Intercropping kura clover with prairie cordgrass mitigates soil greenhouse gas fluxes

Prairie cordgrass (PCG) (Spartina pectinata Link) has a high tolerance to soil salinity and waterlogging, therefore, it can thrive on marginal lands. Optimizing the nitrogen (N) input is crucial to achieving desirable biomass production of PCG without negatively impacting the environment. Thus, this study was based on the hypothesis that the use of legumes such as kura clover (Trifolium ambiguum M. Bieb.) (KC) as an intercrop with PCG can provide extra N to the crop reducing the additional N fertilizer and mitigating soil surface greenhouse gas (GHG) emissions. Specific objective of the study was to assess the impact of PCG managed with different N rates [0 kg N ha−1 (PCG-0N), 75 kg N ha−1 (PCG-75N), 150 kg N ha−1 (PCG-150N), and 225 kg N ha−1 (PCG-255N)], and PCG intercropped with KC (PCG-KC) on GHG fluxes and biomass yield. The experimental site was established in 2010 in South Dakota under a marginally yielding cropland. The GHG fluxes were measured from 2014 through 2018 growing seasons using the static chamber. Net global warming potential (GWP) was calculated. Data showed that cumulative CH4 and CO2 fluxes were similar for all the treatments over the study period. However, the PCG-KC, PCG-0N, and PCG-75N recorded lower cumulative N2O fluxes (384, 402, and 499 g N ha−1, respectively) than the PCG-150N (644 g N ha−1) and PCG-255N (697 g N ha−1). The PCG-KC produced 85% and 39% higher yield than the PCG-0N in 2016 and 2017, respectively, and similar yield to the other treatments (PCG-75N, PCG-150N, and PCG-255N) in these years. Net GWP was 52% lower for the PCG-KC (112.38 kg CO2-eq ha−1) compared to the PCG-225N (227.78 kg CO2-eq ha−1), but similar to other treatments. Soil total N was 15%% and 13% higher under PCG-KC (3.7 g kg−1) than that under PCG-0N (3.2 g kg−1) and PCG-75N (3.3 g kg−1), respectively. This study concludes that intercropping prairie cordgrass with kura clover can enhance biomass yield and reduce fertilizer-derived N2O emissions and net global warming potential.

. The range of SOC was from 30.9 to 37.1 g kg −1 in both years. Nitrogen fertilization rates and intercropping of KC did not show any significant impact on SOC in both years (Fig. 2). However, SOC did not change within the N fertilization rates over time; yet, it was 13.5% higher in 2018 than in 2011 under PCG-KC plots. Total N ranged from 2.9 to 4.0 g kg −1 in both years. Total N content over time (from 2011 to 2018) did not significantly change for each treatment; however, N fertilization rates and intercropping of KC in 2018 significantly affected TN (P = 0.0314). In 2018, PCG-KC recorded significantly 15% and 13% www.nature.com/scientificreports www.nature.com/scientificreports/ higher soil TN (3.7 g kg −1 ) than PCG-0N (3.2 g kg −1 ) and PCG-75N (3.3 g kg −1 ), respectively, but similar TN to PCG-150N (3.5 g kg −1 ) and PCG-225N (3.6 g kg −1 ). Biomass yield. Biomass yields of PCG in 2015, 2016 and 2018 are listed in Table 1. Overall, biomass production ranged from 4.3 to 7.3 Mg ha −1 in 2015, 4.1 to 15.1 Mg ha −1 in 2016, and 5.3 to 13.9 Mg ha −1 in 2018. In 2015, no significant differences on yield among treatments were observed; however, the yield was significantly affected by N fertilization rates and intercropping of KC in 2016 and 2018 (Table 1). In 2016, PCG-KC recorded 85% higher yield (11.19 Mg ha −1 ) than PCG-0N (6.05 Mg ha −1 ), but similar yield as PCG-75N (9.86 Mg ha −1 ), PCG-150N (10.88 Mg ha −1 ), and PCG-225N (14.08 Mg ha −1 ). Similar trends were also observed in 2018, with PCG-KC recording 39% higher yield (10.41 Mg ha −1 ) than PCG-0N (7.5 Mg ha −1 ), but similar yield as PCG-75N  Fig. 3. As it was expected, only positive values of CO 2 fluxes were observed under all treatments in all years. Generally, N fertilization rates and KC intercropping emitted higher CO 2 flux at the middle and end of the growing seasons compared to the beginning of the growing seasons, with peaks of this gas being observed in all years. Carbon dioxide fluxes under N fertilization rates and KC intercropping generally showed a similar pattern in 2016 and 2018. However, the PCG-225N seemed to emit higher CO 2 fluxes than all the other treatments at most dates in 2014 and 2015, but no significant differences on daily CO 2 fluxes were recorded between PCG-225N and other treatments at all dates (P ≤ 0.05 for all years). The cumulative CO 2 fluxes under N fertilization rates and KC intercropping are listed in Table 2. Differences on cumulative CO 2 fluxes among treatments were not significant, with PCG-KC, PCG-0N, PCG-75N, PCG-125N, and PCG-225N recording cumulative CO 2 fluxes of 2731.47 kg ha −1 , 2053.09 kg ha −1 , 2170.36 kg ha −1 , 2337.64 kg ha −1 , and 2661.64 kg ha −1 , respectively, ( Table 2).
