McCartor, M.M., and F.M. Rouquette. 1977. Grazing pressures and animal performance from pearl m illet. Agron. J. 69(6): 983 987 McCaughey, W.P., K. Wittenberg, and D. Corrigan. 1997. Methane production by steers on pasture. Can. J. Anim. Sci. (77): 519 524. McCurdy, J.D., J.S. McElroy, E.A. Guertal, and C.W. Wood. 2013. Dynamics of white clover decomposition in a Southeastern ber mudagrass lawn. Agron. J. 105(5): 1277 1282. McSwiney, C.P., S. Bohm, P.R. Grace, and G.P. Robertson. 2010. Greenhouse gas emissions calculator for grain and biofuel farming s ystems. J. Nat. Resour. Life Sci. Educ. 39: 125 131. Moe, P.W., a nd H.F. Tyrrel l. 1979. Methane p roduction in d airy c ows. J. Dairy Sci. 62(10): 1583 1586.

PAGE 137

137 Moore, J.E., and G.O. Mott. 1974. Recovery of residual organic matter from in vitro digestion of forages. J. Dairy Sci. 57(10):1258 1259. Mori, A., and M. Hojito. 2015. Methane an d nitrous oxide emissions due to excreta returns from grazing cattle in Nasu, Japan. Grassl. Sci. 61(2): 109 120. Moss, A.R., J. P. Jouany, and J. Newbold. 2000. Me thane production by ruminants: I ts contribution to global warming. Ann. Zootech. 49(3): 231 253. Mottet, A., C. de Haan, A. Falcucci, G. Tempio, C. Opio, and P. Gerber. 2016. Livestock: On our plates or eating at our table? A new analysis of the feed/food debate. Glob. Food Sec. 14: 1 8. Muir, J.P., T.J. Butler, W.R. Ocumpaugh, and C.E. Sim peanut for cool, dry climates. J. Plant Regist. 4(2): 106 108. Muir, J.P., W.D. Pitman, J.C. B. Dubeux, and J.L. Foster. 2014. The future of warm season, tropical and subtropical forage legumes in sustainable pastures and rangelands. African J. Range Forage Sci. 31(3): 187 198. Muir, J.P., W.D. Pitman, and J.L. Foster. 2011. Sustainable, low input, warm season, grass leg ume grassland mixtures: M ission (nearly) impossible? Grass Forage Sci. 66:301 315. Mullenix, M.K., L.E. Sollenberger, M.O. Wallau, A.R. Blount, J.M.B. Vendramini, and M.L. Silveira. 2016a. Herbage accumulation, nutritive value, and persistence responses of rhizoma peanut cultivars and germplasm to grazing management. Crop Sci. 56(2): 907 915. Mullenix, M.K., L.E. Sollenberger, M.O. Wallau, D.L. Rowland, A.R. Blount, J.M.B. Vendramini, and M.L. Silveira. 2016 b Sward structure, light interception, and rhizome root responses of rhizoma peanut cultivars and germplasm to grazing manageme nt. Crop Sc i. 56:899 906. Murray, B.Y.R.M., A.M. Bryant, and R.A. Leng. 1976. Rates of production of methan e in the rumen and large intestine of sheep. Br. J. Nutr. 36(1): 1 14. Murray, P.J., A. Moss, D.R. Lockyer, and S.C. Jarvis. 1999. A comparison of systems fo r measuring methane e missions from sheep. J. Agric. S cience 133: 439 444. Naumann, H.D., L.O. Tedeschi, J.P. Muir, B.D. Lambert, and M.M. Kothmann. 2013. Effect of molecular weight of condensed tannins from warm season perennial legumes on ruminal methane production in vitro. Biochem. Syst. Ecol. 50: 154 162. Nichols, K.L., S.J. Del Grosso, J.D. Derner, R.F. Follett, S.L. Archibeque, C.E. Stewart, and K.H. Paustian. 2016. Nitrous oxide and methane fluxes from cattle excrement on C3 pasture and C4 dominated shortgrass steppe. Agric. Ecosyst. Environ. 225: 104 115.

