LONG TERM IMPACTS OF MANAGEMENT INTENSIFICATION ON SOIL CARBON DYNAMICS IN SUBTROPICAL GRASSLANDS By JULIUS B ADEWOPO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
2 2014 Julius Babatunde Adewopo
3 To everyone who has been a part of this exciting journey. To all friends, acquaintances, and family who have encouraged, inspired, and challenged me. To my wife (Phebean) and daughters (Pearl and Jewel) who cheered me on, and whose impeccable support kept me going through the highs and the lows.
4 ACKNOWLEDGMENTS my application and acceptance into the Soil and Water Science department. I thank the graduate school for sele cting me as a recipient of the prestigious Graduate Alumni A ward which covered the costs related to my tuition and sustenance. I am most grateful to my advisor, Dr. Maria Silveira, my co advisor, Dr. Stefan Gerber, and other members of my graduate committe e, Dr. Lynn Sollenberger and Dr. Tim Martin, who supported me immensely through the twists and turns of my doctoral study. This research was funded in part by USDA Sustainable Agriculture term m anagement impact on soil administrative and technical s upport of the staff at UF Range Cattle Research and Education Center at Ona Florida Their dedicated support (including, assistance with tools, troubleshooting of field and laboratory equipment, processing of relevant documents etc. ) enabled me to accomplish research and academic tasks in a timely manner. I would also like to recognize the contribution and help of colleagues and interns who worked with me under intense field conditions to ensure that scheduled tasks we re completed within the designated timeframe. Furthermore, the dedication of the University st aff, including secretaries and j anitors provided a n e nvironment conducive for my learning and professional development. My heartfelt gra titude goes to my wife and twin daughters, who persevered through my countless hours of absence from home during days and nights, and motivated me to complete this race. Fin ally, although this will continually be on my lips, my heart magnifies the Father of grace who raised me up to abound in every good thing, and leads me on step by step.
5 TABLE OF CONTENTS P age ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 LIST OF ABBREVIATIONS ................................ ................................ ................................ ........ 11 ABSTRACT ................................ ................................ ................................ ................................ ... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 14 2 MANAGEMENT INTENSIFICATION IMPACTS ON SOIL AND ECOSYSTEM CARBON STOCKS IN SUBTROPICAL GRASSLANDS ................................ ................... 20 Background ................................ ................................ ................................ ............................. 20 Materials and Methods ................................ ................................ ................................ ........... 22 Study Site ................................ ................................ ................................ ......................... 22 Native rangeland ................................ ................................ ................................ ...... 22 Silvopasture: ................................ ................................ ................................ ............. 23 Sown pasture ................................ ................................ ................................ ............ 23 Experimental Design and Sampling Protocol ................................ ................................ .. 24 Below ground sampling and analysis ................................ ................................ ...... 24 Above ground biomass sampling ................................ ................................ ............. 26 Statistical Analysis ................................ ................................ ................................ .......... 28 Results ................................ ................................ ................................ ................................ ..... 28 SOC Stocks ................................ ................................ ................................ ...................... 28 Soil N Stocks ................................ ................................ ................................ ................... 29 Particulate and Mineral Associated Organic C and N ................................ .................... 30 Root Biomass C ................................ ................................ ................................ ............... 30 Aboveground Biomass C ................................ ................................ ................................ 31 Ecosystem C ................................ ................................ ................................ .................... 31 Discussion ................................ ................................ ................................ ............................... 31 Impact of Grasslan d Intensification on SOC Stocks and Soil Bulk Density ................... 32 Changes in SOC Fractions and their N Content with Management Intensifica tion ........ 33 Management Intensification Impact on Live Biomass and Ecosystem C ....................... 35 Summary ................................ ................................ ................................ ................................ 37
6 3 IMPACT OF MANAGEMENT INTENSIFICATION ON PARTICLE SIZE SOIL CARBON FRACTIONS IN SUBTROPICAL GRASSLANDS: EVIDENCE FROM 13 C NATURAL ABUNDANCE ................................ ................................ ................................ ... 43 Background ................................ ................................ ................................ ............................. 43 Materials and Methods ................................ ................................ ................................ ........... 45 Study Site ................................ ................................ ................................ ......................... 45 Experimental Desig n ................................ ................................ ................................ ....... 47 Soil Sampling and Analyses ................................ ................................ ............................ 47 Statistical Analysis ................................ ................................ ................................ .......... 49 Results ................................ ................................ ................................ ................................ ..... 49 Total Organic C, POC and Cmin Concentration ................................ ............................. 49 13 C) of Bulk Soil, POC, and Cmin Fractions ............................ 50 C 3 and C 4 derived C Composition of Bulk Soil, POC and Cmin C Fractions ............... 51 Discussion and Conclusion ................................ ................................ ................................ ..... 51 Management Intensification Effects on Soil C ................................ ................................ 51 Management Intensifica 13 C) Signature in Soil C Fractions ....... 53 Changes in Stability and Source of C Allocated to Particle size Pools .......................... 54 4 MANAGEMENT INTENSIFICATION EFFECTS ON AUTOTROPHIC AND HETEROTROPHIC SOIL RESPIRATION IN SUBTROPICAL GRASSLANDS .............. 62 Background ................................ ................................ ................................ ............................. 62 Materials and Methods ................................ ................................ ................................ ........... 65 Study Sites ................................ ................................ ................................ ....................... 65 Experimental Design ................................ ................................ ................................ ....... 65 Partitio ning and Measurement of Soil Respiration ................................ ......................... 66 Statistical Analysis ................................ ................................ ................................ .......... 67 Results ................................ ................................ ................................ ................................ ..... 68 Daily and Seasonal Ambient Temperature and Rainfall ................................ ................. 68 Soil Respiration under Different Grassland Ecosystems ................................ ................. 69 Total soil respiration (R S ) ................................ ................................ ......................... 69 Heterotrophic respiration (R H ) ................................ ................................ ................. 69 Autotrophic respiration (R A ) ................................ ................................ .................... 70 Changes in Soil Temperature (S Temp ) and Moisture (S Moist ) ................................ ............ 70 Measured S Temp and S Moist ................................ ................................ ........................ 70 Variability of R S R H and R A with S Temp and S Moist ................................ ................. 71 Discussion and Conclusion ................................ ................................ ................................ ..... 72 Ma nagement Intensification Effects on Total Soil Respiration Rates ............................. 72 Response of Heterotrophic Respiration to Management Intensification ......................... 73 Response of Soil Autotrophic Respiration to Management Intensification .................... 75 Management Intensification Impacts Respiration Sensitivity to Soil Temperature and Moisture. ................................ ................................ ................................ ............... 76
7 5 APPLICATION OF THE PROCESS BASED DNDC MODEL FOR PREDICTING IMPACT OF MANAGEMENT INTENSIFICATION ON SOIL CARBON DYNAMICS IN SUBTROPICAL GRASSLAND ECOSYSTEMS ................................ ..... 88 Background ................................ ................................ ................................ ............................. 88 Materials and Meth ods ................................ ................................ ................................ ........... 90 Model Selection and Description ................................ ................................ .................... 90 Study Area ................................ ................................ ................................ ....................... 91 Field Observation of Soil Respiration and Abiotic Variables ................................ ......... 92 Model Validation and Application Baseline and Intensified Management Conditions ................................ ................................ ................................ .................... 93 Model Validation Performance Metrics ................................ ................................ .......... 94 Resu lts and Discussion ................................ ................................ ................................ ........... 95 DNDC Validation under Native Rangelands (baseline) and Sown Pasture (Intensified Management) Conditions ................................ ................................ ......... 95 Predictive Performance of DNDC Relative to Field Observations ................................ 99 Long term R S and Soil C Sequestration under Alternative Management Scenarios. .... 100 Summary ................................ ................................ ................................ ............................... 103 6 CONCLUSIONS AND SYNTHESIS ................................ ................................ .................. 115 LIST OF REFERENCES ................................ ................................ ................................ ............. 121 BIOGRAPH ICAL SKETCH ................................ ................................ ................................ ....... 138
8 LIST OF TABLES Table P age 1 1 Soil carbon (C) pools, soil C fractions, forms measured and sensitivity to management change ................................ ................................ ................................ ........... 19 2 1 S oil organic C (SOC) and N stocks at different depths under native rangeland, sown pasture, and silvopasture ecosystems ................................ ................................ ................. 38 3 1 Soil carbon (C) in bulk soil, particulate organic C, and miner al associated C fractions at different soil depths as affected by grassland management intensification. .................. 57 3 2 Percent C 3 and C 4 derived soil carbon (C) in bulk soil, particulate organic C, and mineral associated C fractions as affected by grassland management intensification. ..... 59 4 1 Impact of grassland management intensification on in situ soil respiration, soil temperature, and soil moisture. ................................ ................................ .......................... 80 4 2 Effect of grassland management intensification on relationship of soil respiration with soil temperature and moisture. ................................ ................................ ................... 81 4 3 Grassland management intensification effects on temperature sensitivity (Q 10 ) of soil respiration. ................................ ................................ ................................ ......................... 82 5 1 Main DNDC input data requirement and sources from which they were derived ........... 106 5 2 Soil related input parameters adopted for DNDC s imulation ................................ .......... 107 5 3 Definition of DNDC modeled management intensity scenarios and corresponding management practices ................................ ................................ ................................ ...... 108 5 4 Descriptive statistics of DNDC simulated and field observed soil respiration and abiotic control factors in subtropical native rangeland (baseline) and sown pasture (intensified management). ................................ ................................ ................................ 109 5 5 Performance of DNDC in simulating soil respiration and abiotic control variables under native rangeland (baseline) and sown pasture (intensified managemen t) conditions ................................ ................................ ................................ ......................... 110
9 LIST OF FIGURES Figure P age 2 1 Aerial imagery of the study fields obtained from Google Earth and layout of the 5 sampling quadrats (20 m 20 m) on a diagonal transect. ................................ ................. 39 2 2 Bulk density of soil samples collected at three profile depths (0 10 cm, 10 20 cm, 20 30 cm) in native rangeland, sown pasture, and silvopasture ecosystems.. ................... 40 2 3 Root biomass carbon (C) at three soil profile depths (0 10 cm, 10 20 cm, 20 30 cm) depth in native rangeland, sown pasture, and silvopasture ecosystems.. ........................... 41 2 4 Ecosystem carbon (C) stocks under native rangeland, sown pasture, and silvopasture ecosystems.. ................................ ................................ ................................ ....................... 42 3 1 Isotopic 13 13 C) of carbon in bulk soil at different soil depths (0 10, 10 20, and 20 30 cm), as affected by grassland management intensification. ....... 60 3 2 Isotopic 13 13 C) of mineral associated (<53m Cmin) and particulate sized (>53m POC) soil carbon fractions across the sampled so il depth (0 30 cm), as affected by grassland management intensification. ................................ .... 61 4 1 Images showing the fabricated exclusion boxes for partitioning soil respiration components, and the field installation of the boxes (only native rangeland shown). ........ 83 4 2 Images showing field set up for measuring soil respiration and abiotic control variables.. ................................ ................................ ................................ ........................... 84 4 3 Weather data and measured soil climate variables in a management intensity gradient of subtropical grassland ecosystems. ................................ ................................ ................ 85 4 4 Winter soil respiration rates under a management intensity gradient of subtropical grassland ecosystems.. ................................ ................................ ................................ ....... 86 4 5 Summer soil respiration rates in a management intensity gradient of subtropical grassland ecosystems. ................................ ................................ ................................ ........ 87 5 1 The structure of Denitrification and Decomposition (DNDC) model showing two component parts.. ................................ ................................ ................................ ............. 111 5 2 DNDC simulation of soil carbon (C) dynamics during spin up run (year 1 240), 30 year period after conversion from baseline to intensively managed sown pasture (year 210 240). ................................ ................................ ................................ .............. 112 5 3 Daily outputs of soil respiration (total soil respiration R S heterotrophic soil respiration R H and autotrophic soil respiration R A ), and abiotic control factors (soil temperature S Temp and soil moisture S Moist ). ................................ ........................ 113
10 5 4 Measured and DNDC si mulated soil respiration variables (total soil respiration R S heterotrophic soil respiration R H and autotrophic soil respiration R A ), and abiotic control factors (soil temperature S Temp and soil moisture S Moist ). ................................ 114
11 LIST OF ABBREVIATIONS Cmin Mineral associated carbon (Mg C ha 1 and g kg 1 ) EF Modeling efficiency POC Particulate organic carbon (Mg C ha 1 and g kg 1 ) PON Particulate organic nitrogen (Mg C ha 1 and g kg 1 ) R 2 Coefficient of determination R A Autotrophic soil respiration ( g CO 2 m 2 h 1 ) R H Heterotrophic soil respiration ( g CO 2 m 2 h 1 ) and ( M g C ha 1 yr 1 ) RMSE Root mean square error RPD Ratio of prediction to deviation R S Total soil respiration ( g CO 2 m 2 h 1 ) and ( M g C ha 1 yr 1 ) S Moist Soil volumetric moisture content (%) SOC S oil organic carbon (Mg C ha 1 and g kg 1 ) S Temp Soil temperature (C)
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy LONG TERM IMPACTS OF MANAGEMENT INTENSIFICATION ON SOIL CARBON DYNAMICS IN SUBTROPICAL GRASSLANDS By Julius B. Adewopo May 2014 Chair: Maria L. Silveira Coc hair: Stefan Gerber Major: Soil and Water Science Proper management of grassland ecosystems for improved productivity can enhance their potential to sequester atmospheric CO 2 in the soil. However, soil C dynamics in subtr opical grassland ecosystems is poorly understood, and knowledge is lacking about the long term impacts of grassland management intensification on soil C within this environmental condition. This study was conducted to assess long term (> 22 years) effect o f grassland management intensification on ecosystem and soil C dynamics (including stocks, particle size fractions, and losses), based on a gradient of management intensities ranging from native rangeland (lowest), silvopasture (intermediate), to sown past ure (highest). The replicated (n = 2) experimental sites Cattle Research and Education Center, Ona, Florida ( ) and were characterized by homogeneous topography and the same predominant soil series and climatic conditions. The silvopasture and sown pastures were established by clearing native rangeland fields ~35 years ago, and were managed to mimic moderate and optimal management intensity (including N fertilization lev els and grazing practices), commonly practiced by grassland managers within the region. The impacts of grassland intensification on soil and ecosystem C stocks (above and below ground biomass )
13 soil C fraction 13 C natural isotopic abundance and in si tu soil respiration fluxes were determined The process based denitrification decomposition (DNDC) model was tested for predicting soil C dynamics in this biome. Management intensification increased soil C stocks (0 to 30 cm depth) from 41 Mg ha 1 in the native rangeland to 62 and 69 Mg ha 1 in sown pasture and silvopasture, respectively. Silvopastoral ecosystem favored C sequestration in the more stable mineral associated C pool (41.8 Mg ha 1 ; ~1 00% increase) compared with the native rangeland ( 20 Mg ha 1 ) Although the mixed C 3 C 4 composition of silvopasture limited full elucidation of C sources, the loss of relic stable C fraction (decreased from 13.2 to 10.6 Mg ha 1 ) and the increase in recent (C 4 derived) C (increased from 6.7 to 17.5 Mg ha 1 ) in sown pasture suggests that C input s from the sown grass species (b ahiagrass Paspalum notatum Flgge) contributed to stable C sequestrati on. Increasing management intensity from native rangeland to sown pasture elevated soil respiration. For instance, m ean sum mer heterotrophic respiration increased from 0.34 g CO 2 m 2 hr 1 to 0.46 g CO 2 m 2 hr 1 while temperature sensitivity ( Q 10 ) increased from 1.48 to 2.29, respectively, suggesting a potential for management induced positive temperature feedback on CO 2 Despite its potential as a tool for assessing changes in soil C dynamics under intensified management condition (coefficient of determination, R 2 DNDC perform ance in predicting measured soil respiration and abiotic control variables indica tes need for further model recalibration ( ratio of performance to deviation, RPD < 1.8 and modeling eff iciency, EF < 0.68 ). This study shows that grassland management intensification has enhance d soil C sequestration and the use of strategic management p ractices such as integration of trees can improve soil C stability and reduce soil C loss under similar subtropical conditions. These findings are important to inform and support management decisions and policies that can promote long term sustainability o f subtropical grassland biome s.
14 CHAPTER 1 INTRODUCTION General Introduction Terrestrial ecosystems are major sink s for carbon (C), and they mitigat e the increasing concentration of carbon dioxide (CO 2 ) in the atmosphere (IPCC, 2007) Terrestrial plants absorb atmospheric CO 2 and reduce its C to organic forms through photosynthesis (Field et al., 1989; Steinbeiss et al., 2008; Fynn et al., 20 09) About 50% of the photosynthesized C is returned to the atmosphere as CO 2 through plant respiration, while the remaining portion is incorporated in leaves, stems, roots, and subsequently transferred to the soil (Brady and Weil, 2008; Hillel and Rosenzweig, 2009; Fynn et al., 2009) For undisturbed systems, most of the biomass stored in the living plant material is eventually transferred to soil organic matter pools via abov e and below ground inputs (Sobecki et al., 2001; Zhu, 2010) Global estimates of terrestrial C range from 1500 to 2200 Pg C stored within the 1 m surface depth of soil (Eswaran et al., 1993; Sundquist, 1993; Batjes, 1996; Paustian, 2007) and 600 Pg in the vegetation (Schimel, 1995; Houghton, 2012) totaling up to three fold that of the atmospheric C (Sundquist, 1993; Schuman et al., 2002; Powers et al., 2011) Soil C is constituted by organic and inorganic C forms In the top 1m of soil, average soil organic C stocks (which contains a range of organic materials) are ~1500Pg C, while the inorganic C stocks (made up of carbonates and bicarbonates) are ~720 Pg C (Sombroek et al., 1993) Soil inorganic C is characterized by very slow turnover rate, hence, it is less responsive to management than soil organic C and constitutes a relatively low percentage of the total C pool in most cultivated soils (Izaurralde, 2005; Fynn et al., 2009) Soil organic C (a lso SOC) constitutes ~ 48 to 58% of soil organic matter (SOM), and is made up of diverse biological materials, living micro and meso fauna, fresh plant residues, humus from decomposition, inert (humic and char) substances, and
15 silica occluded plant C or phytoliths (Gregorich et al., 1994; Parr and Sullivan, 2005; Wilke, 2005) Although there exists a wide range of estimates, research findings have indicated that soil contains the largest terr estrial pool of organic C, and each ton of carbon stored in soils removes ~ 3.67 tons of CO 2 from the atmosphere (Batjes, 1996; Fynn et al., 2009) A conceptual designation of C pools has been used by researchers to establish difference between C compounds that cycle at different rates in the ecosystem (Fynn et al., 2009) Therefore, the different components of SOC are partitioned into three broad pools based on the turnover rates or mean residence time (MRT) (Allen et al., 2010 ; Table 1 1) The active or labile C pool has a MRT of < 10 years and is made up of fresh plant residues such as fine roots, microbial bio mass, and particulate organic C. The slow C pool is constituted by humus or c lay sorbed C and forms 30 60% of SOC depending on the soil type, clay content and mineralogy (Allen et al., 2010) It has a MRT of 10 to 200 years depending on dominant climatic conditions The relatively slow turnover has been attributed to the increasing aromaticity of the C, spatial inaccessibility to micro organisms and extracellular enzymes, and sorption of SOM on mineral surfaces and its interaction with mineral particles (Sollins et al., 1996; von Ltzow et al., 2007) The resistant or passive soil organic C has a MRT of >100 years (up to thousands of years) and is ma inly composed of charcoal C (30%), organo mineral metal complexed organic C, and the silica occluded C (Skjernstad et al., 1999; Parr and Sullivan, 2005; Kgel Knab ner et al., 2008) Soil C sequestration is determined by the net balance between C inputs and outputs (Singh and Gupta, 1977; Sombroek et al., 1993; Fynn et al., 2009) ther efore changes over time reflect the overall dynamics between litter input and decomposition processes. Land use management influences the input from plants, and they can also facilitate changes in the chemical composition and decomposition rates of C (Batjes, 1999; Post and Kwon, 2000;
16 Dawson and Smith, 2007) The extent of such influence is determined by the type of management practices and intensity, (i.e. plant bio mass harvest, nutrient application, soil tillage practices, and fire management). Unique combinations of these factors have varying impacts on the long term fate of C dynamics in different ecosystems, including grassland ecosystems. Grassland Management an d Soil C The U.S. farm security and rural investment act of 2008 defined grassland as land on which the vegetation is dominated by grasses, grass like plants, shrubs, and forbs (NRCS GRP, 2009) Grasslands provide and s upport important landscape functions and values, including provision of food and fiber. The physical, chemical, and biological a ttributes of grasslands enable them to support diverse ecosystem functions such as nutrient cycling, hydrologic cycling, biodive rsity, livestock and wildlife support (Parton et al., 1995; NRCS GRP, 2009) Unlike forested ecosystems where a large proportion of the C is stored above ground (Shaffer and Ma, 2001; Pregitzer and Euskirchen, 2004; Arevalo et al., 2009; Peichl et al., 2012) most (up to 90%) of the C in grasslands is stored in the soil (Parton et al., 1995; Schuman et al., 2001; Schuman et al., 2002; Pucheta et al., 2004; McSherry and Ritchie, 2013) Soil organi c matter is a critical component of grassland sustainability and productivity and its main functions include i) serving as source for plant nutrients ii) improvement of water h olding capacity of the soil, iii ) formation and stabilization of soil aggregat es and iv) provision of habitable condition s for soil microbial diversity (Follett et al., 2001; Weil and Magdoff, 2004) These functions are promoted by the ability of SOM to buffer soil temper ature and pH, regulate water quality and hydrology, increase the porosity and surface area of soils, and enhance biomass decomposition and nutrient cycling in the pedos p here (Evrendilek et al., 2004; Pattanayak et al., 2005; Fynn et al., 2009)
17 Similar to other ecosystems, C stored in grassland vegetation and soil can change in response to a wide array of management and environmental factors (Schuman et al., 2002; Franzluebbers, 2010; Derner and Jin, 2012) However quantifying such management related changes is imperative for accurate assessment of C storage within grassla nds, and to reliably account for national (or regional) C inventory and global C balance (Sobecki et al., 2001; Follett et al., 2001; Fynn et al., 2009; U.S. EPA, 2012) In Florida, grassland ecosystems cover 2.5 million ha (17.5% of total land) and support 1.71 million cattle and calves (USDA NASS, 2009) Florida is characterized by subtropical climate with over 42% of its area covered by forests and approximately 25% covered by crops and pasture (Mulkey, 2007) However, there are growing concerns over current and predicted shifts in land use in Florida as population and competition among alternative land use s continues to increase (Zwick and Carr, 2006; Mulkey, 2007) Hence, similar to other subtropical grasslands (White et al., 2000) Fl (rangelands) ecosystems have been converted to extensive agriculture or subjected to more intensive grassland management in order to support the cow calf industry, which is the 10 th largest in the U.S. (USDA NASS, 2009) The uniqueness of climatic characteristics in this region, such as relatively high precipitation and warm temperatures associated with fluctuating water tables, flat topography, and sand y soil texture favor rapid decomposition of SOM and leaching of organic compounds (Nair et al., 2007; Haile et al., 2010; Stephenson, 2011) Hence, management strategies and land use changes that have the potential to enhance soil C sequestration and preserve soil resources are essential for environmental sustainability of the subtropical grassland biome within the region. Numerous researchers have evaluated grassland man agement intensification impacts on soil C, mainly involving conversion from one management system to another, or increasing the
18 intensity of adopted management practices within a management system (Biondini et al., 1998; Mazancourt et al., 1998; Conant et al., 2001; Billings et al., 2006; Dubeux et al., 2007) In a global review of 115 studies, Conant et al., (2001) concluded that changes in management practices (such as fertilization, grazing management, conversion from native vegetation, sowing of legume, introduction of earthworm, and irrigation) increased soil C sequestration, with reported rates ranging from 0.11 to 3.04 Mg C ha 1 yr 1 depending on vegetation and climate. Most of these studies are focused on temperate grasslands with few findings from tropical grasslands, and it is uncertain how management inte nsification will impact long term soil C within subtropical grassland biomes. Th e main objective of this dissertation research was to provide a better understanding of the long term impact of grassland management intensification on soil C dynamics in a su btropical grassland biome The specific goals of this study were to: i ) evaluate changes in soil and ecosystem C stocks in response to grassland intensification (conversion of low intensity native rangeland in to high intensity sown pasture systems ), ii) ch aracterize soil C distribution into various pools as affected by increasing management intensity iii ) evaluate the impacts of grassland intensification on soil C loss through respiration and iv ) test the applicability of a process based model for predicting management induced changes in soil C dynamics.
19 Table 1 1 Soil carbon (C) pools, soil C fractions, forms measured and sensitivity to management change Soil C pool Soil C fractio n Pool C/total C (%) Form measured Turnover period (years) Sensitivity to management change Labile (active) C Soluble fresh residues. Living micro and meso flora and fauna Particulate organic C Light fraction 0.5 5 1 10 1 40 1 30 Microbial and root exudates Microbial biomass >53m <1.6 2g/cm 3 <0.1 <5 <10 <10 Very rapid Rapid Rapid Rapid Slow C Humus Clay complexed C 30 50 30 60 Total organic C particulate organic C <2m 10 200 10 100 Medium Medium Resistant (Passive) C Charcoal C Phytoliths Carbonates 1 30 1 30 0 30 Resistant to chemical oxidation Oxidized at ~1300C Release of CO 2 by acid treatment >100 Millenia >1000 Slow Very slow Very slow Source: Allen et al., 2010.
20 CHAPTER 2 MANAGEMENT INTENSIFICATION IMPACTS ON SOIL AND ECOSYSTEM CARBON STOCKS IN SUBTROPICAL GRASSLANDS Background Globally, grasslands support complex and interrelated ecosystem functions such as nutrient cycling, water storage and quality preservation, biodiversity, livestock production, and wildli fe (Parton et al., 1995; NRCS GRP, 2009) Within the past 2 to 3 decades, grasslands have been subjected to more intensive management in order to meet higher productivity demands, driven by growing human population (FAO, 1993; White et al., 2000; Mulkey, 2007) In the 21 st century, land use conversion has become more societally sensitive (White e t al., 2000) and increasing demand for agricultural products is fostering expansion of intensive agro ecosystems into subtropical rangelands, as predicted by FAO (1993) Rapid changes in land use and management inten sity within grasslands can lead to detrimental consequences such as soil degradation (Bruce et al., 1999; Fernndez et al., 2010) loss of biodiversity (White et al., 2000; Weil and Magdoff, 2004) and enhanced greenhouse gas emissions (Conant et al., 2000; Schuman et al., 2001) Depending on the management techniques adop ted, grassland can act as a source or sink of atmospheric CO 2 (Schuman et al., 2001; Follet et al., 2001; Sobecki et al., 2001; Penman et al., 2003) M oreover, both intensity and type of management (e.g. frequency of prescribed fire, fertilization regime, grazing intensity, stocking management, and conversion of native lands) applied to optimize grassland productivity, can affect ecosystem C and soil organic matter storage (Cambardella and Elliott, 1992; Conant et al., 2001; Fynn et al., 20 09 ) For example, management intensification may favor the buildup of labile and short lived soil C fractions in lieu of stable C pools (Buyanovsky and Wagner, 1995; Batjes and Sombroek, 1997) thereby limiting the potential for sustained C sequestration and climate mitigation. Hence, it is imperative to develop managem ent strategies focused on
21 maintaining and improving soil resources as well as overall ecosystem sustainability (Batjes and Sombroek, 1997) A significant proportion (~30%) of the global soil organic C is sequestered in subtropical and tropical ecosystems (Dalal and Carter, 1999) and little is known about C and N dynamics in relation to the broader context of grassland management systems. Most research studying the effects of gr assland management on soil organic matter is limited to temperate regions (Batjes and Sombroek, 1997; Silveira et al., 2013) often focused on individual management practices (Conant et al., 2001) and predominantly conducted as short term experiments (Dubeux et al., 2006a; Chan et al., 201 0; Peichl et al., 2012; Silveira et al., 2013) Several researchers have attempted to elucidate changes in C and N stocks after land use conversion (Corre et al., 1999; Guo and Gifford, 2002; Pucheta et al., 2004; Sharrow and Ismail, 2004; McLauchlan et al., 2006) but they focused primarily on transitions from broad agricultural or forest land cover type to grassland pastur e and vice versa. There is limited knowledge of the impact of land use change on ecosystem C within grassland ecosystems, especially in relation to the increasing global trend of conversion of native grasslands to more intensively managed grassland ecosys tems (White et al., 2000) Addressing this knowledge gap is not only critical for improved accuracy in modeling of ecosystem C responses to global land use changes (Batjes and Sombroek, 1997; Fynn et al., 2009) but it is also crucial for development of sustainable management systems that can improve C sequestration and mitigation of atmospheric CO 2 concentration (Sobecki et al., 2001) The objective of this study was to evaluate the long term impacts (20+ years) of grassland intensification on SOC and overall ecosystem C stocks, taking advantage of a unique long term experimental setup in a native subtropical rangeland.
