Greenhouse gas emissions from beef cattle grazing systems in Florida

MISSING IMAGE

Material Information

Title:
Greenhouse gas emissions from beef cattle grazing systems in Florida
Physical Description:
1 online resource (121 p.)
Language:
english
Creator:
Kohmann, Marta Moura
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Agricultural and Biological Engineering
Committee Chair:
FRAISSE,CLYDE WILLIAM
Committee Co-Chair:
SOLLENBERGER,LYNN E
Committee Members:
ADESOGAN,ADEGBOLA TOLULOPE
DILORENZO,NICOLAS
ASSENG,SENTHOLD

Subjects

Subjects / Keywords:
carbon -- methane -- sensitivity -- sf6
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre:
Agricultural and Biological Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Pastoral systems are of crucial importance for the economy of Florida. The cattle industry in this state is composed mostly of cow-calf operations that rely heavily on grazing systems using tropical grass species such as bahiagrass (Paspalum notatum Flugge). Agricultural operations are also an important source of greenhouse gas (GHG) emissions.Although the scientific community, governmental organizations and public opinion have increased their interest in environmental issues related to the production of food, little is known about the emission of GHG from beef cattle in Florida. The objectives of this study were to estimate GHG emissions (carbon footprint) from a typical low- input cow-calf operation in Florida, examine the model used for estimation of animal methane (CH4) production and measure animal CH4 production using the sulfur hexafluoride (SF6) tracer technique. The model developed by IPCC with emission factors specific for the USA or Florida was used when available from EPA. The greatest source of GHG in the cow- calf operation studied was from enteric fermentation followed by manure. A sensitivity analysis of the model used for estimating enteric CH4 production was performed with Morris, Fourier Amplitude Sensitivity Test and the vary- one- at- a- time methodologies. All analysis showed that average daily gain was the most important factor influencing the model’s output for growing animals independent of feed. A field experiment was carried out with three stocking rates (1.2, 2.4 and 3.6 AU ha-1) of animals grazing bahiagrass. Forage quantity and nutritive value were measured, as well as animal performance. Production of CH4 was measured using the SF6 technique. No effect of treatment was found in CH4 emissions or in other animal variables. Emissions averaged 393 g CH4 animal-1day-1. This was the first CH4 measurements for grazing beef cattle in Florida. Based on our results the IPCC Tier 2 and Tier 1 approaches seem to underestimate emissions of CH4 by grazing cattle,with values of 121 and 145 g CH4 animal-1 day-1,respectively. Due to the great importance of agriculture in Florida’s economy,it is essential to obtain more information about emissions from different agriculture-related sources. This information may help not only to improve model use but also to provide a better understanding of alternative management approaches that can reduce or avoid GHG emissions.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Marta Moura Kohmann.
Thesis:
Thesis (M.S.)--University of Florida, 2013.
Local:
Adviser: FRAISSE,CLYDE WILLIAM.
Local:
Co-adviser: SOLLENBERGER,LYNN E.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-06-30

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0046399:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EUJOS88YN_PBXHLW INGEST_TIME 2014-05-13T21:27:45Z PACKAGE UFE0046399_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

4 GREENHOUSE GAS EMISSIONS FROM BEEF CATTLE GRAZING SYSTEMS IN FLORIDA By MARTA MOURA KOHMANN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Marta Moura Kohmann

PAGE 3

3 To God, the best Agricultural Engineer I know, and to my family

PAGE 4

4 ACKNOWLEDGMENTS I would like to start thanking God for giving me a very long list of people for which to be thank ful First of all, I would like to thank my wonderful family for the love and en couragement they gave me even from a distance. Thanks to my parents for helping with words of wisdom and inspiration and for helping me keep my feet on the ground. To my siblings, thank you for listening to me and making me laugh. To the Turma in Por to Alegre, thank you for the unconditional support. Second, I would like to thank Dr. Clyde W. Fraisse, my committee chair, for the outstanding guidance program and for being an example of how leadership, innovation and interdisciplinar y work can achieve great accomplishments. I would like to express my deep appreciation for the crucial help of Dr. DiLorenzo and Dr. Sollenberger in designing and performing the field experiment. I would also like to thank Dr. Adesogan and Dr. Asseng for the insights and direction through this process. My most sincere gratitude is extended to Francine Messias Ciraco Silva Darren D. Henry, Martin Ruiz Moreno Mi quias Barbosa and to everyone at the North Florida Research and Education Center Marianna for the ir help with the field and laboratory measurements. Thanks for keeping good spirits in the hot Florida summer! I would also like to thank the AgroClimate Team (Verona Oliveira Montone, Eduardo Gelcer, Jos Henrique Debastiani Andreis, Tiago Zortea, Ana Paula Luz Wagner, Carlos Torres, Hermes Cuervo and Daniel R. Dourte). We faced many challenges together and I am glad I had your friendship to encourage me. I would like to thank my gaucho family (Marcelo Osrio Wallau, Alisson Pacheco Kovaleski and Bruno Casamali), for being available to share companionship and mate. I am also very grateful to my American family: to Monica Moss, Cameron and Bryant atitude

PAGE 5

5 to UFRGS for providing m e gr eat, free education. I wo uld like to thank particularly Dr. Moacir Antonio Berlato (and the Agromet group) and Dr. Paulo C sar de Faccio Carvalho (and the Grupo de Estudo de Ecologia em Pastejo) for first introducing me to research and for inspiring me to always ask questions and work hard.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 10 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 CARBON FOOTPRINT ESTIMATION ................................ ................................ .... 15 Literature Review ................................ ................................ ................................ .... 15 Materials and Methods ................................ ................................ ............................ 18 Site Description ................................ ................................ ................................ 18 Identification of GHG Sources ................................ ................................ .......... 19 Equations ................................ ................................ ................................ ......... 21 Enteric fermentation ................................ ................................ ................... 21 Manure ................................ ................................ ................................ ....... 24 Burning of pasture ................................ ................................ ...................... 28 Nitrogen fertilizer ................................ ................................ ........................ 28 Lime ................................ ................................ ................................ ........... 30 Emissions during production, transportation, storage and transfer ............ 31 Feed concentrate ................................ ................................ ....................... 31 Fuel ................................ ................................ ................................ ............ 32 Results ................................ ................................ ................................ .................... 3 2 Discussion ................................ ................................ ................................ .............. 33 Co nclusions ................................ ................................ ................................ ............ 37 2 SENSITIVITY ANALYSIS OF ENTERIC FERMENTATION EMISSION MODEL ................................ ................................ ................................ ................... 49 Literature Review ................................ ................................ ................................ .... 49 Materials and Methods ................................ ................................ ............................ 53 Model ................................ ................................ ................................ ................ 53 Data Source ................................ ................................ ................................ ..... 56 Vary one at a time (OAT) ................................ ................................ ................. 57 Morris ................................ ................................ ................................ ............... 57 FAST ................................ ................................ ................................ ................ 59 Results ................................ ................................ ................................ .................... 60 Discussion ................................ ................................ ................................ .............. 61 Conclusions ................................ ................................ ................................ ............ 67 3 RUMINAL METHANE EMISSIONS, FORAGE AND ANIMAL PERFORMANCE RESPONSES TO DIFFERENT STOCKING RATES ON CONTINUOUSLY STOCKED BAHIAGRASS PASTURES ................................ ..... 80 Literature Review ................................ ................................ ................................ .... 80

PAGE 7

7 Materials and Methods ................................ ................................ ............................ 88 Experi mental Site ................................ ................................ ............................. 88 Treatments and Design ................................ ................................ .................... 89 Pasture and Animal Management ................................ ................................ .... 89 Ruminal Methane Emissions Measurements ................................ ................... 92 Statistical Analysis ................................ ................................ ............................ 94 Results ................................ ................................ ................................ .................... 94 Discussion ................................ ................................ ................................ .............. 94 Conclusions ................................ ................................ ................................ ............ 99 LIST OF REFERENCES ................................ ................................ ............................. 112 BIOGRAPH ICAL SKETCH ................................ ................................ .......................... 121

PAGE 8

8 LIST OF TABLES Table page 1 1 Concentration of GHG in the atmosphere before the Industrial Revolution and in 2005. Source: IPCC, 2007a. ................................ ................................ .... 38 1 2 Emission of greenhouse gases by economic sector, million metric tons CO 2 e. Source: EPA, 2013a. ................................ ................................ ............... 39 1 3 Emissions of GHG Agriculture in the USA (Tg CO 2 e year 1 ), 1990 to 2011. Source: EPA, 2013a. ................................ ................................ .......................... 40 1 4 Herd information from Buck Island Ranch (BIR), period 1998 to 2008. .............. 41 1 5 Area burned (ha) on Buck Island Ranch (BIR), period between 1998 and 2008. ................................ ................................ ................................ .................. 41 1 6 Average above ground biomass available for burning in Buck Island Ranch (BIR), average from period 1998 to 2008. ................................ ............... 41 1 7 Lime, fertilizer, molasses, feed concentrate and fuel used at Buck Island Ranch (BIR), 1998 to 2008. ................................ ................................ ................ 42 1 8 Source categories and GHG emitted in the production system at in Buck Island Ranch (BIR), 1998 to 2008. ................................ ................................ ..... 43 1 9 Glob al Warming Potential (GWP) of GHG. ................................ ......................... 43 1 10 Data and emissions factor values, units and sources. ................................ ........ 44 1 11 Data and emissions factor values, units and sources (continuation). ................. 45 2 1 Default values for parameters not evaluated in sensitivity analysis. ................... 68 2 2 Digestible energy (DE, %) and methane conversion rate (Ym, %) values and sources used in the sensitivity analyses. ................................ ..................... 69 2 3 Average daily gain (ADG, kg/day) values and sources used in the sensitivity analyses. ................................ ................................ ............................ 69 2 4 OAT, FAST and Morris indexes for the enteric fermentation emission model (IPCC, 2006). ADG = average daily gain, kg animal 1 day 1 ; DE = digestible energy, %; Ym = methane conversion rate, %. ................................ .. 70 2 5 fermentation model. ................................ ................................ ............................ 70 3 1 Soil types in experimental sites A and B. Source: USDA, 2013. ....................... 101 3 2 Herbage mass double sample regression equations. Period 1: 25 June to 22 July; Period 2: 23 July to 19 Aug.; Period 3: 20 Aug to 18 Sep. .................. 101

PAGE 9

9 3 3 Effect of stocking rate (1.2, 2.4 and 3.6 AU ha 1 ) on response variables in 2012 period and treatment x period interaction on experimental variables. ...... 101 3 4 Forage herbage mass (HM) (kg ha 1 ) and herbage accumulation rate ( HAR) (kg ha 1 day 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ ................................ ....... 102 3 5 Forage chemical composition response given by CP (%), NDF (%) and ADF(%) to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ ................................ ................................ ............ 102 3 6 Average daily gain (ADG, kg animal 1 day 1 ) BW (kg animal 1 ), DMI (kg animal 1 day 1 ) and DMI (kg BWl 1 day 1 ) response to three stocking rates (1.2, 2.4, 3.6 AU ha 1 ) in 2012. Period 1 : June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ ..................... 103 3 7 Response of CH 4 emissions expressed as g CH 4 day 1 g kg BW 1 and g kg DMI 1 to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ ................................ ................................ ............ 104

PAGE 10

10 LIST OF FIGURES Figure page 1 1 Map of the state of Florida. "A" refers to the location of Buck Island Ranch (BIR). ................................ ................................ ................................ .................. 46 1 2 Cale ndar of animals' reproductive stage. ................................ ............................ 46 1 3 Emissions from BIR, ton CO 2 e year 1 ................................ ................................ 47 1 4 GHG emissions from BIR per category, % over average of all years, 1998 to 2008. ................................ ................................ ................................ .............. 48 1 5 Average GHG emissions from synthetic N fertilizer and lime before and after they are applied on the farm. ................................ ................................ ...... 48 2 1 OAT experimental design used in the sensitivity analysis. ................................ 71 2 2 Evolution of the experimental design in Morris sensitivity analysis (a ................... 72 2 3 Experimental design in the FAST method for animals on pasture and on feedlot. ADG= average daily gain, kg animal 1 day 1 ; DE= digestible energy, %; Ym= methane conversion rate, %. ................................ ................... 73 2 4 Experimental design in the FAST sensitivity analysis. ADG = average daily gain, kg animal 1 day 1 ; DE = digestible energy, %; Ym = methane conversion rate, %. ................................ ................................ ............................. 74 2 5 Output of enteric fermentation methane emission model (IPCC, 2006) in kg CH 4 animal 1 year 1 as a function of parameter values in the OAT sensitivity anal ysis method. ................................ ................................ ................ 75 2 6 OAT index for sensitivity analysis of simulations on pasture and feedlot situations. ................................ ................................ ................................ ........... 76 2 7 FAST indexes for the enteric fermentation emission model (IPCC, 2006). ADG = average daily gain, kg animal 1 day 1 ; DE = digestible energy, %; Ym = methane conversion rate, %. ................................ ................................ .... 77 2 8 Morris indexes for the enteric fermentation emission model (IPCC, 2006). ADG = average daily g ain, kg animal 1 day 1 ; DE = digestible energy, %; Ym = methane conversion rate, %. ................................ ................................ .... 78 2 9 Energy partitioning in animals. Adapt from: (Minson, 1990, Van Soest, 1982). ................................ ................................ ................................ ................. 79 2 10 Relationship between ADG (kg animal 1 day 1 ) and methane production (g CH 4 ADG 1 ) for simulations made with Tier 2 enteric fermentation model (IPCC, 2006) for animals on pasture and feedlot. ................................ ............... 79

PAGE 11

11 3 1 Map of experim ental sites A and B located at the North Florida Research and Education Center (NFREC), Marianna, Florida. ................................ ........ 105 3 2 Animal with CH 4 collection device. A: capillary tube placed on halter; B: collecting canister. Picture by Marta Moura Kohmann, 2012. ........................... 105 3 3 Herbage mass (kg ha 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Pe riod 3: Aug.20 to Sep. 18. ................................ ................................ ....... 106 3 4 Herbage accumulation rate (HAR) (kg ha 1 day 1 ) i response to three stocking rates (1.2, 2.4 a nd 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug. 20 to Sep. 18. ................ 106 3 5 Dry ma tter intake (DMI) (kg animal 1 day 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ............................... 107 3 6 Dry matter intake (DMI) per kg body weight (BW) (kg kg 1 day 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ........ 107 3 7 Average daily gain (ADG) (kg animal 1 day 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ............................... 108 3 8 Acid detergent fiber (ADF) (%) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ .......................... 108 3 9 Neutral detergent fiber (NDF) (%) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ ..................... 109 3 10 Crude protein (CP) (%) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ ................................ ....... 109 3 11 Methane production (g C H 4 animal 1 day 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ............................... 110 3 12 Methane production (g CH 4 kg BW 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 t o Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ .............. 110 3 13 Methane production (g CH 4 kg DMI 1 ) response to three stocking rates (1.2, 2.4 an d 3.6 AU ha 1 ) in 2012. Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ .............. 111 3 14 Measured and simulated methane production (g CH4 animal 1 day 1 ) in three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. The value for

PAGE 12

12 IPCC Tier 1 is a default value for beef cattle in the United States (IPCC, 2006), while values for IPCC Tier 2 were s 2 methodology (IPCC, 2006) using the average daily gain from Periods 1 and 3 in the field experiment presented in this chapter (Table 3 6). Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; Period 3: Aug.20 to Sep. 18. ................................ ................................ ................................ ............ 111

PAGE 13

13 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science GREENHOUSE GAS EMISSIONS FROM BEEF CATTLE GRAZING SYSTEMS IN FLORIDA By Marta Moura Kohmann December 2013 Chair: Clyde W. Fraisse Major: Agricultural and Biological Engineer ing Pastoral systems are of crucial importance for the economy of Florida. The cattle industry in this s tate is composed mostly of cow calf operation s that rely heavily on grazing systems using tropical grass species such as bahiagrass ( Paspalum notatum Flugg e). Agricultural operations are also an important source of greenhouse gas (GHG) emissions. Although the scientific community, governmental organizations and public opinion have increased their interest in environmental issues related to the production of food, little is known about the emission of GHG from beef cattle in Florida. The objective s of this study were to estimate GHG emissions (carbon footprint) from a typical low input cow calf operation in Florida, examine the model used for estimation of an imal methane (CH 4 ) production and measure animal CH 4 production using the sulfur hexafluoride (SF 6 ) tracer technique The model developed by IPCC with emission factors specific for the USA or Florida was used when available from EPA. The greatest source of GHG in the cow c alf operation studied was from enteric fermentation followed by manure. A sensitivity analysis of the model used for estimating ente ric CH 4 production was performed with Morris Fourier Amplitude Sensitivity Test and th e vary one at a time methodologies. All analysis showed that average daily gain was the most important

PAGE 14

14 experiment was carried out with three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) of animals grazing bahiagrass. Forage quantity and nutritive value were measured, as well as animal performance. Production of CH 4 was measured using the SF 6 technique. No effect of treatment was found i n CH 4 emissions or i n other animal variables. Emissions averaged 39 3 g CH 4 animal 1 day 1 This was the first CH 4 measurements for grazing beef cattle in Florida. Based on our results t he IPCC Tier 2 and Tier 1 approaches seem to underestimate emissions of CH 4 by grazing cattle, with values of 121 and 145 g CH 4 animal 1 day 1 respectively. Due to the great essential to obtain more information about emissions from different agriculture related sources. This information may help not only t o improve model use but also to provide a better understanding of alternative management approaches that can reduce or avoid GHG emissions.

