IMPACTS OF CHANGING PRECIPITATION ON NITROGEN CYCLING IN DIFFERENT LANDSCAPE POSITIONS AND CROPPING SYSTEMS By Kathryn Glanville A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Crop and Soil Sciences - Doctor of Philosophy 2020 ABSTRACT IMPACTS OF CHANGING PRECIPITATION ON NITROGEN CYCLING IN DIFFERENT LANDSCAPE POSITONS AND CROPPING SYSTEMS By Kathryn Glanville Soil nitrogen (N) influences crop yields and can interact with climate change. Soil N has many transformations and transfers that are hard to quantify and control. These N transformations and transfers are mediated by many factors, including temperature, water, and carbon. Thus, impending climate change may strongly affect N cycling across cropping systems. To minimize N losses and increase crop production, we must maximize N use efficiency (NUE). Past research shows precipitation and soil moisture act as the primary physical drivers of terrestrial N cycling and losses. To improve NUE with changing precipitation patterns, controls on N cycling in terrestrial systems must be identified. Thus, experiments to elucidate the linkage between hydrological and biogeochemical controls are valuable (Chapter 1). Many aspects of the N cycle are influenced by a changing climate - two are especially important: nitrous oxide fluxes (N2O) and biological nitrogen fixation (BNF). N2O is a powerful greenhouse gas with over 250 times the radiative forcing of CO2. In Chapter 2, I test the hypothesis that changing rainfall patterns strongly alter N2O fluxes in agricultural soils as modulated by cropping system. I use rainfall manipulation shelters to expose soils to the same amount of rainfall delivered at different intervals (3-days, 14-days, and 28- days). Results from the 2016 and 2017 field seasons show cumulative N2O fluxes were 1.4 to 2 times higher when rainfall occurred in 28-day rather than shorter intervals in corn systems. Fluxes were related to changes in denitrifier enzyme activity for both years. In switchgrass systems N2O emissions were not significantly affected by rainfall intervals. ii In Chapter 3, I test the hypothesis that changing rainfall patterns that alter N2O fluxes will be modulated by landscape position as landscape position affects soil texture and carbon. Over two field seasons cumulative N2O fluxes were higher in toeslope positions than in summit positions, and longer rainfall intervals had higher fluxes in summits only, consistent with higher soil carbon and finer soil texture in toeslope positions. Knowledge of these landscape patterns deserve inclusion in models of current and future climate change effects in order to better quantify and mitigate agricultural N2O fluxes. In Chapter 4, I test the hypothesis that BNF is particularly vulnerable to changing rainfall patterns in till vs. no-till and in summit vs. toeslope positions due to differences in texture and organic matter. Results reinforce the importance of topographic position for predicting soybean BNF and show that summit positions are more sensitive to additional rainfall. Results also show changes in rainfall intensity affect BNF in tilled differently than in no-till soils. Models that incorporate these interactions will be better able to characterize legume crop performance and N fixation across landscapes and improve global estimates for BNF. Understanding these interactions in the agricultural US Midwest may help us improve sustainability of N use in cropping systems with a changing climate. iii TABLE OF CONTENTS LIST OF TABLES ........................................................................................................................ vi LIST OF FIGURES ..................................................................................................................... vii Chapter 1: Introduction ................................................................................................................. 1 Background ............................................................................................................................... 1 Chapter Overview ..................................................................................................................... 3 Chapter 2: Consequences of Changing Rainfall Patterns on Nitrous Oxide (N2O) Fluxes in an Annual vs. Perennial Cropping System ........................................................................................ 5 Abstract ...................................................................................................................................... 5 Introduction ............................................................................................................................... 6 Methods ...................................................................................................................................... 9 Experimental Site .................................................................................................................... 9 Rainfall Manipulation Shelters ............................................................................................. 10 Rainfall Intervals ................................................................................................................... 11 Gas sampling ......................................................................................................................... 12 Soil Sampling ........................................................................................................................ 12 Statistical Analysis ................................................................................................................ 14 Results ...................................................................................................................................... 14 N2O emissions ....................................................................................................................... 14 DEA Assays .......................................................................................................................... 16 Soil Moisture and Temperature ............................................................................................ 17 Discussion ................................................................................................................................ 18 N2O dynamics by cropping system ....................................................................................... 19 N2O emissions as affected by rainfall pattern ....................................................................... 19 Seasonal patterns of N2O in response to changing rainfall regimes ..................................... 20 Denitrification rates and NOS enzyme persistence .............................................................. 20 Implications for predicting future effects of changing rainfall patterns on N2O emissions . 22 Chapter 3: Consequences of Changing Rainfall Patterns on Nitrous Oxide Fluxes at Different Landscape Positions ..................................................................................................................... 23 Abstract .................................................................................................................................... 23 Introduction ............................................................................................................................. 24 Methods .................................................................................................................................... 28 Experimental Site .................................................................................................................. 28 Rainfall Manipulation Shelters ............................................................................................. 29 Rainfall Intervals ................................................................................................................... 30 Gas sampling ......................................................................................................................... 30 Soil Sampling ........................................................................................................................ 31 Statistical Analysis ................................................................................................................ 32 iv Results ...................................................................................................................................... 33 N2O emissions ....................................................................................................................... 33 Soil Properties: Soil texture, C:N, moisture, and temperature .............................................. 34 Discussion ................................................................................................................................ 36 N2O dynamics by landscape position (Question 1) .............................................................. 36 N2O emissions as affected by rainfall pattern (Questions 2 and 3) ...................................... 37 Factors associated with N2O emissions ................................................................................ 37 Implications for predicting future effects of changing rainfall ............................................. 38 Chapter 4: Impacts of Changing Precipitation Patterns on Soybean (Glycine Max L.) Biological Nitrogen Fixation as Mediated by Landscape Position and Tillage ........................ 39 Abstract .................................................................................................................................... 39 Introduction ............................................................................................................................. 41 Experimental Design and Site Description ........................................................................... 44 Toposequence Experiment (Question 1) ............................................................................... 46 Rainfall Amounts by Landscape Position Experiment (Question 2) .................................... 47 Rainfall Intervals by Tillage Experiment (Question 3) ........................................................ 48 Methods .................................................................................................................................... 49 Biological Nitrogen Fixation ................................................................................................ 49 Soil Analyses ........................................................................................................................ 49 Plant Analysis ....................................................................................................................... 51 BNF calculations ................................................................................................................... 52 Statistical Analysis ................................................................................................................ 52 Results ...................................................................................................................................... 54 Toposequence Experiment (Question 1) ............................................................................... 54 Rainfall Amounts by Landscape Position Experiment (Question 2) .................................... 56 Rainfall Intervals by Tillage Experiment (Question 3) ........................................................ 56 Soil Properties: Soil Texture, OM, Mo, Fe, K, P, and pH .................................................... 58 Discussion ................................................................................................................................ 59 BNF by Toposequence (Question 1) .................................................................................... 59 BNF with Changes in Rainfall Amounts by Landscape Position (Question 2) .................... 62 BNF with Changes in Rainfall Intervals under Different Tillage (Question 3) .................... 63 BNF validity ......................................................................................................................... 64 Remaining Questions ............................................................................................................ 65 Implications .......................................................................................................................... 67 REFERENCES ............................................................................................................................ 69 v LIST OF TABLES Table 3.1: Field identification, summit elevation, toeslope elevation, length of slope, and slope degree of toposequences. ...................................................................................................... 29 Table 3.2: Mean values for five fields with standard error for total C, total N, C:N, and inorganic N. ........................................................................................................................................... 35 Table 4.1: Field identification, summit elevation, toeslope elevation, length of slope, and slope degree of fields used. ............................................................................................................ 47 Table 4.2: Planting date, harvest date, ambient rainfall, and additional irrigation for 2016 and 2017....................................................................................................................................... 48 Table 4.3: Planting date, harvest date, ambient rainfall, and additional irrigation for 2015 and 2018....................................................................................................................................... 49 Table 4.4: Experiment of each year with associated measurements. ........................................... 50 vi LIST OF FIGURES Figure 1.1: An illustration showing differences in precipitation patterns (historical and intensified), in landscape positions (toeslope and summit), and in cropping systems (corn, switchgrass, and soybeans). This is intended to show some factors explored in the dissertation. ............................................................................................................................. 3 Figure 2.1: Simulated rainfall in experimental treatments by day of experiment. Bar heights are relative to the amounts of water applied (shown in parentheses). All treatments received the same total amount of water (80 mm per four weeks). .......................................................... 11 Figure 2.2: Daily N2O response to different rainfall patterns in the no-till corn and switchgrass systems in 2016 (left) and 2017 (right). Error bars represent standard errors of the mean, n=4 replicate plots (each with three rainfall intervals). Red arrows denote 28-day watering events. Black arrow denotes fertilizer application. ............................................................... 15 Figure 2.3: Cumulative N2O responses to different rainfall patterns in the no-till corn and switchgrass systems in 2016 and 2017. Error bars represent standard errors of the mean, n=4 replicate plots (each with three rainfall intervals). ................................................................ 16 Figure 2.4: Composite NOS enzyme activity for before and after rewetting in June 2017. Error bars represent standard errors of the mean, n=4 replicate blocks. ........................................ 17 Figure 2.5: Soil moisture in corn and switchgrass for 2017. ....................................................... 18 Figure 2.6: Soil temperature for corn and switchgrass for 2017. ................................................ 18 Figure 3.1: Sketch of experimental design a) from a profile view of two landscape positions in a field and b) outline of three shelters side by side. ................................................................. 29 Figure 3.2: Daily N2O response to different rainfall patterns in the summit and toeslope landscape positions in 2016 and 2017. Error bars represent standard errors of the mean, n=4 replicate plots (each with three rainfall intervals). Red arrows denote 28-day watering events. Months omitted lack temporal resolution. ................................................................ 33 Figure 3.3: Cumulative N2O responses to different rainfall patterns in the summit and toeslope positions in 2016. Error bars represent standard errors of the mean, n=4 replicate plots (each with three rainfall intervals). ................................................................................................. 34 Figure 3.4: Percent clay + silt by average N2O flux for the 3-day rainfall interval. Error bars are omitted for clarity. The dashed line represents a linear regression had significant effects (P<0.05). ............................................................................................................................... 35 Figure 3.5: Soil temperature fluctuations in summit and toeslope positions. .............................. 36 vii Figure 4.1: Kellogg Biological Station aerial view with fields used in 2015 in red; fields used in 2016 in green; fields used in 2017 outlined in yellow; and fields used in 2018 outlined in light blue. .............................................................................................................................. 45 Figure 4.2: A illustration showing positions of transect, landscape positions, and soybean plot locations for the toposequence experiment. .......................................................................... 46 Figure 4.3: A illustration showing positions the landscape positions and rainfall addition plot locations for the rainfall amounts by landscape position experiment. .................................. 47 Figure 4.4: A illustration showing how rainfall intervals were conducted, with and without a shelter, in tilled and no tillage soils for the rainfall intervals by tillage experiment. ............ 48 Figure 4.5: Percent BNF by landscape position (a), NPP by landscape position (b), and %NPP from BNF (c). Error bars represent standard errors of the mean, n=4 replicate plots. Asterisks indicate significant effects (P<0.05). .................................................................... 55 Figure 4.6: N mineralization potentials at toeslope, backslope, and summit positions. Error bars represent standard errors of the mean, n=4 replicate plots. Asterisks indicate significant effects (P<0.05) with soil texture. ......................................................................................... 55 Figure 4.7: BNF (%) increased at summit positions with more precipitation. Error bars represent standard errors of the mean, n=4 replicate plots. Asterisks indicate significant effects (P<0.05) with soil texture. .................................................................................................... 56 Figure 4.8: Percent BNF in no-till and conventional tillage with ambient and 2-week rainfall intervals. Error bars represent standard errors of the mean, n=4 replicate plots. ................. 57 Figure 4.9: Percent sand, silt and clay by %BNF. Error bars are omitted for clarity, n=4 replicate plots. Asterisks indicate significant effects (P<0.05) with soil texture. ............................... 58 Figure 4.10: Percent BNF by %OM. Error bars are omitted for clarity, n=4 replicate plots. P value = 0.0116 ....................................................................................................................... 