22 Ma terials and Methods Study Site The experimental site is located at the University of Florida Range Cattle Research and relatively homogenous slope (<5%), and subtropica l climate with 10 year average annual precipitation of ~1206 mm and temperature of ~21.5C. The study was conducted on adjacent fields of three grassland management systems that were consistently maintained for over 20 years. The three management systems c onsist of a gradient of management intensities ranging from native rangeland (lowest), silvopasture (intermediate), to sown pasture (highest). Prior to the establishment of the adjacent silvopasture and sown pasture, the entire site was native rangeland fo r livestock grazing. Each collocated ecological unit was replicated twice (6 ha each; Figure 2 1). The dominant soil series was the same across the sites and consisted of Ona and Immokalee fine sand (sandy siliceous, hyperthermic Typic Alaquods). This soil was developed on parent material of sandy marine deposits (Soil Survey Staff, 1999; NRCS Websoil Survey, 2013) Native rangeland The native rangeland ecosystem consisted primarily o f saw palmetto ( Serenoa repens Bartr.) and a wide variety of grass genera including Andropogon, Panicum, Aristida, and Schizachyrium spp. (Kalmbacher et al., 1984) This ecosystem was never fertilized, but it had been subjected to periodic burning (every 3 years), occasional livestock grazing activities (<60 days per year), and herbivory by wildlife, all typical features of rangeland in this region. Each experimental unit was grazed only during winter at rate of 1 25 animal unit days ha 1 yr 1 (a 500 kg animal grazing for 1 day equals 1 animal unit day) A nimal s were fed a daily supplement of
23 warm season grass hay and sugarcane molasses at 1.5 to 1.9 kg cow 1 day 1 and 0.7 kg cow 1 day 1 respectively, throughout the grazing period. Silvopasture: The silvopasture system has been managed for ~ 22 years and consists of slash pine ( Pinus elliottii ) trees planted in double rows (1.2 m along 2.4 m between rows), and bahiagrass ( Paspalum notatum ) planted in alleys (12.2 m wide). The vegetation of the silvopasture was established in order of bahiagrass first, followed by slash pine, on a previously native rangeland. Native vegetation was suppressed by burning followed by plowing (~ 45 cm deep) and disking with a dual tand em disk harrow until there was no vegetation on the soil surface. These experimental units received periodic applications of 67 kg N ha 1 yr 1 as ammonium nitrate. No fertilizer was applied in the years of 1993 1997, 2000, 2002, 2008, 2009, and 2011. Grazi ng of the silvopasture began in March 1993, 18 months after planting the trees, and has continued from March September every year. Each experimental unit was rotationally stocked for 7 months each year, with a 2 week grazing period followed by a 5 week re sting period. Stocking rate was 207 animal unit days ha 1 yr 1 Animals were supplemented in the pasture with warm season grass hay and sugarcane molasses at 1.5 to 1.9 kg cow 1 day 1 and 0.7 kg cow 1 day 1 respectively, from January to April. The managem ent conditions that were applied represent limited inputs of fertilizer and moderate levels of biomass removal through grazing. Sown pasture The sown pasture system consisted of 32 year old bahiagrass stand, which was managed to provide forage for grazin g livestock. Each experimental unit was stocked at a rate of 360 animal unit days ha 1 yr 1 Grazing occurred on a rotational basis for 7 months, with 1 week residence period followed by 1 week rest period between each grazing event. A nimal were fed a dai ly supplement of warm season grass hay and sugarcane molasses at 1.5 to 1.9 kg cow 1 day 1
24 and 0.7 kg cow 1 day 1 respectively, from January to April. Ammonium nitrate fertilizer was applied at an annual rate of 67 kg N ha 1 yr 1 and dolomite was applied in 2001 and 2008 at a rate of 730 and 550 kg ha 1 respectively. These management conditions are typical of the beef cow calf production system in Florida. Experimental Design and Sampling Protocol The study was based on a comparative mensurative experime ntal design (Hurlbert, 1984; Arevalo et al., 2009) because the two replicate fields (6 ha each) of each management ecosystems are collocated; in other words, the treatment replicates were not spatially dispersed. This experimental design has an underlying assumption that the soil properties of the three e cosystems (native rangeland, silvopasture, and sown pasture) were similar prior to the conversion and designation of each ecological unit (Hurlbert, 1984; Arevalo et al., 2009) Fa ctors such as the uniformity of initial land use, the flat terrain, and the limited potential for variation in the sandy textural class of the soil which were all formed from the same parent material (Kalmbacher et al., 1984; Kalmbacher et al., 1993) lends credence to the assumption that the soils were similar before the management changes were imposed. However, similar to general limitation of most long term ecological studies (Janzen, 1995; Millar and Anderson, 2004) no data were available on the soil properties before the land use conversion (i.e. conversion to more intensively managed ecosystems), and some spatial var iability is bound to be inherent within the study area. Below ground sampling and analysis Soil sampling was conducted in June/July 2012 which corresponds to the period of maximum ecosystem productivity and biochemical cycling rates (decomposition/mineral ization). An initial offset (30 m) was established from the edge of each field in order to avoid potential edge effects, and five quadrats (20 m 20 m), spaced ~75 m apart, were marked out along a
25 diagonal transect within the inner boundary (Figure 2 1). Four randomly located soil cores (diameter of 2.2 cm) were sampled from each quadrat at soil depths of 0 10, 10 20, and 20 30 cm and composited within a depth for C and N analysis, while one random soil core was sampled in each quadrat for bulk density det ermination. Soil samples were air dried and sieved (2 mm sized sieve), and the modified version of the procedure described by Cambardella and Elliott (1992) was adopted to separate SOC and N into two particle size fractions: particulate organic C (POC) an d N (PON) which corresponded to the > 53m sized particles, and mineral associated C and N which corresponded to the < 53m sized particles. Carbon and N concentrations were determined by dry combustion using a ThermoFlash EA 1112 elemental analyzer. Soil C and N stocks in each ecological unit were calculated based on the C and N concentration and the bulk density at each depth. For bulk density, the soil samples collected at each soil depth interval were dried at 105C until constant weight, and weighed. Bulk density was computed by dividing the dry weight by the soil core volume. Below ground root biomass samples were collected using the AMS hydraulic powerprobe (Arts Manufacturing and Supply Inc., American Falls, ID) equipped with soil coring drills (dia meter of 5 cm). Three soil cores were randomly sampled in each quadrat at the 0 10, 10 20 and 20 30 cm depths. Samples were air dried and subsequently dispersed in water to separate roots from soil particles through sieving (sieve size = 250m). Root sa mples were dried to constant weight at 65C, and final weight was recorded after 3 days. The dried root samples were combusted at 550C for 5 hours to determine ash concentration. Carbon and N concentrations were determined by dry combustion using a Thermo Flash EA 1112 elemental analyzer. Root biomass C presented here is expressed on ash free basis.
26 Above ground biomass sampling The above ground biomass measurement was conducted in mid summer, in order to estimate the biomass during the period of peak annual primary production. Sampling was strategically conducted to capture the peak production under each management system such th at: 1) sown pasture was sampled 5 days after the cows had been moved out of the pastures, i.e. two days before the next herd moved in; 2) silvopasture was sampled ~ 4 weeks after the cattle had been moved out, and 3) native rangeland was sampled after ~3 m onths of not being grazed by cattle. Different sampling strategies were employed across the ecological units because of the different vegetation composition. Above ground biomass C in sown pasture: A double sampling method (Wilm, 1944) was adopted to quantify above ground biomass in the sown bahiagrass pastures. The compressed vegetation height was estimated at 40 stratified random sampling locations (~ 30 m apart) using a disc meter (diameter = ~ 0.6 m), while double sa mpling included measurement of both disc height and biomass under the disc at 12 locations across the experimental units. At each double with a cordless grass shea r to soil level (Homelite, Anderson, SC), raked and collected into cloth bags. The collected biomass samples were oven dried to constant weight (at 65C for 5 days). The dry biomass was regressed on the disc height for all double samples from a given treat ment to develop a calibration equation (R 2 = 0.8) for prediction of total above ground biomass in each experimental unit based on the average of the 40 disc heights taken on that experimental unit. Biomass C was calculated by multiplying the biomass by a s tandard factor of 0.5 (Poorter et al., 1997; Pregitzer and Euskirchen, 2004; Arevalo et al., 2009)
27 Above ground biomass C in silvopasture field: Due to the mo rphological differences of the tree and grass components of the silvopasture fields, the biomass C content of these 2 components were quantified separately. To derive the biomass C content of the bahiagrass component, the same procedure adopted for the sow n pasture (as described above) was applied. To estimate the biomass of the tree component, the diameter at breast height (DBH) of all trees in each field was measured with a girth tape. Published allometric equations that relate biomass to this parameter (Santantonio et al., 1977; Gonzalez Benecke et al., 2010) were used to estimate the total above ground biomass of the tree (Equation 1), and the estimated biomass was corrected f or logarithmic bias (Baskerville, 1972) The summation of the standing tree and the bahiagrass biomass C constituted the above ground biomass C of each field. Biomass C was also calculated by multiplying the bio mass by a standard factor of 0.5. ln(B AGT ) = 2.5563158 + 2.5209397 ln( DBH ) (2 1) CF = exp (MSE/2) (2 2) Where, B AGT = aboveground biomass of trees, DBH = diameter at breast height, CF = correction factor for logarithmic bias, MSE = estimated mean square error of the allometric equation (Baskerville, 1972) Above ground bioma ss C in native rangeland: A different double sampling strategy was adopted to quantify above ground biomass in native rangeland (Ebrahimi et al., 2008) Twelve quadrats (4 m 2 ), 30 m apart, were sampled along the diag onal transect. The percentage of woody and non woody vegetation coverage in each quadrat were determined by visual/ocular observation, using a standard reference (Barry, 1998) In order to minimize estimation bias, t wo independent observers recorded the percent cover and the final values were obtained by averaging the two observations for each quadrat. Percentage cover was adopted as the proxy (and
28 high frequency) sample, while vegetation biomass was quantified at eve ry 3 rd quadrat along the sampling transect. The woody and shrub biomass were stored in separate perforated poly ethylene bags and dried (to constant weight) for 3 weeks at 65C. The period of drying was prolonged to ensure adequate drying of the woody biom ass. Regression equations were developed to relate the weight of the dry biomass to the estimated percent vegetation cover (i.e. total, woody, and non woody). The equation ( R 2 = 0.7) was applied to estimate the total biomass in each field based on the per cent vegetation (woody and shrubs) coverage per quadrat. Biomass C was also calculated by multiplying the biomass by a standard factor of 0.5. Statistical Analysis Descriptive and statistical analyses were conducted using SAS 9.2 software (SAS, 2001). The impact of grassland biome on above and below ground C parameters was evaluated using the one way ANOVA procedure. Grassland biomes were considered as the independent treatment effect while the bulk density, above ground biomass, root biomass, SOC, and so il N stocks were the dependent response variable. Although multiple subsamples were collected within each experimental unit, values were averaged per field replicate within each grassland biome to avoid pseudo replication bias. Treatments were considered d ifferent when F test P values were < 0.05. Mean separation was performed using Tukey studentized range test. Results SOC Stocks Intensification of grassland by converting native rangeland to silvopasture and sown pasture had significant positive effects on SOC stock at all soil depth intervals (Table 2 1). In comparison to native rangeland, SOC under silvopasture was 38% higher but not different than in sown pasture at the 0 10 cm soil depth interval. At 10 20 cm and 20 30 cm depth interval, SOC increased (88% and 68%, respectively) following conversion to sown pasture, while it
29 increased to a greater degree following conversion to silvopasture (117% and 108%, respectively). The cumulative (0 30 cm) SOC stock also increased after conversion of native rangeland to sown and silvop asture (52 and 69% increase, respectively). However, the cumulative SOC was not different between sown pasture and silvopasture (62 and 69.2 Mg C ha 1 respectively). Differences in soil bulk density a mong the gras sland biomes were observed at 0 10 cm depth but not at deeper (10 30 cm) soil depths (Figure 2 2). Soil bulk density generally increased with increase in soil profile depth and minimum and maximum values were observed in sown pasture, ranging from ~ 1.0 g cm 3 at 0 10 cm, to 1.5 g cm 3 at 20 30 cm. Soil N Stocks Comparison of the three management systems showed differences in soil N stock in the upper 0 10 cm and 10 20 cm soil depths (Table 2 1). At both depth intervals, higher N stock was obse rved in the sown pasture (2.01 and 1.02 Mg N ha 1 respectively) and silvopasture (1.93 and 1.13 Mg N ha 1 respectively), compared with native rangeland (1.36 and 0.45 Mg N ha 1 respectively). Similar results were obtained for the cumulative N across the sampled soil profile (0 30 cm) where higher N stock was observed in both sown pasture and silvopasture (3.83 Mg N ha 1 and 3.94 Mg N ha 1 respectively), compared with the native rangeland (2.40 Mg N ha 1 ). Soil C:N ratios ranged between 14 (in native ra ngeland at 20 30 cm soil depth) to 21 (in native rangeland at 10 2 0 cm soil depth). Soil C:N ratio was reduced in sown pasture at the upper 0 10 cm while it was reduced in silvopasture at 10 20 cm soil depths, relative to native rangeland. However, there w as no significant difference in the mean C:N ratio of all ecosystems across the sampled soil depth (0 30 cm; Table 2 1).
30 Particulate and Mineral Associated Organic C and N The quantity of POC only increased in sown pasture but not in silvopasture, compare d to the reference native rangeland (Table 2 1). POC fraction in the sow n pasture increased at 0 10 and 10 20 cm depths (7.3 and 4.5 Mg C ha 1 increase, respectively), but there was no difference at 20 30 cm. Across all depths (0 30 cm), POC in native rang eland (21 Mg C ha 1 ) d id not differ significantly from silvopasture (27 Mg C ha 1 ) but it was higher in sown pasture (34 Mg C ha 1 ). Sown pasture and silvopasture were not different at all depth intervals (Table 2 1). Similar to POC, the PON content and the relative percentage to total N decreased with depth across all ecosystems (Table 2 1). Sown pasture contained the highest PON at 0 10 cm and 10 20 cm depths (1.47 and 0.45 Mg N ha 1 respectively), and across the sampled soil depth (2.09 Mg N ha 1 ; Tab le 2 1). The quantity and relative percentage of C and N associated with the mineral fraction increased in the silvopasture ecosystem at upper (0 10 and 10 20cm) soil depth intervals, but did not change in sown pastures, compared to the reference native r angeland (Table 2 1). In contrast to the sown pasture which contained the highest POC at all depths (except 20 30 cm), silvopasture contained highest quantity and percentage of mineral associated C and N fraction at all soil depth intervals (up to 15.1 Mg C ha 1 and 86% of SOC at 20 30 cm depth). Root Biomass C Root biomass C decreased at deeper soil depth (10 30 cm) as management intensity increased. At the 10 20 cm depth, greatest root biomass C was associated with the native rangeland (7.1 Mg C ha 1 ) a s compared with the sown pasture (3.1 Mg C ha 1 ) and silvopasture (2.8 Mg C ha 1 ). A similar trend was observed at the 20 30 cm depth. Across the sampled soil profile (0 30 cm), relative to root C in native rangeland (23.9 Mg C ha 1 ), root biomass C in sow n pasture ( 18.8 Mg C ha 1 ) was not different, while it was lower (15.4 Mg C ha 1 ) in
31 silvopasture (Figure 2 3 and 2 4). However, root C in sown pasture and silvopasture was similar. Generally, the highest proportion of the total root biomass C in all ec osystems accrued in the 0 10 cm depth and declined at lower depths. Aboveground Biomass C The total aboveground biomass C (woody + non woody) was markedly greater in silvopasture (59 Mg C ha 1 ), compared to native rangeland and sown pasture (4.2 Mg C ha 1 and 2.1 Mg C ha 1 respectively; Figure 2 4). Woody vegetation components of the native rangeland (saw palmetto) and silvopasture (slash pine) accounted for 70 and 98% of their total aboveground vegetation biomass, respectively. Ecosystem C Total ecosystem C was 69 Mg C ha 1 in native rangeland, 83 Mg C ha 1 in sown pasture, and 144 Mg C ha 1 in silvopasture (Figure 2 4). SOC constitutes a larger percentage (~75%) of total ecosystem C in sown pasture, but constitutes a lesser proportion in native r angeland and silvopasture (~59 and ~48%, respectively; Figure 2 4). Belowground C (root C and SOC) accounted for ~95% of ecosystem C in both native rangeland and sown pasture, and ~59% of ecosystem C in silvopasture. The total below ground C stocks were co mparable in silvopasture (85 Mg C ha 1 in) and sown pasture (81 Mg C ha 1 ), but relatively higher than the reference native rangeland (65 Mg C ha 1 ). Discussion The impact of management intensification on soil and ecosystem C in grassland biomes that have been established and consistently managed for over 22 years was evaluated. Despite the limitations associated with the experimental approach (i.e., lack of randomization and limited number of replicates), the long term nature of this study offers an uniqu e opportunity to assess the impact of only management system changes on ecosystem C stocks without the influence of
32 varying site or environmental factors such as elevation, climate, or soil properties at field scale. In addition, this approach was based on change in management systems rather than the conventional approach of focusing on individual management practices. Although assessing impacts of different management practices (such as stocking rate, fertilizer application, and fire) separately on ecologi cal properties are often beneficial within the context of specific management system, comparison of distinctly different management systems often preclude the possibility of determining the effect of individual management practices, as in this study. This is because each management system was characterized by different combination of management practices which likely interact to uniquely influence SOC and N dynamics. For instance, in intensively managed pastures, grazing and fire interact to facilitate nutr ient cycling in soils, even though they have distinctly different pathway and magnitude of influence when their effect on grasslands are isolated (Hobbs et al., 1991; Johnson and Matche tt, 2001) Impact of Grassland Intensification on SOC Stocks and Soil Bulk Density According to a review of studies focused on SOC changes due to application of different management practices or conversion from native vegetation (Conant et al., 2001) more intensive management practices have been documented to enhance SOC sequestration at a rate of 0.11 to 3.04 Mg C ha 1 yr 1 This study showed that conversion of native rangeland to sown pasture and silvopast ure enhanced total SOC sequestration to a depth of 30 cm at a rate of 0.66 and 0.88 Mg C ha 1 yr 1 respectively. The SOC sequestration rate observed in this study, which is at the lower end of reported sequestration rates, may be attributed to the potenti al influence of the biome and climate (Conant et al., 2001) and the long term duration of the study. It should be recognized that potential for high SOC sequestration is limited under the joint influence of sandy soil texture, high rainfall, and high temperature within the ecoregion of our study (Silveira et al., 2013) all of which favor fast processing of SOC and prevent large amount of C accumulation. Hence, the
33 overa ll reported SOC are at the lower end of reported range (within similar soil depth) for grassland ecosystems in sub humid southeastern Australia (Chan and McCoy, 2010) and in humid tropical Panama Canal Island (Schwendenmann and Pendall, 2006) The observed similarity in SOC between sown pasture and silvopasture across the sampled soil depth (0 30 cm) is consistent with the findings of other studies that indicated comp arable short and mid term C storage of silvopasture and sown pasture systems after conversion from or to an integrated tree grass ecosystem (Sharrow and Ismail, 2004; Peichl et al., 2012) b ut contrasts the findings of Haile et al. (2008) and Martens et al. (2004) who reported that silvopasture and forested pastureland accreted higher SOC compared wit h a baseline sown pasture. Disparity with the findings of these studies may be related to the difference in the management strategies applied, soil type, and duration of the study. Research has shown that land use conversion and intensification of grasslan d management affects SOC chemical composition, physical structure, stability and function (Bruce et al., 1999; Conant et al., 2001; Huang et al., 2011) The similarit y of soil bulk density across the study site (except at 0 10 cm soil depth) suggested that observed changes in SOC stocks reflected changes in C stocks rather than soil compaction. The dominant soil series of the study site, mainly composed of fine sand, m ay have limited the possibility of soil compaction (and its confounding influence on interpreting SOC changes) which is often associated with increased stocking rate in high clay content soils. Generally, our observed bulk density values are similar to pre viously reported values under different grazing management practices on sandy soils (Tate et al., 2004) and under slash pine plantation in Florida (Gholz and Fisher, 1982) C hanges in SOC Fractions and their N Content with Management Intensification The particle size SOC fractions unveil a contrasting mechanism of SOC sequestration in silvopasture and sown pasture after conversion from native rangeland. Relative to the native
34 rangeland, the accretion of higher proportion of POM (i.e. POC and PON) in sown pasture at 0 20 cm and overall (0 30 cm), and converse accretion of higher proportion of mineral associated SOM (i.e. C and N) in the silvopastoral ecosystem at 10 20 cm and ov erall (0 30 cm) suggests that the conversion of native rangeland to sown pasture favors the accumulation of the labile SOC pool, while the silvopastoral ecosystem favors greater accretion of the stable pools (mineral associated SOC). These findings are i n accordance with the conclusions of researchers (Feller and Beare, 1997; Desjardins et al., 2004; Schwendenmann and Pendall, 2006 ) who have reported that SOC wa s accreted to a greater degree in the coarse (>53 m) POC fraction under converted or sown grasslands. Fertilizer application and greater fecal deposition associated with higher stocking rate in the sown pasture can potentially drive higher net primary pro duction of the constituent bahiagrass and, consequently, favor more allocation of C to the POC fraction (Batjes and Sombroek, 1997) The similarity in chemical and structural composition (lignified woody C 3 species) of the vegetation in both native rangeland and silvopasture may not suffice to elucidate similarity in the allocation of C into the POC fraction. However, the formation of more stable mineral associated C in silvopasture is likely attributable to the comp lexing of organic matter with tree derived phenolic compounds (Jastrow et al., 2007) and potential in situ modification of microclimate under the tree stands which could favor abiotic humification. Research into potential influence of in situ microclimate and the influence of tree derived phenolic compounds may be helpful to unravel the process of stable C formation under silvopasture. As expected, the similarity between the trend of changes in soil N stocks and SOC reflected the coupled response after land use conversion. Intensification of grassland increased soil N stocks due to higher quantity of applied fertilizer and increased potential for animal
35 derived N deposition. The observed soil N stock (0 30 cm dept h) across the treatments (ranging from 2.4 Mg N ha 1 to 3.94 Mg N ha 1 ) was in contrast with other studies in temperate and tropical regions but comparable to studies conducted under similar subtropical condition. For example, Peichl et al. (2012) reported ~7.5 Mg N ha 1 under 5 year old forest and adjacent grassland in Ireland, and Schwendenmann and Pendall (2006 ) reported 9 and 8 Mg N ha 1 in 90 year old a tr opical forest (0 50 cm) and a converted grassland, respectively, while Silveira et al. (2013) reported a range of 1.4 1.7 Mg N ha 1 under different pasture grazing and fertilization regimes in central Florida Typically, greater C:N ratio in particulate fractions, and lower ratio in the heavier, mineral associated fraction would indicate stronger resistance to decomposition (Sollins et al., 1984; Mar tens, 2000) hence, the resistance or ease of decomposition of the SOC fractions was altered after conversion but the impact was generally limited to top 0 20 cm. Management Intensification Impact on Live Biomass and Ecosystem C Despite the similarity in the root biomass C and grass derived biomass C after conversion from native rangeland to sown pasture, the contribution of the woody biomass (2.9 Mg C ha 1 ) offers an advantage for aboveground ecosystem C sequestration in native rangeland. Consistent wi th the similarity of SOC in silvopasture and sown pasture, similarity in biomass C levels further suggests that the bahiagrass component may have a strong impact on SOC though it constitutes a marginal percentage (3%) of aboveground biomass in the silvopas ture treatment (De Groot, 1990; Sharrow and Ismail, 2004) The observed similarity in grass associated biomass C may be attributed to the relatively lower grazing pressure in silvopasture which allowed the tree based system to accrue similar quantity of grass biomass despite the higher potential for livestock mediated nutrient cycling and greater forage production in the sown pasture ecosystem.
36 Compared with trees or woody vegetation, grass speci es generally develop shallow root systems or alternatively allocate the main root biomass to the uppermost soil layers, even though single roots can reach depths of several meters (Haile et al., 2010) The native rangeland ecosystem is dominated by rhizomatous saw palmetto species which have roots concentrated around the rachis, the plant part that is developed ~30 cm beneath and above soil surface (Fish er and Jayachandran, 1999; Duever, 2011) Although, the slash pine trees in the silvopasture ecosystem are also characterized by shallow rooting with about 85% of the total root C accruing to the 0 30 cm soil depth (T ang and Tang, 1989) limitations of the sampling technique (core diameter = 5 cm) may favor sampling of fine roots over large diameter tree roots and possibly influence the representation of tree derived root biomass C under silvopasture, especially at the lower 10 30 cm soil depth. However, the limited potential for tree associated fine root contribution at the upper (0 30 cm) soil depth (Jackson et al., 1997) may allow for dominance of the shallower bahiagrass r oots on the belowground processes (within the sampled depth) thereby leading to the observed similarities with sown pasture. This may potentially translate into similarity in root turnover within silvopasture and sown pasture and overall accretion of compa rable total SOC in both ecosystems in the long term. Further study on root turnover will be beneficial to elucidate the role of roots in regulating the long term sequestration within these ecosystems. Generally, ecosystem C stocks include the aboveground w oody and non woody biomass C, root biomass C, and SOC. Increase in ecosystem C after converting native rangeland to silvopasture (108%) and sown pasture (20%) indicates that the C sink capacity of grasslands can be greatly enhanced in the long term through intensified management systems (Conant et al., 2001; Sharrow and Ismail, 2004) Similar to the findings of Arevalo et al. (2009) in a tra jectory
37 of land use changes, including grasslands, it is interesting to note that the overall difference in belowground C stock (root C + SOC) between sown pasture and silvopasture was apparently marginal, but the dynamics are different in terms of the con tribution from root C and SOC. The disparate biomass allocation into these ecosystem pools may have implications for modeling biogeochemical processes (including turnover of C, N, and other nutrients) and accurate assessment of ecosystem C balance in respo nse to changes in management system (Sharrow and Ismail, 2004; Peichl et al., 2012) Also, the understanding of management induced shifts in biomass allocation is important for strategic adopt ion of management approach that enhances C sequestration opportunities and reduces potential losses in grassland ecosystems. Summary The findings from this study suggested that after > 22 years, conversion from native rangeland to tree integrated silvopas ture or more intensively managed sown pasture increased ecosystem C stocks due to increase in SOC, while silvopasture offered the additional benefit of increased above ground woody biomass C pools within this subtropical ecoregion. Results provided eviden ce of long term allocation of higher quantity of stable SOC in silvopasture compared to baseline native rangeland suggesting a long term C sink. Data suggested that increasing intensity of grassland management is beneficial for soil and ecosystem C sequest ration in the long term, but, shifts in the allocation of C into ecosystem pools will likely have implication for biochemical cycling within this subtropical ecoregion. However, it should be reckoned that these findings may be most applicable to ecological sites with similar site conditions, due to the constraints associated with the non dispersal of replicates in space under the studied long term ecological units.