PAGE 15

15 CHAPTER 1 CARBON FOOTPRINT ESTIMATION Literature Review Changes in climate have become increasing ly importan t to s ociety. In 1998, the United Nations Environmental Programme (UNEP) and the World Meteorological Organization (WMO) instituted a scientific body responsible for reviewing scientific, social, economic and technical information regarding climate change, call ed Intergovernmental Panel on Climate Change (IPCC). Thi s scientific body is composed of scientists from 195 countries and, because of its intergovernmental character, produces policy neutral report s Among its publications, IPCC produces reports to explai n the scientific ba s is of changes in climate, provide information on climate change ri sk management and adaptation strategies, and establish guidelines for estimating g reenhous e g as (GHG) emissions (IPCC, 2013) In the USA, these guidelines are used by the Environmental Protection Agency ( EPA) to estimate GHG emissions o n a national scale. According to IPCC (2007a) climate change exists when there is a statistical decade or more, whether it be natural or originate from anthropogenic actions. There are several factors pointing 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, 2007a) Some of the climate modifying agents includes greenhouse gases (GHG), aerosols, solar radiation and surface cover. These factors change the energy balance positively or negatively and the intensity of this modification is called radiative forcing, measured in W m 2 A nthrop ogen ic activities can be a source of GHG such as carbon dioxide (CO 2 ), methane (CH 4 ), nitrous oxide (N 2 O) and

PAGE 16

16 halocarbons The concentrations of the first three before the Industrial Revolution and in 2005 are presented in Table 1 1 where we can observe a significant increase in their c oncentration in the atmosphere. Anthropogenic activities are estimated to have had a positive radiative forcing in the order of 0.6 to 2.4 W m 2 since 1750 (I PCC, 2007a) There are several sources of GHG related to human activities. In 2011, the USA emitted 6,702 million metric tons of CO 2 equivalent. C onsidering the carbon sequestration occurring in land use, land use change and forestry in the US, net emi ssion was 5,797 million metric tons of CO 2 equivalent in 2011 These emissions were composed of different GHG, including 83.7% CO 2 8.8 % CH 4 5.3 % N 2 O and 2.2% HFC`s, PFC`s and SF 6 Several economic sectors contribute to the GHG emissions in the USA ( Table 1 2 ). In 2011, a griculture contributed 6.9% of total emissions in the USA (EPA, 2013a) Agriculture is an important economic activity in Florida. In 2010, agricultural products in Florida were wo rth 7.81 billion dollars, and cow calf operations alone accounted for 6.4% of this value In fact, in January 2012 all cattle and calves in Florida totaled 1,700,000 head, from which 940,000 were beef cattle. In 2011, 890,000 calves were born in Florida (F lorida Department of Agriculture and Consumer Services,2012) I n this context, it is clear that assessing GHG emissions from the beef industry in Florida is important and the use of models is of particular importance when analyzing production systems at b oth a farm scale (Beauchemin et al., 2010) and at larger scales (Storm et al., 2012) Many studies have been conducted regarding estimating GHG emissions from the processes involved in the production of various produces. In fact, when searching for Wiedman

PAGE 17

17 and Minx (2008) there were 42 hits of which 31 happened in 2007. Studying the number of publications on Life Cycle Assessment (LCA) applied to agriculture products between 2001 and 2011, Ruviaro et al. ( 2012) found a remarkable increase in the number of studies produced particularly after 2007 probably related to governmental and public inquiries regarding anthro pogenic influence o n global GHG emissions. The approach used to calculate the carbon footprint c an vary greatly. Although some authors define carbon footprint as the amount of CO 2 solely emitted directly and indirectly by an activity or through life stages of products (Wiedman and Minx, 2008) the carbon footprint calculation for agriculture products usually accounts for all of the different GHG involved in their production and transforms them into CO 2 equivalent (CO 2 e) according to each gas global warming potential (GWP) (Rs et al., 2013) These estimations are important to identify the main sources of GHG in a production system and point to possible areas o f emission mitigation, as well as energy use inefficiency (Lash and Wellington, 2007) With the increase of concern regarding environment al conservation, having to access environmental records also present s a competitive advantage and can drive purchase decisions (Lash and Wellington, 2007) Considering the importance of cow calf production in Florida and the necessity to account for the GHG emissions related to this production system in order to satisfy governmental and soc responsibility, the objective of this study was to quantify the carbon footprint of a cow calf operation in Florida.

PAGE 18

18 Materials and Methods Site D escription Buck Island Ranch (BIR) is located in Lake Placid, sou th ce ntral Florida, northwest of Lake Okeechobee ( Figure 1 1 ). It has been managed by the Archbold Biological Station since 1988 and, as a division of Archbold Expeditions, the MacArthur Agro eco logy Research Center operates at BIR where it promotes long term ecological research It also runs a commercial cow calf operation on an area of 4,200 h a, approximately 3000 Braham cows and 250 Angus bulls ( Table 1 4 ). Management makes use of natural service during 5 to 7 months of the year (usually from January to May, but the breeding season can be extended until June or July). Calves are born during th e period of November to March and ar e sold or transferred to other s tates at 7 months of age ( Figure 1 2 ). Approximately half of the grazing area is planted to bahiag rass, while the other half i s in semi native pasture. The bahiagrass is occasionally managed with burning. According to Kottek et al. ( 2006) the climate in the region is classified as Cfa. In this classification Cf stands for climates with warm temperatures and fully humid, with minimum temperatures betwee n 3 and 18 C. The a refers to hot summers with maximum temperatures above 22 C. There are two soil types occurring the ranch They are classified as Felda fine s and which is subjected to frequent flooding and Ona l oamy s and where gro und water levels v ary between 25 and 10 0 cm throughout the year (USDA, 1989) Production records from 1998 to 200 8 were considered in this case study. According to the IPCC (2006) guidelines, GHG emissions can be estimated using diff erent levels of data and detail The methodology used can be classified as Tier 1,

PAGE 19

19 Tier 2 and Tier 3 to characterize increasing level of information needed to estimate GHG emissions and accuracy of the predictions The higher the Tier used, the smaller the uncertainty. Livestock categories to be considered in a GHG emission inventory should include all of those which are importan t to a country or region. Categ ory is described as the animal species and it can be separated into subcategories according to age, type of production and gender (IPCC, 2006) In this study the livestock category considered was cattle, with three subcategor ies: cows, bulls and calves. Identification of GHG S ources The first step in calculating the carbon footprint of a production system involves identifying the sources of GHG involved in the process. The source categories and GHG emission considered in the c alculation were identified according to IPCC (2006) while considering specific emission factors available for Florida or the USA in EPA (2013 a; 2013b ) The sources and GHG emitted at BIR in the period of 1998 to 2008 are delineated i n Table 1 8 A brief description follows. Enteric f ermentation. Enteric fermentation refers to a digestive process where the ferment s fee d and produce s CH 4 as a by product. Ruminant livestock, including cattle, sheep and goats, have greater rates of enteric fermentation because of their unique digestive system, which includes a large rumen or fore stomach where enteric fermentation takes place. P roduction of CH 4 depends o digestive system and quality and quantity of food (EPA, 2013a) IPCC (2006) only considers emissions from animals older than 7 months of age. Livestock waste. Manure can be managed in storage or treatment sys tems or spread on fields in lieu of long term storage or it can be deposited directly on

PAGE 20

20 grazed lands. The management of livestock manure can produce CH 4 and N 2 O Production of CH 4 is a natural process in anaerobic decomposition of livestock manure (EPA, 2013a) E missions of N 2 O from lives tock waste depend on the composition of manure and urine, the type of bacteria involved in the process and the amount of oxygen and water in the manure system. Direct N 2 O emissions are produced as part of the N cycle through nitrification and denitrificati on of the organic N in livestock manure or urine. Indirect N 2 O emissions are produced as result of the volatilization of N as ammonia (NH 3 ) and oxides of nitrogen (NO x ) and runoff and leaching of N during treatment, storage, and transportation (IPCC, 2006) Pasture burning. Improved and native pastures in BIR are burned primarily between December and February according to a burning schedule Burns may sometime s occur as lat e in the year as April in order to enhance biological diversity including endangered and threatened species, reduce fire hazards, mimic natural processes and provide educational and research opportunities (Main and Menges, 1997) When burning fields, CO 2 is not considered to be released since it is largely balanced by the CO 2 that is reincorporated back into biom ass via photosynthetic activity within weeks or a few years after burning. Non CO 2 emissions, particula rly carbon monoxide (CO), CH 4 N 2 O and other kinds of nitrogen oxides (NO x ) that result from incomplete combustion of biomass in managed grassland are reported (IPCC, 2006) However, since there is no agreement regarding the GWP value and signal for NO x t his gas was not considered in the evaluation (IPCC, 2006). Fertilization (with synthetic fertilizer) and liming of pasture s and crops. Emissions from these processes include those from manufacturing, storage, transfer and transportation as well as direct emissions of fertilizers and lime applied to the field. The methodology used for accounting for manufacturing, storage and

PAGE 21

21 transportation of fertilizers and lime was developed by Lal (2004) It estimates carbon equivalent emissions by converting the energy or volume involved in the production processes to kg of carbon equivalent (kg CE). After application to the field, f ertilizers also release N 2 O directly through the microbial processes of nitrification and denitrification or indire ctly by volatilization of ammonium (NH 4 + ) and nitrate (NO 3 ) or leaching and runoff mainly of nitrate, which can later go through nitrification and denitrification (IPCC, 2006). Tractor o perations. This category includes all activities that require tractor operations such as tilling, planting, harvesting, and application of agrochemicals. These processes mostly release CO 2 but also release CH 4 a nd N 2 O However, for this paper the methodology used is that develope d by the EPA (2005) which considered only CO 2 emissions based on the amount of carbon present in fuels. Equations Enteric fermentation The first step necessary to calculate emissions from enteric fe rmentation is to determine the different population subcategories. The total number of cows, pregnancy rate, the number of calves and of bulls was obtained from the BIR database. For the seven months after calving, we assumed that the number of lactating c ows equal ed t he number of calves. T he cows pregnant in the next breeding season were considered to be lactating. Remaining cows were considered to be neither pregnant nor lactating. Therefore, there were five sub categories of animals considered to emit CH 4 from enteric fermentation: pregnant cows, lactating cows, cows that were both pregnant and lactating, cows that were neither pregnant nor lactating and bulls. Calves are not considered to emit CH 4 from enteric fermentation (IPCC, 2006).

PAGE 22

22 Gross Energy (GE) is the energy the animal needs for maintenance and to perform activities such as lactation. It was considered that all animals were adults and did not perform any work (for example, pulling plows for working the soil, etc.). Therefore, net energy for work and growth were considered zero and REG was not calculated. GE = [( ) + ( )] / ( ) where GE= gross energy, MJ day 1 NEm= n et energy required by the animal for maintenance, MJ day 1 NEa= n et energy for animal activity, MJ day 1 NEl= n et energy for lactation, MJ day 1 NEwork= n et energy for work, MJ day 1 NEp= n et energy required for pregnancy, MJ day 1 REM= ratio of net energy available in a diet for maintenance to digestible consumed NEg= n et energy needed for growth, MJ day 1 REG= r atio of n et energy available for growth in a diet to digestible energy consumed DE= digestible energy expressed as a percent of gross energy (percent) Net energy for work and growth were considered zero. Formulas used to calculate each of th e factors involved in GE calculation are as follows: NEm= Cfi x (weight) 0.75 where NEm= n et energy required for the animal for maintenance, MJ day 1 Cfi= coefficient, varies for each animal category, MJ day 1 kg 1 Weight= live weight of animal, in kg

PAGE 23

23 NEl= Milk x (1.47+0.40 x Fat) where NEl= n et energy for lactation, MJ day 1 ; Milk= amount of milk produced, (kg of milk) day 1 ; Fat= fat content of milk, % by weight NEp= C pregnancy x NEm where: NEp= n et energy required for pregnancy, MJ/day 1 ; C pregnancy = pregnancy coefficient; NEm= n et energy required for the animal for mantainance, MJ day 1 The NEm used here was that considering pregnancy. NEa= Ca x NEm where NEa= n et energy for animal activity, MJ day 1 ; Ca= activity coefficient corresponding to the dimen sionless ; NEm= n et energy required for the animal for maintenance, MJ day 1. REM= [1.123 (4.092 x 10 3 x DE%) + [1.126 x 10 5 x (DE%) 2 ] ( )] where REM= ratio of net energy available in a diet for maintenance to digestible consumed; DE= digestible energy expressed as a percentage of gross energy. After the calculation of GE a daily emission factor for each category was calculated with the formula below.

PAGE 24

24 DayEmit = [ ] where DayEmit = emission factor, kg CH 4 1 head 1 day 1 ; GE = gross energy intake, MJ day 1 ; Ym = CH 4 conversion rate, which is the fraction of gross energy in feed converted to CH 4 (%); 55.65 = a factor for the energy content of CH 4 MJ (kg CH 4 ) 1 To determine yearly emissions for each category, the formula below was used. Emissions = DayEmit x 365 where Emissions = total emissions in a month for the category, kg CH 4 year 1 ; DayEmit = emission factor for the category, kg CH 4 head 1 day 1 ; 365 = day s in the year. Emissions from enteric fermentation Emissions = EF 1 x N where Emissions = CH 4 emissions from enteric fermentation, kg CH 4 year 1 ; EF 1 = emission factor for the defined population, kg CH 4 head 1 year 1 ; N = number of animals in the subcategory. Manure Manure is a source of both N 2 O and CH 4 Methane Emission of CH 4 from manure was calculated with the equations below:

PAGE 25

25 EF(T) = VS (T) x [ Bo (T) S,k /100) x MS (T,S,k) ] where EF (T) = annual CH 4 emission factor for category T, kg CH 4 animal 1 year 1 ; VS (T) = daily volatile solid excreted for category T, kg dry matter animal 1 year 1 ; for calves, 210 days (7 months) were considered; Bo (T) = maximum CH 4 producing capacity for manure procedure by l ivestock category T, m 3 CH 4 (kg of VS excreted) 1 ; 0.67 = conversion factor of m 3 CH 4 to kilogram CH 4 ; MCF S,k = CH 4 conversion factors for each manure management system S by climate region k, %; MS (T,S,k) management system S in climate region k, dimensionless. Nitrous oxide N 2 O direct emissions Direct N 2 O produced by the manure was calculated with the equations below. N 2 Odirect = N 2 O N PR = F PRP x EF 3 PRP where N 2 Odirect N = annual direct N 2 O N emissions produced from livestock waste, kg N 2 O N year 1 ; N 2 O NPR = annual direct N 2 O N emissions from urine and dung inputs to grazed soils, kg N 2 O N year 1 ; F PRP = annual amount of urine and dung N deposited by grazing animals on pasture, range and paddock, kg N year 1 ; EF 3 PRP = emission factor for N 2 O emissions from urine and dung N deposited on pasture, range and paddock by grazing animals, kg N 2 O N (kg N input) 1 ;

PAGE 26

26 F PRP = [(N (T) x Nex (T) ) x MS (T,P RP) ] where N (T) = number of head of livestock species (category T) 1 ; N ex(T) = annual average N excretion per head of species (category T) 1 (see Table 1 7 ) MS (T, PRP) = fraction of total annual N excretion for each livestock species (category T) 1 that is deposited on pasture, range and paddock. It was considered 1 in this case because all the excretion was deposited on pasture. N ex(T) = E N x weight x days where N ex(T) = annual average N excretion per head of species (category T) 1 ; E N = excretion of nitrogen, kg day 1 (1000 kg) 1 ; weight = average animal weight during the period, kg; days = number of days spent in the farm 365 for cows and bulls and 210 for calves. To convert the results from kg N 2 Odirect N to kg N 2 O, the following formula was used: N 2 O = N 2 O Ndirect x (44/28) N 2 O indirect emissions Indirect emissions of N 2 O have two sources calculated separately: volatilization, and leaching and runoff. To estimate volatilization, the following equation is used: N 2 O (ATD) N = [(F ON + F PRP ) x Frac GASM )] x EF 4 where: N 2 O (ATD) N = annual amount of N 2 O N produced from atmospheric deposition of N volatilized from managed soils, kg N 2 O N year 1 ;

PAGE 27

27 F ON = annual amount of managed animal manure, compost, sewage sludge and other organic N additions applied to soils, kg N year 1 considered zero in this situation; F PRP = annual amount of urine and d ung N deposited by grazing animals on pasture, range and paddock, kg N year 1 Same as the one used to calculate direct N 2 O em issions from manure management ; Frac GASM = fraction of applied organic N fertilizer material (F ON ) and of urine and dung N deposit ed by grazing animals (F PRP ) that volatilizes as NH 4 and NO 4 kg N volatilized (kg of N applied or deposited) 1 ; EF 4 = emission factor for N 2 O emissions from atmospheric deposition of N on soils and water surfaces, kg N 2 O N (kg NH 3 N + NO x N) 1 volatili zed. To estimate leaching and runoff, the following equation was used: N 2 O (L) N = [(F ON + F PRP ) x Frac LEACH (H) )] x EF 5 where N 2 O (L) N = annual amount N 2 O N produced from leaching and runoff of N additions to managed soils in regions where leaching/run off occurs, kg N 2 O N year 1 ; Frac LEACH (H) = fraction of all N added in regions where leaching/runoff occurs that is lost through leaching and runoff, kg N (kg on N added) 1 ; EF 5 = emission factor for N 2 O emissions from N leaching and runoff, kg N 2 O N (kg N leached and runoff) 1 After that, both N from volatilization and from leaching and runoff are summed, as showed below. N 2 Oindirect N = N 2 O (L) N + N2O (ATD) N where

PAGE 28

28 N 2 Oindirect N = annual indirect N 2 O N emissions from urine and dung inputs to graz ed soils, kg N 2 O N year 1 To convert the results from kg N 2 Oindirect N to kg N 2 O, the following formula was used: N 2 O = N 2 O Nindirect x (44/28) Burning of p asture The equation used to estimate emissions from pas ture burning is shown below. L fire = A x M B x C f x G ef x 10 3 where L fire = amount of GHG emissions from fire, tones of each GHG. Indirect GWP for CO was considered 1.9 (IPCC, 2006) ; A= area burnt, ha ( Table 1 5 ); M B = mass of fuel available for combustion, ton ha 1 It includes biomass, ground litter and dead wood, but when using the Tier 1 method ground litter and dead wood are considered zero except when there is land use change. The average of above ground mass available for all native or all improved pastures in a specific month was used ( Table 1 6 ). C f = combustion factor, dimensionless; G ef = emission factor, g (kg dry matter burnt) 1 Nitrogen f ertilizer After nitrogen fertilizers are applied to the soil, they release N 2 O directly and indirectly, with the same methodology used to calculate N 2 O emissions from animal waste. Direct emissions

PAGE 29

29 N 2 O direct N = N 2 O imput = F SN x EF 1 where N 2 O direct N = annual direct N 2 O N emissions produced from managed soils, kg N 2 O N year 1 ; N 2 O imput N = annual direct N 2 O N emissions from N inputs to managed soils, kg N 2 O N year 1 ; N 2 O direct N = annual direct N 2 O N emissions produced from managed soils, kg N 2 O N year 1 ; N 2 O imput N = annual direct N 2 O N emissions from N inputs to managed soils, kg N 2 O N year 1 ; F SN = annual amount of synthetic fertilizer N applied to soils, kg N year 1 ; EF 1 = emission factor for N 2 O emissions from N inputs, kg N 2 O N (kg N input) 1 Indirect emissions Similarly to manure management, indirect emis sions from synthetic N fertilizer app lication happened through volatilization, and leaching and runoff. To estimate vol atilization, the formula used was : N 2 O (ATD) N = F SN x Frac GASF x EF 4 where N 2 O (ATD) N = annual amount of N 2 O N produced from atmospheric deposition of N volatilized from m anaged soils, kg N 2 O N year 1 ; F SN = annual amount of synthetic fertilizer N applied to soils, kg N year 1 ; Frac GASF = fraction of synthetic fertilizer N that volatilizes as NH 3 and NO x (kg N volatilized) (kg N applied) 1 ; EF 4 = emission factor for N 2 O emissions from atmospheric deposition of N on soils and water surfaces, kg N 2 O N (kg NH 3 N + NO x N) 1 volatilized.

PAGE 30

30 To estimate leaching and runoff, the formula used wa s: N 2 O (L) N = F SN x Frac LEACH (H) x EF 5 where N 2 O (L) N = annual amount N 2 O N produced from leaching and runoff of N additions to managed soils in regions where leaching/runoff occurs, kg N 2 O N year 1 ; F SN = annual amount of synthetic fertilizer N applied to soils, kg N year 1 ; Frac LEACH (H) = fraction of all N added in regions wh ere leaching/runoff occurs that is lost through leaching and runoff, kg N (kg of N applied) 1 ; EF 5 = emission factor for N 2 O emissions from N leaching and runoff, kg N 2 O N (kg N leached and runoff) 1 To estimate total indirect emissions from synthetic fe rtilizers, emissions from leaching and runoff and volatilization were summed as show n below. N 2 Oindirect N = N 2 O (L) N + N2O (ATD) N where N 2 Oindirect N = annual indirect N 2 O N emissions from urine and dung inputs to grazed soils, kg N 2 O N year 1 To convert the results from kg N 2 Oindirect N to kg N 2 O, the following formula was used: N 2 O = N 2 O Nindirect x (44/28) Lime The equation used to estimate the emission from lime (dolomite in this case) af ter it was applied to the soil wa s: CO 2 C Emissions = (M dolomite x EF dolomite ) CO 2 C Emissions = annual carbon emissions from lime application, tones C year 1 M dolomite = annual amount of calcic dolomite, tons year 1 ;

PAGE 31

31 EF dolomite = emission factor, tons of CO 2 ( ton of dolomite lime) 1 Emissions d uring productio n, transportation, storage and transfer Off farm emissions are an important source of GHG M any steps are involved before agrochemicals arrive at a farm or ranch. Lal (2004a) developed a methodology to account for GHG emissions fro m production, transportation, storage and transfer of agrochemicals. In this case study, these emissions are those related to the use fertilizer and lime. For synthetic N fertilizer, the formula used was Carbon emission = F SN x Equivalent Carbon Emission x (44/28) / 1000 where Carbon emissi on = emissions, in kg CO 2 e year 1 ; F SN = annual amount of synthetic fertilizer N applied to soils, kg N year 1 ; Equivalent Carbon Emission = C emission in relation to production, packaging, storage and distribution of fertilizers, kg CE (kg N) 1 transfer, the equation used was Carbon emission = M dolomite x Equivalent Carbon Emission x (44/28) / 1000 where Carbon emission = emissions, in t ons CO 2 e year 1 ; M dolomite = annual amount of synthetic fertilizer N applied to soils, kg year 1 ; Equivalent Carbon Emission = C emission in relation to production, packaging, storage and distribution of fertilizers, kg CE (kg N) 1 Feed c oncentrate For feed c oncentrate, the value of 780 kg CO 2 e t 1 of feed concentrate according to Casey and Holden ( 2006) Amount of feed concentrate used is presented on Table 1 7

PAGE 32

32 Fuel The formula used to obtain CO 2 emissions from gasoline and diesel used o n BIR is shown below. It was considered that the emissions from molasses should be accounted for in sugarcane production and processing, so that emissions regarding the use of molasses are only th ose ones related to transportation. EmFuel = Fuel x E fuel where EmFuel= CO 2 e emissions, kg CO 2 year 1 ; Fuel= amount of fuel used, gallons year 1 ; E fuel = em ission factor, kg CO 2 e gallon 1 After emissions were calculated they were expressed as CO 2 equivalent (CO 2 e) emitted per unit of live weight produced. The CO 2 e is a measure used to compare the emissions from various greenhouse gases based upon their glob al warming potential. For example, the global warming potential for CH 4 over 100 years is 25. This means that emissions of one metric ton of CH 4 are equivalent to emissions of 25 metric tons of CO 2 (IPCC, 2007b) Emission factors and other data used for calculating the emissions above described are on Table 1 10 and Table 1 11 Production data including the use of fertilizer, lime and fuel are on Table 1 7 Results Results from the carbon footprint calculation from BIR are show n in Figure 1 3 and Figure 1 4 On average, annual emissions in BIR were of 10, 470 tons CO 2 e year 1 and ranged from 8,750 tons CO 2 e year 1 in 1999 and 12, 36 0 tons CO 2 e year 1 in 2004. The main source of GHG in this production system is enteric fermentation (55 %) fol lowed by animal was te (27 %), corresponding to 5,77 0 and 2, 79 0 tons

PAGE 33

33 CO 2 e year 1 respectively. Application of fertilizer and lime contribute on average with 11 % of total annual emissions in BIR. In evaluating the variation in amou nt of GHG emitted yearly at BIR ( Figure 1 3 ) i t is possible to determine that the amount of GHG from the main contributors to total emissions (enteric fermentation and animal waste) did not show great fluctuation from one year to the next. The variation observed, however, may be related to the use of management practices such as pasture burning and application of synthetic N fertilizer and lime. The emissions from the use of synthetic N fertilizer and lime can be separated in to on farm and pre farm emission ( Figure 1 5 ). E missions related to production, transportation, storage and transfer can be very important when considering the use of agrochemicals. In fact, when looking at the emissions from the use of N fertilizer a lone, 43 % of the GHG emissions occur red before its use. The proportion of emissions occurring during production, transportation, storage and transfer of lime is very si gnificant and accounted for 71 % of total em issions connected to its utilization. In a more intensively managed system where liming and application of fertilizers may be more frequent the pre farm GHG production should have greater importance than in the current case study. The objective of produ ction in BIR is weaned calves, which after 7 months of age are sold or transferred to other regions of the US. Expressing GHG emissions as a function of product originated in a production system can be a useful way to compare different production systems. On average, 22.1 kg CO 2 e (kg calf LW) 1 was emitted at BIR. Discussion In this case study, enteric fermentation was the largest contributor to total GHG emissions in a cow calf production system ( Figure 1 4 ). This is in accordance

PAGE 34

34 with other studies performed on GHG emissions from beef production systems Beauchemin et al. ( 2010) perform ed a Life Cycle Assessment (LCA) of beef production in western Canada and found that 63% of emissions had their source in enter ic fermentation. In the same study, CH 4 and N 2 O emissions from beef manure accounted for 28% of total GHG emissions. These results are very similar to those presented in this study where 55 % if emissions came from enteric fermentation and 27 % from manure Basarab et al. ( 2012) found that enteric ferme ntation was responsible for 53 to 54% of total emissions when evaluating the full production cycle of a beef herd When analyzing different beef production systems in Ireland, Foley et al. ( 2011) found that 46 to 5 3 % of total emissions came from enteric fermentation. It is important to notice that the emissions from enteric fermentation in this study are not directly related to the production of meat since it mainly considers the adult c ows. In fact, the productive cows in a full herd production cycle can account for up to 70% of total GHG emissions (Basarab et al., 2012) Other authors have also highli ghted the relevance of GHG emissions coming from animals in the cow calf production phase, which require high levels of feed for maintenance and low production relative to other production phases in the beef industry (Johnson et al., 2001a) Management strategies can be used to decrease GHG emissions. The use of growth implants, for example, can red uce the carbon footprint of beef production by 5 % (Basarab et al., 2012) Feed management can also strongly influence the carbon foo tprint Animals fed high concentrate diets may have less energy lost as CH 4 (Kurihara et al., 1999; Beauchemin and Mcginn, 2005) while low quality feed can result in higher emissions from enteric f ermentation (Phetteplace et al., 2001) However, Yan et al. ( 2010) emphasize d that at the farm level it is necessary to consider several aspects associated with maintaining high level production animals

PAGE 35

35 including emissions associated with soil management, feed production and use of fe rtilizers. This is an important component of more intensive ly managed systems. Although in this case study the amount of fertilizers and lime applied is not large, it is crucial to acknowledge the importance of GHG emissions occurring before application of these and other products. As Figure 1 5 shows, pre farm emissions can account for a considerabl e part of emissions related to the use of agrochemicals, particularly for lime. Although the intensification of management practices can increase total GHG emissions, it can also reduce emissions per unit of product. When analyzing production scenarios varying in management practices in Ireland, Foley et al. ( 2011) found that inputs required for higher production levels resulted in higher carbon footprint when expressed as t ons CO 2 e year 1 However, when expressed as kg CO 2 e (kg beef) 1 high level p roduction systems had a lower carbon footprint because of associated improvements in fertilizer use and animal growth. Basarab et al. ( 2012) highlight the fact that time is an i mportant factor when analyzing carbon footprint particularly regarding the comparison of different production systems. The authors affi rm that expressing results of carbon footprint as kg CO 2 e (kg beef) 1 can undere stimate the disparity between carbon footprint if a time correction is not made, since higher productivity can be associated with greater animal production per unit of time Another relevant aspect of pasture based production is the ability of the pasture to fix atmospheric carbon (Soussana et al., 2004) and this was not considered in this simulation. The use of pasture as animal feed has several environmental advantages bes ides potentially reducing the carbon footprint of beef production, including the decrease in e missions from manure and return of nutrients to the soil and making use of the to convert high fiber material into

PAGE 36

36 high quality protein (Beauchemin et al., 2010; Pelletier et al., 2010) If considering carbon sequestration from pastures, carbon loss in annual cropping and land used i n the produc tion of hay the carbon footprint in a beef production system was reduced from 11 to 16% (Basarab et al., 2012) It has also been suggested that best management practices such as intensifying the produ ction system can decrease the carbon fo o tprint through carbon fixation by 15 to 30% while maintaining production levels (Phetteplace et al., 2001) Using a life cycle assessment (LCA) in western Canada, Beauchemin et al. ( 2010) found that 83% of GHG emissions in the beef production chain of th e region originated in the cow calf phase. In t he US beef production system, the cow calf phase of beef production was found to be responsible for 76% of GHG emissions (Johnson et al., 2001a) Calves leave BIR with an average weight of 210 kg and, on average, GHG e missions per product are of 22 kg CO 2 e (kg calf LW) 1 This is in agreement with a study performed in t he US using information from Alabama, Texas, Uta h, Virginia and Wisconsin where G HG emissions per product were 21 kg CO 2 e (kg calf LW) 1 (Phetteplace et al., 2001) The studies conducted by Johnson et al. (2001a) and Beauchemin et al. (2010) demonstrated that most of the GHG emissions in beef production systems (76 and 83%, respectively) come from the cow calf phase. However, most of the weight gain of animals occur s after they leave the cow calf phase. Therefore, when expressing the carbon footprint per unit of beef produced as CO 2 e (kg beef) 1 higher values are found for the cow calf phase than for the whole system. If we consider that the animals leaving BI R may achieve 490 kg at slaughter and have 62% of this weight as hot carcass (Miller et al., 1996) we find that emissions for the whole beef production cycle would be of 12 kg CO 2 e (kg b eef) 1 on a live weight basis or 19 kg

PAGE 37

37 CO 2 e (kg beef) 1 on carcass weight basis. This agree s with many other studies performed in similar production systems. In the US beef production system, the emission of 13 to 16 CO 2 e (kg beef) 1 on a live weight basis was reported (Johnson et al., 2001a) while Beauchemin et al. ( 2010) found 13 kg CO 2 e (kg beef) 1 on a live weight basis and 22 kg CO 2 e (kg beef) 1 on a carcass weight basis in a case study made for beef production in western Canada. Basarab et al. ( 2012) reported similar values of 12 to 13 kg CO 2 e (kg beef) 1 on a live weight basis and 20 to 2 3 kg CO 2 e (kg beef) 1 on a carcass weight basis. A range of 2 2 to 26 kg CO 2 e (kg beef) 1 on a carcass weight basis was reported (Foley et al., 2011) for scenarios varying in feed management in Ireland In that study, Foley et al. (2011) found higher total emissions in production systems with higher productivity and therefore input requirements. However, this same scenario presented higher efficiency of fertilizer use and animal performance, resulting in lower emissions relative to beef production. Production efficiency is, therefore, an important aspect to consider when evaluating management strategies to reduce carbon footprint. Expressing carbon foot print relative to produce is a good indicati ve to prod uction efficiency. Conclusion s Emissions occurring on pre farm (before use of specific products in the system evaluated ) are relevant and must be accounted for when assessing the carbon footprint of agricultural production systems. In this case study, CH 4 emissi ons form enteric fermentation were t he largest contributor to the carbon footprint Emissions from BIR varied from one year to another mostly because of management practices such as lime and fertilizer application and pasture burning. On average, BI R emitted 10, 500 tons CO 2 e year 1 and 22 kg CO 2 e (kg calf LW) 1 These values agree with other similar studies performed in the US and other developed countries.

PAGE 38

38 Table 1 1 Concentration of GHG in the atmosphere before the Indus trial Revolution and in 2005. Source: IPCC, 2007a GHG gas Concentration before the Industrial Revolution Concentration in 2005 C O 2 280 ppm 379 ppm C H 4 715 ppb 1774 ppb N 2 O 270 ppb 319 ppb

PAGE 39

39 Table 1 2 Emission of greenhouse gases by economic sector, million metric tons CO 2 e. Source: EPA, 2013a Chapter/ IPCC Sector 1990 2005 2007 2008 2009 2010 2011 2011 (%) Energy 5,267.3 6,251.6 6,266.9 6,096.2 5,699.2 5,889.1 5,745.7 85.7 Industrial Processes 316.1 330.8 347.2 318.7 265.3 303.4 326.5 4.9 Solvent and Other Product Use 4.4 4.4 4.4 4.4 4.4 4.4 4.4 0.1 Agriculture 413.9 446.2 470.9 463.6 459.2 462.3 461.5 6.9 Land Use Change and Forestry 13.7 25.4 37.3 27.2 20.4 19.7 36.6 0.5 Waste 167.8 136.9 136.5 138.6 138.1 131.4 127.7 1.9 Total Emissions 6,183.3 7,195.3 7,263.2 7,048.8 6,586.6 6,810.3 6,702.3 Land Use Change and Forestry (Sinks) 794.5 997.8 929.2 902.6 882.6 888.8 905 .0 Net Emissions (Emissions and Sinks) 5,388.7 6,197.4 6,334.0 6,146.2 5,704.0 5,921.5 5,797.3

PAGE 40

40 Table 1 3 Emissions of GHG Agriculture in the USA (Tg CO 2 e year 1 ) 1990 to 2011. Source: EPA 2013 a Gas/Source 1990 2005 2007 2008 2009 2010 2011 2011 (% of total) CH 4 171.5 191.5 200.5 200.3 198.6 199.9 196.3 Enteric Fermentation 132.7 137 141.8 141.4 140.6 139.3 137.4 29.8 Manure Management 31.5 47.6 52.4 51.5 50.5 51.8 52 11.3 Rice Cultivation 7.1 6.8 6.2 7.2 7.3 8.6 6.6 1.4 Field Burning of Agricultural Residues 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.0 N 2 O 242.3 254.7 270.4 263.3 260.6 262.4 265.2 Agricultural Soil Management 227.9 237.5 252.3 245.4 242.8 244.5 247.2 53.6 Manure Management 14.4 17.1 18 17.8 17.7 17.8 18 3.9 Field Burning of Agricultural 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.0 Total 413.9 446.2 470.9 463.6 459.2 462.3 461.5

PAGE 41

41 Table 1 4 He rd information from Buck Island Ranch (BIR) period 1998 to 2008. Year Cows Pregnant Cows Bulls Calves Average weight of calves (kg) 1998 2933 2410 250 2228 197.8 1999 2865 2036 250 1814 212.7 2000 3106 2312 250 2099 225 0 2001 3213 2640 250 2361 208.2 2002 3205 2666 250 2418 212.3 2003 3114 2596 250 2375 199.6 2004 3014 2043 250 1910 223.2 2005 3209 2337 250 2288 203.7 2006 3215 2625 250 2458 217.3 2007 3306 2838 250 2643 215.5 2008 3414 2687 250 2481 194.6 Table 1 5 Area burned (ha) o n Buck Island Ranch (BIR) period between 1998 and 2008. Area (ha) month, year Improved Native January, 2002 911. 5 0.0 February, 2002 20.2 530.3 January, 2003 461.7 607.6 February, 2003 210.5 0.0 January, 2004 115.5 251.1 February, 2004 418.0 414.7 December, 2004 0.0 146.5 January, 2005 481.2 384.2 February, 2005 0.00 104.0 January, 2006 1159.8 1476.5 April, 2006 0.00 342.4 Table 1 6 Average above ground biomass available for burning in Buck Island Ranch (BIR) average from period 1998 to 2008. Average biomass (ton ha 1 ) M onth I mproved Non improved January 3.4 0.4 February 4.6 4.8 April 2.0 1.7 December 5.4 4 0

PAGE 42

42 Table 1 7 Lime, fertilizer, molasses feed concentrate and fuel used at Buck Island Ranch (BIR) 1998 to 2008. Year Lime (ton year 1 ) Synthetic fertilizer (ton N year 1 ) Molasses (ton year 1 ) Diesel (gallons year 1 ) for molasses transportation Gasoline (gallons year 1 ) Diesel (gallons year 1 ) Feed concentrate (t ) 1998 0.0 105.4 468.7 1245.3 4624.4 13021.1 25.0 1999 0.0 68.3 417.7 1109.9 3136.6 13312.4 144.5 2000 0.0 47.6 564.9 1500.9 3208.0 23339.0 166.3 2001 1154.0 42.6 758.7 2015.9 4041.4 17729.2 384.6 2002 1354.0 45.7 524.4 1393.5 4628.6 15156.5 136.9 2003 1802.0 22.6 874.9 2324.6 5578.5 13068.5 243.6 2004 2392.0 94.1 631.3 1677.4 5680.3 12140.0 241.7 2005 524.0 12.5 885.4 2352.4 3332.9 13844.0 511.5 2006 1624.0 0.0 286.7 761.7 3555.6 14175.4 598.0 2007 0.0 0.0 793.9 2109.5 3732.3 17991.5 812.0 2008 0.0 0.0 723.6 1922.6 3566.8 11743.5 272.6

PAGE 43

43 Table 1 8 Source categories and GHG emitted in the production system at in Buck Island Ranch (BIR) 1998 to 2008. Category GHG Methodology Enteric fermentation (cows) CH 4 IPCC ( 2006) ,Tier 2 Enteric fermentati on (bulls) CH 4 IPCC ( 2006) ,Tier 1 Animal waste CH 4 N 2 O IPCC ( 2006) ,Tier 1 Urea for NNP N 2 O IPCC ( 2 006) ,Tier 1 Pasture fertilization N 2 O IPCC ( 2006) ,Tier 1 Pasture lime CO 2 IPCC ( 2006) ,Tier 1 Production, transportation, storage and transfer CO 2 Lal ( 2004) Burning of pasture CH 4 CO N 2 O NO x IPCC ( 2006) ,Tier 1 Diesel CO 2 CH 4 N 2 O EPA ( 2005 ) Gasoline CO 2 CH 4 N 2 O EPA ( 2005) Table 1 9 Global Warming Potential (GWP) of GHG. GHG GWP Source CO 2 1 IPCC ( 2007b) CH 4 25 IPCC ( 2007b) N 2 O 298 IPCC ( 2007b) CO 1.9 IPCC ( 2007b)

PAGE 44

44 Table 1 10 Data and emissions factor values, units and sources. Enteric fermentation Factor Value Unit Source DE 62.6 % of GE EPA ( 2013 b ) Cfi cows 0.322 MJ day 1 kg 1 IPCC ( 2006) Cfi lactating cows 0.386 MJ day 1 kg 1 IPCC ( 2006) Ca 0.17 dimensionless IPCC ( 2006) C pregnancy 0.1 dimensionless IPCC ( 2006) Ym 6.5 % of GE EPA ( 2013 b ) EF1 53 (kg CH 4 ) head 1 year 1 IPCC ( 2006) Milk yield 6.4; 6.7; 5.6; 5.5; 4.4; 4.0; 3.1 (kg of milk) day 1 Minick et al. ( 2001 ) average Milk fat 3.7 % Marston et al. (1992), average Manure (CH 4 ) Factor Value Unit Source VS (T) bulls 1721 (kg dry matter) animal 1 year 1 EPA (2013b) VS (T) calves 7.7 (kg dry matter) (1000 kg) 1 day 1 EPA (2013b) Bo (T) 0.17 (m 3 CH 4 ) (kg of VS excreted) 1 EPA (2013b) MCF S,k 1.5 % IPCC ( 2006) Manure (N 2 O) Factor Value Unit Source EF 3 PRP 0.02 kg N 2 O N (kg N input) 1 IPCC ( 2006) E N cows 0.33 kg N day 1 (1000 kg) 1 EPA ( 2013 b ) E N bull 0.31 kg N day 1 (1000 kg) 1 EP A ( 2013 b ) E N calves 0.30 kg N day 1 (1000 kg) 1 EP A ( 2013 b ) Frac GASM 0.20 kg N vol (kg of N added) 1 IPCC ( 2006) EF 4 0.01 kg N 2 O N (kg NH 3 N + NO x N) 1 vol. IPCC ( 2006) Frac LEACH (H) 0.30 kg N leach. and run (kg on N added) 1 IPCC ( 2006) EF 5 0.0075 kg N 2 O N (kg N leached and runoff) 1 IPCC ( 2006)

PAGE 45

45 Table 1 11 Data and emissions factor values, units and sources (continuation). Pasture burning Factor Value Unit Source C f 0.74 dimensionless IPCC ( 2006) G ef N 2 O 0.21 g (kg dry matter burnt) 1 IPCC ( 2006) G ef CH 4 2.3 g (kg dry matter burnt) 1 IPCC ( 2006) G ef CO 65.0 g (kg dry matter burnt) 1 IPCC ( 2006) Synthetic N fertilizer Factor Value Unit Source EF 1 0.01 kg N 2 O N (kg N input) 1 IPCC ( 2006) F SN Table 1 7 kg N year 1 BIR a Frac GASF 0.1 (kg N volatilized) (kg N applied) 1 IPCC ( 2006) EF 4 0.01 kg N 2 O N (kg NH 3 N + NO x N) 1 vol. IPCC ( 2006) Frac LEACH (H) 0.30 kg N leach. and run (kg on N added) 1 IPCC ( 2006) EF 5 0.0075 kg N 2 O N (kg N leached and runoff) 1 IPCC ( 2006) Dolomitic lime Factor Value Unit Source M dolomite Table 1 7 kg lime year 1 BIR a EF dolomite 0.064 ton s of CO 2 ( ton of dolomitic lime) 1 EPA ( 2013 a ) Production, Transportation, Storage and Transfer Factor Value Unit Source F SN Table 1 7 kg N year 1 BIR a M dolomite Table 1 7 kg lime year 1 BIR a Equivalent Carbon Emission, N sythetic fertilizer 1.3 kg CE kg 1 Lal ( 2004) Equivalent Carbon Emission, lime 0.16 kg CE kg 1 Lal ( 2004) Fuel Factor Value Unit Source Fuel Gasoline Table 1 7 gallons BIR a Fuel Diesel Table 1 7 gallons BIR a E fuel Gasoline 8.8 kg gallon 1 EPA ( 2005) E fuel Diesel 10.1 kg gallon 1 EPA ( 2005) a BIR: Buck Island Ranch data

PAGE 46

46 Figure 1 1 Map of the state of Florida. "A" refers to the location of Buck Island Ranch (BIR) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Pregnant Lacta t ing Figure 1 2 Calendar of animals' reproductive stage.

PAGE 47

47 Figure 1 3 Emissions from BIR, ton CO 2 e year 1

PAGE 48

48 Figure 1 4 GHG emissions from BIR per category, % over average of all years, 1998 to 2008. Figure 1 5 Average GHG emissions from synthetic N fertilizer and lime before and after they are applied on the farm.

PAGE 49

49 CHAPTER 2 SENSITIVITY ANALYSIS OF ENTERIC FERMENTATION EMISSION MODEL Literature Review A mathematical portrayal of a system is named a model and is usually composed of inputs and outputs (Jones and Luyten, 1998), where one or more equation s are used r (France and Thornley, 1984). Inputs include constants, fixed values throughout all model runs, and parameters, values that change each time the model runs (Fran ce and Thornley, 1984). Outputs refer to time dependen t values that express the status of the system under evaluation (Jones and Luyten, 1998). Models differ from each other according to the process used for their creation and according to the purpose for which they were built. Empirical models fit mathematical equations to data through statistical methods, while mechanistic or analytical models have their equations built based on bio logical systems concepts (Keen and Spain, 1992), although empirical knowl edge is also used to build mechanistic models to some extent (France and Thornley, 1984). The assumptions used for building mechanistic models establish some restrain t s that result in their inferior ability to fit results to data sets when compared to empi rical models (France and Thornley, 1984). Models can also be classified according to th eir target use. Models used as decision making tool s are called engineering or functional models, while those focused on explaining physiological and environmental relat ionships are named scientific or mechanistic (Passioura, 1996). Models can be extremely useful when estimating GHG product ion o n a large scale (Storm et al., 2012), but it should be considered that the use of models along with experiments is crucial when s tudying different alternatives of a system (Tittonell et al., 2012).

PAGE 50

50 Regarding CH 4 enteric fermentation, several models have been reported in the literature. These models either relate CH 4 (mechanistic) or to nutrient s consumed by the animals (empirical) (Kebreab et al., 2008) Moe and Tyrrell (1979) for example, developed equations that relate dry matter intake (DMI) to CH 4 emissions at low intake levels and DMI and carbohydrate type at high intake levels for dairy cattle. Another mechanistic model was developed by Dijkstra et al. (1992) following a Michaelis Menten mass flux dynamic that considers microbial dynamic s in the rumen including volatile fatty acids (VFA) absorption depending on rumen VFA. However, EPA and similar organizations in other countries use the model presented by IPCC (2006) to develop inventories of GHG on national levels (Nijdam et al., 2012) Kebreab et al. (2008) evaluated several models and their ability to predict CH 4 output by both dairy and beef cattle. The authors concluded that the IPCC (2006) model estimated CH 4 agreed fairly well with measured CH 4 values, although it was not the most accurate of the models evaluated. However, authors also emphasize that the national GHG emission inventories level, the difficulty of using mechanistic m odels might prevent their use. Sensiti vity analysis is performed to ass will affect its output and is an important mechanism for analysis of mode l behavio r Sensitivity analysis is also useful in the process of building a model, since it provides information regarding t he importance or irrelevance of considering a parameter in t he simulation This procedure helps identify interactions between inputs and parameters as well as irrelevant inputs (Saltelli et al., 2004; Monod et al., 2006) and, particularly when using global sensitivity analysis, to determine the inputs that should be most a ccurately measured (Monod et al., 2006) Therefore, performing a sensitivity

PAGE 51

51 outputs (Cukier, 1973) Sensitivity analysis methods can be separated in to two groups: local and global. Local sensitivity analysis evaluates the importance of parameters or variables in a model by performing derivati ve calculations of the output in relation to these factors, separated in small intervals not related to their uncertainty. The result is a measurement of how intensely the output varies around inputs (Monod et al., 2006) Global sensitivit y analysis allows the evaluation of different parameters at the same time and is variance based, meaning that the factors under evaluation are varied within the limits of their uncertainty (Monod et al., 2006) Results are averaged over t he variation of all inputs (Saltelli et al., 1999) Several sensitivity analysis techniques exist, differing in their sampling approaches and evaluation. A description of three of these techn iques used in this study follows below. The Bauer and Hamby (1991) sensitivity analysis creates qualitative indexes the remaining parameters at their nominal values, one at a time. This technique is more likely to be successful when used with medium linear models, since it does not identify non linear or extreme interactions in small models, therefore under estimating sensitivity indexes. In large, complex mode ls it may take too much computing time to be processed (Monod et al., 2006) Morris is a model independent sensitivity analysis method, meaning that its use is independent fro m previous assumptions of input effects on the output and that it can be used in non linear, non monotonic models (Saltelli et al., 2004) Monotonicity is a property occurring when a factor has the same signal effect in the

PAGE 52

52 output (Campolongo et al., 2007) This analysis originally aimed at defining which inputs have negligible, additive and linear, or nonlinear or interactive effects in the output, named elementary effects. For this, a computational experimental design is built where inputs are randomly varied in a one at a time fashion so that output variations due to each input are individually evaluated (Morris, 1991) The Morris sensitivity analysis is interpreted based on the indexes (or ) and which are the mean and the standard deviation of (Saltelli et al., 2004) d its interaction and non linear effects, respectively (Campolongo et al., 2007) Campolongo et al. (2007) proposed a new sampling strategy to factor at a time is varied in a random order. The same authors also indicate the use of absolute values for the estimation of and to avoid the cancellation of effects when the model in non monotonic, i. e., when an input may vary the signal of its elementary effect on the output. In the Fourier Amplitude Sensitivity Test (FAST), parameters under evaluation are altered simult aneously in a determined frequency within their probability detecting unimportant parameters, helping to eliminate unnecessary equations in complex models and, due to the exposes interactions between parameters. Results of this analysis represent an (Cukier, 1973; Cukier et al., 1978) where the output variance is broken down into partial variances conferred by each parameter and the ratio s of these partial variances are used to determine (Xu and Gertner, 2011) This

PAGE 53

53 technique needs less computer runs than those that vary one parameter at a time (Cukier, 1973; Cukier et al., 1975) and can be used in a wide range of equation numbers, types and systems (Cukier, 1973) including nonlinear and non monotonic models (Xu and Gertner, 2011) Material s and Methods Model The model evaluated was the one presented in 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC, 2006) for Tier 2 level calculations. This model requires more information than the Tier 1 approach which uses default emission factors and its results are more properly sensitive to changes in animal production systems (Lassey, 2007) The sensitivi ty analysis was performed regarding three parameters in the IPCC (2006) enteric fermentation CH 4 production model considering animals on pasture or in feedlots. Although the case study was performed using an example of adult animals, the parameters re lated to pregnancy and lactation did not offer a large range of variation. Therefore, the SA was performed regarding growing animals to also consider weight gain. Net energy for work refers to energy spent by animals performing activities such as pulling plows, etc., and since such activities are not performed by the animals in Florida it was also considered zero. Remaining parameters (not under evaluation) were used at their default values as presented in the IPCC (2006) ( Table 2 1 ) enteric fermentation equations were transformed into a function in R using the function() command in order to perform the sensitivity analysis. GE=[( ) +( )]/( ) where GE= gross energy, M J day 1

PAGE 54

54 NEm= Net energy required for animal maintenance, MJ day 1 NEa= Net energy for animal activity, MJ day 1 NEl= Net energy for lactation, MJ day 1 NEwork= Net energy for work, MJ day 1 NEp= Net energy required for pregnancy, MJ day 1 REM= ratio of net energy available in a diet for maintenance to digestible consumed NEg= Net energy needed for growth, MJ day 1 REG= Ratio of n et energy available for growth in a diet to digestible energy consumed DE= digestible energy expressed as a percent of gross energ y (percent) Formulas used to calculate each of the factors involved in GE calculation are as follows: NEm= Cfi x (weight) 0.75 where NEm= Net energy required for the animal for maintenance, MJ day 1 Cfi= coefficient, varies for each animal category, MJ da y 1 kg 1 Weight= live weight of animal, in kg Weight= iw + (WG x days) where iw= initial weight, considered to be 200 kg WG= average daily weight gain of the animals in the population, kg day 1 days= days since the beginning of the production cycle NEa= Ca x NEm where NEa= Net energy for animal activity, MJ day 1

PAGE 55

55 NEm= Net energy required for the animal for maintenance, MJ day 1 NEg= 22.02 x ( ) 0.75 x ADG 1.09 7 where NEg= net energy needed for growth, MJ day 1 BW= average body weight of the animals in the population, kg C= coefficient with values of 0.8 for females, 1.0 for castrate d animals and 1.2 for bulls. In this case 0.8 was used. MW= mature live weight of an adult female cow in moderate body condition, kg ADG= average daily weight gain of the animals in the population, kg day 1 REM= [1.123 (4.092 x 10 3 x DE%) + [1.126 x 10 5 x (DE%) 2 ] ( )] where REM= ratio of net energy available in a diet for maintenance to digestible consumed DE= digestible energy expressed as a percentage of gross energy REG= [1.164 (5.160 x 10 3 x DE%) + [1.308 x 10 5 x (DE%) 2 ] ( )] where REG= Ratio of n et energy available for growth in a diet to digestible en ergy consumed DE= digestible energy expressed as a percentage of gross energy. The value used was the same as in the REM calculation After that, a daily emission factor for each category was calculated with the formula below.

PAGE 56

56 DayEmit = [ ] where: DayEmit = emission factor, kg CH 4 1 head 1 day 1 GE = gross energy intake, MJ day 1 Ym = CH 4 conversion rate, which is the fraction of gross energy in feed converted to CH 4 (%). 55.65 = a factor for the energy content of CH 4 MJ (kg CH 4 ) 1 To determine yearly emissions for each category, the formula below was used. Emissions = DayEmit x 365 where: Emissions = total emissions in a month for the category, kg CH 4 year 1 DayEmit = emission factor for the category, kg CH 4 head 1 day 1 365 = days in the year Data Source The sensitivity analysis was performed regarding three parameters in the IPCC (2006) ruminal CH 4 production model considering finishing animals on pasture or in feedlots. For this, a literature review was performed in order to obtai n the range of possible values for the para meters under evaluation ( Table 2 2 and Table 2 3 ) Since the values differ considerably between the two feeding situation s (animals on pasture or on feedlot), sensitivity analyse s were made separately. Three global sensitivity analysis me thodologies were applied: vary one a t a time (Bauer and Hamby, 1991 ) Morris (Morris, 1991) and Fourier Amplitude Sensitivity Test (FAST ; Cukier, 1973 ).

PAGE 57

57 The analyses were made using different methods from the R software that will be described below. R is a software built for statistical computation, including linear and non linear regression models, time series analysis, smoothing, etc., with a flex ible graphical environment. R comes with several basic packages, but packages for specific functions can be downloaded ( Venables et al., 2011 ). Vary one at a time (OAT) The OAT method was programmed using the R language (Venables et al., 2011) One parameter was varied at a time in an equally spaced domain as shown in Figure 2 1 R emaining parameters were maintained at their average values. The index of relative sensitivity was calculated by the following: I= [max y((zi)) min y((zi))] / max y((zi)) where zi= parameter value ; Mo rris For the Morris sensitivity analysis, the parameters were sampled 2560 times within the experimental space delimited by the values i n Table 2 2 and Table 2 3 This analy sis was made using the package sensitivity ( Pujol et al., 2013) library morris() In this method, an OAT (one at a time) experimental design is developed where r is the number of elementary effects for each i nput ; levels refers to the number of levels in the design, i. e., how many times the model runs ; grid.jump is the number of jumps in the experimental grid, usually half the number of levels ;

PAGE 58

58 scale is true if the inputs are scaled to vary between 0 and 1, t o avoid misinterpretation when inputs have contrasting orders of magnitude. In this analysis, r = 2560; levels= 2560; grid.jump= 1280; scale= T (true). The development of the experimental design can be observed in Figure 2 2 where we can see an example of how the random variation evolves and covers the domain of the three inputs under evaluation. The experimental design for animals on pasture and feedlot can be observ ed in f ast99( ) An important feature of this function is that it runs the extended FAST method, which allows for the consideration of both (Saltelli et al., 1999). In this function the arguments are model refers to the model under evaluation; factors output; n is the sample size (Cukier et al., 1978) ; M is the interference p arameter (Cukier et al., 1978) used to decrease the error. Default value of 4 was used (Cukier, 1973) ; omega is the frequency used to sample each factor ; q is the distribution of inputs, set as uniform in this analysis ; q.arg refers to t ls, defined in Table 2 2 and Table 2 3 ; x is the vector wher e input values are stored ; y is a vector where the model responses are stored

PAGE 59

59 FAST returns two values for each parameter evaluated, the main effect of the parameter and the interaction of the parameter with other factors. All of the values summed are equal to 1. The function returns three values: Morris (1991) ; posed by Campolongo et al. (2007) ; : standard FAST For the FAST sensitivity analysis, the parameters were sampled within the experimental space delimited by the values i n Table 2 2 and Table 2 3 shown in Figure 2 6 This analysis was made using the package sensitivi ty ( Pujol et al., 2013) fast99() An important feature of this function is that it runs the extended FAST method, which allows for the consideration of both parameters ns with other factors (Saltelli et al., 1999). In this function the arguments are model refers to the model under evaluation; factors output; n is the sample size (Cukier et al., 1978) ; M is the interference parameter (Cukier et al., 1978) used to decrease the error. Default value of 4 was used (Cukier, 1973) ; omega is the frequency used to sample each factor ; q is the distribution of inputs, set as uniform in this analysis ; q.arg refers to n Table 2 2 and Table 2 3 ;

