CONTRIBUTION OF SOIL PORES TO THE PROCESSING AND PROTECTION OF SOIL CARBON AT MICRO -SCALE By Michelle Yvonne Quigley A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Crop and Soil Sciences Œ Doctor of Philosophy 2019 ABSTRACT CONTRIBUTION OF SOIL PORES TO THE PROCESSING AND PROTECTION OF SOIL CARBON AT MICRO -SCALE By Michelle Yvonne Quigley Soil carbon has the potential to increas e crop yield and mitigate climate change . As the largest terrestrial carbon stock , gains and losses of soil carbon can have a great impact on atmospheric CO 2 concentrations. Additionally, many beneficial soil properties for agricultural sustainability are tied to soil ca rbon. This makes understanding the mechanics of soil carbon vital to accurate climate change modeling and management recommendations. However, current soil carbon models, relying on bulk charac teristics, can vary widely in their results and current recommendations for improving soil carbon do not work in all circumstances. Micro -scale processes, the scale at which carbon protection occurs, are currently not well understood. Improving the underst anding of micro -scale processes would improve both climate models and management recommendations. Carbon processes at micro -scale are believed to occur in diverse microenvironments. However, it is soil pores that, through transport of gasses, water, nutrients and microorganisms, may ultimately control the formation of these microenvironments. Therefore, understanding the relationship between soil pores and carbon is potentially vital to understanding micro -scale carbon processes. To und erstand the relationship between soil pores and carbon I employed carbon. I investigated the spatial variability of soil carbon within the soil matrix of differen t soil managements and how pores of different origin contributed to this variability to explore the effect of management and pore origin on the creation of microenvironments. Then I investigated the effect of pore size distribution on carbon addition durin g growth of cereal rye ( Secale cereale L.) and usage during a subsequent incubation using natural abundance stable carbon isotopes. I investigated the role of management history on the effect of pore size distribution during new carbon addition and usage u sing e nriched stable carbon isotopes. I found managements that build carbon have higher spatial variability of grayscale values biological pores, due to their lar which would impact variability greater. The influence of biological and mechanical pores on adjacent carbon concentrations was found to be independent of management. Pores of 15-40 bon protection after incubation, match ing previously reported results, indicating a universal mechanism for carbon protection, possibly related to fungi, in these pores. From both natural abundance and enriched stable carbo n isotope studies, I found that 40 - large losses of new carbon in the subsequent incubations. I found important relationships between pore origin , pore sizes, and carbon , speci fically, that biological pores exert more influence on the carbon concentrations adjacent to them than mechanical pores. A technique to measure this influence using osmium staining of organic matter and grayscale gradients of images was developed. I found that 40 - important avenues of carbon addition, but also are associated with carbon losses. However, the reasons for these easy gains and losses is yet unclear, requiring further research , but it is believed to be associated with small plant roots . iv ACKNOWLEDGEMENTS It is with the utmost gratitude that I thank my advisor, Dr. Sasha Kravchenko, without whose support and patience this thesis would not be possible , in addition to her amazing research talent . I would also like to thank the members of my committee: Dr. Alvin Smucker for his wisdom and vast literature knowledge, Dr. Phil Robertson for his fibig picturefl view and expertise, and Dr. Dirk Colbry for his vast knowledge of anything programming and as tounding willingness to help. It has been a really pleasure working with all of you. Humbly , I also thank the Department of Plant, Soil, and Microbial Sciences for all their wonderful support and the things they do for us students every day. I would also like to thank all the lab members that have helped me on this journey: Dr. Moslem Ladoni, Dr. Ehsan Toosi, Dr. Kusay Alani, Dr. Erin Anders, Dr. Turgut Kutlu, Richard Price, Jessica Fry, Jordan Beehler, Kyungmin fiAlyssafl Kim, Maxwell Oerther, and the many visiting scholars from Pakistan and China I had the privilege to get to know. Additionally, I would like to thank the friends and acquaintances I meet during my thesis work, whether through teaching, statistical consulting, or were just on the same journe y. Lastly, but certainly not in the least, I would like to thank my family: my husband, Nathanial Quigley, for his unending support and quick wit; my children, Danius, Kai, and Cora, who ma ke everyday an adventure; and our dear friend and roommate Keiran Fallon, who puts up with us. v TABLE OF CONTENTS LIST OF TABLES ........................................................................................ .............................. vii i LIST OF FIGURES ........................................................................................................................ x KEY TO ABBREVIATIONS .............................................. ....................................................... xii i CHAPT ER 1: Introduction ............................................................................................................. 1 REFERENCES ............................................................................................................................. 8 CHAPTER 2: Patterns and Sources of Spatial Heterogeneity in Soil Matrix from Contrasting Long Term Management Practices ......................................................................................................... 16 Abstract .......................................................................................................... ............................. 16 2.1 Introduction .................................................................................................................... 17 2.2 Materials and methods ................................................................................................... 22 2.2.1 Soil Collection and Imaging .............................................................................................. 22 2.2.2 Geostatistical Analysis ...................................................................... ............................... 23 2.2.3 Os Gradients ...................................................................................... ............................... 24 2.2.4 Grayscale Gradients .......................................................................... ............................... 27 2.2.5 Analysis of pores below image resolution (2 - ........................................................ 29 2.2.6 Statistical Analysis ............................................................................................................ 30 2.3 Results ............................................................................................................................. . 31 2.3.1 Geostatistical analysis of grayscale spatial patterns ........................................................ 31 2.3.2 Os levels as a function of distance from soil pores ........................................................... 32 2.3.3 Grayscale levels as a function of distance from soil pores ............................................... 33 2.3.4 Pores below image resolution ............................................................ ............................... 36 2.4 Discussion ..................................................................................................................... ... 37 2.4.1 Image grayscale values as a proxy for SOM patterns ....................................................... 37 2.4.2 Spatial patterns of grayscale values .................................................................................. 41 2.4.3 SOM pattern in relation to soil pores ................................................. ............................... 44 2.4.4 Effect of management practices on SOM pattern in relation to soil pores ....................... 44 2.5 Conclusion ....................................................................................................................... 45 Funding ........................................................................................................................................ 46 REFERENCES ........................................................................................................................... 47 CHAPTER 3: Influence of Pore Characteristics on the Fate and Di stribution of Newly Added Carbon ...................................................................................................................... .................... 53 Abstract ........................................................................ ............................................................... 53 3.1 Introduction .................................................................................................................... 54 3.2 Materials and methods ................................................................................................... 57 3.2.1 Greenhouse experimental setup ........................................................................................ 57 3.2.2 ................................................................................... 59 3.2.3 Incubation experimental set up ......................................................................................... 60 3.2.4 Soil fragment cutting and chemical analyses .................................................................... 61 vi 3.2.5 Grayscale Gradients ......................................................................................................... 62 3.2.6 Statistical Analysis ............................................................................................................ 63 3.3 Results ............................................................................................................................. . 65 3.3.1 Soil and plant characteristics ............................................................................................ 65 3.3.2 Pore characteristics ......................................................................................... ................. 66 3.3.3 Associations between pores and chemical characteristics ................................................ 69 3.3.4 Incubation CO 2 ................................................................................................................. 71 3.3.5 Grayscale Gradients ......................................................................................................... 73 3.3.6 Canonical Correlations .................................................................................. .................. 74 3.4 Discussion ........................................................................................................................ 75 3.4.1 Relationship between C3 carbon and 40 - ...................................................... 75 3.4.2 Relationship between carbon and 6.5 - - ................... 77 3.4.3 Additional considerations ................................................................................................. 80 3.5 Conclusion ....................................................................................................................... 81 Funding ........................................................................................................................................ 81 REFERENCES ........................................................................................................................... 83 CHAPTER 4: Effect of M anagement and Pore Size Distribution on the Input and Persistence of New Carbon ............................................................................................................................. ..... 90 Abstract ....................................................................................................... ................................ 90 4.1. Introduction .................................................................................................................... 91 4.2. Materials and methods ................................................................................................... 94 4.2.1 Soil collection ................................................................................................................... 94 4.2.2 Pulse labeling ................................................................................................................... 95 4.2.3 Sample collection ............................................................................................................ .. 97 4.2.4 .................................................................................................. 99 4.2.5 Incubation experimental design ...................................................................................... 100 4.2.6 13 C analyses ............................................................................................... 101 4.2.7 Determination of POM and root presence ...................................... ................................ 101 4.2.8 Total N, nitrate, and ammonium ..................................................................................... 102 2.9 Statistical analysis .................................................................................................... ...... 102 4.3. Results ............................................................................................................................ 103 4.3.1 Soil and plant characteristics .......................................................................................... 103 4.3.2 Pore characteristics ........................................................................................................ 105 4.3.3 Associations between pores and new carbon ........................................... ....................... 106 4.3.4 Utilization of carbon during incubation .......................................................................... 109 4.4. Discussion ....................................................................................... ............................... 111 4.4.1 Carbon addition during rye growth ................................................................................ 111 4.4.2 Carbon utilization during incubation .............................................................................. 113 4.4.3 POM, roots, and nitrogen ............................................................................................... 115 4.4.4 Soil aggregates vs. intact soil cores ................................................................................ 116 4.5 Conclusion ..................................................................................................................... 117 Funding ...................................................................................................................................... 118 REFERENCES ......................................................................................................................... 119 CHAPTE R 5: Conclusion ........................................................................................................... 129 vii APPENDI X ....................................................................................................................... ......... 131 viii LIST OF TABLES Table 2.1: Characteristics of the variograms of soil material and variance of the histograms and standard errors (in parentheses) calculated based on a total of 157 subsection cubes from 32 aggregates. Different letters within each row denote statistically significant differences among the . ............................................................................................................... 32 Table 2.2: Effective distance of pore influence (EDPI) for the three studied pore types averaged across all studied aggregates. Means were ca lculated based on 32 aggregates with 3 POM -NS, 3 POM -Root, and 5 non -biological pores from each aggregate. Standard errors are shown in .......... 35 Table 3.1: Means of soil bulk density (n=2) and characteristics of rye roots (n=4) from the studied treatments. Standard errors are shown in parentheses. Letters indicate significant differences ................................. 65 Table 3.2: Means of soil carbon and nitrogen characteristics for the three studied treatments Pre and Post. Standard errors are shown in parentheses. Means and standard errors in each treatment are calculated based on 2 -6 aggregates with 1 -13 sections per aggregate. Letters indicate cases where there was a statistically significant difference between Pre and Post results withi n each ........... ........................... 66 13 C with relative abundances of 6.5 - - - - 13 C indicate more new carbon was associated with . .................. 69 Table 4 .1: Bulk density in g/cm 3 13 C (n=12), total C (n=12), and total N (n=12) from the soil before rye planting. Shown are means and standard errors (in parenthesis). Letters indicate ................................................................ 104 Table 4.2: Means of the chemical characteristics of pulse labeled rye roots (n=43). Standard errors ..................................................................................................................................................... 104 Table 4.3: Total C (n=40), Total N (n=40), ammonia (n=16), and nitrate (n=16) means from the soil after rye growth. Standard errors are shown in parenthesis. Letters indicate significant ...................................................... ............................ 105 Table 4.4: Mean of POM (n=16) and roots (n=8) i images. Standard errors are shown in parenthesis. Letters indicate significant differences within .................................................................................... ........................... 105 ix errors are shown in parenthesis. No statistical differences were found. ...................................... 111 Table A.1: Slopes calculated by ANCOVA for the relationship between the amount of pores of the 13 ................. 132 x LIST OF FIGURES Figure 2.1: Workflow for the geostatistical analysi s. From each whole aggregate (A), 5 cubes were selected (B), (C) and then a 3D variogram obtained (D). The whole aggregate is 5 mm in size, . ......................................... ...... 24 the sample was 1 mm. (B) Image of a slice of an Os stained sample a bove the K -edge (74 keV). (C) Image of a slice of an Os stained sample below the K -edge (73.8 keV). (D) Difference between above and below K -edge images with non -biological pore (E), POM -NS (G), and POM -Root (F) expanded. Total image size is 8 x 8 mm for ( B), (C), and (D). .......................... ........................... 26 Figure 2.3: Examples of selected non -biological (A, D), POM -Root (B, E), and POM -NS pores -biological pores were chosen so that no organic m atter was visible in the pores and the pores were not round or oval in shape. POM -Root pores were chosen such that organic material was visible in the pores and were root shaped, i.e. round or oval with an elongated shape. POM -NS was chosen such that organ ic material was visible within the pores and the pore did not have a root like shape. ............................................. ............................ 27 Figure 2.4: Workflow for grayscale gradients. The whole aggregate (A) has all pores identifie d (B). Individual pores are then identified (C). Layers are collected for analysis of grayscale gradient (D). Each color represents a different layer, while the white in the middle is the actual pore and the black layer adjacent to the white accounts for p artial volume effects. ......... ........................... 28 .... 30 Figure 2.6: (A) Mean difference from the background level for Os stained samples as a function of distance from pores of biological origin with plant roots (POM -Root) and with non -root derived POM (POM -NS) and from non -biological pores (n=6). The samples a re from the biologically based system. Positive values indicate increased presence of Os labeled SOM, while negative values indicate a decrease in Os labeled SOM. (B) Normalized grayscale values for all three studied pore types from all three management pr actices. Dots are averages and the standard errors are equal to the size of the dots at each distance. The solid black lines represent the background Os labeled level (on (A)) and the background grayscale value (on (B)). ....................... ............... 33 Figure 2.7: Normalized Os values from Os stained samples (n=6) and grayscale values (n=96) from non -stained samples as a function of distance from POM -NS and POM -Root pores in the soil of the biologically based management ........................... ............................................................... 34 Figure 2.8: Normalized grayscale values as a function of distance from POM -NS pores for the three management practices (n=96). Error bars represent standard errors. The solid black line rep resents the background grayscale value, i.e. the average grayscale value of the entire aggregate. ....................................................................................................................................................... 36 xi Fig ure 2.9: Percentage of 2 - differences in 2 - represent standard errors. ....................................................................................................................................................... 37 Figure 2.10: Percentage of 2 - res of biological origin with plant roots (POM -Root) and with non -root derived POM (POM -NS) and from non -biological pores (n=8). Error bars are standard errors at each distance. ............................ ............................. 39 Figure 2.11: Examples from the three different managements of how 3DMA missed pore material adjacent to non -biological pores, but identifies POM -Root and POM -NS pores correctly. (A) is from the conventional management, (B) is from the biologically based, a nd (C) is from the early successional. The blue outlines are the pores identified by 3DMA. Red arrows on each figure indicate an example of missed porosity on each figure. ................................... ............................. 40 Figure 3.1: Exam ple of a selected root used for grayscale gradient analyses. The color overlay layer. .......................................................................... ................................................................... 63 Figure 3.2: Relative abundances of 6.5 -15, 15-40, 40 - fragments of the three studied treatments before and after incubation. Relative pore abun dance refers to the percent of medial axes per total soil volume as determined from 3DMA -Rock software. Bars represent standard errors. Letters indicate significant differences between differences between Pre and .............................................................. ............................. 67 Figure 3.3: Representative slices of the same soil fragment for Pre -Intact (A), Post -Intact (B) , Pre -Destroyed (C), and Post -Destroyed (D). Red arrows highlight an area where porosity visibly increased during incubation. Each soil fragment is approximately 5 mm across. ......................... 68 Figure 3.4: Correlations between total C (%C) and re lative abundances of 40 - 13 C and 40 - -structure treatment and destroyed -structure treatment for both Pre and Post. Relative pore abundances refer to the percent of medial axes per total soil volume as de termined from 3DMA -Rock software. Gray area indicates 95% confidence interval. Correlation coefficients are shown in Table 3.3. ................................ ............................. 70 Figure 3.5: (A) Cumulative CO 2 and (B) average isotopic signatu re of CO 2 respired during 28 day 13 C values of the soil fragment sections prior to incubation for each treatment while boxes indicate the standard errors ( -22.0±0.1›, -21.4±0.1›, and -21.6±0.1› for destroyed -structure, intact -structure, and 13 C of the emitted CO 2 and the soil secti ons Pre at .......................................................................................................................................... 72 dist ance from 40 - -structure and destroyed -structure at Pre and Post. Error bars indicate standard error. Note that values of the normalized grayscale reflect a combination of contributors, including atomic numbers of the elements and dens ity of the material located within xii an image voxel. Specifically, lower normalized values here correspond to lower atomic number elements and lower densities, while higher values correspond to higher atomic number elements and higher densities. As such, low er values roughly represent more carbon in the soil matrix, while higher values represent less carbon in the soil matrix and/or denser soil matrix. ......................... 73 Figure 3.7: Canonical correlation of pore sizes with total C (%C), total N (%N) 13 C. The first define the latent variables for each axes are shown on the right. The sign indicates the direction of correlation and the number indicate s the amount each observed variable contributes to the latent variable. Lines indicate the (0,1) line (A) and (0, -1) line (B), while letters indicate treatment (destroyed -structure (D) and intact -structure (I)). ......................................................................... 74 Figure 3.8: Correlation between total carbon and abundance of 15 - 2013, blue reproduced with permission from Elsevier) or 15 - y-axis is presented as total C , g/kg instead of %C to align with the original Anayeva et al (2013) graph. ............................................................................................................................................ 79 Figure 4.1: Image of the pulse label ing plexiglass chamber and pulse labeling set up. Note the fan 13 C enriched CO 2. Samples were rotated between rack positions for each pulse labeling event. ............................................................ ............................. 97 Figure 4.2: Pictorial schematic of soil sampling using the sample device. First, the top 1.5 mm of and the 5 mm sample (B). The soil sampling device (C) was then aligned with the red mark (D) and five samples collected simultaneously (E). The samples were then placed into tins for total 13 C analysis (F). ...................................................................... .................................. 99 Figure 4.3: Relative abundances of 4 Œ15, 15Œ40, 40Œ management. Relative pore abundance refers to the percent of medial axes per total soil volume as determined from 3DMA -Rock soft ware. Bars represent standard errors. Letters indicate significant differences at a = 0.05. ..................................................................................... ........................... 106 ive abundances of 4 Œ - pores (B), 40 - -structure with root for both Pre and Post. Relative pore abundances refer to the percent of medial axes per total soil volume as determined from 3DMA -Rock software. Outliers removed from analysis are circled in orange. confidence interval. ......................................................................................... ........................... 107 Œ - pores (B), 40 - pores (D) for destroyed -structure soils when roots are present during both Pre and Post. Relative pore abundances refer to the percent of medial axes per total soil volume as determined from 3DMA -Rock software. Stars next to the end of the lines indicate s .................. 108 2 during incubation (B). Bars indicate standard errors. Stars indicate significant differences between Pre and Post for soil carbon and intact - and destroyed -structure for CO 2 ................. ............................ 110 xiii KEY TO ABBREVIATIONS CO2 Œ Carbon dioxide CT Œ Computed microtomography CT Œ Computed tomography SOM Œ Soil organic matter Os Œ Osmium POM Œ Particulate organic matter DOM Œ Dissolved organic matter APS Œ Advanced Photon Source kGy Œ Kilogray Si Œ Silicon OsO 4 Œ Osmium tetraoxide EDPI Œ Effective distance of pore influence C Œ Carbon N Œ Nitrogen VPD B Œ Vienna PeeDee Belemite H2SO4 Œ Sulfuric acid CaCO 3 Œ Calcium carbonate 1 Chapter 1: Introduction Comprising three times the amount of CO 2 in the atmosphere and as the largest terrestrial pool of carbon, soil carbon has the potential to be a prominent player in global climate change (Batjes, 1996; Lal, 1999; Swift, 2001; Paustian et al, 2016). Soils are known to loose carbon when land is con verted to agriculture (Grandy and Robertson, 2007; Ruan and Robertson, 2013; Abraha et al, 2018). However, through diligent management, some of this loss can be recovered (Senthilkumar et al, 2009; Syswerda et al, 2011; Paustian et al, 2016) , resulting in a reduction in atmospheric CO 2. Additionally, it is projected that agricultural production needs to double current production levels by 2050 (Ray et al, 2013). Boosting soil carbon stocks in agricultural soils is known to improve yields (Melsted, 1954; Bau er and Black, 1994; Lal, 2006). Therefore, improving soil carbon stocks in agricultural soils would mitigate global climate change and help to improve global food security. Current models of soil carbon cycling, while good at predicting long term soil carb on stocks, are inadequate for modeling rapid changes, such as those that may arise due to agricultural management changes (Jenkinson et al, 1991; Parton et al, 1998; Crow and Sierra, 2018). This leads uncertainty when model ing soil carbon and CO 2 emissions , especially if the use of conservation managements becomes more widespread . As both soil carbon and CO 2 emissions are crucial input for climate change models , such as the HadCM3 , this uncertainty propagates to those models. One reason for this uncertainty is that many important processes in soil carbon dynamics occur at micro -scale, but very few models take this scale into account (Young and Crawford, 2004; Kravchenko and Guber, 2017). Including micro -scale parameters in models has already been shown to in crease model accuracy in CO 2 emission estimates (Falconer et al, 2015) and hydrodynamic soil properties (Smet et al, 2015), demonstrating that a better understanding of micro -scale carbon dynamics can result in more accurate models. 