38 Table 2 1. S oil organic C (SOC) and N stocks at different depths under native rangeland, sown pasture, and silvopasture ecosystems Management System SOC (Mg C ha 1 ) Soil N (Mg N ha 1 ) Soil C:N Ratios Total Mineral fraction POC Total Mineral fraction PON Total Mineral fraction POC Depth1 (0 10cm) Native rangeland 23.8b 7.96b (33.5) 15.9b (66.5) 1.36b 0.57b (41.9) 0.79c (58.1) 19.8a 13.9b 20.1a Silvopastu re 32.9a 12.9a (39.2) 20.0ab (60.8) 1.93a 0.80a (41.5) 1.14b (58.6) 17.6ab 16.2a 17.7ab Sown pasture 31.5ab 8.31b (26.4) 23.2a (73.6) 2.01a 0.55b (27.4) 1.47a (72.6) 15.8b 14.8ab 15.9b P value 0.028 0.011 0.013 0.006 0.049 <0.001 0.024 0.006 0.02 Depth2 (10 20cm) Native rangeland 8.67b 5.31b (61.2) 3.37b (38.8) 0.45b 0.30b (66.7) 0.16c (33.3) 20.9a 17.8 21.5a Silvopasture 18.8a 13.9a (73.9) 4.93b (26.1) 1.13a 0.85a (75.2) 0.29b (24.8) 16.8b 16.4 16.9b Sown pasture 16.3a 8.50ab (52.1) 7.83a (47.9) 1.02a 0.57ab (55.9) 0.45a (44.1) 17.2ab 15.4 17.4ab P value <0.002 0.003 <0.001 <0.001 0.002 <0.001 0.040 0.091 ns 0.035 Depth3 (20 30cm) Native rangeland 8.45b 6.75 (79.9) 1.70 (20.1) 0.58 0.46 (79.3) 0.13 (20.7) 13.8 15.9 13.9 Silvopasture 17.6a 15.1 (85.8) 2.51 (14.2) 0.87 0.75 (86.2) 0.12 (13.8) 20.4 26.5 20.7 Sown pasture 14.2ab 11.3 (79.6) 2.94 (20.4) 0.80 0.63 (78.8) 0.17 (21.2) 16.6 17.4 16.5 P value 0.047 0.062 ns 0.135 ns 0.179 ns 0.183 ns 0.060 ns 0.073 ns 0.321 ns 0.148 ns Cumulative depth (0 30cm) Native rangeland 40.9b 20.0b (48.9) 20.9b (51.1) 2.40b 1.32b (55) 1.08c (45) 17.3 15.9 18.5 Silvopasture 69.2a 41.8a (60.4) 27.4ab (39.6) 3.94a 2.39a (60.7) 1.55b (39.3) 18.2 19.7 18.4 Sown pasture 62.0a 28.1b (45.3) 33.9a (54.7) 3.83a 1.75ab (45.7) 2.09a (54.3) 16.1 15.9 16.6 P value 0.001 0.001 0.001 0.001 0.010 <0.001 0.384 ns 0.229 ns 0.432 ns
39 PON = particulate organic C and N, respectively. Means followed by the same letter within column (by depth) are not Values in parentheses represent percent age of C and N associated with the mineral and particulate fraction relative to the total. Figure 2 1. Aerial i magery of the study fields obtained from Google Earth and layout of the 5 sampling quadrats (20 m 20 m) on a diagonal transect.
40 Figure 2 2. Bulk density of soil samples collected at three profile depths (0 10 cm, 10 20 cm, 20 30 cm) in native rangeland, sown pasture, and silvopasture ecosystems. Ecosystems followed by the same letter within depth are not significantly different according
41 Figure 2 3. Root biomass carbon (C) at three soil profile depths (0 10 cm, 10 20 cm, 20 30 cm) depth in native rangeland, sown pasture, and silvopasture ecosystems. Bars followed by the same letter within de pth are not error of mean (n = 2).
42 Figure 2 4. Ecosystem carbon (C) stocks under native rangeland, sown pasture, and silvopasture ecosystems. B ars followed by the same letter within ecosystem C component are not significantly different according to significantly different.
43 CHAPTER 3 IMPACT OF MANAGEMENT INTENSIFICATION ON PARTICLE SIZE SOIL CARBON FRACTIONS IN SUBTROPICAL GRASSLANDS: EVIDENCE FROM 13 C NATURAL ABUNDANCE Background Grasslands occupy one third of the total managed land area in the United States and 30 d area (Lal et al., 2003; U.S. EPA, 2012) They provide feed, fiber, and diverse ecosystem services, accounting for ~30% (~330 Pg) of soil organic carbon (C) storage globally and serving as an important C sink for mitigation of rising atmospheric CO 2 concentration (Eswaran et al., 1993; Batjes, 1999; Guo and Gifford, 2002; Lal et al., 2003; Derner and Jin, 2012) T he major mechanisms for improving C sequestration in grasslands are increasing C inputs to the soil through fertilization, irrigation, sowing of legumes or more productive grass species, and reducing soil disturbance throug h the use of improved grazing management (Conant et al., 2001; Guo and Gifford, 2002; Franzluebbers, 2010) These management practices can promote changes in net pri mary productivity, plant tissue chemistry and morphology, and litter decomposit ion with subsequent effects on soil C turnover and stability (Bruce et al., 1999; McL auchlan et al., 2006; Jastrow et al., 2007) Grassland management intensification such as conversion of native ecosystems in to sown pastures and increased use of nitrogen ( N ) inputs have been driven by increasing human population and demand for greater food production per unit land area (FAO, 1993; White et al., 2000; Mulkey, 2007) Improper management following la nd use conversion may compromise the long term sustainability of grasslands through degradation of soil quality as measured by critical indicators such as soil C (Conant et al., 2001; Silv eira et al., 2013) The extent of soil C changes in response to improved grassland management or land use conversion varies depending upon the region or management practice Research has shown that the changes in soil C can
44 range from 10 g C m 2 yr 1 in rainforest biomes to 100 g C m 2 yr 1 in native grassland and woodland biomes (Conant et al., 2001) Assessing the changes in the natural abundance of 13 13 C), or comparing the 13 C between initial (or legacy) soil C and newer inputs of C, can be used to determine the responses of soil C to changes in vegetation cover (including management intensification) (Farquhar et al., 1982; Boutton et al., 1998; Ehleringer et al., 2000; Biedenbender et al., 2004) The isotopic signature of soil reflects the relative contribution of vegetation from C 3 versus C 4 photosynthetic pathways to soil C and is often used to investigate the impacts of land use conversion on structural and functional characteristics of soil C (Boutton et al., 1998; Diels et al., 2001 ; Biedenbender et al., 2004) Furthermore, the persistence or rate of depletion of soil C provides a direct measure of the turnover rate under the new vegetation (Bout ton et al., 1998; Diels et al., 2001; Wang et al., 2011a) However, the application of isotopic analysis to understand the rate of depletion of relic soil C is limited to land use changes that involve clear transition from C 3 dominated to C 4 dominated v egetation, or vice versa (Roberts, 2001; Haile et al., 2010) While evaluat ion of C responses to land use changes in bulk soil s can provide useful information on overall rate of soil C sequest ration or depletion density or particle size separation combined with determination of the isotopic 13 C natural abundance can be a better indicator of management effects on the stability and source of soil C (Cambardella and Elliott, 1992; Six et al., 2002; Conant et al., 2004; Dubeux et al., 2006a) Given that the introduction of improved grass species and nutrient addition to i ntensively managed grasslands favors higher ecosystem productivity compared to native ecosystems it is expected that management intensification will increase the rate of litter return and accumulation of labile soil C forms such as particulate organic matter (POM) (Billings et al., 2006; Silveira et
45 al., 2013) Conversely, litter derived from sown grass species is often more readily decomposable than native vegetation (Dubeux et al., 2006b; Dubeux et al., 2007) and may consequently promote turnover of litter C into the more stable soil C fraction in the long term. In essence, intensively managed grasslands may not only favor accumulation of recent labile soil C fraction, but may also contribute significantly to the stable mineral associated fraction to offset losses in the relic soil C. Understanding these potential changes in sources and stability of soil C is crucial to infer the long term sustainability of intensively managed grasslands. The obj ectives of this study were to i ) assess the long term (> 20 years) impact of grassland management intensification on soil C fractions after conversion of native rangelands to s ilvopasture and sown pasture ecosystems, and ii) determine the contribution of sown grass species to soil C sequestration in both the labile and more stable soil C fractions. Materials and Methods Study Site This study was conducted at the University of Fl orida Range Cattle Research and subtropical climate with average annual precipitation of ~1650 mm, average minimum daily temperature of ~16.7C, and average maximum dai ly temperature of ~28.2C. The site has a relatively homogeneous slope (<5%) and the grassland ecosystems have been established and consistently maintained for over 22 years. The three ecosystems are native rangeland, silvopasture, and sown pasture, repres enting increasing order of management intensity as commonly practiced within the region. Each grassland ecosystem was replicated twice to form six collocated fields (i.e. 3 grassland ecosystems 2 replicate fields). The dominant soil series across the en tire site were Ona and Immokalee fine sand s (sandy siliceous, hyperthermic Typic
46 Alaquods), which developed on sandy marine deposits parent material (Soil Survey Staff, 1999; NRCS Websoil Survey, 2013) The entire experimental area was historically used as a native rangeland winter pasture and consisted primarily of saw palmetto [ Serenoa repens (Bartram) Small] and a wide variety of grasses such as creeping bluestem ( Schizachyrium stolo niferum Nash) and maidencane ( Panicum hemitomon Schult.) (Kalmbacher et al., 1984) The sown pasture and silvopasture fields were established by clearing native rangeland and successively sowing bahiagrass ( Paspa lum notatum Flgge) in both treatments in 1985 and slash pine ( Pinus elliottii Engelm. var. elliottii ) in the silvopasture in 1991. Hence, the native rangelands fields represent the baseline native vegetation while conversion to silvopasture and sown pastu re represent different levels of management intensification. The native rangeland was never fertilized, but it had been subjected to periodic burning (every 3 years), occasional livestock grazing activities (< 60 days per year), and herbivory by wildlife, all typical features of rangeland in this region. The silvopasture ecosystem received periodic applications of 67 kg N ha 1 yr 1 as ammonium nitrate. No fertilizer was applied in the years of 1993 1997, 2000, 2002, 2008, 2009, and 2011. Grazing of the silv opasture began in March 1993, 18 months after planting the trees, and has continued from March September every year on a 4 week rotational basis. The sown pasture was subjected to higher frequency of fertilization and grazing. Ammonium nitrate fertilizer w as applied at an annual rate of 67 kg N ha 1 yr 1 and dolomite was applied in 2001 and 2008 at a rate of 730 and 550 kg ha 1 respectively, while grazing occurred on a rotational basis for 7 months, with 1 week residence period followed by 1 week rest period between each grazing event. Detailed information about the experimental sites and the specific management practices within the each ecosystem are presented in Chapter 2.
47 Experimental Design The study was based on a comparative mensurative experimental design (Hurlbert, 1984; Arevalo et al., 2009) bec ause the two replicate fields (6 ha) of each management ecosystems are collocated; in other words, the treatment replicates are not spatially dispersed. This experimental design has an underlying assumption that the soil properties of the three ecosystems were similar prior to the conversion and designation of each ecological unit (Hurlbert, 1984; Arevalo et al., 2009) Factors such as the uniformity of initial land use, the flat te rrain, and the limited potential for variation in the textural class of the soil (Kalmbacher et al., 1984; Kalmbacher et al., 1993) lends credence to the assumption that the soils were similar before the management changes were imposed. However, similar to general limitation of most long term ecological studies (Janzen, 1995; Millar and Anderson, 2004) there were no data available on the soil properties before the land use conversions, and some spatial variability is bound to be inherent within the study area. Soil Sampling and Analyses Prior to sampling, an initial offset (30 m) from the edge of each field was establish ed in order to avoid potential edge effects, and five quadrats (20 m 20 m), ~75 m apart, were marked out along a diagonal transect. Five random soil cores (diameter of 2.2 cm) were sampled from each quadrat at the 0 to 10, 10 to 20, and 20 to 30 cm depth s and composited within each depth for soil C determination. One random soil core was sampled in each quadrat at the same depth intervals for bulk density determination. Soil samples collected for bulk density determination were dried at 105C until consta nt weight, and weighed. Bulk density was computed by dividing the dry weight by the soil core volume. Soil samples for C analysis were air dried and passed through a 2 mm sized sieve. Particle size separation was performed using the modified procedure desc ribed by Cambardella and Elliot (1992) Briefly, 10 g of air dried soil was shaken in 30 mL
48 of 5 g L 1 sodium hexametaphosphate solution on a reciprocal shaker at 200 rpm for 15 h. The slurry was gently wash ed with distilled de ionized water and filtered through a 53 m sieve. The fractions retained in the sieve (> 53m) and the fraction that passed through (< 53 m) were transferred to drying dishes, dried at 55C for 72 h, and weighed. The fraction retained in the 53 m sieve corresponds to POC, while the material < 53 m (silt and clay sized particles) correspond to the Cmin. Carbon concentration in each fraction and the bulk soil was determined by dry combustion using a ThermoFlash EA 1112 elemental analyz er. 13 C) of the POC and Cmin fractions were determined on a Thermo Finnigan MAT Delta Plus XL Isotope Ratio Mass Spectrometer (IRMS) interfaced via a Conflo III device to a Costech ECS 4010 elemental analyzer (Costech Valencia, 13 C PDB standard which expresses whether a sample has a higher or lower 13 C / 12 C isotopic ratio compared to the calcium carbonate standard known as Vienna Pee Dee Belemnite (V PDB) (Coplen, 1994, Equation 1; Boutton et al., 1998) (3 1) Carbon derived from initial native ran geland vegetation (C 3 derived C) and from newer sown bahiagrass (C 4 derived C) were calculated based on mass balance equa tions (Equations 3 2 and 3 3), described by Follett et al. (2009) By biomass composition, the native rangeland was predominantly saw palmetto (70%), but also contains diverse mixture of native rangeland forbs and grasses (such as Andropogon spp. and bluestem grass), some of which are C 4 characterized. (3 2) (3 3)
49 13 C C4 13 C C3 13 C values of the C 3 and C 4 vegetation/plant material, respectively; 13 C value of the sample being assessed for C 3 and C 4 content. 13 C values for vegetation were de rived from published literature on the study site: saw palmetto = (Silveira et al., 2013 ) slash pine = (Haile et al., 2010) and bahiagrass = (Haile et al., 2010; Silveira et al., 2013) Statistical Analysis One way ANOVA was performed using SAS 9.2 software (SAS, 2001). Grassland ecosystem was considered a fixed effect because other types of ecosystems are not reckoned in this research study Soil C responses (total soil C stocks, POC and C 13 C sig nature) were response variabl es. Treatments were considered different when F test P values were < 0.05. Mean separation was performed using the Tukey studentized range test. Results Total O rganic C, POC and Cmin C oncentration Over 22 years following land u se conversion from native rangeland to silvopasture or sown pasture total soil C concentration and stock ( 0 30 cm depth) were greater under silvopasture and sown pasture than native rangeland (Table 3 1). Major increas es in soil C occurred in the 0 to 10 cm soil depth of sown pasture (up to 43%) and in the 10 to 20 cm depth of silvopasture (up to 100%) but no changes were observed at the 20 to 30 cm depth ( P = 0.08). Within the 0 to 10 cm depth, the sown pasture contai ned greater soil C (32.1 g C kg 1 ) compared with the baseline native rangeland (22.4 g C kg 1 ), but no differences were observed between silvopasture and native rangeland at this depth Conversely, at the 10 to 20 cm depth, soil C in silvopasture (14.0 g C kg 1 ) was greater in comparison with native rangeland (7.0 g C kg 1 ), while soil C in sown pasture (12.7 g C kg 1 ) was not different from either ecosystem (Table 3 1). Soil C stocks declined with depth for all treatments (Table 3 1).
50 Similar to bulk soi l C, management intensification influenced the allocation of soil C into the particle size fractions at all sampled soil depth intervals except the 20 to 30 cm depth (Table 3 1). At the 0 to 10 cm depth conversion from native rangeland to sown pasture fav ored greater C accumulation in the POC fraction, but no effect was observed in the C min fraction ( P = 0.07). At the 10 to 20 cm soil depth, greater POC in the sown pasture was observed ( P < 0.01), while the Cmin fraction increased in silvopasture (10.4 g C kg 1 ) compared with native rangeland (4.4 g C kg 1 ). Moreover, overall POC concentration (0 30 cm depth) was comparable in native rangeland and silvopasture, but the Cmin increased from 5.7 to 10.1 g C kg 1 respectively. In contrast, conversion to sown pasture increased the POC concentration (10.6 g C kg 1 ) compared with native rangeland ( 6.3 g C kg 1 ) but no effect was observed in the C min fraction (Table 3 1). 13 C) of Bulk Soil, P O C, and C min Fractions Results show that conversion of native rangeland to silvopasture and sown pasture altered 13 C signatur e of the bulk soil and particle size fractions at all sampled depth intervals (Figure 3 1 and Figure 3 13 C values were observed in both the silvopasture and sown pasture, which indicates the contribution of newly added C 4 derived C inputs to the 13 C values of POC across the 0 to 30 cm soil depth became less depleted at all depths after conversion, following the order, native rangeland ( 13 C = 13 C = 13 C = pine tree 13 C value ( 13 C = pronounced influence of the introduced C 4 13 C value (
51 C 3 and C 4 derived C Composition of Bulk Soil, POC and Cmin C Fractions Generally, the percent C 3 derived and C 4 derived C in bulk soil, POC and Cmin fractions were different in the silvopasture and sown pasture compared with the reference native rangeland (Table 3 2). Across the entire soil profile (0 30 cm depth), native rangeland contained a higher proportion of C 3 derived C in the bulk soil (~70%), POC (~74%) and Cmin fractions (~66%), while the sown pasture contained a higher proportion of C 4 derived C (~70%, ~70%, and 62%, respectively). This response is consistent with the predominance of saw palmetto (C 3 species) vegetation within the native rangeland, and bahiagrass (C 4 species) within the sown pasture. The relative contribution of C 3 derived C decreased as grassland management intensity increased. The C 3 derived C accounted for ~66 to 74% of total soil C in native rangeland, which contrasts with ~53 to 59% in silvopasture, and ~30 to 38% in the sown pasture. As expected under the influence of constituent C 4 vegetation (bahiagrass), the percent C 4 derived C in the POC and Cmin fractions increased in the s own pasture and silvopasture compared with the baseline native rangeland (Table 3 2). The highest increase in percent C 4 derived C was observed in sown pasture at the 0 10 cm soil depth (C 4 derive d POC and Cmin fraction were 84 and 76%, respectively). Disc ussion and Conclusion Management Intensification Effects on Soil C Data indicated that conversion of native rangeland to silvopasture and sown pasture enhanced the total soil C stocks in the 0 to 30 cm depth after > 22 years of change in management system (in Chapter 2). The observed increase in soil C stocks in both silvopasture and sown pasture can be related to the characteristically greater plant derived (above and below ground) C inputs and influence of grazing animals in these ecosystems, compared w ith the native rangelands. Given that up to 50% of grazed biomass (including live and dead plant materials) is
52 deposited on the soil as dung (McSherry and Ritchie, 2013) grazers have the potential to incorporate C from above ground into the soil, thus enhancing C sequestration potential with increasing management intensity. This is in agreement with reported increases in soil C after conversion of native vegetation to ma naged grasslands in various ecological regions (Conant et al., 2001; Conant et al., 2004; Chan and McCoy, 2010) For example, results indicate d a 43% increase in bulk s oil C within the 0 10 cm depth after conversion of native rangelands to sown pasture, and this is consistent with the mean increase of 37% reported by Conant et al. (2001) across 17 studies focused on conversi on of native vegetation to managed grasslands. Considering the allocation of soil C at different depth intervals, the differences between silvopasture and sown pasture suggest that the management system imposed after conversion from native rangeland affect s the long term dynamics of soil C. For instance, conversion of native rangeland to sown pasture resulted in a marked increase of soil C concentrations in the 0 to 10 cm soil depth, while silvopasture favored greater co ncentration of soil C in the 10 to 20 cm depth. P otential underlying factors that may be responsible for disparat e allocation of C may include i ) the level of management related disturbance which can favor more rapid mixing of plant litter (Pietola et al., 2005; Fernndez et al., 2010) into the top (0 10 cm) soil layer in sown pasture, ii ) the influence of phenolic root exudates in silvopasture which can enhance chemical protection of translocated soil C against micr obial decomposition in the lower (10 20 cm) soil depth (Jastrow et al., 2007; Baldock and Skjemstad, 2000) and iii) the deeper rooting characteristics of slash pine and bahiagrass (Tang and Tang, 1989; Jackson et al., 1996) which may influence the contribution of root to soil C. This contrasting depth based allocation of soil C after land use conversion has important imp lications for long term soil C sequestration because of its influence on ease of microbial access, potential loss through decomposition, and overall
53 residence time of sequestered C (Baldock and Skjemstad, 2000; Jones and Don nelly, 2004 ; Jastrow et al., 2007) 13 C) Signature in Soil C Fractions 13 C signature of the soil C fractions (Figure 3 2), bahiag rass exerted marked influence on the isotopic signature of the soil in both silvopasture and sown pasture ecosystems. Given that increased accretion of soil C was observed in both POC and C Min fractions after conversion from native rangelands (Table 3 1), the results of the isotopic analysis indicate that the C 4 grass, which is less depleted in 13 C, contributed significantly to C accretion within the particle size fractions over the past > 22 years Despite the potential contribution of tree derived biomass to soil C in silvopasture (Archer et al., 2001; Yelenik et al., 2004; Haile et al., 2008) the resu lts show that the grass component is contributing significantly to soil organic C formation 13 C values of the soil in silvopasture (ranging from 2) which are almost mi d way 13 C for bahiagrass ( 2013) and for slash pine ( 2010). I nformation on the bio chemistry of tree based silvopastoral ecosystems is still insufficient to elucidate the interactions and role of the grass component in co regulating soil C dynamics with the tree component. However, it is noteworthy that the grass component affects soi l C fractions despite the high tree to grass aboveground biomass ratios under silvopasture (Archer et al., 2001; Hibbard et al., 2001) It is likely that grass contributes significantly to soil C through greater, more readily incorporated, and more readily decomposed litter inputs compared with slash pine needles which likely undergo prolonged biochemical breakdown before incorporation into the soil (Kalbitz et al., 2003; Dubeux et al., 2007) Furthermore, potential differences in root stru ctural and chemical composition may influence root C turnover and relative contribution of the tree and grass vegetation components to soil C.
