Potential regional productivity and greenhouse gas emissions of fertilized and irrigated switchgrass in a Mediterranean climate

Agriculture, Ecosystems and Environment 212 (2015) 64–74

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment journal homepage: www.elsevier.com/locate/agee

Potential regional productivity and greenhouse gas emissions of fertilized and irrigated switchgrass in a Mediterranean climate Juhwan Leea,* , Gabriel Pedrosob , Chris van Kesselb , Johan Sixa a b

Department of Environmental Systems Science, Swiss Federal Institute of Technology, ETH-Zurich, Zurich 8092, Switzerland Department of Plant Sciences, University of California, Davis, CA 95616, USA

A R T I C L E I N F O

A B S T R A C T

Article history: Received 27 January 2015 Received in revised form 20 May 2015 Accepted 23 June 2015 Available online 13 July 2015

The potential of switchgrass (Panicum virgatum L.) to offset large-scale greenhouse gas (GHG) emissions depends on optimizing external inputs when the crop is primarily managed as a sustainable source for renewable energy production. Due to the heterogeneity of climate and soil conditions and the complexity of agriculture, an evaluation of the effect of adopting switchgrass as a new biofuel crop into agriculture needs to be done at the regional scale. The objective of the study was to predict long-term (100-yr) GHG emissions under different N fertilization (0, 112, and 224 kg N ha
1) and irrigation application (0, 25, 50, 75, and 99 cm H2O) levels across the Central Valley of California using the DAYCENT model. Six cultivars (Alamo, Kanlow, Cave-in-Rock, Blackwell, Sunburst, and Trailblazer) were selected. The model results suggest that switchgrass productivity is primarily constrained by N inputs when no or low water stress is expected in a Mediterranean climate. In the short-term (the first decade after establishment), soil organic carbon (SOC) stocks (0–20 cm) increased by 0.42–0.92 Mg C ha
1 yr
1 and N2O emissions were 1.37– 2.48 kg N2O–N ha
1 yr
1 across the cultivars with baseline input rates of 224 kg N ha
1 yr
1 and 99 cm H2O. All cultivars were net CO2 sinks in the near term and the potential decreased by 0.09– 0.30 Mg C ha
1 yr
1 (15.5–52.8%) with reduced N input from baseline under varying irrigation rates. There was a reduction in N2O emissions by 47.2–61.6% by applying less N fertilizer when irrigated at rates 75 cm H2O per year over time. In general, higher-yielding cultivars (e.g., Alamo) tended to sequester more CO2 but also led to higher N2O emissions. In the near term, the use of N fertilizer and irrigation is needed for switchgrass systems to be a soil GHG sink, but for longer-term GHG mitigation strategies reducing both N fertilization and irrigation inputs is required. ã 2015 Elsevier B.V. All rights reserved.

Keywords: Managed switchgrass Greenhouse gas emissions Mediterranean climate DAYCENT

1. Introduction Cultivation of biofuel crops and biomass conversion to biofuels (e.g., ethanol and biodiesel) can reduce CO2 emissions by displacing fossil fuels and sequestering C into soil (Bransby et al., 1998; Adler et al., 2007). Biofuel crops, particularly perennials compared to annuals, can also reduce soil N losses that have the potential to affect direct and indirect N2O emissions (Bransby et al., 1998). This has led to renewed interest in biofuel crops as a renewable fuel and an important element in a portfolio of GHG mitigation technologies for the next decades (Pacala and Socolow, 2004; Sims et al., 2006). However, the feasibility of large-scale biofuel production for energy and as a GHG mitigation option is still debated. First of all, it requires large-scale land-use change for sufficient land, possibly increasing a risk for initial soil C loss and

* Corresponding author. E-mail address: [email protected] (J. Lee). http://dx.doi.org/10.1016/j.agee.2015.06.015 0167-8809/ã 2015 Elsevier B.V. All rights reserved.

poor soil fertility (Lemus and Lal, 2005; Searchinger et al., 2008). It also assumes no biophysical constraints from the availability of external inputs, such as N and water, and minimal environmental impacts (Giampietro et al., 1997). The ecological benefits and risks of candidate biofuel crops have been previously found to vary with plant characteristics, nutrient demand, soil C inputs, N retention, and biomass quality (Anderson-Teixeira et al., 2009). Therefore, large-scale assessment of biofuel crops and management effects is a prerequisite to determine environmental and socio-economic effects of converting biomass to energy when adopting a new biofuel crop into agriculture. In the USA, switchgrass (Panicum virgatum L.) has been extensively evaluated as one of promising biofuel crops (McLaughlin and Walsh, 1998; Wullschleger et al., 2010). Switchgrass is a perennial C4 grass that is successfully grown across environmental conditions and a diverse range of geographical regions of North America. Switchgrass presents several agronomic advantages. For example, switchgrass is high-yielding, characterized by high nutrient and water-use efficiency as well as broad tolerance to

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

disturbance compared to other perennial herbaceous grasses (Lewandowski et al., 2003; Wright, 2007), which can be cultivarspecific (Wullschleger et al., 2010). To reduce C and other soil emissions, Lal (2005) recommended biofuel crops that produce a minimum of 10–15 Mg ha
1 yr
1 of biomass when 30–40% of crop residues is assumed to be removed annually. The characteristics of switchgrass generally meet many important selection criteria for producing energy. However, switchgrass is nonnative to the Pacific Coast in the USA, including California, and has never been grown commercially in this region. California, and specifically the Central Valley, is one of the most productive agricultural regions in the world and leads national production and sales of many crop commodities, such as almonds, cotton, grapes, hay, rice, and tomatoes. Presumably, these high-value commodities are less likely to be replaced with switchgrass. Barney and DiTomaso (2008) reported that switchgrass has a highly invasive potential in the region and therefore further evaluation of invasive characters is necessary under various environmental conditions. Clearly, it is relevant to determine if switchgrass adoption in California induces direct land use changes and is environmentally sound over time. Switchgrass is considered an option to restore some of the soil C previously lost by conventional agricultural production (Mensah et al., 2003; Skinner and Adler, 2010). Compared to annual crops, the larger root system of switchgrass significantly increases potential belowground C input, although increased C loss would be expected by soil respiration (Al-Kaisi and Grote, 2007). The potential to mitigate CO2 emissions depends on how it is managed. Established switchgrass is generally suitable for the recovery of N from fertilizer or other sources via its existing root system, and direct N2O emissions under switchgrass tend to vary under different N fertilizer rates (Nikièma et al., 2011). Sufficient N inputs are still required to sustain switchgrass productivity and soil N balance (Boehmel et al., 2008; Monti et al., 2012). Therefore, N fertilization management is an important factor for growing switchgrass. In addition, application of inorganic fertilizer or manure is an important consideration if switchgrass is grown for C sequestration (Lee et al., 2007; Liebig et al., 2008). However, optimizing N rates is difficult because there are large uncertainties in the extent of trade-offs due to variation in environmental conditions. California switchgrass cultivar trials were established at four sites (El Centro, Five Points, Davis, and Tulelake) in 2007 and total 11 different cultivars have been evaluated under different management practices until 2010 (Pedroso et al., 2011, 2013). The field trials data provide empirical data on switchgrass growth across distinct environmental and management conditions of California (Fig. 1). However, the field data accounted for limited combinations of climate, land use, and management systems. The potential for regional and longer-term switchgrass productivity and effects on the environment remain largely unknown under California conditions. So far, Pedroso et al. (2014b) is the only field study, which was conducted in a Mediterranean climate and reported the response of switchgrass to N fertilization and irrigation interactions. No measurements of soil C changes or emission of N2O and CH4 have been made for switchgrass systems in California. External N fertilization is required especially for establishing switchgrass due to poor soil fertility. In the Mediterranean regions of California, irrigation demand is also expected to be high for switchgrass management and likely affect soil C and N dynamics. It is therefore important to assess the interactive effects of N fertilization and irrigation on potential soil emissions. From a practical point of view, it is difficult to consider wide variation in management practices, soils, and microclimates to measure switchgrass productivity and soil GHG emissions. Most frequently, biogeochemical models can be used to predict regional