59 viii Background Chapter 1: Introduction Nitrogen (N) is an important element for crop growth and known contributor to environmental pollution. Humans’ ability to transform inert N to reactive N (Nr) has been a double-edged development with positive and negative effects. Society has increased Nr inputs to the biosphere to ~5 times those of pre-industrial times, mainly to cropping systems (Houlton et al., 2013). These N inputs, through fertilizer, biological nitrogen fixation (BNF), and deposition, have allowed us to support an ever-increasing need for food, fuel, and fiber for a growing human population. However, the majority of N applied is not taken up by plants but lost to the environment, far exceeding planetary boundaries (Rockström et al., 2009) and even N taken up by plants makes its way back to the environment through animal and human waste. Nr from crop production is released as different N species either to the atmosphere or to ground and surface waters via leaching and runoff. Negative impacts include climate change, eutrophication, biodiversity declines, and smog formation. These losses are influenced by complex soil N cycle transformations and transfers - which operate differentially across the environment at different temporal and spatial scales. Together, water and carbon (C) greatly influence soil N cycling, and in the US Midwest water dynamics are changing with climate change. Climate change in the US Midwest is causing more extreme precipitation events with longer dry periods in between (Pryor et al., 2014) and more overall annual precipitation, at least in Michigan (Frankson and Kunkel, 2017). The Midwest is a very important agricultural region and projected to be increasingly impacted by climate change. In this dissertation, I aim to examine how changes in rainfall patterns influence N loss and N acquisition by crops in different cropping systems and in different topographical positions. 1 There are several aspects of the N cycle likely to be particularly affected by changing rainfall; two of the most interesting and important are soil nitrous oxide (N2O) emissions and biological nitrogen fixation (BNF). Approximately two-thirds of Nr created by contemporary human activity is lost to air as opposed to water. This includes nitrogen oxides (NOs), ammonia, dinitrogen, and N2O. N2O is a potent greenhouse gas that contributes greatly to climate change through radiative forcing. A majority of anthropogenic N2O emissions are from agricultural soils (IPCC 2014). Additionally, changing rainfall patterns are known for their capacity to affect N2O emissions (Bergsma et al., 2002; Gelfand et al., 2015; Kim et al., 2012). A second important N-cycle process likely to be affected by climate change due to soil N transformations is BNF, such as in soybean (Glycine max L.) and other legumes. BNF is the capture of atmospheric N2 and its transformation to ammonia by symbiotic N2-fixing rhizobia. N acquired by crops is removed in harvest, decomposed, or stored in the cropping system in soil organic matter. Legumes have the potential to build soil fertility and provide N to subsequent crops. BNF is affected by many soil factors including texture, pH, and inorganic N pools (Bottomley and Myrold, 2015; Salvagiotti et al., 2008). Leguminous N-fixing crops are important worldwide as BNF also occurs in other important crops, including peas (Vigna spp.), clovers (Trifolium spp.), vetches (Vica spp.), and perennial legumes such as alfalfa (Medicago sativa L.). BNF may be even more relevant in the future with the potential introduction of N2- fixing corn (Zea mays), wheat (Triticum aestivum L.), and rice (Oryza sativa) to reduce the use of fertilizer, which may influence Nr in the environment (Bottomley and Myrold, 2015). Understanding controls on BNF can help farmers make management decisions. 2 Chapter Overview In the remainder of this dissertation I describe three projects to explore how changing rainfall and topography will affect N2O emissions, BNF, and farmer vs researcher perspectives in Michigan cropping systems. Figure 1.1 illustrates the factors explored in this dissertation. Figure 1.1: An illustration showing differences in precipitation patterns (historical and intensified), in landscape positions (toeslope and summit), and in cropping systems (corn, switchgrass, and soybeans). This is intended to show some factors explored in the dissertation. Research on N2O emissions appears in two chapters. In Chapter 2, I test the hypothesis that changing rainfall patterns strongly alter N2O fluxes in agricultural soils as modulated by cropping system. I used rainfall manipulation shelters to expose soils to the same amount of rainfall delivered at different intervals (3-days, 14-days, and 28-days) and quantified mineral N, nitrous oxide reductase, and labile C. In Chapter 3, I test the hypothesis that changing rainfall patterns that alter N2O fluxes will be modulated by landscape position insofar as landscape position affects soil texture and carbon content. I again used rainfall manipulation shelters to expose soils to the same amounts of rainfall delivered at different intervals (3-days, 14-days, and 28-days) placed at both summit and toeslope positions within conventionally tilled corn fields. 3 In Chapter 4, I test the hypothesis that BNF is particularly sensitive to changing rainfall patterns in till vs. no-till and in summit vs. toeslope positions due to landscape differences in texture and organic matter. I measured soybean BNF by 15N natural abundance using nodulating and non-nodulating isolines to test effects of landscape position, and tillage management, and their potential interaction with rainfall. The results of both the N2O and BNF experiments elucidate how the N cycle will be influenced by changing rainfall patterns. The findings challenge the assumptions that longer dry periods with less consistent rainfall will lower N2O emissions and decrease BNF consistently. N2O results show this may not be the case due to accelerated emissions upon drought cessation, especially for annual systems with high fertilizer inputs. BNF results show that changes in rainfall will differ by landscape position and tillage management. My conclusions point to potential feedbacks and challenges of changing rainfall patterns for farmers and society. The controls identified will be important in designing models, adapting practices, and creating mitigation plans for the intensifying precipitation patterns predicted. 4 Chapter 2: Consequences of Changing Rainfall Patterns on Nitrous Oxide (N2O) Fluxes in an Annual vs. Perennial Cropping System Abstract Nitrous oxide (N2O) is the dominant natural ozone-consuming substance in the stratosphere and a strong greenhouse gas with 265-300 times the radiative forcing of CO2. N2O is produced by soil microbes and fluxes are closely linked to soil moisture. Consequently, documented and future changing rainfall patterns due to climate change will likely influence N2O fluxes. Since most anthropogenic N2O produced globally is from agricultural soils, where fluxes are controlled by numerous factors including oxygen, nitrate, and carbon availability (all of which are strongly tied to soil moisture status), it seems important to understand the impact of more extreme precipitation patterns on N2O emissions. I tested the hypothesis that changing rainfall patterns strongly alter N2O fluxes in agricultural soils as modulated by cropping system. I used rainfall manipulation shelters to expose soils to the same amount of rainfall delivered at different intervals (3-days, 14-days, and 28-days). The experiment was conducted for 10-weeks during two growing seasons in replicated no-till continuous-corn and switchgrass systems at a site in the upper US Midwest. Results show cumulative N2O fluxes were 1.4 to 2.0 times higher when rainfall occurred in 28-day rather than shorter intervals in corn systems. Response patterns were related to drought-induced changes in denitrifier enzyme activity for both years. In switchgrass systems N2O emissions were not significantly affected by different rainfall intervals. Understanding patterns and mechanisms for N2O fluxes from managed soils is important for achieving more sustainable agriculture, developing mitigation practices, and parameterizing global biogeochemical models. 5 Introduction Climate change is causing shifts in U.S. temperatures and precipitation, resulting in increasingly more intense and frequent precipitation events, longer frost-free seasons, and higher night-time temperatures across much of the Midwest U.S. (Pryor et al., 2014). Associated effects on biodiversity, human health, and ecosystem productivity are projected to intensify under all projected future emissions scenarios (IPCC, 2007). Rainfall intensification in the US Midwest, an area especially important to agriculture, can be expressed as more variable growing season rainfall with longer dry periods between rain events. Rainfall intensification is particularly worrisome for farming, which depends on predictable rainfall for high yields and efficient nitrogen (N) use. Rainfall dramatically affects soil biogeochemical processes, especially for carbon (C) and N flows (Fierer and Schimel, 2002), and rainfall intensification, often quantified as the heaviest 1% of events, is an important attribute of precipitation regimes, especially for agriculture. Future N management decisions will require a thorough understanding of N responses to altered rainfall (Bowles et al., 2018; Robertson et al., 2013), which may lead to even greater losses of N to groundwater, waterways, and the atmosphere. Nitrogen gas losses from agricultural systems include the nonreactive gas dinitrogen (N2) as well as nitric oxides (NOx) and nitrous oxide (N2O). N2O is a main cause of stratospheric ozone depletion and is a powerful greenhouse gas (Ravishankara et al., 2009), with 265 to 300 times the radiative forcing of CO2 (IPCC, 2014). Agriculture is a major global source of N2O and is responsible for most of the rise in atmospheric N2O concentrations since the 1950s (Fowler et al., 2009). About 60% of agricultural N2O emissions are emitted from soil, where N2O is produced by microbes primarily during denitrification and nitrification (Zhu-Barker et al., 2015). 6 Denitrification appears to be the dominant source of N2O in most mesic cropping systems (Aronson and Allison, 2012). Heterotrophic denitrifiers, mainly bacteria but also archaea and fungi, use soil organic C as an energy source and nitrate, which is thermodynamically similar to oxygen, as a terminal electron acceptor when oxygen is unavailable. Denitrification is often enhanced after precipitation (Bergsma et al., 2002; Gelfand et al., 2015) and especially in soil aggregates, rhizospheres, and decomposing litter, where soil microbes have ample access to C but limited access to oxygen (Khalil et al., 2004; Kravchenko et al., 2017; Loecke et al., 2008; Sexstone et al., 1988). The denitrification pathway products nitrite, nitric oxide, N2O, and dinitrogen are formed sequentially (Cooper and Smith, 1963), such that N2O is released to the soil atmosphere when denitrification is not carried out completely. Although considered a minor source of N2O in comparison to denitrification, nitrifying archaea and bacteria also produce N2O from hydroxylamine decomposition following ammonia oxidation (Bremner and Blackmer, 1978). Since nitrifiers require oxygen (O2) to oxidize nitrite to NO3-, nitrification proceeds fastest in aerobic conditions, but can occur in soils with up to 60% water filled pore space (WFPS) (Abbasi and Adams, 2000). Nitrifier denitrification and coupled nitrification-denitrification can also produce N2O (Wrage et al., 2001). Most N2O produced by nitrifiers appears to be from nitrifier denitrification stimulated by wet-dry cycles (Wrage-Mönnig et al., 2018). Thus, both nitrification and denitrification are strongly affected by soil moisture, so likely to be influenced by changing rainfall patterns (G P Robertson and Groffman, 2015). The effects of longer dry periods during the growing season on C and N will likely influence nitrous oxide (N2O) emissions in inconsistent ways (Kim et al., 2012). Changing rainfall patterns are an aspect of climate change that may shift controls on biogeochemical mechanisms, leading to spatial 7 variability and episodic moments that have higher than average influence on N fluxes (Bernhardt et al., 2017). Plants influence N2O fluxes by competing with microbes for available N and this likely differs in annual and perennial cropping systems. Plants in annual cropping systems are not active for the entire growing season – as compared to perennial systems – and as well cannot re- translocate N to roots for use the following year, so might leave more soil N for microbes to transform into N2O (Oates et al., 2016). Perennial systems can shift the microbial community composition by improving niches for functional groups (Liang et al., 2012) and change the physiological ability of the soil community towards enzyme production (Hargreaves and Hofmockel, 2014). How might changing rainfall intensification, with its effects on C and N biogeochemistry, affect N2O fluxes? Some research shows higher N2O emissions following rewetting of dry soil in the field (Gelfand et al., 2015). Numerous lab experiments have also reported N2O emissions stimulated by wetting dry as compared to wetting wet soils (e.g. Bergsma et al., 2002; Christensen and Prieme, 2001). Groffman and Tiedje (1988) suggest that denitrification is stimulated by reduced oxygen following rewetting due to accelerated respiration. Others hypothesize that pulses of N2O are likely caused by mineralization of both C and N due to the hysteretic effects of microbial biomass' increasing substrate availability and denitrifier activity (Congreves et al., 2018). Here I examine the degree to which changing rainfall patterns might accelerate N2O fluxes in different cropping systems. I test the hypothesis that longer dry periods between rainfall events will strongly increase N2O fluxes in agricultural soils and that these increases will be especially pronounced in annual (versus perennial) cropping systems where more N is more 8 available to microbes. Understanding these patterns and the mechanisms that underlie them may be crucial for predicting reactive N responses to climate change and for designing mitigation strategies to forestall potential increases in N2O production. Methods Experimental Site I conducted this study in the Great Lakes Bioenergy Research Center’s Bioenergy Cropping System Experiment at the Kellogg Biological Station (KBS) Long-Term Ecological Research site, Hickory Corners, Michigan, USA (42° 24'N, 85° 23'W 288 m elevation). Average temperatures at KBS range from -3.8°C (January) to 22.9°C (July) with an annual mean of 10.1°C. Yearly precipitation is 1005 mm and the 30-year average in monthly precipitation ranges from 40 mm in February to 100 mm in May (Peters et al., 2013). About 17% of precipitation occurs during the winter months from January to April, with the rest evenly distributed throughout the growing season. Over the past 30 years, precipitation events from April to December were, on average, 2.5 mm and ~3.6 days apart; around eight events per season were greater than 25 mm (https://lter.kbs.msu.edu/datatables/7). Potential evapotranspiration exceeds rainfall for four months out of the year (Crum et al., 1990; Hamilton, 2015). Soils at KBS are well-drained mesic Typic Hapludalfs developed from glacial till and outwash consisting of co-mingled Kalamazoo (fine-loamy, mixed, semiactive) and Oshtemo (coarse-loamy, mixed, active) series (Crum and Collins, 1995; G Philip Robertson and Groffman, 2015) with intermixed loess (Luehmann et al., 2016). The predominant land use in the region consists of cultivated and successional fields, woodlots, residences, lakes, and wetlands. Field experiments were conducted in the Great Lakes Bioenergy Research Center’s Biofuel Cropping System Experiment experimental design as described in Sanford et al. (2016). 9 I used two of ten cropping systems established in 2008 in a randomized complete block design with five replicate blocks. I used four of the replicate blocks. Treatment plots were 27 × 43 m (0.12 ha) with 12-m alleyways between adjacent plots and planted to either continuous no-till corn (Zea mays) or switchgrass (Panicum virgatum, variety “Cave-in-Rock”) to represent annual and perennial crops, respectively. Both crops were managed as per common practice in the region. Corn was planted in 76 cm rows and fertilized at 150 kg N ha-1 in two applications: 25 kg N at planting and the remainder side-dressed ~6 weeks later when plants were at the V6 stage. Switchgrass was fertilized at 56 kg N ha-1 in late spring. About 50% of aboveground corn residue was harvested in the fall along with grain (Gelfand et al., 2020); switchgrass was harvested each fall following first frost. Detailed agronomic protocols are available online (https://lter.kbs.msu.edu/protocols/122). Rainfall Manipulation Shelters Small rainfall manipulation shelters were used to create rainfall environments in which soils were exposed to different rainfall intervals but the same amounts of total rainfall, which was added to each shelter at different times. Well-vented shelters were 61 × 122 × 40 cm high and placed between corn rows and switchgrass crowns; each shelter contained a cylindrical stainless-steel gas sampling frame (28.5 cm diameter × 30 cm high) and sufficient soil surface area to allow periodic soil sampling. Shelters were constructed of polycarbonate (roofs), plexiglas (sides), and sheet metal (bottom edging). Bottom edging on the sides allowed 3 cm insertion into the soil to eliminate the potential for storm-derived lateral surface flow, which was in any case low due to high infiltration rates in these well-drained soils. Three large holes 6.4 cm diameter drilled in the sides of the shelters allowed free airflow and eliminated humidity buildup. Before implementing the experiment, N2O flux measurements taken inside and outside small 10 shelters placed under a larger (10 x 5m) shelter. The small shelters showed no significant shelter effects on fluxes during either low or high flux events (Glanville in prep). Roofs were hinged at the back to allow for periodic soil and gas sampling. Each shelter covered a 0.6 m2 area. Rainfall Intervals Rainfall additions were based on a 30-year regional average calculated from 1984 to 2014. To each of three shelters per plot, rainfall was applied at either 3-day (long-term average for the region), 14-day, or 28-day intervals (Figure 2.1). The 3-day average was 7.1 mm with a 1.3 mm standard deviation not accounting for interception by the canopy. Interception of the canopy was accounted for, lowering the average of 7.1 mm to 6.6 mm rainfall. All treatments received 80 mm of precipitation per month. Since the soil under each shelter was protected from ambient rainfall, the amount of water added was independent of droughts or unusually wet weather; thus, the experiment was repeatable across field seasons because water was not limiting. Figure 2.1: Simulated rainfall in experimental treatments by day of experiment. Bar heights are relative to the amounts of water applied (shown in parentheses). All treatments received the same total amount of water (80 mm per four weeks). 