Daily trend of CH 4 under N fertilization rates and KC intercropping in the four years are shown in Fig. 4. Methane flux was different from the CO 2 flux, where both negative and positive CH 4 fluxes were observed in all  Global warming potential. Global warming potential (GWP) for CH 4 (GWP CH4 ) and N 2 O (GWP N2O ) and net GWP can be seen in Table 3. Similar to cumulative CH 4 flux, GWP CH4 was similar under N fertilization rates and KC intercropping. However, N rates and KC intercropping affected GWP N2O , with PCG-KC (104.37 kg CO 2 -eq ha −1 ) recording 52% lower GWP N2O than PCG-225N (219.59 kg CO 2 -eq ha −1 ), but similar GWP N2O as other treatments ( Table 3). The inputs of GWP N2O in the net GWP was higher than the inputs of GWP CH4 over the study period. Similar to GWP N2O, the net GWP under PCG-KC (112.38 kg CO 2 -eq ha −1 ) was 52% lower than under PCG-225N (227.78 kg CO 2 -eq ha −1 ), but similar to other treatments.   www.nature.com/scientificreports www.nature.com/scientificreports/ Discussion Effects of N rates and kura clover intercropping on soil water content and temperature. This study showed that soil water content and temperature were similar under all treatments (N rates and KC intercropping). Li et al. 4 reported that N inputs ranged from 0 to 84 kg N ha −1 did not affect soil water content and soil temperature when applied to other bioenergy crops like camelina (Camelina sativa L.) and carinata (Brassica carinata). Intercropping reed canarygrass with galega did not affect soil water content as compared to reed canarygrass plots as reported by Epie et al. 14 . Nitrogen fertilizer additions to switchgrass (Panicum virgatum L.) at a rate of 112 kg ha −1 showed similar soil water content and temperature compared to none-fertilized soils 24,25 . Soil organic C was not affected by N rates and KC intercropping in 2018 (Fig. 2), perhaps explaining why these soil parameters were not affected in this study. Increased SOC can improve soil porosity, which in turn can increase soil water holding capacity 26 .