PAGE 138

138 Nyfeler, D., O. Huguenin Elie, M. Suter, E. Frossard, J. Connolly, and A. Lscher. 2009. Strong mixture effects among four species in fertilized agricultural grassland led to persistent and consistent t r ansgressive overyielding. J. Appl. Ecol. 46:683 691. Nyfeler, D., O. Huguenin Elie, M. Suter, E. Frossard, and A. Lscher. 2011. Grass legume mixtures can yield more nitrogen than legume pure stands due to mutual stimulation of nitrogen uptake from symbiotic and non symbiotic sources. Agric. Ecosyst. Environ. 140:155 163. Oades, J.M., M.A. Kirkman, and G.H. Wagner. 1970. The use of gas liquid chromatography for the determination of sugars extracted from soils by sulfuric acid. Soil Sci. Soc. Am. J. 34(2): 230 235. Oenema, O., G.L. Velthof, S. Yamulki, and S.C. Jarvis. 1997. Nitrous oxide emissions from grazed grassland. Soil Use Manag. 13:288 295. Oenema, O., N. Wrage, G.L. Velthof, J.W. van Groenigen, J. Dolfing, and P.J. Kuikman. 2005 Trends in global nitrous oxide emissions from animal production systems. Nutr. Cycl. Agroecosystems 72: 51 65. rskov, E.R., W.P. Flatt, an d P.W. Moe. 1968. Fermentation b a lance approach to estimate e xtent of fermentation and efficiency of volatile fatt y acid formation in r uminants. J. Dairy Sci. 51(9): 1429 1435. Ortega S., J.A., L.E. Sollenberger, K.H. Quesenberry, J.A. Cornell, and C.S. Jones Jr.. 1992. Productivity and persistence of rhizoma peanut p a stures under different grazing managements. Agron J 84:799 804. Pal, P., T.J. Clough, F.M. Kelliher, and R.R. Sherlock. 2013. Nitrous oxide emissions from in situ deposition of N labeled ryegrass litter in a pasture soil. J. Environ. Qual. 42: 323 331. Pal, P., T.J. Clough, F.M. Kelliher, C. van Koten, and R.R. Sherlock. 2012. Intensive cattle grazin g affects pasture litter fall: A n unrecognized nitrous oxide source. J. Environ. Qual. 41: 444 448. Parkin, T.B., and R.T. Venterea. 2010. Chamber based trace gas flux measurements Sampling protocol s p. 1 39. In USDA ARS GRACEnet project protocols, Beltsville, M.D. Pelletier, N., R. Pirog, and R. Rasmussen. 2010. Comparative life cycle environmental impacts of three beef production strategies in the Upper Midwestern United States. Agric. Syst. 103 (6): 380 389. Peters, M., M. Herrero, M. Fisher, K H. Erb, I. Rao, G.V. Subbarao, et al. 2013. Challenges and opportunities for improving eco efficiency of tropical forage based systems to mitigate greenhouse gas emissions. p. 1251 1260. In Proc. 22nd Int. Grassl. Cong.