11 Gas sampling N2O, CH4, and CO2 fluxes were measured three times per week on average to capture responses to rainfall events, using a static chamber method in which airtight lids were clamped to the chamber top and headspace N2O and CH4 concentrations measured periodically (Holland et. al 1999) using one of two techniques. First, for the "GC method" I took four chamber headspace gas samples (20 mL) every 15 minutes for 60 minutes. The samples were stored over-pressurized in glass vials and analyzed for N2O, CO2, and CH4 concentrations in the lab using a gas chromatograph (7890A Agilent Technologies Inc., DE, USA) equipped with an ECD for N2O, an FID for CH4, and an infrared gas absorption analyzer (LI-820, LI-COR, NE, USA) for CO2 (http://lter.kbs.msu.edu/protocols/113). The second technique was the "MICOS (Mobile Integrated Cavity Output Spectroscopy) method" that used a quantum cascade laser (QCL) analyzer (Los Gatos Research Inc., CA, USA) in the field for continuous chamber headspace measurements of N2O and CH4 concentrations and a closed-path infrared gas analyzer (LI-7000, LI-COR, NE, USA) for CO2 concentrations. During the measurement period, the headspace air inside each chamber was circulated through the QCL analyzer for 5 mins and gas concentrations were analyzed at 1s intervals. Soil Sampling Soil moisture and temperature were measured in every shelter every 10 minutes with sensors (5TM, Decagon, WA, USA) installed vertically at the surface and stored on data loggers (EM50, Decagon, WA, USA). Soil was sampled around co-occurring rewetting events to 20 cm depth for several assays: denitrification enzyme activity (DEA), gravimetric soil moisture, inorganic N, and labile and dissolved organic C. Soil texture to 20 cm depth was also measured once from each shelter area in both 2016 and 2017. 12 The DEA assay is a short-term slurry incubation wherein C as glucose and NO3- was provided in excess (Groffman et al. 1999). For each assay 10 g of fresh soil were added to 150 mL airtight jars. I evacuated the headspace of each jar then added either N2 or N2 plus 10% acetylene. Acetylene inhibits the last step of denitrification wherein N2O is converted to dinitrogen. While jars were gently shaken on shaker tables, headspace N2O concentrations were measured as described earlier over a 90 minute period and production rates were compared with and without the acetylene block to provide a measure of nitrous oxide reductase (NOS) activity (https://lter.kbs.msu.edu/protocols/193). For gravimetric soil moisture 10 g of fresh soil were oven-dried at 60°C for 48 hours (https://lter.kbs.msu.edu/protocols/24) Inorganic N was measured via a 1 M KCl 100 mL extraction of 10g of fresh soil performed in duplicate. Soils were shaken on a shaker table overnight and allowed to settle before filtering. Extracts were frozen until analysis. A flow injector analyzer (QuickChem 8500 Series 2, Lachat Instruments, CO, USA) was used to determine nitrate and ammonium concentrations in extracts (https://lter.kbs.msu.edu/protocols/33). Labile C was measured with a carbon mineralization assay wherein 10 g of dry soil were rewetted to 55% WFPS in a 150 mL jar (Franzluebbers et al., 2000): CO2 concentrations were measured every hour for four hours on an infrared gas analyzer (LI-7000, LI-COR, NE, USA). Dissolved organic C was measured via a water extraction with a 1:3 soil water ratio: The soil water combination was shaken overnight then centrifuged and the supernatant then filtered through a 0.2 µm membrane. The extracts were then analyzed for C on a total organic carbon analyzer (TOC-VWS and ASI-V autosampler, Shimadzu; Columbia, Maryland). Soil texture was measured in the shelter locations within each plot using the hydrometer method (https://lter.kbs.msu.edu/protocols/108). Here, 40 g of soil were shaken for 24 hours in 13 200 mL jars using hexametaphosphate as a dispersant. The slurry was then put in 1000 mL cylinders and supplemented with water. Hydrometer and temperature readings were taken over eight hours. Since sand falls out of solution too quickly to accurately record density changes, to supplement hydrometer readings sand was sieved out separately with a 53 µm sieve. Statistical Analysis All statistical analyses were performed using R software version 3.6.1 (R Development Core Team 2019) and with a significance value of P<0.05. The statistical model included two cropping systems, two years, and three rainfall-interval treatments with the interactions among them considered fixed factors. Blocks, crops nested within blocks, and rainfall-interval treatments nested within blocks and crops were considered random factors. Analysis of variance was used by considering cropping system as a whole plot factor and rainfall-interval treatments and years as subplot and sub-subplot factors. Normality of residuals was visually observed, and no violations of assumptions were found. Homogeneity of variance assumptions were checked and no heterogeneous variance was detected by Levene's test. DEA means were compared for significance using t-tests. Regression analysis was used to determine the relationship between N2O flux and temperature, moisture, and soil texture. Results N2O emissions Daily N2O emissions in corn and switchgrass for 2016 and 2017 are shown in Figure 2.2. In corn, fluxes ranged from 3.1 to 1000 μg N2O-N m-2 min-1 in 2016 and from 2.1 to 258 μg N2O-N m-2 min-1in 2017. In corn, higher fluxes occurred during the summer months of June and July before declining to below ~20 μg N2O-N m-2 min-1 for August through December of both years. 14 In switchgrass, N2O fluxes ranged from 1.7 to 183 in 2016 and 0.4 to 105 μg N2O-N m-2 min-1 in 2017, with fluxes highest in May, June, and July, and lower in August and September. Flux variability was greatest in April and May. Flux measurements were not taken past September in either year. Corn Switchgrass 2016 2017 ) 1 - n m i 2 - - m N O 2 N g µ ( x u F O 2 N l 1000 800 600 400 200 0 27 3 Day 14 Day 28 Day 6 1 June 11 16 Date 21 26 6 1 July ) 1 - n m i 2 - - m N O 2 N g µ ( x u F O 2 N l 200 150 100 50 3 Day 14 Day 28 Day 0 27 6 1 June 11 16 21 26 6 1 July ) 1 - n m i 2 - - m N O 2 N g µ ( x u F O 2 N l ) 1 - n m i 2 - - m N O 2 N g µ ( x u F O 2 N l 2500 2000 1500 1000 500 0 1000 750 500 250 0 3 day 14 day 28 day June 1 July 1 Sep 1 Oct 1 Aug 1 Date 3 day 14 day 28 day June 1 July 1 Aug 1 Date Sep 1 Oct 1 Date Figure 2.2: Daily N2O response to different rainfall patterns in the no-till corn and switchgrass systems in 2016 (left) and 2017 (right). Error bars represent standard errors of the mean, n=4 replicate plots (each with three rainfall intervals). Red arrows denote 28-day watering events. Black arrow denotes fertilizer application. Cumulative N2O emissions are sums of measured fluxes; values are not integrated. Cumulative N2O emissions in corn and switchgrass yielded similar trends for 2016 and 2017 (Figure 2.3), however, the difference between the 28-day interval and the other intervals in corn was smaller in 2017 than in 2016. For corn, cumulative N2O in 2016 was highest under the 28- day rainfall interval (2798 ± 633 μg N2O-N m-2 min-2); fluxes under the 3-day and 14-day 15 intervals were not significantly different from one another (1709 ± 240 and 1559 ± 55 μg N2O-N m-2 min-1, respectively). In 2016 and 2017 switchgrass N2O fluxes under the 3-day, 14-day, and 28-day intervals were not significantly different from one another (in 2016: 807 ± 166, 624 ± 57, and 807 ± 65 μg N2O-N m-2 min-1, respectively). Corn Switchgrass 2016 ) 2 - - m N O 2 N g µ ( x u l f O 2 N ) 2 - - m N O 2 N g µ ( x u l f O 2 N 4000 3000 2000 1000 0 1000 800 600 400 200 0 b 2017 a a 3 day 14 day 28 day Rainfall Interval 3 day 14 day 28 day Rainfall Interval ) 2 - - m N O 2 N g µ ( x u l f O 2 N 10000 8000 6000 4000 2000 0 3 day 14 day 28 day Rainfall Interval Figure 2.3: Cumulative N2O responses to different rainfall patterns in the no-till corn and switchgrass systems in 2016 and 2017. Error bars represent standard errors of the mean, n=4 replicate plots (each with three rainfall intervals). DEA Assays Total DEA rates were similar for both cropping systems and did not differ between 3-day and 28-day rainfall intervals for the June 2017 rewetting interval (Figure 2.4). Denitrification rates in corn were similar (30.9 ± 7.9 and 28.0 ± 2.3 μg N gas kg-1 soil h-1 for the different 16 intervals, respectively). However, the proportion of N2O converted to N2 differed between the 3 day and 28-day intervals in corn with NOS converting N2O into N2 at 76% and 8% efficiencies, respectively. There was no difference in the proportion of N2O converted to N2 in switchgrass. Corn Switchgrass ) 1 - h l i o s 1 - g k N g µ ( n o i t c u d o r p s a g N 40 30 20 10 0 3 day 28 day Rainfall Interval N2 component N2O component Figure 2.4: Composite NOS enzyme activity for before and after rewetting in June 2017. Error bars represent standard errors of the mean, n=4 replicate blocks. Soil Moisture and Temperature Soil moisture in corn and switchgrass for 2017 are shown in figure 2.5. In corn, soil moisture VWC ranged from 0 to 0.332 cm3 cm-3 in 2017. In corn, higher water content was observed during the months of April and October, congruent when the crop is not using water. In switchgrass, water content was highest in the 3-day rainfall interval, but lower in both the 14-day interval and 28-day interval. In switchgrass, soil moisture VWC ranged from 0 to 0.333 cm3 cm- 3. Soil temperature shown in figure 2.6 was slightly cooler in switchgrass than corn (18.12 ± 0.09 and 17.49 ± 0.03 μg N gas kg-1 soil h-1 for the different intervals, respectively). 17 3 day 14 day 28 day Corn ) ³ - m c ³ m c ( C W V 0.4 0.3 0.2 0.1 4 r p A 1 y a M 1 n u J 1 l u J 1 g u A 1 p e S 1 t c O Date Switchgrass ) ³ - m c ³ m c ( C W V 0.4 0.3 0.2 0.1 4 r p A 1 y a M 1 n u J 1 l u J 1 g u A 1 p e S 1 t c O 3 day 14 day 28 day Figure 2.5: Soil moisture in corn and switchgrass for 2017. Corn Switchgrass ) ³ - m c ³ m c ( C W V 60 40 20 0 4 r p A 1 y a M 1 n u J 1 l u J 3 day 14 day 28 day 40 30 20 10 ) C ° ( e r u t a r e p m e T 0 4 r p A 1 y a M 1 n u J 1 g u A 1 p e S 1 t c O Figure 2.6: Soil temperature for corn and switchgrass for 2017. Date Date 1 g u A 1 p e S 1 t c O 3 day 14 day 28 day 1 l u J Date Discussion With longer dry periods between rainfall events one might predict lower N2O fluxes due to less consistent anoxic conditions. I found the converse: with longer dry periods during the portion of the year with normally high N2O fluxes, a 28-day drought increased rather than decreased cumulative N2O fluxes. Overall results, then, support the hypothesis that longer dry periods between rainfall events increase N2O fluxes, but only for the annual (continuous corn) cropping system. In this system cumulative fluxes after 28 d of rainfall exclusion during the peak growing season were 1.4 times greater than cumulative fluxes from the control (3 d) rainfall interval in 2016, and 2.0 times greater in 2017. In the perennial (switchgrass) system rainfall interval changes had no detectable effects on N2O fluxes in either year. Reponses were more muted in other parts of the growing season. Response differences can best be explained by 18 differences in NOS enzyme dynamics; my hypothesis that differences are due to soil C and N pools was not supported. N2O dynamics by cropping system Overall, the perennial system had lower N2O emissions than the annual system, though in 2017, the cumulative emissions were more similar. Across both cropping systems, fluxes were comparable to those measured for the unsheltered portions of plots in prior years (Oates et al., 2016). N2O trends differed each year even though the rainfall pattern was the same. In corn systems, fluxes peaked in mid-June following fertilization in both years. Variations in N2O production between the two years did not appear related to inorganic N or temperature patterns over the period; presumably some factors other than temperature, soil N, and rainfall interacted to affect fluxes. Though increases in N input accelerate N2O emissions (Shcherbak et al., 2014), microbial community differences between corn and switchgrass cropping systems may also impact the N2O produced, independent of N inputs. Plant species can affect N cycling by shaping the microbial community composition (Hawkes et al., 2005) including composition of the denitrifier community (Cavigelli and Robertson, 2001), and perennial systems can increase enzyme production and nitrogen mining (Hargreaves and Hofmockel, 2014). This experiment did not address the direct effects of microbial community composition, but it is known to be effected by drying/ wetting treatments (Peralta et al., 2013). N2O emissions as affected by rainfall pattern The 28-day rainfall interval in corn expressed the highest emissions and there were no cumulative differences for other rainfall treatments. In 2017 the magnitude of N2O fluxes for 19 switchgrass was higher than in 2016. There are responses to individual water additions at all rainfall intervals in both cropping systems. Seasonal patterns of N2O in response to changing rainfall regimes Effects of rainfall patterns were most pronounced in the summer, when soil N levels and soil temperatures were highest. In April and May, intensified rainfall patterns had little if any effect on N2O fluxes. In June and July, in contrast, the same rainfall patterns had very discernable effects. By early fall, however, effects were again small and in early winter effects were again not discernable. This is likely due to a different hierarchy of controlling factors by season in each cropping system. N2O production mostly comes from nitrification and denitrification, which are largely influenced by soil mineral N (Davidson and Swank 1986), so it makes sense that I see the highest N2O production in the summer months after fertilizer is applied and when temperature is highest. During the summer the lack of response, when seen, may be due to plant N uptake’s depressing soil N pools, as that is when plant N demand is the highest. In the spring and fall, when soil temperature is lower and soil moisture is greater (due to higher water reserves in the spring and less ET in the winter), microbial activity is lower, and the NOS enzyme may not be depleted as it was in summer months. Denitrification rates and NOS enzyme persistence NOS enzyme activity might explain different rainfall responses. Denitrification enzymes do not persist indefinitely and can disappear when not used for long periods of time such as during droughts when soil moisture is not amenable to denitrification and denitrifiers do not allocate resource to enzyme maintenance (G P Robertson and Groffman, 2015). The induction of NOS, which transforms N2O to N2, lags NIR induction, such that N2O could be produced before 20 NOS is able to further reduce N2O to N2 (Cooper and Smith, 1963). This would result in high N2O emissions in treatments with depressed NOS when conditions next become amenable to denitrification. I saw this in corn with the 28-day rainfall interval. Presumably, the 14-day interval was not sufficient to make NOS disappear in corn systems (I did not assay NOS activity for 14-day intervals). In contrast, in switchgrass systems NOS was persistent even through a 28- day interval, shown by the lack of N2O emitted after rewetting, despite high NIR activity as indicated by total denitrification in DEA results (Figure 2.4). Many of the other soil variables I measured were not strongly correlated with N2O production rates and this contrasts with many other studies (Burger et al., 2005; Liang et al., 2016). NOS enzyme was best related to N2O emissions being greater in 2017 in the co-occurring June rewetting intervals. While both nitrification and denitrification can produce N2O, denitrification is accelerated when there are anaerobic conditions that I expect to see after rewetting soils. For example, in a nearby experiment, I found that 80-90% of N2O is produced through denitrification after rewetting from isotopomer results not included in this dissertation. I saw the same total denitrification rates in switchgrass and corn treatments using the acetylene block method (Groffman et al. 1999) though for many reasons (Groffman et al., 2006) this method may underestimate actual denitrification and perhaps differentially by treatment. Pulses of CO2 after rewetting is termed the "Birch effect", after its discoverer (Birch 1958). These pulses of CO2 are thought to be caused by both physical and biological factors, namely the changes in diffusion-limited substrates under drought and changes in microbial biomass (Evans et al. 2016). The same may be true for these pulses of N2O following rewetting. 21 Implications for predicting future effects of changing rainfall patterns on N2O emissions The mechanisms that control N2O emissions in the future may shift with intensified precipitation patterns. Although N2O emissions are small compared to other nitrogen fluxes, they have a profound effect on atmospheric chemistry because N2O is the main stratospheric ozone depleting agent and a powerful greenhouse gas. Thus, intensified precipitation patterns could change the global warming impacts of future cropping systems. This work challenges the assumption that with longer and more drier conditions, one might expect lower cumulative N2O emissions due to less consistent anoxic conditions. Results here show this may not be the case for long droughts in annual systems with high fertilizer inputs due to accelerated emissions upon drought cessation. Finding the controls on specific processes leading to N2O emissions in cropping systems will be critical to improving predictions and modeling in the response to more intensified precipitation regimes. 22 Chapter 3: Consequences of Changing Rainfall Patterns on Nitrous Oxide Fluxes at Different Abstract Landscape Positions Fluxes of nitrous oxide (N2O), a strong greenhouse gas produced by microbes in agricultural soils, are closely linked with soil moisture as moisture affects oxygen, nitrate, and carbon availability. Changing rainfall patterns due to climate change are thus likely to influence N2O fluxes. I tested the hypothesis that changing rainfall patterns will alter N2O fluxes in agricultural soils but with responses modulated by landscape position as positions affects soil texture and carbon contents. I used rainfall manipulation shelters to expose soils to the same amounts of rainfall delivered at different intervals (3-days, 14-days, and 28-days). Shelters were placed at both summit and toeslope positions within conventionally tilled corn fields in southwest Michigan, USA and N2O fluxes were measured three times per week during the summer for two years. Results show cumulative N2O fluxes under normal conditions were 2.3x higher in toeslope positions than in summit positions, consistent with higher soil carbon and finer soil texture in toeslope positions. Large rewetting events (longer dry intervals) increased cumulative N2O fluxes 32% in summit positions but did not change fluxes in toeslope positions. Changes in fluxes were associated with soil texture, inorganic N, and total C, but not C: N ratios. Knowledge of landscape patterns deserve inclusion in models of current and future climate change effects in order to better quantify and mitigate agricultural N2O fluxes. 