54 Changes in Stability and Source of C Allocated to Particle size Pools Particulate organic C (POC) fraction is considered as mostly unprotected, labile, and susceptible to microbial decomposition (Six et al., 2002; Sleutel et al., 2006) while the Cmin fraction is generally considered as biochemically prot ected against losses and stabilized in microaggregates which are formed through clay and silt based aggregation (Martens, 2000; Post and Kwon, 2000; Six et al., 2002; Martens et al., 2004; Schwendenmann and Pendall, 2006 ) Hence, the observed decline in relic Cmin fraction (Table 3 2) associated with conversion of native rangeland to sown pasture can be attributed to the management related perturbation of the ecosystem which can foster both the initial destabilization and ultimate reconstitution of microaggregates within the upper (0 30 cm) soil stratum (Beare et al., 1994; Jastrow et al., 1996; Blanco Canqui and Lal, 2004) Due to the limited applicability of 13 C natural abundance to infer changes in soil C when photosynthetic pathway (C 3 /C 4 ) is not clearly differentiated (Roberts, 2001; Haile et al., 2010) changes in relic C 3 or C 4 derived carbon under the silvopasture ecosystem was not quantifiable Generally, c hanges in Cmin pool over this relatively short dura tion raises an important consideration for conceptual and mechanistic modeling of soil C dynamics after land use change under this ecological condition. The significant reduction in percent C 3 Cmin and the concomitant increase in the percent C 4 Cmin (e.g. C 3 derived Cmin decreased from 13.2 Mg ha 1 to 10.6 Mg ha 1 while C 4 derived Cmin increased from 6.7 to 17.5 Mg ha 1 across 0 30 cm depth; Table 1 2) suggests that the silt plus clay associated C fraction may be more susceptible to change than currently presumed, and may increase readily with improved grassland management strategies in this environment. This is supported by the findings of Silveira et al. (2013) who used the 13 C natural abundance of soil C to show that < 53 m sized soil fraction decreased after 2 years of imposing more intensive grazing treatments in a subtropical grassland
55 in southeastern USA. Similarly, Solomon et al. (2002) concluded that the s oil C held in silt clay (< 20 m) particle size under an Ethiopian sub humid agroecosystem is more responsive to management than would be expected under temperate conditions. Hence, the changes observed in silt+clay sized (i.e. Cmin) fraction within ~35 ye ars of establishing the sown pasture may be attributed to favorable sub humid ecological condition for microbial accessibility and biochemical cycling. It is ina ccurate to interpret the similar ity of bulk soil C concentration in silvopasture and sown pastu re as an i ndication of comparable soil C accre tion potential under both ecosystems after conversion from native rangelands. This is because the 13 C natural abundance unravels important underlying management induced contrasts in the proportion of soil C derived from C 3 and C 4 sources. The observed change in composition of C 3 derived C (and increase in C 4 derived C) was more pronounced in so wn pasture than in silvopasture. This reflects the continued incorporation of C 3 derived C and potentially lesser con tribution of C 4 derived C in the silvopasture compared to sown pasture, after conversion from native rangeland It is probable that the C 3 tree component in the silvopastoral ecosystem is not only fostering the addition of more C 3 derived C from tree litte r fall, but may also have micro climate and soil chemical effects that may protect relic C (Smith and Johnson, 2004) This research indicates that soil C concentration increased with conversion of native rangeland to more productive grassland ecosystems. This has implications for soil quality and sustainable pasture management within this ecological biome. Additionally, contrary to the expectation that relic stable soil C fractions will be depleted and replaced by m ore labile forms with management intensification (Billings et al., 2006; Dubeux et al., 2006a; Filley et al., 2008) this research provides evidence that the bahi agrass vegetation in both silvopasture and sown
56 pasture has contributed significantly to increased soil C in both the labile and more stable fractions. These findings suggest that the role of sown C 4 grass in promoting long term soil C sequestration after land use conversion may be undervalued, while the changes in composition and sources of soil C may be strongly influenced by the target vegetation after conversion. In effect, manipulation of vegetation dynamics (including structure and composition) may be a strategic management approach for improving overall soil C sequestration potential in similar subtropical grassland ecosystems. However, it should be rec ognized that our findings may be most applicable to ecological sites with similar conditions (Janzen, 1995; Arevalo et al., 2009) due to the constraints associated with the non dispersal of replicates in space under the studied management systems
57 Table 3 1. Soil carbon (C) in bulk soil, particulate organic C, and mineral associated C fractions at different soil depth s as affected by grassland management intensification. Grassland e cosystem Bulk soil Particulate organic C Mineral associated C g kg 1 Mg ha 1 g kg 1 Mg ha 1 g kg 1 Mg ha 1 0 10 cm depth Native r angeland 22.4b 2.1 23.8b 2.0 15.0b 1.9 15.9b 1.6 (66.8) 7.4ns 1.0 7.9b 1.1 (33.2) Silvopasture 27.5ab 1.9 32.9a 2.5 17.0b 1.3 20ab 1.7 (60.8) 11ns 1.0 12.9a 1.3 (39.2) Sown p asture 32.1a 3.1 31.5ab 2.7 23.7a 2.0 23.2a 1.5 (73.7) 8.4ns 1.2 8.3b 1.2 (26.3) 10 20 cm depth Native r angeland 7.0b 1.5 8.67b 1.7 2.6b 0.4 3.4b 0.4 (39.2) 4.38b 1.3 5.3b 1.5 (61.1) Silvopasture 14.0a 1.4 18.8a 1.7 3.6b 0.5 4.9b 0.7 (26.1) 10.4a 1.4 13.9a 1.6 (73.9) Sown p asture 12.7ab 2.0 16.3a 2.2 6.0a 0.6 7.8a 0.7 (47.9) 6.7ab 1.5 8.5ab 1.7 (52.1) 20 30 cm depth Native r angeland 6.6ns 1.3 8.45b 1.6 1.3ns 0.2 1.7ns 0.3 (20.1) 5.3ns 1.2 6.8b 1.4 (80.9) Silvopasture 10.7ns 1.1 17.6a 3.5 1.9ns 0.3 2.5ns 0.3 (14.2) 8.8ns 1.0 15.1a 3.5 (85.8) Sown p asture 10.0ns 1.5 14.2ab 2.0 2.1ns 0.5 2.9ns 0.6 (20.4) 7.9ns 1.2 11.3ab 1.5 (79.6) Cumulative (0 30 cm) Native r angeland 12b 1.2 40.9b 4.0 6.3b 0.7 20.9b 1.8 (51.1) 5.7b 1.0 20b 3.5 (48.9) Silvopasture 17.6a 1.4 69.2a 4.5 7.5b 0.6 27.4ab 2.5 (39.6) 10.1a 1.0 41.8a 3.9 (60.4) Sown p asture 18.3a 1.9 62.0a 5.5 10.6a 0.9 33.9a 2.2 (54.7) 7.7ab 1.0 28.1b 3.4 (45.3)
58 Values represent meanstandard error (n=2); Values in parenthesis is the proportion (%) of bulk soil C; Means followed by dif ferent alphabets within soil depth are significantly differ ent according to Tukey studentiz ed range test ( P different.
59 Table 3 2 P ercent C 3 and C 4 derived soil carbon (C) in bulk soil, particulate organic C, and mine ral associated C fractions as affected by grassland management intensification. Grassland e cosystem Bulk soil C Particulate organic C Mineral associated C %C 3 %C 4 %C 3 %C 4 %C 3 %C 4 0 10 cm depth Native r angeland 64.7a 35.3c 70.2a 29.8c 65.8a 34.2c Silvopasture 49.6b 50.4b 57.5b 42.5b 50.4b 49.6b Sown p asture 9.4c 90.6a 16.8c 83.2a 24.0c 76.0a 10 20 cm depth Native r angeland 66.9a 33.1c 74.5a 25.5c 66.5a 33.5c Silvopasture 54.0b 46.0b 58.0b 42.0b 52.8b 47.2b Sown p asture 38.8c 61.2a 33.7c 66.3a 41.6c 58.4a 20 30 cm depth Native r angeland 71.9a 28.1c 76.9a 23.1c 66.5a 33.5c Silvopasture 59.0b 41.0b 61.8b 38.2b 56.0b 44.0b Sown p asture 42.5c 57.5a 38.2c 61.8a 47.4c 52.6a Cumulative (0 30 cm) Native r angeland 67.9a 32.1c 73.9a 26.1c 66.3a 33.7c Silvopasture 54.2b 45.8b 59.2b 40.8b 53.1b 46.9b Sown p asture 30.2c 69.8a 29.6c 70.4a 37.7c 62.4a Values represent mean (n=2); Means followed by different letter within soil depth are significantly different according to Tukey studenti z ed range test ( P
60 Figure 3 1 Isotopic 13 13 C) of carbon in bulk soil at different soil depths (0 10, 10 20, and 20 30 cm), as affected by grassland management intensification. Me ans are compared within each depth interval. Error bar indicates standard error of mean. Different letters denotes significant difference according to Tukey studentized range test ( P 0.05).
61 Figure 3 2 Isotopic 13 13 C) of mineral associated (<53m Cmin) and particulate sized (>53m POC) soil carbon fractions across the sampled soil depth (0 30 cm), as affected by grassland management intensification. Me ans are compared within each fraction. Different letters denotes significant difference according to Tukey studentized range test ( P 0.05).
62 CHAPTER 4 MANAGEMENT INTENSIFICATION EFFECTS ON AUTOTROPHIC AND HETEROTROPHIC SOIL RESPIRATION IN SUBTROPICAL GRASSLAND S Background Grassland ecosystems constitute about 40% of the global land area and they play a significant role in the global terrestrial carbon (C) cycle (K napp et al., 1998; White et al., 2000; Svejcar et al., 2008; Wang and Fang, 2009) Grasslands are recognized for their great potential as net sink for atmospheric CO 2 because they sequester C by allocating a significant portion of their biomass to rel atively more protected below ground components (Owensby, 1998; Follett et al., 2001; U.S. EPA, 2012) Management practices in grassland ecosystems can impact soil C dynamics by influencing the quantity and quality of litter incorporated into the soil, and subsequent soil C stabilization (Conant et al., 2001; Conant et al., 2004; Sharrow and Ismail, 2004; Dubeux et al., 2006a; Fynn et al., 2009) Globally, soils contain 1500 Pg C which is twice the amount of atmospheric C pool (Schlesinger and Andrews, 2000) with grasslands containing 12% of this terrestrial soil C pool (Schlesinger, 1977) Intensification of land use in the United States, mainly by converting native grasslands into more intensively managed ecosystems, has fostered the loss of ~500 Tg of soil organic C to the atmosphere (Davidson and Ackerman, 1993a; Kern, 1994; Conant et al., 2001) O n a global scale, CO 2 efflux from soils to the atmosphere (i.e. soil respiration) is recognized as a significant flux in the global cycle (77 Pg C yr 1 ), exceeding CO 2 release from fossil fuels by an order of magnitude (Houghton et al., 1990; Gates, 1993; Schlesinger and Andrews, 2000) M arginal changes in soil C fluxes can significantly alter atmospheric CO 2 concentration. Long term studies suggest that land use management is an important factor controlling the ability of grasslands to either serve as a sink or source of CO 2 to the atmosphere (Owensby, 1998; Bloom, 2010) Research across a variety o f ecological condition s indicate d that improved
63 grassland management can favor soil organic matter accretion and therefore increase C sequestration in soils (Bruce et al., 1999; Conant et al., 2001; Conant et al., 2004; Jones and Donnelly, 2004; Dubeux et al., 2006a; Dubeux et al., 2007) Changes in litter composition and tissue chemistry with g rassland management intensification can alter the rate of biochemical cycling and mineralization of C through increased microbial activity (Buyanovski and Wagner, 1995; Batjes and Sombroek, 199 7) accelerate decomposition and respiration rates, and consequently shorten the C residence time in the soil. Many studies have reported increased accumulation of less stable and easily decomposable soil organic C fractions in intensively managed grass lands (Dubeux et al., 2006a; Silveira et al., 2013) Increased decomposability may increase soil C turnover through respiration and anthropogenic CO 2 with concomitant effect on global warming. Understanding such potential changes in magnitude of soil C loss through respiration is critical for a comprehensive assessment of land use conversion and grassland management impacts on terrestrial C dynamic s across different eco regions. This is especially important in subtropical ecosystems, where there is limited knowledge about soil C dynamics and the potential effect of land use on ecosystem C balance (Franzluebbers, 2005; Franzluebbers, 2010; Silveira et al., 2013) Subtropical ecosystems are characterized by unique precipitation and temperature regimes which can have distinct influence on soil dry wet cycle and subsequent impact s on soil biological processes. Soil temperature (S Temp ) and soil moisture (S Moist ) have been reported as the dominant abiotic factors controlling soil CO 2 efflux in both temperate and tropical regions (Lloyd and Taylor, 1994; Adachi et al., 2006; Suh et al., 2009; Fu et al., 2010) Furthermore, grassland ma nagement strategies that change quantity and quality of aboveground biomas s can also influence S Temp with effects on decomposition and respiration (Flanagan and Johnson, 2005;
64 Sakurai et al., 2012) Understanding the effect of management intensity on the sens itivity of soil respiration to abiotic factors is crucial for assessing vulnerability of soil C to climate induced losses in managed grasslands, and it is a necessity towards sustainable management of subtropical grassland ecosystems. Furthermore, such kno wledge may foster improved process based ecosystem modeling of C dynamics in subtropical grasslands, through better representation of relationships between abiotic variables and decomposition as a function of heterotrophic respiration (Holland et al., 2000) This study was conducted in a unique subtropical biome to assess the effect of grassland management intensification on soil respiration. Previous research conducted in this region indicated that grassland intensification favors the accretion of total and particulate organic soil C in the long term (as reported in Chapter 2 and 3). Also, the increased accumulation of particulate organic C (a labile soil C fraction) with management intensification may increas e the sensitivity of soil respiration to critical controlling factors due to ease of microbe mediated decomposition and turnover of this pool, especially under suitable temperature and moisture conditions (Cambardella and Elliott, 1992; Cadisch et al., 1996; Franzluebbers and Stuedemann, 2002) Changes in the dynamics of soil respiration with management intensification were investigated by i) quantifying the rate of total soil respiration (R S ), heterotrophic soil respiration (R H ), and autotrophic soil respiration (R A ) in a gradient of management intensities ranging from native rangeland (lowest), silvopasture (intermediate), to sown pasture (highest) and ii ) assessi ng differences in sensitivity of the soil respiration variables to measured abiotic factors (S Temp and S Moist ) after > 22 years of grassland conversion.
65 Materials and Methods Study Sites This study was conducted on experimental sites at the University of F lorida Range Cattle relatively flat terrain (slope < 5%), under a subtropical climate with 10 year average annual precipitation of ~1650 mm, average minimum daily temperature of ~16.7C, and average maximum daily temperature of ~28.2C. The dominant soil series across the sites are Ona and Immokalee fine sand s (sandy siliceous, hyperthermic Typic Alaquods), which w ere developed on parent material of sandy marine dep osits (Soil Survey Staff, 1999; NRCS Websoil Survey, 2013) The three grassland ecosystems consisted of a gradient of management intensities including Florida native rangelands dominat ed by saw palmetto ( Serenoa repens ), slash pine ( Pinus elliotti ) bahiagrass ( Paspalum notatum ) silvopastures, and bahiagrass sown pastures. These ecosystems represent minimum, moderate, and high management intensities, respectively, as practiced within t his region. The silvopasture and sown p astures were both established (> 22 years ago) by converting fields of native rangelands, which is the natural vegetation within the region. Detailed information about the experimental fields, including the management practices, is presented in Chapter 2. Experimental Design The study was based on a comparative mensurative experimental design (Hurlbert, 1984; Arevalo et al., 2009) because the t wo replicate fields (6 ha each) of each management ecosystems are collocated; in other words, the treatment replicates are not spatially dispersed. This experimental design has an underlying assumption that the soil properties of the three ecosystems (nati ve rangeland, silvopasture, and sown pasture) were similar prior to the conversion and designation of each ecological unit (Hurlbert, 1984; Arevalo et al., 2009) Factors
66 s uch as t he uni formity of initial land use, flat terrain, and limited potential for variation in the textural class of the soil (Kalmbacher et al., 1984; Kalmbacher et al., 1993) support s the as sumption that the soil properties were similar before the management changes were imposed. Similar to most long term ecological research (e.g. Janzen, 1995; Millar and Anderson, 2004) prior data on soil properties before or immediately after conversion of the native rangelands are unavailable, and it is probable that some inherent but unquantified variability exists across the study area. Partitioning and Measurement of Soil Respir ation A modified box exclusion method (Hanson et al., 2000; Luo and Zhou, 2006) was used to partition heterotrophic and autotrophic respiration under each grassland ecosystem. In each sampling locati on, open ended fabricated aluminum boxes (30 x 30 cm) we re installed by carefully digging trenches (with minimal disturbance to the formed soil column) to sever the roots up to 30 cm soil depth (Figure 4 1). The boxes totally excluded roots from growing la terally into the soil column within this soil depth range where over 85% of roots are concentrated (as reported in Chapter 2). In June 2012, the vegetation within the boxes was clipped to ground level in order to truncate photosynthetic production and avoi d allocation of carbohydrate to the roots from aboveground production, thereby leading to eventual death and decay of existing roots. Measurement of respiration was delayed for 6 months to ensure the re stabilization of soil conditions within the exclusion box, and reduce disturbance effect accruing to the installation. The weekly in situ measurements of soil respiration (R S and R H ) were conducted at two random locations (~10 m apart) within each field Soil CO 2 measurement s were conducted weekly (between 8 and 11 am) during the winter ( 1 January to 31 March ) and summer ( 15 May 15 August) season in 2013. R S and R H were measured in situ with environmental gas monitor
67 (EGM 2) portable infra red gas analyzer (PP Systems, Amesbury, MA), equi pped with a soil respiration chamber (SRC 1) (Figure 4 2). The EGM 2 was recalibrated by the manufacturer to match the operational efficiency and accuracy of EGM 4. Before each weekly measurement event, two (2) polyvinyl chloride ( PVC ) collars (height = 2. 5 cm) were installed at each sampling location (one within and one outside the exclusion boxes) to ensure a snug fit of the SRC 1 chamber during measurement, and avoid errors related to potential leakage of respired C. The first PVC (PVC M) was placed with in the exclusion box and designated for measurement of R H Given that the vegetation is non existent and the roots are dead, the CO 2 efflux from this collar is attributable to microbial organisms that are metabolizing the available and accessible soil C wi thin the excluded soil column. The second PVC (PVC RM) was placed ~0.4 m away from each exclusion box to capture the total soil respiration (R S ). The vegetation within the PVC RM collar was consistently clipped to ground level before each measurement to av oid any influence of shoot respiration. Hence, there was no regrowth of vegetation between clipping and measurement. Autotrophic respiration was calculated as the difference between soil respiration and heterotrophic respiration. At each instance of soil respiration measurement, S Temp (C) and S Moist (% volumetric water content) at surface 10cm soil depth were also recorded, using HI 98331 temperature probe (Hanna Instruments, Carrolton, TX) and VG METER 200 soil moisture meter (Vegetronix Inc, Riverton, UT), respectively (Figure 4 2). Statistical Analysis Descriptive and statistical analyses were conducted using SAS 9.2 software. T wo way repeated measure ANOVA was used to determine significant differences in R S R H R A S Temp and S Moist between the different management systems and across seasons. Tukey test was performed to separate means when treatments were different Multiple li near regression (MLR) analyses were performed (Equation 4 1) to assess seasonal relationship of soil respiration with
68 controlling factors (S Temp S Moist and their interactions), based on change in management system. The explanatory power of each variable in the MLR model was assessed based on significance of significant variables are excluded from the final fitted model. Temperature sensitivity (Q 10 ) was derived from Equa tion 4 2 (Smith and Johnson, 2004) and the slopes of the linear regressions of log transformed R (S,A,H) vs. S Temp were compared to determine if the Q 10 significantly changed in response to management intensification. (4 1) (4 2) Where, Y is the log transformed respiration variable (R S R H, and R A ) ; b, c and d are slope coefficients for S Temp S Moist and S Temp S Moist interaction, respectively; a is a constant (intercept); and Q 10 is the quotient indicating rate of change in respiration with 10C change in temperature. Results Daily and Seasonal Ambient Temperature and Rainfall Daily minimum temperature during the measurement period r anged from 0 .1 C during the winter period to 24.5C during the summer period, respectively, while cumulative precipitation was 4.5 and 59cm during the 3 month winter and 3 month summer measurement period s respectively (FAWN, 2013; Figure 4 3). The ambient temperature during each round of measurement in the morning varied more in winter compared with summer, and this appears to influence the trend of measured S Temp under the three ecosystems.
69 Soil Respiration under Different Grassland Ecosystems Tot al soil r espiration (R S ) Measured rates of R S increased with grassland management intensity in the winter and summer ( P < 0.001 ) indicating that conversion from native rangelands to silvopasture and sown pasture generally changed the rate of soil respirat ion (Table 4 1). While increase in R S was observed with management intensification during the winter, R S only increased in sown pasture (100%) but not in silvopasture during the summer compared to the baseline native rangeland ecosystem Greater variability of R S was observed in the sown pasture ecosystems compared with native rangeland, but the temporal trend and variability was similar in native rangelands and silvopasture during both seasons (Figures 4 4 and 4 5). There was significant effect of season management interaction on R S observed ( P < 0.001) such that the magnitude of increase in R S from summer to winter was higher in sown pasture (~200%) compared to native rangeland and silvopasture, which both increased by similar magnitude of ~91% Heterotrophic r espiration (R H ) G rassland management intensification had effect on R H during the winter and summer season s ( P 1). Mean winter R H was higher in sown pasture and silvopasture compared with native rangelands (19% increase) while it increased with management intensity during summer (up to 35% increase in sown pasture) Over the summer period, the pattern of R H was fairly similar across all the ecosystems (Figure 4 5). The R H reached a maximum and then declined between mid May and early June in native rangeland while it was elevated in sown pasture after the third week of measurement, which coincides with period of peak net primary productivity (June to August) within this ecosystem T he average rate of R H increased from winter to summer in the order, silvopasture (109 % increase) < native rangeland (140% increase)
70 < sown pasture ecosystem (17 4 % increase). However, the re was significant interaction effect of season and management on R S ( P <0.001) Autotrophic r espiration (R A ) Winter and summer soil R A were greater in sown pasture (0.24 and 0.82 g CO 2 m 2 h 1 P = 0.03 and P < 0.001, respectively) co mpared with the baseline native rangeland (0.16 and 0.30 g CO 2 m 2 h 1 respectively). However R A was comparable in native rangela nd and silvopasture in both seasons ( P = 0.23 and > 0.1 respectively; Table 4 1). Generally, the average rate of R A increased from winter to summer. The magnitude of change was great er in the sown pasture ecos ystem (240% increase) compared with the native rangeland (97% increase), while the seasonal change was of similar magnitude in silvopasture (74% increase) and rangeland Furthermore, the temporal trend of R A during winter and summer was similar to that of R S (Figures 4 4 and 4 5) and there was significant effect of interaction between season and management Changes in Soil Temperature (S Temp ) and Moisture (S Moist ) Measured S Temp and S Moist During the winter period, management intensification from native rangeland to silvopasture increased S Temp ( ~6%), but no effect was observed in sown pasture ( Table 4 1) Within the same period, S Moist increased in response to management intensification to sown pasture (11.6%, P = 0.006) compared to native rangeland (10.2%) but no change was observed in silvopa sture ecosystem (10.3%; P > 0.1 ) During the summer period average S Temp was comparable between silvopasture and basel ine native rangeland but was great er in sown pasture ( P = 0.003). S Moist was lower in silvopasture (29.8%), but no change was observed in sown pasture (40.8%) compared with native rangeland (44.8%) ( Table 4 1). Generally, average S Temp and S Moist increased from winter to summer, and the magnitude of increase changed with
71 grassland management intensity (Figure 4 3; Table 4 1). Seasonal change in average S Temp was margina lly great er in sown pasture (8.8 C), but lesser in s ilvopasture (7.8C), compared with native rangeland (8.5C). The magnitude of change in average S Moist from winter to summer was 350% for native rangeland, 250% for sown pasture, and 190% for silvopasture Variability of R S R H and R A with S Temp and S Moist During the winter, the variability of R S and R H explained by S Temp and S Moist was fairly comparable between the baseline native rangeland and silvopasture with a marginal R 2 difference of 0.03 0.05, but the explained variability was notably lower in sown pasture (R S and R H lowered by 0.14 an d 0.36, respectively) compared with nat ive rangelands (Table 4 2). The variability of R A explain ed by both abiotic factors was near ly the same across the three ecosystems. Generally, in the fitted multiple linear regression model, there was no significant effect of the S Temp S Moist interaction (coefficient d, P > 0.1 ) on the variability of R S R H and R A Also, temperature sensitivity of soil respiration variables (calculated using Equation 4 2) changed slightly with management intensification. The Q 10 value of R S declined slightly from 1.69 in native rangelands, to 1.58 and 1.59 in silvopasture and sown pasture, respectively. Autotrophic Q 10 ranged from 1.72 in native rangeland to 1.62 in the silvopasture ecosystem, while heterotrophic Q 10 was highest in native rangeland (Q 10 = 1.65) and lowe st in the silvopasture ecosystem (Q 10 = 1.44) (Table 4 3 ). In the summer, the variability of R S explained by S Moist and S Temp changed but not always linearly with management intensity (Table 4 2). Bas ed on the best fit MLR model, great er variability of R S and R A was explained by S Moist and S Temp in silvopasture (R 2 increased by 0.26 and 0.21, respectively), compared with native rangeland, but this declined in the sown pasture (R 2 decreased by 0.20 and 0.14, respectively). The variability of R H with the abi otic factors declined in more intensively managed ecosystems (R 2 = 0.24 and 0.42 in silvopasture and sown
72 pasture), compared with native rangeland ( R 2 = 0.61). There were instances of significant S Temp S Moist interactions but there was no clear indication that such interactions were accentuated or attenuated by management intensification (Table 4 2). The temperature sensitivity of R S and R H increased with management intensification, with a highest Q 10 of 1.55 and 2.29, in sown pasture, compared to Q 10 valu es of 1.09 and 1.48 in native rangelands (Table 4 3 ). However, autotrophic Q 10 followed the patter native rangeland > sown pasture > silvopasture. Discussion and Conclusion Management Intensification Effects on Total Soil Respiration Rates F ield measuremen ts of soil respiration were conducte d only in the winter and summer hence, data were insufficient to estimate an average annual respiration rate. Overall soil respiration rates ranged between 0.30 1.28 g CO 2 m 2 h 1 The range of soil respiration rates in this study is similar to recently reported rates under comparable subtropical ecological condition s in southeastern U.S such as range of 0.01 0.8 g CO 2 m 2 h 1 soil respiration under increasing disturbance of a forest ecosystem in Georgia (Silveira et al., 2010) and 0.33 0.49 g CO 2 m 2 h 1 under control and irrigated grassland sites, in Subtropical Texas (McCulley et al., 20 07) Overall soil respiration is dependent on autotrophic (mainly root) and heterotrophic (mainly soil microbe) respiration. Both of these are strongly influenced by environmental conditions (mainly temperature and moisture) and vegetation (mainly plant productivity and type of land cover) (Raich and Schlesinger, 1992; Ma et al., 2005; Luo and Zhou, 2006) Despite the availability of very limited number of direct comparisons of so il respiration rates in natural and disturbed vegetation (Raich and Schlesinger, 1992; Raich and Tufekciogul, 2000) the findings from this research indicate that conversion from native vegetat ion to a more intensive (and disturbed) vegetation increased soil respiration, poss ibly due to increased ecosystem productivity (Raich and Tufekciogul, 2000; Carlisle et al., 2006) and soil C accretion (Chapter 2)
73 There was also an underlying seasonal response of soil respiration to management intensification. The persistently low winter respiration rates in the winter period is likely attributable to the strong control of climatic factors especially temperature and precipi tation, on ecological factors such as productivity and decomposition (Buyanovsky and Wagner, 1995; Xu and Qi, 2001; Saiz et al., 2006; Oyonarte et al., 2012) However, high net primary productivity and biochemical activity of soil microbes ( and roots ) during the summer are most likely responsible for the observed elevated soil respiration rate, relative to winter (Buyanovsky and Wagner, 1995; Luo and Zhou, 2006) The observation of greater soil respiration rates in the sown pasture compared with native rangeland, may be a result of complex interactions betw e en several factors, including, i ) elevated plant productivity in response to fertilization and change in predominant grass species (Raich and Tufekciogul, 2000) which can enhance both ro ot respiration and litter i nput s, and ii) likelihood of great er microbial activity due to the influence of grazing animals (including fecal returns) on nutrient cycling (Conant et al., 2001; McLauchlan et al., 2006) Interestingly, the magnitude of seasonal change in soil res piration and its components (i.e. effect of season management interaction) increased as management intensity increased and may be attributed to the increased prevalence of summer thrivi ng C 4 grass species. This also suggest s that grassland intensification within this region may foster increased sensitivity of ecological processes to abiotic or environmental factors in the long term Response of Heterotrophic Respiration to Management Int ensification Disentangling the various sources of CO 2 within the soil is essential for improved understanding of the impact of management intensification on soil C losses through respiration. The effect of management intensification on soil C losses through the respiration pathway is hinged on the role of so il heterotrophs. Although heterotrophic respiration increased with management intensification in winter and summer (Table 4 1) potentially lim ited substrate
74 availability and influence of critical edaphic factors, mainly temperature and moisture, may have posed constraints on enzymatic and extracellular processes of soil heterotrophs during winter (Luo and Zhou, 2006; Anderson, 2013) This may explain the observed similarity of heterotrophic resp iration in silvopasture and sown pasture during the winter, and conversely, the difference between both ecosystems in the summer. Generally, the increase in mean heterotrophic respiration rate (140 175 % higher) from winter to summer, which coincides with inreasing soil temperature and moi sture, further indicates that these edaphic factors may strongly leverage response of soil C respiration to changes in management intensity (Schi mel et al., 2001; Xu and Qi, 2001; Wang and Fang, 2009) Despite the comparable quantity of particulate organic C after converting native rangeland to silvopastoral ecosyst ems ( in Chapter 2 ) which can potentially foster similar ity in rates and ease of decomposition, the observed increase in heterotrophic respiration suggests that ecological conditions may have been more favorable for microbial activities under silvopasture than in native rangelands However, this contrasts with the fi ndings of Jenkins et al. (2010) who reported that in situ heterotrophic respiration was comparable under two adjacent woody vegetation regimes (with shrub and grass understory) in subalpine Australian ecosyste ms It is probable that changes in management practices or subtle physiological differences in vegetation type or structure (such as from native rangeland to silvopasture) may alter productivity and su b s trate availability Furthermore, changes in vegetatio n can facilitate shifts and succession in microbial communities (Reeder et al., 2000; Luo and Zhou, 2006; Anderson, 2013) which can result in notable increase or decrease i n overall heterotrophic respiration rates. Given that the grass component in silvopasture contributes to both the stable and labile C pools (in Chapter 3),
75 it will be relevant to determine the sources of heterotrophic C respired in order to fully elucidate the infl uence how heterotrophic respiration is modified in silvopasture. The observation of peak heterotrophic respiration under sown pasture is likely due to a number of factors that are uni que to this ecosystem such as i ) potential for higher bahiagrass derived net primary productivity and litter returns (Obour et al., 2009) ii) potential ease of microbial decomposition of the litter returns (Conant et al., 2004) and iii ) the net effect of higher cattle stocking rate on nutrient cycling to stimulate microbial activity (Liu et al., 2012) These findings contrast with reported research findings under different ecological conditions wher e lower heterotrophic soil respiration was observed in grass dominated vegetation (pastures) compared with reference less intensive woody grass/shrubland vegetation (Jenkins and Adams, 2010; Bes ar et al., 2011) Disparity in findings may be attributed to the uniqueness of several underlying factors within each ecoregion, including productivity, soil organic C stock, soil structural and chemical properties, and soil climate interactions, and ag e of studied ecosystem (Raich and Tufekciogul, 2000; Gavrichkova, 2008; Anderson, 2013) Response of Soil Autotrophic Respiration to Management Intensification Autotrophic respiration (mainly from roots) constitutes a major proportion of overall soil respiration. In this study, the contribution of root to overall soil respiration ranged from 47 to 64% with increasing management intensity of the grassland ecosyste ms A wider range (10 90% ) has been reported for general non forest vegetation (Hanson et al., 2000) and a narrower range (17 40%) for grasslands (Buyanovsky and Wagner, 1995; Raich and Tufekciogul, 2000) However, our result s agree with the fairly recent range for grasslands (8 64%), reported by Wang & Fang (2009) Given that increased root respiration i s mainly associated with improved production of root biomass quantity (Vose et al., 1995) autotrophic CO 2 efflux is often influenced by the rooting characteristics which modulates the rate of intercellular diffusion a nd
76 exchange of gases within the root zone (Hendricks et al., 1993; Graham et al., 2012) These two factors may explain the observation of similar autotrophic respiration rates ~25 years af ter conversion of woody native rangelands to woody silvopastures, and its increase with conversion to non woody sown pasture. It is important to note that root respiration does not im ply loss of stored soil C, but predominantly indicate s the rate of consumption of organic compounds (mainly photosynthate) supplied by aboveground plant organs (Horwath et al., 1994; Hanson et al., 2000; Luo and Zhou, 2006) Hence t he fertilized and optimally stocked sown pasture ecosystem may allocate more photosynthate (derived through higher productivity) to belowground parts than the woody native rangeland vegetation, and foster the observed increase in autotrophic respiration (up to 170% increase; Table 4 1). Similar to the findings in this study, Smith and Johnson (2004) reporte d that root respiration was greater er in grasslands than adjacent woodlands, although to a lesser extent (61%), under a temperate mid continental climate, and they concluded that root biomass could not explain the lower respiration rates in the woody vegetation. Management Intensification Impacts Respiration Sensitivity to Soil Temperature and M oisture. Impac ts of management intensification on sensitivity of soil respiration to abiotic factors may be complex and dependent on trajectory of vegetation conversion (Raich and Tufekciogul, 2000; Luo and Zhou, 2 006) as observed in this study. For instance, soil moisture was greater in sown pasture compared with native rangeland during winter, but this only translated to great er autotrophic respiration while no change was observed in heterotrophic respiration and overall soil respiration. Potential canopy effect (including interception of solar radiation and precipitation, and trapping of heat and moisture) under different vegetation types can readily modify soil microclimate in response to management intensifi cation or land use conversion.