65

Fig. 1. Location of 537 grid cells (12 km  12 km) established within the Central Valley of California. Points show the location of California switchgrass cultivar trials.

long-term changes in plant productivity and soil quality. Previously, the DAYCENT model was calibrated and validated for GHG emissions from agricultural soils in California by considering typical crop rotations (De Gryze et al., 2010, 2011; Lee et al., 2011). Lee et al. (2012) further calibrated and validated the DAYCENT model for six selected switchgrass cultivars: Alamo (southern lowland); Kanlow (northern lowland); Blackwell and Cave-in-Rock (southern upland); Sunburst and Trailblazer (northern upland). Genetic differences in biomass yield among cultivars were simulated reasonably well across the four sites representing diverse ecoregions of California. A recent analysis suggests that switchgrass production is economically feasible and can replace other crops in some part of the Central Valley (Yi et al., 2013). The objective of this study was to predict the long-term (100-yr) effects of different inorganic N fertilizers applications and irrigation intensities on biomass yield, SOC changes, and N2O emissions of switchgrass in the Central Valley of California using the DAYCENT model. 2. Materials and methods 2.1. Model description We used the DAYCENT model (version 4.5), a fully resolved biogeochemical model that simulates the major processes associated with the dynamics of C, N, soil temperature, and water (Del Grosso et al., 2001). Key model outputs include plant growth, soil organic matter (above 20 cm depth), daily flux of N gases (N2O, NOX, and N2), and CH4 oxidation. Phenology, net primary productivity, shoot:root ratio, and the C:N ratio of biomass in plant components are species-specific and determined by soil and air temperature and soil-water stress. Soil-water availability is a function of current soil water, precipitation, irrigation water, and potential evapotranspiration. The effects of N from soil organic matter pools or fertilizer on plant growth are determined by specific grass/crop requirements. The type and timing of each management event can be specified, including tillage, fertilization, organic matter (e.g., manure) addition, harvest (with variable residue removal), drainage, irrigation, burning, and grazing intensity. If the growing degree days following submodel is

66

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

implemented, germination/beginning of growing season is a function of soil temperature and senescence is a function of accumulated growing degree days following germination or regrowth. Decomposition potential of litter and soil organic matter and nutrient supply are determined by the amount and quality of residue returned to the soil, the size of the soil organic matter pools, and temperature and water controls. Soil N2O and N2 fluxes from nitrification and denitrification are controlled by ammonium and nitrate concentrations, labile C availability, water content, temperature, pH, and texture in the soil profile (Parton et al., 1996). Crop and fixed parameters controlling plant growth, soil organic matter dynamics, and fluxes of soil trace gases were previously calibrated and validated to Mediterranean California conditions at the site and regional scale (De Gryze et al., 2010; De Gryze et al., 2011). Recently, Lee et al. (2012) further parameterized and validated the model for the selected switchgrass cultivars using published productivity data across the USA (excluding California) and data generated from four switchgrass field trials in California from 2007 to 2009 (Pedroso et al., 2011, 2013, 2014a). The calibration data were obtained from 37 field sites, where the stands had been generally maintained for up to 10 years under various environmental and management conditions. Since no measurement on soil emissions under switchgrass is available for validating the parameters, DAYCENT was validated using the biomass yield data. The model validated based on productivity has been used to reliably predict GHG emissions from soils (Duval et al., 2013). 2.2. Regional extent We assumed that switchgrass as a new biofuel crop can be grown on any prime or marginal cropland within the Central Valley of California. The simulations covered approximately 2.65 million ha of areas under agricultural land use (e.g., grain and hay crops, field crops, pasture, and others), except rice. A common grid cell (approximately 12 km  12 km) was superimposed on the Central Valley, and intersected with land use, soil and county maps, resulting in 537 grid cells (Fig. 1). 2.3. Input data We obtained daily maximum and minimum temperatures and precipitation data for 2000 until 2009 from different CIMIS stations (wwwcimis.water.ca.gov). Other weather drivers (e.g., solar radiation, wind speed, and relative humidity) were also available in the CIMIS database. The spatially explicit daily weather data for each grid cell were obtained from the nearest CIMIS station to the grid cell. Daily surface weather data for 1980 until 2003 were also obtained from the Daymet data set for the coordinates of each grid cell (Thornton et al., 2014). The Daymet data were used in the historical simulations. Estimates of soil parameters were obtained from the Soil Survey Geographic (SSURGO) database of the Natural Resources Conservation Service. Specifically, soil texture class, bulk density, hydraulic properties, and pH to the depth of 1.5 m were obtained. Hydraulic properties, such as field capacity and wilting point, were estimated using the pedo-transfer functions based on soil texture if not available (Saxton et al., 1986). The soil layer structure is required to simulate daily soil water fluxes, soil temperature distribution, and N trace gas fluxes throughout the soil profile. The land use data were obtained from the California Department of Water Resources (www.water.ca.gov), which was derived from exhaustive analyses of aerial photos and field surveys. The SSURGO database was geographically intersected with land use data within all individual grid cells.