23 Introduction Climate change influences biogeochemical cycles through alterations in temperature and precipitation, at times unpredictably (Greaver et al., 2016). Throughout the Midwest, rising temperatures and more extreme rainfall patterns have been observed and are expected to continue (Pryor et al., 2014), impacting potentially important biogeochemical fluxes. Though all greenhouse gas emission scenarios project intensified temperature and precipitation trends (IPCC, 2014), climate model projections for temperature are more certain than projections for precipitation (Pryor et al., 2014). Likewise, models predicting the biogeochemical responses to the effects of temperature are more certain than the models predicting responses to precipitation change (Robertson et al., 2013). Thus, it seems especially critical to better understand biogeochemical impacts of changing rainfall regimes. Agricultural nitrogen (N) cycles are likely to be particularly affected by changing rainfall patterns. Farmers manage soil nitrogen (N) from fertilizer and biological sources against losses from leaching, microbial processes, and plant uptake (Robertson and Vitousek, 2009), and N pollution to groundwater, surface waters, and the atmosphere represent major perturbations to the biosphere (Rockström et al., 2009). Dry periods and precipitation influence all soil biogeochemical processes, especially those that involve N (Greaver et al., 2016). Understanding N responses to changes in precipitation is important for designing future management strategies for mitigation and adaptation (Bowles et al., 2018; Robertson et al., 2013). Nitrous oxide (N2O) as a greenhouse gas is an important source of contemporary radiative forcing (IPCC 2014), and agricultural soils represent the largest anthropogenic source. Soil N2O fluxes exhibit extremely high spatial and temporal variability, related to different 24 combinations of soil attributes including carbon (C), nitrogen, and oxygen availabilities, microbial community composition, and plant-microbe-soil interactions. While we know much about the factors that control denitrification in the laboratory (primarily oxygen, nitrate, and available C), the complex regulation of these factors in situ creates variability that is difficult to quantify and model at field scales. High variability creates particular problems when scaling chamber-based measurements to the field or regional scale, which limits our ability to determine treatments that might best mitigate future emissions (Kravchenko and Robertson, 2015; McDaniel et al., 2014). In many ecosystems, these soil factors vary predictably by landscape position, with higher positions characterized by coarser soils with lower C contents and sometimes plant productivity, as compared to lower positions that are typically higher in clay, carbon, and associated properties (Beehler et al., 2017; Ladoni et al., 2016; Wickings et al., 2016). Izaurralde et al. (2004), for example, found along agricultural toposequences in Alberta and Saskatchewan greater N2O emissions in fine-textured depressions than in coarse-textured summit positions. Along a catena in Puerto Rico, McSwiney et al. (2001) found that N2O fluxes increased downslope as a function of redox status. Other factors known to affect N2O production do not necessarily differ by landscape position. In Iowa, heterogeneous soil landscapes had higher microbial biomass nitrogen at depressions and toe slope positions, but not nitrate pools (Kaleita et al., 2017). And Groffman (1993) found across a tall grass prairie landscape in Kansas that denitrification was not affected by slope position, although N2O-only production was not evaluated. And a study in southwest Michigan showed higher N2O fluxes in depressions, with no additional effect of cover crop presence (Negassa et al., 2015). 25 Soil texture may be particularly important as it has also been found to strongly influence N2O fluxes in laboratory settings (Zhu et al., 2013). After flooding soil columns representing three landscape positions from Pennsylvania, matric potential was shown to be the strongest predictor of the timing of N2O fluxes across soils that differed in texture, structure, and bulk density (Castellano et al., 2010). Also in the laboratory, Krichels et al. (2019) showed that the ability of a soil to drain or pond water influenced N2O fluxes. In an Illinois laboratory study of N2O emissions, iron redox reactions were found to strongly influence N2O fluxes along micro- topographical gradients (Krichels et al., 2019). Denitrification appears to be the dominant source of N2O in most mesic cropping systems (Aronson and Allison, 2012). Heterotrophic denitrifiers, mainly bacteria but also archaea and fungi, use soil organic C as an energy source and nitrate, which is thermodynamically similar to oxygen, as a terminal electron acceptor when oxygen is unavailable. Denitrification is often enhanced after precipitation (Bergsma et al., 2002; Gelfand et al., 2015) and especially in soil aggregates, rhizospheres, and decomposing litter, where soil microbes have ample access to C but limited access to oxygen (Khalil et al., 2004; A N Kravchenko et al., 2017; Loecke et al., 2008; Sexstone et al., 1988). The denitrification pathway products nitrite, nitric oxide, N2O, and dinitrogen are formed sequentially (Cooper and Smith, 1963), such that N2O is released to the soil atmosphere when denitrification is not carried out completely. Although considered a minor source of N2O in comparison to denitrification, nitrifying archaea and bacteria also produce N2O, from either nitrifier denitrification or hydroxylamine decomposition following ammonia oxidation (Bremner and Blackmer, 1978). Since nitrifiers require oxygen (O2) to oxidize nitrite to NO3-, nitrification proceeds fastest in aerobic conditions but can occur in soils with up to 60% water filled pore space (WFPS) (Abbasi and Adams, 26 2000). Nitrifier denitrification and coupled nitrification-denitrification can also produce N2O, most of which appears to be from nitrifier denitrification stimulated by wet-dry cycles (Wrage- Mönnig et al., 2018; Wrage et al., 2001). Regardless of its source, since both denitrification and nitrification are strongly affected by soil moisture (G P Robertson and Groffman, 2015), N2O fluxes are likely to be influenced by changing rainfall patterns. In particular, the effects of longer dry periods during the growing season on C and N will likely influence N2O emissions in inconsistent ways (Kim et al., 2012). Changing rainfall patterns are an aspect of climate change that may shift controls on biogeochemical mechanisms, leading to spatial variability and episodic events with higher than average influence on N fluxes (Bernhardt et al., 2017). Understanding interactions of changing rainfall patterns and landscape positions may be useful for better predicting the effects of future climate changes on greenhouse gas fluxes from agricultural soils, especially if different landscape positions respond uniquely to changing rainfall patterns. In general, experiments exploring topographical positions have shown greater emissions in fine-textured depressions than in coarse-textured summit positions, as noted above. But little data is available to elucidate the effects of changing rainfall patterns on soil N2O emissions across differing topographic positions. In this study, I hypothesize that under more intensified rainfall patterns, the greater organic matter, moisture, nitrate, and proportion of smaller pore spaces in toeslope positions will support higher rates of N2O production. Three specific questions follow: 1. How will N2O fluxes differ between toeslope and summit positions that differ in OM and soil texture; 2. Will rainfall intensification magnify fluxes; and, if so 27 3. Will landscape position influence magnification differently in toeslope and summit positions? Methods Experimental Site I conducted this study at the Kellogg Biological Station (KBS) Long-Term Ecological Research site in Hickory Corners, Michigan, USA (42° 24'N, 85° 23'W 288 m elevation). Average temperatures at KBS range from -3.8°C (January) to 22.9°C (July) with an annual mean of 10.1°C. Yearly precipitation is 1005 mm and the 30-year average (from 1981) for monthly precipitation ranges from 40 mm in February to 100 mm in May (Robertson and Hamilton, 2015). About 17% of precipitation occurs during the winter months from January to April, with the rest evenly distributed throughout the growing season. Over the past 30 years, precipitation events from April to December were, on average, 2.5 mm and ~3.6 days apart; around eight events per season were greater than 25 mm (https://lter.kbs.msu.edu/datatables/7). Potential evapotranspiration exceeds rainfall for four months out of the year (Crum et al. 1990). Soils at the site are well-drained mesic Typic Hapludalfs developed from glacial till and outwash (Collins and Crum 1995) with intermixed loess (Luehmann et al. 2016), and consist of co-mingled Kalamazoo (fine-loamy, mixed, semiactive) and Oshtemo (coarse-loamy, mixed, active) series soils (Robertson and Hamilton 2015). The predominant land use in the region is cultivated and successional fields, woodlots, residences, lakes, and wetlands. Experiments were conducted in fields of at least 2 hectares growing corn, soybean, wheat, and alfalfa in various rotations. I used fields in the corn phase of these rotations in both 2016 (three replicate fields) and 2017 (four replicate fields). Corn was managed as per common practice in the region, planted in 76 cm rows and fertilized at 150 kg N ha-1 in two applications: 28 25 kg N at planting and the remainder side-dressed ~6 weeks later when plants were at the V6 stage. Experimental treatments were established at two topographical positions in each field: summit and toeslope, as determined using a LiDAR-based digital elevation model and using terrain parameters of slope, profile curvature, and relative elevation (Table 3.1). Highest Elevation Lowest elevation Field F80 F83 F85 F91 A0S F98-04-01 (m) 280.6 283.6 287.7 284.8 285.0 285.2 Rainfall Manipulation Shelters Length of slope (horizonal distance from highest to lowest elevation) (m) 202.5 123.5 126.6 152.1 70.0 80.6 Slope (°) 3.3 2.4 3.5 8.0 3.6 4.2 (m) 274.0 280.7 283.3 272.5 282.5 281.8 Table 3.1: Field identification, summit elevation, toeslope elevation, length of slope, and slope degree of toposequences. Small rainfall manipulation shelters were used to create rainfall environments whereby soils were exposed to different rainfall intervals, such that the same amounts of water were added to each shelter but at different times. There were three shelters at each landscape position in each field. The distance between each shelter was ~10 cm and rainfall treatments were assigned at random (Figure 3.1). a b Figure 3.1: Sketch of experimental design a) from a profile view of two landscape positions in a field and b) outline of three shelters side by side. 29 Shelters were well-vented and 61 × 122 × 40 cm high and placed between corn rows; each shelter contained a cylindrical gas sampling frame (28.5 cm diameter × 30 cm high; 19.1 L) and sufficient soil surface area to allow periodic soil sampling. Shelters were constructed of polycarbonate (roofs), plexiglas (sides), and sheet metal (bottom edging). Bottom edging allowed 3 cm insertion into the soil to eliminate the potential for storm-derived lateral surface flow, which was in any case low to absent due to high infiltration rates in these well-drained soils. Three large holes 6.4 cm diameter drilled in the front and back sides of the shelters allowed airflow and eliminated humidity buildup. N2O flux measurements taken inside and outside small shelters placed under a larger (10 x 5m) shelter showed no significant small shelter effects on fluxes during either low or high flux events (Glanville, in prep). Roofs were hinged at the back to allow for periodic soil and gas sampling. Each shelter covered a 0.6 m2 area. Rainfall Intervals Rainfall additions were based on a 30-year regional average calculated from 1984 to 2014. For each transect, to each of three shelters, per landscape position, rainfall was applied at either 3-day (long-term average for the region), 14-day, or 28-day intervals. The 3-day average was 7.1 ±1.3 mm (standard deviation) not accounting for interception by the canopy. Since the soil under each shelter was protected from ambient rainfall, the amount of water added was independent of unusually dry or wet weather; thus, the experiment was repeatable across two field seasons. Gas sampling N2O, CH4, and CO2 gas fluxes were measured an average of three times per week to capture responses to rainfall events using the static chamber method (Holland et al. 1999). N2O and CH4 were measured using two techniques. First, for the "GC method" I took four chamber 30 headspace gas samples (20 mL) every 15 minutes for 60 minutes. The samples were stored over- pressurized in glass vials and analyzed for N2O, CO2, and CH4 concentrations in the lab using a gas chromatograph (7890A Agilent Technologies Inc., DE, USA) equipped with an ECD for N2O and an infrared gas absorption analyzer (LI-820, LI-COR, NE, USA) for CO2 (http://lter.kbs.msu.edu/protocols/113). Fluxes were calculated as the linear increase in N2O concentrations over the 60-minute closure period. The second technique was the "MICOS (Mobile Integrated Cavity Output Spectroscopy) method" that used a portable quantum cascade laser (QCL) analyzer (Los Gatos Research Inc., CA, USA) in the field for continuous chamber headspace measurements of N2O and CH4 concentrations, and a closed-path infrared gas analyzer (IRGA; LI-7000, LI-COR, NE, USA) for CO2 concentrations. During the measurement period, the headspace air inside each chamber was circulated through the QCL and IRGA analyzers for 5 minutes and gas concentrations were analyzed at 1 second intervals. Soil Sampling Soil moisture and temperature were measured every 10 minutes with sensors (Model 5TM, Decagon, WA, USA) installed vertically at the surface with readings stored on data loggers (EM50, Decagon, WA, USA). Soil was sampled to 20 cm depth for inorganic N, labile C, and dissolved organic C around co-occurring rewetting events, as described below. Soil texture was also measured in the shelter locations within each plot using the hydrometer method (https://lter.kbs.msu.edu/protocols/108). Here, 40 g of soil were shaken with 100 mL water for 24 hours using hexametaphosphate as a dispersant. The slurry was then put into cylinders, brought to 1000 mL with supplemental water, and shaken to suspend soil particles. Hydrometer and temperature readings were then taken over 8 hours. Since sand falls out of solution too 31 quickly to accurately record density changes, sand was determined separately with a 53 µm sieve. Once during the season, inorganic N was measured via 1 M KCl 100 mL extractions of 10g of fresh soil performed in duplicate. Soils were shaken on a shaker table overnight and allowed to settle before filtering. Extracts were frozen until analysis. A flow injector analyzer (QuickChem 8500 Series 2, Lachat Instruments, CO, USA) was used to determine nitrate and ammonium concentrations in extracts (https://lter.kbs.msu.edu/protocols/33). For gravimetric soil moisture 10 g of fresh soil were oven-dried at 60°C for 48 hours (https://lter.kbs.msu.edu/protocols/24). For total C and N analyses soil samples were finely ground and analyzed in triplicate by combustion with an elemental combustion system (ECS 4010, Costech Analytical Technologies, Valencia CA, USA) using acetanilide as a standard (https://lter.kbs.msu.edu/protocols/148). Statistical Analysis All statistical analyses were performed using R software version 3.6.1 (R Development Core Team 2019) and with a significance value of P<0.05. The statistical model included 2 landscape positions × 3 rainfall-interval treatments × 2 years and the interactions among them were considered fixed factors. Fields were considered a random factor. Landscape position is specified as the whole plot factor, the interaction between fields and landscape position is considered a random factor, and this interaction is used to test landscape position effects. The interaction among field, landscape position, and rainfall-interval treatment is specified as a random factor, in order to test for rainfall-interval effects. Rainfall-interval is the subplot factor. Analysis of variance was performed with the “lmer” function in the “lme4” package by considering landscape positions as a whole plot factor, and rainfall-interval treatments and years 32 as subplot and sub-subplot factors. Normality of residuals was checked by plotting residuals against the fitted values, and no violations of assumptions were found. Homogeneity of variance assumptions were examined by the “leveneTest” function in the “car” package and no heterogeneous variance was detected. Regression analysis was used to explore the relationships between N2O flux and inorganic N, total C, and soil texture. Results N2O emissions Daily N2O emissions in summits and toeslopes for 2016 and 2017 are shown in Figure 3.1. In summits, fluxes ranged from 0 to 245 μg N2O-N m-2 min-1 in 2016 and 0 to 251 μg N2O- N m-2 min-1 in 2017. In summits, higher fluxes occurred during the summer months of May, June, and July before declining to below ~1 μg N2O-N m-2 min-1 for August. The toeslope position fluxes ranged from 5.1 to 678 μg N2O-N m-2 min-1 in 2016 and 0 to 718 μg N2O-N m-2 min-1 in 2017, with fluxes highest in May, June, and July, and lower in August. Flux measurements were not taken past August in either year since earlier experiments (Chapter 2) had shown no rainfall N2O responses outside of the main growing season. Summit Toeslope ) e n l x u F O 2 N Δ i l e s a b m o r f ( 30 20 10 0 -10 Date August 15 September 15 3 day 14 day 28 day ) e n l x u F O 2 N Δ i l e s a b m o r f ( 10 5 0 -5 -10 3 day 14 day 28 day August 15 September 15 Date Figure 3.2: Daily N2O response to different rainfall patterns in the summit and toeslope landscape positions in 2016 and 2017. Error bars represent standard errors of the mean, n=4 replicate plots (each with three rainfall intervals). Red arrows denote 28-day watering events. Months omitted lack temporal resolution. 