77 Also, changes in net primary productivity and transpiration rates with management intensification can impact soil plant moisture dynamics, and account for increased winter soil moisture in sown pasture compared with native ra ngeland. Consistent with our findings, Gavrichkova (2008) also reported an increase of soil moisture after subjecting natural Mediterranean grasslands to management. The Q 10 value is the rate of increase in res piration when temperature increases by 10C and it is an important parameter for estimating soil respiration changes (Lloyd and Taylor, 1994; Fang et al., 1998; Xu and Qi 2001; Luo and Zhou, 2006) The range of autotrophic and heterotrophic Q 10 values in this study (0.92 2.29) is well within the range reported (0.9 4.6) across tropical and temperate grassland ecosystems (Wang and F ang, 2009) and specifically the reported range (0.9 2.9) under similar subtropical conditions (Fang et al., 1998; Holland et al., 2000) The observed change in temperature sensitivity of autotrophic and heterotrophic respiration with management intensification seems to differ between seasons. This is likely due to potential acclimation of soil respiration under suitable weather conditions, especially during summer (Luo et al., 2001; Gavrichkova et al., 2008; Wang and Fang, 2009) Also, in addition to observed increased i n mean soil temperature with intensification (Table 4 1), differences in vegetation chara cteristics and soil C stocks (as earlier reported in chapter 2), and their potential influence on litter recalcitrance and microbial activity, can impart changes o n temperature sensitivity (Raich an d Tufekciogul, 2000; Luo et al., 2001) Results from this study are similar to the findings of Gavrichkova et al., (2008) who also reported great er heterotrophic Q 10 in managed grasslands compared with unman aged grassland, and vice versa for autotrophic respiration. As suggested by Luo et al. (2001) ecosystems that contain more soil C are likely to become more sensitive, and less likely to acclimatize, to projected global warm ing scenarios.
78 This is because great er soil C will induce physiological and ecological adjustments including potential changes in litter dynamics, prolifer ation of microbial community and increased microbial activity, and potential increase in soil respiratory capacity (Atkin et al., 2000; Luo et al., 2001) Therefore, the reported increase in soil C after conversion of native rangelands to more intensively managed sown pasture ecosystems (Chapter 2 and 3) may explain the increased temperature sensitivity during warmer season. The moisture sensitivity of autotrophic and heterotrophic respiration coul d be confounding due to the high variability of soil moisture in time and space (Luo and Zhou, 2006; Qi et al., 2011; Anderson, 2013) This limits the possibility of clearly eluc idating the effect of management intensification on moisture sensitivity of soil respiration in this study. However, the seasonal changes can amplify or dampen moisture sensitivity of both autotrophic and heterotrophic respiration, such that soil moisture explained significant variability in soil respiration across all ecosystems during winte r, but not in summer (Table 4 2 ). This is supported by numerous studies which have suggested that moisture sensitivity can decline greatly when moisture conditions are not limiting, especially during high rainfall periods during summer (Fang et al., 1998; Knapp et al., 1998; Raich and Tufekciogul, 2000; Smith and Johns on, 2004) Soil temperature and moisture are expected to interact strongly in influencing soil respiration rates due to their joint control on root and microbial processes (Conant et al., 2000; Luo and Zhou, 2006) However, non sensitivity of soil respiration variables to temperature moisture interaction across all ecosystems during the winter suggests that potential for interaction of these abiotic variables may be limited when climatic conditions ar e not optimal for soil biochemical processes.
79 In conclusion, this study indicates that increasing management intensity by converting native rangeland to woody tree integrated silvopasture and sown pasture management system s increas e d the long term soil res piration rate in this subtropical grassland biome Also, based on the net change in heterotrophic temperature sensitivity between winter (lowered by 13%) and summer (increased by 55%), the previously reported increase in soil C stock with intensification ( Chapter 2 ) may be susceptible to faster turnover under warming climate scenarios It has been reported that overall soil C stocks w ere similar in both silvopasture and sown pasture after conversion from native rangelands (in C hapter 2), however, the observ ed decline of temperature moisture sensitivity (in silvopasture) suggests a potential for improved soil C resilience under this integrated tree grass ecosystem. Improved resilience (and longer residence) of add itionally sequestered soil C after intensification may be crucial for long term ecological resilience, especially with changing climatic conditions (IPCC, 2007) These findings are relevant for sustainable grassland management, especially within subtr opical ecoregions, and add to the understanding of changes that may occur in rates of soil C losses as native grasslands are converted to more productive grassland ecosystems.
80 Table 4 1 Impact of grassland management intensification on in situ soil respiration, soil temperature, and soil moisture. Measured variables Winter Summer Native r angeland Silvopasture Sown p asture Native r angeland Silvopasture Sown p asture R S (g CO 2 m 2 hr 1 ) 0.300.03 c 0.340.04 b 0.410.04 a 0.640.02B 0.680.03B 1.280.06A R H (g CO 2 m 2 hr 1 ) 0.140.02 b 0.160.01 a 0.170.01 a 0.340.03 C 0.360.03B 0.460.03A R A (g CO 2 m 2 hr 1 ) 0.150.02 b 0.180.02 b 0.240.03 a 0.300.03B 0.320.03B 0.820.05A S Temp (C) 16.00.9 b 16.90.7 a 16.60.7 a 24.50.4 B 24.70.3 B 25.40.2 A S Moist (%) 10.20.2b 10.30.3b 11.60.4a 44.87.1 A 29.93.2 B 40.754.6 A Values represent Meanstandard error (n=2) of total soil respiration (R S ), autotrophic respiration (R A ), and heterotrophic respiration (R H ), soil temperature at 10cm depth (S Temp ), and soil moisture at 10cm depth (S Moist ). Significantly different ecosystems within each season are denoted by different letters, according to the Tukey test (P<0.05). Difference in letter case denotes significant difference between season s.
81 Table 4 2 Effect of grassland management intensification on relationship of soil respiration with soil temperature and moisture. Ecosystem Soil r espiration p artition ing R S R H R A R 2 a b c d R 2 a b c d R 2 a b c d Winter Native r angeland 0.92 0.95 *** 0.39 *** 0.63 *** n s 0.96 0.39 *** 0.52 *** 0.53 *** ns 0.75 0.56 ** ns 0.67 ** ns Silvopasture 0.95 1.01 *** 0.46 *** 0.72 *** n s 0.91 0.44 *** 0.42 *** 0.72 *** ns 0.79 0.54 *** 0.44 ** 0.64 *** ns Sown p asture 0.78 0.77 *** 0.41 ** 0.61 *** n s 0.62 ns 0.79 *** ns ns 0.77 0.61 *** ns 0.69 *** ns Summer Native r angeland 0.45 ns 1.07 ** 0.99 ** n s 0.61 3.03 ** 1.48 14.7 ** 15.3 ** 0.30 ns 0.55 ns ns Silvopasture 0.71 6.02 ** 1.68 22.1 *** 23.3 ** 0.24 ns ns ns ns 0.51 6.14 1.9 20.7 ** 21.8 ** Sown p asture 0.25 ns 0.50 ns n s 0.42 1.58 0.64 ** ns ns 0.16 ns ns ns ns Total soil respiration (R S ); Heterotrophic respiration (R H ); Autotrophic respiration (R A ); a, b, c and d are coefficients of intercept, temperature effect, moisture effect, and temperature moisture interaction, respectively ; ns denotes non indicates that the coefficient of all regression variables were not significant.
82 Table 4 3 Grassland management intensification effects on temperature sensitivity (Q 10 ) of soil respiration. Ecosystem Q 10 values R S R H R A Native r angeland 1.69 1.65 1.72 Winter Silvopasture 1.58 1.49 1.64 Sown asture 1.59 1.44 1.71 *** *** *** Native r angeland 1.09 1.48 2.02 Summer Silvopasture 1.36 1.95 0.92 Sown p asture 1.55 2.29 1.43 *** *** *** Total soil respiration (R S ); Autotrophic respiration (R A ); Heterotrophic respiration (R H ); *** denotes significant difference of regression slopes across grassland ecosystems ( P < 0.01).
83 Figure 4 1 Images showing the fabricated exclusion boxes for pa rti ti oning soil respiration compo nents, and the field installation of the boxes (only native rangeland shown). (Photo credits to the author)
84 A B C D Figure 4 2 Images showing field set up for measuring soil respiration and abiotic control variables. A) PVC collar for measuring heterotrophic respiration. B), PVC collar for measuring total soil respiration. C) Measurement with portable EGM 2. D) Closed up view of soil respiration chamber, soil temperature probe, and soil moisture sensor. (Photo credits to the author).
85 Figure 4 3 Weather data and measured soil climate variables in a management intensity gradient of subtropical grassland ecosystems. A) Average ambient and soil temperature in winter. B) Average ambient and soil temperature in summer. C) Precipitation and soil mois ture during the winter. D) Precipitation and soil moisture during the summer. Each data point represents the mean of four measurements within two field replicates. Ambient daily minimum temperature (C) and precipitation (cm) were obtained from the databas e of Florida Automated Weather Network (FAWN) station located at the experimental site http://fawn.ifas.ufl.edu/ Native rangeland is represented in circles, silvopasture in triangles, and sown pasture in squar es. Grey filled legend symbol represents summer season measurement while open symbols represent winter. Temperature (C) Moisture (%) Julian Day Winter Summer Precipitation (cm) A B C D
86 Soil CO 2 Efflux (g CO 2 m 2 hr 1 ) Julian Day Figure 4 4. Winter soil respiration rates under a management intensity gradient of subtropical grassland ecosystems. A ) T otal soil respiration. B) Heterotrophic soil respiration. C) Autotrophic soil respiration. Each data point represents mean of four sampling instances in two field replicates. A B C
87 Soil CO 2 Efflux (g CO 2 m 2 hr 1 ) Julian Day Figure 4 5 Summer soil respiration rates in a management intensity gradient of subtropical grassland ecosystems A ) T otal soil respiration. B) Heterotrophic soil respiration. C) Autotrophic soil respiration Each data point represents mean of four sampling instances in two field replicates. B A C
88 CHAPTER 5 APPLICATION OF THE PROCESS BASED DNDC MODEL FOR PREDICTING IMPACT OF MANAGEMENT INTENSIFICATION ON SOIL CARBON DYNAMICS IN SUBTROPICAL GRASSLAND ECOSYSTEMS Background The unprecedented rise of atmospheric carbondioxide (CO 2 ) since 1850 has been attributable to anthropogenic emissions from various source s including land use (Schimel, 1995; IPCC, 2007) In order to understand the role of various potential carbon (C) sinks in mitigating rising atmospheric CO 2 (McSherry and Ritchie, 2013) it is necessary to understand how changes in land use management can potentially impact ecosystem C dynamics. Grassland ecosystems are critical component s of the terrestrial C sink because they cover about 40% of the ea rth surface (White et al., 2000; Svejcar et al., 2008; Wang and Fang, 2009) contain about one third of global C reserves (Schlesinger, 1977; Hall et al., 2000) and store a relatively high amount (up to 90%) of ecosystem C in soil (Schuman et al., 2001; Liu et al., 2011) Similar to other terrestrial e cosystems, management of subtropical grasslands is increasingly being intensified to increase productivity per unit land area (White et al., 2000; O'Mara, 2012) These ecosystems are sensitive to anthropogenic disruptions, with potential feedback on global C cycles and biogeochemistry (Janzen et al., 2011; IPCC, 2007). Consequently, the growing trend of land use conversion and intensification of management has been identified as one of the major d rivers of global climate change due to loss of sequestered soil C which elevates atmospheric CO 2 concentration (Post et al., 1 982; Post and Kwon, 2000; Conant et al., 2001; Guo and Gifford, 2002; McSherry and Ritchie, 2013) The assessment of long term changes in ecosystem C stocks and fluxes to determine impacts of adopted management practices is often cost prohibitive, and r esource intensive. Hence, ecosystem process based models (also mechanistic models) are becoming prominent to
89 support measured data. They integrate the understanding of ecological processes, such as soil C (and N) dynamics by utilizing and predicting measured ecological properties, such as C and N stocks and/or fluxes under varying management regimes and time scales (Christensen, 1996; Fynn et al., 2009) Generally, mechanistic models of soil organic matter dynamics function based on the differential dynamics of C (and N) pools in relation to decomposition and turnover rates under the influence of different bio tic and abiotic factors such as temperature, water, pH, nutrients, oxygen, cl ay content, cation exchange capacity, type of crop/plant cover, and tillage method (Smith et al., 1997; Beheydt et al., 2007) Different models incorporate these factors in a variety of ways eithe r in single module or several modules (multi compartment) forms. Thus, existing models have inherent strengths and weaknesses depending on factors such as, i ) applicability to the physical and biological condition s of the region under study; ii ) the soundn ess of experimental data incorporated into the parameterization of model; and iii ) richness of boundary and initial conditions, including climate data, land use and geographical information available for the analysis (P ost et al., 2001) Many process based models are developed for specific ecosystem s or biome s and their application is often limited to regions that share identical or similar conditions (especially climate and abiotic parameters ). However, due to incr easing need for reliable prediction of ecosystem C dynamics at field, regional or global scales (Paustian et al., 1997; Post et al., 2001) it is imperative to improve the applicability of pro cess based models across varying ecological condition s through continuous improvements of model structure, recalibration, and validation based on observations in different ecosystems (Smith et al., 1997) In contrast to other climatic regions few studies have focused on assessing management induced changes in C dynamics within subtropical grasslands (Liu et al., 2011; Silveira et al.,
90 2013) Hitherto, few researchers have tested process based models for assessing C dynamics in some subtropical ecosystems (Batjes and Sombroek, 1997; Trettin, 2001; Wang et al., 2011b) however, process based soil C models are yet to be developed or adapted for quantifying soil C changes in response to management intensification in subtropical grasslands ecosystem s Subtropical ecoregion s represent a unique transition between t emperate and tropical climates, hence, the validation and application of process based models within this ecoregion is important for both regional and global C balance analysis. Th is study is focused on evaluating the performance of a process based denitr ification and decomposition (DNDC) model (Li et al., 1992 ; Li et al., 1997) for assessing soil C dynamics in subtropical grassland biome s The objective was to determine the applicability of the model for quantifying the long term effect of varying grassland management intensity on soil C dynamics. Model testing was conducted by validating model outputs against observed soil respiration (R S ) and its components (heterotrophic R H and autotrop hic R A respiration), and abiotic control factors (soil temperature S Temp and moisture S Moist ). Such model validation and testing routines are critical for reliable process based evaluation of management strategies that may enhance soil C sequestration and foster ecological sustainability in subtropical grasslands (Li et al., 1992; Batjes and Sombroek, 1997; Smith et al., 1997) Materials and Methods Model Selection and Description The DNDC model (Li et al., 1992) is a freely available process based ecosystem C model (http://www.dndc.sr.unh.edu) and it was selected for this study based on its potential to simulate biogeochemical interactions and C dynamics under warm wet ecological conditions (Trettin, 2001; Dai et al., 2012) which is a characteristic of our study site Criteria considered include, i ) detailed level of incorporation of re levant environmental variables ii ) differentiation of s oil
91 layers iii ) ability to capture redox reaction which may influence denitrification processes under anaerobic conditions and iv ) presence of alternating hydroperiod submo dels to capture dynamics of transiently wet moisture condition s (Smith et al., 1997; Trettin, 2001) The DNDC model contains five interacting sub models namely, soil climate and thermal hydraul ic, aerobic decomposition, denitrification, fermentation, and plant growth (Li et al., 1994; Li, 2000; Giltrap et al., 2010) The model partitions soil C into four pools: labile litter, resistant litter, humads, and humus pools, and it simulate s biogeochemical processes on hourly time step s with daily outputs ( Figure 5 1 ). The DNDC model simulates soil processes at five discrete horizontal soil layers (0 50 cm at 10 cm inc rement s ) and allows for allocation of plant derived C along the profile. Soil physical properties such as bulk density, porosity and hydraulic parameters are assumed to be constant across all soil layers while most soil properties such as soil moisture, te mperature, pH, C and N pools can vary between layers (Giltrap et al., 2010) Similar to most existing process based models, DNDC adheres to the principle of first order kinetics to simulate decomposition and adjusts rate constants according to soil water content, soil temperature, and soil texture (Li et al., 1992; Li et al., 1994) It was initially developed for modeling C and N dynamics in temperate agroecosystems, but it has been tested and improved for a range of ecosystems within temperate and tropical ecoregio ns (Li et al., 2003; Giltrap et al., 2010) Study Area This study is based on experimental subtropical grassland ecosystems at University of Florida Range Cattle Research and Education Cen The site is within a subtropical climat ic zone with average annual precipitation of ~1650mm, average minimum daily temperature of ~16.7C, and average maximum daily temperature of ~28.2C. The study site i s chara cterized by seasonally fluctuating water table and the dominant soil series is Ona fine sand (sandy, siliceous, hyperthermic Aeric Alaquod) formed from marine
92 deposit (Soil Survey Staff, 1999) The experimental f ields include native rangelands, dominated by saw palmetto [ Serenoa repens (Bartram) Small], and adjacent b ahiagrass ( Paspalum notatum Flgge) fields which were established by clearing a portion of the native rangelands ~30 years ago. The adjacent fields r epresent a transition from a baseline low management intensity native rangeland to a higher management intensity and more productive sown pasture grassland. The native rangeland ecosystem has never been fertilized and was used only for winter grazing whil e the bahiagrass pasture wa s fertilized annually (~67kg N ha 1 yr 1 ) and stocked rotationally (one week on, one week off) for 7 months per year. Detailed information about the experimental sites and the specific management practices are presented in Chapte r 2. Field Observation of Soil Respiration and Abiotic Variables Field observations of selected variables are required to validate DNDC in the baseline (relative to native rangeland) and management intensification (conversion to sown pasture) mode. Hence, in a different study on this site (Chapter 4) EGM 2 portable infra red gas analyzer equipped with a soil respiration chamber (PP Systems, Amesbury, MA) was used to measure weekly in situ soil respiration (including heterotrophic and autotrophic sources) during winter (Jan uary to Mar ch ) and summer (May to Aug ust ) of 2013. Briefly, 2 random locations were selected in each field, and root exclusion aluminum boxes were installed (to 30 cm soil depth) 6 months prior to measurements. PVC collars (diameter = 10 cm, height = 2.5 cm) were installed within and outside the exclusion boxes. The measured soil respiration within and outside the box represents R H and R S respectively, while R A was calculated as the difference between R S and R H Soil temperature (S Temp ) and soil moisture (S Moist ) were measured at each instance of measuring soil respiration. R S and its components (R A and R H ) are critical output variables of soil C dynamics, while S Temp and S Moist are critical ecological control variables that influence
93 und erlying processes of soil C changes. Further details of in situ observation of the variables and results are presented in Chapter 4. Model Validation and Application Baseline and Intensified Management Condition s An initial 21 0 year spin up (also baselin e simulation ) o f the model was first conducted to ensure that the simulated soil C attains steady state as would be expected after long term management of an ecosystem (Li et al., 1994 ) Since the native rangeland is dominated by shrub vegetation, mainly s aw palmetto interspersed with a mixture of different species of grasses and shrubs (Kalmbacher et al., 1984) cr op parameter were designated as the vegetation for the baseline scenario All other required input values for model specified parameters were gleaned from background literatures on the study site ( Tables 5 1 and 5 2 ). Complete on site weather dataset was only available from 1999 2013, hence, the 15 years daily we ather data was recycled for the entire simulation period The recycled multi year weather data allowed for the incorporation variability in historical weather with in model simulation period despite unavailabili ty of weather data prior to 1999. Input weather variables include daily minimum and maximum temperature (C), precipitation (cm), windspeed (m/s), solar radiation (MJ/m 2 /day), and relative humidity (%). After the spin up run, t he model set up was further refined by extending the simulation period for 30 years (i.e. year 210 240) to represent i) conversion of baseline native rangeland to sown pasture condition and ii) continuation of the baseline condition For the sown pasture condition the crop type (perennial grass) and management practices (grazing, fertilization, and tillage) corresponding to the new and current management condition after conversion of the native rangelands were specified All other parameters were th e same as the baseline scenario /native rangeland condition, and daily outputs of C dynamics were generated for each year simulated. In a separate run of the model, the default field capacity ( FC ) and wilting point
94 ( WP ) values were adjusted to field observed values (data not reported), b ased on previously reported sensitivity of DNDC to WFPS FC and WFPS WP (Beheydt et al., 2007; Krbel et al., 2010 ) However, the adjustments to WFPS did not improve the accuracy in simulation of observed R S R H R A S Temp or S Moist At the end of the 30 year till present period (i.e. year 240 in simulation), the model simulation period was further extended to include a 75 year (year 240 315) simulation of soil C dynamics under different futur e management scenarios Five scenarios were defined including, i) continuation of the baseline native rangeland condition BBB, ii) restoration from current sown pasture condition to native rangeland condition BSB iii) reducing current/conventional sow n pasture management intensity by half BSH, iv) continuation of current sown pasture management intensity BSS, and v) doubling of current sown pasture management intensity and later restored to native rangeland condition (Table 5 3 ) Model Validation Performance Metrics The model simulated outputs for each Julian day (in simulation year 240) corresponding to the day (and year) of field observation were extracted, and different model performance metrics were used to assess the accuracy and precision of DNDC simulation, as defined in several similar model testing studies (Loague and Green, 1991; Smith et al ., 1997; Chang and Laird, 2002; Beheydt et al., 2007; Krbel et al., 2010; Abdalla et al., 2011) The metrics (Equations 5 1 5 4), include coefficient of determination (R 2 ), root mean square error (RMSE), ratio of prediction to deviation (RPD), modeli ng efficiency (EF). For a perfect model fit, R 2 = 1, RMSE = 0, RPD >2.0, and EF = 1. (5 1)
95 (5 2) (5 3) (5 4) Where, is the observed value, is the DNDC predicted value, is the mean observed value, SD is the standard deviation of p rediction and is the number of years within the simulation period. Results and Discussion DNDC Validation under Native Rangelands (baseline) and Sown Pasture (Intensified Management) C onditions The baseline spin up run ( 0 210 years ) showed that the model was able to attain a steady state under defined set of management conditions, as shown by the soil C stock (0 30cm) and total annual soil CO 2 efflux which equilibrated at ~41.6 Mg C ha 1 and ~10.2 Mg C ha 1 yr 1 respectively ( Figure 5 2 ). Further simulation of C dynamics for the period elapsed after conversion from baseline conditions to more intensified sown pasture management system resulted in a linear increase (~ 1 Mg C ha 1 yr 1 ) of soil C stock throughout the 30 year perio d (i.e. simulation year 210 240 ). Meanwhile, soil C stock remained unchanged (equilibrium maintained) with the continuation of the baseline (native rangeland) conditions through the 30 year period At the end of the 30 year period (year 240), the cu mulative soil C stock (0 30 cm) in the native rangeland and sown pasture scenarios was 4 2 Mg C ha 1 and 72 Mg C ha 1 respectively (Figure 5 2) Although simulated soil C (at 0 30cm depth) under native rangeland condition is
96 closely in agreement with fie ld observation of soil C stocks at the same depth in this study site (i.e. 41 Mg C ha 1 ; Chapter 2) the simulated soil C stock in sown pastures was fairly higher (~16%) than the observed quantity (62 Mg C ha 1 ) Several researchers (Li et al., 2003; Liu et al., 2006; Tang et al., 2006; Wang et al., 2008) have reported that DNDC closely simulated the observed soil C stocks in different cropping systems and climatic regions. However, allocations of soil C at different soil depth intervals are often not reported. I n this study, there was disparity between t he simulated and observed distribution of soil C at different soil dep th intervals. For instance, under native rangeland management system the simulated soil C stocks at 10cm, 20cm, and 30cm soil depths was 9 3, 18.4 and 14.3 Mg C ha 1 respectively, but the observed soil C stocks at these depths were 23.8, 8.7, and 8.5 Mg C ha 1 (as presented in chapter 2). Similar observations were made in the sown pasture where simulated soil C stocks at 10cm, 2 0cm, and 30cm soil depths was 21.1, 33.5 and 17 9 Mg C ha 1 respectively, and the observed soil C stocks at these depths were 31.1, 12.7, and 10 Mg C ha 1 respectively. This considerable disparity between simulated and observed soil C stocks at different depths, especially at the top 10cm, can alter the magnit ude of microbial activity by influencing quantity of accessible C substrate within the well aerated layer of the soil (Batjes and Sombroek, 1997) and may influence other soil processes related to transformation and turnover of soil C. The daily simulation outputs of R S R H R A S Temp and S Moist for the simulation year which coincides with the year of in situ field observatio ns (year 240) are presented in Figure 5 3. Generally, R S and R A followed fairly similar trend year round, due to low magnitude of simulated R H The model simulation showed seasonality in the daily respiration rates (and abiotic variables), however, this was more strongly expressed in simulated R S and R A than R H Under the baseline condition, R A accounted for about ~90% of the total R S while it accounts for ~75%