2.4. Modeling procedures We assumed four periods of historical land use and management changes: (1) C3 temperate grasses with low-intensity grazing between years 0–1869 (Paruelo and Lauenroth, 1996), (2) initiation of cropping between years 1870–1949, (3) introduction of irrigation and inorganic fertilizer between years 1950–1969, and (4) modern agriculture from 1970 to 1999. Major crops grown were assumed to be maize, winter wheat, and tomatoes, which represents the average history of land use and management in the Central Valley of California (De Gryze et al., 2010). Thus, the same initial conditions were used for all individual grid cells within the region. Switchgrass typically has a life span of 10–15 years under cultivated conditions, but adequate management can keep stands productive permanently (Parrish and Fike, 2005). In this study, six common switchgrass cultivars were selected: Alamo (southern lowland); Kanlow (northern lowland); Blackwell and Cave-in-Rock (southern upland); Sunburst and Trailblazer (northern upland). For each cultivar, biomass yields, soil organic carbon (SOC) storage, and N2O emissions were simulated over the 100-yr simulation period under current climate and management options. Since management recommendations have not been made yet for California switchgrass, we selected annual N fertilizer rates (112 and 224 kg N ha
1) and irrigation rates (0, 25, 50, 75, and 99 cm H2O) based on the field trials (Pedroso et al., 2013). In the field trials, switchgrass stands were fertilized at 56 kg N ha
1 with one cut per year (November) in the establishment year and then 224 kg N ha
1 with two cuts per year (July and November) in the following years. Therefore, the baseline fertilizer rate was set to be 224 kg N ha
1. The baseline irrigation rate for switchgrass was assumed to be 99 cm H2O per year, approximately corresponding to 100% of its measured evapotranspiration in the Central Valley. The maximum N fertilizer and irrigation rates for switchgrass were comparable with the ones for common crops in California, according to recent Cost and Return Studies of Agricultural & Resource Economics at UC Davis (coststudies.ucdavis.edu). Due to insufficient data under California conditions, scenarios of varying timing or number of harvest or other management changes were not considered. Possible differences in management practices by cultivar or location were not considered. Total 11 switchgrass management scenarios, including a zero input scenario (with no N fertilization and irrigation input), were simulated for each cultivar. Yields were based on 95% aboveground residue removal. For all cultivars and grid cells, productive stands were assumed to be fully developed from the second year after planting. 2.5. Calculations and statistical analyses In this study, we considered partial soil components for GHG emissions: changes in SOC content (in the 0–20 cm depth) and N2O fluxes. For each individual grid cell, net soil GHG flux was calculated as: 44 GHG ¼
 DSOC þ 298  ½N2 O þ 25  ½CH4  12 where DSOC is the annual change in soil organic carbon content in Mg C ha
1 yr
1, [N2O] is the flux of N2O in Mg N2O ha
1 yr
1, and [CH4] is the flux of CH4 in Mg CH4 ha
1 yr
1. The net soil GHG flux (Mg CO2-equivalent ha
1 yr
1) was presented on a 100-yr time horizon global warming potential. The radiative forcing constants were obtained from Forster et al. (2007). It should be noted that the model is not able to simulate CH4 production and gives insufficient confidence in CH4 fluxes based on modeled oxidation capacity. Therefore, CH4 fluxes were assumed to be negative. Indirect N2O

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

emissions were not considered due to great uncertainties associated with those emissions and insufficient calibration and validation data. No database of nitrate leaching in cropping systems has been compiled for California (T. Rosenstock, personal communication). Annual yields and the emissions of CO2 and N2O were averaged, weighted by the unit area of cropland in each grid. An 11-yr moving average was then calculated to consider trends in variance over the whole simulation period, excluding the establishment year. In this study, the “near-term” and “long-term” periods represent 2–12 and 90–100 years after planting, respectively. Overall trends in annual averages were reported by cultivar or management, unless otherwise noted. The coefficient of variation (CV) around model predictions due to regional variation in input parameters was quantified. We reported the standard deviation (SD) around the modeled C and N fluxes of different grid cells that were simulated individually. The SD represented regional scale spatial variation, not modeling error. The ‘absolute’ uncertainty in the difference (Ui) by input rate change from baseline (i = N fertilizer or H2O) was also calculated:

67

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi U i ¼ SD2baseline þ SD2alternative

3. Results and discussion 3.1. Regional biomass yields In the Central Valley of California, sustainable switchgrass production will depend on resource-use efficiency as most crop production requires adequate inputs of N fertilizer and irrigation water (Kaffka, 2009). For the selected cultivars, the annual yield ranged from 1.1 to 3.0 Mg ha
1 yr
1 with no N fertilization and irrigation input (Fig. 2). With baseline input rates (224 kg N ha
1 and 99 cm H2O), Alamo and Kanlow yields were 23.9  2.2 (standard deviation) and 24.5  2.4 Mg ha
1 yr
1 in the near term and 28.6  3.7 and 28.1  3.7 Mg ha
1 yr
1 in the long term, respectively (Figs. 2 and 3). The lowland cultivars had relatively high average yields compared to the upland cultivars. Sunburst yields (16.6  1.6 and 18.9  2.2 Mg ha
1 yr
1 in the near and long

Fig. 2. Changes in switchgrass yield by N fertilizer application (224 and 112 kg N ha
1) and irrigation (0, 25, 50, 75, and 99 cm H2O).

68

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

Fig. 3. Switchgrass yield in the near and long term (2–12 years and 90–100 years after planting, respectively) as affected by N fertilizer (224 and 112 kg N ha
1) and irrigation rates (0, 25, 50, 75, and 99 cm H2O). The error bars shows standard deviation around modeled results. A = Alamo; K = Kanlow; CIR = Cave-in-Rock; B = Blackwell; S = Sunburst; T = Trailblazer.

term, respectively) showed consistently the lowest yield over the period. The CV of the baseline switchgrass yields across the cultivars were 8.6–12.0% in the near term and 10.4–14.9% in the long term. Alamo and Kanlow showed almost no differences in yield along a latitudinal gradient, while the yields of the upland cultivars generally decreased toward low latitudes (Lee et al., 2012). This is in agreement with other studies showing distinct adaptability of cultivars at the regional scale (Hopkins et al., 1995; Casler et al., 2004). In general, common upland cultivars have better potential productivity at higher latitudes but faster maturity results in lower yields at more southern locations (Fike et al., 2006). The latitudinal differences in yield were attributed to the cultivar-specific temperature response for switchgrass growth under California conditions (Lee et al., 2012). All selected cultivars had potentially better biomass productivity in the Central Valley of California compared to the data measured (mostly 10–14 Mg ha
1 yr
1) or simulated in the other regions of the USA (Wullschleger et al., 2010; Kang et al., 2014). The modeled switchgrass yields responded nonlinearly to N fertilization and irrigation interactions. For all cultivars and irrigation rates (including no irrigation), the yields decreased by 1.0–8.3 Mg ha
1 yr
1 (11.9–33.9%) in the near term and by 0.1– 9.4 Mg ha
1 yr
1 (1.5–32.8%) in the long term by applying half of the baseline N fertilizer rate (Fig. 3). The yield differences by the change in N rate had the uncertainty of 2.0–4.7 Mg ha
1 yr
1 in the near term, which continued to increase to 2.5–6.3 Mg ha
1 yr
1 in the long term. Similarly, Pedroso et al. (2014b) found significant effects of N fertilizer inputs on Alamo switchgrass yield, which were interacted with other input management. Ma et al. (2001) also showed that shoot biomass was about 22% lower when the fertilizer rate decreased from 224 to 112 kg N ha
1. By decreasing irrigation rates from 99 to 50 cm H2O, the yields decreased by 5.9– 9.5 Mg ha
1 yr
1 among the cultivars with the baseline N input and by 3.7–5.5 Mg ha
1 yr
1 with reduced N input. The yield declines increased up to 15.7 Mg ha
1 yr
1, with the corresponding uncertainty of 4.6 Mg ha
1 yr
1, when no or less irrigation was considered. This suggests that applied N is a principal yieldlimiting factor under the conditions where no or low water stress is