33 Cumulative N2O emissions are sums of measured fluxes. Cumulative N2O emissions in summit and toeslope positions yielded similar trends for 2016 and 2017 (Figure 3.2). For summits, cumulative N2O in 2016 was highest under the 28-day rainfall interval (5420 ± 2351 μg N2O-N m-2 min-2); fluxes under the 3-day and 14-day intervals were not significantly different from one another (3351 ± 1073 and 4229 ± 1040 μg N2O-N m-2 min-1, respectively). In 2016 and 2017 toeslopes N2O fluxes under the 3-day, 14-day, and 28-day intervals were not significantly different from one another (in 2016: 8053.3 ± 954.9, 8281.2 ± 890.0, and 7924.4 ± 949.2 μg N2O-N m-2 min-1, respectively). - m N O 2 N g µ ( x u l f ) 2 - O 2 N 80 60 40 20 0 Toeslopes b b b Summits b a a 3 day 14 day 28 day 3 day 14 day 28 day Rainfall Interval Figure 3.3: Cumulative N2O responses to different rainfall patterns in the summit and toeslope positions in 2016. Error bars represent standard errors of the mean, n=4 replicate plots (each with three rainfall intervals). Soil Properties: Soil texture, C:N, moisture, and temperature N2O fluxes increased linearly with % clay + silt content; in general, there was a higher percent sand in summit positions and lower percent sand in toeslope positions, resulting in a positive linear regression (r2 = 0.63, P<0.05; Figure 3.3). C: N rations ranged from 8.6 to 13.6 (Table 3.2) 34 with no detectable relationship to N2O flux. Soil temperature ranged from 9.7 to 29.4 and 8.7 and 28.4 °C, in summits and toeslopes respectively. (Figure 3.4). For soil moisture (Figure 3.5) in summit positions, the 14-day interval treatment had both the maximum and minimum values of 0.266 and 0.062 cm3 cm-3, respectively. The 3-day interval had the highest mean VWC at 0.129 cm3 cm-3 and the 28-day interval had the lowest mean at 0.107 cm3 cm-3. In toeslope positions, the lowest VWC was 0.003 cm3 cm-3 in the 28-day interval while the maximum was 0.289 cm3 cm-3, also in the 28-day interval. The 3-day and 14-day intervals had the highest mean VWC at 0.122 and 0.125 cm3 cm-3 respectively, while the 28-day interval in the toeslope had the lowest mean VWC at 0.077 cm3 cm-3. Total N (%) Total C (%) C:N Toeslopes Summits 0.09±0.01 0.07±0.01 0.90±0.14 0.73±0.09 10.0±0.13 10.7±0.29 Inorganic N (mg kg-1) NO3-N 6.1± 4.5 1.9 ± 0.5 NH4-N 14.5 ± 2.1 3.1 ± 1.6 Figure 3.4: Percent clay + silt by average N2O flux for the 3-day rainfall interval. Error bars are omitted for clarity. The dashed line represents a linear regression had significant effects (P<0.05). Table 3.2: Mean values for five fields with standard error for total C, total N, C:N, and inorganic N. 35 Summit Toeslope 30 20 10 ) C ° ( e r u t a r e p m e T 0 Aug 15 Sept 15 Date Figure 3.5: Soil temperature fluctuations in summit and toeslope positions. Discussion N2O dynamics by landscape position (Question 1) How do N2O fluxes differ between toeslope and summit positions as influenced by OM and soil texture? Average N2O fluxes at toeslope positions were higher than for summits, by a factor of 2.2 for normal rainfall intervals and 1.7 for 28-day intervals. Across both positions, fluxes were comparable to those measured for the unsheltered portions of neighboring fields planted to corn in prior years (Gelfand et al., 2016; Oates et al., 2016). Higher fluxes in toeslope positions appears associated with higher total C (Table 3.2) and higher % clay and silt (Figure 3.3). These results agree with other studies that found higher OM and N2O emissions from toeslopes than from summits at the same site (Negassa et al., 2015) and as well over larger regions (Florinsky et al., 2009). That soil texture is highly correlated with average N2O flux in the 3-day rainfall interval is consistent with laboratory incubation studies (Castellano et al., 2010; Zhu et al., 2013). This is likely because of textural effects on WFPS, which in turn controls aeration (Schjønning et al., 2003). 36 N2O emissions as affected by rainfall pattern (Questions 2 and 3) Will rainfall intensification magnify fluxes and, if so, will landscape position influence this magnification differently in toeslope positions? The 28-day rainfall interval in corn accelerated N2O emissions – cumulative fluxes were 2x higher than cumulative fluxes for the normal rainfall interval, but only for summit positions. The higher N2O fluxes in toeslope positions were not affected by rainfall intervals. Figure 3.1 shows responses to individual water additions at all rainfall intervals in both landscape positions. Though my work did not elucidate why these soils responded differently by landscape position upon rewetting, other studies show the response is largely dependent on drainage ability and matric potential (Castellano et al., 2010; A. Krichels et al., 2019; Syväsalo et al., 2004). However, these studies did not vary rewetting treatments by amount or interval, so more can be done to understand the impact of changing rainfall patterns. Factors associated with N2O emissions Variations in N2O production between the two years and positions did not appear related to temperature or C:N patterns over the period examined. A similar absence of relationship to C:N and N2O fluxes have been found in other studies (Castellano et al., 2010; Syväsalo et al., 2004). However, I found higher inorganic soil N in toeslope positions (Table 3.2), where N2O emissions were high, than in summit positions. Greater resource for nitrification, which produces N2O through the oxidation of ammonia (NH3), and for denitrification, which produces N2O through the reduction of nitrate (NO3-), could both produce more N2O (Saha et al., 2017). However, since N2O fluxes increase after rewetting when anaerobic microsites proliferate, I infer that higher toeslope position fluxes are from denitrification. Li et al. (2018) found topography to explain the greatest amount of variation in dentification potential in the Delmarva Peninsula. 37 Implications for predicting future effects of changing rainfall In corn systems, N2O emissions will be influenced by predicted intensified precipitation regimes but differently across a field. In corn systems, N2O fluxes may be small in comparison to other N losses like total denitrification or nitrate leaching, but N2O fluxes have a disproportionally large impact on the radiative forcing of the atmosphere since N2O is a powerful greenhouse gas. Results also suggest that including information on landscape position and rainfall intensification in calculations of field-scale N2O fluxes would lead to more accurate landscape and regional contributions of N2O emissions in corn systems. Additionally, the greatest opportunity for N2O mitigation from fields with varying topography is in toeslope positions, where N2O emissions are high. 38 Chapter 4: Impacts of Changing Precipitation Patterns on Soybean (Glycine Max L.) Biological Nitrogen Fixation as Mediated by Landscape Position and Tillage Abstract Cropping systems receive N additions primarily from organic amendments, synthetic fertilizers, and biological N2 fixation (BNF). These processes, most notably the use of fertilizer, have accounted for dramatic anthropogenic increases in reactive nitrogen (Nr) on Earth over the past 100 years; currently ~25% of terrestrial BNF is from soybean cultivation. BNF differs across cropped fields, which may be related to soil properties that covary with soil texture and water availability. Changes in rainfall patterns associated with climate change will create changes in the timing of N-mineralization and other N transformations, potentially promoting or suppressing BNF. Given these changes and the importance of BNF to legume crop productivity, soil fertility, the productivity of subsequent non-leguminous crops, and Nr in the environment, it seems prudent to understand potential effects of future changes in rainfall on BNF. Since BNF may be affected by soil organic matter (OM) and texture, two primary determinants of soil moisture, one might expect the strength of BNF responses to changes in rainfall to vary by tillage and topographic position as affected by OM and texture. In particular, there may be larger impacts of rainfall changes at summits and in tilled systems where there is less OM and less water holding capacity relative to toeslopes and no-till systems, respectively, and therefore greater potential impacts on BNF. Here I test the hypotheses that BNF is sensitive to changing rainfall patterns in till vs. no- till and in summit vs. toeslope positions due to associative relationships in texture and OM. I measured soybean BNF by 15N natural abundance using nodulating and non-nodulating isolines 39 to test effects of landscape position, its potential interaction with rainfall, and tillage management. Soybean BNF was 41% higher at summit than toeslope positions, consistent with soil OM and texture patterns. Percent BNF (%BNF) is the fraction of N from BNF divided by the total N in the plant. When precipitation was increased by 20%, BNF decreased from 75.7 to 51.7 %BNF at summit positions, though was unaffected at toeslope positions. BNF response to 3- week rainfall intervals also differed by tillage. In no-till plots, %BNF decreased 15% with less intense rainfall patterns in conventionally tilled fields, whereas BNF increased 14% with less intense rainfall in no-till plots. Results reinforce the importance of topographic position for predicting soybean BNF responses to changing rainfall patterns and show that summit positions are more sensitive to additional rainfall. Results also show changes in rainfall intensity affect BNF in tilled differently than in no-till soils, perhaps because greater retention of organic matter in no-till systems results in less available plant N. Models that incorporate these interactions will be better able to characterize legume crop performance and N use across landscapes and improve global estimates for BNF. 40 Introduction Cropping systems receive nitrogen (N) additions primarily from organic amendments, manufactured fertilizers, and biological N2 fixation (BNF). These inputs, most notably the use of fertilizer, account for dramatic increases in reactive N (Nr) on Earth over the last 100 years (Vitousek et al., 2013). When not taken up by plants or soil microorganisms, Nr can be lost from cropping systems and become pollutants in water and the atmosphere. BNF contributes ~25% to Nr inputs to the biosphere (Vitousek et al., 2013); soybeans (Glycine max L.) in particular contribute 10.4 Tg Nr yr-1, representing ~18% of total global BNF inputs (Gelfand and Robertson, 2015). Soybean BNF can substitute for N fertilizer application and has a lower environmental cost since systems with BNF as a major source of Nr require less N fertilizer and tend to have lower hydrologic Nr losses (Blesh and Drinkwater, 2013; Syswerda et al., 2012). Thus, the sustainability of food production systems can be aided by BNF management and associated feedbacks (Pearson, 2007). BNF transforms N2 gas from the atmosphere into a form that can be incorporated into the tissue of certain plants, and global rates of BNF have increased due to agricultural intensification (Herridge et al., 2008). BNF is an energetically expensive process for plants since it requires breaking the N2 molecule’s triple bond when transforming it to NH3 and then to amino acids. In legumes such as soybeans, BNF occurs through a symbiotic association between plants and bacteria such as Bradyrhizobium japonicum (Bottomley and Myrold, 2015) and thus depends on interacting factors that affect crop growth and associated N demand (Salvagiotti et al., 2008). Long-term studies of farms and fields planted to soybeans show a wide range of soybean BNF rates across management and fertilization gradients, with greater N availability’s generally 41 suppressing BNF (Gelfand and Robertson, 2015; Grossman et al., 2011; Oberson et al., 2007; Tamagno et al., 2018). Soil physical characteristics influencing BNF include redox state and water availability. Chemical controls include the availability of inorganic N and N fixation cofactors such as iron (Fe), molybdenum (Mo), and vanadium (V). Biological factors include plant genotype and growth rates; rhizobia numbers, diversity, and effectiveness; and pests and pathogens (Unkovich et al., 2008). Thus, BNF is influenced by many of the same agronomic and ecological factors that affect plant nutrition, water availability, herbivory, and disease (Valentine et al., 2010). BNF can supplement N availability to future crops but can also result in decreased N fertility (Bottomley and Myrold, 2015; Salvagiotti et al., 2008) when N exported through legume harvest is greater than N fixed through BNF (Ciampitti and Salvagiotti, 2018). Global calculations of BNF assume that BNF rates are homogenous over individual fields, but soybean BNF has been shown to vary with soil properties. In southwestern Victoria, Australia, for example, in a survey of 71 dairy pasture sites rates of BNF in white clover (Trifolium repens) ranged from 0 to 100% across different soil textures, though were 7% higher on light-textured soils (Riffkin et al., 1999). In a Denmark pea (Pisum sativum L.) field sampled at 56 points, BNF ranged from 26% to 81% of total plant N (Hauggaard-Nielsen et al., 2010). At sites in Central New York, USA, soybean BNF at various farms differed was affected controls in different soil types (Schipanski et al., 2010). Climate change, particularly changes in rainfall intensity and amount, may also affect BNF. The US Midwest, an intensive agricultural region responsible for >80% of US soybean production (USDA NASS 2018), has experienced increasingly more intense and frequent heavy rainfall over the past century (Pryor et al., 2014). Furthermore, global circulation models predict 42 that both the increasing length of dry intervals between precipitation events and the amount of precipitation falling in single events will further increase (Pryor et al., 2014). In Michigan, both the number of extreme precipitation events and observed annual precipitation amounts are increasing (Frankson and Kunkel, 2017). Changes in rainfall will likely be accompanied by changes in the timing of N-mineralization and other N transformations (Bowles et al., 2018; Robertson et al., 2013), potentially promoting or suppressing BNF. Given these changes and the importance of BNF to legume crop productivity and as well to soil fertility and subsequent non-leguminous crops, it seems prudent to examine potential effects of future climates on BNF. Since BNF appears related to soil organic matter (OM) and texture, two primary determinants of soil moisture, one might expect the strength of BNF responses to changing rainfall patterns to vary across a field as OM and texture vary, particularly along toposequences. Impacts of rainfall changes at summits may be greater than at toeslopes, for example, where there is usually more OM, clay, and water holding capacity and thus more buffering against rainfall variability. In general, BNF rates are lower with drought conditions due to nodule sensitivity to decreased phloem flow (Serraj et al., 1999). Studies investigating the interactive effects of N fixation and drought have mostly been performed in controlled greenhouse settings, where plants are not exposed to normal environmental weather and soil conditions and the rhizobia-plant symbiosis does not reflect the impact of rainfall changes. Furthermore, growth in artificial environments alters nodules, as well as nodule depth and distribution (Pueppke, 1986). Field experiments are needed to more accurately quantify fixation responses to changes in topography, rainfall, and tillage. This will aid models for predicting BNF under the varying crop, soil, and 43 meteorological conditions that may account for the spatial variation of BNF changes with rainfall. Here I test the hypotheses that BNF is vulnerable to changing rainfall regimes in summit vs. toeslope topographic positions and, separately, in till vs. no-till management as associated with differences in texture and OM. I address three specific questions: 1. How does soybean BNF vary by topographic position as affected by OM and its influence on N-mineralization and water availability; and 2. Does BNF in summit and toeslope positions differ in response to added precipitation; 3. Do changes in precipitation intensity, in particular the length of time between rainfall events, influence soybean BNF in tilled plots differently from no-tilled plots? Experimental Design and Site Description The study was conducted over four field seasons (2015-2018) at the Kellogg Biological Station (KBS) Long-Term Ecological Research site, Hickory Corners, Michigan, USA (42° 24'N, 85° 23'W 288 m elevation). Average temperatures at KBS range from -3.8°C (January) to 22.9°C (July) with an annual mean of 10.1°C. Yearly precipitation is 1005 mm on average (30- year mean) and the 30-year monthly precipitation rate ranges from 40 mm month-1 in February to 100 mm month-1 in May (Peters et al., 2013). About 17% of precipitation occurs during the winter months from January to April, with the rest evenly distributed throughout the growing season. Over the past 30 years, precipitation events from April to December were, on average, 2.5 mm and ~3.6 days apart; around eight events per season were greater than 25 mm (https://lter.kbs.msu.edu/datatables/12). Potential evapotranspiration exceeds rainfall for four months out of the year (Crum et al., 1990). 44 Soils at KBS are well-drained mesic Typic Hapludalfs developed from glacial till and outwash (Collins and Crum 1995) with intermixed loess (Luehmann et al., 2016) and of co- mingled Kalamazoo (fine-loamy, mixed, semiactive) and Oshtemo (coarse-loamy, mixed, active) series (Crum and Collins, 1995). The predominant land use in the region is cultivated and successional fields, woodlots, residences, lakes, and wetlands. Fields used are outlined in Figure 4.1. The experiments were conducted on conventionally tilled fields used for corn, soybean, and wheat planted to soybeans during the year of this study. These rainfed fields were managed as per regional practice, including fertilization with potassium (K2O), phosphorus (PO4-), and lime as recommended by soil tests. All fields were planted at 150,000 seeds ha-1 in 15-inch rows. Glyphosate was used to control weeds during soybean years. For all experiments I quantified soybean BNF by 15N natural abundance using nodulating and non-nodulating isolines, as described below. Figure 4.1: Kellogg Biological Station aerial view with fields used in 2015 in red; fields used in 2016 in green; fields used in 2017 outlined in yellow; and fields used in 2018 outlined in light blue. 45 Toposequence Experiment (Question 1) Figure 4.2: A illustration showing positions of transect, landscape positions, and soybean plot locations for the toposequence experiment. For the toposequence experiment, I used three fields of 6-8 ha each (KBS fields 30-1, 38, and 79-8 south). Soybeans were planted in 2015 along four transects in three landscape positions (summit, midslope, and toeslope) in each field (Figure 4.2; Table 4.1). Plots were 1 × 1 m with five non-nodulating plants per plot, at normal planting density and the same density as in the rest of the field. For this 2015 topography experiment, at physiological maturity plants were harvested both above and below ground and separated into above-ground vegetative biomass, belowground biomass, and seeds as detailed below. 