2 Soil carbon dynamics at micro -scale is believed to take place within microenvironments in the soil. The greater the diversity of these microenvironments, it is hypothesized, the greater soil carbon protection in a soil (Kuzyakov and Blagodatskaya, 2015) . This diversity of microenvironments might be able to be estimated through spatial characteristics of the soil matrix with larger variability indicating greater microenvironment diversity. Microenvironments themselves are created through the distributions of microorganisms, water, gases, and nutrients in the soil ( Young et al , 2001; Ekschmitt et al , 2005, 2008; Kravchenko and Guber, 2017; Rabot et al , 2018). These distributions, in turn, are controlled by the soil pore network via their control of water, air, and nutrient flo w. It is, therefore, soil pores that may ultimately control soil carbon dynamics. Even though soil pores control soil carbon dynamics , most research on soil carbon involves soil aggregates, specifically microaggregates -within -macroaggregates ( Six et al, 19 99, 2000). Originally proposed by Tisdall a nd Oades (198 2), aggregates are hypothesized to have a hierarchical structure with small microaggregates combining to form macroaggregates. Six et al (2000) developed methods to isolate both macroaggregates and mi croaggregates, as well as microaggregates -within -macroaggregates , and has been widely used ( Six and Paustian, 2014 and references within ) for understanding carbon retention in soils . However, using microaggregates -within -macroaggregates in modeling is prob lematic for three reasons, 1) aggregate selection is purely based on the amount of energy used to break apart the soil and therefor e, can be somewhat arbitrary, 2) the soil pore information, while theoretically, inversely related, is not clear and 3) infor mation on connections between aggregates is lost. Six and Paustian (2014) likened it to filooking at the walls of a house to understand what is happening in the living -roomfl. 3 Until the advent of computed tomography ( CT) and computed micro tomography (CT) imaging, however, direct nondestructive measurements of soil pores in situ was impossible (Dathe and Thullner, 2005; Gibson et al, 2006 ). Studies utilizing CT and CT are increasing, but rarely focus on the relationship of pores and carbon, instead focusi ng on pore creation (De Gryze et al, 2006; Schl üter and Vogel, 2016 ), connectivity ( Jarvis et al, 2017 ), and flow dynamics (Wildenschild and Sheppard, 2013; Koestel and Larsbo, 2014 ). Furthermore, CT imaging, currently, is not uti lizable for routine and l arge volume studies , while aggregate separation is . This has led to a gap in connecting aggregate observations to pores and, therefore, mechanisms . Pores can be created either through mechanical or biological means and are believed to have different effect s on soil carbon dynamics (Park et al, 2007; Peng et al, 2007) . Mechanical pores are created through (i) the shrinking and swelling of clay minerals, specifically 2:1 clays, during wetting and drying cycles (Peng et al, 2007) and (ii) the expansion of soil pores due to freeze/thaw action (Parker et al, 1982; Jabro et al, 2013). The size of these pores is usually related to the clay content of a soil, with high clay contents being required for the formation of large pores. The creation of new mechanical pores is believed to allow access to previously physically protected carbon, resulting in carbon losses (Sørensen, 1974; Denef et al , 2001; Smucker et al , 2007). This should result in less soil carbon adjacent to mechanical pores , although the extent of this influence is currently unknown. Biological pores are created primarily through root action, but can also be the result of macrofauna activity within the soil. Unlike mechanical pores, biological pores are regarded as sources o f carbon and result in carbon additions. These carbon additions, if from roots, can be in the form of root biomass or root exudates. Root biomass consists primarily of more difficult to decompose materials, such as lignin and tannin (Rasse et al , 2005; Jac kson et al , 2017). In 4 contrast, root exudates tend to be smaller organic compounds , of lower molecular weight, and more easily decomposable, such as small organic acids, carbohydrates, and amino acids (Dungait et al , 2012). Biological pores also may play a vital role in carbon protection as a large portion of carbon stored in soils is derived from root sources; up to 75% by some estimates (Balesdent and Balabane, 1996; Gale and Cambardella, 2000; Rasse et al, 2005; Clemmensen et al, 2013; Mazzilli et al, 20 15). Due to the close contact between mineral surfaces and root materials, protection of soil carbon through occlusion on mineral surfaces is believed to be enhanced (Kiem and Kögel -Knabner, 2002; Six et al, 2002, Dungait et al, 2012). This protection is b elieved to be due to the electrostatic forces in the organo -mineral complexes being greater than the enzyme binding energy, protecting the carbon from microbial attack (Dungait et al, 2012). This enhanced protection of carbon from biological pores would re sult in higher soil carbon concentrations adjacent to these pores, but the extent of influence on these soil carbon concentrations is unclear. Isotopic studies have indicated transport of decomposition materials on the order of 5 -10 mm, but this transport is not limited to individual pores (Gaillard et al , 1999, 2003; Toosi et al , 2017). NanoSIMS studies have indicated transport of only a few microns adjacent to a carbon source at nanometer resolution, although higher spatial ranges occurring at higher spat ial scales was suggested (Mueller et al, 2012). Agricultural management has been shown to have a profound effect on soil carbon dynamics (Senthilkumar et al, 2009; Syswerda et al, 2011). Management practices such as reducing tillage and increasing residue biomass are known to increase soil carbon (Follett, 2001, De Gryze et al, 2004; Ogle et al, 2012, Poeplau and Don, 2015). However, cover crop management has been shown to increase soil carbon stocks, despite utilization of heavy tillage 5 and relatively low biomass inputs (Syswerda et al, 2011; Paul et al, 2015). Additionally, some systems with abundant biomass inputs , such as monoculture switchgrass, and no tillage have unexpectedly low soil carbon accrual (Garten and Wullschleger, 1999; Chimento et al, 2016 ; Sprunger and Robertson, 2018). It is currently unknown what mechanisms account for these observations as management can affect many soil properties (pore size distribution, microorganism activity, and nutrient concentration, for example). Distinguishing between the effects of these different soil properties is difficult as it is often problematic to manipulate one factor without affecting the others. Additionally, measurements of soil carbon additions are routinely are done on microaggregates -within -micro aggregates. These microaggregates -within -macroaggregates have proven an excellent indicator of soil carbon preservation under different agricultural managements ( Six et al, 1999; Denef et al, 2004; 2007 ). However, as noted previously, this is inadequate fo r understanding mechanisms , which are currently unknown . Agricultural management can have a substantial effect on pore size distributions. Wang et al (2012) found that conventional management had a higher proportion of 37.5 Œ while early fewer 32Œ These different pore sizes can be equated to different biological pro cesses. Fungi can push aside silt particles to create 20- et al , 1993; Bearden, 2001; Emerson and McGarry, 2003; Six et al, 2006). Water content of pores and by extension oxygen concentrations is also tied to soil pore size. drain faster, resulting in high oxygen con tents, but water limitation at typical field moisture 6 levels. Pores of 10 - moisture. conditions (Schurgers et al, 2006). The gradients of oxygen availability and water content affects decomposition, which can be 1/10 th in anaerobic sites as compared to aerobic sites (Keiluweit et al, 2017). The effect of pore size on carbon dynamics has been establi shed previously. Ananyeva et al (2013) found increased carbon concentrations in soil aggregates with an increased presence of 15Œ Œ pores. Bailey et al (2017) discovered pores. Decom position has also been shown to vary depending on pore size. Killham et al (1993) found increased decomposition in pores of 6 - Strong et al (2004) found higher decomposition rates in 15 - and 60 -300 Negassa et al (2015) found that decomposition occurs faster in soil with larger pores (up to 2 mm) than smaller pores. These differences in decomposition rate s are believed to be driven by increased microbial activity in certain pores, possi bly as a function of water availability (Thomsen et al, 1999; Ruamps et al, 2011; Wang et al, 2013). These studies indicate that pores of different sizes play different roles in carbon protection. However, how pore sizes affect distributions of new carbon within soil and the actual utilization of new car bon sources are still nebulous. 7 My main study objective is to quantify the role that pores of different sizes play in influencing carbon dynamics in soils from different land use and management practices. The specific objectives were: To explore the spatial characteristics of soil matrix from different agricultural soil organic matter ( SOM ) levels of these practices . To explore the effect of pore origin (biological or mechanical) on SOM levels adjacent to the pores and relate these changes to spatial characteristics. To examine the relationships between newly added carbon and pore size distribution after addition and subsequent incubation of carbo n to understand the effect of pore size on carbon addition and usage. 8 REFERENCES 9 REFERENCES Abraha, M., Hamilton, S. K., Chen, J., and Robertson, G. P. (2018) . Ecosystem carbon exchange on conversion of Conservation Reserve Program grasslands to annual and perennial cropping systems. Agric For Meteorol. 253-254, 151-160. doi: 10.1016/j.agrformet.2018.02.016 Ananyeva, K., Wang, W., Smucker, A. J. M., Rivers, M. L ., and Kravchenko, A. N. (2013). Can intra -aggregate pore structures affect the aggregate™s effectiveness in protecting carbon? Soil Biol. Biochem. 57, 868-875. doi: 10.1016/j.soilbio.2012.10.019 Bailey, V. L., Smith, A. P., Tfaily, M., Fansler, S. J., and Bond -Lamberty, B. (2017). Differences in soluble organic carbon chemistry in pore waters sampled from different pore size domains. Soil Biol. Biochem. 107, 133-143. doi: 10.1016/j.soilbio.2016.11.025 Balesdent, J. and Balabane, M. (1996) . Major contributi on of roots to soil carbon storage inferred from maize cultivated soils. Soil Biol. Biochem. 28, 1261-1263. doi: 10.1016/0038 -0717(96)00112 -5 Batjes, N. H. (1996) . Total carbon and nitrogen in soils of the world. Eur. J. Soil Sci. 47, 151-163. doi: 10.1111 /j.1365 -2389.1996.tb01386.x Bauer, A. and Black, A. L. (1994) . Quantification of the effect of soil organic matter content on soil productivity. Soil Sci. Soc. Am. J. 58, 185-193. doi: 10.2136/sssaj1994.03615995005800010027x Bearden, B. N. (2001). Influenc e of arbuscular mycorrhizal fungi on soil structure and soil water characteristics of vertisols. Plant Soil 229, 245Œ 258. doi: 10.1023/A:1004835328943 Cannell, R. Q. (1977). Soil aeration and compaction in relation to root growth and soil management. Appl . Biol . 2, 1-86. Chimento, C., Almagro, M., and Amaducci, S. (2016) . Carbon sequestration potential in perennial bioenergy crops: the importance of organic matter in puts and its physical protection. GCB Bioenergy 8, 111-121. doi: 10.1111/gcbb.12232 Clemmen sen, K. E., Bahr, A., Ovaskainen, O., Dahlberg, A., Ekblad, A., Wallander, H., et al. (2013) . Roots and associated fungi drive long -term carbon sequestration in boreal forest. Science 339, 1615-1618. doi: 10.1126/science.1231923 Crow, S. E. and Sierra, C. A. (2018) . Dynamic, intermediate soil carbon pools may drive future responsiveness to environmental change. J. Environ. Qual. 47, 607-616. doi:10.2134/jeq2017.07.0280 Dathe , A.and Thullner, M. (2005). The relationship between fractal properties of solid ma trix and pore space in porous media. Geoderma 134, 318 -326. doi:10.1016/j.geoderma.2005.01.003 10 De Gryze, S., Six, J., Paustin, K., Morris, S. J., Paul, E. A., and Merckx, R. (2004) . Soil organic carbon pool changes following land -use conversions. Global Ch ange Biol. 10, 1120-1132. doi: 10.1111/j.1365 -2486.2004.00786.x De Gryze, S., Jassogne, L., Six, J., Bossuyt, H., Wevers, M., and Merckx, R. (2006). Pore structure changes during decomposition of fresh residue: X -ray tomography analyses. Geoderma 134, 82-96. doi: 10.1016/j.geoderma.2005.09.002 Denef, K., Six, J., Bossuyt, H., Frey, S. D., Elliot, T. E., Merckx, R., et al. (2001). Influence of dry -wet cycles on the interrelationship between aggregate, particulate organic matter, and microbial community dynam ics. Soil Biol. Biochem. 33, 1599-1611. doi: 10.1016/S0038 -0717(01)00076 -1 Denef, K., Six, J., Merckx, R ., and Paustian, K. (2004). Carbon sequestration in micro -aggregates of no -tillage soils with different clay mineralogy. Soil Sci. Soc. Am. J. 68, 1935-1944. doi: 10.2136/sssaj2004.1935 Denef, K., Zotarelli, L., Boddey, R. M., and Six, J. (2007). Microaggregate -associated C as a diagnostic fraction for management -induced changes in soil organic carbon in two Oxisols. Soil Biol. Biochem . 39, 1165-1172. doi: 10.1016/j.soilbio.2006.12.024 Dorioz, J. M., Robert, M., and Chenu, C. (1993). The role of roots, fungi and bacteria on clay particle organization. An experimental approach. Geoderma 56, 179Œ 194. doi:10.1016/0016 -7061(93)90109 -X Dungait, J. A. J., Hopki ns, D. W., Gregory, A. S., and Whitmore, A. P. (2012). Soil organic matter turnover is governed by accessibility not recalcitrance. Global Change Biol. 18, 1781- 1796. doi: 10.1111/j.1365 -2486.2012.02665.x Ekschmitt, K., Kandeler, E., Poll, C., Brune, A., B uscot, F., Friedrich, M., et al. (2008). Soil -carbon preservation through habitat constraints and biological limitations on decomposer activity. J. Plant Nutr. Soil Sci. 171, 27-35. doi: 10.1002/jpln.200700051 Ekschmitt, K., Liu, M., Vetter, S., Fox, O., a nd Wolters, V. (2005). Strategies used by soil biota to overcome soil organic matter stability Œ why is dead organic matter left over in soil? Geoderma 128, 167-176. doi: 10.1016/j.geoderma.2004.12.024 Emerson, W. W. and McGarry, D. (2003). Organic carbon and soil porosity. Aust. J. Soil Res. 41, 107Œ118. doi: 10.1071/SR01064 Falconer, R. E., Battaia, G., Schmidt, S., Baveye, P., Chenu, C., and Otten, W. (2015). Microscale heterogeneity explains experimental variability and non -linearity in soil organic mat ter mineralization. PloS One . 10(5) : e0123774. doi: 10.1371/journal.pone.0123774 Follet, R. F. (2001) . Soil management concepts and carbon sequestration in cropland soils. Soil Tillage Res. 61, 77-92. doi: 10.1016/S0167 -1987(01)00180 -5 11 Gaillard, V., Chenu, C., and Recous, S. (2003). Carbon mineralization in soil adjacent to plant residues of contrasting biochemical quality. Soil Biol. Biochem. 35, 93-99. doi: 10.1016/S0038 -0717(02)00241 -9 Gaillard, V., Chenu, C., Recous, S. and Richard, G. (1999). Carbon, n itrogen and microbial gradients induced by plant residues decomposing in soil. Eur. J. Soil Sci. 50, 567 -578. doi: 10.1046/j.1365 -2389.1999.00266.x Gale, W. J. and Cambardella, C. A. (2000) . Carbon dynamics of surface residue - and root -derived organic matt er under simulated no -till. Soil Sci. Soc. Am. J. 64, 190 -195. doi: 10.2136/sssaj2000.641190x Garten, C. T., and Wullschleger, S. D. (2000) . Soil carbon dynamics beneath switchgrass as indicated by stable isotope an alysis. J. Envrion. Qual. 29, 645-653. doi: 10.2134/jeq2000.00472425002900020036x Gibson, J. R., Lin, H., and Bruns, M. A. (2006). A comparison of fractal analytical methods on 2- and 3 -dimensional computed tomographic scans of soil aggregates. Geoderma . 134, 335-348. doi: 10.1016/j.geoderma.2006 .03.052 Grandy, A. S. and Robertson , G. P. (2007). Land -use intensity effects on soil organic carbon accumulation rates and mechanisms. Ecosystems 10, 58-73. doi: 10.1007/s10021 -006-9010-y Jabro, J. D, Iverson, W. M., Evans, R. G., Allen, B. L., and Steven s, W. B. (2013) . Repeated freeze -thaw cycle effects on soil compaction in a clay loam in northeastern Montana. Soil Sci. Soc. Am. J. 78:737 -744. doi: 10.2136/sssaj2013.07.0280 Jackson, R. B., Lajtha, K., Crow, S. E., Hugelius, G., Kramer, M. G., and Piñeir o, G. (2017). The ecology of soil carbon: pools, vulnerabilities, and biotic and abiotic controls. Annu Rev Ecol Evol Syst . 48, 419-445. doi: 10.1146/annurev -ecolsys -112414-054234 Jarvis, N., Koestel, J., and Larsbo, M. (2017). Understanding preferential f low in vadose zone: Recent advances and future prospects. Vadose Zone 15 doi: 10.2136/vzj2016.09.0075 Jenkinson, D. S., Adams, D. E., and Wild, A. (1991) . Model estimates of CO 2 emissions from soil in response to global warming. Nature 351, 304-306. doi: 10.1038/351304a0 Keiluweit, M., Wanzek, T., Kleber, M., Nico, P., and Fendorf, S. (2017). Anaerobic microsites have an unaccounted role in soil carbon stabilization. Nat. Commun. 8: 1771. doi: 10.1038/s41467 -017-01406-6 Kiem, R., and Kögel -Knabner, I. (200 2). Refractory organic carbon in particle -size fractions of arable soils II: organic carbon in relation to mineral surface area and iron oxides in fractions <6 Org. Geochem. 33, 1699 -1713. doi: 10.1016/S0146 -6380(02)00112 -2 Killham, K., Amato, M., and Ladd, J. N. (1993) . Effect of substrate locations in soil and soil pore -water regime on carbon turnover. Soil Biol. Biochem. 25, 57-62. doi: 10.1016/0038 -0717(93)90241 -3 12 Koestel, J. and Larsbo, M. (2014). Imaging and quantification of preferential solute t ransport in soil macropores. Water Resour Res 50, 4357-4378. doi: 10.1002/2014WR015351 Kravchenko, A. N ., and Guber, A. K. (2017). Soil pores and their contributions to soil carbon processes. Geoderma 287, 31 -39. doi: 10.1016/j.geoderma.2016.06.027 Kravchenko, A. N., Negassa, W. C., Guber, A. K., Hildebrandt, B., Marsh, T. L., and Rivers, M. L. (2014). Intra -aggregate pore structure influences phylogenetic composition of bacterial community in macroaggregates. Soil Sci. Soc. Am. J. 78, 1924-1939. doi: 10.2136/sssaj2014.07.0308 Kuzyakov, Y., and Blagodatskaya, E. (2015). Microbial hotspots and hot moments in soil: concept & review. Soil Biol. Biochem. 83, 184-199. doi: 10.1016/j.soilbio.2015.01.025 Lal, R. (1999). Soil management and restoration for ca rbon sequestration to mitigate the accelerated greenhouse effect. Prog. Env. Sci. 1, 307 -326. doi: 10.1055/s -2008-1072270 Lal, R . (2006) . Enhancing crop yields in the developing countries through restoration of the soil organic carbon pool through restorat ion of the soil organic carbon pool in agricultural lands. Land Degrad. Develop. 17, 197-209. doi: 10.1002/ldr.696 Mazzilli, S. R., Kemanian, A. R., Ernst, O. R., Jackson, R. B., and Piñero, G. (2015) . Greater humification of belowground than aboveground b iomass carbon into particulate soil organic matter in no -till corn and soybean crops. Soil Bio. Biochem. 85, 22-30. doi: 10.1016/j.soilbio.2015.02.014 Melsted, S. W. (1954) . New concepts of management of corn belt soils. Adv. Agron. 6, 121-142. doi: 10.101 6/S0065 -2113(08)60383 -1 Mueller, C. W., Kölbl, A., Hoeschen, C., Hillion, F., Heister, K., Herrmann, A. M ., et al. (2012) . Submicron scale imaging of soil organic matter dynamics using NanoSIMS -from single particles to intact aggregates. Org Geochem. 42, 1476-1488. doi: 10.1016/j.orggeochem.2011.06.003 Negassa, W. C., Guber, A. K., Kravchenko, A. N., Marsh, T. L., Hildebrandt, B., and Rivers, M. L. (2015). Properties of soil pore space regulate pathways of plant residue decomposition and community structure of associated bacteria. PLoS One 10(4) : e0123999 doi: 10.1371/journal.pone.0123999 Ogle, S. M., Swan, A., and Paustian, K. (2012) . No-till management impacts on crop productivity, carbon input and soil carbon sequestration. Agric Ecosyst Environ. 149, 37-49. doi: 10.1016/j.agee.2011.12.010 Park, E. J., Sul, W. J., and Smucker, A. J. M. (2007). Glucose additions to aggregates subjected to drying and wetting cycles promote carbon sequestration and aggregate stability. Soil Biol. Biochem. 39, 2758-2768. doi: 10.1016/j.soilbio.2007.06.007 13 Parker, J. C., Amos, D. F., and Zelazny, L. W. (1982) . Water adsorption and swelling of clay minerals in soil systems. Soil Sci. Soc. Am. J. 46, 450-456. doi: 10.2136/sssaj1982.03615995004600030002x Parton, W. J., Hartman, M., Ojima, D., and Schimel, D. (1998) . DAYCEN T and its land surface submodel : description and testing. Global Planet. Change 19, 35-48. doi:10.1016/S0921 -8181(98)00040 -X Paul, E. A., Kravchenko, A., Grandy, S., and Morris, S. (2015). fiSoil organic matter dyna mics: controls and management for sustainable ecosystem functioning,fl in The ecology of agricultural landscapes: long -term research on the path to sustainability , eds. S. K. Hamilton, J. E. Doll, and G. P. Robertson, (New York, New York, USA: Oxford Univer sity Press), 104 -134. Paustian, K., Lehmann, J., Ogle, S., Reay, D., Robertson, G. P., and Smith, P. (2016) . Climate -smart soils. Nature 532, 49-57. doi: 10.1038/nature17174 Peng, X., Horn, R., and Smucker, A. J. M. (2007). Pore -shrinkage dependency of ino rganic and organic soils on wetting and drying cycles. Soil Sci. Soc. Am. J. 71, 1095-1104. doi: 10.2136/sssaj2006.0156 Poeplau, C. and Don, A. (2015). Carbon sequestration in agricultural soils via cultivation of cover crops -a meta -analysis. Agric Ecosyst Environ. 200, 33-41. doi: 10.1016/j.agee.2014.10.024 Rabot, E., Wiesmeier, M., Schlüter, S., and Vogel, H. -J. (2018) . Soil structure as an indicator of soil functions: a review. Geoderma 314, 122-137. doi:10.1016/j.geoderma.2017.11.009 Rasse, D. P., Rumpe l, C. and Dignac, M. -F. (2005) . Is soil carbon mostly root carbon? Mechanisms for a specific stabliz ation. Plant Soil 269, 341-356. doi: 10.1007/s11104 -004-0907-y Ray, D. K., Mueller, N. D., West, P. C., and Foley, J. A. (2013) . Yield trends are insufficie nt to double global crop production. PLoS One 8(6) :e66428. doi:10.1371/journal.pone.0066428 Ruamps, L. S., Nunan, N., and Chenu, C. (2011) . Microbial biogeography at the soil pore scale. Soil Biol. Biochem. 43, 280-286. doi: 10.1016/j.soilbio.2010.10.010 Raun, L. and Robertson, G. P. (2013). Initial nitrous oxide, carbon dioxide, and methane costs of converting Conservation Reserve Program grassland to row crops under no -till vs. conventional tillage. Global Change Biol. 19, 2478-2489. doi: 10.1111/gcb.122 16 Schlüter, S. and Vogel, H. -J. (2016). Analysis of soil str ucture turnover with garnet part icles and X-ray microtomography. PLo S One 11.7:pe0159948. doi: 10.1371/journal.pone.0159948 Schurgers, G., Dörsch, P., Bakken, L., Leffelarr, P., and Haugen, L. E. (2006) . Modelling soil anaerobiosis from water retention characteristics and soil respiration. Soil Biol. Biochem. 38, 2637-2644. doi: 10.1016/j.soilbio.2006.04.016 14 Senthilkumar, S., Basso, B., Kravchenko, A. N., and Robertson, G. P. (2009). Contemporary evidence of soil carbon loss in the U.S. corn belt. Soil Sci. Soc. Am. J. 73, 2078-2086. doi: 10.2136/sssaj2009.0044 Six, J., Conant, R. T., Paul, E. A., Paustian, K. (2002). Stabilization mechanisms of soil organic matter: implications for C -saturation of soils. Plant Soil . 241, 155-176. doi: 10.1023/A:1016125726789 Six, J., Elliot, E. T., and Paustian, K. (1999) Aggregate and soil organic matter dynamics under conventional and no -tillage systems. Soil Sci. Soc. Am. J. 63, 1350-1358. doi: 10.2136/sssaj1999.6351350x Six, J., Elliott, E. T., and Paustian, K. (2000). Soil macroaggregate turnover and microaggregate formation: a mechanism for C sequestration under no -tillage agriculture. Soil Biol. Biochem. 32, 2099-2103. doi: 10.1016/S0038 -0717(00)00179 -6 Six, J., Frey, S. D., Thiet, R. K., and Batten, K. M. (2006). Bacterial and fungal contributions to carbon sequestration in agroecosystems. Soil Sci. Soc. Am. J. 70, 555-569. doi:10.2136/sssaj2004.0347 Six, J. and Paustian, K. (2014). Aggregate -associated soil organic matter as an ecosystem property and measurement tool. Soil Biol. Biochem 68, A4 -A9. doi: 10.1016/j.soilbio.2013.06.014 Smet, S., Beckers, E., Plougonven, E., Léonard, A., and Degré, A. (2018) . Can the pore scale geometry e xplain soil sample scale hydrodynamic properties? Front. Environ. Sci. 6:20. doi: 10.3389/fenvs.2018.00020 Smucker, A. J. M., Park, E. -J., Dorner, J., and Horn, R. (2007). Soil micropore development and contributions to soluble carbon transport within macr oaggregates. Vadose Zone J. 6, 282-290. doi: 10.2136/vzj2007.0031 Sørensen, L. H. (1974). Rate of decomposition of organic matter in soil as influenced by repeated air drying -rewetting and repeated additions of organic material. Soil Biol. Biochem. 6, 287-292. doi: 10.1016/0038 -0717(74)90032 -7 Sprunger C. D. and Robertson, G. P. (2018) . Early accumulation of active fraction soil carbon in newly established cellulosic biofuel systems. Geoderma 318, 42-51. doi: 10.1016/j.geoderma.2017.11.040 Strong, D. T., De Wever, H., Merckx, R., and Recous, S. (2004). Spatial location of carbon decomposition in the soil pore system. Eur. J. Soil Sci. 55, 739-750. doi: 10.1111/j.1365 -2389.2004.00639.x Swift, R. S. (2001). Sequestration of carbon by soil. Soil Sci. 166, 858-871. doi: 10.1097/00010694 -200111000-00010 15 Syswerda, S. P., Corbin, A. T., Mokma, D. L., Kravchenko, A. N., and Robertson, G. P. (2011). Agricultural management and soil carbon storage in surface vs. deep layers. Soil Sci. Soc. Am. J. 75, 92-101. doi: 10.2 136/sssaj2009.0414 Thomsen, I. K, Schjønning, P., Jensen, B., Kristensen, K., and Christensen, B. T. (1999) . Turnover of organic matter in differently textured soils -II. Microbial activity as influenced by soil water regimes. Geoderma 89, 199-218. doi: 10. 1016/S0016 -7061(98)00084 -6 Tisdall, J. M. and Oades, J. M. (1982) . Organic matter and water -stable aggregates in soils. J. Soil Sci. 62, 141-163. doi: 10.1111/j.1365 -2389.1982.tb01755.x Toosi, E. R., Kravchenko, A. N., Guber, A. K., and Rivers, M. L. (2017 ). Pore characteristics regulate priming and fate of carbon from plant residue. Soil Bio. Biochem. 113, 219 -230. doi: 10.1016/j.soilbio.2017.06.014 Wang, W., Kravchenko, A. N., Johnson, T., Srinivasan, S., Ananyeva, K. A., Smucker, A. J. M., et al. (2013). Intra -aggregate pore structures and Escherichia coli distribution by water flow within and movement out of soil macroaggregates. Vadose Zone J. 12:4. doi: 10.2136/vzj2013.01.0012 Wang, W., Kravchenko, A. N., Smucker, A. J. M., Liang, W., and Rivers, M. L. (2012). Intra -aggregate pore characteristics: X -ray computed microtomography analysis. Soil Sci. Soc. Am. J. 76, 1159-1171. doi: 10.2136/sssaj2011.0281 Wiersum, L. K. (1957). The relationship of the size and structural rigidity of pores to their penetrati on by roots. Plant Soil 9, 75-85. doi: 10.1007/BF01343483 Young, I. M., and Crawford, J. W. (2004). Interactions and self -organization in the soil -microbe complex. Science 304, 1634 -1637. doi: 10.1126/science.1097394 Young, I. M., Crawford, J. W., and Rappoldt, C. (2001). New methods and models for characterizing structural heterogeneity of soil. Soil Tillage Res. 61, 33-45. doi: 10.1016/S0167 -1987(01)00188 -X 16 Chapter 2: Patterns and Sources of Spatial Heterogeneity i n Soil Matrix from Contrasting Long Term Management Practices Abstract in situ 3D visualization of soil at micron scale became easily achievable. -based research has focused on visualizat ion and quantification of soil pores, roots, and particulate organic matter (POM), whi le little effort has been invested into exploring the soil matrix itself. This study aims to characterize spatial heterogeneity of soil matrix in macroaggregates from thr ee differing long term managements: conventionally managed and biologically based row -crop agricultural systems and early successional unmanaged system, and explore the utility of using grayscale gradients as a proxy of soil organic matter (SOM). To determ ine spatial characteristics of the soil matrix, I completed a geostatistical analysis of the aggregate matrix. It demonstrated that, while the treatments had the same range of spatial auto correlation, there was much greater overall variability in soil from the biologically based system. Since soil from both managements have the same mineralogy and texture, I hypothesized that greater variability is due to differences in SOM distributions, driven by spatial distribution patterns of soil pores. To test this h ypothesis, I applied osmium (Os) staining to intact micro -cores from the biologically based management, and -biological origin. Biological pores had the highest SOM levels adjacent to the pore, which receded to background levels at distances of 100 - Non -biological pores had lower SOM levels adjacent to the pores and returned to background levels at distances of 30 - This indicates that some of the spatial heterogeneity within the soil matrix can be ascribed to SOM Originally published as: Quigley, M. Y., Rivers, M. L., and Kravchenko, A. N. (2018b). Patterns and sources of spatial heterogeneity in soil matrix from contrasting long term management practices. Front. Environ. Sci. 6:28 doi: 10.3389/fenvs.2018.00028 17 distribution patterns as controlled by pore origins and distributions. Lastly, to determine if the grayscale values could be used as a proxy for SOM levels, gradients of grayscale values from biol ogical and non -biological pores were compared with the Os gradients. Grayscale gradients matched Os gradients for biological pores, but not non -biological pores due to an image processing artifact. Grayscale gradients would, therefore, be a good proxy for SOM gradients near biological origin pores, while their use for non -biological pores should be conducted with caution. As the same pattern was seen for root and non -root pores from all managements, this indicates that SOM distribution is controlled by the same mechanism regardless of management. 