Soil properties as influenced by N rates and kura clover intercropping. Nitrogen fertilizers can
release hydrogen ions during nitrification process 27 , and legumes can acquire N as diatomic N rather than as nitrate 28 , signifying that both management practices can lower the soil pH. In addition, higher rates of fertilizer contain high salt concentrations, which may induce soil salnization 29 . However, N rates and KC intercropping with PCG had similar soil pH and EC in this study. Eight years of N fertilization and KC intercropping may not be enough to cause changes in soil pH and EC. Application of N fertilizer (0-112 kg N ha −1 ) to switchgrass for four years did not affect soil pH compared to the control in South Dakota 24 . Kidd et al. 30 reported that N fertilizer applications for 120 years acidified the soil in the grassland ecosystem. Intercropping corn with bambara groundnut (Vigna subterranea (L) Verdc) for five years 31 and wheat (Triticum aestivum) with faba beans (Vicia faba L.) for nine years 32 did not affect soil pH compared to the sole cropping system. Liebig et al. 33 reported that N fertilizer applications for 16 years affected soil salinization in the western corn belt.  Table 2. Cumulative soil surface CO 2 , CH 4   Similar to pH and EC, treatments did not affect SOC in this study. High N fertilization rates 4 and intercropping with legumes 34 in bioenergy system increased biomass production resulting in an increase in SOC. The increase in biomass with higher N rate and PCG-KC mixture was observed in 2016 and 2018 (Table 1); however, this increased biomass was not translated into an increase in the SOC. The aboveground biomass was harvested in all years, which may caused the lack of effects of N fertilization rates and KC intercropping on SOC. Alvarez et al. 35 reported that N application could increase the SOC only when crop residues were not removed. However, intercropping PCG with KC increased soil C over time, which may be attributed to greater root inputs than other N fertilization rates. Cong et al. 34 conducted a study in China for seven years in a field experiment that compared corn grown as a sole crop with corn and faba beans growing as an intercropping system and reported that C sequestration rate was significantly higher in intercrop than in the sole crop. Higher TN under higher N rate and PCG-KC mixture in this study were attributed to N availability by the former 36 and through N-fixation by the latter 37 . This increase in SOC over time and TN content under KC intercropping in this study emphases the fact that intercropping KC with PCG may be one of the effective ways to improve marginal land soils.
Effects of N rates and kura clover intercropping on biomass yield. Higher biomass under higher N fertilization rates and KC intercropping was probably a result of increased N availability from these two management systems. This suggests that intercropping KC with PCG increased the N availability to PCG through N fixation 12 and resulting in increased grass biomass yield. The PCG-KC had higher soil TN than the PCG-0N which can confirm the N availability to PCG through kura clover intercropping. These results concur with the findings of Gulwa et al. 38 , who reported that a mixture of red clover (Trifolium pretense) and native grasses in South Africa produced more yield in comparison with the grass only. Similarly, Lee et al. 39 reported that 120 kg N ha −1 increased the biomass of Miscanthus × giganteus in Illinois compared to the none fertilized soils. This improvement in PCG biomass due to its intercropping with KC can provide more feedstock for the bioenergy from marginal lands, which in turn can help in lowering our dependency on fossil fuels.
Trends and cumulative GHG fluxes as affected by N rates and kura clover intercropping. Higher N fertilizer rate and KC intercropping, which increased biomass yield in 2016 and 2018 (Table 1), were expected to promote daily CO 2 fluxes during the active crop growth due to higher root and microbial respiration compared to lower N fertilizers rate; however, no increase in the daily CO 2 fluxes due to these two managements were observed in these two years. Carbon dioxide fluxes varied from one year to another in this study, which may have caused the lack effect of N fertilization rates and KC intercropping on the cumulative CO 2 fluxes. Other researchers also reported that N fertilization rates and intercropping of legumes did not affect cumulative CO 2 fluxes. For example, Nikièma et al. 25 and Li et al. 4 reported that N fertilization rates did not affect CO 2 emission compared to the unfertilized soil in switchgrass, camelina and carinata. Shen et al. 18 reported that intercropping corn with soybean in China did not affect CO 2 emissions compared to the corn alone. This non significant impact due to treatments on cumulative CO 2 fluxes was also attributed to the absence effects of N fertilizer rate and kura clover intercropping on the measured soil properties (pH, EC and SOC). For example, Abagandura et al. 40 reported that microbial activity which plays a major role in decomposing soil organic matter and CO 2 emissions could be affected by soil pH. Increasing soil EC due to management practices can decrease microbial activity resulting in a reduction in soil respiration 41 . The change in SOC can influence soil respiration as reported by Jin et al. 42 . In addition, soil temperature and water content were similar under all N fertilizer rates and intercropping KC, which may cause similar CO 2 flux from these managements in this study 43 .
The experimental field received lower precipitation in 2014 and 2015 compared to 2016 and 2018 (Table 4), which may have caused lower CH 4 in the former than latter 40 . Higher soil water content can increase CH 4 emission as reported by Mbonimpa et al. 24 . Watts et al. 44 reported lower soil CH 4 fluxes under aerobic soil conditions. Similar soil water content in this reported study may have resulted in similar cumulative CH 4 fluxes among treatments. A study by 4 also reported no significant effect of N fertilizer rates on soil CH 4 fluxes compared with the unfertilized soils in the dryland cropping system. Shen et al. 18 reported that intercropping corn with soybean in China did not affect CH 4 fluxes compared to corn alone. However, Sainju et al. 19 reported that CH 4 Table 3. Global warming potential for CH 4  www.nature.com/scientificreports www.nature.com/scientificreports/ reduced with N fertilizer and legumes, which was attributed to greater root growth under these two managements, which can result in CH 4 absorption.