46 Highest Elevation Lowest Elevation (horizonal distance from highest to lowest elevation) Length of Slope Field 30-1 38 79-8 north 79-8 south 80 82-1 87 89+90 94 (m) 288.3 288.3 288.3 287.7 280.6 282.7 281.3 285.1 282.5 (m) 276.7 282.6 284.4 280.5 274.0 273.4 272.8 279.8 272.1 (m) 194.6 324.4 116.6 310.0 202.5 235.1 136.0 204.0 494.1 Slope (°) 6.0 1.8 3.4 2.3 3.3 4.0 6.2 2.6 2.1 Table 4.1: Field identification, summit elevation, toeslope elevation, length of slope, and slope degree of fields used. Rainfall Amounts by Landscape Position Experiment (Question 2) Figure 4.3: A illustration showing positions the landscape positions and rainfall addition plot locations for the rainfall amounts by landscape position experiment. For the rainfall addition experiment, I used three fields of 2-12 ha each (KBS Fields 79-8 north, 82-1, and 87 in 2016; and 80, 89+90, 79, and 94 in 2017). Soybeans were planted at four toeslope and four summit locations in each field; there were not transects as in the previous experiment (Figure 4.3). Each location had a plot 2 × 2 m with 15 nodulating plants and 15 non- nodulating plants, at the same density as the rest of the field. From planting to harvest one half of 47 each plot was rainfed and the other half was irrigated with 2.5 mm of water when dry for three days. Water application was allowed to infiltrate completely so water did not spread neighboring plants. Planting date, harvest date, ambient rainfall, and additional rainfall are shown in Table 4.1. Soybean plants were harvested aboveground and separated into seed and vegetative biomass. Roots were not sampled. Planting Date June 22 Year 2016 Harvest Date September 28 and 29 Ambient Rainfall (mm) 158 Additional Rainfall (mm) 40 2017 June 22 September 28 and 29 331 37.5 Precipitation Increase (compared to ambient rainfall) (%) 20.2 10.2 Table 4.2: Planting date, harvest date, ambient rainfall, and additional irrigation for 2016 and 2017. Rainfall Intervals by Tillage Experiment (Question 3) Figure 4.4: A illustration showing how rainfall intervals were conducted, with and without a shelter, in tilled and no tillage soils for the rainfall intervals by tillage experiment. I conducted the tillage experiment in four replicate 1 ha plots in each of the conventional and no-till treatments of the KBS LTER main cropping system experiment (MCSE; Robertson and Hamilton, 2015). In 2015, in one 4 x 4 m subplot per replicate I imposed a rainfall interval of 48 two weeks for the period between soybean planting and harvest; adjacent control subplots received ambient rainfall (Figure 4.4). Rainfall was excluded by complete-exclusion rainfall manipulation shelters (5 × 5 × 1.5 m high as described in Hess et al. 2019) and irrigated with captured rainwater delivered via overhead sprinkler heads. The experiment was repeated in 2018 but with a three-week interval. Soybeans plants were harvested aboveground and separated into seed and vegetative biomass. Roots were not sampled. Year 2015 May 19 - 21 Planting Date Harvest Date October 2 2018 May 25 - 28 September 29 Ambient Rainfall (mm) 659 387 Rainfall Interval 2 Weeks 3 Weeks Table 4.3: Planting date, harvest date, ambient rainfall, and additional irrigation for 2015 and 2018. Methods Biological Nitrogen Fixation I quantified soybean BNF by using 15N natural abundance in nodulating (Pioneer P22T69R) and non-nodulating (PI 547695, seed source: 04U-3266; Horosoy variety) soybean isolines obtained from the USDA Soybean Germplasm Collection (USDA-ARS, Urbana, IL). To plant the non-nodulating isolines, I removed nodulating seeds from the soil and replaced them with non-nodulating seeds. Seeds were planted in late May and early June and plants harvested in late September and early October at R6.5 (physiological maturity). Soil Analyses Soil was sampled in all experiments by compositing four 2.5 cm diameter × 25 cm depth push-probe soil cores on each sample date in each subplot replicate. Soils were passed through a 4 mm mesh screen and analyzed for texture, inorganic N, N mineralization, organic matter, gravimetric moisture, and pH at the time of peak N fixation mid-season (Table 4.4). 49 2016 and 2017 Topography Study with Rainfall -BNF of seeds only -soil texture -organic matter -pH 2018 Conventional Tillage vs No-till by 3-day vs 21-day interval -BNF of seeds only -soil texture -organic matter -pH 2015 Topography Study -BNF of all plant parts separately -N mineralization -organic matter -pH -gravimetric moisture Conventional Tillage vs No- till by 3-day vs 14-day interval -BNF of seeds only -soil texture -organic matter -pH Table 4.4: Experiment of each year with associated measurements. Soil texture was measured using the hydrometer method (https://lter.kbs.msu.edu/protocols/108). Forty g of air-dried soil were shaken for 24 hours in 200 mL jars using sodium hexametaphosphate as a dispersant. The slurry was then put in 1 L cylinders and supplemented with water. Hydrometer and temperature readings were taken over 8 hours. Since sand falls out of solution too quickly to accurately record density changes, to supplement hydrometer readings sand from a separate sample was sieved out with a 53 µm mesh. Soil pH was measured for two duplicate subsamples. A slurry of 15 g of field moist soil and 30 mL deionized water was shaken by hand for 10 seconds than allowed to settle for 30 minutes before measuring pH (VWR International, Randor, PA) (https://lter.kbs.msu.edu/protocols/163) For 2015, Fe, K, P contents were analyzed on composite soil samples by atomic absorption following extraction with 0.1 M hydrochloric acid by the MSU Soil and Plant Nutrient Laboratory (East Lansing, MI). Soils were also by comparable treatment and analyzed 50 for Fe, K, P by the Cornell Nutrient Analysis Laboratory (Ithaca, NY) using the Modified Morgan analysis. Inorganic N was measured via a 1 M KCl 100 mL extraction of 10g of fresh soil performed in duplicate. A flow injector analyzer (QuickChem 8500 Series 2, Lachat Instruments, CO, USA) was used to determine nitrate and ammonium concentrations (https://lter.kbs.msu.edu/protocols/33). Potential N mineralization was measured via a 28-day laboratory incubation where 10 g of soil were held at 60% WFPS in a 25 °C incubator. Inorganic N was measured by extracting 10 g of soil in a 100 mL solution of 1 M KC, followed by shaking and filtration. Organic matter was measured via loss on ignition whereby soil was weighed into crucibles and placed in a 500° C oven for two hours. Weights were recorded before and after combustion and adjusted for soil moisture. For gravimetric soil moisture 40 g of fresh soil were oven-dried at 60°C for 48 hours (https://lter.kbs.msu.edu/protocols/24). Plant Analysis Whole plants were harvested at physiological maturity (R6.5) to determine total nitrogen fixed and soybean biomass. For root harvest in 2015 (toposequence experiment), coarse and fine roots were collected within a 25 × 25 × 25 cm soil volume. Soil was then carefully shaken from roots in the field and roots were examined for the presence or absence of nodules. Roots were then placed on a 0.1 mm screen and washed gently with water to remove adhering soil. All plants from the same replicate subplot were composited and dried to a stable weight in a 60° C forced air oven. Plant material was divided into grain, aboveground vegetative biomass (stem, leaves, and pods, less grain), and, when sampled, belowground biomass (roots including nodules). 51 Biomass was weighed and ground to pass through a 1 mm sieve and 3-5 mg of homogenized plant material was weighed into tins. Packed tins were analyzed for 15N and %N at the University California - Davis and Michigan State University isotope facilities. The ash correction was determined by burning subsamples in a muffle furnace at 550° C for four hours and weighing the remaining residue. BNF calculations The 15N abundance method for BNF allows for end-of-season sampling to represent growing season fixation (Unkovich et al., 2008). The 15N is used to distinguish plant N derived from soil vs. air (air contains very little 15N compared to soil). The fraction of N derived from the atmosphere (fNdfa) is: fNdfa = (δ15Nref - δ15Nfix) / (δ15Nref - δ15Nb), where Nref represents tissue from non-N2-fixing soybean isolines, Nfix represents tissue from N2- fixing soybean isolines, and Nb represents tissue from N2-fixing soybean isolines grown with atmospheric N2 as the only N source. I used the δ15Nb value determined by Gelfand and Robertson (2015), who used the same commercial variety grown in N-free sand culture in a KBS greenhouse with N-free Hoagland's solution (0-7-5 NPK with micronutrients; GreenCare Fertilizers, Chicago, IL, USA). Statistical Analysis Project 1: All statistical analyses were performed using R software version 3.6.1 (R Development Core Team 2019) and with a significance value of P<0.05. I fit the BNF data with a linear mixed 52 model using the “lme4” package with landscape positions as a fixed factor, and fields and transects as random factors. To compare BNF differences among landscape positions, pairwise T-tests were conducted with the “lsmeans” package. Project 2: The statistical model included two rainfall treatments and two landscape positions and the interactions between them were considered fixed factors. Fields were considered a random factor. Landscape position was specified as the whole plot factor, the interaction between fields and landscape position were considered a random factor, and this interaction was used to test landscape position effects. Analysis of variance was used by considering landscape positions as a whole plot factor, and rainfall treatments as a subplot factor. Project 3: The statistical model included two rainfall treatments and two tillage treatments and the interactions between them were considered fixed factors. Blocks were considered a random factor. Tillage treatment was specified as the whole plot factor, the interaction between block and tillage treatment was considered a random factor, and this interaction was used to test tillage treatment effects. Analysis of variance was used by considering crops as a whole plot factor, and rainfall-interval treatments as a subplot factor. For all: Normality of residuals was visually checked by plotting residuals against fitted values, and no violations of assumptions were found. Homogeneity of variance assumptions were examined by the “leveneTest” function in the “car” package and no heterogeneous variance was detected by Levene’s test. Simple linear regressions were used to determine the relationship 53 between %BNF and soil texture, OM, pH, and fixation cofactors, holding %BNF as the dependent variable. Results Toposequence Experiment (Question 1) BNF contributed 77.1 ± 4.9% (standard error of the mean) to the N content of grain in summit positions and 36.5 ± 6.6% at toeslope positions. Backslope landscape positions were intermediate with a BNF contribution of 62.6 ± 4.4% (Figure 4.5a). Measured net primary productivity was highest in toeslope positions (115.3 ± 13.5 g plant-1) and lowest in summit positions (47.7 ± 8.1 g plant-1) (Figure 4.5b). I found no differences in BNF contributions to aboveground or belowground vegetative tissues; only in seeds was there a BNF by landscape position effect. The %NPP from BNF showed no significant trends with landscape position (Figure 4.5c). 54 Figure 4.5: Percent BNF by landscape position (a), NPP by landscape position (b), and %NPP from BNF (c). Error bars represent standard errors of the mean, n=4 replicate plots. Asterisks indicate significant effects (P<0.05). Potential N mineralization was lowest in backslope positions at 51.4 ± 14.1 (mg NH4+ and NO3-) N kg-1 soil day-1 and highest at toeslope positions and lowest at 143.0 ± 11.3 (mg NH4+ and NO3-) N kg-1 (Figure 4.6). a b b Figure 4.6: N mineralization potentials at toeslope, backslope, and summit positions. Error bars represent standard errors of the mean, n=4 replicate plots. Asterisks indicate significant effects (P<0.05) with soil texture. 55 Rainfall Amounts by Landscape Position Experiment (Question 2) In rainfed plots, BNF contributed 37.7% to the N content of grain in summit positions and 75.7% in toeslope positions. In irrigated plots, BNF contributed 51.7% to the N content of grain in summit positions and 35.0% in toeslope positions (Figure 4.4). a Figure 4.7: BNF (%) increased at summit positions with more precipitation. Error bars represent standard errors of the mean, n=4 replicate plots. Asterisks indicate significant effects (P<0.05) with soil texture. Rainfall Intervals by Tillage Experiment (Question 3) In 2015, in the ambient rainfall treatment BNF contributed from 66 ± 6% and 64 ± 3% to the N content of grain in tilled and no-till plots, respectively. In soybeans experiencing 2-week rainfall intervals, BNF’s contribution to the N content of grain was 68 ± 4% and 69 ± 4% in tilled and no-till plots, respectively (Figure 4.5). 56 Figure 4.8: Percent BNF in no-till and conventional tillage with ambient and 2-week rainfall intervals. Error bars represent standard errors of the mean, n=4 replicate plots. In 2018, with 3-week rainfall intervals, the BNF contributions to the N content of grain in tilled and no-till plots were 66 ± 3% and 82 ± 1%, respectively. In ambient rainfall intervals, the BNF contributions were correspondingly 81 ± 5% and 68 ± 5%, respectively (Figure 4.5). Figure 4.6: Percent BNF in no-till and conventional tillage with ambient and 3-week rainfall intervals. Error bars represent standard errors of the mean, n=4 replicate plots. Asterisks indicate significant rainfall interval effects (P<0.05) within tillage systems. 57 Soil Properties: Soil Texture, OM, Mo, Fe, K, P, and pH The contribution of BNF to plant N content linearly decreased with soil sand content and increased with silt content (Figure 4.6) and OM (Figure 4.7). There was a higher percent sand in summit positions and lower percent sand in toeslope positions (Figure 4.8). Fe concentrations expressed the least range in concentrations from 0.8 to 5.1 mg kg-1. Likewise, P differences ranged from 0.9 to 8.2 mg kg-1. K concentrations expressed the greatest range, from 26.3 to 141.4 mg kg-1. I found no relationship between BNF and soil Mo, Fe, P, or K concentrations. All pH values for all experiments were between 5.7 and 7.1. Figure 4.9: Percent sand, silt and clay by %BNF. Error bars are omitted for clarity, n=4 replicate plots. Asterisks indicate significant effects (P<0.05) with soil texture. 58 Figure 4.10: Percent BNF by %OM. Error bars are omitted for clarity, n=4 replicate plots. P value = 0.0116 Discussion As hypothesized, the contribution of BNF to total plant N (% BNF) was highest in summit positions and lowest in toeslope positions, coincident with texture and organic matter. More rainfall at summit positions suppressed BNF, as did 3-week rainfall intervals in tilled plots. BNF by Toposequence (Question 1) How does soybean BNF vary by topographic position as affected by OM and its influence on N-mineralization and water availability? Percent BNF ranged from 0 – 94%, reflecting ranges seen in the literature (Salvagiotti et al., 2008; Schipanski et al., 2010) and at a nearby site (Gelfand and Robertson, 2015). Nevertheless, I observed a significant effect of landscape position on BNF, correlated with soil texture and N-mineralization rates. %BNF was higher at summit positions where soils were coarser and N mineralization rates were lower. The differential response of N2-fixing and non-N2-fixing soybeans to the fertility and textural 59 gradients reveals the influence of soil properties along toposequences on BNF across heterogeneous fields. Other studies have also revealed in situ soil properties that influence BNF across individual fields. Riffkin et al. (1999) documented higher rates of BNF on sandier soils in Australia, as did Shipanski et al. (2010) in New York. Riffkin et al. (1999) found differences of 7% between light- and medium-textured soils; this compares to my finding of 50% differences, on average. However, in KBS fields, when BNF contribution is scaled by grain yield, there were no differences between landscape positions in the total amount of Nr supplied through BNF (Figure 4.2c). I found no differences in %BNF in roots by landscape position, though this assumes that %BNF of roots at all depths are similar to those at the 0-25 cm depth that I sampled. However, this may not have sufficiently captured differences since soybean roots grow deeper than 150 cm (Ordóñez et al., 2018) and root distributions have been known to relate to claypan soil properties across toposequences (Myers et al., 2007). Rooting depths at different landscape positions may have differed due either to water availably or low permeability soil layers caused by tillage or geology. Whole plant %BNF reflect changes in all parts of the plant, though trends are especially prevalent in seeds since they are the biggest sink for N and have the highest concentrations (~6%). Vegetative matter and belowground biomass have much lower N contents (~0.8% and ~1.3%, respectively) and %BNF (~4.5% and ~15%). Thus, Nr through soybean production is mostly associated with the grain. Higher %BNF in summit positions is correlated with lower N mineralization rates (Figure 4.3), congruent with other studies that note lower %BNF where mineralization rates are high 60 (Schipanski et al., 2010). Soils with higher sand content tend to have less organic matter and N in comparison to soils with more clay (Six et al., 2000), thus sandier soils would be expected to have lower N mineralization rates. Comparing N assimilation between non-nodulating and N2-fixing soybeans may indicate the ability of N2-fixing soybeans to allocate carbon belowground and induce N mineralization in low fertility soils, such as those at summit positions. I found soil N assimilation by non- nodulating and N2-fixing soybeans was similarly high in toeslope soils with their finer texture, evidenced by lower % BNF. George et al. (1993) found similar differences across an elevation and fertility gradient, with more soil N uptake in N2-fixing soybeans compared to the non- nodulating plants at low soil N availability. This indicates that N2-fixing soybeans may be able to circumvent N limitation, through carbon allocation, even as soil N stocks decline. Roots obtain oxygen from pores in the bulk soil environment. Oxygen availability is an important regulator of nitrogenase activity; legume nodules can have four times the oxygen demand of an equal biomass of roots (Layzell and Hunt, 1990). Soils with lower microbial respiration, then, may have more oxygen-rich environments capable of supporting a high number of nodules, which might promote more BNF. Thus, my finding of lower BNF rates at toeslope positions may also be due to less oxygen where soils have more clay and higher microbial activity. Additionally, disease can be a sign of low oxygen. While I did not observe any evidence of disease pressure, finer textured soils at toeslopes can have an increased prevalence of soybean pathogens (Workneh et al., 1999) due to lack of oxygen. Furthermore, differences in rooting depth between nodulating and non-nodulating plants can influence the accuracy of the methods used, potentially accounting for differences between landscape position (Schipanski et al., 2010). 