2.1 Introduction in situ characterization of the physical structure of soil, specifically, positions, size distributions, and shapes of soil pores (Gibson et al, 2006; Chun et al, 2008; Peth et al, 2008; Papadopoulos et al, 2009; Kravchenko et al, 2011; Wang et al, 2012 ). It has also enabled identification of large organic fragments, including particulate organic matter (POM) ( Kravchenko et al, 2014a ) and intact plant roots (Mooney et al, 2012 ). These advances have led to quantitative insights into the contribution of pore characteristics to residue decomposition, carbon protection, and spatial patterns of bacterial distributions ( De Gryze et al, 2006 ; Ananyeva et al, 2013 ; Wang et al, 2013 ; Kravchenko et al, 2014b; Negassa et al, 2015 ). However, little attention regarding mineral soil matrix, that is, solid material containing no pores or organic fragments visible at material may relate to soil organic matter (SOM) dynamics, specifically SOM protection. correlated to the attenuation of x -rays, which is controlled by the density and atomic number (Z) of the elements occurring within an image voxel ( Ketcham, 2005 ; Peth, 2010 ). Voxels that 18 contain primarily low Z elements, such as nitrogen, carbon, and oxyg en, have lower grayscale silicon, and aluminum, have higher grayscale values (appear brighter). A voxel™s overall grayscale value is the average attenuation of t he elements occurring within that voxel. Spatial variability in grayscale values of the solid material originates from multip le sources, including variations in mineralogy, presence of pores with sizes below image resolution, and SOM distribution patterns. The first two of these factors are important drivers of SOM protection , while SOM distribution patterns can be an indicator of where such protection has occurred . Mineralogy influences SOM protection by affecting organic matter binding via electrostatic forces . Small pores can contribute to SOM protection by a combination of restricting decomposers' access and , due to anaerobic conditions prevalent in these pores , slowing of decomposition (Bailey et al, 2017 ; Keiluweit et al, 2017 ). Thus, overall SOM distribution patterns are likely controlled by a combination of mineralogy and pore architecture, i.e. pore size and connectivity ( Dungait et al, 2012 ; Kravchenko et al, 2015 ). Soil pores function as the soil transport network; regulating the flow of nutri ents, microorganisms, oxygen, and organic material (De Gryze et al , 2006; Young and Crawford, 2004; Kuzyakov and Blagodatskaya, 2015; Negassa et al , 2015). Soil pores are created through either biological or non -biological means. Biological pores are forme d by macrofauna, such as earthworms, or through action of roots (Rasse and Smucker, 1998) and root hairs as they spread and grow. Non -biological pores are primarily produced in a course of wetting/drying and freeze/thaw cycles and are controlled by soil te xture, specifically clay content. Biological and non-biological pores also play different roles in the cycling of organic matter within the soil. Biological pores are generally thought of as a source of new carbon inputs, either through direct 19 organic addi tion, such as decaying roots, or through ancillary organic matter additions, such as root exudates. Organic matter that then diffuses out from biological pores typically occurs as dissolved organic matter (DOM). DOM can be bound to minerals by electrostati c forces (Kiem and Kögel -Knabner, 2002; Six et al, 2002, Dungait et al, 2012), where, due to the electrostatic force being greater than the enzyme binding energy, it can be protected from microbial attack and results in SOM protection (Dungait et al, 2012) . While biological origin pores are sources of organic matter and SOM protection, they also compress adjacent solid material as roots push through the soil resulting in denser material closer to root pores (Bengough et al, 2011; Aravena et al, 2014). Thus, the net effect of biological pores on proximate grayscale densities is uncertain. Formation of non -biological pores, on the other hand, created through the shrinking and swelling of clay minerals, can expose previously inaccessible carbon to microbial att ack, resulting in a net carbon loss (Sorensen, 1974; Denef et al, 2001; Smucker et al, 2007). However, quantitative data on how presence, abundance, and characteristics of pores of different origins influence SOM accrual and protection is currently lacking . Falconer et al (2015) noted that despite identical bulk characteristics, including average porosity, POM turnover rate varied widely due to micro -scale properties. Their results indicated that an understanding of micro -scale pore properties may be vital to achieve more accurate modeling of soil carbon dynamics. insights into spatial patterns of SOM and the associations between such patterns and pores of different origins. A s noted previously, a voxel™s overall grayscale value is the average attenuation of the elements occurring within that voxel. While mineralogy plays the largest role in the spatial characteristics of the solid material and would normally override any spati al characteristics from distribution of SOM and presence of pores with below image resolution 20 sizes, s amples with similar mineralogy c ould allow for the spatial patterns caused by these other factors to be observed. There is some experimental evidence that , in samples with similar communications). Studies have shown that geostatistics is helpful for describing the spatial characteristics of pores and, therefore, can be exp ected to also model well the spatial characteristics of the solid material ( De Gryze et al, 2006 ; Feeney et al, 2006 ; Nunan et al, 2006 ). Therefore, to assess spatial patterns within the solid material, geostatistics will be used to quantify the range of s patial correlation, overall spatial variability, and the contribution of spatial variability that is below image resolution to the overall spatial variability. Here I as a proxy for the spatial patterns of SOM distribution. I will focus on SOM distribution patterns in the vicinity of soil pores of both non -biological and biological origin. This focus is driven by an expectation that, due to the role of biological pores in su pplying new organic inputs and the role of non -biological pores in contributing to carbon losses, their comparison should yield contrasting gradients in SOM distributions. Identification of such gradients will indicate that grayscale values can provide use ful information on SOM distribution patterns within intact soil samples. To further test the utility of grayscale values as indicators of SOM spatial patterns, I will Osmium (Os) stained organic matter from Os dual -energy images. Os staining has been proposed Peth et al , 2014 ) and was applied to estimate SOM spatial patterns ( Rawlins et al , 2016 ). Os strong ly binds with carbon -carbon 21 images. By taking images above and below the K -edge of Os, 3D maps of SOM within an intact soil sample can be constructed. From such maps I can then obtain direct measurements of SOM gradients in the vicinity of non -biological and biological soil pores. Agricultural management is known to have an effect on overall SOM levels ( Oades, 1984; Six et al, 2000 ; Syswerda et al, 2011 ; Paul et al, 2 015), as well as on micro -scale SOM patterns ( Ananyeva et al, 2013 ; Kravchenko et al, 2015). Since pores are known drivers of SOM protection, any change in the spatial pattern of pores would result in a change in SOM spatial patterns and potentially SOM le vels. The distribution of non -biological and biological pores is known to be affected by agricultural management. Wang et al (2012) and Kravchenko et al (2014b) both observed that non -biological pores tended to dominate in systems with tillage, while pores of biological origin tended to dominate in conservation managements with little soil disturbance. I hypothesize d that areas dominated by non -biological pores will have relatively uniform microenvironmental conditions, and thus relatively uniform SOM spatial distribution patterns. Together with lack of point sources of organic matter in such pores, this will lead to smaller SOM gradients in their vicinity. On the other hand, areas dominated by biological pores provide spatially variable SOM inputs as w ell as a more diverse range of microenvironmental conditions for microorganisms. Thus, I expect greater spatial variability in SOM, as well as greater SOM gradients, in the vicinity of biological pores. In addition, I hypothesize that agricultural manageme nt practices that lead to a greater presence of biological pores will increase SOM spatial variability and result in larger SOM gradients than the management practices with greater presence of non -biological pores. 22 The first objective of the study is to ex plore utility of grayscale values of solid material with Os spatial patterns. The second objective is to explore spatial characteristics of the solid images of intact soil samples from three contrasting land use and management practices and to analyze relationships between the spatial characteristics and the SOM levels of these practices. My third objective is to explore SOM and grayscale value gradien ts at distances from pores of different origins and in soils from different managements. 2.2 Materials and methods 2.2.1 Soil Collection and Imaging The studied soil was collected from three different managements at Kellogg Biological Station Long Term Eco logical Research site , Hickory Corners, MI (42°24´N, 85°24´W) . The three managements were a convention al corn -soybean wheat rotation maintained with current best management practices, a biologically based corn -soybean -wheat with rye cover after corn and re d clover interseeded into wheat with no additional inputs and rotary tillage between rows for weed management, and an early successional management, which is burned annually, but otherwise unmanaged. These management practices represent a management gradie nt with a highly managed system (conventional), a conservation management system (biologically based), and an unmanaged system ( early successional). Further details can be found in Kravchenko et al (2015) . The soil (from 0 -15 cm depth) was dry sieved and a ggregates of 4 -6.3 mm were collected -BM-D of the GeoSoilEnvironCARS (GSECARS) at the Advanced Photon Source (APS), Argonne National Laboratory (ANL) in Argonne, Illinois. Two -dimensional projections we re taken at 0.25° rotation angle steps with a one second exposure and combined into a three -dimensional image consisting of 520 slices with 23 696 by 696 pixels per slice for grayscale analysis and 1200 slices with 1920 by 1920 pixels per slice for analysis o f pores below image resolution . using the indicator kriging method in 3DMA -Rock ( Oh and Lindquist, 1999 ; Wang et al, 2011) for grayscale analysis and through simple thresholding with Otsu™s method for analysis of pores below image resolution. 2.2.2 Geostatistical Analysis A total of 32 soil aggregate images were used in the geostatistical analysis, namely, 11 images from conventional and biologically based management and 10 images from early successional management. On each image I x x in size (Figure 2.1). Positions of the cubes were initially randomly selected, with further adjustments made to avoid major overlaps with other cubes, coarse sand grains that would not reflect the overall spatial characteristics of the aggregate, and aggregate boundaries. Soil pores identified by 3DMA -Rock were removed from the cubes prior to geostatistical analysis allowing for analysis of spatial patterns in the solid material only. 3D variograms were obtained using the gstat package in R ( Pebesma, 2004 ) run on the High Perfo rmance Computing Center at Michigan State University. Variograms were fit with an exponential model using PROC NLIN in SAS 9.3 (SAS Inc., 2009 ). Spatial characteristics of the solid material can be determined from the components of the 3D variogram s. The s ill, where the variogram asymptotes, indicat es the total spatial variability with in a sample. The range of a variogram , lag distance at which the sill spatial auto correlation exists in a sample. The nugget , the difference between the zero and the y -interce pt, represents both measurement error and the variability at scales below the image resolution . The nugget to sill ratio describes the relative amount of spatial dependence at the voxel size. 24 Figure 2. 1: Workflow for the geostat istical analysis. From each whole aggregate (A) , 5 cubes were selected (B) , (C) and then a 3D variogram obtained (D) . The whole aggregate is 5 mm in size, while the cubes in the slice (B) and 3D (C) 2.2.3 Os Gradients Soil samples for Os analysis were taken as mini -cores. Only three mini -cores, all from the biological based management practice, were analyzed. The reason for the small number of samples used for this analysis is the very long image collection time for dua l-energy Os scans limits the number of samples that could be processed. I choose biologically based management for these analyses, since I expected that pores of both non -biological and biological origin would be well represented in soil under this managem ent. Samples were taken as 8 mm mini -cores, as opposed to dry sieved aggregates, because of concerns that aggregates would be too fragile for the multiple handling steps required by this method. 25 The mini -cores were taken at 3.5 -5 cm depth using a beveled 3 mL Luer -Lok polypropylene syringe with an 8 mm inner diameter ( BD, Franklin Lakes NJ, USA ). There was minimal interference with Os staining , which binds to carbon -carbon double bonds, from polypropyle ne syringes as polypropylene contains almost no carbon -carbon double bounds. Cores were air dried and exposed to OsO 4 gas in a fume hood for one week. This allowed ample time for the OsO 4 gas to diffuse throughout the soil and to ensure maximum binding of Os to the soil organic material. The cores were then sc anned at beamline 13 -BM-D, GSECARS , APS ANL. T wo-dimensional projections were taken at 0.25° rotation angle steps with a two second exposure and combined into a three -dimensional image consisting of 1200 slices with 1920 by 1920 pixels per slice. Final ima scans, 74 keV, 73.8 keV, and 28 keV. These energies provided, respectively, an image above Os K-edge, an image below Os K -edge, and an image at an energy optimal for soil pore and POM identificati on. By taking the difference between the above and below K -edge images, a map of the stained soil organic materials was created (Figure 2.2). Using the 28 keV images, non - biological pores were identified using simple thresholding, while POM pieces of both root and non-root origin were visually identified by hand. POM of non -root origin was defined as stand -alone organic fragments of round or irregular shape that were not connected to any obvious root remains. 26 Figure 2. 2 (A) A 3D scan of an entire Os stained samples. The thickness of the sample was 1 mm. (B) Image of a slice of an Os stained sample above the K -edge (74 keV). (C) Image of a slice of an Os stained sample below the K-edge (73.8 keV). (D) Difference between above and below K -edge images with non -biological pore (E) , POM -NS (G) , and POM -Root (F) expanded. Total image size is 8 x 8 mm for (B) , (C) , an d (D) . Two pores containing non -root derived POM (POM -NS), two pores containing root -derived POM (POM -Root) and four pores of non -biological origin were identified by hand for the analyses in each mini -core image (Figure 2.3). Identified pores varied in size from 20 to 300 lues of the Os stained map were averaged for each 27 the image background. To ensure comparability among the mini -cores, the Os gradients were standardized by subt racting the Os map™s average grayscale value from each mini -core. Figure 2. 3: Examples of selected non -biological (A, D) , POM -Root (B, E) , and POM -NS pores (C, F) -biological pores were chosen s o that no organic matter was visible in the pores and the pores were not round or oval in shape. POM -Root pores were chosen such that organic material was visible in the pores and were root shaped, i.e. round or oval with an elongated shape. POM -NS was cho sen such that organic material was visible within the pores and the pore did not have a root like shape. 2.2.4 Grayscale Gradients From each of the 32 images used in the geostatistical analysis, I identified three POM -NS, three POM -Root, and five non -biological origin pores (Figure 2.3). Identified pores ranged in sizes by one voxel in all three dimensio ns, thus they did not include the layer of border voxels that contained both pore and solid material. The grayscale value gradients were obtained by 2.4). Averages did not include the 0 and 255 grayscale values as the 0 value was the value of the image background and excluding the 255 value corrected for any overly dense material, such as 28 iron minerals like magnetite or limonite, in the samples that might have skewed the grayscale averages. To enable direct comparisons between images, the grayscale value gradients were Figure 2. 4: Workflow for grayscale gradients. The whole aggregate (A) has all pores identified (B) . Individual pores are t hen identified (C) . Layers are collected for analysis of grayscale gradient (D) . Each color represents a different layer, while the white in the middle is the actual pore and the black layer adjacent to the white accounts for partial volume effects. normalized so that the minimum grayscale value was 0 and the maximum grayscale value was 1 for each gradient. Calculation of the distance over which the gradient had influence was done by 29 fitting the individual gradients using PROC NLIN in SAS 9.4 (SAS Inc., 2009) with the following non -linear model: ()=+()×1 (2.1) where x is distance from the pore, n is the y -intercept, s is the average grayscale value of the image, and d is the distance at which the pore stops affecting the grayscale values or effective distance of pore influence (EDPI). 2.2.5 Analysis of pores below image resolution (2 - gradients of grayscale values. If so, the n it would not be possible to attribute the observed gradients in grayscale values to SOM. In order to explore the potential effect of such pores on the studied grayscale gradients, I aggregates from each management were scanned. I explored the differences among the studied managements of the presence of 2 - pores. The purpose of this analysis was to ensure that the o bserved differences among the management practices were driven by SOM and not by below -resolution pores. For determination of presence of 2 -13 m, four cubes, 140 x 140 x using a selection process identical to that described above for the geostatistical analysis. Using Otsu™s method, the overall porosity of each cube was determined. Binning was then used to image was then determined using the same threshold as the 2 m samples. Subtracting the porosity of the binned images from the un - referred to hereafter as the below image resolution porosity. 30 Figure 2. 5: Exam (A) (B) . Example of (C) (D) . 2.2.6 Statistical Analysis Data analyses for all studied variables were conducted using the mixed model approach implemented in the PROC MIXED procedure of SAS Version 9.4 (SAS Inc., 2009). For comparisons of the geostatistical characteristics and total below image resolution porosity , the statistical model consisted of the fixed effect of ma nagement and the random effect of aggregates nested within management. For investigation of pore type and management effects on the grayscale gradients , the statistical model consisted of the fixed effects of management, pore 31 types, distances from the pores, and their interactions, as well as the random effect of aggregates nested within management and the random effect of individual pores nested within respective pore types and aggregates. In this analysis, distance was treated as a repeated measure fa ctor. Comparisons among pore types for the Os gradients were conducted using the statistical model with the fixed effects of pore type, distance, and their interaction and the random effects of soil core and soil core by pore type interaction. As with the grayscale gradients, the Os gradients distance factor was treated as a repeated measure. Comparisons among managements and pore types for the gradient influence distance were evaluated using the statistical model with the fixed effects of management, pore type, and the interaction s between them and the random effect of aggre gates nested within management. The normality of residuals in all analyses was visually assessed using normal probability plots and stem -and -leaf plots, while equal variances assumption was assessed using Levene™s For all analyses, if the interactions were not significant, pairwise comparisons of the main effects using the LSMEANS statement were used. In the case whe re interactions were significant, slicing using the LSMEANS statement was employed. T-tests were conducted to determine if the mean values differed from zero. 2.3 Results 2.3.1 Geostatistical analysis of grayscale spatial patterns Biologically based and early successional managements had greater overall spatial variability in grayscale values of the solid material than the conventional management system, as indicated by their higher sill values and histogram variance (Table 2.1). Spatial variations at dista nces less than the image resol than 50% of the overall variability, as indicated by nugget -to-sill ratios ranging from 39% in biologically based 32 management to 46 and 48% in early successional and conventional managements. Th e three managements did not differ in terms of their nugget and range values, indicating similarities in present. Table 2.1: Characteristics of the variograms of soil material and variance of the histograms containing no >13 µm pores in three studied land use and management practices. Shown are means and standard errors (in parentheses) calculated based on a total of 157 subsect ion cubes from 32 aggregates. Different letters within each row denote Variogram Characteristic Conventional Biologically Based Early Successional 334.1(1.4)a 309.4(1.4)a 297.7(1.4)a Nugget 313(26)a 321(26)a 357(28)a Sill 641(45)a 822(45)b 778(47)b Nugget to Sill 0.48(0.02)a 0.39(0.02)b 0.46(0.02)a Variance 637(45)a 817(45)b 782(47)b 2.3.2 Os levels as a function of distance from soil pores The Os gradients were markedly different in pores of non -biological and biological origin (Figure 2.6). However, Os gradients did not differ between biological pores associated with POM -NS and POM -Root. Pores of biological origin had a large increase in Os labeled SOM immediate ly adjacent to the pores, which then slowly declined until returning to background levels at distances of 100 - -biological pores, on the other hand, had levels of Os labeled SOM that were statistically lower than background levels (P=0.045) at di stances up to 30 33 Figure 2. 6: (A) Mean difference from the background level for Os stained samples as a function of distance from pores of biological origin with plant roots (POM -Root) and with non -root de rived POM (POM -NS) and from non -biological pores (n=6). The samples are from the biologically based system. Positive values indicate increased presence of Os labeled SOM, while negative values indicate a decrease in Os labeled SOM. (B) Normalized grayscale values for all three studied pore types from all three management practices. Dots are averages and the standard errors are equal to the size of the dots at each distance. The solid black lines represent the background Os labeled level (on (A) ) and the bac kground grayscale value (on (B) ). 2.3.3 Grayscale levels as a function of distance from soil pores The grayscale gradients for all managements and pore types had d ecreased grayscale values adjacent to the p ores and then increased as distance from the pore increased (Figure 2.6). 34 This was similar to the biological pore results from the Os gradients. For biological pores, the grayscale gradients matched the Os gradients almost identically. When both grayscale and Os gradients for POM -NS and POM -Root in biologically based management were normalized to the average grayscale value (Figure 2.7), the overlap between the relationships was almost perfect, indicating that Os and grayscale value gradients were equivalent. Figure 2. 7: Normalized Os values from Os stained samples (n=6) and grayscale values (n=96) from non -stained samples as a function of distance from POM -NS and POM -Root pores in the soil of the biologically based management. The grayscal e levels in the solid material adjacent to non -biological pores were lower than the background grayscale values (average grayscale value of the whole aggregate). However, they increased much faster with increasing distance (Figure 2.6) and reached the back ground levels at much shorter distances than those of biological pores (Table 2.2). For 35 POM -NS and POM - - Table 2.2: Effective dista nce of pore influence (EDPI) for the three studied pore types averaged across all studied aggregates. Means were calculated based on 32 aggregates with 3 POM -NS, 3 POM -Root, and 5 non -biological pores from each aggregate. Standard errors are shown in paren theses. Different letters denote significant differences Pore Type Non -biological 74.2(5.0)a POM -Root 123.3(6.2)b POM -NS 122.7(6.2)b While no differences were observed between POM -NS and POM -Root grayscale gradients for biologically based and early successional managements (P=0.42, 0.23), POM -NS retained decreased grayscale values over longer distances than POM -Root in conventional mana gement (P<0.001) (Figure 2.8). No significant differences were observed between managements for non -biological and POM -Root pores (P=0.85, 0.36), but conventional management again showed a shallower POM -NS grayscale gradient than the other managements (P=0 .0096). EDPI only varied by pore type, indicating that management had no effect on this characteristic of grayscale value distributions (Table 2.2). 36 Figure 2. 8: Normalized grayscale values as a function of distance from POM -NS p ores for the three management practices (n=96). Error bars represent standard errors. The solid black line represents the background grayscale value, i.e. the average grayscale value of the entire aggregate. 2.3.4 Pores below image resolution While t he > differed among the managements (P=0.044 ), the below image resolution porosity (2 - based management than the conventional management, while the early successional management was not significantly different from either management. 37 Figure 2 9: Percentage of 2 - agements. Letters indicate - not statistically significant. Error bars represent standard errors. 2.4 Discussion 2.4.1 Image grayscale values as a proxy for SOM patterns My findings indicate that g rayscale values can be a useful proxy for SOM, however, caution needs to be exercised in such use. Specifically, in the studied soil , a reliable correspondence between SOM gradients as determine d via Os st aining and grayscale value gradients was achieved only for pores of biological origin. As can be seen from Figure 2.7, the Os and grayscale gradients corresponded to each other remarkably well , indicating that grayscale gradients can be used as a suitable proxy when exploring SOM patterns near pores of biological origin. The EDPI distributions in soil of 38 - 38 analyses, although these analyses were not correlated to pores specifically ( Rawlins et al, 2016 ). Previous studies utilizing isotopically labeled materials have reported movement of decomposition products as far as 5 -10 mm from carbon sources during soil in cubations ( Gaillard et al, 1999 , 2003; Toosi et al, 2017 ). However, these studies do not specifically measure transport of DOM from individual pore, but overall transport of isotopic labeled materials from its source, which would account for the larger tra nsport ranges seen in previous studies. Direct imaging of the spatial distribution of SOM near individual pores has previously been achieved using NanoSIMS, however, at spatial scales much lower (nm) than those used in this study. Mueller et al (2012) show from biological pores into the soil matrix in similarly textured soils. For non -biological pores, Os staining and grayscale value gradient trends did not match. The grayscale levels increased with the distance from the pore, which could be interpreted as increased SOM concentrations near the pore. Yet, the Os gradients clearly indicate lower SOM levels in immediate vicinity (<30 micron) of non-biological pores (Figure 2.6) . This result was -biological origin pores. 39 Figure 2. 10: Percentage of 2 - (POM -Root) and with non -root derived POM (POM -NS) and from non -biological pores (n=8). Error bars are standard errors at each distance. A possible explanation for this discrepancy is an artifact of image processing via 3DMA. 3DMA uses indicator kriging as a thresholding method, while the distance measures were conducted using Otsu™s method due to computational and time limits resulting from the smaller image resolution. Indicator kriging performs well for identifying pores well above the image resolution, but may misidentify pores of sizes at or only slightly larger than the image resolution (Figure 2.11). Thus, the decreases in the grayscale values might be due to such missed porosity in non -biological pores. For non -biological pores, indicator kriging fails to identify small visible connections between larger pores that extend for several voxels between adjacent pores due to using the surrounding voxels to help det ermine if a voxel is pore or not, while Otsu™s method correctly identifies these pores because it only uses the raw grayscale value to identify pores . 40 This artifact is less pronounced in biological pores as biological pores have no such small connections a nd, therefore, the true extent of the pore was accounted for. Figure 2. 