PAGE 139

139 Pinares Pati n o, C.S., M.J. Ulyatt, K.R. Lassey, T.N. Barry, and C.W. Holmes. 2003. Rumen function and digestion parameters associated with differences between sheep in methane emissions when fed chaffed lucerne hay. J. Agric. Sci. 140(2) : 205 214. Pinheiro, J., D. Bates, S. DebRoy, D. Sarkar, and R.C Team. 2016. nlme: Linear and nonlinear mixed effects m odels. R package version 3.1 127. Prine, G.M., E.C. French, A.R. Blount, M.J. Williams, and K.H. Quesenberry. 2010. Registration of Ar blick and Ecoturf rhizoma peanut germplasms for ornamental or forage use. J. Plant Regist. 4(2): 145 148. Puchala, R., B.R. Min, A.L. Goetsch, and T. Sahlu. 2005. The effect of a condensed tannin containing forage on methane emission by goats. J. Anim. Sc i. 83: 182 186. Quesenberry, K.H., A.R. Blount, P. Mislevy, E.C. French, M.J. Williams, and G.M. Prine. 2010. yields, persistence, and disease tolerance. J. Plant Regis t. 4(1): 17 21. R Core Team. 2016. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Rafique, R., D. Hennessy, and G. Kiely. 2011. Nitrous oxide emission from grazed grassland under differen t management systems. Ecosystems 14:563 582. Rajabi, A., E.S. Ober, and H. Griffiths. 2009. Genotypic variation for water use efficiency, carbon isotope discrimination, and potential surrogate measures in sugar beet. F. Crop. Res. 112: 172 181. Rezende, C.D.P., R.B. Cantarutti, J.M. Braga, J.A. Gomide, J.M. Pereira, and E. Ferreira. 1999. Litter deposition and disappearance in Brachiaria pastures in the Atlantic forest region of the South of Bahia, Brazil. Nutr. Cycl. Agroecosytems 54:99 112. Rochette, P ., D.R. Chadwick, C.A.M. de Klein, and K. Cameron. 2012. Deployment protocol. p. 34 55. In de Klein, C., Harvey, M. (eds.), Nitrous oxide chamber methodology g uidelines. 1st ed. Ministry for Primary Industries, Wellington, New Zealand. Rs, E., C. Sundberg, P. Tidker, I. Strid, and P. A. Hansson. 2013. Can carbon footprint serve as an indicator of the environm ental impact of meat production ? Ecol. Indic. 24: 573 581. Rovira, P., and V.R. Vallejo. 2002. Labile and recalcitrant pools of carbon and n itrogen in organic matter decomposin g at different depths in soil: a n acid hydrolysis approach. Geoderma 107: 109 141. Russel, J.B. 2002. Rumen microbiology and its role in ruminant nutrition Ithaca, NY.

PAGE 140

140 Rypdal, K., and W. Winiwarter. 2001. Uncertaintie s in greenhouse gas emission inventories evaluation, comparability and implications. Environ. Sci. Policy 4: 107 116. Sanaullah, M., A. Chabbi, J. Leifeld, G. Bardoux, D. Billou, and C. Rumpel. 2011. Decomposition and stabilization of root litter in top and subsoil horizons: W hat is the difference? Plant Soil 338: 127 141. Sasaki, T., T. Suzaki, T. Soyano, M. Kojima, H. Sakakibara, and M. Kawaguchi. 2014. Shoot derived cytokinins systematically regulate root nodulation. Nat. Commun. 5. Schnyder, H., F Locher, and K. Auerswald. 2010. Nutrient redistribution by grazing cattle drives patterns of topsoil N and P stocks in a low input pasture ecosystem. Nutr. Cycl. Agroecosystems 88: 183 195. Scherer Lorenzen, M. 2008. Functional diversity affects decompo sition processes in experimental grasslands. Funct. Ecol. 22:547 555. Schlesinger, W.H. 2013. An estimate of the global sink for nitrous oxide in soils. Glob. Chang. Biol. 19: 2929 2931. Sebaugh, J.L., and P.D. McCray. 2003. Defining the linear portion o f a sigmoidshaped curve: Bend points. Pharm. Stat. 2: 167 174. Seibt, U., A. Rajabi, H. Griffiths, and J.A. Berry. 2008. Carbon isot opes and water use efficiency: S ense and sensitivity. Oecologia 155: 441 454. Shine, K.P., R.G. Derwent, D.J. Wuebbles, and J. J. Morcrette. 1990. Radiative forcing of climate. p. 410. In Report prepared for Intergovernmental Panel on Climate Change by Working Group I. Sigua, G.C., C.C. Chase Jr., and J. Albano. 2014. Soil extractable phosphorus and phosphorus saturation threshold in beef cattle pastures as affected by grazing management and forage type. Environ. Sci. Pollut. Res. iIternational 21: 1691 1700. Silva, H.M S. da J.C. B Dubeux, M.V. Ferreira dos Santos, M. de Andrade Lira, and J.P. Muir. 2012. Signal grass litter decomposition rate increases with inclusio n of Calopo. Crop Sci. 52:1416 1423. Silva, H.M. S. da J.C.B. Dubeux, M.L. Silveira, E.V. de Freitas, M .V.F. d os Santos, M. de Andrade Lira 2015. Stocking rate and nitrogen fertilization affect root decomposition of elephantgrass. Agron. J. 107(4): 1331 1338. Silveira, M.L., J.M.B. Vendramini, H.M.S. da Silva, B.M.M.N. Borges, V.S. Ribeirinho, J.J.J. Lacerda M.V. Azenha, P.R.A. Viegas, A.D. Aguiar. 2017. Potassium and phosphorus fertilization impacts on bermudagrass and limpograss herbage accumulation, nutritive valu e, and persistence. Crop Sci. (in press).