flux was
Increased cumulative N 2 O fluxes under higher N fertilizer rates compared to lower and unfertilized PCG, and PCG-KC was attributed to increasing in the N availability substrate under higher rates enahncing soil microbes activity, thereby increasing N 2 O flux 25,45 . Intercropping KC with PCG decreased cumulative N 2 O flux compared to the higher N fertilizer rates, probably due to N fixation within the KC nodules in the PCG-KC and thus is not freely available in the soil compared to higher N addition to PCG. These finding supported the results reported by Senbayram et al. 46 , who found that intercropping wheat with faba bean reduced cumulative N 2 O emissions compared with N-fertilized wheat. In addition, higher particulate N was found in PCG-225N and PCG-125N than in PCG-KC, PCG-0N and PCG-75N in 2018 (data not shown). This labile N can increase mineralization through enhancement of microbial activity, thus resulting in an increase in N 2 O flux 47 . Effects of N rates and kura clover intercropping on global warming potential. The GWP CH4 was similar under N rates and KC intercropping, which can be attributed to the lack effect of these treatments on CH 4 flux. Higher GWP N2O under higher N rate in this study was attributed to the availability of N substrate under higher N rates increasing microbial N 2 O production 45 as discussed above. Because fluxes of CH 4 was small in this study, the major part of net GWP resulted from GWP N2O. Dryland cropping system always emits a small quantity of CH 4 33,35 . Higher GWP N2O under higher N rates in this study had resulted in higher net GWP from these treatments compared to PCG-KC. In summary, intercropping KC with PCG decreased N 2 O flux and resulting net GWP. This finding concludes that such management may decrease the concern of GHG emissions associated with bioenergy crops.
Management implementation potential and its implication on the environment. Because fossil fuels are non-renewable and negatively impact the environment, the interest of using sustainable sources grown on marginal lands has increased worldwide. The total area of marginal land in the US varies significantly from one study to another 48 . According to Boe et al. 7 , there is more than 210,000 ha in South Dakota that are classified as wetlands and are not suitable for conventional crop production. Prairie cordgrass, which is mostly found in flooded soils in the Northeast and Midwest of the US, could potentially be used as a biofuel feedstock on the marginal lands. One of the challenges in prairie cordgrass cropping systems is N fertilizer management. Although it is accepted that legumes can provide N to the soil, there is no study evaluating how intercropping a legume like kura clover with prairie cordgrass in marginal lands can affect prairie cordgrass yield and GHG emissions. This is the first study evaluating how intercropping kura clover with prairie cordgrass can affect grass biomass, soil properties, GHG emissions and GWP. Results from this study suggested that intercropping kura clover with prairie cordgrass increased total N, SOC over time, increased the yield of prairie cordgrass, and mitigated GHG emissions from marginal land cropping systems. Kura clover and prairie cordgrass intercropping in marginal lands not only improves soil and allows such lands to be productive; it also decreases the concern from bioenergy regarding the use of edible crops and GHG emissions. Considering the environmental damage from fossil fuels and N fertilizers, utilizing prairie cordgrass/kura clover mixtures for bioenergy may mitigate GHG emissions by reducing N fertilizer requirements and improve soil and environment quality.

conclusion
This study evaluated the response of soil surface GHG fluxes and resulting GWP and prairie cordgrass to N fertilization rates and intercropping kura clover with prairie cordgrass. Soil data for 2011 and 2018 showed that N fertilization rates and intercropping K did not affect soil pH, EC and SOC. However, PCG-KC increased total N and stored more C compared to the lower and unfertilized PCG in 2018. The PCG-KC produced a higher yield  www.nature.com/scientificreports www.nature.com/scientificreports/ than the unfertilized PCG, and was similar to the fertilized PCG in 2016 and 2018. The cumulative CO 2 and CH 4 fluxes for 4 years (2014, 2015, 2016 and 2018) were not influenced by the application of N fertilization rates and PCG-KC. However, the cumulative N 2 O fluxes and net GWP for 4 years increased under the higher N fertilization rate compared to the PCG-KC. In general, this study suggests that intercropping kura clover with prairie cordgrass has a good potential for mitigating N 2 O emissions and global warming potential.