11: Examples from the three different managements of how 3DMA missed pore material adjacent to non -biological pores, but identifies POM -Root and POM -NS pores correctly. (A) is from the conventional management, (B) is from the biologically based, and (C) is from the early successional. The blue outlines are the pores identified by 3DMA. Red arrows on each figure indicate an example of missed porosity on each fi gure. The reason for this artifact to be present in pores of non -biological but not biological origin can be explained by the processes that create the pores of these two different types. Biological pores are created through the radial compression of the surrounding matrix as a root 41 or macrofauna pushes through the soil and is then supported by organic binding agents, such as mucilage , mucus , and large amounts of DOM from decomposing organic matter , resulting in a clear boundary between pore and solid mat erial ( Gray and Lissmann, 1938 ; Greacen et al 1972 ; Greacen and Sands, 1980 ; Czarnes et al , 2000; Ruiz et al, 2017 ). Additionally, Helliwell et al (2017) found that while porosity near roots increased initially in sandy soils, which is similar to the texture of these soils, after eight days of growth, porosity was found to decrease adjacent to roots at this image resolu tion. Non -biological pores are created through the shrinking and expanding of clays, resulting in neither clear nor stable boundar ies (Peng et al, 2007 ). However, using a different thresholding method may overcome the artifact effect. 2.4.2 Spatial pattern s of grayscale values The nugget to sill ratio values of the studied samples indicate that approximately 50 -60% distances. This matches the porosity data, where approximately 50% of the porosity occurs at and the highest nugget to sill ratio. Since nugget to sill ratio indicates the relative amount of spatial dependence at voxel size, this may indicate a connection between spatial dependence and porosity at image scale. This would support my hypothesis that spatial variability in the solid material of similar mineralogy is driven by pores. However, more research would be necessary to confirm this connection. The lack of difference in spatial auto correlation ranges among the aggregates from the three management s was surprising. Til lage, utilized both in conventional and biologically based management, homogenize s the soil , which, according to my expectations and previous findings (Garrett 2009 ), should be manifested in greater spatial correlation range values . Lack of such an 42 effect in my samples may indicate that the spatial correlation range in this soil is controlled more by the inherent mineralogy and/or texture, which are similar for the soil of all three managements, than by the management -driven SOM differences . The EDPI was mu ch smaller the SOM distribution was not a driver of the spatial correlation range values. Rawlins et al (2016) investigated the spatial ranges of SOM, pores, While the spatial ranges for SOM (38 - my results. Observed similarity in the nugget values fr om the three studied managements was in agreement with the results of Nunan et al (2006) , who found similar nuggets between three different amendment managements. The lack of differences between nuggets corr oborates the below image resolution porosity meas urements where the below image resolution porosity was similar between all three managements (Figure 2.9). The differences observed in the overall porosity match ed results previously reported for these aggregates with early succession al conventional (Kravchenko et al, 2015 ). Greater overall spati al variability in less intense management practices, i.e., biologically based and early successional managements, manifested via greater sill values (Table 2.1), is likely a result of managem ent -induced changes in SOM. As mentioned above, soil mineralogy and texture of the studied managements were very similar, as well as their below image resolution porosity values (Figure 2.9 & 2.10). Yet, after almost 20 years of implementation, the biologi cally based and early successional management practices resulted in higher SOM than the conventionally managed practice ( Paul et al, 1999 ; Senthilkumar et al, 2009 ). Observed greater variability in grayscale values of biologically based and early successio nal managements 43 suggests that these SOM inputs were not uniformly distributed. This assertion is supported by previous findings of Ananyeva et al (2013) who reported greater variability in soil carbon within the macro -aggregates from early successional man agement as compared to conventional management practice; while Feeney et al (2006) observed that active biota, in particular roots, increased spatial correlation of soil pores. Spatial gradients in SOM associated with pores of biological origin is one poss ible mechanism contributing to the increased variability. The increased spatial variability observed in the grayscale values of the biologically based and early successional managements, if driven by SOM distributions, would indicate increased occurrence of different microenvironments of either increased or decreased amounts of SOM, while conventional management would have less of these differing microenvironment s. Greater presence of biological pores may result in an increased diversity of microenvironmental conditions, including different levels of microbial accessibility, nutrient availability, and potentially water and gas fluxes. Such microenvironmental diffe rences affect microbial activities (Ekschmitt et al , 2005; Ekschmitt, 2008 ; Kravchenko and Guber, 2017 ); and greater SOM decomposition can be expected in microenvironments conducive to high microbial activity, while SOM protection in microenvironments wher e microbial activity is reduced. This increase in microenvironment heterogeneity, and therefore, greater presence of microsites where SOM might not be available to microbial decomposers and/or reduction of microbial decomposition due to anaerobic microsite s (Keiluweit et al, 2016, 2017 ), may be reflective of the increased carbon protection/sequestration observed in the biologically based and early successional managements as compared to conventional management (Paul et al, 1999; Senthilkumar et al, 2009). 44 2.4.3 SOM pattern in relation to soil pores My Os results demonstrated that pores were the drivers of SOM's spatial variability in the studied soil (Figure 2.6). Biological pores had a clear spatial gradient of SOM with highest levels in the vicinity of po res and decreasing when moving into the surrounding solid material. Biological pores are observed more frequently in biologically base and early successional manageme nts. Non -biological pores had a small decrease in SOM adjacent to the pores, but otherwise their presence was not related to SOM distribution patterns. Non -biological pores are observed more frequently in conventional management as non -biological pores ten d to be 40 -90 the decrease in SOM observed adjacent to non -biological pores. In the studied soil, <13 µm pores (below image resolution sizes) are water filled dur ing most of the year. This would hamper the diffusion of oxygen and lead to dominance of anaerobic conditions, which can result in as much as a 10 -fold decrease in decomposition rates (Keiluweit et al, 2017). The likely outcome is, thus, organic matter 's d ecomposition near large (20 -300µm) pore boundaries, where oxygen is available, and organic's preservation in anaerobic (<13 µm pore) zones. 2.4.4 Effect of management practices on SOM pattern in relation to soil pores The effect of management on the SOM gradients as inferred from the grayscale gradient was unanticipated. T here were no differences among managements observed for non -biological and POM -Root pores. However, there were observed differences between the managements in POM -NS pores. Biologically based and early successional managements had no differences between POM -Root and POM -NS pores. Conversely, in conventional management, POM -NS gradients tended to retain decreased grayscale values over longer distances. I inferred that this decrease is rela ted to an increase in SOM content. A possible explanation is that, per visual 45 observations, POM -NS within conventional soil aggregates tended to be located closer to the -NS in biologically based and early successional managements were located closer to the aggregate restriction of microorganism, water, and oxygen access to POM -NS, re sulting in incomplete decomposition. The incomplete decomposition produces decomposition products of a more hydrophobic nature. This hydrophobicity would decrease the ability of these products to be transported by water, resulting in a build -up of organic matter closer to the pore. Toosi et al (2017) observed that as maximum pore size decreases the presence of SOM compounds with fewer oxygen functional groups and greater aromaticity increases; this observation supports my increased hydrophobicity explanatio n. 2.5 Conclusion Analysis of grayscale gradients near pores of biological origin were found to be a useful proxy for assessing SOM spatial distribution patterns at micro -scale. Grayscale gradients of non -biological pores, in contrast, were fo und to not be a useful proxy for SOM due to a pore identification artifact. Utilizing a different thresholding method may overcome this limitation. Os and grayscale value gradients indicate increased SOM concentrations adjoining biological pores, decreasing to backgrou nd levels as distance from the pore increases. The average distance SOM concentrations decreased in the direct vicinity of non -biological pores th en returned to the background levels. The average distance of negative influence of non -biological pores on SOM Soil material without >13 µm pores was more variable in its grayscale values in biologically based and early successional management than in conventional management 46 practice. The greater variability is likely to be driven by SOM spatial distribution patterns, which reflect presence of soil pores, especially, pores of biological origin. This spatial variability likely results in greater variabil ity of microenvironmental conditions for microbial functioning with possible implications for soil carbon protection. Funding Support for this research was provided in part by the USDA -NIFA, Award No. 2016 -67011-24726 fiUsing stable isotopes and computer to mography to determine mechanisms of soil carbon protection in cover crop based agricultural systemsfl and by the US National Science Foundation Long -Term Ecological Research Program (DEB 1027253) at the Kellogg Biological Station and by Michigan State Unive rsity AgBioResearch. Portions of this work were performed at GeoSoilEnviroCARS (The University of Chicago, Sector 13), Advanced Photon Source (APS), Argonne National Laboratory. GeoSoilEnviroCARS is supported by the National Science Foundation - Earth Scie nces (EAR - 1634415) and Department of Energy - GeoSciences (DE -FG02 -94ER14466). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE -AC02 -06CH11357. 47 REFERENCES 48 REFERENCES Ananyeva, K., Wang, W., Smucker, A. J. M., Rivers, M. L., and Kravchenko, A. N. (2013). Can intra -aggregate pore structures affect the aggregate™s effectiveness in protecting carbon? Soil Biol. Biochem. 57, 868-875. doi: 10.1016/j.soilbio.2012.10.019 Aravena, J. E., Berli, M., Ruiz, S., Su árez, F., Ghezzehei, T. A., and Tyler, S. W. (2014). Quantifying coupled deformation and water flow in the rhizosphere using X -ray microtomography and numerical simulations. Plant Soil . 376, 95-110. doi: 10.1007/s11104 -013-1946-z Bailey, V. L., Smith, A. P. , Tfaily, M., Fansler, S. J., and Bond -Lamberty, B. (2017). Differences in soluble organic carbon chemistry in pore waters sampled from different pore size domains. Soil Biol. Biochem. 107, 133-143. doi: 10.1016/j.soilbio.2016.11.025 Bengough, A. G., McKen zie, B. M., Hallett, P. D., and Valentine, T. A. (2011). Root elongation, water stress, and mechanical impedance: a review of limiting stresses and beneficial root tip traits. J. Exp. Bot. 62, 59-68. doi: 10.1093/jxb/erq350 Chun, H. C., Gim énez, D., and Yo on, S. W. (2008). Morphology, lacunarity and entropy of intra -aggregate pores: aggregate size and soil management effects. Geoderma . 146, 83-93. doi: 10.1016/j.geoderma.2008.05.018 Czarnes, S., Hallett, P. D., Bengough, A. G., and Young, I. M. (2000). Root - and microbial -derived mucilages affect soil structure and water transport. Eur. J. Soil Sci. 51, 435-443. doi: 10.1046/j.1365 -2389.2000.00327.x De Gryze, S., Jassogne, L., Six, J., Bossuyt, H., Wevers, M., and Merckx, R. (2006). Pore structure changes du ring decomposition of fresh residue: X -ray tomography analyses. Geoderma . 134, 82-96. doi: 10.1016/j.geoderma.2005.09.002 Denef, K., Six, J., Bossuyt, H., Frey, S. D., Elliot, T. E., Merckx, R., et al. (2001). Influence of dry -wet cycles on the interrelati onship between aggregate, particulate organic matter, and microbial community dynamics. Soil Biol. Biochem. 33, 1599-1611. doi: 10.1016/S0038 -0717(01)00076 -1 Dungait, J. A. J., Hopkins, D. W., Gregory, A. S., and Whitmore, A. P. (2012). Soil organic matter turnover is governed by accessibility not recalcitrance. Global Change Biol. 18, 1781-1796. doi: 10.1111/j.1365 -2486.2012.02665.x Ekschmitt, K., Kandeler, E., Poll, C., Brune, A., Buscot, F., Friedrich, M., et al. (2008). Soil -carbon preservation through habitat constraints and biological limitations on decomposer activity. J. Plant Nutr. Soil Sci. 171, 27-35. doi: 10.1002/jpln.200700051 49 Ekschmitt, K., Liu, M., Vetter, S., Fox, O., and Wolters, V. (2005). Strategies used by soil biota to overcome soil orga nic matter stability Œ why is dead organic matter left over in soil? Geoderma. 128, 167-176. doi: 10.1016/j.geoderma.2004.12.024 Falconer, R. E., Battaia, G., Schmidt, S., Baveye, P., Chenu, C., and Otten, W. (2015). Microscale heterogeneity explains experimental variability and non -linearity in soil organic matter mineralization. PloS One . 10(5) : e0123774 . doi: 10.1371/journal.pone .0123774 Feeney, D. S., Crawford, J. W., Daniell, T., Hallett, P. D., Nunan, N., Ritz, K., et al. (2006). Three -dimensional microorganization of the soil -root -microbe system. Microb. Ecol. 52, 151-158. doi: 10.1007/s00248 -006-9062-8 Gaillard, V., Chenu, C. , and Recous, S. (2003). Carbon mineralization in soil adjacent to plant residues of contrasting biochemical quality. Soil Biol. Biochem. 35, 93-99. doi: 10.1016/S0038 -0717(02)00241 -9 Gaillard, V., Chenu, C., Recous, S. and Richard, G. (1999). Carbon, nitr ogen and microbial gradients induced by plant residues decomposing in soil. Eur. J. Soil Sci. 50, 567 -578. doi: 10.1046/j.1365 -2389.1999.00266.x Garrett, R. G. (2009). Relative spatial soil geochemical variability along two transects across the United Stat es and Canada. Appl. Geochem. 24, 1405 -1415. doi: 10.1016/j.apgeochem.2009.04.011 Gibson, J. R., Lin, H., and Bruns, M. A. (2006). A comparison of fractal analytical methods on 2- and 3 -dimensional computed tomographic scans of soil aggregates. Geoderma . 134, 335-348. doi: 10.1016/j.geoderma.2006.03.052 Greacen, E. L., and Oh, J. S. (1972). Physics of root growth. Nat. New Biol. 235, 24-25. doi: 10.1038/newbio235024a0 Greacen, E. L., and Sands, R. (1980). Compaction of forest soils. a review. Aust. J. So il. Res. 18, 163-189. doi: 10.1071/SR9800163 Gray, J., and Lissmann, H. W. (1938). An apparatus for measuring the propulsive forces of the locomotory muscles of the earthworm and other animals. J. Exp. Biol. 15, 518-521. Helliwell, J. R., Sturrock, C. J., Mairhofer, S., Craigon, J., Ashton, R. W., Miller, A. J., et al. (2017) . The emergent rhizosphere: imaging the development of the porous architecture at the root -soil interface. Sci. Rep. 7: 14875 . doi:10.1038/s41598 -017-14904-w Keiluweit, M., Nico, P. S., Kleber, M., and Fendorf, S. (2016). Are oxygen limitations under recognized regulators of organic carbon turnover in upland soils? Biogeochemistry . 127, 157-171. doi: 10.1007/s10533 -015-0180-6 Keiluweit, M., Wanzek, T., Kleber, M., Nico, P., and Fendorf, S. (2017). Anaerobic microsites have an unaccounted role in soil carbon stabilization. Nat. Commun. 8: 1771 . doi: 10.1038/s41467 -017-01406-6 50 Kiem, R., and Kögel -Knabner, I. (2002). Refractory organic carbon in particle -size fractions of arable soils II: organic carbon in relation to mineral surface area and iron oxides in fractions <6 Org. Geochem. 33, 1699 -1713. doi: 10.1016/S0146 -6380(02)00112 -2 Ketcham, R. A. (2005). Three -dimensional grain fabric measurements using high -resolution X -ray computed tomography. J. Struc. Geol . 27, 1217 -1228. doi: 10.1016/j.jsg.2005.02.006 Kravchenko, A. N ., and Guber, A. K. (2017). Soil pores and their contributions to soil car bon processes. Geoderma . 287, 31 -39. doi: 10.1016/j.geoderma.2016.06.027 Kravchenko, A. N., Negassa, W. C., Guber, A. K., Hildebrandt, B., Marsh, T. L., and Rivers, M. L. (2014b). Intra -aggregate pore structure influences phylogenetic composition of bacter ial community in macroaggregates. Soil Sci. Soc. Am. J. 78, 1924-1939. doi: 10.2136/sssaj2014.07.0308 Kravchenko, A. N., Negassa, W. C., Guber, A. K., and Rivers, M. L. (2015). Protection of soil carbon within macro -aggregates depends on intra -aggregate po re characteristics. Sci. Rep. 5:16261 . doi: 10.1038/srep16261 Kravchenko, A. N., Negassa, W., Guber, A. K., and Schmidt, S. (2014a). New approach to measure soil particulate organic matter in intact samples using X -ray computed microtomography. Soil Sci. S oc. Am. J. 78, 1177-1185. doi: 10.2136/sssaj2014.01.0039 Kravchenko, A. N., Wang, W., Smucker, A. J. M., and Rivers, M. L. (2011). Long -term differences in tillage and land use affect intra -aggregate pore heterogeneity. Soil Sci. Soc. Am. J. 75, 1658-1666. doi: 10.2136/sssaj2011.0096 Kuzyakov, Y., and Blagodatskaya, E. (2015). Microbial hotspots and hot moments in soil: concept & review. Soil Biol. Biochem. 83, 184-199. doi: 10.1016/j.soilbio.2015.01.025 Mooney, S. J., Pridmore, T. P., Heliwell, J., and Ben nett, M. J. (2012). Developing X -ray computed tomography to non -invasively image 3 -D root systems architecture in soil. Plant Soil . 352, 1-22. doi: 10.1007/s11104 -011-1039-9 Mueller, C. W., K ölbl, A., Hoeschen, C., Hillion, F., Heister, K., Herrmann, A. M., et al. (2012). Submicron scale imaging of soil organic matter dynamics using NanoSIMS Œ from single particles to intact aggregates. Org. Geochem. 42, 1476-1488. doi: 10.1016/j.orggeochem.201 1.06.003 Negassa, W. C., Guber, A. K., Kravchenko, A. N., Marsh, T. L., Hildebrandt, B., and Rivers, M. L. (2015). Properties of soil pore space regulate pathways of plant residue decomposition and community structure of associated bacteria. PLoS One . 10(4): e0123999 doi: 10.1371/journal.pone.0123999 Nunan, N., Ritz, K., Rivers, M., Feeney, D. S., and Young, I. M. (2006). Investigating microbial micro -habitat structure using X -ray computed tomography. Geoderma . 133, 398 -407. doi: 10.1016/j.geoderma.2005.08. 004 51 Oades, J. M. (1984). Soil organic matter and structural stability: mechanisms and implications for management. Plant Soil. 76, 319-337. doi: 10.1007/BF02205590 Oh, W., and Lindquist, W. B. (1999). Image thresholding by indicator kriging. IEEE Trans. Pattern Anal. Mach. Intell. 21, 590-602. doi: 10.1109/34.777370 Papadopoulos, A., Bird, N. R. A., Whitmore, A.P., and Mooney, S. J. (2009). Investigating the effects of organic and conventional management on soil aggregate stability using X -ray computed tomo graphy. Eur. J. Soil Sci . 60, 360-368. doi: 10.1111/j.1365 -2389.2009.01126.x Paul, E. A., Harris, D., Collins, H. P., Schulthess, U., and Robertson, G. P. (1999). Evolution of CO2 and soil carbon dynamics in biologically managed, row -crop agroecosystems. Appl. Soil Ecol. 11, 53-65. doi: 10.1016/S0929 -1393(98)00130 -9 Paul, E. A., Kravchenko, A., Grandy, S., and Morris, S. (2015). fiSoil organic matter dynamics: controls and management for sustainable ecosystem functioning,fl in The ecology of agricultural land scapes: long -term research on the path to sustainability , eds. S. K. Hamilton, J. E. Doll, and G. P. Robertson, (New York, New York, USA: Oxford University Press), 104 -134. Pebesma, E. J. (2004). Multivariable geostatistics in S: the gstat package. Comput. Geosci. 30, 683-691. doi: 10.1016/j.cageo.2004.03.012 Peng, X., Horn, R., and Smucker, A. (2007). Pore shrinkage dependency of inorganic and organic soils on wetting and drying cycles. Soil Sci. Soc. Am. J. 71, 1095-1104. doi: 10.2136/sssaj2006.0156 Peth, S. (2010). fiApplications of microtomography in soils and sediments,fl in Developments in Soil Science, Vol. 34 , eds. B. Singh and M. Gr äfe, (The Netherlands: Elsevier B. V.), 73 -101. Peth, S., Chenu, C., Leblond, N., Mordhorst, A., Garnier, P., Nunan, N., et al. (2014). Localization of soil organic matter in soil aggregates using synchrotron -base X -ray microtomography. Soil Biol. Biochem. 78, 189-194. doi: 10.1016/j.soilbio.2014.07.024 Peth, S., Horn, R., Beckmann, F., Donath, T., Fischer, J., and Smucker, A. J. M. (2008). Three -dimensional quantification of intra -aggregate pore -space features using synchrotron -radiation -based microtomography. Soil Sci. Am. J . 72, 897-907. doi: 10.2136/sssaj2007.0130 Rasse, D. P. and Smucker, A. J. M. (1998). Root recolonization of previous root channels in corn and alfalfa rotations. Plant Soil 204, 203-212. doi: 10.1023/A:1004343122448 Rawlins, B. G., Wragg, J., Reinhard, C., Atwood, R. C., Hou ston, A., Lark, R. M., et al. (2016). Three -dimensional soil organic matter distribution, accessibility and microbial respiration in macroaggregates using osmium staining and synchrotron X -ray computed tomography. SOIL . 2, 659-671. doi: 10.5194/soil -2-659-2016 Ruiz, S., Schymanski, S. J., and Or, D. (2017). Mechanics and energetics of soil penetration by earthworms and plant roots: higher rates cost more. Vadose Zone J. 16:8. doi: 10.2136/vzj2017.01.0021 52 SAS Inc. (2009). SAS user™s guide. Vers ion 9.2 . Cary, NC: SAS Inst. Senthilkumar, S., Basso, B., Kravchenko, A. N., and Robertson, G. P. (2009). Contemporary evidence of soil carbon loss in the U.S. corn belt. Soil Sci. Soc. Am. J. 73, 2078-2086. doi: 10.2136/sssaj2009.0044 Six, J., Conant, R. T., Paul, E. A ., Paustian, K. (2002). Stabilization mechanisms of soil organic matter: implications for C -saturation of soils. Plant Soil . 241, 155-176. doi: 10.1023/A:1016125726789 Six, J., Elliott, E. T., and Paustian, K. (2000). Soil macroaggregate turnover and micro aggregate formation: a mechanism for C sequestration under no -tillage agriculture. Soil Biol. Biochem. 32, 2099-2103. doi: 10.1016/S0038 -0717(00)00179 -6 Smucker, A. J. M., Park, E. -J., Dorner, J., and Horn, R. (2007). Soil micropore development and contrib utions to soluble carbon transport within macroaggregates. Vadose Zone J. 6, 282-290. doi: 10.2136/vzj2007.0031 Sørensen, L. H. (1974). Rate of decomposition of organic matter in soil as influenced by repeated air drying -rewetting and repeated additions of organic material. Soil Biol. Biochem. 6, 287-292. doi: 10.1016/0038 -0717(74)90032 -7 Syswerda, S. P., Corbin, A. T ., Mokma, D. L., Kravchenko, A. N., and Robertson, G. P. (2011). Agricultural management and soil carbon storage in surface vs. deep layers. Soil Sci. Soc. Am. J. 75, 92-101. doi: 10.2136/sssaj2009.0414 Toosi, E. R., Kravchenko, A. N., Mao, J., Quigley, M. Y., and Rivers, M. L. (2017). Effects of management and pore characteristics on organic matter composition of macroaggregates: evidence from characterization of organic matter and imaging. Eur. J. Soil Sci. 68, 200-211. doi: 10.1111/ejss.12411 Wang, W., K ravchenko, A. N., Johnson, T., Srinivasan, S., Ananyeva, K. A., Smucker, A. J. M., et al. (2013). Intra -aggregate pore structures and Escherichia coli distribution by water flow within and movement out of soil macroaggregates. Vadose Zone J. 12:4. doi: 10.2136/vzj2013.01.0012 Wang, W., Kravchenko, A. N., Smucker, A. J. M., Liang, W., and Rivers, M. L. (2012). Intra -aggregate pore characteristics: X -ray computed microtomography analysis. Soil Sci. Soc. Am. J. 76, 1159-1171. doi: 10.2136/sssaj2011.0281 Wang, W., Kravchenko, A. N., Smucker, A. J. M., and Rivers, M. L. (2011). Comparison of image segmentation methods in simulated 2D and 3D microtomographic images of soil aggregates. Geoderma . 162, 231 -241. doi: 10.1016/j.geoderma.2011.01.006 Young, I. M., and Cr awford, J. W. (2004). Interactions and self -organization in the soil -microbe complex. Science . 304, 1634 -1637. doi: 10.1126/science.1097394 53 Chapter 3: Influence of Pore Characteristics on the Fate and Distribution of Newly Added Carbon Abstract Pores cr eate a transportation network within a soil matrix, which controls the flow of air, water, nutrients, and movement of microorganisms .The flow of air, water, and movement of the visualization of pore structure at micron scale, but quantitative information on contribution of pores to the fate and p rotection of soil carbon, essential for modeling, is still lacking. I used the natural difference between carbon isotopes of C3 and C4 plants to determine how the presence of pores of different sizes affects spatial distribution patterns of newly added car bon immediately after plant termination and then after one -month incubation. I considered two contrasting soil structure scenarios: soil with the structure kept intact and soil for which the structure was destroyed via sieving. For the experiment, soil was collected from 0 -15 cm depth at a 20 -year continuous maize ( Zea mays L., C4 plant) experiment into which cereal rye ( Secale cereale L., C3 plant) was planted. Intact soil fragments (5 -6 mm) were procured after 3 months rye growth in a greenhouse. Pore cha 13 15 N were measured. The results indicate that, prior to incubation, greater presence of 40 - with higher levels of C3 carbon, pointing to the positive role of these pores in transport of new C inputs. Nevertheless, after incubation, the association became negative, indicating greater losses of newly added C in such pores. These trends were statis tically significant in destroyed -structure soil and numerical in intact -structure soil. In soils of intact -structures, after incubation, higher Originally published as: Quigley, M. Y., Negassa, W. C., Guber, A. K., Rivers, M. L., a nd Kravchenko, A. N. (2018a). Influence of pore characteristics on the fate and distribution of newly added carbon. Front. Environ. Sci. 6:51. doi: 10.3389/fenvs.2018.00051 54 levels of total carbon were associated with greater abundance of 6.5 -15 and 15 - indicating a lower c arbon loss associated with these pores. The results indicate that, in the studied soil, pores of 40 - n. 3.1 Introduction Soils contain twice as much carbon as the atmosphere and have the potential to store even more, especially in agricultural soils (Lal, 1999; Swift, 2001; Dungait et al, 2012; Kell, 2012). Soil carbon content is an important component of soil fertility as it drives several defining criteria of soil quality, including cation exchange capacity, soil aggregation, and water holding capacity (Dou et al, 2014). This makes utilization of agricultural management practices that increase and/or con serve soil carbon vitally important to sustainability (Grandy and Robertson, 2007). One such practice is the utilization of cover crops, a crop that is planted between main crops for the purpose of increasing water infiltration, stabilizing the soil surfac e, preventing erosion, decreasing weeds, and increasing soil fertility. The a ctivity of cover crop roots may benefit the physical protection of new carbon inputs. Physical protection of soil carbon occurs when physical disconnections separate decomposers f rom carbon sources (Dungait et al, 2012). This disconnect is not limited to access of decomposers and their enzymes to soil carbon, but also includes availability of other components necessary for decomposition, such as oxygen and water (Schmidt et al, 201 1, Keiluweit et al, 2017 ). Long -term cover crop based management increases soil aggregation (Tiemann and Grandy , 2015), and soil carbon is known to be better protected within soil aggregates (Six et al, 2000; Grandy and Robertson, 2007). Yet, mechanistically, it is the soil pore -space that controls not only movement of soil microbes, but also air and water fluxes and transport of nutrients 55 necessary for decomposition (DeGryze et al , 2006; Negassa et al , 2015; Young a nd Crawford, 2004; ). Pores within the soil matrix serve as planes of breakage along which the aggregates form; and their sizes and spatial positions not just define soil aggregate -size distributions but determine micro -environmental conditions driving phy sical carbon protection within the aggregates (Young et al, 2001; Ekschmitt et al , 2005, 2008; Kravchenko and Guber, 2017 ; Rabot et al, 2018 ). Pores of different sizes have different origins, connectivities, accessibilities, and hydraulic properties (Park and Smucker, 2005; Peth et al, 2008) . As pore size decreases, higher suction is required to drain the pore . This means that while pores of >10 m sizes may only require gravity to fully or partially drain, under normal soil moisture regimes, pores <10 m remain water filled (Marshall et al, 1996) . Plant root diameters are typically > roots can only access and/or create pores exceeding that size (Wiersum, 1957; Cannell, 1977). Root pores are formed by compressing the soil matrix radia lly as the root pushes through the soil and then their walls are stabilized through mucilage and often refilled by roots of the next generation crops (Gray and Lissmann, 1938; Greacen and Oh 1972; Greacen and Sands, 1980; Rasse and Smucker, 1998; Czarnes et al , 2000; Ruiz et al, 2017). Fungal hyphae are known to create pores of 20 - by pushing aside silt particles and exuding binding agents to buttress the pores (Dorioz et al, 1993; Bearden, 2001, and Emerson and McGarry, 2003). However, fungi are typically excluded Roots provide carbon into the soil system in two ways: as a source of biomass when they die and as a source of easily decomposable carbon via root exudates. Roots tend to consist of more diff icult to decompose molecules (such as lignin and tannin), which, in addition to being 56 harder to decompose, are easier to adsorb to mineral surfaces, sequestering them (Rasse et al, 2005; Jackson et al, 2017). Root exudates, on the other hand, tend to be sm all , soluble , and easily decomposable materials, such as organic acids, carbohydrates, and amino acids (Dungait et al, 2012) or water insoluble materials, such as mucilage (Brimecombe et al , 2001). The easily decomposable materials stimulate microbial grow th, which increases decomposition of native soil organic matter ( SOM ) (Kuzyakov and Blagodatskaya, 2015). There is some