PAGE 141

141 Smith, D.R., N. Dilorenzo, J. Leibovich, M. L. May, M.J. Quinn, J.W. Homm, and M.. L. Galyean. 2010. Effects of sulfur and monensin concentrations on in vitro dry matter disappearance, hydrogen sulfide production, and volatile fatty acid concentrations in batch culture ruminal fermentations. J. Anim Sci. 88: 1503 1512. Smith, V.C., and M.A. Bradford. 2014. Do non additive on decomposition experiments effects result from differences in resource quality between litters? Oikos 102:235 242. Sol lenberger, L.E. 2014. Changing e mphases in soil plant anim al research in p astures. p. 1 38. In Sociedade Brasileira de Zootecnica, Brasilia, B. (ed.), 51st Proc. Brazilian Soc. Animal Sci. Mtg. Barra dos Coqueiros. Sollenberger, L.E., J.C. B. Dubeux, Jr., and J.P. Muir. 2014. Establishment and management of legum e grass pastures. p. 135 177. In O.G. Pereira et al. (ed.) 7th Symposium on Strategic Management of Pasture. 13 15 Nov. 2014. Viosa, Brazil. Sollenberger, L.E., and C.S. Jones. 1989. Beef production from nitrogen fertilized Mott dwarf elephantgrass and P ensacola bahiagrass pastures. Trop. Grasslands 23(3): 129 134. Sollenberger, L.E., C.S. Jones, and G.M. Prine. 1989a. Animal performance on dwarf elephantgrass and rhizoma peanut pasture. p. 1189 1190. In Desroches, R. (ed.), XVI International Grassland C ongress. The French Grassland Society, Versailles Cedex, France., Nice, France. 4 11 Oct.1989. Sollenberger, L.E., G.A. Rusland, C.S. Jones, K.A. Albrecht, and K.L. Gieger. 1989b. Animal bahiagrass pastures. Agron. J. 81: 760 764. Sordi, A., J. Dieckow, C. Bayer, M.A. Alburquerque, J.T. Piva, J.A. Zanatta, M. Tomazi, C.M. da Rosa, and A. de Moraes. 2013. Nitrous oxide emission factors for urine and dung patches in a subtropical Brazilian pastureland. Agric. Ecosyst. Environ. 190: 94 103. Steinfeld, H., and T. Wassenaar. 2007. The role of livestock production in carbon and nitrogen cycles. Annu. Rev. Environ. Resour. 32: 271 294. Stewart, R.L., L.E. Sollenberger, J.M.B. Vendramini, S.M. Interrante, and Y.C. Newman. 2007. Herbage and animal responses to manageme nt intensity of continuously stocked bahiagrass pastures. Agron. J. 99: 107 112. Storm, I.M.L.D., A.L.F. Hellwing, N.I. Nielsen, and J. Madsen. 2012. Methods for measuring and estimating methane e mission from r uminants. Animals 2(4): 160 183. Subbarao, G V., I.M. Rao, K. Nakahara, Y. Ando, K.L. Sahrawat, and T. Tesfamariam. 2013. Nitrogen management in grasslands and forage based production systems Role of biological nitrification inhibition (BNI). Trop. Grasslands 1: 168 174.