61 BNF with Changes in Rainfall Amounts by Landscape Position (Question 2) Does BNF in summit and toeslope positions differ in response to added precipitation? In this experiment %BNF decreased with additional rainfall at all summit positions, but not at toeslope positions. The most likely explanation for suppressed BNF with additional irrigation at summit positions is increased inorganic N supply. Inorganic N pools can suppress BNF (Schipanski et al., 2010), and it's possible that added water stimulated additional N mineralization, which in turn suppressed BNF. Toeslope positions, with their higher ambient water contents and N mineralization potentials (Fig. 4.3), may have likewise had BNF suppressed by soil mineral N pools. Higher %BNF under drier (ambient) conditions in summit positions contrasts with studies that show nodule production, which is closely tied to BNF, generally decreases with reduced precipitation. Thus, one may have predicted additional rainfall to have increased %BNF. BNF is sensitive to soil drying because N2 fixation is more sensitive to soil conditions than to plant stress during drought (Abdelhamid et al. 2011), compounded by dry conditions that can lead to excess solutes in the root zone, restricting water availability to the bacteria (Walsh, 1995). That BNF was not inhibited under ambient conditions, is likely because summit soils were not sufficiently dry: In a study in Champaign, Illinois, Gray et al. (2013) found that drought stress must be greater than 41% of the historical average to inhibit nodulation. Drought stress at KBS was ~15% in 2016 and nil in 2017. The timing of dry conditions can also affect %BNF, which occurs differentially throughout plant stages (Gan et al., 2003). Furthermore, low nodulation response to high or low moisture conditions could persist throughout the season despite improved weather conditions and 62 complex spatial-temporal variation. If dry conditions are severe enough to inhibit BNF, there were will be lower BNF values, thus scaling up BNF values from soybeans under drier conditions can potentially lead to an underestimation of BNF (Gelfand and Robertson, 2015). BNF with Changes in Rainfall Intervals under Different Tillage (Question 3) Do changes in precipitation intensity influence soybean BNF in tilled plots differently from no-tilled plots? The same amounts of BNF occurred in tilled and no-till treatments under 3- day and 2-week rainfall interval treatments in 2015. However, when rainfall was excluded for three weeks in 2018, there was less BNF in no-till plots, where BNF decreased from 82 to 68 %BNF. This response fails to support the hypothesis that no-till management, with its higher organic matter content (Syswerda et al. 2011), will be better buffered against changes in rainfall intensity than conventional tillage management. In fact, counter to expectations, BNF under conventional tillage increased after drought from 68 to 80 %BNF, rather than decreased. Were the hypothesis supported, it would be expected BNF in the no-till to change little following drought and BNF in the tilled soil to decrease. There are several possible partial explanations for 1) BNF’s being greater in no-till than in conventional till systems under ambient intervals; 2) decreasing %BNF in conventional till exposed to the longer dry interval; and 3) increasing %BNF in no-till soybeans exposed to the longer dry interval. First, BNF could be higher in no-till than conventional due to higher N immobilization in no-till soils due to greater retention of organic matter, resulting in less available plant N following drought. This may be less of an issue in the ambient treatment because more consistent rainfall may have promoted more N mineralization. This is corroborated by past work of in situ 63 assays in the same plots, which show higher net N mineralization in no till plots vs. conventionally tilled plots (Millar and Robertson, 2015) Second, biogeochemical processes crate a vertical gradient of 15N through the soil profile with a higher distribution of 15N than in the upper 10 cm than in the lower depths (Natelhoffer and Fry, 1988). However, when soils are homogenized as they are in tilled plots, this disrupts the naturally occurring patterns of δ15N with soil depth in comparison to no-till plots. Since soybean roots are concentrated in the upper 15 cm (e.g. Böhm et al., 1977; Robertson et al., 1980), it is possible that soybeans in no-till systems may have different 15N uptake patterns than soybeans in tilled plots because of associated δ15N patterns. Other explanations are also possible. These include the potential for differences in rhizobia populations as well as tillage-related compaction that may have limited rooting development in tilled plots. In tilled plots, compaction caused by repeated tillage could form a tillage pan, which can be difficult for roots to penetrate. Keisling et al. (1995) found soybean roots in no-till systems followed classical taproot trends, but when tillage pans were present, roots followed old root channels and pan fractures. There is also speculation that increased disturbance, like tillage, will decrease the presence of effective rhizobia, though this remains to be tested specifically (Kiers et al., 2002). Trace element deficiencies could also contribute to lower %BNF, but there appear no trace element deficiencies in either conventional or no-till treatments (http://lter.kbs.msu.edu/datatables/354) BNF validity Various soil and environmental factors can influence BNF rates and the suitability of 15N natural abundance to accurately reflect rates, but three consistent outcomes in this study suggest that the values reported are robust. First, I found no relationship between soil P, K, pH and 64 soybean yield or fixation, which suggests that these factors were not influencing BNF differentially in these fields during the experimental period. Second, calculated BNF values never exceeded 100%, which can happen when isolines do not appropriately represent N uptake under non-BNF conditions. And third, the N content in both nodulating and non-nodulating soybeans responded in the same direction, i.e., increases in the N content of nodulating soybeans were paralleled by increases in the N content of non-nodulating soybeans, further suggesting good correspondence between nodulating and non-nodulating isolines. These three observations warrant connecting the three studies and drawing overall conclusions. Remaining Questions Results here suggest several lines of additional research that could be useful for further understanding spatial patterns on BNF in the field and responses to changing precipitation patterns. Three remaining areas in BNF in order to better understand and manage for topographic, tillage, and climate change issues are: 1) rhizobia strains; 2) other elements of global change; and 3) BNF rate differences in legumes and over space, and time. First, genetic variation among rhizobia strains may be sensitive to topographic position. I did not examine the distribution of strains, which are known to differ regionally (Batzli et al., 1992) and with management (Kiers et al., 2002). Weese et al. (2015), for example, found less- mutualistic rhizobia evolved in long-fertilized fields and likewise, there may be less-mutualistic rhizobia in toeslope positions where N mineralization is high. This may also be the case with different precipitation regimes, as we know that changes in regimes alter microbial communities (Zeglin et al., 2013). A major assumption of this study is that rhizobia strains are consistent across landscape position and with tillage treatments. However, rhizobia inoculants differ, and some are more effective at fixing N2 than others (Thilakarathna and Raizada, 2017). Certain 65 rhizobia genes are needed to establish symbioses and have been identified (Bottomley and Myrold, 2015). Unfortunately, it was not feasible to calculate separate B values for each field or measure rhizobia strains or their effectiveness in different soils. Thus, it would be valuable to understand differences in rhizobia strains at differing landscape positions, potentially justifying rhizobia inoculant for parts of fields. Second, BNF appears sensitive to changes in precipitation amount. Understanding the full relationship between BNF and rainfall requires additional experimentation with different precipitation treatments, including both amounts of precipitation and delivery patterns. Heavy rainfall events, particularly in the spring, can delay planting and lead to anerobic conditions that can limit rhizobia's ability to infect roots and thus lead to reduced nodulation (Layzell and Hunt, 1990). Another global change factor that might affect BNF is elevated concentrations of atmospheric CO2, which stimulates N2 fixation in legumes (van Groenigen et al 2006). However, elevated CO2 is affected by drought differentially since higher levels of CO2 stimulate nodulation and nodule density can help maintain N2 fixation under drought (Parvin et al., 2019). Third and finally, more experimentation is needed to understand how timing of drought influences N2 fixation, i.e. how drought effects may depend on the stages of growth and development (Zahran, 1999). There are few experiments that integrate measures of plant physiology and structure to provide an understanding of whole plant function and fixation throughout the growing season alongside changes in precipitation. However, we know BNF varies with different stages of soybean crop growth (Córdova et al., 2019) . 66 Implications Changing precipitation patterns seem likely to influence BNF in predictable ways depending on landscape position and tillage history. As hypothesized in Question 1, %BNF was highest in summit positions and lowest in toeslope positions, coincident with texture and organic matter: summits had coarser (sandier) textures than toeslopes and less organic matter. Increased precipitation diminished %BNF at summit positions, but not at toeslope positions (Question 2). In no-till plots, %BNF decreased with more rainfall events, but only in no-till fields (Question 3). Results have several implications for global assessments of N fixation and soybean N management. First, results suggest that including information on landscape position and rainfall changes in calculations of field-scale BNF rates would lead to more accurate landscape and regional contributions of BNF to Nr insofar as soybean BNF rates vary greatly by landscape position and rainfall regime. The creation of regional Nr budgets for BNF are currently performed with values from small plots. If BNF varies over a landscape, Nr created through BNF by soybeans may be greatly over or underestimated, especially in areas that are topographically diverse. Differences in %BNF by toposequence, rainfall, and tillage will influence regional and global estimates of BNF contributions to Nr. Second, results suggest that irrigation and organic matter can decrease the need for N fertilizer in soybeans. Despite meta analyses that show response to N supplementation are inconsistent (Mourtzinis et al., 2017), N fertilizer recommendations are promoted in some materials (Pioneer 2020). Results from rainfall and toposequence experiments suggest that soil organic matter is the best predictor of BNF in tilled fields (Figure 4.7). With higher organic matter and more frequent rainfall in tilled fields, there are lower BNF rates. Organic matter and 67 irrigation can supply N through mineralization, thus managing soil organic matter and irrigation can act as substitutes for N fertilizer, avoiding the environmental cost of applying fertilizer N to soybeans. Third and finally, results suggest that management on a site-specific basis would be helpful. Given increasing soybean acreage and yields in the US, in order to limit inputs, it will be practical to be efficient in inoculation, fertilizer, irrigation and organic matter management when transferring to the larger scale, whether that be a field, the Midwest, different geographical regions or globally. Understanding toposequence differences in soybean BNF provides the opportunity to manage inputs more efficiently by slope position. 68 REFERENCES 69 REFERENCES Abbasi, M.K., Adams, W.A., 2000. Gaseous N emission during simultaneous nitrification- denitrification associated with mineral N fertilization to a grassland soil under field conditions. Soil Biology and Biochemistry. doi:10.1016/S0038-0717(00)00042-0 Aronson, E.L., Allison, S.D., 2012. Meta-analysis of environmental impacts on nitrous oxide release in response to N amendment. Frontiers in Microbiology 3, 1–6. doi:10.3389/fmicb.2012.00272 Batzli, J.M., Graves, W.R., Van Berkum, P., 1992. Diversity among rhizobia effective with Robinia pseudoacacia L. Applied and Environmental Microbiology 58, 2137–2143. doi:10.1128/aem.58.7.2137-2143.1992 Beehler, J., Fry, J., Negassa, W., Kravchenko, A., 2017. Impact of cover crop on soil carbon accrual in topographically diverse terrain 72. doi:10.2489/jswc.72.3.272 Bergsma, T.T., Robertson, G.P., Ostrom, N.E., 2002. Influence of soil moisture and land use history on denitrification end-products. Journal of Environmental Quality 717, 711–717. doi:10.2134/jeq2002.0711 Bernhardt, E.S., Blaszczak, J.R., Ficken, C.D., Fork, M.L., Kaiser, K.E., Seybold, E.C., Fork, M.L., Ficken, C.D., Seybold, E.C., Bernhardt, E.S., 2017. Control points in ecosystems : moving beyond the hot spot hot moment concept. Ecosystems 20, 665–682. doi:10.1007/s10021-016-0103-y Birch, H.F., 1958. The effect of soil drying on humus decomposition and nitrogen availability. Plant and Soil 10, 9e31. http://dx.doi.org/10.1007/BF01343734. Blesh, J., Drinkwater, L.E., 2013. The impact of nitrogen source and crop rotation on nitrogen mass balances in the Mississippi River Basin. Ecological Applications 23, 1017–1035. doi:10.1890/12-0132.1 Böhm, W., Maduakor, H., Taylor, H.M., 1977. Comparison of five methods for characterizing soybean rooting density and development. Agronomy Journal 69, 415–419. doi:10.2134/agronj1977.00021962006900030021x Bottomley, P.J., Myrold, D.D., 2015. Chapter 15 - Biological N Inputs, in: Soil Microbiology, Ecology and Biochemistry. Elsevier Inc., pp. 365–387. doi:10.1016/B978-0-08-047514- 1.50018-4 Bowles, T.M., Atallah, S.S., Campbell, E.E., Gaudin, A.C.M., Wieder, W.R., Grandy, A.S., 2018. Addressing agricultural nitrogen losses in a changing climate. Nature Sustainability. doi:10.1038/s41893-018-0106-0 70 Bremner, A.J.M., Blackmer, A.M., 1978. Nitrous oxide : emission from soils during nitrification of fertilizer nitrogen. Science 199, 295–296. doi: 10.1126/science.199.4326.295 Burger, M., Jackson, L.E., Lundquist, E.J., Louie, D.T., Miller, R.L., Rolston, D.E., Scow, K.M., 2005. Microbial responses and nitrous oxide emissions during wetting and drying of organically and conventionally managed soil under tomatoes. Biology and Fertility of Soils 42, 109–118. doi:10.1007/s00374-005-0007-z Castellano, M.J., Schmidt, J.P., Kaye, J.P., Walker, C., Graham, C.B., Lin, H., Dell, C.J., 2010. Hydrological and biogeochemical controls on the timing and magnitude of nitrous oxide flux across an agricultural landscape. Global Change Biology 16, 2711–2720. doi:10.1111/j.1365-2486.2009.02116.x Cavigelli, M.A., Robertson, G.P., 2001. Role of denitrifier diversity in rates of nitrous oxide consumption in a terrestrial ecosystem. Soil Biology and Biochemistry 33, 297–310. doi:10.1016/S0038-0717(00)00141-3 Christensen, S., Prieme, A., 2001. Natural perturbations , drying-wetting and freezing-thawing cycles , and the emission of nitrous oxide , carbon dioxide and methane from farmed organic soils. Soil Biology and Biochemistry 33, 2083–2091. doi:10.1016/S0038- 0717(01)00140-7 Ciampitti, I.A., Salvagiotti, F., 2018. New insights into soybean biological nitrogen fixation. Agronomy Journal 110, 1185–1196. doi:10.2134/agronj2017.06.0348 Congreves, K.A., Wagner-Riddle, C., Si, B.C., Clough, T.J., 2018. Nitrous oxide emissions and biogeochemical responses to soil freezing-thawing and drying-wetting. Soil Biology and Biochemistry. doi:10.1016/j.soilbio.2017.10.040 Cooper, G.S., Smith, R.L., 1963. Sequence of products formed during denitrification in some diverse western soils. Soil Science Society of America Journal 659–662. Córdova, S.C., Castellano, M.J., Dietzel, R., Licht, M.A., Togliatti, K., Martinez-Feria, R., Archontoulis, S. V, 2019. Soybean nitrogen fixation dynamics in Iowa, USA. Field Crops Research 236, 165–176. doi:10.1016/j.fcr.2019.03.018 Crum, J. R. and H. P. Collins. 1995. KBS Soils. Kellogg Biological Station Long-term Ecological Research Special Publication. Zenodo, doi.org/10.5281/zenodo.2581504.. Crum, J.R., Robertson, G.P., Nurenberger, F., 1990. Long-term climate trends and agricultural productivity in Southwestern Michigan., Climate variability and ecosystem response. U.S. Department of Agriculture, U.S. Forest Service, Southeatern Forest Experiment Station, Asheville, North Carolina, USA. Evans, S.E., Wallenstein, M.D., 2012. Soil microbial community response to drying and rewetting stress: does historical precipitation regime matter? Biogeochemistry 109, 101– 116. doi: 10.1016/j.soilbio.2015.10.020 71 Fierer, N., Schimel, J.P., 2002. Effects of drying-rewetting frequency on soil carbon and nitrogen transformations. Soil Biology and Biochemistry 34, 777–787. doi:10.1016/S0038- 0717(02)00007-X Florinsky, I. V, McMahon, S., Burton, D.L., 2009. Topographic factors of nitrous oxide emission, in: Nitrous Oxide Emissions Research Progress. Nitrous Oxide Emissions Research Progress, pp. 105–126. Fowler, D., Pilegaard, K., Sutton, M.A., Ambus, P., Raivonen, M., Duyzer, J., Simpson, D., Fagerli, H., Fuzzi, S., Schjoerring, J.K., Granier, C., Neftel, A., Isaksen, I.S.A., Laj, P., Maione, M., Monks, P.S., Burkhardt, J., Daemmgen, U., Neirynck, J., Personne, E., Wichink-Kruit, R., Butterbach-Bahl, K., Flechard, C., Tuovinen, J.P., Coyle, M., Gerosa, G., Loubet, B., Altimir, N., Gruenhage, L., Ammann, C., Cieslik, S., Paoletti, E., Mikkelsen, T.N., Ro-Poulsen, H., Cellier, P., Cape, J.N., Horváth, L., Loreto, F., Niinemets, Ü., Palmer, P.I., Rinne, J., Misztal, P., Nemitz, E., Nilsson, D., Pryor, S., Gallagher, M.W., Vesala, T., Skiba, U., Brüggemann, N., Zechmeister-Boltenstern, S., Williams, J., O’Dowd, C., Facchini, M.C., de Leeuw, G., Flossman, A., Chaumerliac, N., Erisman, J.W., 2009. Atmospheric composition change: Ecosystems-Atmosphere interactions. Atmospheric Environment. doi:10.1016/j.atmosenv.2009.07.068 Frankson, R., Kunkel, K.E., 2017. Michigan State Climate Summary. Franzluebbers, A.J., Haney, R.L., Honeycutt, C.W., Schomberg, H.H., Hons, F.M., 2000. Flush of carbon dioxide following rewetting of dried soil relates to active organic pools. Soil Science Society of America Journal 64, 613–623. doi:10.2136/sssaj2000.642613x Gan, Y., Stulen, I., Van Keulen, H., Kuiper, P.J.C., 2003. Effect of N fertilizer top-dressing at various reproductive stages on growth, N2 fixation and yield of three soybean (Glycine max (L.) Merr.) genotypes. Field Crops Research 80, 147–155. doi:10.1016/S0378- 4290(02)00171-5 Gelfand, I., Cui, M., Tang, J., Robertson, G.P., 2015. Short-term drought response of N2O and CO2 emissions from mesic agricultural soils in the US Midwest. Agriculture, Ecosystems & Environment 212, 127–133. doi:10.1016/j.agee.2015.07.005 Gelfand, I., Hamilton, S.K., Kravchenko, A.N., Jackson, R.D., Thelen, K.D., Robertson, G.P., 2020. Empirical evidence for the potential climate benefits of decarbonizing light vehicle transport in the U.S. with bioenergy from purpose-grown biomass with and without BECCS. Environmental Science and Technology 54, 2961–2974. doi:10.1021/acs.est.9b07019 Gelfand, I., Robertson, G.P., 2015. A reassessment of the contribution of soybean biological nitrogen fixation to reactive N in the environment. Biogeochemistry 123, 175–184. doi:10.1007/s10533-014-0061-4 Gelfand, I., Shcherbak, I., Millar, N., Kravchenko, A.N., Robertson, G.P., 2016. Long-term nitrous oxide fluxes in annual and perennial agricultural and unmanaged ecosystems in the upper Midwest USA. Global Change Biology 1–14. doi:10.1111/gcb.13426 72 George, T., Singleton, P.W., van Kessel, C., 1993. The use of nitrogen-15 natural abundance and nitrogen yield of non-nodulating isolines to estimate nitrogen fixation by soybeans (Glycine max L.) across three elevations. Biology and Fertility of Soils 15, 81–86. doi:10.1007/BF00336422 Gray, S.B., Strellner, R.S., Puthuval, K.K., Ng, C., Shulman, R.E., Siebers, M.H., Rogers, A., Leakey, A.D.B., 2013. Minirhizotron imaging reveals that nodulation of field-grown soybean is enhanced by free-air CO2 enrichment only when combined with drought stress. Functional Plant Biology 40, 137–147. doi:10.1071/FP12044 Greaver, T.L., Clark, C.M., Compton, J.E., Vallano, D., Talhelm, A.F., Weaver, C.P., Band, L.E., Baron, J.S., Davidson, E.A., Tague, C.L., Felker-Quinn, E., Lynch, J.A., Herrick, J.D., Liu, L., Goodale, C.L., Novak, K.J., Haeuber, R.A., 2016. Key ecological responses to nitrogen are altered by climate change. doi:10.1038/NCLIMATE3088 Groffman, P.M., Altabet, M.A., Böhlke, J.K., Butterbach-bahl, K., David, M.B., Firestone, M.K., Giblin, A.E., Kana, T.M., Nielsen, L.P., Voytek, M.A., Groffman, P.M., Altabet, M.A., Bohlke, J.K., Butterbach-bahl, K., David, M.B., Firestone, M.K., Giblin, A.E., Kana, T.M., Nielsen, L.P., Voytek, M.A., 2006. Methods for measuring denitrification: diverse approaches to a difficult problem. Ecological Applications 