97 under the sown pasture. This contrasts to field observations in the study site (presented in C hapter 4) which indicates that observed proportion of R S attribut able to R A sources ranged from ~45% in native rangeland to 60% in sown pasture ( Table 5 4 ). Within the field observation period, the simulated mean R H in native rangeland was about an order of magnitude lower than mean R A (Figure 5 3; Table 5 4 ) while th e overall mean of simulated R S (0.34 g CO 2 m 2 hr 1 ) was fairly lower than the mean of the observed R S (0.47 g CO 2 m 2 hr 1 ). Also, in sown pasture the mean of simulated R H (0.18 g CO 2 m 2 hr 1 ) was lower than the mean of simulated R A (0.64 g CO 2 m 2 hr 1 ), but the mean of simulated R S was comparable to the observed mean (Table 5 2). Similar observation was made by Li et al. (1994) and Dietiker et al. (2010) who indicated that simulated R S were close to the observed values in temperate grasslands and agroecosystem despite season based over and underestimation of CO 2 fluxes. Given that the contribution of R A to R S under field conditions vary widely among diff erent studies, ranging from ~15% to 90% (Raich and Tufekciogul, 2000; Luo and Zhou, 2006; Wang and Fang, 2009) estimation of fraction of R S attributable to R A so urces falls within the broad range. The simulated proportion of R A under the sown pasture condition is in agreement with the findings of Li et al. (1994 ) who reported that the contribution of R A to R S simulated for grasslands (i n sandy soil ) and croplands ( in silty loam soil ) conditions ranged from 30 70%. T he higher proportion of simulated R A contribution to R S in the native rangeland, compared to field observations seems to be relat ed to the allocation of higher p ro portion of biomass to the root. Generally, model outputs showed that disproportionate percentage (~90%) of annual gross primary productivity (GPP) was allocated to root respiration throughout the simulation period (data not shown) with the remaining fra ction accruing to net primary productivity (NPP). By implication, based on the model structure (Li et al., 1994) it is expected that this would favor
98 more active root processes (including root growth, maintenance, and ion uptake and transport ) at the expense of heterotrophic processes, than would be obse rved under field conditions This explains the unusually low R H at equilibrium since the soil C respired by decomposers is directly propor tional to ecosystem NPP. Changes in DNDC core parameterization by reducing the proportion of GPP allocated to root processes may be an effective means to improve the estimation of R H under this ecological condition. In DNDC, S Temp and S Moist play critical roles in simulating C dynamics because they jointly influence soil C stocks and fluxes in the decomposition sub model. M icrobial activity (and decomposition rate) is regulated by a temperature moisture reduction factor which retards decomposition rates relative to optimum temperature or moisture conditions (Li et al., 1992 ) In the native rangeland, the simulated S Temp within the field observation period varied between 19C and 30C, while S Moist varied between ~ 0.0 1% and 46 %, in winter and summer, respectively (Fig ure 5 4). I n the sown pasture, simulated S Temp was the same as simulated for native rangeland but S Moist ranged from 2% during winter to 39 % during summer. Generally, the simulated mean S Temp was greater than observed (~5C greater), while simulated mean S Moist was lower than the observed in both native rangeland and sown pasture (1 5 % and 8 % difference, Temp and S Moist is modulat ed by the thermal hydraulic sub model which is driven by soil physical properties and weather data (Li et al., 1992 ; Li et al., 1994 ) However, since the inherent thermo hydraulic relationships in the model are mostly drawn from studies focused on temperate ecosystems (Li et al., 1992; Li et al., 1994) its systemic poor performance in simulating the critic al soil climate variables point to the need for r ecali bration of this sub model relative to subtropical climate For instance, net soil thermal conductivity depends on the type of soil and water content, while soil water content depends on
99 soil water tension and unsaturated hydraulic conductivity. The ex ponential formulations for these relationships in DN DC were derived from research findings on limited soil data (Clapp and Hornberger, 1978) which may not represent soil climate relations in subtropical ecosystems. Many studies have attempted to recalibrate DNDC to diverse ecological conditions, however, deviations between observed and simulated soil moisture still persists in application of DNDC to agroecological ecosystems in both temperate and tropical ecoregions (Smith et al., 2008; Abdalla et al., 2009; Krbel et al., 2010; Ludwig et al., 2011) Predictive Performance of DNDC Relative to Field Observations In the native rangeland, the model performed well in simulating the variability of measured R S and S Temp (R 2 = 0.74 and 0.9 5 respectively), average for R H and R A (R 2 = 0. 42 in both cases ), and fair for soil moisture (R 2 = 0. 27 ), however the variability explained were all significant (P<0.01 ; Table 5 5 ). Similarly, in sown pasture, the model seemed to perform well in simulating the variability of observed R S R A S Temp and S Moist (R 2 = 0.75 ~ 0.97, P<0.001), but performed poorly in simulatin g R H (R 2 = 0.07 P=0.16). Generally, the higher RPD values ( 0.78 1.81 ) and EF ( 0.73 0.68) were obtained in simulation of R S under the sown pasture, compared to native rangelands (RPD = 0.50 1.1; and EF = 0.31 0.13) Despite the relatively high RPD and EF value obtained in simulation of R S under sown pasture condition ( Table 5 5 ), it is not sufficient to conclude that DNDC is reliable for simulating C dynamics under the intensified management condition because of the poor RPD and EF in control va riable (such as temperature and moisture ) and particularly with respect to soil respiration partition (such as R H vs. R S ) The performance of DNDC in simulating the dynamics of R S its component s (R A and R H ) and the abiotic control factors (S Temp and S Moist ) were poor under the nat ive rangeland condition (RPD<1.1 ; Table 5 5 ). However, th e im prov ement of the validation metrics (especially RPD and EF) in simulation of these var iables under the sown pasture indicate that
100 further tweaks in the model can improve its predictive performance (Chang and Laird, 2002; Giltrap et al., 2010; Ludwig et al., 2011) Furthermore, there were several instances of season b ased under estimation (e.g. R H in sown pasture) and over estimation (e.g R A in sown pasture), and general over prediction (e.g. S Temp in both native rangelands and sown pasture) of these variables (Figure 5 4). These results are in agreement with several re search findings that have indicated poor performance of DNDC for simulating soil climate variables, especially S Moist in agroecological systems (Tonitto et al., 2007; Giltrap et al., 2010; Dietiker et al., 2010; Ludwig et al., 2011) After multiple recalibration attempts to improve performance of DNDC in predicting WFPS across different agroecosystems, Ludwig et al., (2011) reported lowest RMSE value of 24.7 and highest EF of 0.6, while Tonitto et al., (2007) reported RMSE value of 26 and EF of 0.6. Given that DNDC is multi structured and p rocess driven, poor prediction of control var iables can upset the soil plant relationship and alter processes related to microbial decomposition, litter decay, translocation of nutrients (Li et al., 1992 ; Li et al., 1994; Abdalla et al., 2011) Consequently, the observed poor performance of DNDC in simulating R A and R H may likely be improved by further improving the thermo hydraulic sub model to better simulate soil plant climate interactions. More importantly, there is also critical need to address the allocation of gross primary productivity to growth (NPP) and root respiration. Long term R S and Soil C S equestration under Alternative Management Scenarios. Based on the performance of DNDC in simulating the variability of R S and the close estimation of soil C stock in both native rangeland and improved pasture, the model was applied for long term (7 5 years) simulation of C dynamics under alternative management scenarios to better understand the long term effects of management practices In the scenario based simulation of long term grassland management (year 240 315) soil C initially increased at similar linear rate under the sown pasture scenarios (i.e. BSH, BSS, and BS D ), but notable
101 differences began to occur after 25 years (Fig ure 5 2 ). H owever, the pattern wa s contrasting under the native rangeland condition where the initial baseline equilibrium was maintained under the BBB s cenario, while the BSB scenario resulted in gradual decrease of soil C (at a rate of 0.19 Mg ha 1 yr 1 ) with tendency of re equilibrating soil C to the baseline level if the simulation period were extended Such decrease in soil C after change in land use and vegetation type is attributable to the changes in modeled biochemical processes as it relate s to re initiation of ecological interactions and vegetation dynamics (Houghton, 1991; Davidson and Ackerman, 1993b) I t is plausib le that soil C will attain equilibrium slowly (further beyond the simulation period) due to the aforementioned high C:N ratio of the input crop parameter and the absence of tillage practice (Li et al., 1994) These factors, amidst other possible factors, constrain litter decomposition and delay C turnover in the model simulation (Li et al., 1992; Li et al., 1997) For instance, in explaining the excha nge of C and N at the soil plant interface in DNDC, Li et al. (1994) stated that post harvest litter are assumed to stand inert in the field until the next tillage moves them into the soil, and that decompositi on rates are substantially reduced with the exclusion o f tillage under sandy and sandy loam soil conditions (a feature of our study site). As earlier indicated, s oil C stocks unde r the sown pasture scenarios were apparently similar until around year 25 ( i .e. year 265 in overall simulation) generally reached a maximum around year 6 0 ( i.e. year 300 in overall simulation) and tended to equilibrate at different soil C levels by the end of the 7 5 years prediction period (i.e year 315 in the overall simulatio n period). At the end of the simulation, soil C was lower in the BSD scenario ( 97 Mg C ha 1 yr 1 ) in comparison to the half intensity scenario which contained the highest soil C (117 Mg C ha 1 ) ( Figure 5 2 ). Although it is expected that new soil C equilibrium should be attained within ~20 years of management change (Houghton, 1991; McLauchlan et al., 2006) soil C stocks did not
102 exhibit te ndency to equilibrate under any of the scenarios until ~50 years after change in management intensity This indicates that a time frame of > 50 years is required for equi libration of soil C dynamics under this subtropical grassland ecosystem similar to oth er agroecological ecosystems (Jenkinson et al., 1991; Li et al., 1994; Li et al., 1997) Moreover, the scenario result suggests that soil C in sown pasture ecosystem (in Chapter 2) is yet to attain the upper limit or equilibrium under this climatic condition, and further increases in soil C can be expected within the next 2 3 decades. Consistent with soil C stocks, R S under the sown pasture scenarios re equilibrated after about 60 years under the scenario simulation s but in contrast to soil C stock, the trend and magnitude of R S was similar across all these scenarios ( Fig ure 5 2 ) In the BSB scenario R S declined for about 1 5 years re equilibrated to th e same as level as BBB scenario, and maintained this equilibrium for the rest of the simulation period (Fig ure 5 2 ). In s pite of the poor performance of DNDC in accurately simulating R H (a direct measure of the rate of soil C loss), it is interesting to n ote that most of the readjustment in soil C loss under the BSB scenario occurred within the first 20 years. This is in concord with findings across a range of grassland ecosystems which indicate that most soil C loss occurs within the first 20 years of cha nge in land use management (Houghton, 1991; Davidson and Ackerman, 1993b) Generally, the equilibrium R H increased with managem ent intensity from 0.16 kg CO 2 m 2 yr 1 under the BBB and BSB scenario to 1.4 kg CO 2 m 2 yr 1 under the BSD (most intensively managed) scenario Several factors such as plant photosynthetic capability, nutrient availability, a nd grazing intensity, influence the dynamics of soil C in grasslands (Conant et al., 2001; Schuman et al., 2001; McSherry and Ritchie, 2013) Although, planting of improved grass species, adoption of proper grazing practices, and improving the nutrient status have been reported to increase soil C,
103 these practices also favor increased turnover and loss of soil C (Cambardella and Elliott, 1 992; Hibbard et al., 2001; Dubeux et al., 2006b; Dawson and Smith, 2007) Numerous researchers have indicated that intensification of grassland management promotes the accumulation of labile soil C fraction (Cambardella and Elliott, 1992; Franzluebbers and Stuedemann, 2002; Dubeux et al., 2006a; Silveira et al., 2013) which can be readily decomposed and turned over to the atmospher e by soil microbes (Martens, 2000; Jastrow et al., 2007; Filley et al., 2008) Hence, model simulation of higher R H in the sown pasture scenarios, compared to th e native rangeland scenarios are related to increased availability of labile and readily decomposable litter input. Also, in the BSB scenario, the reduction of plant yield from perennial grass (~9.3 Mg C ha 1 yr 1 ) to shrub + annual grass (2.4 Mg C ha 1 y r 1 ) will reduce supply of C substrate to soil heterotrophs and subsequently contribute to the decline in R S (Knapp et al., 1998; Johnson a nd Matchett, 2001; Adachi et al., 2006; Gavrichkova, 2008) Notwithstanding this reduction of soil C loss, the soil C level did not improve over the sown pasture scenarios, as shown by the simulated soil C stocks (Figure 5 2 ). Therefore it is reasonabl e to infer that soil C input under intensified management scenarios sufficiently compensates for the elevated soil C loss es, with marked improvement in soil C stocks compared to what would be attainable under native rangeland condition. These findings are well supported by field observations of soil C changes following management intensification in this study site (presented in Chapter 2), in subtropical savanna of Texas (Filley et al., 2008) and across grasslands in diverse ecoregions (Conant et al., 2001) Summary The performance of DNDC process based model for assessing impacts of management intensification on soil C dynamics in subtropical grass land ecosystems was evaluated, and the model was applied to predict soil C dynamics under alternative management scenarios Based on the performance metrics, the model performed poorly in simulating observed soil respiration and
104 showed deficits in abiotic soil climate variables under native rangelands and management intensification (to sown pastures) condition. Despite the poor performance in simulation of autotrophic and heterotrophic soil respiration component s and critical control variables (soil tempera ture and moisture), the model yielded close prediction of R S and soil C stocks. Given the ability of DNDC to simulate the variability of these variables (except R H ) under the intensified management condition, DNDC appears promising for comparative studies such as assessment of change in soil C dynamics under alternative intensive management scenarios. However, further adjustment of the model input parameters such as soil properties and vegetation phenology, may be necessary to minimize systemic poor esti mation of soil respiration com ponents and control variables, and improve model precision in simulati n g field observ ed C variables under subtropical grasslands. There were limitations in applying DNDC for modeling soil C dynamics under native rangeland, and these include the absence of fire management parameters, the non representation of the multi specie s vegetation composition, and the crashing of the model when crops are flagged as perennial. These factors, especially fire, play a critical regulatory role in ecosystem and soil C dynamics (Hobbs et al., 1991; Hibbard et al., 2001; Johnson and Matchett, 2001) and they are prominent characteristics of th e native rangeland ecosystems in the study region. Future application of DNDC in this ecosystem may require further parameterization to incorporate fire and multi specie s effects on soil C dynamics. Based on the appl ication of DNDC model to simulate soil C dynamics under alternative management scenarios, it is notewo rthy that model outputs support the finding in field based studies which indicates that optimal management intensification is beneficial for increasing the C sink capacity of grasslands (Mazancourt et al., 1998; Conant et al., 2001; Soussana et al., 2004; Dubeux et al., 2006a; Silve ira et al., 2013) A closer look at the three sown pasture
105 s cenarios suggest that the half intensity (BSH) and conventional intensity (BSS ) scenarios may offer more long term benefit for soil C sequestration in terms of minimizing losses and optimizing C stocks, compared to the double intensity (BSD) scenario. Based on the forecast that the need to maximize pasture productivity per unit land area will become more persistent in the 21 st century (FAO, 1993; White et al., 2000; O'Mara, 2012) the adoption of sustainable management practices that are compatible with long term soil C sequestration is imperative for enhancing C sink capacity of grassland ecosystems (Conant et al., 2001; Sobecki et al., 2001; Franzluebbers, 2010) Therefore, it will likely be more attractive for grassland managers to adopt conventional intensity management approa ch which can foster improved mitigation of rising atmospheric CO 2 improved food production capacity to feed growing human populace, and potentially provide additional incentives for grassland managers through participation in carbon credit markets.
106 Table 5 1 Main DNDC input data requirement and sources from which they were derived Category Items (unit) Source (Reference) Climate Maximum and minimum temperature (C), precipitation (mm), solar radiation (MJ m 2 d 1 ), wind speed (m s 1 ), and relative humidity (%), CO 2 Concentration and annual rate of increase Florida Automated Weather Network ( http://fawn.ifas.ufl.edu/ ) Soil properties Bulk density (g cm 3 ), clay fraction (kg kg 1 ), pH, Initial NO 3 and NH 4 (mg N kg 1 ) Adewopo et al. ( 2014 ) Haile et al., ( 2008 ) Nair et al., ( 2007 ) and Kalmbacher et al., ( 1993 ) Farming m anagement Type and rate of fertilizer application (kg N ha 1 yr 1 ), stocking rate (head ha 1 ), grazing frequency, supplementation rate, crop type and crop parameters Adewopo et al., ( 2014 ) Macoon et al., ( 2011 ) Gholz et al., ( 1999 ) Pitman ( 1993 ) Kalmbacher et al., ( 1984 ) and Duever, ( 2010 )
107 Table 5 2 Soil related input parameters adopted for DNDC simulation Parameter Unit Value Reference Land use Moist grassland/Pasture Kalmbacher et al., 1984 Texture Sand Kalmbacher et al., 1984 Bulk density g/cm 3 1.1 Measured data Soil pH 4.5 Kalmbacher et al., 1993 Field capacity % 0.3 Measured Wilting point % 0.1 Measured Clay fraction 0.03 Adapted from DNDC Hydro conductivity m/hr 0.63 Adapted from DNDC Porosity 0.4 Adapted from DNDC Depth of water retention layer m 1 Kalmbacher et al., 1984 Drainage efficiency 1 Adapted from DNDC Initial soil organic carbon (SOC) at surface soil kg C/kg soil 0.03 Kalmbacher et al., 1993 Depth of soil with uniform SOC content m 0.08 Field obser ved SOC decrease rate below top soil 1.4 Measured data Resistant litter partition 0.4 Measured data Humad partition 0.2 Measured data Humus partition 0.4 Measured data Resistant litter C/N 100 Adapted from DNDC Humad C/N 10 Adapted from DNDC Humus C/N 10 Adapted from DNDC Modification factor for SOC pool decomposition 1 Adapted from DNDC Initial N concentration at surface soil (nitrate) mg N/kg 5 Adapted from DNDC Initial N concentration at surface soil (ammonium) mg N/kg 8 Adapted from DNDC Microbial activity index 1 Adapted from DNDC Slope degree 5 Kalmbacher et al., 1984 Soil salinity index 0 Adapted from DNDC Rain water collection index 1 Adapted from DNDC
108 Table 5 3 Definition of DNDC modeled management intensity scenarios and corresponding management practices Modeled s cenario Definition Management practices BBB Baseline unconverted native rangelands remaining as native rangeland Shrubs dominated, winter graz ing (15 days yr 1 ; 1 head ha 1 ), no fertilization BSB Restoration of intensively managed sown pasture to the initial native rangelands (baseline) condition Shrubs dom inated, winter grazing (15 days yr 1 ; 1 head ha 1 ), no fertilization BSH Reducing the intensity of current industry practiced sown pasture management to minimize ecological impacts Perennial grass sown, week ly rotational grazing (7 months yr 1 ; 1 head ha 1 ), 34 kg N ha 1 yr 1 of NH 4 NO 3 BSS Maintaining the intensity of current industry practiced sown pasture management to optimize grazing land productivity Perennial grass sown, week ly rotational grazing (7 months yr 1 ; 2 head ha 1 ), 67 kg N ha 1 yr 1 of NH 4 NO 3 BSD Doubling the intensity of current industry practiced sown pasture management to maximize grazing land productivity Perennial grass sown, week ly rotational grazing (7 months yr 1 ; 4 head ha 1 ), 134 kg N ha 1 yr 1 of NH 4 NO 3
109 Table 5 4 Descriptive statistics of DNDC simulated and field observed soil respiration and abiotic control factors in subtropical native rangeland (baseline) and sown pasture (intensified management). Statistics DNDC Simulated Field Observed R S R H R A S Temp S Moist R S R H R A S Temp S Moist (g CO 2 m 2 hr 1 ) (C) (%) (g CO 2 m 2 hr 1 ) (C) (%) Native rangeland management Mean 0.34 0.02 0.32 25.0 12 0 0.47 0.24 0.23 20.3 27.5 SE 0.02 0.01 0.02 0.76 2.61 0.04 0.03 0.02 1.0 4.99 SD 0.09 0.0 2 0.08 3.72 1 2.8 0.20 0.12 0.15 4.90 24.4 Min 0.12 0.00 0.12 18.5 0.59 0.15 0.07 0.05 11.1 9.45 Max 0.45 0.06 0.41 29.8 45.6 0.76 0.50 0.5 26.9 80.5 S o wn pasture management Mean 0.82 0.18 0.6 4 25.0 18.1 0.85 0.32 0.53 21.0 26.2 SE 0.06 0.02 0.06 0.73 2.24 0.10 0.03 0.07 0.98 3.78 SD 0.29 0.09 0.2 7 3.57 11.0 0.48 0.17 0.32 4.81 18.5 Min 0.31 0.0 3 0.23 19.2 2.34 0.19 0.09 0.08 13.8 9.68 Max 1 .20 0.35 1.02 29.7 39.2 1.61 0.59 1.15 26.4 60.3 R S total soil respiration, R H heterotrophic soil respiration, R A autotrophic soil respiration S Temp soil temperature, S Moist soil moisture, SE standard error, SD standard deviation, min minimum, max maximum.
110 Table 5 5 Performance of DNDC in simulatin g soil respiration and abiotic control variables under native rangeland (baseline) and sown pasture (intensified management) conditions significance, R S total soil respiration, R H heterotrophic soil respiration, R A autotrophic soil respiration. S Temp soil temperature, S Moist soil moisture, R 2 coefficient of determination, RMSE root mean square error, RPD ratio of prediction to deviation, EF modeling efficiency. For a perfect model performance, R 2 = 1, RMSE = 0, RPD > 2.0, EF = 1. Metrics Variables R S R H R A S Temp S Moist Native rangeland management R 2 0.74 0.42 0.42 0.9 5 0.27* RMSE 0.18 0.25 0.23 4.9 2 5 .7 RPD 1 1 0.50 0.91 1.0 0.95 EF 0.13 2.11 0.26 0.05 0.15 Sown pasture management R 2 0.75 0.07 ns 0.67 0.97* 0.66* RMSE 0.26 0.21 0.21 4.22 13.9 RPD 1.81 0.78 1.52 1.14 1.33 EF 0.68 0.73 0.55 0.20 0.41
111 Figur e 5 1. The structure of Denitrification and Decomposition (DNDC) model showing two component parts. The first component, consisting of the soil climate, plant growth, and decomposition submodels, predicts effects of climate, soil physical properties, veget ation, and anthropogenic activity on soil temperature, moisture, pH, Eh, and substrate concentration profiles. The second component, consisting of the nitrification, denitrification, and fermentation submodels, predicts NO, N 2 O, CH 4 and NH 3 fluxes by simu lating impacts of soil environmental conditions on the relevant geochemical and biochemical reactions. Source : Li (2000)
112 Baseline/Spin up Simulation Futuristic Simulation 30 year period after conversion from native rangeland to sown pasture R H Figure 5 2 DNDC simulation of soil carbon (C) dynamics during spin up run (year 1 240), 30 year period after conversion from baseline to intensively managed sown pasture (year 210 240), and under 75 year futuristic management scenarios (year 240 315). A) Sim ulated soil carbon stocks B) Simulated soil respiration. Futuristic management scenarios include i) baseline native rangeland condition remaining unconverted BBB, ii) restoration of intensively managed sown pasture to native rangeland BSB iii) reducing current intensity of sown pasture management by half BSH iv) maintaining currently adopted intensity of sown pasture management BSS, and v) doubling the intensity of sown pasture management BSD Year of conversion/continuation of baseli ne condition Year of field observation B A
113 Baseline Intensified Management Figure 5 3. Daily outputs of soil respiration ( total soil respiration R S heterotrophic soil respiration R H and autotrophic soil respiration R A ), and abiotic control factors (soil temperature S Temp and soil moisture S Moist ) after 30 years of managing subtropical native rangeland (baseline condition ) and sown pasture (intensified management condition ) as simulated by denitrification and decomposition model (DNDC) Julian Days
1 14 Baseline Intensified Management Figure 5 4 Measured and DNDC simulated soil respiration variables ( total soil respiration R S heterotrophic soil respiration R H and autotrophic soil respiration R A ), and abiotic control factors (soil temperature S Temp and soil moisture S Moist ) under subtrop ical native rangeland (baseline ) and sown pasture (intensified management). Broken line represents a transition from winter to summer.