Materials and Methods
Site description and treatment details. The  Seeds of PCG and KC were obtained from a natural population in South Dakota and Ag Research New Zealand, respectively. Seedlings of PCG and KC were grown individually in containers (Stuewe, Inc, Corvallis, OR) during early spring 2010 in the greenhouse and transplanted in the field in later spring at the same year. The size of each field plot was 3.0 m wide and 5.7 m long. There have been nine rows in each plot. Kura clover seedling was transplanted on 30 cm centers within rows in the field. The density of K was 111,111 plants ha −1 . After that, PCG seedlings (intercropping and monoculture) were transplanted in the field on 60 cm centers, with a density of 26,896 plants ha −1 . In October 2015, 2016, and 2018, the yield was determined for all plots. From each plot, all biomass was harvested using a sickle-bar mower. To measure the grass fresh weight, all fresh harvested biomass was weighed, and then biomass subsamples from each plot were dried at 60 °C for 72 h to measure the grass dry weight.
Soil analyses. Soil samples were taken in 2011 and in 2018 using a core with a diameter of 3.2-cm. Seven soil samples from each plot at a depth of 0-15 cm were taken from random spots. These soil samples were composited to represent one sample for each plot. Soil pH and electrical conductivity (EC) were measured using pH and EC meter. Dry combustion was used to measure SOC and TN. To remove inorganic C from SOC samples before combustion, the soils were pretreated with acid.
Soil GHG sampling. Soil GHG fluxes were measured every week from May to October in 2014, 2015, 2016, and to 2018. Static chamber technique was used for GHG measurements. Details about the GHG measurements can be found in our previous papers 4, 40 . Briefly, PVC chambers (one chamber in each plot) were installed to measure GHG fluxes. During GHG sampling, PVC lids were used to cover the installed chambers. To ensure no gases were released from the chambers; the lids closed the chambers firmly 50 . GHG sampling was measured at an interval time of 0, 20, and 40-minute using a syringe to extract these gases 28 . Gas samples were transferred to the lab, and GHG (CO 2 , CH 4 , and N 2 O) concentrations were analyzed using a gas chromatograph (GC-2014, Shimadzu Co., Ltd., Japan). Soil GHG fluxes were calculated as the change in headspace gas concentration over time within www.nature.com/scientificreports www.nature.com/scientificreports/ the enclosed chamber volume 42 . Cumulative fluxes were calculated using linear interpolation 40 . Global warming potential (GWP) in CO 2 -equivalent was estimated using radiative forcing of 296 and 23 for N 2 O and CH 4 , respectively, and because stock of SOC changes was not evaluated in this study, GWP for CO 2 was not calculated 48 . The summation of GWP for CO 2 and N 2 O (Net GWP) was also determined.
At the time of GHG measurements, soil water content and temperature were also measured. Climatic data (minimum and maximum air temperature and precipitation) in 2014, 2015, 2016, and 2018 growing seasons were obtained from Northeast Regional Climate Center 51 and are shown in Table 4. Total precipitation in the 2016 growing season (459.7 mm) was 277% higher than in 2014 (121.8 mm), 315% higher than in 2015 (110.7 mm), and 47% higher than in 2018 (313.0 mm) ( Table 4). The trend of air temperature was similar in all years, with the beginning of the growing season having higher air temperature than the middle and the end of the season (Table 4).

Statistical analyses.
Repeated measures analysis in PROC MIXED were used to analyze daily, and cumulative GHG fluxes, with treatments and replications, were considered as fixed and random variables, respectively. Sampling date in daily GHG fluxes and year in cumulative fluxes were considered as repeated measure variables. For treatment comparisons, Tukey's test was used at 0.05 probability. Multiple regression (soil temperature and soil water content as explanatory variables and CO 2 and N 2 O fluxes as dependent variables) was conducted in Sigma Plot 14.0 (Fig. 5).