expected, while the productivity was not affected by N input with no or restricted (i.e., at 50 cm H2O) irrigation in the near term. Limited irrigation led to greater yield declines in the long term, but the effects were highly uncertain (the CV of the yields increased up to 60.5%). The response of switchgrass yields to N fertilization can be explained by a simple balance between N input and removal by harvest. In upper-southeastern USA, 43–233 kg N ha
1 was annually removed in biomass harvested for Alamo, which was fertilized at 50–100 kg N ha
1 and harvested once or twice (Fike et al., 2006). Lemus et al. (2008) showed that 76–127 kg N ha
1 yr
1 was removed in biomass harvested for Cave-in-Rock, which was fertilized with 0–270 kg N ha
1 and harvested twice per year. Parrish and Fike (2005) and Vogel et al. (2002) also reported similar ranges of N removal in harvested biomass at sites in the Midwest. However, the effect of applied N on yield was insignificant at three sites in South Dakota where yields were below 4 Mg ha
1 yr
1 (Mulkey et al., 2006). Data from the California switchgrass cultivar trials showed that the total amount of harvested biomass N for Trailblazer was 66–363 kg N ha
1 yr
1 at the fertilizer rates of 112–225 kg N ha
1 (Pedroso et al., 2013). The model estimated that the harvested biomass N was 117– 368 kg N ha
1 yr
1 for Trailblazer each year at similar N fertilizer and irrigation rates (data not shown). Crop N removal was significantly affected by N management (Pedroso et al., 2014b), showing that low-input systems (i.e., low intensity irrigation and harvest) with increasing N fertilizer rates tended to lead to a more positive soil N balance. In general, rain-fed switchgrass has a stronger response to N inputs than soil-water availability (Heaton et al., 2004). It is important to provide and maintain adequate soil fertility for sustainable switchgrass cropping systems over time, especially in the near term, but it is economically impractical to depend on the use of inorganic fertilizers only. Once established, soil inputs from residue return and constant root biomass contribution to SOC may contribute to soil fertility. Nevertheless, alternative N management options are needed due to unavoidable irrigation requirements for switchgrass production in a Mediterranean climate. Nitrogen in switchgrass can be provided, for

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

example, through biological N2 fixation by legume species with enhanced water stress tolerance when grown in mixture (Wang et al., 2010), but the economic feasibility would have to be assessed (Griffith et al., 2011). Studies suggest that biomass yields generally decrease with stand age after several years as soil-water is depleted (Jager et al., 2010; Tulbure et al., 2012). However, our model results did not show such yield declines under N or water deficit after a typical life span of 10–15 years if N or water input was continuously maintained at a low level. It is possible that environmental tolerance of switchgrass may be overestimated in our simulations. Thus, the long-term effects of management practices that lead to sub-optimal biomass productivity should be interpreted with caution as little is known about long-term yield trends that can neither confirm nor reject this expectation. 3.2. Soil organic carbon stocks and changes During the establishment year, SOC stock was around 24.5  6.1 Mg C ha
1. Under the baseline management practices, Alamo had

69

the highest C sequestration potential (0.92  0.15 Mg C ha
1 yr
1) among the cultivars in the near term (Figs. 4 and 5). The other cultivars were also a net atmospheric CO2 sink by sequestering C, ranging from 0.42  0.13 to 0.84  0.16 Mg C ha
1 yr
1 across the same period. The CV of the baseline C sequestration rates were 16.9–30.8% in the near term. Only in the establishment year (data not shown), Sunburst and Trailblazer had an initial loss of SOC, at
0.28 and
0.23, respectively. In comparison, long-term soil C sequestration (CV = 70.6–99.1%) was not predicted under the baseline management conditions. The modeled SOC changes in the near term were comparable to those measured under field conditions. Changes in soil C stock ranged from
1.20 to 1.74 Mg C ha
1 yr
1 at sites in the central and northern Great Plains, USA after five years under switchgrass, including Cave-in-Rock, Sunburst, and Trailblazer (Liebig et al., 2008). Anderson-Teixeira et al. (2009) reported that soil C under switchgrass increased at a rate of about 0.4 Mg C ha
1 yr
1 in the 0– 30 cm depth. In addition, there was a considerable increase in soil C in deeper depths below 20 cm over time (Liebig et al., 2005; Lee et al., 2007; Liebig et al., 2008). Frank et al. (2004) reported that soil

Fig. 4. Changes in soil organic carbon content over time by N fertilizer application (224 and 112 kg N ha
1) and irrigation (0, 25, 50, 75, and 99 cm H2O).

70

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

Fig. 5. Soil carbon sequestration under switchgrass in the near and long term (2–12 years and 90–100 years after planting, respectively) as affected by N fertilizer (224 and 112 kg N ha
1) and irrigation rates (0, 25, 50, 75, and 99 cm H2O). The error bars shows standard deviation around modeled results. A = Alamo; K = Kanlow; CIR = Cave-in-Rock; B = Blackwell; S = Sunburst; T = Trailblazer.

C stocks in the 0–90 cm depth under Sunburst and Dacotah increased by 7.58 Mg C ha
1 yr
1 at Mandan, North Dakota, USA, with greater changes in 30–90 cm than in 0–30 cm depths. Skinner and Adler (2010) showed that ecosystem respiration decreased over time under Cave-in-Rock in the northeastern USA, while harvested biomass increased in a few years after planting. In comparison, longer-term changes in soil C can vary greatly as the pool of soil C increases under switchgrass. Garten (2011) estimated that soil C sequestration rates after 30 years varied from
0.28 to 1.14 Mg C ha
1 yr
1 at different levels of initial soil C stocks and N fertilization. Lemus and Lal (2005) showed that soil C sequestration still leads to a significant reduction of net soil GHG fluxes over 40–60 years of switchgrass production. This suggests that switchgrass has the limited, non-permanent potential to sequester atmospheric CO2. Soil organic C changes under switchgrass were affected by N fertilization (Fig. 4). The modeled C sequestration rate then decreased by 0.09–0.30 Mg C ha
1 yr
1 (15.5–52.8%) in the near term when N fertilization was reduced from 224 to 112 kg N ha
1 across the different irrigation rates (Fig. 5). The uncertainty of the differences ranged from 0.16 to 0.32 Mg C ha
1 yr
1 in the near term. For all selected cultivars, the soil C sequestration rate did not differ much when irrigated with 50 cm H2O in the near term when switchgrass was fertilized at each fixed rate. The soil potential to increase C stocks under fertilized switchgrass in the near term was greatly reduced by no irrigation or when irrigated at 25 cm H2O. In comparison, there was no obvious N fertilization effect on soil C sequestration in the long term. The long-term soil C sequestration rates tended to increase with reducing irrigation rates, although highly variable. The model simulated some inconsistent effect of N fertilization on soil C sequestration over time, presumably as it reached an steady-state soil C level (Garten, 2011). In addition, soil C sequestration was more sensitive to N fertilization for Trailblazer and Sunburst than the other cultivars in the near term. This was likely because upland ecotypes tended to be less subject to drought than lowland ecotypes (Stroup et al., 2003) as well as have relatively higher N requirements (Porter, 1966). Since most of the aboveground biomass was harvested, the

capacity of each cultivar to sequester C was expected to depend on changes in belowground C inputs from roots and rhizodeposition. This was in agreement with many studies, showing the significant effect of belowground inputs on soil C change under switchgrass (Ma et al., 2000a,b; Lemus and Lal, 2005; Liebig et al., 2008). Changes in soil C, however, may not clearly differ by cultivar due to confounding effects from types of land use conversion even when significant differences in root biomass or quality exist (Sanderson, 2008). Regardless of cultivar, the increases in soil C stock were expected to be high at intermediate or low levels of yields (Silver et al., 2010). The amount of soil C input needed to maintain total C levels was estimated to be about 310 g C m
2 yr
1 in tilled soils specifically in California (Kong et al., 2005), when both aboveground residue and root C were considered. The changes in SOC under switchgrass were almost negligible with no N and water input (Fig. 4). Otherwise, soil C sequestration under fertilized and irrigated switchgrass was expected with the amount of soil C input ranging from 108 to 757 g C m
2 in the near term (data not shown). However, long-term C sequestration was not related to the levels of soil C input. Little long-term changes in soil C was predicted (Fig. 5), once the soil C stocks increased to be close to levels of C stocks in annual grassland soils of California (Silver et al., 2010). Similarly, Chamberlain et al. (2011) showed that the change in soil C under switchgrass would depend on fertilization or historical land use. Despite hotter summer and frequent irrigation, switchgrass sequestered soil C in the Central Valley of California. Our model results suggest that soil C sequestration is a near-term component for a net GHG sink in a soil particularly with high initial soil C stock, as found in other studies (McLaughlin et al., 2002; Adler et al., 2007). However, a further gain in soil C can be made by increasing residue retention at harvest, as our simulations considered 95% residue removal at harvest. 3.3. N2O emissions If switchgrass is intensively fertilized, N2O emissions from the soil can negate any reduction in warming potential due to

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

increased C sequestration (Melillo et al., 2009). Uncertainty associated with N2O emissions can be predominantly influenced by N fertilization among sources of N input (Hutchinson et al., 2007; Meyer-Aurich et al., 2012), hence contributing to the overall variation of net soil GHG mitigation potential. Baseline N2O emissions ranged from 1.4  0.6 to 2.50.9 kg N2O–N ha
1 yr
1 in the near term and from 1.8  1.1 to 5.2  2.3 kg N2O–N ha
1 yr
1 in the long term for all cultivars (Figs. 6 and 7). The baseline emissions increased over time in the established stands, with the CV of 37.7– 46.3% in the near term and 42.1–62.0% in the long term across the cultivars. Emissions of N2O increased over time with increasing organic matter pools in the fully irrigated soil. This trend was presumably due to the contribution from increasing organic C availability (Lesschen et al., 2011). The lowland and southern upland cultivars had similar N2O emissions, but the northern upland cultivars had relatively smaller emissions particularly in the long term. The differences were mainly because the northern upland cultivars tended to have higher biomass N allocation than the other ecotypes. By decreasing N fertilization from 224 to 112 kg N ha
1, modeled N2O fluxes were between 0.4  0.2 and

71

1.2  0.7 kg N2O–N ha
1 yr
1 in the near term, which was reduced by 0.4–1.6 kg N2O–N ha
1 yr
1 across the irrigation rates. In the long term, N2O emissions ranged from 0.70.5 to 2.61.3 kg N2O-N ha
1 yr
1 with 112 kg N ha
1 applied, corresponding to the reduction by 0.4–3.8 kg N2O–N ha
1 yr
1. The near- and long-term uncertainties in the difference had the range of 0.5–1.2 kg N2O– N ha
1 yr
1 and 1.1–2.4 kg N2O–N ha
1 yr
1, respectively. Different N inputs from fertilization and frequent irrigation can substantially affect N2O emissions under switchgrass in the Central Valley of California. The model showed greater reduction in N2O emissions for all cultivars by applying less N fertilizer when irrigated at rates 75 cm H2O per year over time (Fig. 6). In comparison, the reduction in the N2O emissions by decreased N input was not apparent for irrigation at 50 cm H2O. This suggests that the biophysical potential of less fertilized switchgrass to reduce N losses can be constrained by the availability of water. When fertilized at 224 kg N ha
1, for example, changes in N2O emission from baseline due to different irrigation rates were between
1.5 and 0.4 kg N2O–N ha
1 yr
1, while the uncertainty around the differences was estimated to be around 0.8–1.4 kg N2O–

Fig. 6. Changes in nitrous oxide flux over time by N fertilizer application (224 and 112 kg N ha
1) and irrigation (0, 25, 50, 75, and 99 cm H2O).

72

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

Fig. 7. Nitrous oxide fluxes under switchgrass in the near and long term (2–12 years and 90–100 years after planting, respectively) as affected by N fertilizer (224 and 112 kg N ha
1) and irrigation rates (0, 25, 50, 75, and 99 cm H2O). The error bars shows standard deviation around modeled results. A = Alamo; K = Kanlow; CIR = Cave-in-Rock; B = Blackwell; S = Sunburst; T = Trailblazer.

N ha
1 yr
1. The magnitude of reduced emissions was in an agreement with other simulation studies (Adler et al., 2007; Chamberlain et al., 2011), which reported 0.1–0.6 kg N2O–N ha
1 yr
1 when fertilized at 90 kg N ha
1. Overall, N2O emissions for both levels of N fertilization tended to increase over time at each irrigation rate, potentially counterbalancing mitigation potential from soil C sequestration. This was presumably because microbial nitrification and denitrification may be less constrained by soil N availability while stands reached full yield potential. Initial decreases in nitrate leaching may also contribute to this pattern, which should be further tested under field conditions. For all cultivars, modeled nitrate leaching ranged from 56.9 to 87.9 kg N ha
1 in the establishment year and then decreased to 0.4– 14.3 kg N ha
1 over the following five years under the baseline management practices (data not shown). McLaughlin et al. (2002) reported that nitrate–N loss by annual runoff decreased from 10.7 to 0.3 kg N ha
1 over the first three years for switchgrass in northern Alabama, USA. The baseline nitrate leaching tended to increase annually following the initial decline for the lowland and

southern upland cultivars, but otherwise it did not change much. When switchgrass growth was not limited by drought stress, our model results suggest that the levels of modeled N2O emissions depended on the fertilizer rate but the overall emission changes tended to follow SOC changes over time. Also, the soil may have increased risk for N leaching losses due to the accumulation of soil organic matter N pools with faster turnover if switchgrass growth is limited by drought stress. This highlights the importance of careful irrigation management for controlling N2O emissions. Applying less N fertilizer can effectively reduce N2O emissions in both the near and long term, but optimizing the use efficiency of both N fertilizer and water inputs can have more important implications for long-term N2O mitigation. In addition, soil fertility and belowground biomass N stocks (or switchgrass productivity) are likely to be important factors in order to adjust N fertilizer rates over time. The importance of N2O emissions in global warming potential is well acknowledged as how direct and indirect N2O emissions are estimated will eventually justify life-cycle GHG balances of switchgrass (Adler et al., 2007). It is still a challenge to

Fig. 8. Net soil GHG fluxes or global warming potentials (GWP) under switchgrass in the near and long term (2–12 years and 90–100 years after planting, respectively) as affected by N fertilizer (224 and 112 kg N ha
1) and irrigation rates (0, 25, 50, 75, and 99 cm H2O). The points represent cultivar-specific averages by period.

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

quantify uncertainty in N2O emission estimates under switchgrass based on long-term measurements covering a wide range of environmental and management conditions. 3.4. Net soil GHG balance All cultivars had negative net soil GHG fluxes (sink), ranging from
0.79  0.39 to
2.62  0.51 Mg CO2-eq ha
1 yr
1 in the near term under the baseline management practices (Fig. 8). Soil GHG emissions under switchgrass became positive (source) in the long term, varying from 0.62  0.48 to 2.26  1.02 kg CO2-eq ha
1 yr
1 across the cultivars. For the other alternative input practices, the same pattern was simulated for each period. The only exception was long-term effects of no irrigation system on soil emissions, mostly being GHG neutral. In the near term, a relatively small GHG sink was expected when fertilized at 112 kg N ha
1 compared to 224 kg N ha
1, by 0.09–0.45 Mg CO2-eq ha
1 yr
1. However, there was no change in the overall GHG balance by irrigation at either N fertilizer rate. In addition, the soil potential as a GHG sink appeared to differ by cultivar. This suggests that the near-term GHG sink was primarily determined by the level of SOC accumulation potential. These estimates was similar to net GHG reduction under switchgrass fertilized with 56 kg N ha
1, which was simulated at the site scale (Hudiburg et al., 2015). In contrast, Qin et al. (2015) estimated that switchgrass had slightly positive GHG balance of 0.3 Mg CO2-eq ha
1 yr
1 under annual N input of 67 kg N ha
1. In comparison, the long-term GHG emissions increased with irrigation rates when switchgrass was fertilized at 224 kg N ha
1. When N input was reduced to 112 kg N ha
1, there were some but marginal differences in the GHG emissions by irrigation. Cultivarspecific differences still existed with the input of 224 kg N ha
1 and 75 cm H2O. Therefore, the long-term net soil GHG fluxes were determined by N effects on N2O emissions. This is particularly relevant because fertilized switchgrass may have an increased risk for direct and indirect N2O emissions in the long term. 4. Conclusions Our model results showed that switchgrass should be fertilized and irrigated for effective regional biomass production under California conditions. Despite high yield variability among the cultivars, the lowland cultivars can be generally recommended if optimizing biomass productivity is targeted. However, it is important to optimize the input rates of external N and irrigation water to balance between biomass yields and soil GHG mitigation. Our model results indicate that in the near term switchgrass should be managed for increasing soil C sequestration. Especially to be a net GHG sink in the near term, switchgrass requires N fertilizer although its potential to mitigate GHG emissions would not be altered substantially by N fertilizer or irrigation rate. However, since soil C sequestration is not permanent, the mitigation of N2O emissions would become more important for longer-term GHG mitigation. Adequate irrigation and reduced N fertilizer input could support sufficient switchgrass growth and reduce N2O emissions to maintain a GHG neutral switchgrass cropping system. Acknowledgment This research was funded by a grant from Chevron Technology Ventures, USA. References Adler, P.R., Del Grosso, S.J., Parton, W.J., 2007. Life-cycle assessment of net greenhouse-gas flux for bioenergy cropping systems. Ecol. Appl. 17, 675–691.

73

Al-Kaisi, M.M., Grote, J.B., 2007. Cropping systems effects on improving soil carbon stocks of exposed subsoil. Soil Sci. Soc. Am. J. 71, 1381–1388. Anderson-Teixeira, K.J., Davis, S.C., Masters, M.D., Delucia, E.H., 2009. Changes in soil organic carbon under biofuel crops. Glob. Change Biol. Bioenergy 1, 75–96. Barney, J.N., DiTomaso, J.M., 2008. Nonnative species and bioenergy: are we cultivating the next invader? Bioscience 58, 64–70. Boehmel, C., Lewandowski, I., Claupein, W., 2008. Comparing annual and perennial energy cropping systems with different management intensities. Agric. Syst. 96, 224–236. Bransby, D.I., McLaughlin, S.B., Parrish, D.J., 1998. A review of carbon and nitrogen balances in switchgrass grown for energy. Biomass Bioenergy 14, 379–384. Casler, M.D., Vogel, K.P., Taliaferro, C.M., Wynia, R.L., 2004. Latitudinal adaptation of switchgrass populations. Crop Sci. 44, 293–303. Chamberlain, J.F., Miller, S.A., Frederick, J.R., 2011. Using DAYCENT to quantify onfarm GHG emissions and N dynamics of land use conversion to N-managed switchgrass in the Southern U.S. Agric. Ecosyst. Environ. 141, 332–341. De Gryze, S., Wolf, A., Kaffka, S.R., Mitchell, J., Rolston, D.E., Temple, S.R., Lee, J., Six, J., 2010. Simulating greenhouse gas budgets of four California cropping systems under conventional and alternative management. Ecol. Appl. 20, 1805–1819. De Gryze, S., Lee, J., Ogle, S., Paustian, K., Six, J., 2011. Assessing the potential for greenhouse gas mitigation in intensively managed annual cropping systems at the regional scale. Agric. Ecosyst. Environ. 144, 150–158. Del Grosso, S.J., Parton, W.J., Mosier, A.R., Hartman, M.D., Brenner, J., Ojima, D.S., Schimel, D.S., 2001. Simulated interaction of carbon dynamics and nitrogen trace gas fluxes using the daycent model. In: Schaffer, M., Ma, L., Hansen, S. (Eds.), Modeling Carbon and Nitrogen Dynamics for Soil Management. CRC Press, Boca Raton, FL, pp. 303–332. Duval, B.D., Anderson-Teixeira, K.J., Davis, S.C., Keogh, C., Long, S.P., Parton, W.J., DeLucia, E.H., 2013. Predicting greenhouse gas emissions and soil carbon from changing pasture to an energy crop. PLoS One 8. Fike, J.H., Parrish, D.J., Wolf, D.D., Balasko, J.A., Green, J.T., Rasnake, M., Reynolds, J.H., 2006. Long-term yield potential of switchgrass-for-biofuel systems. Biomass Bioenergy 30, 198–206. Forster, P., Ramaswamy, V., Artaxo, P., Berntsen, T., Betts, R., Fahey, D.W., Haywood, J., Lean, J., Lowe, D.C., Myhre, G., Nganga, J., Prinn, R., Raga, G., Schulz, M., Dorland, R.V., 2007. 2007: changes in atmospheric constituents and in radiative forcing. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge, United Kingdom and New York, NY, USA. Frank, A.B., Berdahl, J.D., Hanson, J.D., Liebig, M.A., Johnson, H.A., 2004. Biomass and carbon partitioning in switchgrass. Crop Sci. 44, 1391–1396. Garten, C.T., 2011. Review and model-based analysis of factors influencing soil carbon sequestration beneath switchgrass (Panicum virgatum). Bioenerg. Res. 5, 124–138. Giampietro, M., Ulgiati, S., Pimentel, D., 1997. Feasibility of large-scale biofuel production—does an enlargement of scale change the picture? Bioscience 47, 587–600. Griffith, A.P., Epplin, F.M., Fuhlendorf, S.D., Gillen, R., 2011. A comparison of perennial polycultures and monocultures for producing biomass for biorefinery feedstock. Agron. J. 103, 617–627. Heaton, E., Voigt, T., Long, S.P., 2004. A quantitative review comparing the yields of two candidate C4 perennial biomass crops in relation to nitrogen, temperature and water. Biomass Bioenergy 27, 21–30. Hopkins, A.A., Vogel, K.P., Moore, K.J., Johnson, K.D., Carlson, I.T., 1995. Genotypic variability and genotype X environment interactions among switchgrass accessions from the midwestern USA. Crop Sci. 35, 565–571. Hudiburg, T.W., Davis, S.C., Parton, W., Delucia, E.H., 2015. Bioenergy crop greenhouse gas mitigation potential under a range of management practices. Glob. Change Biol. Bioenergy 7, 366–374. Hutchinson, J.J., Grant, B.B., Smith, W.N., Desjardins, R.L., Campbell, C.A., Worth, D.E., Verge, X.R., 2007. Estimates of direct nitrous oxide emissions from Canadian agroecosystems and their uncertainties. Can. J. Soil Sci. 87, 141–152. Jager, H.I., Baskaran, L.M., Brandt, C.C., Davis, E.B., Gunderson, C.A., Wullschleger, S. D., 2010. Empirical geographic modeling of switchgrass yields in the United States. Glob. Change Biol. Bioenergy 2, 248–257. Kaffka, S.R., 2009. Can feedstock production for biofuels be sustainable in California. Calif. Agric. 63, 202–207. Kang, S.J., Nair, S.S., Kline, K.L., Nichols, J.A., Wang, D.L., Post, W.M., Brandt, C.C., Wullschleger, S.D., Singh, N., Wei, Y.X., 2014. Global simulation of bioenergy crop productivity: analytical framework and case study for switchgrass. Glob. Change Biol. Bioenergy 6, 14–25. Kong, A.Y.Y., Six, J., Bryant, D.C., Denison, R.F., van Kessel, C., 2005. The relationship between carbon input, aggregation, and soil organic carbon stabilization in sustainable cropping systems. Soil Sci. Soc. Am. J. 69, 1078–1085. Lal, R., 2005. World crop residues production and implications of its use as a biofuel. Environ. Int. 31, 575–584. Lee, D.K., Owens, V.N., Doolittle, J.J., 2007. Switchgrass and soil carbon sequestration response to ammonium nitrate, manure, and harvest frequency on conservation reserve program land. Agron. J. 99, 462–468. Lee, J., De Gryze, S., Six, J., 2011. Effect of climate change on field crop production in California’s Central Valley. Clim. Change 109, 335–353. Lee, J., Pedroso, G., Linquist, B.A., Putnam, D., van Kessel, C., Six, J., 2012. Simulating switchgrass biomass production across ecoregions using the DAYCENT model. Glob. Change Biol. Bioenergy 4, 521–533.