indication that microorganisms can also stimulate root growth and exudation (Neumann et al, 2014). Agricultural management influences po re size distributions. Wang et al (2012) showed that soil under long -term conventional tillage had more pores of 37.5 - early succession agricultural management Kravchenko et al (2014) foun d that organic management with cover crops had fewer 32- pores and a greater amount of study, the difference in pores from organic cover crop management were attributed to increased root acti vity, while conventional management promoted 32 - wetting/drying cycles. Ananyeva et al (2013) found that higher carbon concentrations were found in sections of soil aggregates with an increased presence of 15 - The presence of 37.5 - was associated with aggregate sections containing less carbon. Stable carbon isotopic signatures can be used to track carbon within a system. Plants preferentially incorporate 12 C into their tissues, but the extent of 12 C inco rporation depends on which metabolic pathway the plant utilizes. Plants that utilize the C3 photosynthetic pathway incorporate more 12 C than plants utilizing the C4 photosynthetic pathway. Therefore, it is possible to differentiate between carbon derived f rom C3 and C4 plants isotopically due to this 57 differences between the 13 C/12 C ratio of the sample compared to a standard: = / 1000 (3.1) 13 13 C values of the original 13 C of a soil reflects the C3/C4 history of the soil (Ehleringer et al, 2000; Bowling et al, 2008 ). Experiments that utilize C3/C4 transitions have been used extensively for determination of soil C turnover rates (Bernoux et al, 1998; Derrien and Amelung, 2011) and for analyses of the fresh carbon input distribution within soi l aggregates (Smucker et al, 2007; Urbanek et al, 2011). The goal of this study was to determine how the abundance of different pore sizes relates to the preservation or loss of newly added carbon. I utilized a C3/C4 natural abundance greenhouse experiment with soil collected from a long term C4 cropping system and planted a C3 plant, cereal rye ( Secale cereale L.), which is one of the most commonly used cover crops in the US Midwest. The first objective of this study was to examine the relationships betwee n newly added carbon and soil pores of different sizes . I used 13 C to fitrackfl newly added C3 carbon and determined pore characteristics via computed microtomography (CT) . The second objective was to examine the relationships between the decomposition of carbon and soil pores sizes after incubating the studied soil. 3.2 Materials and methods 3.2.1 Greenhouse experimental setup Soil for the greenhouse study was obtained in the summer of 2013 from the Living Field Lab (LFL) experiment established in 1993 at W. K. Kellogg Biological Site , Hickory Corners, MI (42°24´N, 85°24´W). The soil is a fine -loamy, mixed mesic Typic Hapludalf (Oshtemo and Kalamazoo series) developed on glacial outwash. Soil was collected from the LFL's conventional 58 management continuous m aize ( Zea mays L) treatment. This treatment has been planted with maize, a C4 plant, and no other crop since 1993. Detailed management and site information is available at http://lter.kbs.msu.edu/Data/LTER and https://lter.kbs.msu.edu/research/long -term -experiments/living -field -lab/ . Six soil blocks of 40 cm x 26 cm x 15 cm size were collected at 0-15 cm depth. Three of the blocks were placed directly into plastic bins with as little disturbance as possible to r etain intact soil structure, and are referred to hereafter as intact -structure treatment. However, I was concerned that , due to the tendency of roots to follow existing pore structure, the root effects generated during out experiment might be masked by the legacy of the existing pores . Therefore, soil from the other three blocks was crushed and sieved through a 1 mm sieve to destroy the existing soil structure, and is referred to hereafter as destroyed -structure treatment. One of the intact soil bins was le ft unplanted as a control, and the remaining bins had cereal rye ( Secale cereale L.) hand planted at a depth of 3 cm and a plant density of ~23.5 plants per m 2. Rye was grown in the greenhouse for 3 months and watered daily to allow for the development of a good stand of rye biomass ; the control bin was watered along with the rest . Soil bulk density was taken in each tub using a 7.5 cm brass ring. After 3 months of rye growth, a pproximately an eighth of the soil was taken from a random side in each bin was removed using a trowel and allowed to air dry. The soil was allowed to break along natural planes of weakness through manual crushing. Intact soil fragments of 5 mm size were hand selected (n=5, 11, and 11 for control, destroyed -structure and intact -struct ure treatments, respectively) for analyses and incubation. Soil fragments were selected based on proximity to rye roots to best determine the effect of rye root growth on the aggregates. Two rye roots per bin were hand collected for isotope analysis from i ntact plants 59 from the soil used for soil fragment selection. Selected intact soil fragments were mounted on top of plastic stands using rubber cement for subsequent scanning and incubation. The experiment and data collection are briefly summarized here and then described in detail in the sections below. First, all intact soil fragments were air - scanning (Section 2.2). Then half of the intact soil fragments were physically cut into ~0.5 -1 mm3 sections, with the physical positions o f the procured section s matching their virtual I 13 15 N, total C (%C), and total N ( %N). These intact soil fragments are hereafter referred to as Pre for preincubation soil. The re maining intact fragments were subjected to a 28 day incubation during which respired CO 2 13 C analysis (Section 2.3). After incubation, the intact soil fragments were re -scanned and also cut into sections, then 13 15 N, total C (%C), and total N ( %N) measurements were taken. These intact soil fragments are hereafter referred to as Post for post incubation soil. 3.2.2 on the bending magnet beam line, station 13 -BM-D of the GeoSoilEnvironCARS (GSECARS) at the Advanced Photon Source (APS), Argonne National Laboratory (ANL), IL. Images were collected with the Si (111) double crystal monochromator tuned to 28 keV incident energy, the distance from sample to source was approximately 55 m, and the X -ray do se is estimated to be 1 kGy. Two -dimensional projections were taken at 0.25° rotation angle steps with a one second exposure and combined into a three -dimensional image cons isting of 1040 slices with 1392 by 1392 pixels per -processed by 60 correcting for dark current and flat field and reconstructed using the GridRec fast Fourier transform reconstruction algorithm (Rivers, 2012). Pore/solid segmentation of the images was conducted using the indicator kriging method in 3DMA -Rock software (Oh and Lindquist, 1999; Wang et al, 2011). Based on the analysis of the segmented data I obtained the total porosity of the intact soil fragments, the total image porosity of each intact fragment was calculated using the dry weight of the fragment and its percent of pore voxels within the total intact soil fragment's voxels. Size distribution of image identified pores was determined using the burn number distribution approach in 3DMA -Rock software (Lindquist et al, 2000, Ananyeva et al, 2013). Briefly, the burn number represents the shortest distance from the pore medial axis to the pore wall. Fo r clarity, burn numbers have been converted into pore diameters. I specifically focused the data analyses on the pores of the following four diameter size ranges: 6.5 - - - sizes were chosen to match pore sizes p reviously studied in macro -aggregates and which demonstrated strong associations with carbon (Wang et al, 2012, 2013; Ananyeva et al, 2013; Kravchenko et al, 2014, 2015). 3.2.3 Incubation experimental set up The soil fragments used in incubation ( Post soil ) consisted of two intact soil fragments from control treatment, 6 intact soil fragments from destroyed -structure treatment, and 5 intact soil fragments from intact -structure treatment. Water was added to the fragments to achieve 60% of water filled pore s pace. The fragments were then placed into 10 ml vacutainers (BD Franklin Lakes NJ, USA) with 1 mL of de -ionized water added to the bottom to maintain high humidity. Incubations were carried out for 28 days at 22.4±0.1°C. CO 2 emission measurements were take n 61 on days 1, 2.5, 4, 8, 13, 19, and 28. Gas samples for isotope analysis were collected on days 13, 19, and 28 only. The CO 2 emission measurements were conducted using an LI -820 CO 2 infrared gas analyzer (Lincoln, Nebraska, USA). After each sampling, the r emaining gas in the headspace was flushed with CO 2-free air. Flushing was found to dry out the soil, so de -ionized water (~10 - -structure fragment and four dest royed -structure fragments broke during the incubation and, while chemical analyses were possible, the broken fragments could not be re -scanned. 3.2.4 Soil fragment cutting and chemical analyses 13 15 N, total C and total N and thei r relation ship to pore characteristics, each intact soil fragment was cut into 5 to 13 sections. This was done to account for variation between the soil fragments. The number of sections into which the fragment was cut depended on its size and shape. To fa cilitate cutting, de -ionized water was added to fill 30% of the pore volume immediately prior to cutting. Cutting was performed with a #11 scalpel and a 24x magnifying glass. The relative position of each cut was recorded. Then , the relative positions were yielded regions in the three -dimensional tomographic images that corresponded to the physically cut sections. Image based porosity and pore size distributions were deter mined in each virtual section of each soil fragment. Prior to chemical analyses, visibly identifiable particulate organic matter (POM) was separated from physically cut sections and analyzed separately. The identifiable POM consisted primarily of plant roo t remains, but occasional plant residue fragments of unknown origin were also observed. Soil from cut sections, POM from cut sections, and separately collected rye roots 13 15 N, total C, and total N at the Stable Isotope Facility at t he University of California Davis. Fragment sections were analyzed using an Elementar Vario EL Cube or 62 Micro Cube elemental analyzer (Elementar Analysensysteme GmbH, Hanau, Germany) interfaced to a PDZ Europa 20 -20 isotope ratio mass spectrometer (Sercon L td., Cheshire, UK). POM material and rye roots were analyzed using a PDZ Europa ANCA -GSL elemental analyzer interfaced to a PDZ Europa 20 -20 isotope ratio mass spectrometer (Sercon Ltd., Cheshire, UK). 13 C at the Stable Isoto pe Facility at the University of California Davis. Gas samples were analyzed using a ThermoScientific GasBench system interfaced to a ThermoScientific Delta V Plus isotope ratio mass spectrometer (ThermoScientific, Bremen, Germany). The carbon isotopes are reported relative to Vienna PeeDee Belemnite (VPDB) with a 0.1› standard deviation for solid samples and 0.02› standard deviation for gas samples. The nitrogen isotopes are reported relative to AIR and had a standard deviation of 0.1›. 3.2.5 Grayscale Gra dients Grayscale gradients were used to identify spatial patterns in the soil matrix adjacent to root pores of 40 - The g rayscale value of an individual voxel from a is a function of the atomic number and relative density of the material within the voxel. Higher atomic number elements, such as iron, have higher grayscale values on images, while lower atomic number elements, such as carbon and nitrogen, have lower grayscale values on i mages. Therefore, the value of each grayscale voxel reflects elements present within it . Quigley et al (2018) showed that spatial gradients in grayscale values adjacent to the pores formed through plant root activities matched well with SOM gradients deter mined by the osmium staining method (Peth et al, 2014; Rawlins et al, 2016) . Thus, in this study I will use the grayscale gradients adjacent to the root pores as indicators of SOM distributions. Three root pores of 40 - range were identified on Pr e and Post images from 4 soil fragments . The root pores were then 3D dilated by one voxel to exclude any voxels 63 and the grayscale gradien ts were obtained by averaging the grayscale values of each layer (Figure 3.1). Averages excluded 0 values that represented the background and 255 values to prevent skewing the gradients by the occasional presence of inclusions of high atomic number element s, e.g., Fe . For direct comparison of the images, the values were normalized such that the lowest average grayscale value within the gradient was 0 and the highest average grayscale value within the gradient was 1. Figure 3. 1: Example of a selected root used for grayscale gradient analyses. The color overlay indicates the extent of 3.2.6 Statistical Analysis Comparisons between intact - and destroyed -structure treatments as well as between Pre 13 15 N, total C , and total N were conducted using the mixed model approach implemented in the PROC MIXED procedure of SAS Version 64 9.4 (SAS, 2009). The statistica l model for the analyses consisted of the fixed effects of treatment, Pre and Post, and their interaction; and a random effect of soil fragments nested within treatment and Pre and Post. The normality was visually assessed using normal probability plots an d stem -and -leaf plots, while equal variances was assessed using Levene™s test. Where the equal variance assumption was violated, analysis with unequal variances was conducted (Milliken and Johnson, 2009). 13 CO2 and CO 2 data obtained during soil fragment incubations, the statistical model consisted of the fixed effects of treatment, time, and their interaction. Time was treated as a repeatedly measured fixed factor using the REPEATED statement of PROC MIXED. 13 CO2 an 13 C of the fragments prior to incubation were 13 C of the soil in each treatment as a control and analyzed using Dunnett™s comparison -with -control test. The significant differences at the 0.05 level were reported, while tre nds are reported at the 0.1 level. 13 C and total C were conducted using the PROC REG procedure in SAS Version 9.4 (SAS, 2009). The significant slopes at the 0.05 level were reported. To investigate the correlation between the pore sizes (6.5 -15, 15-40, 40- 13 C, total C , and total N ), canonical correlation analysis was conducted using the cancor function in R (R Core Team, 2013). Canonical correlation compares how one set of variables is correlated to another set of variables in multidimensional space. The correlations are described through axes, which can be represented as orthogonal planes of maximum correlation, known as correlation axes. Each correlation axis is de fined by canonical variates . Canonical variates are latent variables , which are not observed, but derived from a 65 combination of the observed variables. Collinearity within the observed variables was checked through the calculation of the determinant prior to canonical correlation analysis. As canonical correlations requires a larger data set, only the Pre data set was used for canonical correlations due to the small sample size of the Post data set. 3.3 Results 3.3.1 Soil and plant characteristics Soil bulk density was lower in the treatments with rye as compared to control treatment (Table 3.1). The lower bulk density was the result of root growth and carbon addition. The 13 C values of particulate organic matter (POM), that is, the visible root rem ains and unidentifiable plant fragments isolated from intact soil fragments during their cutting, showed that the control treatment had significantly more C4 POM than the destroyed -structure treatment and numerically more C4 POM than the intact -structure t reatment (Table 3.1). The destroyed -structure and intact - 13 C of rye roots grown in destroyed -structure soil were depleted by ~1.5› more than rye roots grown in intact - 15 N of rye roots in destroyed -structure was depleted by ~3.3› as compared to intact (Table 3.1). Table 3. 1: Means of soil bulk density (n=2) and characteristics of rye roots (n=4) from the studi ed treatments. Treatment Bulk Density, g/cm 3 13 C roots 15 N roots C:N roots 13 C POM Control 1.40(0.1) a NA NA NA -18.4(2.3) a Intact 1.13(0.1) b -28.6(0.3)b 2.3(0.4) b 18.6(1.5)b -22.5(1.3)b Destroyed 1.16(0.1) b -30.1(0.4)a -1.0(0.3) a 12.1(1.8)a -26.0(1.8 )ab Prior to incubation, intact -structure and destroyed -structure soil had significantly higher total C than the control soil (Table 3.2). However, after incubation, this significance disappeared. 66 The C:N ratio was significantly lower for control soil fragment s than for the fragments from the intact -structure treatment both in Pre and Post 13 C values in destroyed -structure soil significantly increased, indicating losses in C3 carbon during incubation . The numerical trend in 15 N of Destroyed>Intact>Contro l observed in the Pre increased and became statistically significant post -incubation. The total N was not affected by either treatment or incubation. Table 3. 2: Means of soil carbon and nitrogen characteristics for the three studied treatments Pre and Post. Standard errors are shown in parentheses. Means and standard errors in each treatment are calculated based on 2 -6 aggregates with 1 -13 sections per aggregate. Letters indicate significant differences among treatments within Pre and Post Time Treatment Total C 13 C Total N 15 N C:N ratio Pre Intact 0.87(0.03)a -21.4(0.2) 0.10(0.01) 4.0(0.2) 9.1(0.4)a Destroyed 0.87(0.05)a -22.0(0.2)* 0.12(0.01) 4.2(0.2) 8.0(0.4)ab Control 0.74(0.04)b -21.5(0.2) 0.10(0.01) 3.6(0.3) 7.7(0.6)b Post Intact 0.86(0.04) -21.4(0.2) 0.10(0.01) 4.3(0.2)a 8.9(0.4)a Destroyed 0.80(0.03) -21.2(0.2) 0.10(0.01) 4.8(0.2)b 8.3(0.4)ab Control 0.77(0.07) -21.3(0.2) 0.11(0.02) 3.2(0.3)c 7.1(0.7)b 3.3.2 Pore characteristics Total porosity of individual soil fragments ranged from 10 -30% for all three treatments. in diameter, was around 12% in fragments from control and 20% in fragments from rye treatments ( Figure 3.2). After incubation, pore abundances tended to numerically increase in soils from all three treatments (Figure 3.3), however, the increases were only statistically significant for the pores from the 6.5 - group ( Figure 3.2). Pores with diameters >9 0 treatment, followed by the intact -structure and destroyed -structure soils. This tendency was observed in the soils prior to incubation and remained after incubation. Differences between treatments were onl y observed for the >90 m pores. 67 Figure 3. 2: Relative abundances of 6.5 -15, 15 -40, 40 - three studied treatments before and after incubation. Relative pore abundance ref ers to the percent of medial axes per total soil volume as determined from 3DMA -Rock software. Bars represent standard errors. Letters indicate es 68 Figure 3. 3: Representative slices of the same soil fragment for Pre -Intact (A) , Post -Intact (B) , Pre -Destroyed (C) , and Post -Destroyed (D) . Red arrows highlight an area where porosity visibly increased during incubation. Each soil fragment is approximately 5 mm across. 69 3.3.3 Associations between pores and chemical characteristics In soil from the control treatment there were no significant correlation s observed between the two studied carbon characteristics ( total C 13 C) and pore abundances of any of the studied sizes, either before or after incubation. There was also no correlation observed between the two nitrogen characteristics ( total N 15 N) and pore abundances (results not shown). 13 C and abundance of pores of different sizes in the Pre destroyed -structure treatment. Correlation with 6.5 - positive, no correlation was obse rved with 15 - 40- 3.4, Table 3.3). This indicates that in the soil from destroyed -structure Table 3. 3: Correlation coefficients for Pre and Post soil for total C a 13C with relative abundances of 6.5 - 15- - - 13C indicate more new carbon was associated with a higher presence of specified pore. Stars indicate significant Structure Pore size, Incubation Total C 13 C Destroyed 6.5-15 Pre -0.25 0.33* Post -0.75* 0.76* 15-40 Pre -0.28 0.17 Post -0.78* 0.79* 40-90 Pre 0.05 -0.39* Post -0.47* 0.40* >90 Pre 0.25 -0.07 Post 0.1 -0.19 Intact 6.5-15 Pre 0.19 0.02 Post 0.50* 0.52* 15-40 Pre 0.18 0.01 Post 0.37* 0.45* 40-90 Pre 0.26 -0.18 Post 0.11 0.17 >90 Pre 0.12 -0.14 Post 0.09 -0.08 treatment prior to its incubation, the sections with greater abundance of 6.5 - to have less C3 carbon while the sections with greater abundance of 40 - 70 Figure 3. 4: Correlations between total C (%C) and relative abundances of 40 - (A) 13C and 40- (B) for intact -structure treatment and destroyed -structure treatment for both Pre and Post. Relative pore abundances refer to the percent of medial axes per total soil volume as determined from 3DMA -Rock software. Gray area indicates 95% confidence interval. Correlation coefficients are shown in Table 3.3. 13 C was positively correlated with 6.5 -15, 15-40, and 40 - ores, indicating that the sections with greater abundance of pores of all three sizes tended to have less C3 carbon after incubation. The trend of negative correlations Pre and positive correlations Post 13 C and 40 - gnificant in destroyed -structure soil and 71 numeric in intact -structure soil. In the soil from the intact - 13 C was positively correlated to 6.5 -15 and 15 - There was no significant correlations between total C and any por e sizes in either intact -structure or destroyed -structure soils Pre (Table 3.3). In P ost intact -structure soil total C was positively correlated with 6.5 - - -structure soil total C was negatively correlated with these pores. 3.3.4 Incubation CO 2 The cumulative amount of CO 2 emitted from the soil fragments during the 28 -day incubation was the highest in the soil from the intact -structure treatment, followed by the destroyed -structure and control treatments ( Figure 3.5 13 C values of the CO 2 emitted during the incubation indicate that microorganisms preferentially used more C3 carbon in the destroyed -structure and intact -structure treatments than in the control, but the differen ce was only statistically significant on day 28 ( Figure 3.5b). 13 C values of the CO 2 emitted during the incubation indicate that during the last three measurements (days 13, 19, and 28), the CO 2 gas became more depleted for all three treatments. 72 Figure 3. 5: (A) Cumulative CO 2 and (B) average isotopic signature of CO 2 respired during 28 day incubation 13C values of the soil fragment sections prior to incubation for each treatment while boxes indicate the standard errors ( -22.0±0.1›, -21.4±0.1›, and -21.6±0.1› for destroyed -structure, intact -structure, and control, respectively). Letters indicate significant 13C of the emitted CO 2 73 3.3.5 Grayscale Gradients Both intact - and destroyed -structure Pre and Post soils had similar general patterns of very high grayscale val ues directly adjacent to the pores, followed by a sudden decrease (Figure 3.6). Then, the grayscale values slowly increased until reaching a plateau at 120 - distances from the pore. The plateau grayscale value roughly corresponded to the background grayscale value. However, the differences in Pre and Post grayscale gradients had opposite signs in the two treatments. In destroyed -structure soil, Pre soils had lower grayscale values than Post at the same distance, while Pre intact -structure soil had hig her grayscale values than the Post soil. Figure 3. 6: Normalized grayscale values from µCT images of soil fragments as a function of distance from 40 -90 -structure and destroyed -structure at Pre and Post. Error bars indicate standard error. Note that values of the normalized grayscale reflect a combination of contributors, including atomic numbers of the elements and density of the material located within an image voxel. Specifically, lower normalized values her e correspond to lower atomic number elements and lower densities, while higher values correspond to higher atomic number elements and higher densities. As such, lower values roughly represent more carbon in the soil matrix, while higher values represent le ss carbon in the soil matrix and/or denser soil matrix. 74 3.3.6 Canonical Correlations The first two canonical correlation axes were significant at the 0.05 level ( Figure 3.7). The first canonical variates can be described by the relationship between total C (negatively correlated) and total N (positively correlated) with 6.5 - -40 - higher C:N ratios while 15 - ated with lower C:N ratios. There was a treatment difference observed in this axis between destroyed -structure and intact -structure soils: the destroyed -structure soil contained more carbon with lower C:N ratios and a higher abundance of 15 - n intact -structure soil. Figure 3. 7: 13C. The first two canonical e the latent variables for each axes are shown on the right. The sign indicates the direction of correlation and the number indicates the amount each observed variable contributes to the latent variable. Lines indicate the (0,1) line (A) and (0, -1) line (B) , while letters indicate treatment (destroyed -structure (D) and intact -structure (I)). 75 13 C (positive correlation) and total N (positive correlation) with 40 - ive - have newer carbon with higher nitrogen concentrations carbon with lower nitrogen concentrations . There was no effect of treatment observed in the second canonical correlation axis. 3.4 Discussion Three months of rye growth increased total C and the C:N ratio within both the intact -structure and destroyed -structured soils. However, in the subsequent incubation, gains of total C tended to disappear 13 C results, the carbon losses, at least in th e destroyed -structure fragments , were dominated by losses in C3 carbon . Gains and losses of C3 and of total carbon were associated with presence of soil pores. However, the relationships between carbon and pores differed for different pore sizes, suggesting different microscale mechanisms by which these pores contribute to carbon accrual processes. 3.4.1 Relationship between C3 carbon and 40 - The corre 13 C and pores of the studied size ranges had similar signs in both intact - and destroyed -structure soils, but in the intact -structure soil, the correlations were not statistically significant (Table 3.3). This is likely the outcome of the legacy of soil pore architecture of the intact -structure soil, which contributed to greater variability, thus lowering statistical significance in that treatment, as well as differences in decomposability of plant root material in the two treatments (as di scussed below). Negative correlation between 13 C and 40 -90 , indicated that greater levels of C3 were associated with the presence of 40- Figure 3.4, Table 3.3). It is assumed that the increase in C3 carbon is associated with the newly added carbon. I surmise that a 76 possible cause for this association is that many of the 40 -90 m pores , especially those in the destroyed -structure soil, were created by fine plant r oots. Since old root pores were destroyed during the sieving process, any 40 - in the destroyed -structure soil, which were of root origin , would have been directly produced by the growth of the rye. On the contrary, in the intact -structure treatm ent such pores could have been produced by both new and historically grown plants, increasing variability that weakened the correlation. After incubation, the gains in new carbon in the destroyed -structure soil in relation to the abundance of 40 - es were quickly lost. The 40 - correlated with new carbon in Pre to being negatively correlated with new carbon in Post. The grayscale gradients in the Post destroyed -structure soil had higher grayscale values than in t he Pre (Figure 3.6). This further supports the notion that, while prior to incubation the SOM levels in the vicinity of such newly formed pores were relatively high, in samples subjected to incubation the SOM levels adjacent to 40 - Gre ater decomposition of newly added carbon in 40 - more labile nature of the new carbon and from greater microbial activities in these pores. The second canonical correlation axis (Figure 3.7) shows that the 40 - tend to have newer carbon and a higher concentration of nitrogen, thus possibly, containing more decomposable organic compounds. Indeed, the small plant roots located within such pores could have been more easily decomposable since fine roots tend to have less lignin and a lower lignin:N ratio is an indication of root decomposability (Rasse et al, 2005). Bailey et al (2017) observed that water decomposable, than lignin a nd tannin, which are more difficult to decompose. Moreover, the increased decomposition/carbon loss in such pores was reported as related to greater microbial 77 presence, transport, and activity in 40 - Kravch enko et al, 2014). Some of the differences between the intact - and destroyed -structure treatments in terms of pore associations with new carbon might be related to differences in root decomposability. The intact -structure roots had a higher C:N ratio, as w 15 13 C (Table 3. 15 N of plant roots is controlled by the nitrogen use efficiency. Large differences between root and 15 N values can result from pooling of nitrate in plant roots (Kalcsits et al, 2015; Kalcsits and Guy, 2016). T 15 N was 2.08› for rye samples collected from both intact -structure and destroyed -structure soils, but, while the intact -structure roots were similar to the shoot values, destroyed -structure roots were ~3› more depleted. This suggests that pooling of nitrate could have taken place in the destroyed -structure roots, lowering C:N ratio and increasing 13 C values of roots from the destroyed -structure soil would make it slightly easier for microorganisms to de compose them than intact -structure 13 C of C3 plants are related to water availability with more depleted values occurring where water is more plentiful (Farquhar et al, 1989, Stewart et al, 1995). The differences in overall p ore size distributions of the two treatments could be the cause for the differences in nitrate and water availability. However, since the normal range of values for C3 plants is from -24 to -34›, the difference between intact -structure and destroyed -struct ure roots observed in this study can be regarded as relatively small. 