115 CHAPTER 6 CONCLUSIONS AND SYNTHESIS Summary Management intensification of grassland ecosystems in the 21 st century has become societally acceptable due to the need for increased productivity per unit land area to support food and fiber needs of growing human population However, m anagement strategies that can enhance soil C sequestration and preserve soil resources in the long term are essential for ecological balance anthropogenic C mitigation, and ecosystem sustainability while optimizing land productivity. The dynamics of soil C in subtropical grasslands is poorly understood, and the long term understanding of the management intensification impacts on soil C is currently lacking. Addressing this research gap is not only important for improved national C accounting and accurate g lobal C assessment, but can inform management decisions towards supporting sustainable grassland and livestock production in the subtropical region. Hence, this study was conducted to assess long term (>22 years) impacts of management intensification on so il C dynamics in subtropical grassland ecosystems. The assessment of C stocks (in C hapter 2) indicated that management intensification has affected the quantity and dynamics of both ecosystem and soil C stocks. It was evident that ecosystem and soil C sequestered under native rangeland (baseline management condition) was lower in compari son to silvopastures and sown pastures, which is characterized by great er intensity of management practices such as fertilizer application, stocking rate, grazing frequency, and sowing of more productive grass species ( e.g., bahiagrass). For instance, soil C was 41 Mg ha 1 in native rangeland, 69 Mg ha 1 in silvopasture, and 62 Mg ha 1 in sown pasture. A s imilar pattern was observed for ecosystem C, however, the silvopastoral ecosystem sequestered far greater amounts of C in the above ground biomass (59 Mg C ha 1 ), compared with native
116 rangeland and sown pasture (4.2 and 2.1 Mg ha 1 respectively). Furthermore, results indicated that although the silvopasture and sown pasture ecosystem s contained comparable amount of soil C, there were major contrasts in the allocation of C into different particle size fractions. The relatively stable mineral associated C (Cmin) fraction constitutes ~60% (42 Mg ha 1 ) of soil C in silvopasture, while it constitutes ~45% (28 Mg ha 1 ) in sown pasture (within 0 30 cm soil depth ), therefore suggesting that silvopasture is more beneficial for accretion of more stable soil C fraction, compared with sown pastures. This patter was reversed for the labile particulate organic C (POC) fraction, which constitute d a notably greater propor tion of C sequestered in the sown pasture ecosystem. In C hapter 3, the contribution of introduced C 4 grass species (bahiagrass) to the particle size soil C fractions were evaluated by integration of soil C fractionation with 13 C isotopic analysis technique. The stable isotope ratios further unveil the underlying contrast in the dynamics of soil C sequestration as influenced by management intensity. Both POC and Cmin became less depleted in 13 C (less negative) with increasing man age ment intensity, but the most striking changes were observed in the 13 C values ranging from native rangelands to contribution of the introduced bahi agrass component to soil C accretion under this subtropical condition. Although the percent contribution of C 3 derived C generally decreased as management intensity increased, it was quite striking to note that under the sown pasture, C 4 derived C accounted for 76% of the stable Cmin fraction. This contradicts the hypothesis that introduction of more productive grass species results in depletion of the stable C fraction. Rather, it is likely that the sown C 4 grass promotes occlusion (and stabilization) of soil C and formation
117 of soil microaggregates as the grass derived litter becomes readily decomposable under the warm wet climatic condition. A third experiment was conducted to measure in situ measurements of soil respiration R S (and it s components) and the dominant abiotic factors affecting R S (Chapter 4) The specific objectives were to i) assess impacts of management intensification on long term rate of soil C loss (based on heterotrophic respiration R H ), and ii) determine the effec t of management intensification on sensitivity of soil respiration (and its components) to abiotic control factors (soil temperature S Temp and moisture S Moist ). During the winter and summer, management intensification (especially to sown pasture) incre ased the magnitude of R S and its components. For instance soil C loss through R H increased under the sown pasture by ~19% during winter and ~35% during summer compared with native rangeland Furthermore, greater magnitude of increase ( ~ 100%) in R S was ob served due to the elevated contribution of roots under the sown pasture, compared to native rangelands. The effect of management intensification on relationship s between R S and abiotic factors were not straightforward. The variability of R H (and R S ) explai ned by abiotic factors (including intera ction) declined in sown pasture during both seasons, but the response was seasonally contrasting in silvopasture. The temperature sensitivity quotient (Q 10 ) also declined from 1.69 in native rangeland to ~1.58 in bot h silvopasture and sown pasture, during the winter. However, during the summer, Q 10 of R H and R S increased with management intensity from 1.09 in native rangeland to 2.29 in sown pasture. Hence, turnover of soil C to the atmosphere through respiration is l ikely to be accelerated with warming temperature under more intensively managed grassland ecosystems. The last part of this research was focused on testing the applicability of process based denitrification and decomposition (DNDC) model for predicting management induced changes
118 in soil C within subtropical grasslands. The model output was evaluated against in situ field observation data o f R S R H (and autotrophic respiration R A ), S Temp and S Moist presented in Chapter 4. The DNDC simulated the seasonality of the respiration and abiotic variables under baseline and intensified management conditions, however there were inadequacies in th e simulation of R H R A S Temp and S Moist compared to field observations. Despite the acceptable performance metrics for DNDC simulation of R S after management intensification (R 2 = 0.75, RPD = 1.8, and EF = 0.68 ) the poor perfor mance metrics in simulat ion of R H R A S Temp and S Moist indicates a major need for improved c alibration and parameterization of the model. For example the model performed well in simulat ing the variability of these variables (except R H ) under the sown pasture condition, but the accuracy was impaired by season based or general systemic over and under estimations. Furthermore t he notable improvement of the performance metrics under the intensified management condition compared to native rangeland condition, suggests tha t DNDC may be more suitable for application within the context of intensively managed/ mono specie grassland ecosystem than within multi species and highly variable ecosystem such as native rangeland d disparity in distribution of soil C at different soil depth intervals relative to field observation, the model simula ted s oil C stocks (0 30cm) for the year of field observation (42 and 72 Mg C ha 1 in native rangelands and sown pasture, respectively) a nd compares fairly well with field measured values (41 and 62 Mg C ha 1 respectively). A pplying DNDC model to simulate 7 5 years R S and soil C stocks under five alternative management inte nsity scenarios showed that longer timeframe (>50 years) is required for soil C to attain full equilibrium after the initiation of new management system. Therefore, the current field observed soil C within the sown pasture may not represent the
119 equilibrium or upper soil C limit under the ecological condition. Mo deling results showed that soil C stock in the sown pasture scenarios (including half conventional and double intensity of current management practice) equilibrated at higher level ( 96 116 Mg C ha 1 ) compared to the restoration scenario ( 57 Mg C ha 1 ) which tended to re equilibrate to the baseline unconverted native rangeland condition. Although the simulated soil C stocks ( and R S ) we re similar between sown pasture scenarios during the initial ~20years of the 75 year futuristic simulation period s oil C stocks later declined with management intensity At the end of the simulation period, equilibrium soil C stocks were lower in sown pasture scenarios representing conventional and double intensity management systems compared to scenario representing reduction of management intensity by half. However, R S was comparable across all the sown pasture scenarios This field scale research under a unique setting of varied grassland management intensity gradient, without the influence of potential confounding factors (such as elevation, climate, and soil type) offers a valuable means to assess the long term impact of management intensification on soil C dynamics within a subtropical ecoregion. The analysis of soil C stocks suggests that management intensificat ion is beneficial for improved C sequestration in this biome, while isotopic analysis of soil C sources reveals that introduced sown grass species does not only contribute to the increase in the labile C pool, but also contribute towards increasing the rel atively stable C pool. Data also suggested that there are major differences in the stability and allocation of soil C at different depth intervals among the grassland management systems. Further evidence of changes in soil C dynamics with management intensification was revealed by changes in magnitude of soil C loss ( through heterotrophic respiration), which increased in sown pasture, possibly due to increase in productivity and accretion of labile soil C
120 fraction. Also, the changes in sensitivity of respiration variables to abiotic control factors constitutes an important implication for assessing the impacts of current and future global warming trends. For instance, the increase of Q 10 va lues with management intensification during the summer suggests the potential for a faster turnover and release soil C into atmospheric C pool (in form of CO 2 ) with increase in global temperatures, especially during the summer (Luo et al., 2001; Smith and Johnson, 2004; Luo and Zhou, 2006; IPCC, 2007; Bloom, 2010) Assessment of the predictive performance of DNDC model in this study is crucial f or assessing potential long term impacts of alternative management practices/systems on soil C dynamics under this warm wet subtropical biome. The poor performance of DNDC in simulating observed soil C and abiotic control variables indicates that further r ecalibration and/or parameterization is crucial to fully develop the model into a tool for understanding how diverse management practices and varying intensities of these practices will enhance or compromise the C sequestration potential of subtropical gra ssland biome. More so, i mprovement of DNDC through recalibration (and further parameterization) is a necessity to enhance the versatility across diverse ecoregions. Given, the improved performance of DNDC under the sown pasture condition, opportuni ty exists for model improvement through field based studies of abiotic processes and vegetation dynamics within the unique long term ecological and experimental setting for this study. Such improvement of process based understanding and modeling of soil C dynamics is an indispensable step towards improving accuracy of regional and national C budgets, and for assessing the impacts of human induced ecological changes on terrestrial C cycle.
121 LIST OF REFERENCES Abdalla, M., Kumar, S., Jones, M., Burke, J., Williams, M., 2011. Testing DNDC model for simulating soil respiration and assessing the effects of climate change on the CO2 gas flux from irish agriculture. Global and Planetary Change 78, 106 115. Abdalla, M., Wattenbach, M., Smith, P., Ambus, P., Jones, M., Williams, M., 2009. Application of the DNDC model to predict emissions of N 2 O from I rish agriculture. Geoderma 151, 327 337. Adachi, M., Bekku, Y.S., Rashidah, W., Okuda, T., Koizumi, H., 2006. Differen ces in soil respiration between different tropical ecosystems. Applied Soil Ecology 34, 258 265. Adewopo, B.J., Silveira, M.L., Xu, S., Gerber, S., Sollenberger, L.E., Martin, T.A. 2014. Management intensification impacts on Soil and Ecosystem Carbon in S ubtropical Grasslands. Soil Science Society of America Journal. doi:10.2136/sssaj2013.12.0523 Allen, D.E., Pringle, M.J., Page, K.L., Dalal, R.C., 2010. A review of sampling designs for the measurement of soil organic carbon in australian grazing lands. Th e Rangeland Journal 32, 227 246. Anderson, O.R., 2013. Soil Respiration, Climate Change, and the Role of Microbial Communities. In: Anonymous Molecular Microbial Ecology of the Rhizosphere. John Wiley & Sons, Inc., pp. 1055 1061. Archer, S., TW Boutton, KA Hibbard, 2001. Trees in Grasslands: Biogeochemical Consequences of Woody Plant Expansion. In: ED Schulze, SP Harrison, M Heimann, EA Holl, J Lloyd, IC Prentice, D Schimel (Eds.), Global Biogeochemical Cycles in the Climate System. Academic Press, San D iego, pp. 115 137. Arevalo, C.B.M., Bhatti, J.S., Chang, S.X., Sidders, D., 2009. Ecosystem carbon stocks and distributio n under different land uses in North c entral Alberta, C anada. Forest Ecology and Management 257, 1776 1785. Atkin, O.K., Edwards, E.J. Loveys, B.R., 2000. Response of root respiration to changes in temperature and its relevance to global warming. New Phytologist 147, 141 154. Baldock, J.A., Skjemstad, J.O., 2000. Role of the soil matrix and minerals in protecting natural organic materi als against biological attack. Organic Geochemistry 31, 697 710. Barry, S., 1998. Monitoring Vegetation Cover. In: Anonymous Animal Science Factsheet. University of California, Davis Cooperative Extension, CA, pp. 1 9. Baskerville, G.L., 1972. Use of loga rithmic regression in the estimation of plant biomass. Canadian Journal of Forest Resources 2, 49 53.
122 Batjes, N.H., 1999. Management options for reducing CO 2 concentrations in the atmosphere by increasing carbon sequestration in the soil. Report No 410 2 00 031, Dutch National Research Programme on Global Air Pollution and Climate Change Technical Paper 30, 1 114. Batjes, N.H., 1996. Total carbon and nitrogen in the soils of the world. European Journal of Soil Science 47, 151 163. Batjes, N.H., Sombroek, W.G., 1997. Possibilities for carbon sequestration in tropical and subtropical soils. Global Change Biology 3, 161 173. Beare, M.H., Hendrix, P.F., Cabrera, M.L., Coleman, D.C., 1994. Aggregate protected and unprotected organic matter pools in convention al and no tillage soils. Soil Science Society of America Journal 58, 787 795. Beheydt, D., Boeckx, P., Sleutel, S., Li, C., Van Cleemput, O., 2007. Validation of DNDC for 22 long term N 2 O field emission measurements. Atmospheric Environment 41, 6196 6211 Besar, N.A., Knoblauch, C., Pfeiffer, E., 2011. Carbon sequestration in coastal soils under different land use in schleswig holstein, northern germany. Environment and Natural Resources Research 1, 53 62. Biedenbender, S.H., Mcclaran, M.P., Quade, J., Weltz, M.A., 2004. Landscape patterns of vegetation change indicated by soil carbon isotope composition. Geoderma 119, 69 83. Billings, S.A., Brewer, C.M., Foster, B.L., 2006. Incorporation of plant residues into soil organic matter fractions with grassland management practices in the north american midwest. Ecosystems 9, 805 815. Biondini, M.E., Patton, B.D., Nyren, P.E., 1998. Grazing intensity and ecosystem processes in a northern mixed grass prairie, USA. Ecological Applications 8, 469 479. Blanco Canqui, H., Lal, R., 2004. Mechanisms of carbon sequestration in soil aggregates. Critical Reviews in Plant Sciences 23, 481 504. Bloom, A.J., 2010. Global Climate Change: Convergence of Disciplines. Sinauer As sociates Inc., Sunderland, MA. 398 p. organic carbon and their use in documenting vegetation change in a subtropical savanna ecosystem. Geoderma 82, 5 41. Brady, N.C., Weil, R.R., 2008. The Nature and Properties of Soil. Prentice Hall Technology Resources, Upper Saddle River, NJ., 975 pp. Bruce, J.P., Frome, M., Haites, E., Janzen, H., Lal, R., Paustian, K., 1999. Carbon sequestration in soils. Journal of Soil and Water Conservation 54, 382 389.
123 Buyanovski, G.A., Wagner, G.H., 1995. Soil Respiration and Carbon Dynamics in Parallel Native and Cultivated Ecosystems. In: Lal, R., Kimble, J., Levine, E., Stewart, B.A. (Eds.), Soils and Global Change. Lewis Publishers, Boca Raton, Fl., pp. 209 217. Buyanovsky, G.A., Wagner, G.H., 1995. Soil Respiration and Carbon Dynamics in Parallel Native and Cultivated Ecosystems. In: Lal, R., Kimble, J., Levine, E., Stewart, B.A. (Eds.), Advances in Soil Science: Soils and Global Cha nge. CRC Press Inc., Boca Raton, Fl., pp. 209 217. and nitrogen mineralization potential of particulate light soil organic matter after rainforest clearing. So il Biology and Biochemistry 28, 1555 1567. Cambardella, C.A., Elliott, E.T., 1992. Particulate soil organic matter changes across a grassland cultivation sequence. Soil Science Society of America Journal 56, 777 783. Carlisle, E.A., Steenwerth, K.L., Sma rt, D.R., 2006. Effects of land use on soil respiration. Journal of environmental quality 35, 1396 1404. Chan, K., McCoy, D., 2010. Soil carbon storage potential under perennial pastures in the mid north coast of new south wales, australia. Tropical Grass lands 44, 184 191. Chan, K.Y., Oates, A., Li, G.D., Conyers, M.K., Prangnell, R.J., Poile, G., Liu, D.L., Barchia, I.M., 2010. Soil carbon stocks under different pastures and pasture management in the higher rainfall areas of south eastern australia. Aust ralian Journal of Soil Research 48, 7 15. Chang, C., Laird, D., 2002. Near infrared reflectance spectroscopic analysis of soil C and N. Soil Science 167, 110 116. Christensen, B., 1996. Matching Measurable Soil Organic Matter Fractions with Conceptual Po ols in Simulation Models of Carbon Turnover: Revision of Model Structure. In: Powlson, D., Smith, P., Smith, J. (Eds.), Springer Berlin Heidelberg, pp. 143 159. Clapp, R.B., Hornberger, G.M., 1978. Empirical equations for some soil hydraulic properties. Water Resources Research 14, 601 604. Conant, R.T., Paustian, K., Elliott, E.T., 2001. Grassland management and conversion into grassland: Effects on soil carbon. Ecological Applications 11, 343 355. Conant, R., Six, J., Paustian, K., 2004. Land use effe cts on soil carbon fractions in the southeastern unit ed states II. C hanges in soil carbon fractions along a forest to pasture chronosequence. Biology and Fertility of Soils 40, 194 200. Conant, R., Klopatek, J., Klopatek, C., 2000. Environmental factors c ontrolling soil respiration in three semiarid ecosystems. Soil Science Society of America Journal 64, 383 390.
124 Coplen, T.B., 1994. Reporting of stable carbon, hydrogen, and oxygen isotopic abundances. Pure and Applied Chemistry 66, 273 276. Corre, M.D., Schnabel, R.R., Shaffer, J.A., 1999. Evaluation of soil organic carbon under forests, cool season and warm season grasses in the northeastern US. Soil Biology and Biochemistry 31, 1531 1539. Dai, Z., Trettin, C., Li, C., Li, H., Sun, G., Amatya, D., 2012. Effect of assessment scale on spatial and temporal variations in CH4, CO2, and N2O fluxes in a forested wetland. Water, Air, & Soil Pollution 223, 253 265. Dalal, R.C., Carter, J.O., 1999. Soil Organic Matter Dynamics and Carbon Sequestration in Australi an Tropical Soils. In: Lal, R., Kimble, J.M., Stewart, B.A. (Eds.), Global Climate Change and Tropical Ecosystems. CRC Press, LLC, Boca Raton, FL., pp. 285 314. Davidson, E.A., Ackerman, I.L., 1993a. Changes in soil carbon inventories following cultivation of previously untilled soils. Biogeochemistry 20, 161 193. Davidson, E.A., Ackerman, I.L., 1993b. Changes in soil carbon inventories following cultivation of previously untilled soils. Biogeochemistry 20, 161 193. Dawson, J.J.C., Smith, P., 2007. Carbon losses from soil and its consequences for land use management. Science of The Total Environment 382, 165 190. De Groot, P., 1990. Are we missing the grass for the trees? New Scientist 125, 29 30. Derner, J.D., Jin, V.L., 2012. Chapter 6 Soil Carbon Dynamics and Rangeland Management. In: Anonymous Managing Agricultural Greenhouse Gases. Academic Press, San Diego, pp. 79 92. Desjardins, T., Barros, E., Sarrazin, M., Girardin, C., Mariotti, A., 2004. Effects of forest conversion to pasture on soil carbo n content and dynamics in brazilian amazonia. Agriculture, Ecosystems & Environment 103, 365 373. Diels, J., Vanlauwe, B., Sanginga, N., Coolen, E., Merckx, R., 2001. Temporal variations in 13 C technique in long term soil organic matter studies. Soil Biology and Biochemistry 33, 1245 1251. Dietiker, D., Buchmann, N., Eugster, W., 2010. Testing the ability of the DNDC model to predict CO2 and water vapour fluxes of a swiss cropland site. Agriculture, Ecosyst ems & Environment 139, 396 401. Dubeux, J.C.B., Sollenberger, L.E., Comerford, N.B., Scholberg, J.M., Ruggieri, A.C., Vendramini, J.M.B., Interrante, S.M., Portier, K.M., 2006a. Management intensity affects density fractions of soil organic matter from gr azed bahiagrass swards. Soil Biology and Biochemistry 38, 2705 2711.
125 Dubeux, J.C.B., Sollenberger, L.E., Interrante, S.M., Vendramini, J.M.B., Stewart, R.L., 2006b. Litter decomposition and mineralization in bahia grass pastures managed at different inten sities. Crop Science 46, 1305 1310. Dubeux, J.C.B., Sollenberger, L.E., Mathews, B.W., Scholberg, J.M., Santos, H.Q., 2007. Nutrient cycling in warm climate grasslands. Crop Science 47, 915 928. Duever, L.C., 2011. Ecology and management of saw palmetto. Report submitted to florida fisha nd wildlife conservation commission. 1 81. Ebrahimi, A., Bossuyt, B., Hoffmann, M., 2008. Effects of species aggregation, habitat and season on the accuracy of double sampling to measure herbage mass in a lowland grassla nd ecosystem. Grass and Forage Science 63, 79 85. Ehleringer, J.R., Buchmann, R., Flanagan, L.B., 2000. Carbon isotope ratios in belowground carbon cycle process. Ecological Applications 10, 412 422. Eswaran, H., van Den Berg, E., Reich, P., 1993. Organi c carbon in soils of the world. Soil Science Society of America Journal 57, 192 194. Evrendilek, F., Celik, I., Kilic, S., 2004. Changes in soil organic carbon and other physical soil properties along adjacent mediterranean forest, grassland, and cropland ecosystems in turkey. Journal of Arid Environments 59, 743 752. Fang, C., Moncrieff, J., Gholz, H., Clark, K., 1998. Soil CO 2 efflux and its spatial variation in a florida slash pine plantation. Plant and Soil 205, 135 146. FAO, 1993. Key aspects of str ategies for the sustainable development of drylands. T0752, 1 80. Farquhar, G.D., O'Leary, M.H., Berry, J.A., 1982. On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Australian Journal of Plant Physiology 9, 121 137. Feller, C., Beare, M.H., 1997. Physical control of soil organic matter dynamics in the tropics. Geoderma 79, 69 116. Fernndez, P.L., lvarez, C.R., Schindler, V., Taboada, M.A., 2010. Changes in topsoil bulk density after grazing crop residues under no till farming. Geoderma 159, 24 30. Field, C.B., Ball, J.T., Berry, J.A., 1989. Photosynthesis, Principles and Field Techniques. In: Pearcy, R.W., Ehleringer, J., Mooney, H.A., Rundel, P.W. (Eds.), Plant Physiolological Ecol ogy, Field Methods and Instrumentation. Chapman and Hall, New York, pp. 209 253. Filley, T.R., Boutton, T.W., Liao, J.D., Jastrow, J.D., Gamblin, D.E., 2008. Chemical changes to nonaggregated particulate soil organic matter following grassland to woodland transition in a subtropical savanna. Journal of Geophysical Research: Biogeosciences 113, G3 G03009.
126 Fisher, J., Jayachandran, K., 1999. Root structure and arbuscular mycorrhizal colonization of the palm serenoa repens under field conditions. Plant and So il 217, 229 241. Flanagan, L.B., Johnson, B.G., 2005. Interacting effects of temperature, soil moisture and plant biomass production on ecosystem respiration in a northern temperate grassland. Agricultural and Forest Meteorology 130, 237 253. Follet, R.F ., Kimble, J.M., Lal, R., 2001. The Potential of U.S. Grazing Lands to Sequester Soil Carbon. In: Follet, R.F., Kimble, J.M., Lal, R. (Eds.), The Potential of US Grazing Lands to Sequester Carbon and Mitigate the Greenhouse Effect. Lewis Publishers, Boca Raton, Fl, pp. 420 449. Follett, R.F., Kimble, J.M., Lal, R., 2001. The Potential of U.S. Grazing Lands to Sequester Soil Carbon. In: Follet, R.F., Kimble, J.M., Lal, R. (Eds.), The Potential of US Grazing Lands to Sequester Carbon and Mitigate the Greenh ouse Effect. Lewis Publishers, Boca Raton,FL., pp. 420 449. Follett, R.F., Varvel, G.E., Kimble, J.M., Vogel, K.P., 2009. No till corn after bromegrass: Effect on soil carbon and soil aggregates. Agronomy Journal 101, 261 268. Franzluebbers, A.J., 2010. S oil organic carbon in managed pastures of the southeastern united states of america. Integrated Crop Management 11, 163 175. Franzluebbers, A., Stuedemann, J., 2002. Particulate and non particulate fractions of soil organic carbon under pastures in the so uthern piedmont USA. Environmental Pollution 116, S53 S62. Franzluebbers, A.J., 2005. Soil organic carbon sequestration and agricultural greenhouse gas emissions in the southeastern USA. Soil and Tillage Research 83, 120 147. Fu, W., Huang, M.S., Horton, R., 2010. Soil CO 2 efflux of two shrubs in response to plant density in northern loess plateau of china. African Journal of Biotechnology 9, 6916 6926. Fynn, A.J., Alvarez, P., Brown, J.R., George, M.R., Kustin, C., Laca, E.A., Oldfield, J.T., Schohr, T. Neely, C.L., Wong, C.P., 2009 Soil carbon sequestration in U.S. rangelands. issues paper for protocol development. 57 p Gates, D.M ., 1993. Climate Change and its Biological Consequences. Sinauer, Sunderland, MA. Gavrichkova, O., Moscatelli, C., Grego, M.S., Valentini, R., 2008. Soil carbon mineralization in a mediterranean pasture : Effect of grazing and mowing management practices. 52, 285 296. Gavrichkova, O., 2008. Drivers of soil respiration of root and microbial origin in grasslands. 1 166. Ghol z, H.L., Fisher, R.F., 1982. Organic matter production and distribution i n slash pine (P inus elliottii) plantations. Ecology 63, 1827 1839.
127 Giltrap, D.L., Li, C., Saggar, S., 2010. DNDC: A process based model of greenhouse gas fluxes from agricultural soi ls. Agriculture, Ecosystems & Environment 136, 292 300. Gonzalez Benecke, C.A., Martin, T.A., Cropper Jr., W.P., Bracho, R., 2010. Forest management effects on in situ and ex situ slash pine forest carbon balance. Forest Ecology and Management 260, 795 80 5. Graham, S.L., Millard, P., Hunt, J.E., Rogers, G.N., Whitehead, D., 2012. Roots affect the response of heterotrophic soil respiration to temperature in tussock grass microcosms. Annals of botany 110, 253 258. Gregorich, E.G., Monreal, C.M., Carter, M. R., Angers, D.A., Ellert, B.H., 1994. Towards a minimum data set to assess soil organic matter quality in agricultural soils. Canadian Journal of Soil Science 74, 367 385. Guo, L.B., Gifford, R.M., 2002. Soil carbon stocks and land use change: A meta anal ysis. Global Change Biology 8, 345 360. Haile, S.G., Nair, P.K.R., Nair, V.D., 2008. Carbon storage of different soil size fractions in florida silvopastoral systems. Journal of environmental quality 37, 1789 1797. Haile, S.G., Nair, V.D., Nair, P.K.R., 2010. Contribution of trees to carbon storage in soils of silvopastoral systems in florida, USA. Global Change Biology 16, 427 438. Hall, D.O., Scurlock, J.M.O., Ojima, D.S., Parton, W.J., 2000. Grasslands and the Global Carbon Cycle: Modeling the Effects of Climate Change. In: Wigley, T.M.L., Schimel, D.S. (Eds.), The Carbon Cycle. Cambridge University Press, pp. 102 114. Hanson, P.J., Edwards, N.T., Garten, C.T., Andrews, J.A., 2000. Separating root and soil microbial contributions to soil respiration: A review of methods and observations. Biogeochemistry 48, 115 146. Hendricks, J.J., Nadelhoffer, K.J., Aber, J.D., 1993. Assessing the role of fine roots in carbon and nutrient cycling. Trends in Ecology & Evolution 8, 174 178. Hibbard, K.A., Archer, S. Schimel, D.S., Valentine, D.W., 2001. Biogeochemical changes accompanying woody plant encroachment in a subtropical savanna. Ecology 82, 1999 2011. Hillel, D., Rosenzweig, C., 2009. Carbon exchange in the terrestrial domain and the role of agriculture. CSA News 54, 4 11. Hobbs, T.N., Schimel, D.S., Owensby, C.E., Ojima, D.S., 1991. Fire and grazing in the tallgrass prairie: Contingent effects on nitrogen budgets. Ecology 72, 1374 1382.