PAGE 60

60 x is the vector where input values are stored ; y is a vector where the model responses are stored FAST returns two values for each parameter evaluated, the main effect of the parameter and the interaction of the parameter with other factors. All of the va lues summed are equal to 1. Results Results can be observed i n Table 2 4 and Figure 2 6 through Figure 2 8 One of the advantages of using the OAT analysis is that, particularly when using graphic resources, one can observe the basic relationship between output and inputs. We can observe in Figure 2 6 that CH 4 production (kg animal 1 day 1 ) increases with the increasing values of CH 4 conversion rate (Ym, %) and ADG (average daily gain, kg animal 1 day 1 ) and decreases with greater values of digestible energy (DE, %). The association between CH 4 emissions (kg animal 1 year 1 ) with digestible energy (DE, %) is negative and slightly quadratic and ADG (kg animal 1 day 1 ) had a quadratic relationship with CH 4 emissions (kg animal 1 year 1 ) ( Figure 2 6 ). The rel ationship of CH 4 emissions (kg animal 1 year 1 ) with CH 4 conversion rate (Ym, %) was positive and linear. This factor is used only once in the estimation of CH 4 production, when multiplied by the gross energy (GE, MJ day 1 ), thus the linear relationship: DayEmit= [ ] Observing the OAT indexes obtained for each parameter ( Figure 2 6 and Table 2 4 ) we can see that ADG is the most important parameter influencing the output of the mode l being evaluated, independent of feeding situation. However, when the model is used to simulate CH 4 emissions of animals on pasture, digestible

PAGE 61

61 energy available for animals influences the output more intensely than CH 4 conversion rate Ym. Results from th e sensitivity analysis performed using the FAST method can be observed in Figure 2 7 One of the important aspects of this analysis is that separates direct effects f rom a parameter in the output and its interaction with other factors, where the sum of the FAST index should be 1. The r anked importance of parameters in each simulation scenario was not different from the previous analysis, where ADG (kg animal 1 day 1 ) w as considered the most import parameter to (%) was the second most important factor influencing CH 4 emissions on pasture conditions, similar to what was found in the analysis with OAT and Morris methodologies ( Table 2 5 ) Interactions between parameters and other factors used in the model do not seem to have a major importance in the model according to the FAST analysis However, when we observe the results from the Morris analysis ( Figure 2 8 ) we can see that, according to the values, interactions and non linearity play a n important role in the behavio r of the model particularly regarding the digestible energy (DE, %) in simulations made for animals on pasture, the CH 4 conversion rate (Ym, %) for simulation of animals on feedlo t and, for both feeding situations, average daily gain (ADG, kg animal 1 day 1 ). Discussion Energy partitioning of consumed feed in animals follows the scheme below ( Figure 2 9 ), where we can see that consumed gross energy (GE) is lost in feces, urine and CH 4 Total consumed energy minus energy lost in the feces is referred to as digestible energy (DE), and digestib le energy minus energy lost in urine and as CH 4 is called metabolizable energy (ME). Remaining energy, also referred to as net

PAGE 62

62 energy (NE), is used by animals for maintenance, growth and production ( Minson, 1990 ; Van Soest, 1982) The model evaluated by this sensitivity analysis is based on the concepts described above. In the first instance, the GE is calculated for animals based on their energy requirement for maintenance, growth, milk production and reproduction. This calculation is based on production data or production estimates (ADG) associated DE usually refers to soluble carbohydrates, organic acids and structur al carbohydrates that are not coated by lignin layers (Minson, 1990) and is closely related to dry matter digestibility (Moir, 1961) and organic matter digestibility (Rittenhouse et al., 1971) endogenous waste and microbial material (Van Soest, 1982) Methane (CH 4 ) conversion rate (Ym) is a ratio between CH 4 produced by the animals and feed intake, both in units of energy of combustion. In grazing trials, assessing animal intake may be troublesome and lead to uncertainty in individual Ym estimations, while o n a population scale the Ym can have uncertainty from the inter animal variation (Lassey, 2007) These ch aracteristics may difficult retrieving reliable data of Ym to be used in national scale inventories. Decreasing the amount of energy used in the production of CH 4 can increase use efficiency and consequently lead to an improvement in econo mical competence (Beauchemin and Mcginn, 2006; Kurihara et al., 1999) Analyzing data from 20 e nergy metabolism studies with dairy cattle, Yan et al. ( 2010) found that energy partitioning, use efficiency, metaboli sm an d animal productivity can have an effect on the proportion of energy spent in CH 4 production.

PAGE 63

63 DE was found to be negatively related to dry matter intake (Nkrumah et al., 2006) Ku rihara et al. (1999) evaluating CH 4 production and energy partitioning in Brahman cattle fed two tropical forage and one high grain diets, found that less CH 4 was produced as DE (%) increased. In that experiment, DE ranged from 44 to 59% for the forage s and achieved 70% in the high grain diets. A decrease in DE was reported by Beauchemin and Mcginn (2006) when feeding animals unsaturated fat using canola oil and resulted in a reduction in daily CH 4 emissions. In that experiment, however, the reduction in DE (%) occurred along with a reduction in total GE and the authors suggest that the use of fats in animal diet migh t not be feasible because of their negative effect o n total DE intake by the animals. For grazing beef cattle, EPA calculates the DE values used in the Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990 2011 (EPA, 2013a) as a using a distinct value for the western regions to consider its general lower quality forage. For animals on feedlots, one v alue of DE was used for the entire country but it was considered to vary annually according to feed type nutritional information and availability (EPA, 2013b) All the sensitivity analysis perf ormed show that in the model ass essed DE is of crucial importance in the output when simulations are made for animals on pastu re. Considering the large population of cattle in the U.S. kept on pasture, rangeland and meadow, which reached 49.5 million head in 2011 (EPA, 2013b) gathering precise information on DE of feed will significantly affect GHG estimation of the beef production sector. Animals fed high concentrate diets may have less en ergy lost as CH 4 (Beauchemin and Mcginn, 20 05; Kurihara et al., 1999) However, changing animal feed from pasture to concentrate should consider the emissions associated with

PAGE 64

64 production of feed at a farm level (Yan et al., 2010) Also, in the context of climate change and its future consequen ces to agricultural practices and production, human consumption should not be ignored (Beauchemin et al., 2010) The amount of total energy consumed used in the production of CH 4 can be affected by s everal factors. Yan et al. (2010) reviewing 20 studies made with dairy cows fed fresh grass and grass silage found that les s energy was used in the production of CH 4 when feed intake and energy used in milk production increased. The authors also concluded that CH 4 emissions and milk production are negatively related, i.e., an increase in the value of energy converted into CH 4 was related to a decrease in energy use efficiency For simulations made for animals on feedlot CH 4 conversion rate Ym (%) particular importance considering the differen ce found in Ym from experimental studies and the values used in national scale inventories. Beauchemin and Mcginn (2005) studying CH 4 emissions from cattle fed corn or barley diets, found that Ym from animals on barley feed was greater than the ones indicated by IPCC ( 2006) A sorghum based high grain diet fed to cattle by Kurihar a et al. (1999) however, presented a Ym similar to the one suggested by IPCC (2006). For simulations made for animals on pasture it was observed that Ym was of lesser relevance in the use of the model. However, this does not imply Ym is irrelevant when estimating emissions from beef production on pasture. In fact, t he extrapolation of Ym values from animal populations to national scale GHG inventories can introduce considering amount of uncertainty to the result (Lassey, 2007)

PAGE 65

65 for estimating CH 4 emissions from enteric ferment ation from non dairy cattle showed that Ym and the coefficient for calculating net energy for maintenance were the most relevant sources of uncertainty (Karimi Zindashty et al., 2012) This is valuab le information considering the diffi culty involved in obtaining these data from field experiments on pasture, to which only the SF 6 technique can currently be used (Lassey, 2007; Storm et al., 2012) Kebreab et al. (2008) evaluated different empirical and mechanistic models developed to estimate CH 4 emissions, including the IPCC (2006) model evaluated in this study. The authors conclude d that mechanistic models are better CH 4 predictors, howeve r due to the difficulty related to their use the authors suggest that these model s should be used in order to more accurately estimate Ym and its change due to modifications in feed. For simulation s made for animals on pasture, the USDA uses a fixed value of 6.5%. For animals on feedlot, the mechanistic model MOLLY as described in Kebreab et al. (2008) is used to estimate Ym. An important aspect of the m odel evaluation is that (Kebreab et al., 2008) did not evaluate prediction ability for animals maintained on pasture due to lack of reliable information related not only to CH 4 emissions but also t o the characteristics that may affect animals in pastoral environments. An appropriate way to evaluate and compare different production systems is to express CH 4 emissions per unit of animal production (Kurihara et al., 1999) particularly when considering that CH 4 production relates to energy channeled not into production but lost to the environment (Lassey, 2007) In beef production, ADG is an important parameter to evaluate production efficiency. Dividing CH 4 emissions from enteric ferm entation by ADG used in the simulation ( Figure 2 10 ) we can see that increases in productivity reduce CH 4 ADG 1 supporting the concept that increases in productivity enhance GHG efficiency. In a study where Brahman cattle

PAGE 66

66 were fed diets varyi ng in digestibility and N content animal weight change was found to have a negative quadratic relationship with CH 4 emissions when these were expressed as g CH 4 (kg live weight gain) 1 This implies that reduction in CH 4 may only be achieved in animals tha t present low weight gains (Kurihara et al., 1999) The IPCC (2006) model for enteric fermentation seems to agree with this concept. A ccording to Figure 2 10 animals on pasture seem to have higher potential to increase their CH 4 use efficiency. In fact, the high Ym found in tropical forages suggests considerable potential to improve animal productivity in the Tropics (Kurihara et al., 1999) This increase in productivity would the refore not only benefit the farmer, but also be beneficial to the environment by reducing GHG emissions per unit of animal product (Lassey, 2013) Interactions among parameters and other inputs in the model are evaluated by both the FAST and the Morris analysis. The FAST analysis returns one value for the main effect of th interactions with other factors. Similarly, Morris analysis returns two values, one referring to main effects of the parameter in the output () and another to the l inearity ( ). Campolongo et al. (2007) highligh t the importance of considering both and when evaluating the importance of a parameter because is slightly disposed to Type II error, i. e., when a parameter is significant but the an alysis is unable to detect it. Looking at the ranking of the param eters evaluated in this sensitivity analysis, we observe that the main effect o f parameters o the three sensitivity analysis Table 2 5 The interactions of the parameters follow the the ones that show interaction. The FAST analysis shows that i nteraction plays a

PAGE 67

67 minor role in the use of the model. Therefore, the high values for found by the Morris analysis for ADG in both pasture and feedlot simulations and in DE for linea r relationship with the output in CH4 animal 1 year 1 also observed i n the OAT sensitivity analysis Conclusions Interactions seem to be of minor importance in the use of this model. The most important parameter influencing the output in the enteric ferme ntation model evaluated is ADG (kg animal 1 day 1 ) for simulations made for anima l s on both pasture and feedlot. The implication for this is that more accurate values of ADG should greatly influence the results found with the model. Considerin g its general ized use on large scale modeling such as national GHG inventories, it may be of importance to more precisely acquire information on ADG to improve the use of the model. for simulations on pasture is DE (%) followed by Ym (%). For simulations made for animals on feedlot, Ym (%) is the second most important parameter influencing the considerable diffi culty to be measured on experimental basis and include the need for expensive equipment and trained personnel. However, if one aims at improving the use of models, more research is necessary to collect reliable information on these parameters.

PAGE 68

68 Table 2 1 Default values for parameters not evaluated in sensitivity analysis. Parameter Description Feedlot Pasture Unit Source MW mature body weight of an adult female in moderate body condition 450 450 kg IPCC ( 2006 ) C coefficient with a value of 0.8 for females, 1.0 for castrates and 1.2 for bulls 0.8 0.8 dimensionless IPCC ( 2006 ) Cfi coefficient varying for each animal category 0.322 0.322 MJ d 1 kg 1 IPCC ( 2006 ) Ca activity coefficient corresponding to animal's feeding situation 0 0.17 dimensionless IPCC ( 2006 )

PAGE 69

69 Table 2 2 Digestible energy (DE, %) and methane conversion rate (Ym, %) values and sources used in the sensitivity analyse s. Feeding conditions Parameters Grazing Feedlot Source DE 55 75 75 85 IPCC ( 2006 ) Ym 5.5 7.5 2 4 IPCC ( 2006 ) Table 2 3 Average daily gain (ADG, kg/day) values and s ources used in the sensitivity analyses ADG Diet composition Source 0.06 0.70 Limpograss Sollenberger et al. ( 1997) 0.28 0.34 Limpograss Stewart Jr. et al., ( 2007) 0.30 0.48 Bermudagrass Mislevy and Dunavin ( 1993) 0.43 0.62 Florico Stargrass 0.37 0.49 Florona Stargrass 0.31 0.33 Pensacola Bahiagrass 0.19 0.24 Tifton 9 0.25 0.56 Floralta Limpograss Sollenberger et al. ( 1989) 0.17 0.56 Pensacola Bahiagrass 0.49 1.14 4.5% molasses 3.5% cottonseed meal 0.75% limestone 82.5% ground corn 8% ground alfalfa hay 0.75% urea Phillips et al. ( 2006)

PAGE 70

70 Table 2 4 OAT, FAST and Morris indexes for the enteric fermentation emission model (IPCC, 2006). ADG = average daily gain, kg animal 1 day 1 ; DE = digestible energy, %; Ym = methane conversion rate, %. Parameter OAT Main Effect Interactions Pasture Feedlot Pasture Feedlot Ym 0.27 0.5 DE 0.41 0.16 ADG 0.58 0.68 Parameter FAST Main Effect Interactions Ym 0.08 0.46 0.01 0.02 DE 0.25 0.03 0.03 0.00 ADG 0.64 0.48 0.03 0.02 Parameter Morris Main Effect ( ) Interactions and non monotonicity ( ) Ym 14.4 15.5 4.2 3.1 DE 24.2 4.1 8.8 1.2 ADG 39.8 15.8 8.7 3.2 Table 2 5 Ranking of parameters' influence in the fermentation model. Pasture Feedlot OAT FAST Morris OAT FAST Morris Ym 3 3 3 2 2 2 DE 2 2 2 3 3 3 ADG 1 1 1 1 1 1

PAGE 71

71 Figure 2 1 OAT experimental design used in the sensitivity analysis.

PAGE 72

72 Figure 2 2 Evolution of the experimental design in Morris sensitivity analysis (a

PAGE 73

73 Figure 2 3 Experimental design in the FAST method for animals on pasture and on feedlot. ADG= average daily gain, kg animal 1 day 1 ; DE= digestible energy, %; Ym= methane conversion rate, %.

PAGE 74

74 Figure 2 4 Experimental design in the FAST sensitivity analysis. ADG = average daily gain, kg animal 1 day 1 ; DE = digestible energy, %; Ym = methane conversion rate, %.

PAGE 75

75 Figure 2 5 Output of enteric fermentation methane emission model (IPCC, 2006) in kg CH 4 animal 1 year 1 as a function of parameter values in the OAT sensitivity analysis method

PAGE 76

76 Figure 2 6 OAT index for sensitivity analysis of simulations on pasture and feedlot situations.

PAGE 77

77 Figure 2 7 FAST indexes for the enteric fermentation emission model (IPCC, 2006). ADG = average daily gain, kg animal 1 day 1 ; DE = digestible energy, %; Ym = methane conversion rate, %.

PAGE 78

78 Figure 2 8 Morris indexes for the enteric fermentation emission model (IPCC, 2006). ADG = average daily gain, kg animal 1 day 1 ; DE = digestible energy, %; Ym = methane conversion rate, %.

PAGE 79

79 Figure 2 9 Energy partitioning in animals. Adapt from: (Minson, 1990 Van Soest, 1982) Figure 2 10 Relationship between ADG (kg animal 1 day 1 ) and methane production (g CH 4 ADG 1 ) for simulations made with Tier 2 enteric fermentation model (IPCC, 2006) for animals on pasture and feedlot.

PAGE 80

80 CHAPTER 3 RUMINAL METHANE EMISSIONS, FORAGE AND ANIMAL PERFORMANCE RESPONSES TO DIFFERENT STOCKING RATES ON CONTINUOUSLY STOCKED BAHIAGRASS PASTURES Literature Review Bahiagrass ( Paspalum notatum ) is an important pasture species in the US and is cultivated in a larg e area in the country, from southern regions to east Texas and central Tennessee. This species covers over 2 million hectares in the southern USA (Newman et al., 2011) and over one million hectares in Florida (Chambliss and Sollenberger, 1991) In north Florida, bahiagrass growing season extend s from April to November while in the south ern part of the state bahiagrass grow mainly from March to mid December (Chambliss and Sollenberger, 1991) Bahiagrass is adapted to a wide variety of soil types and soil pH (Chambliss and Sollenberger, 1991; Twidwell et al., 1998) and climates (Chambliss and Sollenberger, 1991) It is a warm season perennia l grass that reproduces by seed (Burton, 1955) and it can achieve 225 to 560 kg of se ed ha 1 when properly managed (Chambliss and Sollenberger, 19 91) It has a prostrate habit with shallow rhizomes (Newman et al., 2011) Bahiagrass can be used for the production of hay and grazing and is able to support heavy grazing at low soil fertility levels (Twidwell et al., 1998) presenting high persistence and being easily established (Burton, 1955) In Florida, bahiagrass is mostly used for beef cattle (Chambliss and Sollenberger, 1991) Sollenberger et al. ( 1989) found an ADG of 0.41 kg head 1 day 1 for beef steers on bahiagrass but encountered a large decrease in ADG in August and September. Ruminants have one forestomach separated in to three compartments called reticulum, rumen and omasum In the rumen, the anaerobic fermentation of feed occurs as different microbes use the feed to produc e volatile fatty aci ds (VFA) that

PAGE 81

81 are absorbed by ruminants They also possess a secretory stomach called the abomasum where protein is hydrolyzed and bacteria are digested The ferme ntation that occurs in the forestomach breaks down cellulose, which is used as a nutrient by ruminants, and allows for access to cell content s (Leek, 2004) Methane accounts for 30 to 40 % of gases produced during enteri c fermentation in ruminants (Leek, 2004) and it can be eliminated through eructation, the lungs or the anus. Around 94 96% of the CH 4 produced in the rumen is eliminated through eructation, whereas the remaining CH 4 is eliminated through the lungs. The CH 4 hind gut sum s to 10% of total CH 4 produced and is mainly excreted through the lungs (89%), with a small part being eliminated via the anus (11%) (Murray et al., 1976) The microbial population inside the rumen is composed of d ifferent microorganisms. The primary bacteria degrade cellulose or starch, while secondary bacteria use end products from primary bacteria to obtain energy. Protozoa use several substrates as the energy source s including starch, polyunsaturated fatty acids and ruminal bacteria. The function of fungi in the rumen is yet not well understood. Microbes use mainly hydrolysis and anaerobic oxidation (remov al of hydrogen) reactions to obtain energy, creating a need for other reactions that use this hydrogen. This need is fulfilled by methanogenic reactions (Leek, 2004) mediated by obligate anaerobic microorganisms from th e Archaea family In the rumen, hydrolysis of cellulose and hemicellulose releases glucose, which is then fermented (Demeyer and Fievez, 2000) Because this reaction occurs in an anaerobic environment, it needs electron receptors other than oxygen for regenerating reaction co factors like NAD + NADP + and FAD + through oxidation, which guarantees that the substrate wi ll continue to be oxidized. However, the