16, 2091–2122. Groffman, P.M., Holland, E.A., Myrold, D.D., Robertson, G.P., Zou, X., 1999. Denitri- fication. in Robertson, G.P., Coleman, D.C., Bledsoe, C.S., Sollins, P. (Eds.), Standard Soil Methods for Long-Term Ecological Research. Oxford University Press, Oxford, UK, pp. 272e288. Groffman, P.M., Rice, C.W., Tiedje, J.M., 1993. Denitrification in a tallgrass prairie landscape. Ecology 74, 855–862. doi:10.2307/1940811 Groffman, P.M., Tiedje, J.M., 1988. Denitrification hysteresis during wetting and drying cycles in soil. Soil Science Society of America Journal 52, 1626. doi:10.2136/sssaj1988.03615995005200060022x Grossman, J.M., Schipanski, M.E., Sooksanguan, T., Seehaver, S., Drinkwater, L.E., 2011. Diversity of rhizobia in soybean [Glycine max (Vinton)] nodules varies under organic and conventional management. Applied Soil Ecology 50, 14–20. doi:10.1016/j.apsoil.2011.08.003 Hamilton, S. K. 2015. Water quality and movement in agricultural landscapes. Pages 275-309 in S. K. Hamilton, J. E. Doll, and G. P. Robertson, editors. The Ecology of Agricultural Landscapes: Long-Term Research on the Path to Sustainability. Oxford University Press, New York, New York, USA. WaterHargreaves, S.K., Hofmockel, K.S., 2014. Physiological shifts in the microbial community drive changes in enzyme activity in a perennial agroecosystem. Biogeochemistry 117, 67– 79. doi:10.1007/s10533-013-9893-6 73 Hauggaard-Nielsen, H., Holdensen, L., Wulfsohn, D., Jensen, E.S., 2010. Spatial variation of N2-fixation in field pea (Pisum sativum L.) at the field scale determined by the 15N natural abundance method. Plant and Soil 327, 167–184. doi:10.1007/s11104-009-0043-9 Hawkes, C. V., Wren, I.F., Herman, D.J., Firestone, M.K., 2005. Plant invasion alters nitrogen cycling by modifying the soil nitrifying community. Ecology Letters 8, 976–985. doi:10.1111/j.1461-0248.2005.00802.x Herridge, D.F., Peoples, M.B., Boddey, R.M., 2008. Global inputs of biological nitrogen fixation in agricultural systems. Plant and Soil. doi:10.1007/s11104-008-9668-3 Holland, E. A., G. P. Robertson, J. Greenberg, P. M. Groffman, R. D. Boone, and J. R. Gosz. 1999. Soil CO2, N2O, and CH4 exchange. Pages 185-201 in G. P. Robertson, D. C. Coleman, C. S. Bledsoe, and P. Sollins, editors. Standard Soil Methods for Long-term Ecological Research. Oxford University Press, New York, New York, USA. Houlton, B.Z., Boyer, E., Finzi, A., Galloway, J., Leach, A., Liptzin, D., Melillo, J., Rosenstock, T.S., Sobota, D., Townsend, A.R., 2013. Intentional versus unintentional nitrogen use in the United States: Trends, efficiency and implications. Biogeochemistry 114, 11–23. doi:10.1007/s10533-012-9801-5 Intergovernmental Panel on Climate Change (IPCC). Climate Change 2007: The Physical Science Basis; Cambridge University Press: Cambridge, U.K., 2007. (84) Intergovernmental Panel on Climate Change (IPCC). Climate Change 2014: Mitigation of Climate Change; Cambridge University Press: Cambridge, U.K., 2014. Izaurralde, R.C., Lemke, R.L., Goddard, T.W., McConkey, B., Zhang, Z., 2004. Nitrous oxide emissions from agricultural toposequences in Alberta and Saskatchewan. Soil Sci. Soc. Am. J. 1285–1294. Kaleita, A.L., Schott, L.R., Hargreaves, S.K., Hofmockel, K.S., 2017. Differences in soil biological activity by terrain types at the sub-field scale in central Iowa US. PLoS ONE 12, 1–13. doi:10.1371/journal.pone.0180596 Keisling, T.C., Batchelor, J.T., Porter, O.A., 1995. Soybean root morphology in soils with and without tillage pans in the lower mississippi river valley. Journal of Plant Nutrition 18, 373–384. doi:10.1080/01904169509364910 Khalil, K., Mary, B., Renault, P., 2004. Nitrous oxide production by nitrification and denitrification in soil aggregates as affected by O2 concentration. Soil Biology and Biochemistry 36, 687–699. doi:10.1016/j.soilbio.2004.01.004 Kiers, E.T., West, S.A., Denison, R.F., 2002. Mediating mutualisms: farm management practices and evolutionary changes in symbiont co-operation. Journal of Applied Ecology 39, 745– 754. doi:10.1046/j.1365-2664.2002.00755.x 74 Kim, D.G., Vargas, R., Bond-Lamberty, B., Turetsky, M.R., 2012. Effects of soil rewetting and thawing on soil gas fluxes: a review of current literature and suggestions for future research. Biogeosciences 9, 2459–2483. doi:10.5194/bg-9-2459-2012 Kravchenko, a. N., Robertson, G.P., 2015. Statistical challenges in analyses of chamber-based soil CO2 and N2O emissions data. Soil Science Society of America Journal 79, 200. doi:10.2136/sssaj2014.08.0325 Kravchenko A.N., Snapp, S.S., Robertson, G.P., 2017. Field-scale experiments reveal persistent yield gaps in low-input and organic cropping systems. Proceedings of the National Academy of Sciences 114, 926–931. doi:10.1073/pnas.1612311114 Kravchenko, A N, Toosi, E.R., Guber, A.K., Ostrom, N.E., Yu, J., Azeem, K., Rivers, M.L., Robertson, G.P., 2017. Hotspots of soil N2O emission enhanced through water absorption by plant residue. Nature Geoscience 10, 496–500. doi:10.1038/ngeo2963 Krichels, A.N., DeLucia, E.H., Sanford, R., Chee-Sanford, J., Yang, W.H., 2019. Historical soil drainage mediates the response of soil greenhouse gas emissions to intense precipitation events. Biogeochemistry 142, 425–442. doi:10.1007/s10533-019-00544-x Krichels, A.H., Sipic, E., Yang, W.H., 2019. Iron redox reactions can drive microtopographic variation in upland soil carbon dioxide and nitrous oxide emissions. Soil Systems 3, 60. doi:10.3390/soilsystems3030060 Ladoni, M., Basir, A., Robertson, P.G., Kravchenko, A.N., 2016. Scaling-up: Cover crops differentially influence soil carbon in agricultural fields with diverse topography. Agriculture, Ecosystems and Environment 225, 93–103. doi:10.1016/j.agee.2016.03.021 Layzell, D.B., Hunt, S., 1990. Oxygen and the regulation of nitrogen fixation in legume nodules. Physiologia Plantarum 80, 322–327. doi:10.1111/j.1399-3054.1990.tb04414.x Li, X., McCarty, G.W., Lang, M., Ducey, T., Hunt, P., Miller, J., 2018. Topographic and physicochemical controls on soil denitrification in prior converted croplands located on the Delmarva Peninsula, USA. Geoderma 309, 41–49. doi:10.1016/j.geoderma.2017.09.003 Liang, C., Jesus, E. da C., Duncan, D.S., Jackson, R.D., Tiedje, J.M., Balser, T.C., 2012. Soil microbial communities under model biofuel cropping systems in southern Wisconsin, USA: impact of crop species and soil properties. Applied Soil Ecology 54, 24–31. doi:10.1016/j.apsoil.2011.11.015 Liang, L.L., Grantz, D.A., Jenerette, G.D., 2016. Multivariate regulation of soil CO2 and N2O pulse emissions from agricultural soils. Global Change Biology 22, 1286–1298. doi:10.1111/gcb.13130 Loecke, T.D., Robertson, G.P., Kellogg, W.K., 2008. Soil resource heterogeneity in terms of litter aggregation promotes nitrous oxide fluxes and slows decomposition. Soil Biology and Biochemistry 41, 228–235. doi:10.1016/j.soilbio.2008.10.017 75 Luehmann, M.D., Peter, B.G., Connallon, C.B., Schaetzl, R.J., Smidt, S.J., Liu, W., Kincare, K.A., Walkowiak, T.A., Thorlund, E., Holler, M.S., Schaetzl, J., 2016. Loamy, two-storied soils on the outwash plains of Southwestern Lower Michigan: pedoturbation of loess with the underlying sand. Annals of the American Association of Geographers 106, 551–572. doi:10.1080/00045608.2015.1115388 McDaniel, M.D., Kaye, J.P., Kaye, M.W., 2014. Do “hot moments” become hotter under climate change? Soil nitrogen dynamics from a climate manipulation experiment in a post-harvest forest. Biogeochemistry 121, 339–354. doi:10.1007/s10533-014-0001-3 Mcswiney, C.P., Mcdowell, W.H., Keller, M., 2001. Distribution of nitrous oxide and regulators of its production across a tropical rainforest catena in the Luquillo Experimental Forest, Puerto Rico. Biogeochemistry 56, 265–286. doi:10.1023/A:1013191010244 Millar, N., Robertson, G.P., 2015. Nitrogen transfers and transformations in row-crop ecosystems, in: The Ecology of Agricultural Landscapes: Long-Term Research on the Path to Sustainability. pp. 213–251. Mourtzinis, S., Kaur, G., Orlowski, J.M., Shapiro, C.A., Lee, C.D., Wortmann, C., Holshouser, D., Nafziger, E.D., Kandel, H., Niekamp, J., Ross, W.J., Lofton, J., Vonk, J., Roozeboom, K.L., Thelen, K.D., Lindsey, L.E., Staton, M., Naeve, S.L., Casteel, S.N., Wiebold, W.J., Conley, S.P., 2017. Soybean response to nitrogen application across the United States: A synthesis-analysis. doi:10.1016/j.fcr.2017.09.035 Myers, D.B., Kitchen, N.R., Sudduth, K.A., Sharp, R.E., Miles, R.J., 2007. Soybean root distribution related to claypan soil properties and apparent soil electrical conductivity. Crop Science 47, 1498–1509. doi:10.2135/cropsci2006.07.0460 Natelhoffer, K.J., Fry, B., 1988. Controls on natural nitrogen-15 and carbon-13 abundances in forest soil organic matter. Soil Science of America Journal 52, 1633–1640. doi:10.2136/sssaj1988.03615995005200060024x Negassa, W., Price, R.F., Basir, A., Snapp, S.S., Kravchenko, A., 2015. Cover crop and tillage systems effect on soil CO2 and N2O fluxes in contrasting topographic positions. Soil and Tillage Research 154, 64–74. doi:10.1016/j.still.2015.06.015 Oates, L.G., Duncan, D.S., Gelfand, I., Millar, N., Robertson, G.P., Jackson, R.D., 2016. Nitrous oxide emissions during establishment of eight alternative cellulosic bioenergy cropping systems in the North Central United States. GCB Bioenergy 8, 539–549. doi:10.1111/gcbb.12268 Oberson, A., Bosshard, C., Nanzer, S., Maeder, P., Frossard, E., Dubois, D., 2007. Symbiotic N2 fixation by soybean in organic and conventional cropping systems estimated by 15N dilution and 15N natural abundance. Plant and Soil 290, 69–83. doi:10.1007/s11104-006- 9122-3 Ordóñez, R.A., Castellano, M.J., Hatfield, J.L., Helmers, M.J., Licht, M.A., Liebman, M., Dietzel, R., Martinez-Feria, R., Iqbal, J., Puntel, L.A., Córdova, S.C., Togliatti, K., Wright, 76 E.E., Archontoulis, S. V, 2018. Maize and soybean root front velocity and maximum depth in Iowa, USA. Field Crops Research 215, 122–131. doi:10.1016/j.fcr.2017.09.003 Parvin, S., Uddin, S., Tausz-Posch, S., Fitzgerald, G., Armstrong, R., Tausz, M., 2019. Elevated CO2 improves yield and N2 fixation but not grain N concentration of faba bean (Vicia faba L.) subjected to terminal drought. Environmental and Experimental Botany 165, 161–173. doi:10.1016/j.envexpbot.2019.06.003 Pearson, C.J., 2007. Regenerative, semiclosed systems: a priority for twenty-first-century agriculture. BioScience 57, 409–418. doi:10.1641/b570506 Peralta, A.L., Ludmer, S., Kent, A.D., 2013. Hydrologic history influences microbial community composition and nitrogen cycling under experimental drying/wetting treatments. doi:10.1016/j.soilbio.2013.06.019 Peters, D.P.C., Laney, C.M., Lugo, A.E., Collins, S.L., Driscoll, C.T., Grove, J.M., Knapp, A.K., Kratz, T.K., Ohman, M.D., Waide, R.B., Yao, J., 2013. Long-term trends in ecological systems: a basis for understanding responses to global change. Agricultural Research Service Technical Bulletin 378. Pryor, S., Scavia, D., Downer, C., Gaden, M., Iverson, L., Nordstrom, R., Patz, J., Roberson, G., 2014. Midwest. Climate Change Impacts in the United States: The Third National Climate Assessment. 418–440. doi:10.7930/J0J1012N.On Pueppke, S.G., 1986. Nodule distribution on legume roots: Specificity and response to the presence of soil. Soil Biology and Biochemistry 18, 601–606. doi:10.1016/0038- 0717(86)90082-9 Ravishankara, A.R.., Daniel, J.S.., Portmann, R.W., 2009. Nitrous oxide (N2O): the dominant ozone-depleting substance emitted in the 21st Century. Science 326, 123–125. doi:10.1126/science.1176985 Riffkin, Peoples, Kearney, Quigley, Cameron, Gault;, Thies, 1999. Factors associated with biological nitrogen fixation in dairy pastures in south- western Victoria. Australian Journal of Plant Physiology 27, 645. doi:10.1071/PP97167 Robertson, G.P., Bruulsema, T.W., Gehl, R.J., Kanter, D., Mauzerall, D.L., Rotz, C.A., Williams, C.O., 2013. Nitrogen-climate interactions in US agriculture. Biogeochemistry 114, 41–70. doi:10.1007/s10533-012-9802-4 Robertson, G Philip, Groffman, P., 2015. Nitrogen transformations, 4th ed, Soil Microbiology Biochemistry and Ecology. Elsevier Inc. doi:10.1016/B978-0-12-415955-6.00014-1 Robertson, G P, Groffman, P.M., 2015. Chapter 14 - Nitrogen Transformations, in: Soil Microbiology Ecology and Biochemistry. Elsevier Inc., pp. 421–446. doi:10.1016/B978-0- 12-415955-6.00014-1 77 Robertson, G. P. and S. K. Hamilton. 2015. Long-term ecological research in agricultural landscapes at the Kellogg Biological Station LTER site: conceptual and experimental framework. Pages 1-32 in S. K. Hamilton, J. E. Doll, and G. P. Robertson, editors. The Ecology of Agricultural Landscapes: Long-Term Research on the Path to Sustainability. Oxford University Press, New York, New York, USA Robertson, G.P., Vitousek, P.M., 2009. Nitrogen in Agriculture: Balancing the Cost of an Essential Resource. Annual Review of Environment and Resources 34, 97–125. doi:10.1146/annurev.environ.032108.105046 Robertson, W.K., Hammond, L.C., Johnson, J.T., Boote, K.J., 1980. Effects of Plant-Water Stress on Root Distribution of Corn, Soybeans, and Peanuts in Sandy Soil. Agronomy Journal 72, 548. doi:10.2134/agronj1980.00021962007200030033x Rockström, J., W. Steffen, K. Noone, Å. Persson, F. S. Chapin, E. F. Lambin, T. M. Lenton, M. Scheffer, C. Folke, H. J. Schellnhuber, B. Nykvist, C. A. de Wit, T. Hughes, S. van der Leeuw, H. Rodhe, S. Sörlin, P. K. Snyder, R. Costanza, U. Svedin, M. Falkenmark, L. Karlberg, R. W. Corell, V. J. Fabry, J. Hansen, B. Walker, D. Liverman, K. Richardson, P. Crutzen, J. A. Foley, 2009. A safe operation space for humanity. Nature 461, 472–475. Saha, D., Rau, B.M., Kaye, J.P., Montes, F., Adler, P.R., Kemanian, A.R., 2017. Landscape control of nitrous oxide emissions during the transition from conservation reserve program to perennial grasses for bioenergy. GCB Bioenergy 9, 783–795. doi:10.1111/gcbb.12395 Salvagiotti, F., Cassman, K.G., Specht, J.E., Walters, D.T., Weiss, a., Dobermann, a., 2008. Nitrogen uptake, fixation and response to fertilizer N in soybeans: a review. Field Crops Research 108, 1–13. doi:10.1016/j.fcr.2008.03.001 Sanford, G.R., Oates, L.G., Jasrotia, P., Thelen, K.D., Robertson, G.P., Jackson, R.D., 2016. Comparative productivity of alternative cellulosic bioenergy cropping systems in the North Central USA. Agriculture, Ecosystems and Environment 216, 344–355. doi:10.1016/j.agee.2015.10.018 Schipanski, M.E., Drinkwater, L.E., Russelle, M.P., 2010. Understanding the variability in soybean nitrogen fixation across agroecosystems. Plant and Soil 329, 379–397. doi:10.1007/s11104-009-0165-0 Schjønning, P., Thomsen, I.K., Moldrup, P., Christensen, B.T., 2003. Linking soil microbial activity to water- and air-phase contents and diffusivities. Soil Science Society of America Journal 67, 156–165. doi:10.2136/sssaj2003.1560 Serraj, R., Sinclair, T.R., Purcell, L.C., 1999. Symbiotic N2 fixation response to drought. Journal of Experimental Botany. doi:10.1093/jxb/50.331.143 Sexstone, A.J., Parkin, T.B., Tiedje, J.M., 1988. Denitrification response to soil wetting in aggregated and unaggregated soil. Soil Biology and Biochemistry 20, 767–769. doi:10.1016/0038-0717(88)90165-4 78 Shcherbak, I., Millar, N., Robertson, G.P., 2014. Global metaanalysis of the nonlinear response of soil nitrous oxide (N2O) emissions to fertilizer nitrogen. Proceedings of the National Academy of Sciences 111, 9199–9204. doi:10.1073/pnas.1322434111 Six, J., Paustian, K., Elliott, E.T., Combrink, C., 2000. Soil structure and organic matter I. distribution of aggregate-size classes and aggregate-associated carbon. Soil Science Society of America Journal 64, 681–689. doi:10.2136/sssaj2000.642681x Syswerda, S.P., Basso, B., Hamilton, S.K., Tausig, J.B., Robertson, G.P., 2012. Long-term nitrate loss along an agricultural intensity gradient in the Upper Midwest USA. Agriculture, Ecosystems and Environment 149, 10–19. doi:10.1016/j.agee.2011.12.007 Syväsalo, E., Regina, K., Pihlatie, M., Esala, M., 2004. Emissions of nitrous oxide from boreal agricultural clay and loamy sand soils. Nutrient Cycling in Agroecosystems 69, 155–165. doi:10.1023/B:FRES.0000029675.24465.fc Tamagno, S., Sadras, V.O., Haegele, J.W., Armstrong, P.R., Ciampitti, I.A., 2018. Interplay between nitrogen fertilizer and biological nitrogen fixation in soybean: implications on seed yield and biomass allocation. Scientific Reports 8, 17502. doi:10.1038/s41598-018- 35672-1 Thilakarathna, M.S., Raizada, M.N., 2017. A meta-analysis of the effectiveness of diverse rhizobia inoculants on soybean traits under field conditions. Soil Biology and Biochemistry 105, 177–196. doi:10.1016/j.soilbio.2016.11.022 Unkovich, M., Herridge, D., Peoples, M., 2008. Measuring plant-associated nitrogen fixation in agricultural systems. Australian Centre for International Agricultural Research 258. doi:citeulike-article-id:9530718 Valentine, A.J., Benedito, V.A., Kang, Y., 2010. Legume Nitrogen Fixation and Soil Abiotic Stress: From Physiology to Genomics and Beyond, in: Nitrogen Metabolism in Plants in the Post-Genomic Era. pp. 207–248. doi:10.1002/9781444328608.ch9 Vitousek, P.M., Menge, D.N.L., Reed, S.C., Cleveland, C.C., 2013. Biological nitrogen fixation: rates, patterns, and ecological controls in terrestrial ecosystems. Philisophical Transactions of the Royal Society of London, B Biological Sciences 368, 20130119. Walsh, K.B., 1995. Physiology of the legume nodule and its response to stress. Soil Biology and Biochemistry 27, 637–655. doi:10.1016/0038-0717(95)98644-4 Weese, D.J., Heath, K.D., Dentinger, B.T.M., Lau, J. a., 2015. Long-term nitrogen addition causes the evolution of less-cooperative mutualists. Evolution 69, 631-642. doi:10.1111/evo.12594 Wickings, K., Grandy, A.S., Kravchenko, A.N., 2016. Going with the flow: Landscape position drives differences in microbial biomass and activity in conventional, low input, and organic agricultural systems in the Midwestern U.S. Agriculture, Ecosystems and Environment 218, 1–10. doi:10.1016/j.agee.2015.11.005 79 Workneh, F., Yang, X. B., and Tylka, G. L. 1999. Soybean brown stem rot, Phytophthora sojae, and Heterodera glycines affected by soil texture and tillage relations. Phytopathology 89:844-850. Wrage-Mönnig, N., Horn, M.A., Well, R., Müller, C., Velthof, G., Oenema, O., 2018. The role of nitrifier denitrification in the production of nitrous oxide revisited. Soil Biology and Biochemistry. doi:10.1016/j.soilbio.2018.03.020 Wrage, N., Velthof, G.L., Beusichem, M.L. Van, Oenema, O., 2001. Role of nitrifier denitrication in the production of nitrous oxide. Soil Biology and Biochemistry 33, 1723– 1732. Zahran, H.H., 1999. Rhizobium-legume symbiosis and nitrogen fixation under severe conditions and in an arid climate. Microbiology and Molecular Biology Reviews 63, 968–989. doi:10.1128/mmbr.63.4.968-989.1999 Zeglin, L.H., Bottomley, P.J., Jumpponen, A., Rice, C.W., Arango, M., Lindsley, A., McGowan, A., Mfombep, P., Myrold, D.D., 2013. Altered precipitation regime affects the function and composition of soil microbial communities on multiple time scales. Ecology 94, 2334– 2345. doi:10.1890/12-2018.1 Zhu-Barker, X., Cavazos, A.R., Ostrom, N.E., Horwath, W.R., Glass, J.B., 2015. The importance of abiotic reactions for nitrous oxide production. Biogeochemistry 126, 251–267. doi:10.1007/s10533-015-0166-4 Zhu, X., Burger, M., Doane, T.A., Horwath, W.R., 2013. Ammonia oxidation pathways and nitrifier denitrification are significant sources of N2O and NO under low oxygen availability. Proceedings of the National Academy of Sciences. 110, 6328–6333. doi: 10.1073/pnas.1219993110 80