PAGE 142

142 Ta, T., and M. Faris. 198 7. Effects of alfalfa proportions and clipping frequencies on timothy alfalfa mixtures. II. Nitrogen fixation and transfer. Agron. J. 79: 820 824. Ta, T.C., F.D.H. Macdowall, and M.A. Faris. 1986. Excretion of nitrogen assimilated from N 2 fixed by nodulat ed roots of alfalfa ( Medicago sativa ). Can. J. Bot. 64: 2063 2067. Thomas, R.J., and N.M. Asakawa. 1993. Decomposition of leaf litter from tropical forage grasses and legumes. Soil Biol. Biochem. 25:1351 1361. Twidwell, E., W.D. Pitman, and G.J. Cuomo. 1 998. Bahiagrass production and management. LSU publication # 2697. Uchida, Y., T.J. Clough, F.M. Kelliher, J.E. Hunt, and R.R. Sherlock. 2011. Effects of bovine urine, plants and temperature on N 2 O and CO 2 emissions from a sub tropical soil. Plant Soil 345: 171 186. United States Department of Agriculture. 2012. Florida agriculture by the numbers. U nited S tates D epartment of A griculture 1998. Soil bulk density/ Moisture/ Aeration. Valencia, E., M.J. Wil liams, C.C. Chase, L.E. Sollenberger, A.C. Hammond, R.S. Kalmbacher, and W.E. Kunkle. 2001. Pasture management effects on diet composition and cattle performance on continuously stocked rhizoma peanut mixed grass swards. J. Anim. Sci. 79: 2456 2464. van d er Weerden, T.J., J. Luo, C. a. M. de Klein, C.J. Hoogendoorn, R.P. Littlejohn, and G.J. Rys. 2011. Disaggregating nitrous oxide emission factors for ruminant urine and dung deposited onto pastoral soils. Agric. Ecosyst. Environ. 141:426 436. van Groenigen, J.W., P.J. Kuikman, W.J.M. de Groot, and G.L. Velthof. 2005a. Nitrous oxide emission from urine treated soil as influenced by urine composition and soil physical conditions. Soil Biol. Biochem. 37: 463 473. van Groenigen, J.W., G.L. Velthof, F. J.E. van der Bolt, A. Vos, and P.J. Kuikman. 2005b. Seasonal variation in N 2 O emissions from urine patches: Effects of urine concentration, soil compaction and dung. Plant Soil 273: 15 27. Vendramini, J. M.B. 2010. Forage evaluation and quality in Florida. p. 92 106. In The 21st Annual Florida Ruminant Nutrition. Gainesville. Vendramini, J.M.B., A.T. Adesogan, M.L.A. Silveira, L.E. Sollenberger, O.C.M. Queiroz, and W.F. Anderson. 2010. Nutritive value and fermentation parameters of warm season grass silage Prof. Anim. Sci. 26: 193 200. Vendramini, J.M.B., J.C.B. Dubeux, and M.L. Silveira. 2014. Nutrient cycling in tropical pasture ecosystems. Rev. Bras. Ciencias Agrar. 9(2): 308 315.