PAGE 82

82 oxidation of the co factor is inhibited by the presence of H 2 in the rumen. Methanogenic bacteria use the hydrogen to produce CH 4 thus eliminating it from the ruminal environment and allowing for the re oxidation o f co factors (Demeyer and Fievez, 2000) Production of H 2 also occurs from the oxidation of pyruvate i nto acetate by the S organisms (Bryant et al., 1967) and is used by methanogenic bacteria to reduce CO 2 and produce CH 4 Experiments suggest that the presence of methanogenic bacteria improves the metabolism of pyruvate (Reddy et al., 1972) The CO 2 used in this reaction comes either from the decarboxylation reactions that occur during fermentation or from the neutralization of the H + by HCO 3 (Leek, 2004) The reactions that produce or use hydrogen inside the rumen are as follows (rskov et al., 1968 ; Moss et al. 2000) : Glucose 2 pyruvate + 4H Pyruvate + H 2 O acetate + CO 2 + 2H Pyruvate + 4H propionate + H 2 O 2 acetate + 4H butyrate + 2H 2 O CO 2 + 8H CH 4 + 2H 2 O One mole of glucose can form two moles of acetic acid, two moles of propionic acid or 1 mole of butyric acid, as seen above. The production of CH 4 in the rumen requires the presence of hydrogen to reduce CO 2 The formation of propionic acid utilizes hydrogen hence reducing CH 4 pr oduction, whilst increasing acetate formation in the rumen leads to an increase in available hydrogen for CH 4 production (rskov et al., 1968) The acetate to propionate ratio is a good indicator of methanogenic potential in diets. Forage fed cattle can have an acetate: propionate ratio up to 4:1, while animals on grain diets have an acetate: propionate ratio of 1:1 (Russel, 2002)

PAGE 83

83 There are different techniques used to measure CH 4 emissions from ruminants, each having advantages and disadvantages and being more adequate for specific experimental conditions (Storm et al., 2012) These are the SF 6 tracer gas technique (Johnson et al., 1994) the open circuit chamber technique (Brown et al., 1984) the tunnel technique (Murray et al., 19 99) micrometeorological mass difference (Harper et al., 2011) and the in vitro t echnique (Storm et al., 2012) A brief description of these follows. The sulfur hexafluoride (SF 6 ) tracer technique is used to estimate CH 4 emission rates by ruminants in production conditions, including grazing (J ohnson et al., 2007) and is useful to assess several aspects of feeding and nutrition such as influence of chemical and physical feed composition or additives on CH 4 emissions (Storm et al., 2012) It is important t o notice that this technique does not measure CH 4 produced in the hindgut that is not absorbed by the bloodstream (Johnson and John son, 1995) Some of the advantages of the SF 6 tracer technique summarized by Storm et al. ( 2012) are that animals are allowed to move freely when carrying the equipment, so that this technique can be used during grazing trials on which animal performance data is also measured Also, the results are useful to analyze variation in CH 4 emission rates within and between animals and can be related to production aspects such as mil k yield and weight gain The technique can be used to assess almost all feeding aspects important in animal production. Studies comparing the SF 6 tracer technique with chamber measurements (technique described below) found no significant difference among t hese methods (Johnson et al., 1994, 2007) In addition, animal acclimation for the use of the SF 6 tracer technique takes some days to a week, a short period of time when compared with the weeks necessary for the use of calorimetry chambers (Johnson et al., 1994) However, Storm et al. ( 2012)

PAGE 84

84 emphasize that this technique is not suitable to evaluate differences in the emissions through the day and that the results are more variable than in experiments made using chambers, hence mo re animals are necessary when performing animal CH 4 measurements using this technique. Chambers used in CH 4 measurements can be of the open or close circuit types (Storm et al., 2012) Open circuit indirect calorimetry chambers are used to analyze how modifications in nutritional state, activities or environmental temperature affect the metabolic rate of living beings during a short period (Brown et al., 1984) When used to measure CH 4 emissions by animals, this method is useful to analyze nu trition and feeding influence on emissions and can be used to obtain results of emissions of CH 4 through out the day (Storm et al., 2012) With this methodology, measurements are obtained using infra red or paramagnetic analyzers that give results proportional to the molecular density of the gas. These measurements are used in a series of equations to estimate volume of gas and corrected to conditions of temperature and pressure (STP). The rate of increase of the volume of a giv en gas inside the chamber is calculated as the sum of rate of volume production and flow of the gas inside the chamber minus the rate of volume flow outside the chamber (Brown et al., 1984) When measuring CH 4 emission rates using open circuit respiration calorimetry chambers, the concentration of CH 4 is analyzed in the air that comes in and out the chamber constantly during the experimental period using infrared analyzers (Johnson et al., 1994) Among the difficulties presented b y the use of the chamber methods in CH 4 emission evaluations are the training ess that can take several weeks and t restrictions to the

PAGE 85

85 behaviour may mak e this method unsuitable for extrapolation to real production situations (Johnson et al., 1994) Similar to the chamber s, the tunnel methodology analyz es the air f low ing in and out of a tunnel for CH 4 concentration, air speed and temperature (Murray et al., 1999) This system was designed to provide animals with an environment as similar to natural conditions as possible (Lockyer and Jarvis, 1995) When using this methodology CH 4 emissions by animals seem to be lower than other estimates (Lockyer and Jarvis, 1995; Murray et al., 1999) Another technique useful to measure CH 4 emissions from animals in either grazing or feedlot conditions is the micrometeorological mass difference approach where the flux of gases of interest in the free atmosphere is measured while maintaining animals in their natural environment (Ha rper et al., 2011) This technique requires less animal handling, therefore the animals maintain their natural behavior (Harper et al., 2011) In this technique CH 4 concentration in the air is analyzed using infrared gas analysis or gas chromatography at different heights around the area frequented by animals under study. This informat ion combined with data on wind speed and direction provides values of CH 4 flux (Harper et al., 1999) The micrometeorological technique is useful for analyzing production systems however little is known about its reliability (Stor m et al., 2012) The in vitro technique is a rapid and low cost approach when assessing how additives influence CH 4 emissions because it is fast and relatively inexpensive In this procedure, feeds are incubated in a container with a medium combining rumen fluid, a buffer solution and minerals D ata on CH 4 production are obtained from the analysis of gas composition and production 24 hours after incubation usually Degradation of feed must be analyzed to guarantee that CH 4 production is not

PAGE 86

86 decreased because of decreased feed degradation (Storm et al., 2012) Bhatta et al. ( 2008) analyzed CH 4 emissions by goats fed at 1.1 times maintenance with nineteen different diets using open circuit respiration chambers and in vitro methodology. The authors found that emissions from both methodologies were quite similar particularly for diets with high proportions of structural carbohy drates. Feeding Holstein cows at a maintenance level with five different diets, Bhatta et al. ( 2006) found correlation values between the SF 6 tracer and in vitro gas production technique of 0.75 and 0.94 for incubations of 24 and 48 hours, respectively Many factors affect CH 4 production by cattle, including manipulation of the diet fed to animals (processing, addition of lipid to diet, type of carbohydrate offered in feed), modification of ruminal microflora and management of animal intake (Johnson and Johnson, 1995) on which animal pe rformance is greatly dependent (Poppi et al., 1997) Martin et al. (2009) affirm aspects that should be considered in mitigat ion strategies are the decrease d in H 2 production in the rumen while maintaining feed digestion, modifying the end products from the H 2 utilization react ions and inhibiting methanogenic bacteria A nimal breeding has also been suggested as an alternative to reduce CH 4 emissi ons from ruminants not only directly, but also by improving animal efficiency (conversion of feed into products like meat and milk), productivity (reducing the number of animals needed for production) and reduction of losses such as empty reproductive cycl es (Wall et al., 2009) Some work ha s focused on feed manipulation for decreasing CH 4 production in ruminants by the use of additive s Bhatta et al. ( 2008) found several by products that reduced CH 4 sweet potato vine silage. Tannin has also been studied and resulted in lesser CH 4 emissions in go ats (Puchala et al., 2005) and sheep (Carulla et al., 2005) but tannin

PAGE 87

87 had no effect on cattle in reducing emissions, where it had bound protein (Beauchemin et al., 2007) Other additives like canola oil reduce CH 4 production in ruminants but negati vely affect animal intake and fiber digestibility, thus possibly affecting animal performance (Beauchemin and Mcginn, 2006) Moe and Tyrrell (1979) suggest ed that the natu re of carbohydrates in ruminant diets can affect production of CH 4 particularly at intakes higher than 1.5 time s main tenance. Increases in digestible cellulose in diets can drive fermentation from the formation of propionate to methanogenesis. According to Russel (2002) access to cell content s microbial cellulases cannot break. The presence of lignin can also hinder rumination and therefore decrease fe ed particle reduction. P roduction of CH 4 was found to be positively related to apparent cellulose digestibility (Pinares Patio et al., 2003) and to type of carbohydrates in high intak e diets (Moe and Tyr rell, 1979) Animal CH 4 measurements have also been performed in grazing trials, where management practices and pasture species effect on CH 4 emissions were investigated. McCaughey et al. (1999) found lower emissions for animals grazing alfalfa ( Medicago sativa L.)/grass (meadow bromegrass, Bromus biebersteinii Roem and Schult) than for animals grazing on grass only (meadow b romegrass, Bromus bieberstein ii Roem and Schult). Kurihara et al. (1999) found that emissions were hi gher for heifer s eating mature angleton grass ( Dicanthium aristatum ) with low er N content than for those grazing on r hodes grass ( Chloris gayana ) with medium effect on CH 4 emissions. The use of B MPs (intensive grazing, rotational stocking rate, overseeding ryegrass and fertilization) was found to result in lower CH 4 production when animals were grazing on bahiagrass and bermudagrass (Kurihara

PAGE 88

88 et al., 1999) McCaughey et al. (1997) found a significant difference in CH 4 emissions when studying animals on con tinuously stocked alfalfa grass (Roem and Schult, Bromus biebersteinii ; Russian wildrye, Psathyrostachys juncea (Fisch.) Nevski) at two stocking rates. In that study, the treatment with higher stocking rate had lower CH 4 day 1 when compared to lower stocking rates. However, management strategies and their effect on CH 4 emission still need further investigation D espite the importance of the beef industry in Florida, no measures of CH 4 emissions for animals on pasture exist in t his state. The objective of this study was to measure CH 4 emissions, animal performance and forage characteristics of animals continuously stocked on bahiagrass pastures at three stocking rates. Materials and Methods Experimental Site This experiment was carried out during the summer of 2012 at the North Florida Research and Education Center (NFREC) in Marianna (Jackson County), Florida ( ). The area used for the experiment had bahiagrass established in 2010 and was previously used for evaluations of herbage production under different N applications. In April 2012 the area was evenly fertilized with 57 kg ha 1 of N from ammonium nitrate (NH 4 NO 3 ). The experimental area was located in two experimental sites Figure 3 1 Soil types at the experimental sites are described on Table 3 1 The climate in the region where the experiment was conducted is classified as warm temperate, fully humid with hot summers (Cfa) according to the Kppen Geiger classification (Kottek et al., 2006)

PAGE 89

89 Treatments and Design The treatments in this experiment were defined as stocking rates (SR) of 1.2, 2.4 and 3.6 AU ha 1 w h ere one animal unit (AU) is 470 kg. The bahiagrass pastures were continuously stocked under the described fixed stocking rates during the grazing season of 2012, meaning that a fixed number of animals was kept in each experimental unit during the enti r e ex nitial liveweight was 347 29 kg an d it was estimated that animal avera ge daily weight gain would be approximately 0.46 kg/day (Stewart, 2003) during a 84 day period, which would result in an average weight of 366 kg in th e experimental period. Each experimental unit was 1.3 ha in area To achieve the target SR s of 1.2, 2.4 and 3.6 AU ha 1 the number of animals used was 2, 4 and 6 animals pasture 1 These treatments were chosen with the objective of achieving differences in animal intake and were determined on the estimated carrying capacity of the pastures according to the N fertilization rates applied in the experimental area (57 kg N ha 1 ), based on Twidwell et al. ( 1998) and Stewart ( 2003) The experiment was initiated o n 25 June and ended o n 18 September 2012 This time included three equal periods of 28 days each (P eriod 1: 25 June to 22 July; P eriod 2: 23 July to 19 August; P eriod 3: 20 August to 18 September). The experimental units were arranged in two blocks located in adjoining areas. Each block contained 6 pastures of 1.3 ha and so the total number of experimental units was 12. The three treatments were replicated four times such that two replicates of a treatment were randomly assign ed to each block (randomized complete block). Pasture and Animal Management Herbage mass, herbage accumulation and herbage nutritive value were measured. Herbage mass was estimated using the double sampling technique which

PAGE 90

90 associates direct and indirect m easurements of herbage mass (Santillan et al., 1979) The indirect measurement was taken using a disk meter, where a 0.25 m 2 aluminum disk was dropped always from the same height and settling height recorded. Every 28 days, three sites representing low, me dium and high herbage mass were sampled per experimental unit. At these sites, disk meter settling height was measured (indirect measurement) and with the help of a circular quadrat with the same area of the disk the forage beneath the disk was clipped (di rect measurement). The p asture samples were dried at 60 C until constant weight (48 to 72 hours). Calibration equations were obtained for each of the three sampling dates with disk height as a dependent variable to herbage mass ( Table 3 2 ). Every 14 days, disk meter settling height was measured at 30 location s per pasture, with sampling units separated by the same number of steps in each experimental unit. Average disk meter settling height was inputted in the calibration equation to estimate herbage mass. To estimate herbage accumulation rate (HAR), three 1 m 2 exclusion cages were installed per pasture at the beginni ng of the experimental period at sites where the dis k meter measure ment s were 1 cm from average disk meter measures for that pasture. Every 14 days, disk height was measured inside the cage. The cages were then relocated to sites in the pasture that had the same disk settling height (1 cm) of the rest of the pasture. Herbage mass was estimated using the calibration equation for the period. Herbage accumulation rate was calculated as follows: HAR (kg ha 1 day 1 ) = [Cage HM (kg ha 1 ) day 2 Cage HM (kg ha 1 ) day 1 ] / days Every 14 days, one sample composed of 20 25 hand plu cked subsamples per pasture was collected and used to measure nutritive value (crude protein, CP; neural detergent fiber, NDF; acid detergent fiber, ADF). After collection, the samples

PAGE 91

91 were dried at 60 C for 48 to 72 hours and ground to pass through a 1 mm screen. Data regarding herbage information for each period was obtain ed by taking the mean of values on days 1, 14 and 28 for each period. Animals (cross bred heifers) were weighted every 28 days in the morning after 16 hours of fasti ng. Data were then used to calculate animal performance as animal daily weight gain (ADG, kg day 1 ) as ADG (kg day 1 ) = (final weight initial weight) / 28 days Animal intake was estimated using the disappearance of herbage mass technique (Moore and Soll enberger, 1997), where the difference in dry herbage mass (HM) in grazed and ungrazed areas for a specific period of time are calculated to estimate how much forage is taken in by the group of animals in the pasture: Animal daily DM intake = {[(Available H M) (Residual HM)] / (period) / (SR) A nimal in take was estimated every 14 days. M easurements of herbage mass taken inside the exclusion cages represented ungrazed areas or available HM and measurements of herbage mass taken outside the exclusion cages wer e from grazed areas or residual dry HM. Therefore Animal daily DM intake = {[(Available DHM) (Residual DHM)]/ (period)/ (SR) where Animal daily DM intake = kg day 1 ; Available HM = kg HM ha 1 ; Residual HM = kg HM ha 1 ; Periods = 14 days; SR = stocking rate, AU ha 1 T hree weeks after the beginning of the experiment (by 17 July) the treatments did not show difference s in forage characteristics Since the objective of the study

PAGE 92

92 was to use SR to create differences in fora ge characteristics and study their effect s on CH 4 production by animals and animals performance, 1, 2 and 3 AU ha 1 were added to the experiment in the treatments 1.2, 2.4 and 3.6 ha 1 respectively. These animals remained in the experimental units for three weeks. Animal weight data from t hese animals were not considered in the analysis of average daily gain. Ruminal Methane Emissions Measurements Two CH 4 collection periods occurred starting 8 July and 10 September 2012 using the technique described in Johnson et al. ( 2007) One collection period consisted of five consecutive days of measurements and results of CH 4 production per day were averaged to obtain one value per animal per period. Methane emission rate was also averaged between the animals within each experimental unit. Twenty four animals previously used in another experiment for CH 4 measurements were used in this study so that two animals per experimental unit were used for the CH 4 measurements. In these animals, one permeatio n tube with a known SF 6 released from Washington State University, where they were calibrated in hot water bath at 39 C to achieve a constant release rate of between 1 and 2 mg d ay 1 Average SF 6 release rate was of 1 759 7 m g day 1 and nostrils wa s made through a capillary tube place on the top of the This capillary was attached to a halter and is connected through a valve to a PVC canister evacuated to 27 mm Hg ( Figure 3 2 ). Samples we re collected to a pressure of 13.5 mm Hg. Data from canisters returning with pressures above 10 mm Hg or below 20 mm Hg were not considered. Any animals which did not have at least three days of collection were not considered in the analysis. Collection time to achie ve the final pressure was 24 h and was regulated by the length of the capillary

PAGE 93

93 t ube. In the first day of each collection period a halter and an evacuated yoke were placed on the animals. After 24 h, the yoke collecting the sample was retrieved and a new, evacuated canister was placed on the animals, connected to the yoke, and the valv e was open ed This procedure was repeated for five consecutive days. On every collection day, two canisters were used to collect air and one placed at the center of each expe rimental area ( A and B Figure 3 1 ) that served to analyze background (ambient) concentrations of CH 4 and SF 6 The canisters containing the sample were taken to the lab for analysis lab and allowed to cool until they reached environmental temperature. They were then pressur iz ed with N until 3 psi and CH 4 and SF 6 concentrations were measured twice from each canister using gas chromatography (Agilent 7820A GC, Agilent Technologies, Palo Alto, CA) equipped with a flame ionization detector and a capillary column (Plo t Fused Silica 25 m 0.32 mm, c oating Molsieve 5A, Varian CP7536). Every collection day, pure standards were used to obtain a calibration curve for CH 4 and SF 6 After this, the canisters were rinsed ten times with pres sured air. To obtain the values of g CH 4 day 1 the following equation was used: QCH 4 = QSF 6 x [(CH 4 CH 4 A ) / (SF 6 SF 6 A )] where QCH 4 is CH 4 emissions rate, g day 1 ; QSF 6 is the SF 6 release rate CH 4 and SF 6 are the concentration in the yoke CH 4 A e SF 6 A are the concentrations in the ambient. The correction for background SF 6 concentrations were followed as suggested by (Lassey, 2013) QCH 4 outliers (mean 2 SD) were eliminated.