74

J. Lee et al. / Agriculture, Ecosystems and Environment 212 (2015) 64–74

Lemus, R., Lal, R., 2005. Bioenergy crops and carbon sequestration. Crit. Rev. Plant Sci. 24, 1–21. Lemus, R., Parrish, D.J., Abaye, O., 2008. Nitrogen-use dynamics in switchgrass grown for biomass. Bioenergy Res. 1, 153–162. Lesschen, J.P., Velthof, G.L., de Vries, W., Kros, J., 2011. Differentiation of nitrous oxide emission factors for agricultural soils. Environ. Pollut. 159, 3215–3222. Lewandowski, I., Scurlock, J.M.O., Lindvall, E., Christou, M., 2003. The development and current status of perennial rhizomatous grasses as energy crops in the US and Europe. Biomass Bioenergy 25, 335–361. Liebig, M.A., Johnson, H.A., Hanson, J.D., Frank, A.B., 2005. Soil carbon under switchgrass stands and cultivated cropland. Biomass Bioenergy 28, 347–354. Liebig, M.A., Schmer, M.R., Vogel, K.P., Mitchell, R.B., 2008. Soil carbon storage by switchgrass grown for bioenergy. Bioenergy Res. 1, 215–222. Ma, Z., Wood, C.W., Bransby, D.I., 2000a. Carbon dynamics subsequent to establishment of switchgrass. Biomass Bioenergy 18, 93–104. Ma, Z., Wood, C.W., Bransby, D.I., 2000b. Soil management impacts on soil carbon sequestration by switchgrass. Biomass Bioenergy 18, 469–477. Ma, Z., Wood, C.W., Bransby, D.I., 2001. Impact of row spacing, nitrogen rate, and time on carbon partitioning of switchgrass. Biomass Bioenergy 20, 413–419. McLaughlin, S.B., Walsh, M.E., 1998. Evaluating environmental consequences of producing herbaceous crops for bioenergy. Biomass Bioenergy 14, 317–324. McLaughlin, S.B., de la Torre Ugarte, D.G., Garten, Jr, Lynd, L.R., Sanderson, M.A., Tolbert, V.R., Wolf, D.D., 2002. High-value renewable energy from prairie grasses. Environ. Sci. Technol. 36, 2122–2129. Melillo, J.M., Reilly, J.M., Kicklighter, D.W., Gurgel, A.C., Cronin, T.W., Paltsev, S., Felzer, B.S., Wang, X.D., Sokolov, A.P., Schlosser, C.A., 2009. Indirect emissions from biofuels: how important. Science 326, 1397–1399. Mensah, F., Schoenau, J.J., Malhi, S.S., 2003. Soil carbon changes in cultivated and excavated land converted to grasses in east-central Saskatchewan. Biogeochemistry 63, 85–92. Meyer-Aurich, A., Schattauer, A., Hellebrand, H.J., Klauss, H., Plochl, M., Berg, W., 2012. Impact of uncertainties on greenhouse gas mitigation potential of biogas production from agricultural resources. Renew. Energy 37, 277–284. Monti, A., Barbanti, L., Zatta, A., Zegada-Lizarazu, W., 2012. The contribution of switchgrass in reducing GHG emissions. Glob. Change Biol. Bioenergy 4, 420– 434. Mulkey, V.R., Owens, V.N., Lee, D.K., 2006. Management of switchgrass-dominated conservation reserve program lands for biomass production in South Dakota. Crop Sci. 46, 712–720. Nikièma, P., Rothstein, D.E., Min, D.H., Kapp, C.J., 2011. Nitrogen fertilization of switchgrass increases biomass yield and improves net greenhouse gas balance in northern Michigan, USA. Biomass Bioenergy 35, 4356–4367. Pacala, S., Socolow, R., 2004. Stabilization wedges: solving the climate problem for the next 50 years with current technologies. Science 305, 968–972. Parrish, D.J., Fike, J.H., 2005. The biology and agronomy of switchgrass for biofuels. Crit. Rev. Plant Sci. 24, 423–459. Parton, W.J., Mosier, A.R., Ojima, D.S., Valentine, D.W., Schimel, D.S., Weier, K., Kulmala, A.E., 1996. Generalized model for N2 and N2O production from nitrification and denitrification. Glob. Biogeochem. Cycles 10, 401–412. Paruelo, J.M., Lauenroth, W.K., 1996. Relative abundance of plant functional types in grasslands and shrublands of north America. Ecol. Appl. 6, 1212–1224. Pedroso, G.M., De Ben, C., Hutmacher, R.B., Orloff, S., Putnam, D., Six, J., van Kessel, C., Wright, S., Linquist, B.A., 2011. Switchgrass is a promising, high-yielding crop for California biofuel. Calif. Agric. 65, 168–173.

Pedroso, G.M., Hutmacher, R.B., Putnam, D., Wright, S.D., Six, J., van Kessel, C., Linquist, B.A., 2013. Yield and nitrogen management of irrigated switchgrass systems in diverse ecoregions. Agron. J. 105, 311–320. Pedroso, G.M., Hutmacher, R.B., Putnam, D., Six, J., van Kessel, C., Linquist, B.A., 2014a. Biomass yield and nitrogen use of potential C4 and C3 dedicated energy crops in a Mediterranean climate. Field Crops Res. 161, 149–157. Pedroso, G.M., van Kessel, C., Six, J., Putnam, D.H., Linquist, B.A., 2014b. Productivity, 15N dynamics and water use efficiency in low- and high-input switchgrass systems. Glob. Change Biol. Bioenergy . Porter, C.L., 1966. An analysis of variation between upland and lowland switchgrass, Panicum Virgatum L. in Central Oklahoma. Ecology 47, 980–992. Qin, Z.C., Zhuang, Q.L., Zhu, X.D., 2015. Carbon and nitrogen dynamics in bioenergy ecosystems: 2. Potential greenhouse gas emissions and global warming intensity in the conterminous United States. Glob. Change Biol. Bioenergy 7, 25–39. Sanderson, M.A., 2008. Upland switchgrass yield, nutritive value, and soil carbon changes under grazing and clipping. Agron. J. 100, 510–516. Saxton, K.E., Rawls, W.J., Romberger, J.S., Papendick, R.I., 1986. Estimating generalized soil-water characteristics from texture. Soil Sci. Soc. Am. J. 50, 1031– 1036. Searchinger, T., Heimlich, R., Houghton, R.A., Dong, F., Elobeid, A., Fabiosa, J., Tokgoz, S., Hayes, D., Yu, T.-H., 2008. Use of U. S. croplands for biofuels increases greenhouse gases through emissions from land-use change. Science 319, 1238– 1240. Silver, W.L., Ryals, R., Eviner, V., 2010. Soil carbon pools in California's annual grassland ecosystems. Rangel. Ecol. Manag. 63, 128–136. Sims, R.E.H., Hastings, A., Schlamadinger, B., Taylor, G., Smith, P., 2006. Energy crops: current status and future prospects. Glob. Change Biol. 12, 2054–2076. Skinner, R.H., Adler, P.R., 2010. Carbon dioxide and water fluxes from switchgrass managed for bioenergy production. Agric. Ecosyst. Environ. 138, 257–264. Stroup, J.A., Sanderson, M.A., Muir, J.P., McFarland, M.J., Reed, R.L., 2003. Comparison of growth and performance in upland and lowland switchgrass types to water and nitrogen stress. Bioresour. Technol. 86, 65–72. Thornton, P.E., Thornton, M.M., Mayer, B.W., Wilhelmi, N., Wei, Y., Devarakonda, R., Cook, R.B., 2014. Daymet: Daily Surface Weather Data on a 1-km Grid for North America, Version 2. Data set. Available on-line [http://daac.ornl.gov] from Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, USA. http://dx.doi.org/10.3334/ORNLDAAC/1219. Tulbure, M.G., Wimberly, M.C., Boe, A., Owens, V.N., 2012. Climatic and genetic controls of yields of switchgrass, a model bioenergy species. Agric. Ecosyst. Environ. 146, 121–129. Vogel, K.P., Brejda, J.J., Walters, D.T., Buxton, D.R., 2002. Switchgrass biomass production in the Midwest USA: harvest and nitrogen management. Agron. J. 94, 413–420. Wang, D., Lebauer, D.S., Dietze, M.C., 2010. A quantitative review comparing the yield of switchgrass in monocultures and mixtures in relation to climate and management factors. Glob. Chamge Biol. Bioenergy 2, 16–25. Wright, L., 2007. Historical Perspective on How and Why Switchgrass Was Selected as a Model High-potential Energy Crop. Oak Ridge National Laboratory, Oak Ridge, TN. Wullschleger, S.D., Davis, E.B., Borsuk, M.E., Gunderson, C.A., Lynd, L.R., 2010. Biomass production in switchgrass across the United States: database description and determinants of yield. Agron. J. 102, 1158–1168. Yi, F., Mérel, P., Lee, J., Hossein Farzin, Y., Six, J., 2013. Switchgrass in California: where, and at what price. Glob. Change Biol. Bioenergy .