PAGE 143

143 Venterea, R.T., T.B. Parkin, L. Cardenas, S.O. Petersen, and A.R. Pedersen. 2012. Data analysis considerations. p. 95 121. In de Klein, C., Harvey, M. (eds.), N itrous oxide chamber methodology guidelines. 1st ed. Ministry for Primary Industries, Wellington New Zealand. Venterea, R.T., K. A. Spokas, and J.M. Baker. 2009. Accuracy and precision analysis of chamber based nitrous oxide gas flux estimates. Soil Sci. Soc. Am. J. 73(4): 1087 1093. Virkajrvi, P., M. Maljanen, K. Saarijrvi, J. Haapala, and P.J. Martikainen. 2010. N 2 O emissions from boreal grass and grass clover pasture soils. Agric. Ecosyst. Environ. 137: 59 67. Wall, E., G. Simm, and D. Moran. 2009. Developing breeding schemes to assist mitigation of greenhouse gas emissions. Animal 4(3): 36 6 376. Werner, R.A., and W.A. Brand. 2001. Referencing strategies and techniques in stable isotope ratio analysis. Rapid Commun. M ass Spectrom. 15: 501 19. White Leech, R., K. Liu, L.E. Sollenberger, K.R. Woodard, and S.M. Interrante. 2013 a Excreta depo sition on grassland. I. Forage harvested, nutritive value, and nitrogen recovery. Crop Sci. 53:688 695. White Leech, R., K. Liu, L.E. Sollenberger, K.R. Woodard, and S .M. Interrante. 2013b. Excreta deposition on g rassland patches. II. Spatial pattern and duration of forage r esponses. Crop Sci. 53: 696 703. Whitehead, D.C. 2000. Nutrient elements in grassland: Soil plant animal relationships (DC Whitehead, e d.). CAB International, Wallingford. Wider, R.K., and G.E. Lang. 1982. A critique of the analytical methods used in examining decomposition data obtained from litter bags. Ecology 63(6):1636 1642. Williams, M.J. 1994. Growth characteristics of rhizoma peanut and nitrogen fertilized bahiagrass swards. Agron. J. 86: 819 823. Williams, M.J., K.H. Quesen berry, G.M. Prine, and C.B. Olson. 2005. Rhizoma peanut more 229. In Proceedings of XX International Grasslands Congress. Dublin, Ireland. Winiwarter, W., and K. Rypdal. 2001. Assessing the uncertainty associa ted with national gree nhouse gas emission inventories: A case study for Austria. Atmos. Environ. 35: 5425 5440. Wrage, N., G.L. Velthof, M.L. Van Beusichem, and O. Oenema. 2001. Role of nitrifier denitrification in the production of nitrous oxide. Soil Bi ol. Biochem. 33(12 13): 1723 1732.

PAGE 144

144 Yan, T., C.S. Mayne, F.G. Gordon, M.G. Porter, R.E. Agnew, D.C. Patterson, C.P. Ferris, and D.J. Kilpatrick. 2010. Mitigation of enteric methane emissions through improving efficiency of energy utilization and productivit y in lactating dairy cows. J. Dairy Sci. 93(6): 2630 8. Yamulki, S., S.C. Jarvis, and P. Owen. 1998. Nitrous oxide emissions from excreta applied in a simulated grazing pattern. Soil Biol. Biochem. 30(4): 491 500. Yamulki, S., S.C. Jarvis, and P. Owen. 1 999. Methane emission and uptake from soils as influenced by excreta deposition from grazing animals. J. Environ. Qual. 28: 676 682. Yang, J., E. Worley, M. Wang, B. Lahner, D.E. Salt, M. Saha, and M. Udvardi. 2009. Natural variation for nutrient use and remobilization efficiencies in switchgrass. Bioenergy Res. 2(4):257 266. Zhang, C., S. Li, L. Zhang, X. Xin, and X. Liu. 2013. Litter mixing significantly affects decomposition in the Hulun Buir meadow steppe of Inner Mongolia, China. J. Plant Ecol. 7(1): 56 67. Zhong, Z., R.L. Lemke, and L.M. Nelson. 2009. Nitrous oxide emissions associated with nitrogen fixation by grain legumes. Soil Biol. Biochem. 41: 2283 2291.

PAGE 145

145 BIOGRAPHICAL SKETCH Marta Moura Kohmann graduated with a degree in Agronomy from the Federal University of Rio Grande do Sul, Brazil in August 2011. During her undergraduate studies, she developed research work in the areas of Agrometeorology and Grazing Ecology. In August 2011, she started her M.Sc. degree in Agricultural and Biologic al Engineering at the University of Florida, working in the area of greenhouse gas emissions from beef production. She was awarded her M.Sc. degree in December of 2013. In May 2014, she started her Ph.D. program in the Agronomy Department at the University of Florida, working in evaluating nutrient cycling and emissions of greenhouse gases from legume grass relative to N fertilizer grazing systems. She co mpleted her Ph.D. in December 2017.