PAGE 94

94 Statistical Analysis Statistical analysis was performed using SAS. Response variables were herbage mass (HM), herbage accumulation rate (HAR), CP, ADF, NDF, average daily gain (ADG), body weight (BW), dry matter intake (DMI) ( kg animal 1 day 1 kg kg BW 1 day 1 ) and CH 4 ( g CH 4 day 1 g CH 4 kg BW 1 and g CH 4 kg DMI 1 ). Data were analyzed using repeated measures analysis of variance using P ROC MIX model of SAS. Periods were considered repeated measures and pasture (experimental unit) was considered to be the experimental unit SR, period and SR x period were fixed effect s Block and interaction s with block were considered random. Results R esults of the effect of treatment (different SR on continuously stocked bahiagrass), period and treatment x per iod interaction are summarized i n Table 3 3 Forage nutritive value was affected by period but no treatment or treatment x period effect were obs erved on animal or plat related response variables There was no effect of treatment or period on animal related variables, including the measures of CH 4 ( g CH 4 day 1 g CH 4 kg BW 1 and g CH 4 kg DMI 1 ). Discussion The values of HM (presented on Figure 3 3 ) found in this study are similar to those previously reported in the literature. Chambliss and Sollenberger ( 1991) repo rted values varying between 2600 and 63 30 for the summer in Florida. There was an increase in HM th rough the grazing season with higher values observed in August. This pattern was different than that observed by others (Johnson et al., 2001b; Stewart 2003) where HM peaked by the end of June and July. Th e difference might be related to the rainfall pattern observed during the experimental period. In Jul y, rainfall was 139 mm and reached 367 mm in August. There was no

PAGE 95

95 effect of treatment or period in HAR. Values of HAR ranged from 57 to 67 kg HM ha 1 day 1 These values are in the range of previously reported studies. Inyang et al. ( 2010) found values of 106, 128 and 118 kg HM ha 1 day 1 on continuously stocked bahiagrass. The lower HAR observed might have been caused by the lower rainfall observed in the months of June and July in the present study. Stewart Jr. ( 2003) evaluated bahiagrass under low, moderate and high management (1.2, 2.4 and 3.6 AU ha 1 with applications of 40, 120 and 360 kg N ha 1 respectively) The author found significantly higher HAR at moderate and high management treatment achieving values of as high as 37 kg ha 1 day 1 At low management levels, average HAR w as 19 kg ha 1 day 1 A tendency to treatment effect and an effect of period were observed for values of CP, but no treat ment x per iod interactions were observed ( Figure 3 10 ). Period also had a positive effect on CP with values 8.8, 9.2 and 9.5% P eriods 1, 2 and 3, respectively. Values of CP between 9.9 and 12% w ere reported for rotationally stock ed bahiagrass, with values increasing after September (Sollenberger et al., 1989 ) On continuously stocked bahiagrass, Stewart (2003) found values of CP varying between 9.2 and 14.3% with significant ly higher values of CP with more intense management (higher SR and N fertilizer applica tion). Cuomo et al. (1996) evaluated the nutritive value of Argentine and Pensacola bahiagrass under different cutting frequencies (20, 30 and 4 0 days) in two years. Values reported by Cuomo et al. (1996) were of 14% and 10 % for CP in early (late May to end of June) and late (mid August to mid September). For NDF, no treatment effect was observed. There was a period effect on NDF which increased through the grazing season ( Figure 3 9 ) Other authors also found an increase in NDF concentration of bahiagrass through grazing seasons.

PAGE 96

96 NDF increased through the grazing season from the range of 73 to 86% to the range of 79 to 82% (Grise et al., 2006) These values are higher than those observed in the present study. However, the values presented by Grise et al. ( 2006) refer to whole plant samples and do not represent the grazed proportion of the pasture. In this case, nutritive value might be underestimated (Sollenberger and Cherney, 1995) T here was no effect of treatment on ADF, but a period effect was observed with increasing ADF concentration as the season progressed Other studies have also found differences in ADF through grazing season in bahiagrass. Cuomo et al. ( 1996) found lower values of ADF for bahiagrass in early summer (May to June). No effect of SR or period on ADG was observed. The ADG was lower than what has been o bserved for continuously stocked bahiagrass with values varying between 0.22 to 0.18. Inyang et al. ( 2010) observed a line ar decrease in ADG with increasing of SR (4, 8 and 12 heifers ha 1 ). Low ADG of animals grazing bahiagrass and bermudagrass was also observed during summer and fall by DeRamus et al. ( 2003) Sollenberger et al. (1989) found an ADG of 0.41 kg head 1 day 1 when crossbred yearling steers grazed o n bahiagrass from April through November but encountered a large decrease in ADG in Augus t and Septem ber. Stewart ( 2003) evaluated bahiagrass under low, moderate and high management (1.2, 2.4 and 3.6 AU ha 1 with applications of 40, 120 and 360 kg N ha 1 respectively ) and found that ADG decreased with higher management levels. However, gain per hectare in that study was sm aller at low management levels. There was no effect of SR or period on DMI express ed as kg animal 1 day 1 or kg (kg BW) 1 day 1 Values of DMI (kg animal 1 day 1 ) found were higher than exp ected achieving values as high as 22 kg animal 1 day 1 Undi et al. ( 2008) compared the use of herbage mass disappearance technique with the use of

PAGE 97

97 alkanes and two prediction equations for measuring intake The author found that estimation of DMI using the herbage disappearance technique was greater than any of the other techniques, achieving values of 14.3 to 18.9 kg DMI day 1 The authors suggest that the trampling of the pasture might be the cause of this overestimation of DMI. It has been found that DMI is linearly related to CH 4 emissions (g animal 1 day 1 ) for cattle grazing tropical forages (Kurihara et al., 1999) However, no linear or quadratic relation ship was found between CH 4 (g day 1 ) and DMI (kg animal 1 day 1 ). There was no effect of treatment or period on CH 4 emissions expressed as g CH 4 day 1 g CH 4 kg BW 1 and g CH 4 kg DMI 1 ( Figure 3 11 Figure 3 12 and Figure 3 13 ) The coefficient of variation (CV, %) within animals was 79 and 86% in the first and second periods of CH 4 collection. Although these CVs are higher than those reported in the literature, l arge variation in CH 4 production between animals has been reported in other studies using the SF 6 technique In sheep fed chaf f ed lucern e hay diets, between individual variation was responsible for 70% of variation in CH 4 production (Pinares Patio et al., 2003) In cattle fed different forage s with different qualities CV (%) valu es for CH 4 emission above 26% were observed in both day to day and animal to animal measurements (Boadi et al., 2004) However, the authors emphasize d that these high values might be related to the diverse feeding conditions and small number of animals. This large CV is one of the reasons why no stat istical difference s between treatments or periods were detected. The values of CH 4 emiss ions varied between 140 to 879 g CH 4 day 1 The values found in the present study were higher than what was previously reported for animal grazing bahiagrass. For anim als grazing bahiagrass under normal or BMP management (overseeding of ryegrass and fertilization), lower values of CH 4 emissions were found under BMP management (DeRamus e t al., 2003) Values of

PAGE 98

98 daily CH 4 emissions for cows and heifers of 120 to 249 g day 1 and 86 to 166 g day 1 were reported (DeRamus et al., 2003) High values of CH 4 emissions were found for dairy cattle and lactating beef cattle. For dairy cattle grazing perennial ryegrass and white clover, Ulyatt et al. ( 2002) found varying values of CH 4 emissions through the season between 137 and 431 g day 1 The seasonal variation was closely related to milk production and estimated animal intake. For lactati ng beef cattle grazing alfalfa grass or bromegrass pastures, CH 4 emissions of 267 293 g day 1 were reported (McCaughey et al., 1999) McCaughey et al. (1997) found a difference in CH 4 emissions when studying animals on continuously stocked alfalf a ( Medicago sativa ) grass (Roem and Schult, Bromus biebersteinii ; Russian wildrye, Psathyrostachys juncea (Fisch.) Nevski) with two stocking rates. In that study, the treatment with higher stocking rate had lower CH 4 day 1 when compared to lower stocking rates. Kurihara et al. ( 1999) suggest ed that the CH 4 c onversion rate of tropica l forages is higher than those from temperate species, i. e., more energy is used in the production of CH 4 when ani mals are fed tropical forages. Although some authors conclude that DM intake can be used to predict CH 4 production by adult cattle on maintenance (Moe and Tyrrell, 1979) more studies are necessary to confirm if the linear relation ship between DM intake and CH 4 emissions (g animal 1 day 1 ) for cattle grazing tropical forages may be used for prediction (Kurihara et al., 1999) In that study, authors found that CH 4 emission rate is reduced at high feed levels and energy intake in comparison to animals kept on maintenance tract and consequent reduction in feed fermentation time. This difference in CH 4 energy Moe and Tyrrell (1979) in dairy cattle, and the authors conclude d that type of carbohydrate influences CH 4 in

PAGE 99

99 dairy cattle at higher intakes but do not significantly affect CH 4 prediction at lower intakes. Another study conducted by Boadi and Wittenberg (2002) where animals were fed hay varying in IVOMD concluded that DM intake could be used to predict CH 4 emissions from animals fed ad libitum and DMI represented 64% of variation in CH 4 emissions. Although several studies show a relationship between CH 4 emissions and intake, ther e was no correlation of these two variables in this study. Information on initial weigh t and ADG of each treatment on P eriods 1 and 3 were used in the IPCC (Tier 2) method to estimate emissions in g CH 4 day 1 Results are presented i n Figure 3 14 where the default value suggested by IPCC (2006) for the Tier 1 method is also presented. It is evident that the values from the field measures were much more variable than t hat obtained using the Tier 2 approach. Results obtained by the Tier 2 approach do not differ greatly due to treatment. This is probably related to the very similar ADG values encountered in the field experiment and used in the simulation, since it was fou nd that the model is very sensitive to the ADG variable. The Tier 1 and Tier 2 approaches seem to underestimate CH 4 emission s in this study however a more detailed study on prediction capacity is necessary. Conclusions The objective of this study was to measure CH 4 emissions, animal performance and forage characteristics of animals grazing continuously stocked bahiagrass under three different stocking rates. The SR applied affected HM and chemical value of bahiagrass, however it did not affect animal res ponse variables. CH 4 emissions were also not affected by SR. On average, animals grazing bahiagrass emitted 393 g CH 4 day 1 and a high variation was found on emissions. This is the first CH 4 measurements for grazing animals in Florida. The IPCC Tier 2

PAGE 100

100 and Tier 1 approaches seem to underestimate emissions of CH 4 by cattle. Due to the information about emissi ons from different agriculture related sources. This information may help not only to improve model use but also give a more clear understanding of management approaches that can be used to reduce or avoid GHG emissions.

PAGE 101

101 Table 3 1 Soil types in experimental sites A and B. Source: USDA, 2013. Experimental site A Map Unit Name Percentage of area Chipola loamy sand, 0 to 5 percent slopes 2.9 Orangeburg loamy sand, 2 to 5 percent slopes 49.3 Red Bay fine sandy loam, 2 to 5 percent slopes 47.9 Experimental site B Fuquay coarse sand, 0 to 5 percent slopes 25.2 Greenville fine sandy loam, 2 to 5 percent slopes 28.0 Orangeburg loamy sand, 2 to 5 percent slopes 46.7 Red Bay fine sandy loam, 2 to 5 percent slopes 0.1 Table 3 2 Herbage mass double sample regression equations. Period 1: 25 June to 22 July ; P eriod 2: 23 July to 19 Aug. ; P eriod 3: 20 Aug to 18 Sep. Period Equation R 2 1 y = 6.3604x + 5.2975 0.87 2 y = 6.6018x 0.9663 0.89 3 y = 6.8917x + 17.804 0.81 Table 3 3 Effect of stocking rate (1.2, 2.4 and 3.6 AU ha 1 ) on response variables in 2012 period and treatment x period interaction on experimental variables. Treatment Period Treatment Period HM (kg ha 1 ) ** ** ** HAR (kg ha 1 day 1 ) NS NS NS ADF (%) NS ** NS NDF (%) NS ** NS CP (%) NS ** NS DMI (kg animal 1 day 1 ) NS NS NS DMI (kg kg BW 1 day 1 ) NS NS NS ADG (kg animal 1 day 1 ) NS NS NS BW (kg) NS NS NS CH 4 (g CH 4 day 1 ) NS NS NS CH 4 (g CH 4 kg BW 1 ) NS NS NS CH 4 (g CH 4 kg DMI 1 ) NS NS NS NS: non significant ** : significant at 1% (P<0.01)

PAGE 102

102 Table 3 4 Forage herbage mass (HM ) (kg ha 1 ) and herbage accumulation rate ( HAR ) (kg ha 1 day 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: Ju ne 25 to July 22; P eriod 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18 Period 1 Period 2 Period 3 Stocking rate (AU ha 1 ) HM (kg ha 1 ) SE 1.2 3235 3956 4787 198.4 2.4 3239 3661 4063 198.4 3.6 2967 2735 2983 198.4 HAR (kg ha 1 day 1 ) 1.2 67 57 57 8.5 2.4 65 61 59 8.5 3.6 62 57 67 8.5 Table 3 5 Forage chemical composition response given by CP (%), NDF (%) and ADF (%) to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: Ju ne 25 to July 22; P eriod 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18 Period 1 Period 2 Period 3 Stocking rate (AU ha 1 ) CP (%) SE 1.2 8.6 8.6 8.7 1.5 2.4 8.5 9.0 9.2 1.5 3.6 9.4 10.0 10.8 1.5 NDF (%) 1.2 63.2 64.8 65.4 0.6 2.4 63.2 65.6 65.4 0.6 3.6 64.5 64.2 63.6 0.6 ADF (%) 1.2 35.1 36.1 37.1 1.0 2.4 35.6 36.9 36.9 1.0 3.6 35.7 36.5 36.4 1.0

PAGE 103

103 Table 3 6 Average daily gain (ADG, kg animal 1 day 1 ) BW (kg animal 1 ), DMI (kg animal 1 day 1 ) and DMI (kg BWl 1 day 1 ) response to three stocking rates (1.2, 2.4, 3.6 AU ha 1 ) in 2012 Period 1: Ju ne 25 to July 22; P eriod 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18 Period 1 Period 2 Period 3 Stocking rate (AU ha 1 ) ADG (kg animal 1 day 1 ) SE 1.2 0.03 0.12 0.18 0.2 2.4 0.22 0.03 0.03 0.2 3.6 0.12 0.02 0.01 0.2 BW (kg animal 1 ) 1.2 344.9 347.1 349.2 4.7 2.4 348.4 351.0 351.1 4.7 3.6 353.3 355.8 351.9 4.7 DMI (kg animal 1 day 1 ) 1.2 20 21 22 3.2 2.4 18 16 17 3.2 3.6 15 14 15 3.2 DMI (kg kg BWl 1 day 1 ) 1.2 5.9 6.0 6.5 0.9 2.4 5.2 4.6 4.9 0.9 3.6 4.1 4.0 4.2 0.9

PAGE 104

104 Table 3 7 Response of CH 4 emissions expressed as g CH 4 day 1 g kg BW 1 and g kg DMI 1 to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012. Period 1: Ju ne 25 to July 22; P eriod 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18 Period 1 Period 3 Stocking rate (AU ha 1 ) CH 4 (g day 1 ) SE 1.2 485.3 237.3 337.3 193.8 2.4 140.3 167.8 333.4 167.8 3.6 180.1 335.6 879.4 193.8 CH 4 (g kg BW 1 ) SE 1.2 1.4 0.7 1.0 0.6 2.4 0.4 0.5 0.9 0.5 3.6 0.5 1.0 2.5 0.6 CH 4 (g kg DMI 1 ) SE 1.2 41.7 20.3 13.6 16.6 2.4 8.3 14.4 22.6 14.4 3.6 12.8 28.7 70.2 16.6

PAGE 105

105 Figure 3 1 Map of experimental sites A and B located at the North Florida Research and Education Center (NFREC), Marianna, Florida Figure 3 2 Animal with CH 4 collection device. A: capillary tube placed on halter; B: collecting canister. Picture by Mar ta Moura Kohmann, 2012. A B

PAGE 106

106 Figure 3 3 Herbage mass (kg ha 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: Ju ne 25 to July 22; P eriod 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18 Figure 3 4 Herbage accumulation rate (HAR) (kg ha 1 day 1 ) i response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: Ju ne 25 to July 22; P eriod 2: July 23 to Aug. 19; Period 3: Aug. 20 to Sep. 18

PAGE 107

107 Figure 3 5 Dry matter intake (DMI) (kg animal 1 day 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18. Figure 3 6 Dry matter intake (DMI) per kg body weight (BW) (kg kg 1 day 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18.

PAGE 108

1 08 Figure 3 7 Average daily gain (ADG) (kg animal 1 day 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to July 22; P eriod 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18. Figure 3 8 Acid detergent fiber (ADF) (%) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18.

PAGE 109

109 Figure 3 9 Neutral detergent fiber (N DF) (%) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18. Figure 3 10 Crude protein (C P) (%) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to Ju ly 22; P eriod 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18.

PAGE 110

110 Figure 3 11 Methane production (g CH 4 animal 1 day 1 ) respon se to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18. Figure 3 12 Methane production (g CH 4 kg BW 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18.

PAGE 111

111 Figure 3 13 Methane production (g CH 4 kg DMI 1 ) response to three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18. Figure 3 14 Measured and s imulated methane production (g CH4 animal 1 day 1 ) in three stocking rates (1.2, 2.4 and 3.6 AU ha 1 ) in 2012 The value for IPCC Tier 1 is a default value for beef cattle in the United States (IPCC, 2006), while values for IPCC Tier 2 were simulated using the Tier 2 methodology (IPCC, 2006) using the average daily gain from Periods 1 and 3 in the field experiment presented in this chapter (Table 3 6). Period 1: June 25 to July 22; Period 2: July 23 to Aug. 19; P eriod 3: Aug.20 to Sep. 18.

PAGE 112

112 LIST OF REFERENCES Archbolt Biologi cal Station. 2013. Archbolt Biological Station. http://www.archbold station.org (accessed 7 Oct 2013). Basarab, J., V. Baron, ". Lpez Campos, J. Aalhus, K. Haugen Kozyra, and E. Okine. 2012. Gr eenhouse Gas Emissions from Calf and Yearling Fed Beef Production Systems, With and Without the Use of Growth Promotants. Animals 2(4): 195 220. Bauer, L.R., and D.M. Hamby. 1991. Relative Sensitivities of Existing and Novel Model Parameters in Atmospheri c Tritium Dose Estimates. Bauer and Ha. Radi at. Prot. Dosimetry 37(4): 253 260. Beauchemin, K.A., J.H. Henry, S.M. Little, T.A. McAllister, and S.M. McGinn. 2010. Life cycle assessment of greenhouse gas emissions from beef production in western Canada: A c ase study. Agric. Syst. 103(6): 371 379. Beauchemin, K.A., and S.M. Mcginn. 2005. Methane emissions from feedlot cattle fed barley or corn diets. J. Anim. Sci. 83: 653 661. ts 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. 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. Agric. 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. Boadi, D.A., and K.M. Wittenberg. 2002. Methane production from dairy and beef heifers fed forages differing in nutrient density using the sulphur hexafluoride ( SF 6 ) tracer gas technique. Can. J. Anim. Sci. 82: 201 206. 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(3): 445 453.