3.4.2 Relationsh ip between carbon and 6.5 - , 15- After incubation, there was a notably decreased association with C3 carbon in both intact -structure and destroyed -structure soil. This implies a preferential utilization of newer carbon in these pores. This preference could be the result of anaerobic conditions that existed within the soil. During incubation, the soil moi sture level was kept at 60% water filled porosity , which 78 would have resulted in water filling the majority of both the 6.5 -15 and 15 - the incubation , resulting in anaerobic conditions prevailing there during incubation. Keiluweit et al (2 017) observed that in anaerobic microsites within upland soils, decomposition rates were reduced by a factor of 10, which may also explain the slower decomposition of materials from these pores as seen in the association with increased amounts of carbon. T he anaerobic conditions may also explain why newer carbon was preferentially used in association with these pores. Newer carbon would likely contain more oxidized functional groups than older carbon. These functional groups would be quickly used under anae robic conditions, resulting in biased decomposition of newer carbon in relation to pores of 6.5 -15 and 15 - The association between total C and 15 - 3.8) was identical to those observed by Ananeyv a et al (2013). The two data se ts, while of the same soil type and collected from the same geographic area, were of two completely different agricultural managements. This study is from a 20 year conventional management continuous corn treatment, while Ananeyva et al (2013) used aggrega tes from a 19 year native succession management , which was essentially unmanaged. This seems to suggest a universal mechanism for the relationship between soil carbon and the presence of 15 - possible driver of this relationship might be the presence of fungi in these pores. The first canonical correlation axis (Figure 3.7), shows a difference in the C:N ratio of the two pore sizes. This potentially could signal a difference in decomposability between 6.5 - - Bailey et a l (2017) and Smith et al 79 Figure 3. 8: Correlation between total carbon and abundance of 15 - reproduced with permission from Elsevier) or 15 - -axis is presented as total C, g/kg instead of %C to align with the original Anayeva et al (2013) graph. pore s y attributed this difference to accessibility of fungi, which preferential decompose more complex organic materials, but , as 2006). Fungi are also known to create pores of 20 - extruding binding agents, which w ould create micro -environments with more decomposable material in these created pores (Dorioz et al, 1993; Bearden, 2001, and Emerson and McGarry, 2003). Another p otential explanation might be the presence of root hairs. Root hairs are also 10 - necessary to explore the cause of this correlation between total C and 15 - Caution, 80 however, should also be applied to these results as some of these correlations might be the result of soil handling, i.e. water addition during incubation, drying during scanning, etc., and not neccesarily management effects. 3.4.3 Additional con siderations The CO 2 results seem to indicate a different story than the soil fragment data. In the soil fragment data, destroyed -structure soil lost the most carbon during incubation, while the intact -structure soil lost a negligible amount of carbon durin g incubation. The CO 2 data, on the other hand, indicates that the intact -structure lost the most carbon as CO 2. This discrepancy is due to 13 15 N measurements. The amount of POM removed from the intact -structure soil was almost twice as large as the amount of POM removed from the destroyed -structure soil. This means that the discrepancy between the CO 2 data and soil fragments was most likely due to the difference in the amount of POM. I recognize that in terms of exploring associations between carbon and soil pores my work is, in essence, an observational study. Thus, it possesses a limitation common to all observational studies, that is, an inability to unequivocally declare cause and effect relationships. Yet, I posit that, at present it is impossible to recreate soi l environments with specific pore characteristics for controlled cause -effect determin ation. Even though creation of artificial soil materials with contrasting pore architecture by either using soil fractions of different sizes or by soil compaction is pos sible (Negassa et al, 2015, Sleutel et al, 2012,Thomson et al, 2010; Stenger et al, 2002; De Neve and Hofman, 2000) such constructions fail to recreate biological conditions. By biological conditions, I refer to the structure and abundance of resident micr obial communities, formed in pores of different sizes in situ and acclimated to specific 81 microenvironments existing there. Since it is microbial activities that largely drive carbon processing, failure to correctly represent them will likely mislead findin gs. This leaves no alternative, but observational studies, such as this study, to explore the role of pores within soil micro -environments. 3.5 Conclusion My findings confirm previous results on the importance of pores in tens of microns size range for pro cessing of organic carbon in soil, specifically in regards to fate and distribution of newly added carbon. I demonstrated that p ores of 40 - intriguing role in new carbon gains as well as its subsequent losses. Such pores seem to be "easy come easy go" locations which receive the greatest amounts of new carbon from growing p lant roots, but then rapidly lo se that newly added carbon. On the other hand, both 6.5-15 and 15 -40 wer carbon. Carbon protection associated with the 6.5-15 m pores could be associated with lack of accessibility by fungal hyphae and pervasiveness of anaerobic conditions when soils are near field capacity . Pores of 15 - size are also associated with a prevalence of anaerobic conditions when soils are above field capacity , but fungal hyphae are not excluded and are potential drivers of carbon dynamics in pores of this size. Funding Support for this research was provided in part by the USDA -NIFA, A ward No. 2016 -67011-24726 fiUsing stable isotopes and computer tomography to determine mechanisms of soil carbon protection in cover crop based agricultural systemsfl, USDA -NIFA, Award No. 2011 -68002-30190 fiCropping Systems Coordinated Agricultural Project ( CAP): Climate Change, Mitigation, and Adaptation in Corn -based Cropping Systemsfl sustainablecorn.org, and by the US National Science Foundation Long -Term Ecological Research Program (DEB 1027253) at the Kellogg 82 Biological Station and by Michigan State Univ ersity AgBioResearch. Portions of this work were performed at GeoSoilEnviroCARS (The University of Chicago, Sector 13), Advanced Photon Source (APS), Argonne National Laboratory. GeoSoilEnviroCARS is supported by the National Science Foundation - Earth Sci ences (EAR - 1634415) and Department of Energy - GeoSciences (DE -FG02 -94ER14466). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE -AC02 -06CH11357. 83 REFERENCES 84 REFERENCES Ananyeva, K., Wang, W., Smucker, A. J. M., Rivers, M. L., and Kravchenko, A. N. (2013). Can intra -aggregate pore structures affect the aggregate™s effectiveness in protecting carbon? Soil Biol. Biochem. 57, 868-875. doi: 10.1016/j.soilbio.2012.10.019 Bailey, V. L., Smith, A. P., Tfaily, M., Fansler, S. J., and Bond -Lamberty, B. (2017). Differences in soluble organic carbon chemistry in pore waters sampled from different pore size domains. Soil Biol. Biochem. 107, 133-143. doi: 10.1016/j.soilbio.2016.11.025 Bearden, B. N. (2001). Influence of arbuscular mycorrhizal fu ngi on soil structure and soil water characteristics of vertisols. Plant Soil 229, 245Œ 258. doi: 10.1023/A:1004835328943 Bernoux, M., Cerri, C. C., Neill, C. and de Moraes, J. F. L. (1998) . The use of stable carbon isotopes for estimating soil organic mat ter turnover rates. Geoderma 82, 43 -58. doi: 10.1016/S0016 -7061(97)00096 -7 Bowling, D. R., Pataki, D. E., and Randerson, J. T. (2008) . Carbon isotopes in terrestrial ecosystem pools and CO 2 fluxes. New Phytol . 178, 24-40. doi: 10.1111/j.1469 -8137.2007.02342.x Brimecombe, M. J., de Leij, F. and Lynch, J. M. (2001). fiThe effect of root exudates on rhizosphere microbial populations,fl in The rhizosphere: biochemistry and organic substances at the soil -plant interface , eds. R. Pinton, Z. Varanini and P. Nannipie ri. (New York, New York, USA: Marcel Dekker), 95 Œ140. Cannell, R. Q. (1977). Soil aeration and compaction in relation to root growth and soil management. Appl. Biol . 2, 1-86. Czarnes, S., Hallett, P. D., Bengough, A. G., and Young, I. M. (2000). Root - and microbial -derived mucilages affect soil structure and water transport. Eur. J. Soil Sci. 51, 435-443. doi: 10.1046/j.1365 -2389.2000.00327.x De Gryze, S., Jassogne, L., Six, J., Bossuy t, H., Wevers, M., and Merckx, R . (2006). Pore structure changes during d ecomposition of fresh residue: X -ray tomography analyses. Geoderma 134, 82-96. doi: 10.1016/j.geoderma.2005.09.002 De Neve, S. and Hofman, G. (2000) . Influence of soil compaction on carbon and nitrogen mineralization of soil organic matter and crop residues. Biol. Fertil. Soils 30, 544 -549. doi: 10.1007/s003740050034 Derrien, D. and Amelung, W. (2011). Computing the mean residence time of soil carbon fractions using stable isotopes: impacts of the model framework. Eur. J. Soil Sci. 62, 237-252. doi: 10.1111/j.1365 -2389.2010.01333.x 85 Dorioz, J. M., Robert, M., and Chenu, C. (1993). The role of roots, fungi and bacteria on clay particle organization. An ex perimental approach. Geoderma 56, 179Œ 194. doi:10.1016/0016 -7061(93)90109 -X Dou F., Hons F. M., Wright A. L., Boutton, T. W. and Yu, X. (2014) . Soil carbon sequestratio n in sorghum cropping systems: evidence from stable isotopes and aggregate -size fractionation. Soil Sci. 179, 68-74. doi: 10.1097/SS.0000000000000045 Dungait, J. A. J., Hopkins, D. W., Gregory, A. S., and Whitmore, A. P. (2012). Soil organic matter turnover is governed by accessibility not recalcitrance. Global Change Biol. 18, 1781-1796. doi: 10.1111/j.1365 -2486.2012.02665.x Ehleringer, J. R., Buchmann, N., and Flanagan, L. B. (2000). Carbon isotope ratios in belowground carbon cycle processes. Ecol. Appl . 10, 412 -422. doi: 10.2307/2641 103 Ekschmitt, K., Kandeler, E., Poll, C., Brune, A., Buscot, F., Friedrich, M., et al. (2008). Soil -carbon preservation through habitat constraints and biological limitations on decomposer activity. J. Plant Nutr. Soil Sci. 171, 27-35. doi: 10.1002/jpln.2 00700051 Ekschmitt, K., Liu, M., Vetter, S., Fox, O., and Wolters, V. (2005). Strategies used by soil biota to overcome soil organic matter stability Œ why is dead organic matter left over in soil? Geoderma 128, 167-176. doi: 10.1016/j.geoderma.2004.12.024 Emerson,W. W. and McGarry, D. (2003). Organic carbon and soil porosity. Aust. J. Soil Res . 41, 107Œ118. doi: 10.1071/SR01064 Farquhar, G. D., Ehleringer, J. R., and Hubick, K. T. (1989). Carbon isotope discrimination and photosynthesis. Annu. Rev. Plant P hysiol. Plant Mol. Biol. 40, 503-537. doi: 10.1146/annurev.pp.40.060189.002443 Grandy, A. S. and Robertson , G. P. (2007) . Land -use intensity effects on soil organic carbon accumulation rates and mechanisms. Ecosystems 10, 58-73. doi: 10.1007/s10021 -006-9010-y Gray, J., and Lissmann, H. W. (1938). An apparatus for measuring the propulsive forces of the locomotory muscles of the earthworm and other animals. J. Exp. Biol. 15, 518-521. Greacen, E. L., and Oh, J. S. (1972). Physics of root growth. Nat. New Biol. 235, 24-25. doi: 10.1038/newbio235024a0 Greacen, E. L., and Sands, R. (1980). Compaction of forest soils. a review. Aust. J. Soil. Res. 18, 163-189. doi: 10.1071/SR9800163 Jackson, R. B., Lajtha, K., Crow, S. E., Hugelius, G., Kramer , M. G., and Pi ñeiro, G. (2017). The ecology of soil carbon: pools, vulnerabilities, and biotic and abiotic controls. Annu Rev Ecol Evol Syst . 48, 419-445. doi: 10.1146/annurev -ecolsys -112414-054234 86 Kalcsits, L. A. and Guy, R. D. (2016). Variation in fluxe s estimated from nitrogen isotope discrimination corresponds with independent measures of nitrogen flux in Populus balsamifera L. Plant Cell Environ. 39, 310 -319. doi: 10.1111/pce.12614 Kalcsits, L. A., Min, X., and Guy, R. D. (2015). Interspecific variati on in leaf -root differences in 15 N among three tree species grown with either nitrate or ammonium. Trees. 29, 1069-1078. doi: 10.1007/s00468 -015-1186-3 Kell, D. B. (2012) . Large -scale sequestration of atmospheric carbon via plant roots in natural and agri cultural ecosystems: why and how. Philos T R Soc B . 367, 1589-1597. doi: 10.1098/rstb.2011.0244 Keiluweit, M., Wanzek, T., Kleber, M., Nico, P., and Fendorf, S. (2017). Anaerobic microsites have an unaccounted role in soil carbon stabilization. Nat. Commun . 8: 1771 . doi: 10.1038/s41467 -017-01406-6 Kravchenko, A. N ., and Guber, A. K. (2017). Soil pores and their contributions to soil carbon processes. Geoderma 287, 31 -39. doi: 10.1016/j.geoderma.2016.06.027 Kravchenko, A. N., Negassa, W. C., Guber, A. K., Hi ldebrandt, B., Marsh, T. L., and Rivers, M. L. (2014). Intra -aggregate pore structure influences phylogenetic composition of bacterial community in macroaggregates. Soil Sci. Soc. Am. J. 78, 1924-1939. doi: 10.2136/sssaj2014.07.0308 Kravchenko, A. N., Negassa, W. C., Guber, A. K., and Rivers, M. L. (2015). Protection of soil carbon within macro -aggregates depends on intra -aggregate pore characteristics. Sci. Rep. 5:16261 . doi: 10.1038/srep16261 Kuzyakov, Y., and Blagodatskaya, E. (201 5). Microbial hotspots and hot moments in soil: concept & review. Soil Biol. Biochem. 83, 184-199. doi: 10.1016/j.soilbio.2015.01.025 Lal, R. (1999) . Soil management and restoration for carbon sequestration to mitigate the accelerated greenhouse effect. Prog. Env. Sci. 1, 307 -326. doi: 10.1055/s -2008-1072270 Lindquist, W. B., Venkatarangan, A., Dunsmuir, J., and Wong, T. (2000). Pore and throat size distributions measured from synchrotron X -ray tomographic images of Fontainebleau sandstornes. J. Geophys. Re s. 105, 21,509-21,527. doi : 10.1029/2000JB900208 Marshall, T. J., Holmes, J. W., and Rose, C. W. (1996). Soil Physics, third ed. Cambridge, England: Cambridge University Press. Milliken, G .A., and Johnson, D. E. (2009). Analysis of Messy Data Volume I: De signed Experiments, second ed. Boca Raton, FL: CRC Press. Negassa, W. C., Guber, A. K., Kravchenko, A. N., Marsh, T. L., Hildebrandt, B., and Rivers, M. L. (2015). Properties of soil pore space regulate pathways of plant residue decomposition and community structure of associated bacteria. PLoS One 10(4) : e0123999 doi: 10.1371/journal.pone.0123999 87 Neumann, G., Bott, S., Ohler, M. A., Mock, H. -P., Lippmann, R., R. Grosch, et al. (2014) . Root exudation and root development of lettuce ( Lactuca sativa L. cv. Ti zian) as affected by different soils. Front. Microbol. 5:2. doi: 10.3389/fmicb.2014.00002 Oh, W., and Lindquist, W. B. (1999). Image thresholding by indicator kriging. IEEE Trans. Pattern Anal. Mach. Intell. 21, 590-602. doi: 10.1109/34.777370 Park, E. J. and Smucker, A. J. M. (2005). Saturated hydraulic conductivity and porosity within macroaggregates modified by tillage. Soil Sci . Soc. Am. J. 69, 38-45. doi: 10.2136/sssaj2005.0038 Peth, S., Horn, R., Beckmann, F., Donath, T., Fischer, J., and Smucker, A. J. M. (2008). Three -dimensional quantification of intra -aggregate pore -space features using synchrotron -radiation -based microtomography. Soil Sci . Soc. Am. J. 72, 897-907. doi: 10.2 136/sssaj2007.0130 Peth, S., Chenu, C., Leblond, N., Mordhorst, A., Garnier, P., Nunan, N., et al. (2014). Localization of soil organic matter in soil aggregates using synchrotron -base X -ray microtomography. Soil Biol. Biochem. 78, 189-194. doi: 10.1016/j. soilbio.2014.07.024 Quigley, M. Y., Rivers, M. L., and Kravchenko, A. N. (2018) . Patterns and sources of spatial heterogeneity in soil matrix from contrasting long term management practices. Front. Environ. Sci. 6:28 doi: 10.3389/fenvs.2018.00028 R Core Te am (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R -project.org/. Rabot, E., Wiesmeier, M., Schl üter, S., and Vogel, H. -J. (2018) . Soil structure as an indicator of soil functions: a review. Geoderma 314, 122-137. doi:10.1016/j.geoderma.2017.11.009 Rasse, D. P., Rumpel, C. and Dignac, M. -F. (2005) . Is soil carbon mostly root carbon? Mechanisms for a specific stablisation. Plant Soil 269, 341-356. doi: 10.1007/s11104 -004-0907-y Rasse, D. P. and Smucker, A. J. M. (1998). Root recolonization of previous root channels in corn and alfalfa rotations. Plant Soil 204, 203-212. doi: 10.1023/A:1004343122448 Rawlins, B. G., Wragg, J., Reinhard, C., Atwood, R. C., Houston, A., Lark, R. M., et al. (2016). Three -dimensional soil organic matter distribution, accessibility and microbial respiration in macroaggregates using osmium staining and synchrotron X -ray computed tomography. SOIL . 2, 659-671. doi: 10.5194/soil -2-659-2016 Rivers, M. L. (2012). tomoRecon: High -speed tomography reconstruction on workstations using multi -threading. Proc. SPIE 8505:OU. doi:10.1117/12.930022 Ruiz, S., Schymanski, S. J., and Or, D. (2017). Mechanics and energetics of soil penetration by earthworms and plant roots: higher rates cost more. Vadose Zone J. 16:8. doi: 10.2136/vzj2017.01.0021 88 SAS Inc. (2009). SAS user™s guide. Vers ion 9.2 . Cary, NC: SAS Inst. Schmidt, M. W., Torn, M. S., Abiven, S., Dittmar, T., Guggenberger, G., Janssens, I. A., et al. (2011). Pe rsistence of soil organic matter as an ecosystem property. Nature 478, 49-56. doi:10.1038/nature10386 Six, J., Elliott, E. T., and Paustian, K. (2000). Soil macroaggregate turnover and microaggregate formation: a mechanism for C sequestration under no -tillage agriculture. Soil Biol. Biochem . 32, 2099-2103. doi: 10.1016/S0038 -0717(00)00179 -6 Six, J., Frey, S. D., Thiet, R. K., and Batten, K. M. (2006). Bacterial and fungal contributions to carbon sequestration in agroecosystems. Soil Sci . Soc. Am. J. 70, 555-569. doi:10.2136/sssaj2004.0347 Sleutel, S., Bouckaert, L. Buchan, D., Van Loo, D., Cornelis, W. M., and Sanga, H. G. (2012) . Manipulation of the soil pore and microbial community structure in soil mesocosm incubation studies. Soil Biol. Biochem. 45, 40-48. doi: 10.1016/j.soilbio.2011.09.016 Smith, A. P, Bond -Lamberty, B., Benscoter, B. W., Tfaily, M. M., Hinkle, C. R., Liu, C., et al (2017) . Shifts in pore connectivity from precipitation versus groundwater rewetting increases soil carbon loss after drough t. Nature Comm . 8:1335. doi : 10.1038/s41467 -017-01320-x Smucker, A. J. M., Park, E. -J., Dorner, J., and Horn, R. (2007). Soil micropore development and contributions to soluble carbon transport within macroaggregates. Vadose Zone J. 6, 282-290. doi: 10.2136/vzj2007.0031 Stewart, G. R., Turnbull, M. H., Schmidt, S., and Erskine, P. D. (1995). 13 C natural abundance in plant communities along a rainfall gradient: a biological integrator of water availability. Aust. J. Plant Physiol . 22, 51-55. doi: 10.1071/PP9950051 Stenger, R., Barkle, G. F., and Burgess, C. P. (2002) . Mineralisation of organic matter in intact versus sieved/refilled soil cores. Aust. J. Soil Res. 40, 149-160. doi: 10.1071/SR01003 Strong, D. T., De Wever, H., Merckx, R., and Recous, S. ( 2004). Spatial location of carbon decomposition in the soil pore system. Eur. J. Soil Sci. 55, 739-750. doi: 10.1111/j.1365 -2389.2004.00639.x Swift, R. S. (2001). Sequestration of carbon by soil. Soil Sci. 166, 858-871. doi: 10.1097/00010694 -200111000-00010 Thomson, B. C., Ostle, N. J., McNamara, N. P., Whiteley, A. S., and Griffiths, R. I. (2010) . Effects of sieving, drying and rewetting upon soil bacterial community structure and respiration rates. J. Microbiol. Methods 83, 69-73. doi: 10.1016/j.mimet.201 0.07.021 Tiemann, L. K. and Grandy, A. S. (2015). Mechanisms of soil carbon accrual and storage in bioenergy cropping systems. Glob Change Biol Bioenergy 7, 161-174. doi: 10.1111/gcbb.12126 89 Urbanek, E., Smucker, A. J. M., and Horn, R. (2011) . Total and fre sh organic carbon distribution in aggregate size classes and single aggregate regions using natural 13 C/12 C tracer. Geoderma 164, 164-171. doi:10.1016/j.geoderma.2011.05.020 Wang, W., Kravchenko, A. N., Johnson, T., Srinivasan, S., Ananyeva, K. A., Smucker , A. J. M., et al. (2013). Intra -aggregate pore structures and Escherichia coli distribution by water flow within and movement out of soil macroaggregates. Vadose Zone J. 12:4. doi: 10.2136/vzj2013.01.0012 Wang, W., Kravchenko, A. N., Smucker, A. J. M., Li ang, W., and Rivers, M. L. (2012). Intra -aggregate pore characteristics: X -ray computed microtomography analysis. Soil Sci. Soc. Am. J. 76, 1159-1171. doi: 10.2136/sssaj2011.0281 Wang, W., Kravchenko, A. N., Smucker, A. J. M., and Rivers, M. L. (2011). Com parison of image segmentation methods in simulated 2D and 3D microtomographic images of soil aggregates. Geoderma . 162, 231 -241. doi: 10.1016/j.geoderma.2011.01.006 Wiersum, L. K. (1957). The relationship of the size and structural rigidity of pores to their penetration by roots. Plant Soil 9, 75-85. doi: 10.1007/BF01343483 Young, I. M., and Crawford, J. W. (2004). Interactions and self -organization in the soil -microbe complex. Science 304, 1634 -1637. doi: 10.1126/science.1097394 Young, I. M., Crawford, J. W., and Rappoldt, C. (2001). New methods and models for characterizing structural heterogeneity of soil. Soil Tillage Res. 61, 33-45. doi: 10.1016/S0167 -1987(01)00188 -X 90 Chapter 4: Effect of management and pore size distributi on on the input and persistence of new carbon Abstract Agricultural management can have a large eff ect on soil carbon concentrations with some managements triggering losses and others generating accumulation. However, the mechanisms behind these effects are not sufficiently understood, especially at a scale of a few to hundreds of microns where the phys ical protection of soil carbon takes place. Understanding where new carbon is added and how it is used could refine current management recommendations for enhancing C accrual and improving soil health. As soil pores control the movement of gasses, water, a nd microorganisms, soil pores may also control new carbon gains and losses. In order to determine how new carbon is added in relation to abundance and size distribution of soil pore s, I utilized cereal rye ( Secale cereale L.) enriched with 13 C to track the addition of new carbon. Pulse labeling was used to enrich the rye during growth and to track the fate of plant derived C in soil after plant growth and after subsequent 3 week incubation . Computed microtomography was used to characterize pore size distrib utions . Soil from two contrasting agricultural managements, conventional and biologically based, was used . In order to better differentiate the effect of roots, soil was either kept intact or destroyed with a 1 mm sieve. Soil mini -cores were taken after ry e 13 C and total C. Results indicate that in soils with legacy root channels, new carbon is added evenly between pores of different sizes. In soils without pre -existing root channels, new carbon was preferentially added to 15 -40 and 40 - Relationships between new carbon and pore s were lost after incubation, indicating rapid new carbon loss from pores where new carbon was added , although the loss was less severe in 4-40 . This has implications for understanding underlying mechanisms of why some soils retain carbon, while others do not. 91 4.1. Introduction Soil carbon stocks are roughly equivalent to twice the amount of carbon stored in the atmosphere ( Lal, 1999; Swift, 2001; Falkowski et al, 2000 ; Davidson and Ja nssens, 2006 ). While a substantial portion of this carbon is stored in arctic regions, agricultural soils have a large untapped storage capacity that can help with climate change mitigation (Oechel and Vourlitis, 1994; Lal, 2011; Dungait et al , 2012; Kell, 2012). Additionally, soil carbon is strongly linked with higher soil fertility and greater crop yields, thus making increased storage of soil carbon in agricultural soils important to agricultural sustainability and soil health (Melsted, 1954; Bauer and B lack, 1994; Lal, 2006). Agricultural management can have a substantial effect on soil carbon gains and losses (Senthilkumar et al, 2009; Syswerda et al, 2011 ). Conventionally managed systems, i.e., those receiving tillage, chemical fertilizers, and no win ter cover, are associated with carbon losses (Grandy and Robertson, 2007; Ruan and Robertson, 2013; Abraha et al, 2018). Including cover crops in the rotation, i.e. planting of a non -cash crop between cash crops, can provide erosion control, suppress weeds , increase water holding capacity and fertility, as well as enable soil carbon gains; although the gains can take years to be reliably detected (Necp álov á et al, 2014; Rorick and Kladivko, 2017 ). The mechanisms behind carbon gains in agricultural systems w ith cover crops are not fully understood ( Austin et al , 2017). For example, increasing plant biomass inputs is believed to be one of the best ways to improve soil carbon (Paustian et al, 2016). However, crops producing large amounts of biomass do not always lead to substantial carbon gains (Garten and Wullschleger, 1999; Chimento et al, 2016; Sp runger and Robertson, 2018). This observation indicates that not only the amount of C input , but it subsequent protection within the soil matter is required for increasing soil C levels. Protection of soil C is driven by micro -scale soil 92 processes, including accessibility of new C to microbial decomposers th at is achieved via soil pores. Already, t he additio n of microscale information has been shown to increase m odel accuracy over us e of the macroscale characteristics for modelling carbon respiration and hydraulic properti es (Falconer et al, 2015; Smet et al, 2018) . Lack of accounting for microscale processes may explain the wide variability in performance of soil carbon models (Keel et al, 2017). The presence of diverse microenvironments is believed to be the key driver of carbon protection within soils (Kuzyakov and Blagodatskaya, 2015). However, characteristics and properties of soil microenvironments are largel y defined by soil pores that control the fluxes of water and gases, microbial access to carbon sources, as well as microorganism movement and nutrient transport ( Young et al , 2001; Ekschmitt et al , 2005, 2008; Park et al, 2007; Kravchenko and Guber, 2017; Rabot et al , 2018). Through their control of soil microenvironment creation, soil pores are also essential to soil carbon protection. While a link between pores of specific size ranges and soil carbon loss/protection has been established ( Strong et al, 200 4; Ananyeva et al, 2013; Bailey et al, 2017; Quigley et al, 2018a), the mechanisms behind these correlations have yet to be elucidated. Moreover, how roots contribute to these relationships is also poorly understood as root growth can affect the pore size distribution (Graecen et al 1968; Dexter, 1987), but pore size distribution can also affect root growth (Bowen, 1981; Bengough et al, 2006). Numerous studies have shown that roots, as compared to shoots, contribute a disproportionate amount (up to 75%) to soil carbon (Gale et al, 2000; Rasse et al, 2005; Kong and Six, 2010; Mazzilli et al, 2015; Austin et al, 2017). This contribution can be in the form of actual root biomass or through root exudates. Studies have shown that around 30 -50% of 93 belowground biom ass can be attributed to root exudates (Barber and Martin, 1976; Meharg and Killham, 1991; Kuzyakov et al, 2003). The exudates consist of organic acids, amino acids, and other small, highly soluble and easy to decompose compounds, although mucilage and oth er harder to decompose materials can also be produced (Brimecombe et al, 2011; Dungait et al, 2012). The easily decomposable compounds can then be quickly taken up by soil microbes, contributing to microbial biomass. It has been found that up to 25 -30% of microbial biomass carbon can be derived from actively growing plants (Williams et al, 2006; Austin et al, 2017). Processing carbon by soil microorganism is also known to be one of the first steps in soil organic matter production. This processed carbon can then easily attach to mineral particles, where they are protected from further degradation (Grandy and Neff, 2008; Wieder et al, 2014; Kallenbac h et al, 2015, 2016; Jackson et al, 2017). Therefore, spatial patterns in the distribution of roots and their e xudates can also play an important role in soil carbon protection. The distribution of soil pore sizes varies depending on agricultural management. Conventional management has been associated with an increased presence of 40 - (Wang et al, 2012), which has also been linked with carbon losses (Ananeyeva et al, 2013), especially of newer carbon (Quigley et al, 2018a). These pores are created through either mechanical wetting/drying and freeze/thaw cycles or by smaller plant roots. On the other hand, management with continuous presence of live vegetation cover, e.g., cover crops, has been larger roots (Kravchenko et al, 2014). However, existing root pores also are preferentially used by new plants, potentially masking any new root effects (Rasse and Smucker, 1998). I examined the associations between pore size distributions, i.e., abundances of pores of different sizes, and the addition and usage of newl y added carbon in soils from two long -term 94 contrasting agricultural practices: conventional chemical fertilization and biologically based management with cover crops. A greenhouse experiment using soil from these two different agricultural managements was planted with cereal rye ( Secale cereale L.), a common cover crop in the Midwestern US. Two soil treatments were considered, one with the origin al soil structure intact and one where soil structure was destroyed by sieving through 1 mm sieve. Destroying the existing structure destroy s the existing roots pores, allowing for the root effects on soil pore formation and carbon protection to be separated from root legacy effects and to be more easily detected . The rye was enriched with 13 C during 3 months of grow th via pulse labeling. The enriched rye was used to track newly added carbon within the soil after 3 months of rye growth and a subsequent 21 -day incubation. The first objective was to determine the localization of the new carbon added to soil by growing r ye roots. The second objective was to determine if and where this new carbon was lost in a subsequent incubation. The particular focus of the study was on evaluating the role of pores on new carbon localization and subsequent losses. 