128 Holland, E.A., Neff, J.C., Townsend, A.R., McKeown, B., 2000. Unce rtainties in the temperature sensitivity of decomposition in tropical and subtropical ecosystems: Implications for models. Global Biogeochemical Cycles 14, 1137 1151. Horwath, W.R., Pregitzer, K.S., Paul, E.A., 1994. 14C allocation in tree soil systems. T ree physiology 14, 1163 1176. Houghton, R.A., Jenkins, G.J., Ephramus, J.J., 1990. Scientific Assessment of Climate Change. IPCC, Geneva, Switzerland. Houghton, R.A., 2012. Historic Changes in Terrestrial Carbon Storage. In: Lal, R., Lorenz, K., Httl, R. F., Schneider, B.U., von Braun, J. (Eds.), Springer Netherlands, pp. 59 82. Houghton, R.A., 1991. Tropical deforestation and atmospheric carbon dioxide. Climatic Change 19, 99 118. Huang, Z., Davis, M.R., Condron, L.M., Clinton, P.W., 2011. Soil carbon po ols, plant biomarkers and mean carbon residence time after afforestation of grassland with three tree species. Soil Biology and Biochemistry 43, 1341 1349. Hurlbert, S.H., 1984. Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54, 187 211. IPCC, 2007. Climate Change 2007: The Physical Science Basis. Intergovernmental Panel on Climate Change Secretariat, Geneva, Switzerland, 21 pp. Izaurralde, R.C., 2005. Measuring and Monitoring Soil Carbon Sequestration at the Proje ct Level. In: Anonymous CRC Press, pp. 467 500. Jackson, R.B., Mooney, H.A., Schulze, E.D., 1997. A global budget for fine root biomass, sur face area, and nutrientcontents Proc. Natl Acad. Sci. 94, 7362 7366. Jackson, R.B., Canadell, J., Ehleringer, J.R., Mooney, H.A., Sala, O.E., Schulze, E.D., 1996. A global analysis of root distributions for terrestrial biomes. Oecologia 108, 389 411. Janzen, H.H., 1995. The role of long term sites in agroecological researc h: A case study. Canadian Journal of Soil Science 75, 123 133. Jastrow, J.D., Miller, R.M., Boutton, T.W., 1996. Carbon dynamics of aggregate associated organic matter estimated by carbon 13 natural abundance. Soil Science Society of America Journal 60, 8 01 807. Jastrow, J., Amonette, J., Bailey, V., 2007. Mechanisms controlling soil carbon turnover and their potential application for enhancing carbon sequestration. Climatic Change 80, 5 23.
129 Jenkins, M., Adams, M.A., 2010. Vegetation type determines hete rotrophic respiration in subalpine australian ecosystems. Global Change Biology 16, 209 219. Jenkinson, D.S., Adams, D.E., Wild, A., 1991. Model estimates of CO 2 emissions from soil in response to global warming. Nature 351, 304 306. Johnson, L.C., Matchett, J.R., 2001. Fire and grazing regulate belowground processes in tallgrass prairie. Ecology 82, 3377 3389. Jones, M.B., Donnelly, A., 2004 Carbon sequestration in temperate grassland ecosystems and the influence of management, clim ate and elevated CO2. New Phytologist 164, 423 439. Kalbitz, K., Schmerwitz, J., Schwesig, D., Matzner, E., 2003. Biodegradation of soil derived dissolved organic matter as related to its properties. Geoderma 113, 273 291. Kalmbacher, R.S., Long, K.R., J ohnson, M.K., Martin, F.G., 1984. Botanical composition of the diets of cattle grazing south florida rangelands. Journal of Range Management 37, 334 340. Kalmbacher, R.S., Martin, F.G., Rechcigl, J.E., 1993. Effect of N P K fertilization on yield and till er density of creeping bluestem. Journal of Range Management 46, 452 457. Kern, J.S., 1994. Spatial patterns of soil organic matter in the contiguous united states. Soil Science Society of America Journal 58, 200 210. Knapp, A.K., Conard, S.L., Blair, J. M., 1998. Determinants of soil CO 2 flux from a sub humid grassland: Effect of fire and fire history. Ecological Applications 8, 760 770. Kgel Knabner, I., Guggenberger, G., Kleber, M., Kandeler, E., Kalbitz, K., Scheu, S., Eusterhues, K., Leinweber, P., 2008. Organo mineral associations in temperate soils: Integrating biology, mineralogy, and organic matter chemistry. Journal of Plant Nutrition and Soil Science 171, 61 82. Krbel, R., Sun, Q., Ingwersen, J., Chen, X., Zhang, F., Mller, T., Rmheld, V., 2010. Modelling water dynamics with DNDC and DAISY in a soil of the north china plain: A comparative study. Environmental Modelling & Software 25, 583 601. Kuzyakov, Y., Gavrichkova, O., 2010. Review: Time lag between photosynthesis and carbon dioxide eff lux from soil: A review of mechanisms and controls. Global Change Biology 16, 3386 3406. Lal, R., Follett, R.F., Kimble, J.M., 2003. Achieving soil carbon sequestration in the united states: A challenge to the policy makers. Soil Science 168, 827 845. Li C.S., 2000. Modeling trace gas emissions from agricultural ecosystems. Nutrient Cycling in Agroecosystems 58, 259 276.
130 Li, C., Frolking, S., Crocker, G.J., Grace, P.R., Klr, J., Krchens, M., Poulton, P.R., 1997. Simulating trends in soil organic carbo n in long term experiments using the DNDC model. Geoderma 81, 45 60. Li, C., Fro lking, S., Frolking, T.A., 1992 A model of nitrous oxide evolution from soil driven by rainfall events: 1. model structure and sensitivity. Journal of Geophysical Research: A tmospheres 97, 9759 9776. Li, C., Frolking, S., Harriss, R., 1994. Modeling carbon biogeochemistry in agricultural soils. Global Biogeochemical Cycles 8, 237 254. Li, C., Zhuang, Y., Frolking, S., Galloway, J., Harriss, R., III, B.M., Schimel, D., Wang, X., 2003. Modeling soil organic carbon change in croplands of china. Ecological Applications 13, 327 336. Liu, K., Sollenberger, L.E., Silveira, M.L., Vendramini, J.M.B., Newman, Y.C., 2011. Distribution of nutrients among Soil Plant bermudagrass pastures grazed at different intensities. Crop Science 51, 1800 1807. Liu, N., Zhang, Y., Chang, S., Kan, H., Lin, L., 2012. Impact of grazing on soil carbon and microbial biomass in typical steppe and desert steppe of inner mongolia. PLoS ONE 7, e36434. Liu, Y., Yu, Z., Chen, J., Zhang, F., Doluschitz, R., Axmacher, J.C., 2006. Changes of soil organic carbon in an intensively cultivated agricultural region: A denitrification decomposition (DNDC) modelling approach. Science of The Total En vironment 372, 203 214. Lloyd, J., Taylor, J.A., 1994. On the temperature dependence of soil respiration. Functional Ecology 8, 315 323. Loague, K., Green, R.E., 1991. Statistical and graphical methods for evaluating solute transport models: Overview and application. Journal of contaminant hydrology 7, 51 73. Ludwig, B., Jger, N., Priesack, E., Flessa, H., 2011. Application of the DNDC model to predict N 2 O emissions from sandy arable soils with differing fertilization in a long term experiment. Journal of Plant Nutrition and Soil Science 174, 350 358. Luo, Y., Zhou, X., 2006. Soil Respiration and the Environment. Elsevier, Burlington, MA., 315 pp. Luo, Y., Wan, S., Hui, D., Wallace, L.L., 2001. Acclimatization of soil respiration to warming in a tall grass prairie. Nature 413, 622 625. Ma, S., Chen, J., Butnor, J.R., North, M., Euskirchen, E.S., Oakley, B., 2005. Biophysical controls on soil respir ation in the dominant patch types of an old growth, mixed conifer forest. Forest Science 51, 221 232.
131 Martens, D.A., 2000. Plant residue biochemistry regulates soil carbon cycling and carbon sequestration. Soil Biology and Biochemistry 32, 361 369. Marte ns, D.A., Reedy, T.E., Lewis, D.T., 2004. Soil organic carbon content and composition of 130 year crop, pasture and forest land use managements. Global Change Biology 10, 65 78. Mazancourt, C.D ., Loreau, M., Abbadie, L., 1998. Grazing optimization and nut rient cycling: When do herbivores enhance plant production? Ecology 79, 2242 2252. McCulley, R.L., Boutton, T.W., Archer, S.R., 2007. Soil respiration in a subtropical savanna parkland: Response to water additions. Soil Sci.Soc.Am.J. 71, 820 828. McLauch lan, K.K., Hobbie, S.E., Post, W.M., 2006. Conversion from agriculture to grassland builds soil organic matter on decadal timescales. Ecological Applications 16, 143 153. McSherry, M.E., Ritchie, M.E., 2013. Effects of grazing on grassland soil carbon: A global review. Global Change Biology 19, 1347 1357. Millar, R.B., Anderson, M.J., 2004. Remedies for pseudoreplication. Fisheries Research 70, 397 407. Mulkey, S., 200 7. Climate change and land use: interdependencies and opportunities. Available online a t http://www.communicationsmgrcom/projects/1349/docs/Climatechange .pdf 43 p. Accessed, 5/03/ 201 4 Nair, V.D., Nair, P.K.R., Kalmbacher, R.S., Ezenwa, I.V., 2007. Reducing nutrient loss from farms through silvopastoral practices in coarse textured soils of florida, USA. Ecological Engineering 29, 192 199. NRCS Websoil Survey, 2013. Soil survey staff, natural resources conservation service, united states department of agriculture. Natural Resource Conservation Society Web Soil Survey. Available online at http://websoilsurvey.nrcs.usda.gov/. Accessed, 02/15/2013 NRCS GRP, 2009. Grasslands reserve program: Programmatic environmental assessment. 1 92. Obour, A.K., Silveira, M.L., Adjei, M.B., Vendramini, J.M., Rechcigl, J.E., 2009. Cattle manure application strategies effects on bahiagrass yield, nutritive value, and phosphorus recovery. Agronomy Journal 101, 1099 1107. O'Mara, F.P., 2012. The role of grasslands in food security and climate ch ange. Annals of botany 110, 1263 1270. Owensby, C.E., 1998. Role of grasslands as modifiers of global climate change. 9 12.
132 Oyonarte, C., Rey, A., Raimundo, J., Miralles, I., Escribano, P., 2012. The use of soil respiration as an ecological indicator in arid ecosystems of the SE of spain: Spatial variability and controlling factors. Ecological Indicators 14, 40 49. Parr, J.F., Sullivan, L.A., 2005. Soil carbon sequestration in phytoliths. Soil Biology and Biochemistry 37, 117 124. Parton, W.J., Scurlock J.M.O., Ojima, D.S., Schimel, D.S., Hall, D.O., 1995. Impact of climate change on grassland production and soil carbon worldwide. Global change biology 1, 13 22. Pattanayak, S., McCarl, B., Sommer, A., Murray, B., Bondelid, T., Gillig, D., DeAngelo, B., 2005. Water quality co effects of greenhouse gas mitigation in U.S. agriculture. Climatic Change 71, 341 372. Paustian, K., 2007. Organic Matter and Global C Cycle. In: Lal, R. (Ed). Encyclopeadia of Soil Science. Taylor & Francis, pp. 1176 1179. Paustia n, K., Levine, E., Post, W.M., Ryzhova, I.M., 1997. The use of models to integrate information and understanding of soil C at the regional scale. Geoderma 79, 227 260. Peichl, M., Leava, N., Kiely, G., 2012. Above and belowground ecosystem biomass, carbo n and nitrogen allocation in recently afforested grassland and adjacent intensively managed grassland. Plant and Soil 350, 281 296. Penman, J., Gytarsky, M., Hiraishi, T., Kr ug, T., Kruger, D., 2003. Good p ractice guidance for l and use, l and use c hange an d f orestry. Institute for Global Environmental Strategies for the IPCC., Kanagawa,Japan., 590 pp. Pietola, L., Horn, R., Yli Halla, M., 2005. Effects of trampling by cattle on the hydraulic and mechanical properties of soil. Soil and Tillage Research 82, 9 9 108. Poorter, H., Van Berkel, Y., Baxter, R., Den Hertog, J., Dijkstra, P., Gifford, R.M., Griffin, K.L., Roumet, C., Roy, J., Wong, S.C., 1997. The effect of elevated CO 2 on the chemical composition and construction costs of leaves of 27 C 3 species. Pl ant, Cell & Environment 20, 472 482. Post, W.M., Izaurralde, R.C., Mann, L.K., Bliss, N., 2001. Monitoring and verifying changes of organic carbon in soil. Climatic Change 51, 73 99. Post, W.M., Kwon, K.C., 2000. Soil carbon sequestration and land use ch ange: Processes and potential. Global Change Biology 6, 317 327. Post, W.M., Emanuel, W.R., Zinke, P.J., Stangenberger, A.G., 1982. Soil carbon pools and world life zones. Nature 298, 156 159.
133 Powers, J.S., Corre, M.D., Twine, T.E., Veldkamp, E., 2011. Geographic bias of field observations of soil carbon stocks with tropical land use changes precludes spatial extrapolation. Proceedings of the National Academy of Sciences Pregitzer, K.S., Euskirchen, E.S., 2004. Carbon cycling and storage in world fores ts: Biome patterns related to forest age. Global Change Biology 10, 2052 2077. Pucheta, E., Bonamici, I., Cabido, M., Diaz, S., 2004. Below ground biomass and productivity of a grazed site and a neighbouring ungrazed exclosure in a grassland in central a rgentina. Austral Ecology 201 208. Qi, G., Wang, Q., Zhou, W., Ding, H., Wang, X., Qi, L., Wang, Y., Li, S., Dai, L., 2011. Moisture effect on carbon and nitrogen mineralization in topsoil of changbai mountain, northeast china. Journal of Forest Science 5 7, 340 348. Raich, J.W., Schlesinger, W.H., 1992. The global carbon dioxide flux in soil respiration and its relationship to vegetation and climate. Tellus B 44, 81 99. Raich, J., Tufekciogul, A., 2000. Vegetation and soil respiration: Correlations and c ontrols. Biogeochemistry 48, 71 90. Reeder, J.D., Franks, C.D., Milchunas, D.G., 2000. Root biomass and microbial p rocesses. In: Kimble, J.M., Lal, R., Follett, R.F. (Eds.), The Potential of U.S. g razing l ands to sequester carbon and m itigate the greenhou se e ffect. CRC Press LLC., Roberts, M., 2001. Soil Carbon Sequestration for Improved Land Management. FAO, Rome, Italy, 57 pp. Saiz, G., Green, C., Butterbach Bahl, K., Kiese, R., Avitabile, V., Farrell, E., 2006. Seasonal and spatial variability of soil respiration in four sitka spruce stands. Plant and Soil 287, 161 176. Sakurai, G., Jomura, M., Yonemura, S., Iizumi, T., Shirato, Y., Yokozawa, M., 2012. Inversely estimating temperature sensitivity of soil carbon decomposition by assimilating a turnover model and long term field data. Soil Biology and Biochemistry 46, 191 199. Santantonio, D., Hermann, R.K., Overton, W.S., 1977. Root biomass studies in forest ecosystems. Pedobiologia Band 17, 1 31. Schimel, D.S., House, J.I., Hibbard, K.A., Bousquet, P ., Ciais, P., Peylin, P., Braswell, B.H., Apps, M.J., Baker, D., Bondeau, A., Canadell, J., Churkina, G., Cramer, W., Denning, A.S., Field, C.B., Friedlingstein, P., Goodale, C., Heimann, M., Houghton, R.A., Melillo, J.M., Moore, B., Murdiyarso, D., Noble, I., Pacala, S.W., Prentice, I.C., Raupach, M.R., Rayner, P.J., Scholes, R.J., Steffen, W.L., Wirth, C., 2001. Recent patterns and mechanisms of carbon exchange by terrestrial ecosystems. Nature 414, 169 172.
134 Schimel, D.S., 1995. Terrestrial ecosystems an d the carbon cycle. Global Change Biology 1, 77 91. Schlesinger, W.H., 1977. Carbon balance in terrestrial detritus. Annual Review of Ecology and Systematics 8, 51 81. Schlesinger, W., Andrews, J., 2000. Soil respiration and the global carbon cycle. Biog eochemistry 48, 7 20. Schuman, G.E., Herrick, J.E., Janzen, H.H., 2001. The Dynamics of soil carbon in r angelands. In: Follett, R.F., Kimble, J.M., Lal, R. (Eds.), The Potential of U.S. Grazinglands to Sequester Carbon and Mitigate the Greenhouse Effect. CRC Press, Boca Raton, Fl., pp. 267 290. Schuman, G.E., Janzen, H.H., Herrick, J.E., 2002. Soil carbon dynamics and potential carbon sequestration by rangelands. Environmental Pollution 116, 391 396. Schwe ndenmann, L., Pendall, E., 2006 Effects of fores t conversion into grassland on soil aggregate structure and carbon storage in panama: Evidence from soil carbon fractionation and stable isotopes. Plant and Soil 288, 217 232. Shaffer, M., Ma, L., 2001. Carbon and Nitrogen Dynamics in Upland Soils. In: An onymous CRC Press, Sharrow, S.H., Ismail, S., 2004. Carbon and nitrogen storage in agroforests, tree plantations, and pastures in western oregon, USA. Agroforestry Systems 60, 123 130. Silveira, M.L., Comerford, N.B., Reddy, K.R., Prenger, J., DeBusk, W .F., 2010. Influence of military land uses on soil carbon dynamics in forest ecosystems of georgia, USA. Ecological Indicators 10, 905 909. Silveira, M.L., Liu, K., Sollenberger, L.E., Follett, R.F., Vendramini, J.M.B., 2013. Short term effects of grazing intensity and nitrogen fertilization on soil organic carbon pools under perennial grass pastures in the southeastern USA. Soil Biology and Biochemistry 58, 42 49. Singh, J.S., Gupta, S.R., 1977. Plant decomposition and soil respiration in terrestrial eco systems. Botanical Review 43, 449 528. Six, J., Conant, R.T., Paul, E.A., Paustian, K., 2002. Stabilization mechanisms of soil organic matter: Implications for C saturation of soils. Plant and Soil 241, 155 176. Skjernstad, J.O., Taylor, J.A., Smernik, R .J., 1999. Estimation of charcoal (char) in soils. Communications in Soil Science and Plant Analysis 30, 2283 2298. Sleutel, S., De Neve, S., Nmeth, T., Tth, T., Hofman, G., 2006. Effect of manure and fertilizer application on the distribution of organi c carbon in different soil fractions in long term field experiments. European Journal of Agronomy 25, 280 288.
135 Smith, D.L., Johnson, L., 2004. Vegetation mediated changes in microclimate reduce soil respiration as woodlands expand into grasslands. Ecology 85, 3348 3361. Smith, P., Smith, J.U., Powlson, D.S., McGill, W.B., Arah, J.R.M., Chertov, O.G., Coleman, K., Franko, U., Frolking, S., Jenkinson, D.S., Jensen, L.S., Kelly, R.H., Klein Gunnewiek, H., Komarov, A.S., Li, C., Molina, J.A.E., Mueller, T., P arton, W.J., Thornley, J.H.M., Whitmore, A.P., 1997. A comparison of the performance of nine soil organic matter models using datasets from seven long term experiments. Geoderma 81, 153 225. Smith, W.N., Grant, B.B., Desjardins, R.L., Rochette, P., Drury, C.F., Li, C., 2008. Evaluation of two process based models to estimate soil N2O emissions in eastern canada. Canadian Journal of Soil Science 88, 251 260. Sobecki, T., Moffitt, D., Stone, J., Franks, C., Mendenhall, A., 2001. A broad scale perspective on the e xtent, distribution, and characteristics of U.S. grazingl ands. In: Follet, R., Kimble, J., Lal, R. (Eds.), The Potential of US Grazing Lands to Sequester Carbon and Mitigate the Greenhouse Effect. Le wis Publishers, Boca Raton, FL., pp. 420 449. Soil Survey Staff, 1999. Soil taxonomy: A basic system of soil classification for making and interpreting soil surveys. United States Department of Agriculture Natural Resource Conservation Service Agricultural Handbook 436, 1 871. Sollins, P., Spycher, G., Glassman, C.A., 1984. Net nitrogen mineralization from light and heavy fraction forest soil organic matter. Soil Biology and Biochemistry 16, 31 37. Sollins, P., Homann, P., Caldwell, B.A., 1996. Stabiliza tion and destabilization of soil organic matter: Mechanisms and controls. Geoderma 74, 65 105. Solomon, D., Fritzsche, F., Lehmann, J., Tekalign, M., Zech, W., 2002. Soil organic matter dynamics in the subhumid agroecosystems of the ethiopian highlands. S oil Science Society of America Journal 66, 969 978. Sombroek, W.G., Nachtergaele, F.O., Hebel, A., 1993. Amounts, dynamics and sequestering of carbon in tropical and subtropical soils. Ambio 22, 417 426. Soussana, J ., Loiseau, P., Vuichard, N., Ceschia, E., Balesdent, J., Chevallier, T., Arrouays, D., 2004. Carbon cycling and sequestration opportunities in temperate grasslands. Soil Use and Management 20, 219 230. Steinbeiss, S., Temperton, V.M., Gleixner, G., 2008. Mechanisms of short term soil carbon s torage in experimental grasslands. Soil Biology and Biochemistry 40, 2634 2642. Stephenson, K.E., 2011. Distribution of grasslands in 19th century florida. The American Midland Naturalist 165, 50 59.
136 Suh, S., Lee, E., Lee, J., 2009. Temperature and moist ure sensitivities of CO 2 efflux from lowland and alpine meadow soils. Journal of Plant Ecology 2, 225 231. Sundquist, E.T., 1993. The global carbon dioxide budget. Science 259, 934 941. Svejcar, T., Angell, R., Bradford, J.A., Dugas, W., Emmerich, W., Frank, A.B., Gilmanov, T., Haferkamp, M., Johnson, D.A., Mayeux, H., Mielnick, P., Morgan, J., Saliendra, N.Z., Schuman, G.E., Sims, P.L., Snyder, K., 2008. Carbon fluxes on north american r angelands. Rangeland Ecology & Management 61, 465 474. Tang, R., Tang, X., 1989. Root system distribution in a tea plantation with interplanted slash pine. Chinese Journal of Ecology 6 10. Tang, H., Qiu, J., Van Ranst, E., Li, C., 2006. Estimations of so il organic carbon storage in cropland of china based on DNDC model. Geoderma 134, 200 206. Tate, K.W., Dudley, D.M., McDougald, N.K., George, M.R., 2004. Effect of canopy and grazing on soil bulk density. Journal of Range Management 57, 411 417. Tonitto, C., David, M., Drinkwater, L., Li, C., 2007. Application of the DNDC model to tile drained illinois agroecosystems: Model calibration, validation, and uncertainty analysis. Nutrient Cycling in Agroecosystems 78, 51 63. Trettin, C.C., 2001. Existing s oil c arbon models do n ot apply to forested w etlands. USDA Forest Service Southern Resear ch Station, Asheville, N.C., 10 p. U.S. EPA, 2012. Inventory of U.S. greenhouse gas emissions and sinks: 1990 2010. EPA 430 R 12 001, 1 481. USDA NASS, 2009. 2007 census of agriculture: United S tates summary and state data. AC 07 A 51, 1 739. von Ltzow, M., Kgel Knabner, I., Ekschmitt, K., Flessa, H., Guggenberger, G., Matzner, E., Marschner, B., 2007. SOM fractionation methods: Relevance to functional pools and to stab ilization mechanisms. Soil Biology and Biochemistry 39, 2183 2207. Vose, J.M., Elliot, K.J., Johnson, D.W., 1995. Soil CO 2 Flux in Response to Elevated Atmospheric CO 2 and Nitrogen Fertilization: Patterns and Methods. In: Lal, R., Kimble, J., Levine, E., Stewart, B.A. (Eds.), Advances in Soil Science: Soils and Global Change. CRC Press Inc., Boca Raton, Fl., pp. 199 208. Wang, S., Liu, J., Zhang, C., Yi, C., Wu, W., 2011a. Effects of afforestation on soil carbon urnal of Geographical Sciences 21, 118 134. Wang, W., Fang, J., 2009. Soil respiration and human effects on global grasslands. Global and Planetary Change 67, 20 28.
137 Wang, X., Zhu, B., Li, C., Gao, M., Wang, Y., Zhou, Z., Yuan, H., 2011b. Dissecting soil CO 2 fluxes from a subtropical forest in china by integrating field measurements with a modeling approach. Geoderma 161, 88 94. Wang, Z., Han, X., Li, L., 2008. Effects of grassland conversion to croplands on soil organic carbon in the temper ate inner mon golia. Journal of Environmental M anagement 86, 529 534. Weil, R.R., Magdoff, F., 2004. Significance of Soil Organic Matter to Soil Quality and Health. In: Magdoff, F., Weil, R.R. (Eds.), Significance of Soil Organic Matter to Soil Quality and Health. CRC Press, Boca Raton, Fl., pp. 1 43. White, R., Murray, S., Mark, R., 2000. Pilot analysis of global ecosystems: Grassland ecosystems. 1 81. Wilke, B., 2005. Determination of Chemical and Physical Soil Properties. In: Anonymous Springer Berlin Heidelberg, pp. 47 95. Wilm, H.G., 1944. Estimating forage yield by the double sampling method. American Society of Agronomy Journal. 36, 194. Xu, M., Qi, Y., 2001. Soil surface CO2 efflux and its spatial and temporal variations in a young ponderosa pine plantation in northern california. Global Change Biology 7, 667 677. Yelenik, S.G., Stock, W.D., Richardson, D.M., 2004. Ecosystem level impact s of invasive acacia saligna in the south african fynbos. Restoration Ecology 12, 44 51. Zhu, Z. (Ed.), 2010. Method for Assessing Carbon Stocks, Carbon Sequestration, and Greenhouse Gas Fluxes in Ecosystems of the United States Under Present Conditions a nd Future Scenarios. U.S. Geological Sur vey Scientific Investigations, 190 p. Zwick, P.D., Carr, M.H., 2006. Florida 2060 A population distribution scenario for the state of florida. Research project prepared f or "1000 Friends of Florida", 29 p
138 B IOGRAPHICAL SKETCH Julius Babatunde Adewopo was born in Ibadan, Nigeria where he had his primary (elementary) and part of his secondary (high) school education. His intrinsic passion for natural resource sustainability informed his decision to pursue a 5 y ear undergraduate degree in Forest resources at the Federal University of Technology, Akure (FUTA) in Nigeria where he pioneered LC FUTA) and graduated in 2006 with ho nors. He served the government of Nigeria for one year under the auspices of the National Youth Service Corps (NYSC), and later proceeded to University of Arkansas Monticello where he graduated in 2010 with M.Sc. degree in Forest Resources. Based on his st rong motivation to acquire well rounded knowledge about soil plant interactions and ecological sustainability, Julius decided to further pursue a doctorate degree in Soil science. His doctorate research and evolving research interest includes the field bas ed assessment and process based modeling of long term spatio temporal changes in ecosystem C dynamics with changes in vegetation composition and land use intensification Julius has served as expert reviewer for IPCC AR5 WGII& WGIII, invited reviewer for B ioresources journal; and grant reviewer for USDA Sustainable Agriculture Research and Education (SARE). He led youth and children delegation to the 8 th and 9 th session of UN Forum on Forests at U.N. Headquarters (in New York), and served as elected focal p oint for UNFCCC a role in which he led youth delegation and organized side event at UNFCCC COP16 in Cancun, Mexico. Recently, Julius proposed and led a nationwide initiative to identify top priority research questions for soil science in the 21 st century and organized a special session at the Adewopo and they are both parents to twin girls (Pearl and Jewel) coincidentally, both his advisor and co advisor are als o parents to twin girls.
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 EI0IH2GME_RJTF4G INGEST_TIME 2014-10-03T21:46:54Z PACKAGE UFE0046691_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC