Introduction
Agriculture faces the challenge of producing more food with fewer inputs, while simultaneously addressing problems such as soil degradation, loss of biodiversity and unpredictable weather due to climate change (Harvey and Pilgrim, Reference Harvey and Pilgrim2011). Governments also have elevated expectations that agriculture should provide additional ecosystem services such as biomass for sustainable bioenergy production and climate change mitigation (Tilman et al., Reference Tilman, Socolow, Foley, Hill, Larson, Lynd, Pacala, Reilly, Searchinger and Somerville2009; Harvey and Pilgrim, Reference Harvey and Pilgrim2011; Sapp et al., Reference Sapp, Harrison, Hany, Charlton and Thwaites2015). These challenges call for a focus on eco-functional intensification and multifunctionality, i.e., increased efficiency of natural resource use, improved nutrient-cycling techniques and agro-ecological methods for protecting and possibly enhancing biodiversity (Halberg et al., Reference Halberg, Panneerselvam and Treyer2015; Jensen et al., Reference Jensen, Bedoussac, Carlsson, Journet, Justes and Hauggaard-Nielsen2015). A well-planned production system with functional diversity of crops within the field and over the cropping season has the potential to improve the outcome of several of these challenges (Drinkwater and Snapp, Reference Drinkwater and Snapp2007; Niggli et al., Reference Niggli, Slabe, Schmid, Halberg and Schlüter2008; Doré et al., Reference Doré, Makowski, Malézieux, Munier-Jolain, Tchamitchian and Tittonell2011).
Nitrogen (N) is often the most limiting nutrient for crop performance in terms of yield and quality, but can also be a major contributor to pollution of drinking water, eutrophication of surface water and pollution of the atmosphere with the potent greenhouse gas nitrous oxide (N2O) (Baggs et al., Reference Baggs, Rees, Castle, Scott, Smith and Vinten2002; MEA, 2005; Galloway et al., Reference Galloway, Townsend, Erisman, Bekunda, Cai, Freney, Martinelli, Seitzinger and Sutton2008; Foley et al., Reference Foley, Ramankutty, Brauman, Cassidy, Gerber, Johnston, Mueller, O'Connell, Ray, West, Balzer, Bennett, Carpenter, Hill, Monfreda, Polasky, Rockstrom, Sheehan, Siebert, Tilman and Zaks2011; Cohen, Reference Cohen2015). Increased levels of N in natural or semi-natural ecosystems also lead to a reduction in biodiversity (Zillén et al., Reference Zillén, Conley, Andrén, Andrén and Björck2008; Sutton et al., Reference Sutton, Oenema, Erisman, Leip, van Grinsven and Winiwarter2011). Regardless of whether the N is fixed industrially or biologically by legumes, the fixation contributes to the availability of reactive N. Excessive inputs of reactive N lead to disequilibrium of the planetary N cycle and thereby to detrimental effects on ecosystems (Rockström et al., Reference Rockström, Steffen, Noone, Persson, Chapin, Lambin, Lenton, Scheffer, Folke and Schellnhuber2009). Improved retention and recycling of N is, and should continue to be, a highly prioritized goal of policy makers, advisors and farmers (Steffen et al., Reference Steffen, Richardson, Rockström, Cornell, Fetzer, Bennett, Biggs, Carpenter, de Vries and de Wit2015). It is common that farmers supply N in stockless organic systems by including green-manure crops based on N2-fixing legumes (Watson et al., Reference Watson, Atkinson, Gosling, Jackson and Rayns2002). A disadvantage is that growing green manures reduces the amount of land available for food crops. There may also be a high risk of N losses through ammonia (NH3) and N2O volatilization, and/or nitrate (NO3−) leaching, depending on incorporation time and technique (Li, Reference Li2015). Another N supply option is to grow grain legumes for food production, but the organic N left in the field after grain harvest is often not sufficient to cover the needs of the succeeding non-legume crop (Beck et al., Reference Beck, Wery, Saxena and Ayadi1991; Jensen, Reference Jensen1997). Roots with nodules left in the field or additional residual biomass may nevertheless be a valuable addition to soil N.
The harvest of ensiled or anaerobically digested biomass permits target-oriented application of organic nutrients, to fertilize crops with the highest nutrient requirements (Möller and Müller, Reference Möller and Müller2012). The biogas (bio-methane) produced via anaerobic digestion can be used on the farm, or sold to the market. Generally, a larger proportion of the total N is present as mineral N and the C/N ratio is lower in the digestate obtained after anaerobic digestion compared with in fresh or ensiled biomass (Gutser et al., Reference Gutser, Ebertseder, Weber, Schraml and Schmidhalter2005). This is because the bacterial digestion of organic matter results in release of C, mainly as methane (CH4) but also CO2, while most of the organic N is converted to ammonium (NH4+), which remains in the digestate (Möller and Müller, Reference Möller and Müller2012). Several studies have observed an increased yield of cereals fertilized with plant-based digestate compared with un-digested feedstock (Stinner et al., Reference Stinner, Moller and Leithold2008; Frøseth et al., Reference Frøseth, Bakken, Bleken, Riley, Pommeresche, Thorup-Kristensen and Hansen2014). On the other hand, Gunnarsson (Reference Gunnarsson2012) reports a lack of yield increase or even a decreased vegetable yield in response to fertilization with digestate, as compared with undigested biomass harvested from a green-manure ley (Gunnarsson, Reference Gunnarsson2012). The availability of N in biomass and digestate for crop N acquisition also depends on mineralization and immobilization dynamics, which in turn are influenced by many factors such as C/N ratio, temperature and moisture (Trinsoutrot et al., Reference Trinsoutrot, Recous, Bentz, Lineres, Cheneby and Nicolardot2000; Nicolardot et al., Reference Nicolardot, Recous and Mary2001; Cabrera et al., Reference Cabrera, Kissel and Vigil2005). If the mineralization is delayed, the application of biomass or digestate to a few crops in the cropping system can also be expected to increase the biomass yield and N accumulation in cover crops growing after the fertilized main crops (Kumar and Goh, Reference Kumar and Goh2002; Peoples et al., Reference Peoples, Hauggaard-Nielsen and Jensen2009).
The aim of this study was to compare three methods for strategic recycling and application of residual and green-manure biomass N in terms of yield and N concentration of the edible fraction of food crops in an organic stockless cropping system. The crop response after leaving residual biomass resources in situ compared with redistributing the same biomass resources after ensiling or ensiling plus anaerobic digestion was evaluated in a crop rotation. Our main hypotheses were that (1) strategic recycling of the digestate from anaerobic digestion of biomass leads to higher yield of winter rye, white cabbage and red beet, due to a higher concentration of plant-available N in the digestate compared with strategic redistribution of ensiled biomass or in situ incorporation; (2) concentration of N in the edible plant parts of winter rye, white cabbage and red beet increases with strategic recycling of digestate, due to a higher concentration of plant-available N in the digestate compared with biomass redistribution and in situ incorporation; and (3) strategic recycling of ensiled or digestate biomass increases the biomass production of the cover crops following a main crop receiving biomass, compared with after in situ incorporation of biomass. The reason for the third hypothesis is that the targeted addition of a large quantity of silage or digestate will increase the N availability also for the cover crops following the fertilized crops.
Materials and Methods
Study site and soil
The experiment was established in 2012 at the Swedish University of Agricultural Sciences in Alnarp, Sweden (55°39′21″N, 13°03′30″E), on the SITES (Swedish Infrastructure for Ecosystem Science) field research station in Lönnstorp on a sandy loam soil (Table 1) characterized as an Arenosol (Deckers et al., Reference Deckers, Nachtergaele and Spaargaren1998). The land has been organically certified since 1993 and the preceding crop was a 1-yr legume-grass ley. Soil nutrient availability and particle distribution was analyzed at the start of the experiment (Table 1) by a commercial soil analysis laboratory (LMI, Helsingborg, Sweden) using the modified Spurway Lawton method (extraction in 0.1% acetic acid) (Spurway and Lawton, Reference Spurway and Lawton1949).
Table 1. Soil characteristics in the upper 0–30-cm soil layer and the lower 30–60-cm in March 2012.
Climatic data
The region has a typical northern-European maritime climate with mild winter and summer temperatures. Lowest and highest monthly mean temperature and monthly precipitation data from the 3 yr of the field experiment are presented in Figure 1. The 30-yr (1961–1990) average for annual temperature and total annual precipitation were 7.9°C and 666 mm, respectively, measured at the weather station in Lund (55°43′N, 13°12′E). The temperature and precipitation in 2012–2015 were close to the average for the region, except for unusually high temperatures during November to February in 2013–2014 and high rainfall in August 2014 (Fig. 1).
Figure 1. Mean of minimum (light gray line) and maximum (dark gray line) monthly temperatures and monthly accumulated precipitation (histogram) during the field experiment. The data were retrieved from a weather station LantMet, Alnarp (55°40′N, 13°6′E).
Crop rotation
A 6-yr crop rotation was used for the study (Fig. 2), although the experiment was only performed during the three full seasons in 2012–2015 (Fig. 3). Within each treatment and block, the crop rotation was established in six separate plots, so that each of the six main crops in the rotation was grown during each year of the experiment. Since the experiment started in spring 2012 without any autumn-sown crop from the previous year, winter rye (Secale cereale L.) was replaced by spring barley (Hordeum vulgare L.) during the first year.
Figure 2. Crop rotation that was used for the 3 yr crop sequence. The main crops are marked with a circle and the cover crops or overwintering crops as an arrow.
Figure 3. Three-year sequence of the crop rotation investigated in the study, with all six main crops present each year from 2012 to 2014. The light grey sections represent uncultivated stubble after cover crop harvest and the dark grey sections represent a short black fallow after red beet harvest. The vertical arrows show the harvests of residual (food crops) and total aboveground (cover crops and green-manure ley) biomass for redistribution in the subsequent year in the BR and AD treatments. In IS, the same biomass resources were left in situ in the same field plot as they had been growing.
The crops included in the rotation (Table 2) were chosen to optimize several functions, namely the production of food crops, provision of biomass resources for internal recycling of nutrients, biological N2 fixation, weed suppression and enhancing the presence of beneficial insects. The rotation therefore included crops with different functional traits, such as fast stem elongation, variation of leaf architecture, nectar-rich flowers, rapid root growth and efficient nutrient acquisition. The cropping system also included several different crop-management strategies in accordance with the principles of organic agriculture, i.e., hoeing in row crops and frequent cutting of the ley to reduce pest and weed pressures.
Table 2. The components of the crop rotation with main and cover crops.
Intercrops contained legumes to provide symbiotic N2 fixation, promote soil N availability and produce food crops with high-protein concentration. The pea (Pisum sativum L.)/barley and lentil (Lens culinaris Medik)/oat (Avena sativa L.) intercrops were selected, since mixtures with legumes and cereals have been shown to enhance resource use efficiency and reduce weed abundance compared with legume sole crops (Hauggaard-Nielsen et al., Reference Hauggaard-Nielsen, Jørnsgaard, Kinane and Jensen2008; Bedoussac et al., Reference Bedoussac, Journet, Hauggaard-Nielsen, Naudin, Corre-Hellou, Prieur, Jensen and Justes2014). A replacement design (De Wit and Van den Bergh, Reference De Wit and Van den Bergh1965) was employed with the ratio 80/20 for pea/barley and 90/10 for lentil/oat. Winter rye was included in the rotation since it competes well with weeds, retains N and reduces the risk of soil erosion. Row crops [red beet (Beta vulgaris L.) and cabbage (Brassica oleracea L.)] were included during two of 6 yr in the rotation, as examples of high-value food crops that also enable efficient mechanical reduction of weeds between the rows. The six species included in the ley were chosen to add diversity for resilience of biomass production, N2 fixation and provide a food source for beneficial insects. The composition followed a replacement design with 16.7% of recommended sowing density for each species. Each main crop was followed by an autumn- or winter-growing main or cover crop in order to reduce N leaching, reduce weeds and produce biomass during the autumn or winter season. Oilseed radish (Raphanus sativus L.) and lacy phacelia (Phacelia tanacetifolia Beneth) were selected as cover crops for three reasons: they have a high NO3− uptake (Thorup-Kristensen, Reference Thorup-Kristensen2001), oilseed radish has shown partial resistance to clubroot (Plasmodiophora brassicae) (Diederichsen et al., Reference Diederichsen, Frauen, Linders, Hatakeyama and Hirai2009), and lacy phacelia is a valuable food source for beneficial insects such as parasitic wasps and bees (Araj and Wratten, Reference Araj and Wratten2015; Barbir et al., Reference Barbir, Badenes-Pérez, Fernández-Quintanilla and Dorado2015). Both cover crops were grown in combination with buckwheat (Fagopyrum esculentum Moench) (50% of each species’ recommended sowing density) in order to further provide resources for beneficial insects, and since it has been indicated that buckwheat produces compounds that can limit the growth of weeds (Kalinova et al., Reference Kalinova, Vrchotova and Triska2007). The mixture of perennial ryegrass (Lolium perenne L.), red clover (Trifolium pratense L.) and white clover (Trifolium repens L.) was used as a cover crop growing during autumn, winter and spring since these crops can improve soil structure (Breland, Reference Breland1995) and retain NO3− (Askegaard et al., Reference Askegaard, Olesen, Rasmussen and Kristensen2011). The sowing densities of ryegrass, red clover and white clover in this mixture were 73/15/12% of the recommended density for each species as sole crop.
Experimental design
The field experiment comprised in total 72 experimental plots measuring 3 m × 6 m, distributed in four replicate blocks. The experiment started by establishing each of the six main crops, which were followed by cover crops and main crops according to the designed crop rotation (Fig. 2) in the same physical plots during the two subsequent years, thereby providing a 3-yr crop sequence with all six crops present each year (Fig. 3). Within each block, 18 individual plots (six main crops × three treatments) were randomly assigned to one of the following biomass-management treatments applied at the cropping system level, i.e., consistently throughout the 3-yr crop sequence:
IS—in situ incorporation of biomass resources (crop residues, cover crops and green-manure ley), i.e., leaving the biomass after harvest in the same plot as they were grown.
BR—biomass redistribution: storing the biomass resources as silage and redistributing them to cabbage, red beet and rye growing in the same system in the following year.
AD—anaerobic digestion of the biomass resources (after storing them as silage) and redistributing the digestate to cabbage, red beet and rye growing in the same system in the following year.
The residual biomass comprised straw from grain legumes and cereals, leaves from cabbage and red beets, and all aboveground biomass of cover crops. The green manure consisted of ley, from which aboveground biomass was harvested four times. The IS treatment differed from BR and AD already during the first year (2012), since biomass resources were left in situ instead of being removed from the plot, and redistributed in the next year as silage in BR and digested silage in AD. In contrast, the distinction between BR and AD did not start until the second year (2013), when the non-legume crops were fertilized either with silage (BR) or digestate (AD). The May cuttings of the green-manure ley and the ryegrass/clover were stored together with the other residual biomass sources harvested later in the growing season, and redistributed in the following year.
The distribution of N in BR and AD was based on the strategy to use all available biomass resources for redistributing N to the non-legume main crops within the cropping system, in proportions that reflected national recommendations for N fertilization of rye, cabbage and red beet, respectively. Total N content of biomass was measured in subplot samples for each treatment and used to estimate total N in the residual and green-manure biomass (Table 3). The total N content, i.e., the sum of all biomass resources, was similar for the three treatments in 2013, while in 2014, the AD treatment resulted in a lower amount of N applied than in the IS and BR treatments. The differences in total N between AD biomass and AD digestate mean that there have been losses of N during handling of biomass, silage and digestate in the AD treatment. Losses of N from the IS and BR systems were not quantified.
Table 3. Total N content in the residual and green-manure biomass from the previous year, redistributed to rye, cabbage and red beet in the BR and AD treatments and applied in situ at harvest in the IS treatment (kg ha−1).
1 Refers to yield from 6 ha.
Crop management
All crops were sown with a density based on national recommendations in organic farming (Table 2). The row spacing for winter rye was 12.5 cm in 2012 and doubled to 25 cm in 2013 to facilitate spreading the biomass and digestate in the rows. Red beet and cabbage were sown and planted with a row spacing of 50 cm. The variety of red beet was changed from the monogerm type ‘Alvro mono’ in 2012 and 2013 to the multigerm variety ‘Kestrel’ in 2014. The cabbage plants were mechanically transplanted in rows with 50 cm apart and irrigated to assure the establishment of the plants, in order to simulate a large-scale production farm. In 2012, six rows were sown and planted in each plot of red beet and cabbage. They were reduced to five rows in 2013 and 2014, since plants in the border rows were severely stunted in 2012. The green-manure ley and the clover/ryegrass catch crop were undersown in their respective main crops (Table 2) at the same time as the main crop.
At the start of the experiment in spring 2012, the previous crop (ley) was ploughed, and the soil was harrowed twice over two consecutive weeks to control weeds. Subsequent soil management was made with non-inversion tillage (2013 and 2014). At the time of establishment in 2012 (not repeated in the following years), the entire field was fertilized with digestate from a stockless organic farm with biogas production. The digestate (containing 7.1 kg total-N Mg−1 digestate, 5.4 kg NH4+-N Mg−1, 1.3 kg P Mg−1 and 1.7 kg K Mg−1) was applied at a rate of approximately 16 Mg digestate ha−1, to achieve 115 kg N ha−1. The digestate was applied with a 20-m wide boom that had trailing hoses.
The weeds in the row crops were controlled by hand hoeing during each growing season. Winter rye was sown in late September/early October, after red beet harvest in late August. During this short fallow period, the soil was tilled when the weeds emerged and again a few weeks later. No weed control was used in the intercrops or cover crops. The cabbage was covered with an insect net (0.8 mm × 0.8 mm mesh). Hand spraying of Bacillus thuringiensis with knapsack spraying equipment occurred in 2013 and 2014 as an organic pest control measurement of Lepidoptera species. The spraying started at the observation of the larvae on the crop and was repeated two times with an interval of 2 weeks.
Anaerobic digestion and application of biomass resources
The anaerobic digestion of biomass resources in the AD treatment was made using a mesophilic leach bed reactor at the Annenberg research facility (Biotechnology, Lund University, Sweden). In this type of batch reactor, solids are hydrolyzed by adding and circulating water over the biomass (Lehtomäki et al., Reference Lehtomäki, Huttunen, Lehtinen and Rintala2008). Recirculation of the liquid stimulates the microbial digestion of the biomass, due to the continuous redistribution of inoculum, nutrients and dissolved organic matter (Chanakya et al., Reference Chanakya, Venkatsubramaniyam and Modak1997; Lissens et al., Reference Lissens, Vandevivere, De Baere, Biey and Verstraete2001). The silage feedstock in our study had a dry matter content of 24% in both years and was not pre-treated in any other way than mixing the pile of silage with a tractor-carried shovel before loading it into the reactor. The digestion was allowed to run for 2 months in early spring in both 2012–2013 and 2013–2014. The resulting digestate was delivered in a liquid and solid phase (Table 4). The mean C/N ratio of the pooled digestate (liquid + solid) was 12 and 14, in 2013 and 2014, respectively (Table 4). The total N concentration in the pooled digestate was 1.1 kg Mg−1 (fresh weight) in both years, and the NH4+-N concentration of total N was 25% in 2013 and 16% in 2014.
Table 4. Composition of digestate produced from residual and green-manure biomass in the studied cropping system in 2013 (from anaerobic digestion of biomass resources harvested in 2012) and 2014 (from anaerobic digestion of biomass resources harvested in 2013).
Standard deviation of 2–3 samples is presented within brackets. Data are based on fresh weight analyses.
The aim of the study was to measure the effect of redistributing the entire residual and green-manure biomass resource, and thus the total amount of biomass or digestate was divided in specific ratios to the non-legume crops in BR and AD, respectively. The application rate to the different crops (Table 3) was based on a discussion with advisors in organic farming. There was a delay in N analysis of some crops, which made it necessary to make estimates of concentration of N in the BR silage based on the previous year, with the aim of providing the same ratio in total N supply in both BR and AD. The solid phase of the digestate was mixed on a tarpaulin and weighed to achieve the right amount per crop according to defined proportions. The liquid phase was carefully stirred and then measured by volume in watering cans, according to the same proportions as the solid fraction, adding liquid on top of the distributed solid digestate. In the red beet and cabbage plots, applied digestate was incorporated into the soil by non-inversion tillage machinery before planting and sowing. The plants of winter rye had grown too tall to incorporate the digestate with machinery, and it was therefore banded on the soil surface between the rows.
Sampling and harvest
Immediately before crop harvest, samples for analyses of yield and crop quality were obtained by sampling subplots in each main crop. The samples of cereals, legumes and grasses were harvested from an area of 0.25 m2 at a position approximately 1 m from the northern side of each plot. The crops were cut 5 cm above the soil surface and divided in legumes and non-legumes before drying and milling. Samples were dried at 70°C for 24–72 h, depending on water content. The grain legumes and cereal grains were hand-separated from straw. The red beet was sampled by harvesting all the plants from 2 m in a centrally located row, followed by separation of beet roots from leaves by hand. The beet roots were rinsed with water and allowed to dry in room temperature for 30 min before being counted and weighed. A subsample consisting of two small, two large and one medium beet root, each cut in half (discarding one-half of each beet), was dried and milled. This sampling method was chosen to get a representative nutrient subsample from the core to the skin from beets of different sizes. Four adjacent cabbages in a central row were harvested for analysis of the weight of the residue and edible fraction. The edible fraction was defined as a tight smooth head, and the rest of the plant was defined as residue. A 1-cm thick slice was cut all the way into the core as a subsample from all four heads. The sample was weighed, dried and milled.
The crops and biomass resources used for digestion and redistribution were harvested on the entire area of each plot (after subsampling for analyses, as described above) with methods that mimicked commercial farming practices as far as possible. Ley and cover crops were cut with a large-scale lawn mower and the harvest from each plot was collected and weighed in bags. Grain legume/cereal intercrops were harvested with a Sampo Rosenlew plot combine harvester with a bag collecting the straw from each plot for weighing. Red beet leaves and cabbage residues (the outermost layer of leaves) were separated from the beets and heads, and weighed in the field prior to ensiling the residues. The biomass in the BR and AD treatments was collected in separate 1-m3 plastic containers, where it was compressed and covered with a tarpaulin and four 15-kg sandbags. The first biomass was collected in May and the last in October. The cuttings from the May harvest of the green-manure ley and ryegrass/clover cover crop were ensiled, and also digested in AD, in preparation for application in the next growing season (Fig. 3).
The green-manure ley was harvested once in August in 2012, as it was established the same spring, and the yield was expected to be low compared with if the ley is established the previous year by undersowing in a main crop. The second harvest was in May 2013 before tilling and establishing the next crop. The green-manure ley undersown in pea/barley in 2012 was harvested at three consecutive occasions in 2013: in June, July and September, with an additional harvest occasion in May 2014 before soil tilling. Similarly, the green-manure ley undersown to pea/barley in 2013 was harvested at three occasions in 2014 (June, August and September) plus a fourth occasion in May 2015. The grain legumes and cereals were harvested when they were mature, while the harvest of cabbage and red beet was based on optimal timing for yield and quality, but also so that there was sufficient time for establishment and growth of cover and winter crops before the onset of winter. All biomass resources were weighed (total fresh weight per plot) before ensiling, and subsamples were used for analyses of dry matter concentration.
Calculations and statistics
The effect of the different biomass-management systems was measured in terms of yield (food fraction and straw/residual leaves), with the intercrops separated into legumes and non-legumes. Nitrogen concentration in the edible fraction of the crops was measured as a quality parameter, using an elemental analyzer (PDZ Europe ANCA-GSL for the intercrops and Flash 2000, Thermo Scientific for rye, cabbage and red beet). The data were analyzed with a general linear model and Tukey's post hoc analysis at a 5% significance level using the software Minitab 16.
Results
Yield and N concentration of rye, cabbage and red beet
The yield of the edible fraction of rye, cabbage and red beet neither show any statistically significant difference in yield between treatments (Table 5), nor did the treatments result in different concentrations of N in the edible fraction of rye, cabbage and red beet (Table 6) or yield of biomass residue (Table 7).
Table 5. Yield of edible fraction (dry weight), presented as average ± standard error of the mean (n = 4).
IS, in situ incorporation; BR, biomass redistributed to the non-leguminous crops; AD, digested biomass distributed to the non-leguminous crops; NA, data not available.
Intercrops are shown both as total and separate as IC component yields. Means that do not share a letter within a row and year are significantly different. Bold indicates year and crop with significant effect of biomass treatment.
Table 6. Nitrogen concentration (%) of the edible fraction of the crops, presented as average ± standard error of the mean (n = 4).
IS, in situ incorporation; BR, biomass redistributed to the non-leguminous crops grown in pure stand; AD, digested biomass distributed to the non-leguminous crops grown in pure stand.
Table 7. Yield (dry weight) from crop residues, green-manure ley and cover crops presented as average ± standard error of the mean (n = 4, * = n = 3).
IS, in situ incorporation; BR, biomass redistributed to the non-leguminous crops grown in pure stand; AD, digested biomass distributed to the non-leguminous crops grown in pure stand.
Italic numbers represent fractions in intercrops (IC) and species mixtures (green-manure ley and ryegrass/clover). Means that do not share a letter within the same row and year are significantly different. Bold indicates year and crop with significant effect of biomass treatment. The sum of biomass presented at the bottom of the table represent the total amount of biomass resources that would be available if all main crops and associated cover crops were cultivated on 1 ha each.
Yield and N concentration of the intercrops lentil/oat and pea/barley
The lentil grain yield was significantly lower in IS compared with BR in 2013 (Table 5). Data are not available for the grain yield of pea and barley intercrop in 2013, since the crop was severely damaged by rabbits and hares that year. The biomass treatments did not result in any significant difference in the N concentration of grain legume or cereal seeds. The IS treatment resulted in significantly higher yields of oat straw in both years (Table 7).
Yield of cover crops and green-manure ley
The yield of buckwheat/lacy phacelia (grown after rye) was significantly higher in BR compared with IS and AD in both years (Table 7). The redistributed biomass (BR and AD) had no carry-over effect on the other cover crops. The clover proportion of the ryegrass/clover cover crop was exceptionally low in general for all the treatments at harvest in 2013. The legume proportion of the green-manure ley was significantly higher in the BR and AD treatments compared with IS in 2014.
Discussion
As compared with the IS treatment, removal of biomass (AD and BR) resulted in a shift in legume/non-legume proportions in several of the crop mixtures, i.e., higher lentil grain yield in 2013, lower oat straw biomass in both years and higher legume yields in the green-manure ley in 2014. This shift is most likely a result of the removal of N-rich biomass in treatments BR and AD compared with IS, leading to reduced N availability and thereby a lower competitive ability of the oat in the intercrop and the grasses in the ley. Our results thereby confirm previous findings about the effect of N availability on legume/non-legume proportions in crop mixtures (Ledgard and Steele, Reference Ledgard and Steele1992; Jensen, Reference Jensen1996; Hejcman et al., Reference Hejcman, Szaková, Schellberg and Tlustoš2010).
We did not observe a significant effect of the biomass management on yields of rye, cabbage and red beet or the other crops, indicating that biomass removal and extraction of CH4 could be performed without a decrease in yields or N concentration of the food crops. However, the hypotheses that redistributing the digestate from anaerobic digestion of the biomass resources would have a positive effect on crop yields and crop N concentration had to be rejected. It is possible that the higher N availability in the digestate also led to higher losses of N through NH3 volatilization during handling and after field application of the digestate (pH value of digestate in this study: 7.2–7.4). In particular, it is likely that NH3 losses occurred in winter rye in the AD treatment, as it was not possible to inject the digestate into the soil or till after application (Möller and Stinner, Reference Möller and Stinner2009; Möller and Müller, Reference Möller and Müller2012), due to the advanced growth stage of the established crop. There may also have been losses of NH4+ via seepage from silage and when the silage was mixed prior to digestion. Potential N losses in the AD system might thus have counteracted the expected benefits of a higher N availability for crop N acquisition.
Our third hypothesis was supported by the result that the buckwheat/lacy phacelia following winter rye produced higher biomass yields in the BR than in the IS treatment, which could be explained by the higher addition of biomass N to the preceding main crop in BR than in IS. The fact that the corresponding yield effect was not observed in the main crop (rye) receiving the biomass N in BR indicates that the mineralization was delayed, and not in synchrony with the requirements of the main crop, but leading to an increased N availability for the subsequent cover crop. Moreover, the lack of a corresponding increase of the same cover crop biomass yield in the AD treatment, even in 2014 when rye in AD received more N than rye in BR (Table 3), implies that the N dynamics differ if the residual biomass is applied as silage or digestate. As discussed above, potentially higher NH3 emissions after field application of the digestate (Wulf et al., Reference Wulf, Maeting and Clemens2002) compared with silage may explain the lack of yield increase of cover crops in AD.
Since this study was based on recycling, all biomass resources obtained within the cropping system inclusion of N-poor biomasses such as cereal straw contributed to a relatively low total N concentration in the digestate. The digestate also contained water added during the digestion process, which diluted the nutrient concentration expressed on a fresh weight basis (1.1 kg N Mg−1). The C/N ratio (12–14) of our digestate was within the range (7–39) of other plant-based digestate (Möller et al., Reference Möller, Stinner, Deuker and Leithold2008; Gunnarsson et al., Reference Gunnarsson, Bengtsson and Caspersen2010; Gunnarsson et al., Reference Gunnarsson, Lindén and Gertsson2011; Frøseth et al., Reference Frøseth, Bakken, Bleken, Riley, Pommeresche, Thorup-Kristensen and Hansen2014). The NH4+-N proportion of total N (16–25%) and concentration of NH4+-N (0.18–0.27 kg NH4+-N Mg−1 fresh weight) in our digestate were also within, but at the lower end of the range of plant-based digestate from other studies (6–55% NH4+-N of total N; 0.18–1.52 kg NH4+-N Mg−1 fresh weight) (Möller et al., Reference Möller, Stinner, Deuker and Leithold2008; Gunnarsson et al., Reference Gunnarsson, Bengtsson and Caspersen2010; Gunnarsson et al., Reference Gunnarsson, Lindén and Gertsson2011). The relatively low concentration of NH4+-N in the digestate in our study indicates that the digestion of the biomass has not been efficient, which might in turn lead to a slow mineralization of the organic N in the soil. The chemical composition of the digestate depends both on the composition and pre-treatment of the feedstock, and the lack of pre-treatment (e.g., shredding) might also have contributed to a low concentration of NH4+-N in the digestate.
The biomass treatments did not result in different N concentrations in the food fraction of cereals or legumes. There was a trend of normal to high N concentrations in oat in this study, as compared with the mean concentration for the variety when it is grown in similar climate (Hagman et al., Reference Hagman, Halling and Dryler2014). The barley grain N concentration was on average in line with the critical optimum for desirable malting quality, i.e., <1.84% N (Bertholdsson, Reference Bertholdsson1999). The mean N concentration of the winter rye variety ‘Amilo’ in variety tests (Hagman et al., Reference Hagman, Halling and Dryler2014) is similar to results from our experiment in 2013, while all treatments resulted in lower N concentrations in 2014, indicating suboptimal N supply for rye in the second year of the experiment.
Strategic biomass management (BR and AD) maintained levels of food crop yields, with increased biomass production potential of cover crops and an increase in legume proportions in intercrops, green-manure and ryegrass/clover leys. An increased proportion of legume biomass in the green-manure ley is correlated with increased N inputs via N2 fixation (Evans et al., Reference Evans, O'connor, Turner, Coventry, Fettell, Mahoney, Armstrong and Walsgott1989; Carlsson and Huss-Danell, Reference Carlsson and Huss-Danell2003), which leads to a reduced need for external input of N to cover requirements of the following crop. This is essential in stockless organic agriculture as there are few economically viable options to supply N, when there is no access to animal manure.
The possibility of using AD as a treatment of residual and green-manure biomass without losses in yield and quality provides the opportunity of producing bioenergy as an additional source of energy or income for the farmer. Tuomisto and Helenius (Reference Tuomisto and Helenius2008) even argue that a slightly lower crop yield in a bioenergy scenario would be beneficial for the systems energy balance compared with leaving the biomass in situ.
Conclusions
Our results show that food, biomass for bioenergy carriers and digestate can be produced within the same cropping system without reductions in yield and N concentration of the food crops, relative to standard organic farming practices, e.g., green manuring and crop residue incorporation. Maintenance of food crop yields and increased biomass yields, as was found for one of the cover crops, show that strategic redistribution of residual biomass resources has a potential for increasing the overall system productivity and opens up for additional biomass uses in synergy with on-farm nutrient recirculation. The allocation of biomass resources for the additional production of CH4 without yield losses in the AD treatment enhances on-farm self-sufficiency and potentially also farm profitability depending on the energy pricing.
Our results indicate that the anaerobic digestion of biomass resources and field applications of the digestate might be associated with larger N losses than the biomass management in BR and IS. More detailed studies of N losses at each step of the management of biomass resources and digestate as well as at the entire cropping system level are therefore important in order to develop N-efficient cropping systems that provide bioenergy extraction in synergy with food production. An analysis of nutrient balances, energy and economics is also required to gain more knowledge for further developments of biomass resource-management systems for enhanced farm sustainability.
Acknowledgements
The research project has been financed by FORMAS, Swedish University of Agricultural Sciences, Alnarp and Lund University. This study has also been made possible by the Swedish Infrastructure for Ecosystem Science (SITES), in this case the Lönnstorp Research Station in Alnarp, Sweden. Parts of the seeds have been donated by Lantmännen. The lentil seeds were donated by Professor Albert Vandenberg, University of Saskatoon, Canada. The authors thank PhD Emma Kreuger who has produced the digestate and biogas at the facilities at Anneberg, Biotechnology, Lund University, Sweden. The authors also thank Sven-Erik Svensson, Lina Hirsch and the staff at SITES Lönnstorp for skilled technical assistance.