4.2. Materials and met hods 4.2.1 Soil c ollection Soil for the greenhouse experiment was collected from two different management practices established in 1989 at Kellogg Biological Station Long Term Ecological Research site , located in Hickory Corners, MI (42°24´N, 85°24´W) . The soil is a fine -loamy, mixed mesic Typic Hapludalf developed on glacial outwash with an intermixed loess layer of the Oshtemo and Kalamazoo series (Crum and Collins, 1995; Luehmann et al, 2016). The two practices were the conventional and biologically b ased systems. The conventional practice is a corn -soybean -wheat rotation maintained with current best management practices (tillage, chemical fertilizer additions, pesticide and herbicide as needed, no winter cover) . The biologically based practice is 95 a corn-soybean -wheat rotation with rye cover after corn and red clover inter -seeded into wheat. Rotary tillage was used between rows for weed control and no additional inputs were added. Detailed management and site information is available at http://lter.kbs. msu.edu/Data/LTER . Soil was collect ed between rows in May of 2016 during the wheat rotation prior to red clover establishment in the biologically based system. From each practice, soil was collect ed at the 0 -10 cm depth with minimal disturbance and placed into three 30 x 2 1 cm size containers to a depth of 8 cm. This soil will be referred to as intact -structure soil. Additional soil was collected from each treatment and sieved through a 1 mm sieve to destroy the existing soil structure. The sieved soil was placed in three 30 x 2 1 cm containers to an 8 cm depth for each practice. This soil will be referred to as destroyed -structure soil. Two containers from each management and ate a zone free of the immediate influence of plant roots. Rye was then hand planted at a 3 cm depth outside of the root excluding mesh every 4 cm with a total of nine rye plants per container. 4.2.2 Pulse labeling Enriched stable isotopes can be used to e asily track carbon in a system. Stable carbon 13 C/12 C ratio of the sample compared to a standard: = / 1000 (4.1) Pulse labeling is a technique used to uniformly label plant material via enriched 13 CO2 (Thompson, 1996; Bromand et al, 2001; Sangster et al, 2010). Pulse labeled plant material can then be tracked through the system using the following calculation: = () (4.2) 96 where F is the fraction of enriched material in the final sample, 13 Cfinal is the 13 C value of the sample, 13 Cenriched is the 13C of the enriched plant material, and 13 Cnon -enriched is the 13 C of the original material. Decomposition is believed to have a negligible effect on 13 C values, and therefore its effects can be safely ignored. Enriched material has previously been succe ssfully used to track the decomposition of organic material within soils at micron scale ( Gaillard et al, 1999, 2003; Toosi et al, 2017). Rye was grown in the greenhouse for a total of three months. Pulse labeling began two weeks after rye establishment an d was repeated every 10 days until the end of the three month growth period. At each labeling event, the rye containers were moved into a plexiglass chamber. One gram of 99% 13 C enriched CaCO 3 was placed in the chamber with a fan to circulate the 13 C labeled CO 2. The chamber was then sealed with duct tape to create an air tight enclosure (Figure 4.1). Excess H 2SO4 was added to the 99% 13 C enriched CaCO 3 to evolve CO 2. Plants were labeled for 24 hours (Bird et al, 2003; Toosi et al, 2017) and then remov ed from the plexiglass chamber until the next labeling event. A total of eight labeling events occurred during the experiment. The four containers where rye was not planted (1 per treatment combination) were not subjected to pulse labeling, but were watere d throughout the experiment as an isotopic control. 97 Figure 4. 1: Image of the pulse labeling plexiglass chamber and pulse labeling set up. Note the fan in the corner to 13C enriched CO 2. Samples were rotate d between rack positions for each pulse labeling event. 4.2.3 Sample collection At the end of the three month rye growth period, four intact mini -cores were taken from each container at 0 -5 cm depth using a beveled 3 mL Luer -Lok polypropylene syringe with an 8 mm inner diameter (BD, Franklin Lakes NJ, USA). In each container, two cores were taken from 98 the root exclusion zone and two adjacent to rye plants. Two additional cores per container were taken to calculate bulk density. The bulk density was use d to verify if mini -core sampling resulted in significant compaction of the samples. If compaction took place during sampling, the bulk density would be significantly higher than the 1.43 g/cm 3 bulk density of the original soil (Crum and Collins, 1995; Wic kings and Grandy, 2013). The cores were then air dried for Roots and shoots were collected and air 13 C to determine the enrichment of the rye in each container. All mini -cores w 13 C analysis. The other half was incubated for 21 days (see below), 13 C. The two halves will be referred to as Pre and Post (incubation) samples. 13 C analysis were collected using a custom -made soil sampling device (Figure 4. 13 device consisted of five 2 mm diameter and 5 mm deep soil cylinders . Approximately 10 - of soil was collected into each cylinder 13 C and total C analysis. 99 Figure 4. 2: Pictorial schematic of soil sampling using the sample device. First, the top 1.5 mm of the sample was removed (A) (B) . The soil sampling device (C) was then aligned with the red mark (D) and five samples collected simultaneously (E) . The samples were then placed into tins for total ca 13C analysis (F) . 4.2.4 Collection of CT images Images were obtained on the bending magnet beam line, station 13 -BM-D of the GeoSoilEnvironCARS (GSECARS) at the Advanced Photon Source (APS), Argon ne National Laboratory (ANL), IL . Images were collected with the Si (111) double crystal monochromator 100 tuned to 28 keV incident energy, the distance from sample to source was approximately 55 m, and the X -ray do se is estimated to be 1 kGy. Two -dimensional projections were taken at 0.25° rotation angle steps w ith a one second exposure and combined into a three -dimensional image consisting of 1198 slices with 192 0 by 1 920 pixels per slice for Pre scans . This resulted in a voxel size of 4.2 m. Post scans had 1200 slices of 1920 by 1920 pixels and result ed in a voxel size of 4.3 -processed by correcting for dark current and flat field and reconstructed using the GridRec fast Fourier transform reconstruction algorithm (Rivers, 2012). The indicator kriging method was utilized for segmentation of pore/solid in the images using 3DMA -Rock software (Oh and Lindquist, 1999; Wang et al, 2011). Total image porosity (pores > 4 4 were collected from each samp led section . The total image porosity was calculated as the percent of pore voxels withi n the sample voxels. Size distribution of image identified pores was determined using the burn number distribution approach in 3DMA -Rock (Lindquist et al, 2000 ; Ananyev a et al, 2013). Briefly, the burn number represents the shortest distance from the pore medial axis to the pore wall. For clarity, burn numbers have been converted into pore diameters. I specifically focused the data analyses on the pores of the following four diameter size ranges: 4- - 40-90 0 These pore sizes were chosen to match pore sizes previously studied in macro -aggregates that have demonstrated strong associations with carbon in the studied soil (Wang et al, 2012, 2013; Ananyeva et al, 2013; Kravchenko et al, 2014, 2015 ; Quigley et al, 2018a). 4.2.5 Incubation experimental design Prior to incubation, water was added from the top of the cores to achieve 50% of water filled pore space. Mini -cores were sealed at the bottom a nd placed into 10 ml vacutainers (BD 101 Franklin Lakes NJ, USA) with 1 mL of de -ionized water added to the bottom to maintain high humidity and consistent moisture levels. Samples were then incubated at 22.4±0.1°C for 21 days. Emission of CO 2 13 C measure ments were taken at day 1, 3, 7, 14, and 21. An LI -820 CO 2 infrared gas analyzer (Lincoln, Nebraska, USA) was used to take CO 2 measurements. After each CO2 13 C sampling, the headspace was flushed using CO 2 free air. One sample from conventional tillage intact -structure with root exclusion was lost and was not used in the analysis. 4.2.6 Total C and 13 C analyses 13 C and total C at the Stable Isotope Facility at the University of Calif ornia Davis. Soil samples were analyzed using an Elementar Vario EL Cube or Micro Cube elemental analyzer (Elementar Analysensysteme GmbH, Hanau, Germany) interfaced to a PDZ Europa 20 -20 isotope ratio mass spectrometer (Sercon Ltd., Cheshire, UK). Rye roo ts and shoots were analyzed using a PDZ Europa ANCA -GSL elemental analyzer interfaced to a PDZ Europa 20 -20 isotope ratio mass spectrometer (Sercon Ltd., Cheshire, UK). Gas samples were analyzed using a ThermoScientific GasBench system interfaced to a Ther moScientific Delta V Plus isotope ratio mass spectrometer (ThermoScientific, Bremen, Germany). The carbon isotopes are reported relative to Vienna PeeDee Belemnite (VPDB) with a 0.1› standard deviation for all samples . 4.2.7 Determination of POM and root p resence Both the amount of particulate organic matter (POM) and amoun t of roots were determined from , specifically from each image corresponding to the 2 mm section s sampled for 13 C. A piece of organic material was visually identified by graysc ale value, shape, 102 and size. Once designated organic material was identified, it was then classified as POM or root. Organic matter was designated as POM pieces if they were non -root shaped or, if root shaped, did not extend for more than a few hundred micr ons in any direction. In order to determine the true influence of roots on the sampled soil due to root exudation, roots were counted not only in the sample, but also in a 0.5 mm radius around the sample. 4.2.8 Total N, nitrate, and ammonium Samples collec ted for 13 C and total carbon were too small to determine total N, nitrate, and ammonia concentrations. Due to its lower concentration in soil, measurements for total N require more soil than total C; at least 20 g for reliable data. For total N, approximately 20 -40 g of soil was sampled from the side of the hole left by the mini -core sampling and analyzed at the Stable Isotope Facility at the University of California Davis on the same instrumentation as used for total C analysis. Two 10 g samples for nitrate and ammonium determination were collect from the root area of each container. Nitrate and ammonium were then extracted from the soil using a 1M KCl solution. Concentrations of nitrate and ammonium were determi ned using a Biotek Synergy H1 microplate reader. Nitrate was determined via vanadium(III) sulfanilamide and N -(1-naphthyl) -ethylenediamine dihydrocloride (NED) solution (Doane and Horw áth, 2003), while ammonium concentration was determined by salicylate an d cyanurate assay (Sinsabaugh et al, 2000). 4.2.9 Statistical analysis Comparisons between treatments were conducted using the mixed model approach implemented in the PROC MIXED procedure of SAS Version 9.4 (SAS Inc., 2009) . The analysis of 13 CO2 data col lected during the incubation used the statistical model with fixed effects of management history, structure, root presence (samples next to plants and 103 samples from root exclusion area), day (1, 3, 7, 14, and 21), and their interactions. The model also incl uded the random effects of containers nested within management history and structure as well as root by container interaction. Day was treated as a repeated measure in the analysis. The statistical model for the analysis of the soil samples for pores, tota 13 C consisted of the fixed effects of the management history, soil structure, root presence, incubation status (Pre and Post), and their interactions. The model included the random effects of container within management and structure as well as the interaction of root presence with container. The normality assumption was visually assessed using normal probability plots and stem -and -leaf plots, while equal variance assumption was assessed using Levene™s test. Where the equal variance assumptio n was violated, analysis with unequal variances was conducted (Milliken and Johnson, 2009). 13 C were conducted using the ANCOVA approach in the PROC MIXED procedure in SAS Version 9.4 (SAS Inc., 2009). Three data points from the biologically based Pre interaction were excluded from the analysis 13 C values, which were driven by the presence of root material within the samples. The regression s slopes significant at the 0.05 level were reported. Multiple regression analysis was performed with the PROC REG procedure in SAS Version 9.4 (SAS Inc., 2009) and adjusted R2 selection was used to determine the best model. 4.3. Results 4.3.1 Soil and plant characteristics Prior to rye planting, the only observed differences between soil characteristics were due to soil management history (Table 4.1). Soil from biologically based treatment had higher total C 104 and N levels and more depleted 13 C than soil from conventional management. The two soils did not differ in their bulk density. Table 4. 1: Bulk density in g/cm 3 13C (n=12), total C (n=12), and total N (n=12) from the soil before rye planting. Shown are means and sta ndard errors (in parenthesis). Letters indicate significant differences within each Management Structure Bulk Density 13 C Total C (%C) Total N (%N) Biological Intact 1.49(0.05)a -23.2(0.6)a 0.86(0.01)a 0.085(0.005)a Biological Destroyed 1.46(0.05)a -23.9(0.2)a 0.84(0.09)a 0.094(0.002)a Conventional Intact 1.37(0.05)a -22.3(0.6)b 0.59(0.02)b 0.066(0.005)b Conventional Destroyed 1.46(0.05)a -21.9(0.2)b 0.60(0.02)b 0.062(0.002)b The chemical characteristics of the rye roots grown in intact - and destroyed -structure soils were similar 15 N, which was more depleted in intact -structure soils (Table 4.2). The enrichment of the rye roots was uniform for all treatments and structures. Table 4. 2: Means of the chemical characteristics of pulse labeled rye roots (n=43). Standard errors are shown in Structure 13 C Total C (%C) 15 N Total N (%N) Intact 571(103)a 28(1)a 4.5(0.5)b 1.2(0.1)a Destroyed 521(103)a 25(2)a 1.7(0.5)a 0.9(0.1)a There were no observed differences between ammonia and nitrate concentrations in the soils after 3 months of rye growth (Table 4.3). Total nitrogen was higher in biologically based management history samples. Intact -structure with roots in conventional management was lower than the other conventional management history treatments. Intact -structure conventional with roots was the only treatment to not gain nitrogen during rye growth. After 3 months of rye growth, total carbon was higher in the biologically based management with roots, but not without roots compared to the conventional management. 105 Table 4. 3: Total C (n=40), Total N (n=40), ammonia (n=16), and nitrate (n=16) means from the soil after rye growth. Standard errors are shown in parenthesis. Letters indicate significant differences within each column at Management Structure Root Total N (%N) Total C (%C) Ammonia (ppm) Nitrate (ppm) Biologica l Intact Root 0.094(0.007)a 1.00(0.06)a 0.13(0.03)a 0.19(0.02)a Biological Intact No Root 0.108(0.006)a 0.88(0.05)a d Biological Destroyed Root 0.101(0.004)a 0.97(0.02)a 0.15(0.03)a 0.14(0.02)a Biological Destroyed No Root 0.094(0.004)a 0.89(0.06)a d Conventional Intact Root 0.067(0.006)c 0.64(0.04)a b 0.13(0.03)a 0.14(0.02)a Conventional Intact No Root 0.088(0.007)ab 0.79(0.08)b d Conventional Destroyed Root 0.074(0.004)bc 0.77(0.04)b 0.13(0.03)a 0.13(0.02)a Conventional Destroyed No Root 0.072(0.004)bc 0.71(0.02)b d Root presence decreased the amount of POM in all managements and structures (Table 4.4). Biologically based management history had significantly more POM than conventional. The amount of roots did not vary significantly between any studied treatments. Table 4. 4: errors are shown in parenthesis. Letters indicate significant differences within each colum Management Structure Root POM Roots Biological Intact Root 19.1(1.5)ab 6.7(4.0)a Biological Intact No Root 30.4(3.1)c Biological Destroyed Root 21.9(1.5)a 16.1(4.3)a Biological Destroyed No Root 44.2(3.1)d Conventional Intact Root 7.1(0.9)e 19(4.2)a Conventional Intact No Root 15.2(1.5)bd Conventional Destroyed Root 7.9(0.9)e 12(4.0)a Conventional Destroyed No Root 14.5(1.5)d 4.3.2 Pore characteristics Total image porosity of the soil cores ranged from 12 to 32%. Differences between structure and Pre and Post were observed in the 4-15, 15-40, and 90+ (Figure 4.3) . 106 Increases in 4 -15 m pores Post were observed for both structures with increases greater in destroyed -structure soils . Both 15 -40 and 90+ m por es had differences between the structures, but where 90+ m pores saw an increased presence in intact soils, 15 -40 m pores saw a decreased presence. Figure 4. 3: Relative abundances of 4 Œ15, 15 Œ40, 40 Œ incubation and management. Relative pore abundance refers to the percent of medial axes per total soil volume as determined from 3DMA -Rock software. Bars represent standard errors. Letters indicate significant differences at a = 0.05. 4.3.3 Associations between pores and new carbon Significant positive correlations between new carbon and pores were only observed in soils prior to incubation and under the influence of roots. Structure was the main driver of observed associations. In intact -structure soils, increases in new carbon were significant in 107 relation to all observed pore sizes (Figure 4.4). On the other hand, destroyed -structure soils saw significant increases in new carbon only in conjunction with 15 -40 and 40 - 4.5). Figure 4. 4: 13C and relative abundances of 4 Œ (A) , 15 - (B) , 40 - pores (C) (D) for the intact -structure with root for both Pre and Post. Relative pore abundances refer to the percent of medial axes per total soil volume as determined from 3DMA -Rock software. Outliers removed from analysis are circled in orange. Stars next to the end Gray area indicates 95% confidence interval. 108 Figure 4. 5: 13C and relative abundances of 4 Œ (A) , 15 - (B) , 40 - pores (C) , an (D) for destroyed -structure soils when roots are present during both Pre and Post. Relative pore abundances refer to the percent of medial axes per total soil volume as determined from 3DMA -Rock software. Stars next to the end of the lines i interval. After incubation, intact -structure soils lost substantial amounts of new carbon from the two largest pore sizes, but reduced losses were observed in the two smallest pore sizes. In contrast, destroyed -structure soils lost considerable amounts of new carbon from both 15 -40 and 40- 109 Multiple regression indicated that the model explaining the most variation for the 13 C and pore size distributio n for intact -structure soil prior to incubation 13 C and 40 - The equation for this model was: y=1306 * (40 -90) - 6.54 (4.3) This only had an R 2 of 0.39, indicating that this did not model well the variation in 13 C values. For destroyed -structure soil, the model that explained the most variation included 4 -15, 15-40, and 40 -90 m pores. The equation for this model was: y=886 * (40 -90) + 321 * (15 -40) - 40 * (4 -15) + 2.7 (4.4) This model had an R 2 13 C values. Both 4 -15 and 15 - -value of 0.02 and 0.03, respectively). 4.3.4 Util ization of carbon during incubation Prior to the incubation of the soil cores, the only significant factor in regards to amount of new carbon added to the soil by rye was whether roots were present or not. This indicates that root presence was the main con tributor to new carbon incorporation into the soil during rye growth. After incubation, while root presence was still a significant factor, management history and structure played a larger role in whether new carbon was utilized within each soil by microor ganisms during the incubation. For example, the conventional management history with intact -structure and roots exhibited a minimal loss of new carbon, while biologically based intact -structure with roots lost a substantial portion of its new carbon (Figur e 4.6a). In the destroyed -structure, biologically based management history lost an insubstantial amount of new carbon, while the conventional lost a noteworthy amount of new carbon. The amount of new 110 carbon added during rye growth without roots was not sig nificantly different from zero (Table 4.5). Figure 4. 6: (A) and released as CO 2 during incubation (B) . Bars indicate standard errors. Stars indicate significant differences between Pre and Post for soil carbon and intact - and destroyed -structure for CO 2 111 Table 4. 5: parenthesis. No statistical differences were found. Management Structure Incubation g New Carbon Biological Intact Pre 1.1(1.9) Biological Intact Post 0.5(1.7) Biological Destroyed Pre 0.6(1.8) Biological Destroyed Post 0.9(1.7) Conventional Intact Pre 1.6(1.9) Conventional Intact Post -0.9(2.0) Conventional Destroyed Pre 1.3(1.9) Conventional Destroyed Post 1.0(1.7) During the incubation, intact -structure biologically based management history with roots lost the most new carbon, while destroyed -structure biologically based without roots lost the least new carbon (Figure 4.6b). An interesting pattern emerged between the two management histories. Biologically based, with roots present, lost more new carbon in intact -structure soils, while conventional with roots lost the most new carbon in destroyed -structure. However, when roots were not present, biologically based soil lost more new carbon in destroyed treatments, while conventional soil saw no differences between structures. 4.4. Discussion In intact -structure soil, th at is, in soil with legacy root channels, additions of new carbon were equally associated with all studied pore sizes. However, in destroyed -structure soils, where older root channels were not present, new carbon was positively associated only with the 15 -90 of new carbon occurred in those same pores where it was initially preferentially added. 4.4.1 Carbon addition during rye growth In intact -structure soi l new carbon was positively associated with all studied pore sizes (Figure 4.4 ). However, multiple regression indicated that the association with presence of 40 -90 the strongest (Equation 4.3) . s 112 been shown that they can only enter pores of this size or greater (Wiersum, 1957; Cannell, 1977). Therefore, a direct contribution of rye plant roots likely occur red within the 40 - pores. Previous work also has shown these pores to be asso ciated with new carbon additions (Quigley et al, 2018a). However, the R 2 value is low, which indicates substantial contribution of other factors to the observed 13 C values. One of such confounding factors is preferential root growth into alre ady established pores of old root channels (Rasse and Smucker, 1998). If the growth of roots of specific size are a controlling fa ctors in new carbon addition, but the roots grow through pores of larger size , this would uncouple root size and pore size, potentially dis missing pore size limitations on new root carbon additions. This preferential root growth may also explain the association of new carbon Fungal growth and transport is another factor that may explain the inc reases in new carbon associate d with the 15 - (Figure 4.4b, 4.5b) . These pores are too small for plant roots to enter, but fungal mycelia can enter these pores and they have been known to push aside silt particles to create 20 - rioz et al, 1993; Bearden, 2001; Emerson and McGarry, 2003). In intact -structure soil of this study , this relationship would be associated with carbon transport through an existing fungal network. Fungi are known t o be able to transport carbon great distances (Godbold et al, 2006) and some are reliant on plants for lipids and other building materials (Luginbuehl et al, 2017). This means that a well -estab lished fungal network can and does transport new carbon throughout its hyphae network . In destroyed -structure soils, on the other hand, the fungal network was broken by sieving and, therefore, the associations with new carbon and 15 -40 m pores is more likely related to the re -establishment of the fungal network rather than transport through the fungal network. 113 Unlike intact -structure soil, the multiple regression model that explained the most variation in destroyed -structure soils included 4 -15, 15 -40, and 40 - -15 and 15- -40 and 40- pores were positively associated with new carbon, 4 - with new carbon. This indicates that new carbon was unable to enter 4 - Roots would not be able to penetrate 4 -15 m pores and, while fungal mycelia can partially enter these pores, the destruction of the fungal network would limit the extent that fungi could deposit new carbon into these pores. Additionally, these pores would be potential reservoirs of protected carbon due to the prevalence of anaerobic conditions, potentially decreasing new carbon preservation in destroyed soils, which was seen in the conventional management soils (Figure 4.6). Keiluweit et al (2017) found that anaerobic micro -sites in upland soils can retard decomposition by 10-fold, effectively protecting carbon. Pores of 4 - anaerobic conditions ( Schurgers et al, 2006). The creation of new root pores, instead of using older root channels, may have also retarded the transport of new carbon from roots to the surrounding soil . Quigley et al (2018a) found increased grayscale values in the vicinity of root pores in destroyed -structure soils. This was believed to be due to compaction as the root grew. This compacted soil around the root would restrict flow of carbon from new roots to the rest of the soil and, therefore, prevent the spread of new carbon to smaller pores. 4.4.2 Carbon utilization during incubation The obse rved positive relationships between pores and carbon were lost after in cubation (Figure 4.4, 4.5 ). This indicates that either gains of new carbon were lost as CO 2 or that they were distributed more evenly through the soil. Due to the significant loss of ne w carbon observed 114 in soil from biologically based management with intact -structure and in soil from conventional management with destroyed -structure (Fig ure 4.6 ), the loss as CO 2 is more likely the explanation. On the other hand, the relative lack of carbo n loss in biologically based destroyed -structure and conventional management intact -structure (Fig ure 4.6 ) indicate s redistribution in the soil. The difference in biologically based management may relate to availability of older carbon. Soil from biologica lly based management had higher POM as compared to conventional (Table 4.4) . However, this POM would mostly likely be physically protected and inaccessible to decomposers in intact -structure soils, while fresh roots would be more easily available, increasi ng new carbon usage. In destroyed -soils, POM would have lost its physical protection and therefore, was more available for decomposition. This could result in preferential decomposition of POM over fresh roots in these soils, resulting in less new carbon l osses. In soils from conventional management, POM is sparse (Table 4.4) and, therefore, could not explain the difference between new carbon utilization in intact -structure and destroyed -structure soils. With little available POM, new material would prefere ntially be used in both structures. However, the compaction around roots in destroy ed-structure soil would restrict carbon flow out of the 40 - (Chenu et al, 2001; Strong et al, 2004; Ruamps et al, 2011). This restriction to pores with increased microbial activities would promote complete oxidation of root material to CO 2 over preservation of the intermediates. However, in intact -structure soil, roots preferentially grew in old root channels, which Quigley et al (2018b) showed did not have adjacent soil compaction. This would not restrict carbon to 40 - preservation was more likely to occur. 115 While both intact -structure and destroyed -structure lo st new carbon from soil pores during incubation, intact -structure lost substantial amount of new carbon, specifically from 40-90 and >90 , resulting in a flat non -significant associatio n with these pores (Figure 4.4 ). However, 4-15 and 15 -40 m pores were still positively related, potentially indicating preferential new carbon preservation in these pores. Both Quigley et al (2018a) and Ananyeva et al (2013) found increased carbon in association with 15 -40 m pores, which may support this postul ate of preferential carbon preservation in soils with intact -structure within these pores. Destroyed -structured soil preferentially lost new carbon from both 15 -40 and 40 -90 m pores after incubation (Figure 4.5) , which is consistent with previous investig ations of destroyed -structure soils (Quigley et al, 2018a). 4.4.3 POM, roots, and nitrogen The numbers of POM fragments in the studied soil revealed an i nteresting pattern (Table 4.4). Consistent with previous reports, POM was approximately twice as abunda nt in biologically based management compared to conventional (Kravchenko et al, 2014). However, the difference between the treatments with and without roots was consistent with the concept of priming. Priming is the addition of new carbon stimulating the d ecomposition of older carbon (Kuzyakov et al, 2000; Fontaine et al, 2004; Schimel and Schaeffer, 2012; Blagodatskaya et al, 2014). POM in the treatments without roots were approximately twice as prevalent as treatments with roots, indicating priming took p lace during rye growth (Table 4.4) . The finding of similar nitrate and ammonium concentrations were unexpected (Table 4.3). Total soil nitrogen was lower in the conventional management prior to rye planting (Table 4.1). This would indicate less sources of nitrogen available as nitrate and ammonium to the plants, since no fertilizer was added to the soil during rye growth. Additionally, sieving the soil 116 would have exposed protected organic matter and, therefore, potentially increased soil nitrogen mineraliza tion ( Fukumasu and Shaw, 2017 ). However, rye is known as an efficient nitrogen scavenger ( Staver and Brinsfield, 1990; Strock et al, 2004) and, therefore, may have been responsible for the similar nitrate and ammonium concentrations present after its growt h. The pattern of total nitrogen was as predicted in the conventional inta ct-structure soils (Table 4.1, 4.3 ). Rye growth, due to its efficiency as a nitrogen scavenger, was expected to lower nitrogen levels in the soil. This was observed in the convention al intact -structure soil. However, the increases in total nitrogen in the biologically based management soils and destroyed -structure conventional management soils during rye growth was unanticipated. I hypothesize that free living nitrogen fixing bacteria in the soil are responsible for this increase. Increases were more prevalent in the biologically based management. While activity of free living nitrogen fixing bacteria has not been measured in these soils, overall microbial activity is known to be incre ased in the biologically based management (Xue et al, 2013). This would result in more nitrogen fixation occurring in this management. 