PAGE 113

113 Brown, D., T.J. Cole, M.J. Dauncey, R.W. Marrs, and P.R. Murgatroyd. 1984. Analysis of gaseous exchange in open circuit indirect calorimetry. M ed. Biol. Eng. Comput. 22: 333 338. Bryant, M.P., E.A. Wolin, and Wolin, M. J., and R.S. Wolfe. 1967. Methanobacillus omelianskii, a Symbiotic Association of two Species of Bacteria. Arch. Mikrobiol. 59: 20 31. Burton, G.W. 1955. Breeding Pensacola Bahiagrass, Paspalum notatum; I. Method of Reproduction. Agron. J. 47(7): 311 314. Campolongo, F., J. Cariboni, and A. Saltelli. 2007. An effective screening design for sensitivity analysis of large models. Environ. Model. Softw. 22(10): 1509 1518. Carulla, J.E., M. Kreuzer, A. Machmller, and H.D. Hess. 2005. Supplementation of Acacia mearnsii tannins decreases methanogenesis and urinary nitrogen in forage fed sheep. Aus t. J. Agric. Res. 56(9): 961 970. Casey, J.W., and N.M. Holden. 2006. Quantification of GHG emissions from sucker beef production in Ireland. Agric. Syst. 90: 79 98. Chambliss, C.G., and L.E. Sollenberger. 1991. Bahiagrass: The foundation of cow calf nutr ition in Florida. Proc. 40th Florida Beef Cattle Shortcourse Gainesville. Univ. of Florida, Gainesville, p. 74 80 Cukier, R.I. 1973. Study of the sensitivity of coupled reaction systems to uncertainties in rate coefficients. I Theory. J. Chem. Phys. 59( 8): 3873 3878 Cukier, R.I., H.B. Levine, and K.E. Schuler. 1978. Nonlinear Sensitivity Analysis of Multiparameter Model Systems. J. Comput. Phys. 26: 1 42. Cukier, R.I., J.H. Schaibly, and K.E. Shuler. 1975. Study of the sensitivity of coupled reaction sy stems to uncertainties in rate coefficients. III. Analysis of the approximations. J. Chem. Phys. 63(3): 1140 1149 Cuomo, G.J., D.C. Blouin, D.L. Corkern, J.E. Mccoy, and R. Walz. 1996. Plant morphology and forage nutritive value of three bahiagrasses as a ffected by harve st fre quency. Agron. J. 88: 85 89. Ann. Zootech. 49: 95 112. DeRamus, H.A., T.C. Clement, D.D. Giampola, and P.C. Dickison. 2003. Methane emissions of beef cat tle on forages: efficiency of grazing management systems. J. Environ. Qual. 32(1): 269 277.

PAGE 114

114 Dijkstra, J., D.S.C.N. Heather, D.E. Beever, and J. France. 1992. Simulation of Nutrient 22): 2239 2256. E nvironmental P rotection A gency 2005. Emission Facts: Average Carbon Dioxide Emissions Resulting from Gasoline and Diesel Fuel. U. S. Gov. Print. Office Washington DC. E nvironmental P rotection A gency. 2013a. Inventory of U. S Greenhouse Gas 2011. U. S. Gov. Print. Office Washington, DC. Environmental Protection Agency 2013b. ANNEX 3 Methodological Descriptions for Additional Source or Sink Categories. U. S. Gov. Print. Office Washington, DC. Florida Department of Agriculture and Consumer Services. 2012. Florida a griculture by the numbers. http://www.nass.usda.gov/Statistics_by_State/Florida/Publications/ Agriculture_Statistic al_Directory/2012/2012%20FL%20Ag%20by%20the%20Numb ers(FASD).pdf (accessed 12 Sept. 2013). Foley, P. A ., P. Crosson, D.K. Lovett, T. Kenny. 2011. Whole farm systems modelling of greenhouse gas emissions from pastoral suckler beef cow production systems. Agric. Ecosyst. Environ. 142(3 4) : 222 230. Grise, M.M., L. Sollenberger, J.C. Dubeux J r., S. Interrante, R.L. Stewart and A. Pelissari. 2006. Biomass allocation and forage quality in a pensacola bahia grass pasture in north central Florida. Acta Sci. Anim. Sci. 28(4): 375 383. 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., O.T. Denmead, J.R. Freney, and F.M. Byers. 1999. Direct measurements of methane emissions from grazi ng and feedlot cattle. J. Anim. Sci. 77 1401. Inyang, U., J.M.B. Vendramini, L.E. Sollenberger, B. Sellers, A. Adesogan, L. Paiva, and A. Lunpha. 2010. Forage Species and Stocking Rate Eff ects on Animal Performance and H Sci. (20): 1079 1085. IPCC. 2006 2006 IPCC Guidelines for National Greenhouse Gas Inventories In: Eggleston, H.S., L Buendia, K Miwa, T Ngara, and K Tanabe, editors IGES, Japan. IPCC. 2007a. Climate Change 2007: Synthesis Report. Contribution of Working Groups I, II and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. In: Pachauri, R. K.; A. Reisinger editors. Geneva, Switzerland.

PAGE 115

115 IPCC. 2007b. Climate change 2007: the Physical Science Basis. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. editors, Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Summary for Policymakers Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. IPCC. 2013. Organization. http://www.ipcc.ch (accessed 20 Oct. 2013). Johnson, K., M. Huyler, H. Westberg, B. Lamb, and P. Zlmmerman. 1994. Measurement of methane emissions from ruminant livestock using a SF 6 tr acer techn ique. Environ. Sci. Technol. 28: 359 362. Johnson, K.A., and D.E. Johnson. 1995. Me thane emissions from cattle. J. Anim. Sci. 73: 2483 2492. Johnson, D.E., H.W. Phetteplace, and A.F. Seidl. 2001a. Methane, nitrous oxide and carbon dioxide emissions from ruminant production system. In : Chwalibog, A., Jakobsen, K. editors Energy metabolism in animals. Proceedings of the 15th symposium on energy metabolism in animals, Snekkersten, Denmark. p. 161 164 Johnson, C.R., B.A. Reiling, P. Mislevy, and M.B. Hall. 2001b. Effects of nitrogen fertilization and harvest date on yield, digestibility, fiber, and protein fractions of tropical grasses. J Anim. Sci. 79: 2439 2448. Johnson, K.A., H.H. Westberg, J.J. Michal, and M.W. Cossalman. 2007. The SF 6 tracer technique: methane measurement from ruminant S. In : M akkar, H.P.S., Vercoe, P.E. editors Measuring Methane Production from Ruminants. Springer Netherl ands. p. 33 67. Karimi Zindashty, Y., J.D. Macdonald, R.L. Desjardins, D.E. Worth, J.J. Hutchinson, and X.P.C. Verg. 2012. Sources of uncertainty in the IPCC Tier 2 Canadian livestock model. J. Agric. Sci. 150: 556 569. Kebreab, E., K. A. Johnson, S.L. A rchibeque, D. Pape, and T. Wirth. 2008. Model for estimating enteric methane emissions from United States dairy and feedlot cattle. J. Anim. Sci. 86(10): 2738 48. Kottek, M., J. Grieser, C. Beck, B. Rudolf, and F. Rubel. 2006. World Map of the Kppen Geige r climate classification updated. Meteorol. Zeitschrift 15(3): 259 263. 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 emissio n from farm operations. Environ. Int. 30(7): 981 90.

PAGE 116

116 Lash, J., and F. Wellington. 2007. Competitive advantage on a warming planet. Harv. Bus. Rev. 85(3): 94 102, 143. Lassey, K.R. 2007. Livestock methane emission: From the individual grazing animal through national inventories to the global methane cycle. Agric. For. Meteorol. 142: 120 132. Lassey, K.R. 2013. On the importance of background sampling in applications of the SF6 tracer technique to determine ruminant methane emissions. Anim. Feed Sci. Technol. 180: 115 120. Leek, B.F. 2004. Digestion in the Ruminant Stomach. p. 438 474. In Reece, W.O., ed itor New York. Lockyer, D.R., and S.C. Jarvis. 1995. The measurement of methane losses from grazing animals. Environ. Pollut. 90(3): 383 90. Main, K.N., and E.S. Menges. 1997. Archbold Biological Station fire management plan. Archbolt Biological Station. http://www.archbold station.org/station/documents/ publicationspdf/FireMgmtPlan20 04.pdf (accessed 30 Aug. 2013). Martin, C., D.P. Morgavi, and M. Doreau. 2009. Methane mitigation in ruminants: from microbe to the farm scale. Animal 4(3): 351 65. McCaughey, W.P., K. Wittenberg, and D. Corrigan. 1997. Methane production by steers on pasture. Can. J. Anim. Sci. (77): 519 524. McCaughey, W.P., K. Wittenberg, and D. Corrigan. 1999. Impact of pasture type on methane production by lactating beef cows. Can. J. Anim. Sci. (79): 221 226. Miller, R.K., L.C. Rockwell, D.K. Lunt, and G.E. Carst ens. 1996. Determination of the flavor attributes of cooked beef from cross bred Angus steers fed corn or barley based diets. Meat Sci. 44(4): 235 43. Minson, D.J. 1990. Forage in Ruminant Nutrition Academic Press, San Diego, California. Mislevy, P., an d L.S. Dunavin. 1993. Management and Utilization of Bermudagrass and Bahiagrass in South Florida. In: 42nd Beef Cattle Short Course Gainesville. p. 5 7 Moe, P.W., and H.F. Tyrrell. 1979. Methane Production in Dairy Cows. J. Dairy Sci. 62(10): 1583 1586

PAGE 117

117 Moir, R.J. 1961. A note on the relationship between the digestible dry matter and the digestible energy content of ruminant diets. Aust. J. Exp. Agric. Anim. Husb. 1: 24 26. Monod, H., C. Naud, and D. Makowski. 2006. Uncertainty and sensitivity analysis for crop models. p. 55 100. In: Wallach, D., Makowski, D., Jones, J.W. ed itors Working with Dynamic Crop Models. 1st ed. Elsevier The Netherlands p. 55 100. Morris, M.D. 1991. Factorial sampling plans for preliminary computational exper iments. Technometrics 33(2): 161 174. Moss, A.R., J. P. Jouany, and J. Newbold. 2000. Methane production by ruminants: its contribution to global warming. Ann. Zootech. 49(3): 231 253. Murray, B.Y.R.M., A.M. Bryant, and R.A. Leng. 1976. Rates of production of meth an in the rumen and large large intestine o f 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 for measuring methane emissions from sheep. J. Agric. Acience 133: 439 444. Newman, Y., J. Ven dramini, and A. Blount. 2011. Ba hiagrass ( Paspalum notatum ): Overview and Management. http://edis.ifas.ufl.edu/ag342 (accessed 12 July 2013). Nijdam, D., T. Rood, and H. Westhoek. 2012. The price of protein: Review of land use and carbon footprints from life cycle assessments of animal food products and their substitutes. Food Policy 37(6): 760 770. Nkrumah, J.D., E.K. Okine, G.W. Mathison, K. Schmid, C. Li, J.A. Basarab, M.A. Price, Z. Wang, and S.S. Moore. 2 006. Relationships of feedl ot feed efficiency performanc and feedi ng behavior with metabolic rate methane producti on and energy partitioning in beef cattle. J. Anim. Sci. (84): 145 153. rskov, E.R., W.P. Flatt, and P.W. Moe. 1968. Fermentation balance approach to estimate extent of fermentation and efficiency of volatile fatty acid formation in rumin ants. J. Dairy Sci. 51(9): 1429 1435. Pelletier, N., R. Pirog, and R. Rasmussen. 2010. Comparative life cycle environmental impacts of three beef p roduction strategies in the Upper Midwestern United States. Agric. Syst. 103(6): 380 389. Phetteplace, H.W., D.E. Johnson, and A.F. Seidl. 2001. Greenhouse gas emissions from simulated beef and dairy livestock systems in the United States. Nutr. Cycl. Agro ecosystems 60: 99 102.

PAGE 118

118 Phillips, W.A., S.W. Coleman, D.G. Riley, C.C. Chase, and H.S. Mayeux. 2006. Stocker and Feedlot Performance and Carcass Characteristics of Purebred Angus and Romosinuano Steers. Prof. Anim. Sci. 22: 8 13. Pinares Patio, C.S., M.J. U lyatt, 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. Poppi, D.P., S.R. Mclennan, S. Bediy e, A. de Vega, and J. Zorrila Rios. 1997. F orage Smith, L.D. Bailey and L.M.C. Ferreira, editors, P roceedings of the 18th International Grassland Congress. Winnipeg, Manito ba, Canada. p. 307 322. 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. Sci. 83: 182 186. Reddy, C. a, M.P. Bryant, and M.J. Wolin. 1972. Characteristics of S organism isolated from Methanobacillus omelianskii J. Bacteriol. 109(2): 539 45. Rittenhouse, L.R., C.L. Streeter, and D.C. Clanton. 1971. Estimating digestible energy from digestible dry and organic matter in diets of grazing ca t tle. J. Range Manag. 24 (1): 73 75. Rs, E., C. Sundberg, P. Tidker, I. Strid, and P. Hansson. 2013. Can carbon footprint serve as an indicator of the environm ental impact of meat production ? Ecol. Indic. 24: 573 581. Russel, J.B. 2002. Rumen Microbiology and Its Role In Rumina nt Nutrition. Ithaca, NY. Ruviaro, C.F., M. Gianezini, F.S. Brando, C. A Winck, and H. Dewes. 2012. Life cycle assessment in Brazilian agriculture facing worldwide trends. J. Clean. Prod. 28: 9 24. Saltelli, A., S. Tarantola, F. Campolongo, and M. Ratto. 2004. Sensitivity analysis in practice: a guide to assessing scientific m odels. John Wiley & Sons, Ltd. Saltelli, A., S. Tarantola, and K.P. S. Chan. 1999. A quantitative model independent method for global sensitivity analysis of model o utput. Technometi rcs 41(1): 39 56. Sollenberger, L.E., and D.J.R. Cherney. 1995. Evaluating forage production and quality. In: Barnes, R ., Miller, D.A., Ne lson, C.J., editors Forages: The science of grassland agriculture. Iowa State Univ. Press, Ames, IA. p. 97 110. Solle nberger, L.E., G.F. Lima, J.F. Holderbaum, W.E. Kunkle, J.E. Moore, and A.C. Hammond. 1997. C attle weight gain and sward animal nitrogen relationships in

PAGE 119

119 grazed Hemarthria altissima pastures p. 43 44. In: Baker, M.J. Crush, J.R., Humphreys, L.R., ed itors Proc eedings if the 17th Int ernational Grassl and Congress Keeling and Mundy, Ltd, Palmerston North, NZ. p. 43 44. Sollenberger, L.E., G.A. Rusland, C.S. Jones, K.A. Albrecht, and K.L. Gieger. 1989. Animal and forage responses on rotationally graze d li mpograss and bahiagrass p astures. Agron. J. 81: 760 764. Soussana, J. F., P. Loiseau, N. Vuichard, E. Ceschia, J. Balesdent, T. Chevallier, and D. Arrouays. 2004. Carbon cycling and sequestration opportunities in temperate grasslan ds. Soil Use Manag. 20(2): 219 230. Stewart, R.L. 2003. Management intensity effects on animal performance and herbage response in bahiagrass pastures. M. Sc. thesis, University of Florida, Gainesville Stewart, R.L., L.E. Sollenberger, J.M.B. Vendramini, S.M. Interrante, and Y.C. Newman. 2007. Herbage and animal responses to management intensity of continuously stocked bahiagrass p astures. J. Agron. 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 emission from rumin ants. Animals 2(4): 160 183. Tw idwell, E., W.D. Pitman, and G.J. Cuomo. 1998. Bahiagrass production and management. http://www.lsuagcenter.com/NR/rdonlyres/82A096A6 7F6F 4306 853B F49A1D98AFEF/78032/pub2697BahiagrassHIGHRES.pdf (accessed 20 July 2013). Ulyatt, M.J., K.R. Lassey, I.D. Shelton, and C.F. Walker. 2002. Seasonal variation in methan e emission from dairy cows and breeding ewes grazing ryegrass / white clover pasture in New Zealand. New Zeal. J. Agric. 45(4): 217 226. Undi, M., C. Wilson, K.H. Ominski, and K.M. Wittenberg. 2008. Comparison of techniques for estimation of forage dry ma tter intake by grazing beef cattle. Can. J. Anim. Sci. 88(4): 693 701. U nited States D epartment of A griculture 1989. Soil Survey of Hardee County Florida. U. S. Gov. Print. Office Washington, DC. U nited S tates D epartment of A griculture 2013. Web Soil Survey. United States Department of Agriculture. http://websoilsurvey.sc.egov.usda.gov/ App/ WebSoilSurvey.aspx (accessed 10 Oct 2013). Van Soest, P.J. 1982. Nutritional Ecology of the Ruminant: Ruminant metabolism, nutritional strategi es, the cellulolytic fermentation and thechemistry of forages and plant fibers. O & B Books, Inc., Corvallis, Oregon.

PAGE 120

120 Wall, E., G. Simm, and D. Moran. 2009. Developing breeding schemes to assist mitigation of greenhouse gas emissions. Animal 4(3): 366 376. Wiedman, T., and J. Minx. 2008. A D In: Per tsova, C.C., editor Ecological Economics Research Trends. Nova Science Publishers, Hauppauge, NY. p. 1 11. Xu, C., and G. Gertner. 2011. Understanding and comparisons of differ ent sampling approaches for the Fourier Amplitudes Sensitivity Test (FAST). Comput. Stat. Data Anal. 55(1): 184 198. 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 me thane emissions through improving efficiency of energy utilization and productivity in lactating dairy cows. J. Dairy Sci. 93(6): 2630 8.

PAGE 121

121 BIOGRAPHICAL SKETCH Marta Moura Kohmann was born in 1988 in a small town in the countryside of southern Brazil, Carazinho. She decided to e ngineering in the capital of Rio Grande do Sul State, Porto Alegre She finished her undergraduate degree at the Federal University of Rio Grande do Sul (UFRGS) in 2011. During her undergraduate studies, she was approved to go to the University of Florida in a study abroad funded by CAPES and FIPSE for one semester, duri ng which she took classes and worked on research. After her graduation, she started her g raduate studies at the University of Florida in the Department of Agricultural and Biologica l Engineering with emphasis in c limatology. In fall 2013, she completed her Master of Science (M. Sc.) degree at the University of Florida.