4.4.4 Soil aggregates vs. intact soil cores Prior to the use of CT images, research on soil carbon was limited to measuring the distribution of soil aggregates (Six et al, 2000). While useful as an indicator of soil carbon addition ( Six et al, 1999; Denef et al, 2004; 2007 ), underlying mechanisms were unknown as the relationship between pores and aggregates, while the oretically inversely related, was ambiguous (Six and Paustian, 2014 ). Furthermore, it was unknown if soil pores have the same function in aggregates of different size fractions and, therefore, it is unclear if aggregate data can be extrapolated to the int act soil ( Young and Ritz, 2000; Young et al, 2001 ). 117 This study, along with Quigley et al (2018 a), and Ananyeva et al (2013), may help bridge the gap between soil pore studies based on CT images and aggregate studies. The results from all three studies ar e consistent in the relationships between pores and carbon, however, Ananyeva et al (2013) used strict aggregates, Quigley et al (2018) used soil fragments, and this study used intact mini -cores. Consistent results obtained in all three studies suggest tha t pores behave the same, in regards to carbon, regardless of wheth er located in aggregates or intact soil. Furthermore, the behavior of pores was independent of management . This indicates that aggregate distributions, if they can be connect to pore distrib utions, would be useful as a proxy, at least for carbon. 4.5 Conclusion Distributions of new soil carbon in relation to pores was different before and after incubations. Prior to incubation, the disturbace level of soil structure was the major regulator of new soil carbon additions. In intact -structure soils, new carbon addition was associated with all studied pore sizes, potentially directed by the in fluence of legacy root channels and fungal activities with 40 -90 m pores being important new carbon source s. In destroyed -structure soils, new carbon addition was related to 4 -15, 15-40, and 40 - ith 15 -40 and 40 -90 m pores having a positive association while 4 -15 m pores having a negative association with new carbon . This indicates that new carb on is unable to reach the 4 -15 m pores in destroyed -structure soils . Overall, pores play a large role in new carbon addition but can be masked when legacy root channels are present. After incubation, no associations between new carbon and pore sizes were observed. Management history, seemed to play a larger role in whether new carbon was utilized or not. Biologically based management with intact -structure and conventional with destroyed -structure 118 both lost significant amounts of new carbon during the incub ation. Biologically based with destroyed -structure and conventional with intact -structure lost negligible amounts of new carbon. These patterns may relate to POM availability and root carbon mobility. Losses of carbon were most visible in relation to 40 -90 addition and carbon loss. This association was independent of management, indicating a universal mechanism for these pores. Funding Support for this research was provided in part by the USDA -NIFA, Award No. 2016 -67011-24726 fiUsing stable isotopes and computer tomography to determine mechanisms of soil carbon protection in cover crop based agricultural systemsfl and by the US National Science Foundation Long -Term Ecological Research Program (DEB 1 027253) at the Kellogg Biological Station and by Michigan State University AgBioResearch. Portions of this work were performed at GeoSoilEnviroCARS (The University of Chicago, Sector 13), Advanced Photon Source (APS), Argonne National Laboratory. GeoSoilEn viroCARS is supported by the National Science Foundation - Earth Sciences (EAR - 1634415) and Department of Energy - GeoSciences (DE -FG02 -94ER14466). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Sci ence User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE -AC02 -06CH11357. 119 REFERENCES 120 REFERENCES Abraha, M., Hamilton, S. K., Chen, J., and Robertson, G. P. (2018) . Ecosystem carbon exchange on conversion of Conservation Reserve Program grasslands to annual and perennial cropping systems. Agric For Meteorol. 253-254, 151-160. doi: 10.1016/j.agrformet.2018.02.016 Ananyeva, K., Wang, W., Smucker, A. J. M., Rivers, M. L., and Kravchenko, A. N. (2013). Can intra -aggregate pore structures affect the aggregate™s effectiveness in protecting carbon? Soil Biol. Biochem. 57, 868-875. doi: 10.1016/j.soilbio.2012.10.019 Austin, E. E., Wickings, K., McDaniel, M. D., Robertson, G. P., and Grandy, S. (20I7). Cover crop root contributions to soil carbon in a no -till corn bioenergy cropping system. GCB Bioenergy 9, 1252-1263. doi: 10.1111/gcbb.12428 Bailey, V. L., Smith, A. P., Tfaily, M., Fansler, S. J., and Bond -Lamberty, B. (2017). Differences in soluble organic carbon chemistry in pore waters sampled from different pore size domains. Soil Biol. Biochem. 107, 133-143. doi: 10.1016/j.soilbio.2016.11.025 Barber, D. and Martin, J. (1976). Release of organic substances by cereal roots i nto soil. New Phytol . 76, 69-80. doi: 10.1111/j.1469 -8137.1976.tb01439.x Bauer, A. and Black, A. L. (1994) . Quantification of the effect of soil organic matter content on soil productivity. Soil Sci. Soc. Am. J. 58, 185-193. doi: 10.2136/sssaj1994.03615995 005800010027x Bearden, B. N. (2001). Influence of arbuscular mycorrhizal fungi on soil structure and soil water characteristics of vertisols. Plant Soil 229, 245Œ 258. doi: 10.1023/A:1004835328943 Bengough, A. G., Bransby, M. F., Hans, J., McKenna, S. J., Roberts, T. J., and Valentine, T. A. (2006) . Root responses to soil physical conditions; growth dynamics from field to cell. J. Exp. Bot. 57, 437-447. doi: 10.1093/jxb/erj003 Bird, J. A., van Kessel, C. , and Horwath, W. R. (2003). Stabilization of 13 C-carb on and immobilization of 15 N-nitrogen from rice straw in humic fractions. Soil Sci. Soc. Am. J. 67, 806-816. doi: 10.2136/sssaj2003.0806 Blagodatskaya, E., Khomyakov, N., Myachina, O., Bogomolova, I., Blagodatsky, S., and Kuzyakov, Y. (2014). Microbial int eractions affect sources of priming induced by cellulose. Soil Biol. Biochem. 74, 39-49. doi: 10.1016/j.soilbio.2014.02.017 Bowen, H. D. (1981) . fiAlleviating mechanical impedancefl, in Modifying the Root Environment to Reduce Crop Stress , eds. G. F. Arkin a nd H. M. Taylor (St. Joseph, MI: Am. Soc. Agric. Engineers) 21 -57. Brimecombe, M. J., de Leij, F. , and Lynch, J. M. (2001). fiThe effect of root exudates on rhizosphere microbial populations,fl in The rhizosphere: biochemistry and organic substances at 121 the s oil -plant interface , eds. R. Pinton, Z. Varanini and P. Nannipieri. (New York, New York, USA: Marcel Dekker), 95 Œ140. Cannell, R. Q. (1977). Soil aeration and compaction in relation to root growth and soil management. Appl. Biol . 2, 1-86. Chenu, C., Hassin k, J., and Bloem, J. (2001) . Short -term changes in the spatial distribution of microorganisms in soil aggregates as affected by glucose addition. Biol. Fertil. Soils 34, 349-356. doi: 10.1007/s003740100419 Chimento, C., Almagro, M., and Amaducci, S. (2016) . Carbon sequestration potential in perennial bioenergy crops: the importance of organic matter in puts and its physical protection. GCB Bioenergy 8, 111-121. doi: 10.1111/gcbb.12232 Crum, J. R. and Collins, H. P. (1995) . KBS Soils [Online]. W. K. Kellogg B iological Station Long -Term Ecological Research Project, Michigan State University, Hickory Corners, MI. http://www.lter.kbs.msu.edu/soil/characterization. Davidson, E. A. and Janssens, I. A. (2006). Temperature sensitivity of soil carbon decomposition and feedbacks to climate change. Nature 440, 165-173. doi: 10.1038/nature04514 Denef, K., Six, J., Merckx, R ., and Paustian, K. (2004). Carbon sequestration in micro -aggregates of no -tillage soils with different clay mineralogy. Soil Sci. Soc. Am. J. 68, 1935-1944. doi: 10.2136/sssaj2004.1935 Denef, K., Zotarelli, L., Boddey, R. M., and Six, J. (2007). Microaggregate -associated C as a diagnostic fraction for management -induced changes in soil organic carbon in two Oxisols. Soil Biol. Biochem . 39, 1165-1172. doi: 10.1016/j.soilbio.2006.12.024 Dexter, A. R. (1987) . Compression of soil around roots. Plant Soil 97, 401-406. doi: 10.1007/BF02378351 Doane, T. A. and Horw áth, W. R. (2003). Spectroscopic determination of nitrate with a single reagent. Anal. Lett. 36, 2713 -2722. doi: 10.1081/AL -120024647 Dorioz, J. M., Robert, M., and Chenu, C. (1993). The role of roots, fungi and bacteria on clay particle organization. An experimental approach. Geoderma 56, 179Œ 194. doi:10.1016/0016 -7061(93)90109 -X Dungait, J. A. J., Hopkins, D. W., Gregory, A. S., and Whitmore, A. P. (2012). Soil organic matter turnover is governed by accessibility not recalcitrance. Global Change Biol. 18, 1781- 1796. doi: 10.1111/j.1365 -2486.2012.02665.x Ekschmitt, K., Kandeler, E., Poll, C., Br une, A., Buscot, F., Friedrich, M., et al. (2008). Soil -carbon preservation through habitat constraints and biological limitations on decomposer activity. J. Plant Nutr. Soil Sci. 171, 27-35. doi: 10.1002/jpln.200700051 122 Ekschmitt, K., Liu, M., Vetter, S., Fox, O., and Wolters, V. (2005). Strategies used by soil biota to overcome soil organic matter stability Œ why is dead organic matter left over in soil? Geoderma 128, 167-176. doi: 10.1016/j.geoderma.2004.12.024 Emerson, W. W. and McGarry, D. (2003). Organ ic carbon and soil porosity. Aust. J. Soil Res. 41, 107Œ118. doi: 10.1071/SR01064 Falconer, R. E., Battaia, G., Schmidt, S., Baveye, P., Chenu, C., and Otten, W. (2015). Microscale heterogeneity explains experimental variability and non -linearity in soil o rganic matter mineralization. PloS One . 10: e0123774. doi: 10.1371/journal.pone.0123774 Falkowski, P., Scholes, R. J., Boyle, E., Canadell, J., Canfield, D., Elser, J., et al. (2000). The global carbon cycle: a test of our knowledge of Earth as a system. Science . 290, 291 -296. doi: 10.1073/pnas.0811302106 Fontaine, S., Bardoux, G., Abbadie, L., and Mariotti, A. (2004). Carbon input to soil may decrease soil carbon content. Ecol Lett 7, 314-320. doi: 10.1111/j.1461 -0248.2004.00579.x Fukumasu, J. and Shaw, L. (2017). The role of macro -aggregation in regulating enzymatic depolymerization of soil organic nitrogen. Soil Bio. Biochem. 115, 100-108. doi: 10.1016/j.soilbio.2017.08.008 Gaillard, V., Chenu, C., and Recous, S. (2003). Carbon mi neralization in soil adjacent to plant residues of contrasting biochemical quality. Soil Biol. Biochem. 35, 93-99. doi: 10.1016/S0038 -0717(02)00241 -9 Gaillard, V., Chenu, C., Recous, S. , and Richard, G. (1999). Carbon, nitrogen and microbial gradients indu ced by plant residues decomposing in soil. Eur. J. Soil Sci. 50, 567 -578. doi: 10.1046/j.1365 -2389.1999.00266.x Gale, W. J., Cambardell, C. A., and Bailey, T. B. (2000) . Surface residue - and root -derived organic matter under simulated no -till. Soil Sci. So c. Am. J. 64, 196-201. doi: 10.2136/sssaj2000.641196x Garten, C. T. and Wullschleger, S. D. (1999). Soil carbon inventories under a bioenergy crop (switchgrass): measurement limitations. J Environ. Qual 28, 1359-1365. doi: 10.1016/j.biombioe.2010.08.013 Godbold, D. L., Hoosbeek, M. R., Lukac, M., Cotrufo, M. F., Janssens, I. A., Ceulemas, R., et al. (2006). Mycorrhizal hyphal turnover as a dominant process for carbon input into soil organic matter. Plant Soil 281, 15-24. doi: 10.1007/s11104 -005-3701-6 Graec en, E. L., Farrell, D. A., and Cockroft, B. (1968) . Soil resistance to metal probes and plant roots. 9th International Congress of Soil Science , Angus and Roberton, Adelaide. 769 -779. Grandy, A. S. and Neff, J. C. (2008). Molecular C dynamics downstream: t he biochemical decomposition sequence and its impact on soil organic matter structure and function. Sci Total Environ 404, 297-307. doi: 10.1016/j.scitotenv.2007.11.013 123 Grandy, A. S. and Robertson , G. P. (2007). Land -use intensity effects on soil organic c arbon accumulation rates and mechanisms. Ecosystems 10, 58-73. doi: 10.1007/s10021 -006-9010-y Jackson, R. B., Lajtha, K., Crow, S. E., Hugelius, G., Kramer, M. G., and Piñeiro, G. (2017). The ecology of soil carbon: pools, vulnerabilities, and biotic and abiotic controls. Annu Rev Ecol Evol Syst . 48, 419-445. doi: 10.1146/annurev -ecolsys -112414-054234 Kallenbach, C. M., Frey, S. D., and Grandy, A. S. (2016). Direct evidence for microbial -derived soil organic matter formation and its ecophysiological controls. Nat. Commun. 7:13630. doi : 10.1038/ncomms13630 Kallenbach, C. M., Grandy, A. S., Frey, S. D ., and Diefendorf, A. F. (2015) . Microbial physiology and necromass regulate agricultural soil carbon accumulation. Soil Biol. Biochem 9, 279-290. doi: 10.1016/j.soilbio.2015.09.005 Keel, S. G., Liefeld, J., Mayer, J., Taghizadeh -Toosi, A., and Olesen, J. E. (2017). Large uncertainty in soil carbon modelling related to method of calculation of plant carbon input in agricultural systems. Eur. J. Soil Sci. 68, 953-963. doi: 10.1111/ejss.12454 Keiluweit, M., Wanzek, T., Kleber, M., Nico, P., and Fendorf, S. (2 017). Anaerobic microsites have an unaccounted role in soil carbon stabilization. Nat. Commun. 8: 1771. doi: 10.1038/s41467 -017-01406-6 Kell, D. B. (2012) . Large -scale sequestration of atmospheric carbon via plant roots in natural and agricultural ecosyste ms: why and how. Philos T R Soc B . 367, 1589-1597. doi: 10.1098/rstb.2011.0244 Kong, A. Y. Y. and Six, J. (2010) . Tracing root vs. residue carbon into soils from conventional and alternative cropping systems. Soil Sci. Soc. Am. J. 74, 1201-1210. doi: 10.2136/sssaj2009.0346 Kravchenko, A. N ., and Guber, A. K. (2017). Soil pores and their contributions to soil carbon processes. Geoderma 287, 31 -39. doi: 10.1016/j.geoderma.2016.06.027 Kravchenko, A. N., Negassa, W. C., Guber, A. K., Hildebrandt, B., Marsh, T. L., and Rivers, M. L. (2014). Intra -aggregate pore structure influences phylogenetic composition of bacterial community in macroaggregates. Soil Sci. Soc. Am. J. 78, 1924-1939. doi: 10.2136/sssaj2014.07.0308 Kravchenko, A. N., Negassa, W. C., Guber, A. K., and Rivers, M. L. (2015). Protection of soil carbon within macro -aggregates depends on intra -aggregate pore characteristics. Sci. Rep. 5:16261 . doi: 10.1038/srep16261 Kuzyakov, Y., and Blagodatskaya, E. (2015). Microbial hotspots and hot moments in soil: concept & review. Soil Biol. Biochem. 83, 184-199. doi: 10.1016/j.soilbio.2015.01.025 Kuzyakov, Y. , Friedel, J. K., and Stahr, K. (2000). Review of mechanisms and quant ification of priming effects. Soil Biol. Biochem. 32, 1485-1498. doi: 10.1016/S0038 -0717(00)00084 -5 124 Kuzyakov, Y., Leinweber, P., Sapronov, D., and Eckhardt, K. U. (2008). Qualitative assessment of rhizodeposits in non -sterile soil by analytical pyrolysis. J. Plant Nutr. 166, 719-723. doi: 10.1002/jpln.200320363 Lal, R. (1999). Soil management and restoration for carbon sequestration to mitigate the accelerated greenhouse effect. Prog. Env. Sci. 1, 307 -326. doi: 10.1055/s -2008-1072270 Lal, R . (2006) . Enhanci ng crop yields in the developing countries through restoration of the soil organic carbon pool through restoration of the soil organic carbon pool in agricultural lands. Land Degrad. Develop. 17, 197-209. doi: 10.1002/ldr.696 Lal, R. (2011) . Sequestering c arbon in soils of agro -ecosystems. Food Policy . 36, S33 -S39. doi: 10.1016/j.foodpol.2010.12.001 Lindquist, W. B., Venkatarangan, A., Dunsmuir, J., and Wong, T. (2000). Pore and throat size distributions measured from synchrotron X -ray tomographic images of Fontainebleau sandstornes. J. Geophys. Res. 105, 21,509-21,527. doi : 10.1029/2000JB900208 Luehmann, M. D., Peter, B. G ., Connallon, C. B., Schaetzl, R . J., Smidt, S. J., Liu, W., et al. (2016). Loamy, two -storied soils on the outwash plains of southwester n lower Michigan: Pedoturbation of loess and the underlying sand. Ann Am Assoc Geogr. 106, 551 -572. doi: 10.1080/00045608.2015.1115388 Luginbuehl, L. H., Menard, G. N., Kurup, S., Van Erp, H., Radhakrishnan, G. V., Breakspear, A., et al. (2017). Fatty aci ds in arbuscular mycorrhizal fungi are synthesized by the host plant. Science . 356, 1175-1178. doi : 10.1126/science.aan0081 Mazzilli, S. R., Kemanian, A. R., Ernst, O. R., Jackson, R. B., and Piñero, G. (2015) . Greater humification of belowground than abov eground biomass carbon into particulate soil organic matter in no -till corn and soybean crops. Soil Bio. Biochem. 85, 22-30. doi: 10.1016/j.soilbio.2015.02.014 Meharg, A. A. and Killham, K. (1991). A novel me thod of quantifying root exudation in the presen ce of soil microflora. Plant Soil 113, 111-116. doi: 10.1007/BF00010488 Melsted, S. W. (1954) . New concepts of management of corn belt soils. Adv. Agron. 6, 121-142. doi: 10.1016/S0065 -2113(08)60383 -1 Milliken, G .A. and Johnson, D. E. (2009). Analysis of Messy Data Volume I: Designed Experiments, second ed. Boca Raton, FL: CRC Press. Necpálová, M., Anex Jr., R. P., Kravchenko, A. N., Abendroth, L. J., Del Grosso, S. J., Dick, W. A., et al. (2014). What doe s it take to detect a change in soil carbon stock? A regional comparison of minimum detectable difference and experiment duration in the north central United States. J Soil Water Conserv. 69, 517-531. doi: 10.2489/jswc.69.6.517 125 Oechel, W. C. and Vourlitis, G. L. (1994). The effects of climate change on land -atmosphere feedbacks in arctic tundra regions. Trends Ecol Evol . 9, 324-329. doi: 10.1016/0169 -5347(94)90152 -X Oh, W., and Lindquist, W. B. (1999). Image thresholding by indic ator kriging. IEEE Trans. Pattern Anal. Mach. Intell. 21, 590-602. doi: 10.1109/34.777370 Park, E. J., Sul, W. J., and Smucker, A. J. M. (2007). Glucose additions to aggregates subjected to drying and wetting cycles promote carbon sequestration and aggrega te stability. Soil Biol. Biochem. 39, 2758-2768. doi: 10.1016/j.soilbio.2007.06.007 Paustian, K., Lehmann, J., Ogle, S., Reay, D., Robertson, G. P., and Smith, P. (2016) . Climate -smart soils. Nature 532, 49-57. doi: 10.1038/nature17174 Quigley, M. Y., Nega ssa, W. C., Guber, A. K., Rivers, M. L., and Kravchenko, A. N. (2018a). Influence of pore characteristics on the fate and distribution of newly added carbon. Front. Environ. Sci. 6:51. doi: 10.3389/fenvs.2018.00051 Quigley, M. Y., Rivers, M. L., and Kravch enko, A. N. (2018b) . Patterns and sources of spatial heterogeneity in soil matrix from contrasting long term management practices. Front. Environ. Sci. 6:28 doi: 10.3389/fenvs.2018.00028 Raun, L. and Robertson, G. P. (2013). Initial nitrous oxide, carbon d ioxide, and methane costs of converting Conservation Reserve Program grassland to row crops under no -till vs. conventional tillage. Global Change Biol. 19, 2478-2489. doi: 10.1111/gcb.12216 Rabot, E., Wiesmeier, M., Schlüter, S., and Vogel, H. -J. (2018) . Soil structure as an indicator of soil functions: a review. Geoderma 314, 122-137. doi:10.1016/j.geoderma.2017.11.009 Rasse, D. P., Rumpel, C. and Dignac, M. -F. (2005) . Is soil carbon mostly root carbon? Mechanisms for a specific stablisation. Plant Soil 269, 341-356. doi: 10.1007/s11104 -004-0907-y Rasse, D. P. and Smucker, A. J. M. (1998). Root recolonization of previous root channels in corn and alfalfa rotations. Plant Soil 204, 203-212. doi: 10.1023/A:1004343122448 Rivers, M. L. (2012). tomoRecon: Hig h-speed tomography reconstruction on workstations using multi -threading. Proc. SPIE 8505:OU. doi:10.1117/12.930022 Rorick, J. D. and Kladivko, E. J. (2017) . Cereal rye cover crop effects on soil carbon and physical properties in southeastern Indiana. J Soi l Water Conserv. 72, 260-265. doi: 10.2489/jswc.72.3.260 Ruamps, L. S., Nunan, N., and Chenu, C. (2011) . Microbial biogeography at the soil pore scale. Soil Biol. Biochem. 43, 280-286. doi: 10.1016/j.soilbio.2010.10.010 SAS Inc. (2009). SAS user™s guide. V erssion 9.2 . Cary, NC: SAS Inst. 126 Schimel, J. P. and Schaeffer, S. M. (2012). Microbial control over carbon cycling in soil. Front. Microbial 3:348. doi: 10.3389/fmicb.2012.00348 Schurgers, G., Dörsch, P., Bakken, L., Leffelarr, P., and Haugen, L. E. (2006) . Modelling soil anaerobiosis from water retention characteristics and soil respiration. Soil Biol. Biochem. 38, 2637-2644. doi: 10.1016/j.soilbio.2006.04.016 Senthilkumar, S., Basso, B., Kravchenko, A. N., and Robertson, G. P. (2009). Contemporary evidenc e of soil carbon loss in the U.S. corn belt. Soil Sci. Soc. Am. J. 73, 2078-2086. doi: 10.2136/sssaj2009.0044 Sinsabaugh, R. I., Reynolds, H., and Long, T. M. (2000). Rapid assay for amidohydrolase (urease) activity in environmental samples. Soil Biol. Bio chem. 32, 2095-2097. d oi: 10.1016/S0038 -0717(00)00102 -4 Six, J., Elliot, E. T., and Paustian, K. (1999) . Aggregate and soil organic matter dynamics under conventional and no -tillage systems. Soil Sci. Soc. Am. J. 63, 1350-1358. doi: 10.2136/sssaj1999.63513 50x Six, J., Elliott, E. T., and Paustian, K. (2000). Soil macroaggregate turnover and microaggregate formation: a mechanism for C sequestration under no -tillage agriculture. Soil Biol. Biochem. 32, 2099-2103. doi: 10.1016/S0038 -0717(00)00179 -6 Six, J. and Paustian, K. (2014). Aggregate -associated soil organic matter as an ecosystem property and measurement tool. Soil Biol. Biochem 68, A4 -A9. doi: 10.1016/j.soilbio.2013.06.014 Smet, S., Beckers, E., Plougonven, E., Léonard, A., and Degré, A. (2018) . Can the pore scale geometry explain soil sample scale hydrodynamic properties? Front. Environ. Sci. 6:20. doi: 10.3389/fenvs.2018.00020 Sprunger C. D. and Robertson, G. P. (2018) . Early accumulation of active fraction soil carbon in newly established cell ulosic biofuel systems. Geoderma 318, 42-51. doi: 10.1016/j.geoderma.2017.11.040 Staver, K. W. and Brinsdield, R. B. (1990). Patterns of soil nitrate availability in corn production systems: Implications for reducing ground water contamination. J. Soil Wa ter Conserv. 53, 230-240. Strock, J. S., Porter, P. M., and Russelle, M. P. (2004) . Cover cropping to reduce nitrate loss through subsurface drainage in the northern U.S. Corn Belt. J. Environ Qual 33, 1010-1016. doi: 10.2134/jeq2004.1010 Strong, D. T., De Wever, H., Merckx, R., and Recous, S. (2004). Spatial location of carbon decomposition in the soil pore system. Eur. J. Soil Sci. 55, 739-750. doi: 10.1111/j.1365 -2389.2004.00639.x 127 Swift, R. S. (2001). Sequestration of carbon by soil. Soil Sci. 166, 858-871. doi: 10.1097/00010694 -200111000-00010 Syswerda, S. P., Corbin, A. T., Mokma, D. L., Kravchenko, A. N., and Robertson, G. P. (2011). Agricultural management and soil carbon storage in surface vs. deep layers. Soil Sci. Soc. Am. J. 75, 92-101. doi: 10.21 36/sssaj2009.0414 Toosi, E. R., Kravchenko, A. N., Guber, A. K., and Rivers, M. L. (2017). Pore characteristics regulate priming and fate of carbon from plant residue. Soil Biol. Biochem. 113, 219-230. doi: 10.1016/j.soilbio.2017.06.014 Wang, W., Kravchenk o, A. N., Johnson, T., Srinivasan, S., Ananyeva, K. A., Smucker, A. J. M., et al. (2013). Intra -aggregate pore structures and Escherichia coli distribution by water flow within and movement out of soil macroaggregates. Vadose Zone J. 12:4. doi: 10.2136/vzj 2013.01.0012 Wang, W., Kravchenko, A. N., Smucker, A. J. M., Liang, W., and Rivers, M. L. (2012). Intra -aggregate pore characteristics: X -ray computed microtomography analysis. Soil Sci. Soc. Am. J. 76, 1159-1171. doi: 10.2136/sssaj2011.0281 Wang, W., Krav chenko, A. N., Smucker, A. J. M., and Rivers, M. L. (2011). Comparison of image segmentation methods in simulated 2D and 3D microtomographic images of soil aggregates. Geoderma . 162, 231 -241. doi: 10.1016/j.geoderma.2011.01.006 Wickings, K. and Grandy, A. S. (2013) . Management intensity interacts with litter chemistry and climate to drive temporal patterns in arthropod communities during decomposition. Pedobiologia 56, 105-112. doi: 10.1016/j.pedobi.2013.01.001 Wieder, W. R., Grandy, A. S., Kallenback, C. M ., and Bonan, G. B. (2014). Integrating microbial physiology and physiochemical principles in soils with MIcorbial -MIneral carbon stabilization (MIMICS) model. Biogeosci. Discuss. 11, 1147-1185. doi:10.5194/bgd -11-1147-2014 Wiersum, L. K. (1957). The relat ionship of the size and structural rigidity of pores to their penetration by roots. Plant Soil 9, 75-85. doi: 10.1007/BF01343483 Williams, M. A., Myrold, D. D., and Bottomley, P. J. (2006). Distribution and fate of 13C -labeled root and straw residues from ryegrass and crimson clover in soil under western Oregon field conditions. Biol. Fertil. Soils 42, 523-531. doi: 10.1007/s00374 -005-0046-5 Xue, K., Wu, L., Deng, Y., He, Z., Nostrand, J. V., Robertson, G. P., et al. (2013) . Functional gene differences in so il microbial communities from conventional, low -input, and organic farmlands. Appl. Environ. Microbiol. 79, 1284-1292. doi: 10.1128/AEM.03393 -12 Young, I. M., Crawford, J. W., and Rappoldt, C. (2001). New methods and models for characterizing structural he terogeneity of soil. Soil Tillage Res. 61, 33-45. doi: 10.1016/S0167 -1987(01)00188 -X 128 Young, I. M. and Ritz, K. (2000). Tillage, habitat space and function of soil microbes. Soil Till Res. 53, 201-213. doi: 10.1016/S0167 -1987(99)00106 -3 129 Chapter 5 : Conclus ion Improving soil carbon stocks is important to mitigating global climate change as well as increasing crop yields and improving sustainability in agriculture. However, the microscale dynamics of soil carbon preservation, the scale at which soil carbon preservation takes place, are poorly understood. This has resulted in general guidelines for the improvement of soil carbon that fail to perform as predicted in several cases. This study sought to elucidate some of these microscale mechanisms in relation t o specific soil pore sizes and pore origins for a better understanding of microscale soil carbon dynamics. Spatial variability of the soil matrix of conservation managements, specifically biologically based and early successional managements, was greater t han in conventional tillage. It is believed this variability was the result of variability in soil carbon distributions within the matrix. Root pores were found to increase carbon concentrations for 123 m, while non -root pores were found to decrease carbo n concentrations for A mixture of root and non -root pores would result in areas of high and low carbon, possibly creating the variability in biologically based and early successional managements. Conventional management, which has less root pores, w ould be more uniform in soil carbon distributions, resulting in less spatial variability. Increasing the spatial variability of soil carbon would result in more diverse microenvironments, a potential driver of soil carbon dynamics. Pores of 15 - ere associated with increased amounts of carbon. The association found in this study was nearly identical to previously reported results, despite being from two different agricultural managements collected at different times of the year from two different fields. This indicates a universal mechanism for carbon preservation in these soils in 130 association with 15 - necessary to confirm this. Pores of 40 - s where carbon was easily gained, but also easily lost. Both natural abundance and enriched isotopes showed that new carbon was preferentially associated with pores of these sizes prior to incubation, but this association was lost during subsequent incubat ions. This is thought to be connected to root growth in these soils, as 40 - contain very fine plant roots. The findings of this study improve the understanding of microscale carbon dynamics within soil. Increased amounts of root pores lead to higher spatial variability of soil carbon in soil aggregates due to a larger influence on the surrounding soil matrix. Soil carbon preservation is believed to occur through organo -mineral interactions in this matrix. Larger amounts of 15 -40 m pores we re associated with soil carbon gains from a possibly universal mechanism, however, further research would be needed to determine the source of this mechanism. Higher amounts of 40-90 m pores were associated with fast gains and losses in soil. This might b e related to root growth dynamics, but further research is necessary to confirm this. These studies, when combined, show that pores behave similarly regardless of management, indicating that if a pore distribution is known, carbon dynamics may be predictab le and may help bridge the gap between currently used aggregate distributions a nd a more pore centric approach; however, determining if pores behave the same in different soil types would be necessary. While this study improves knowledge of micro -scale car bon dynamics, further research is necessary for full comprehension. 131 APPENDIX 132 Table A. 1: Slopes calculated by ANCOVA for the relationship between the amount of pores of the specified size 13C of the soil. Pore Size Structure Root Pre Post 4-15 m Intact Root 75.5084 26.9310 Intact Control 8.9397 11.3010 Destroyed Root 16.7750 6.8059 Destroyed Control 7.6211 3.5835 15-40 m Intact Root 447.68 231.00 Intact Control -113.90 -80.5010 Destroyed Root 312.14 31.8726 Destroyed Control 131.04 22.7433 40-90 m Intact Root 1306.00 -590.72 Intact Control -330.48 -153.21 Destroyed Root 1573.15 -112.06 Destroyed Control 98.0046 79.6818 90+ m Intact Root 15853 -5965.45 Intact Control -1432.11 -192.13 Destroyed Root 2715.94 -3518.71 Destroyed Control -737.93 2359.21