Effects of two wood-based biochars on the fate of added fertilizer nitrogen—a 15N tracing study

A 15N tracing pot experiment was conducted using two types of wood-based biochars: a regular biochar and a Kon-Tiki-produced nutrient-enriched biochar, at two application rates (1% and 5% (w/w)), in addition to a fertilizer only and a control treatment. Ryegrass was sown in pots, all of which except controls received 15N-labelled fertilizer as either 15NH4NO3 or NH415NO3. We quantified the effect of biochar application on soil N2O emissions, as well as the fate of fertilizer-derived ammonium (NH4+) and nitrate (NO3−) in terms of their leaching from the soil, uptake into plant biomass, and recovery in the soil. We found that application of biochars reduced soil mineral N leaching and N2O emissions. Similarly, the higher biochar application rate of 5% significantly increased aboveground ryegrass biomass yield. However, no differences in N2O emissions and ryegrass biomass yields were observed between regular and nutrient-enriched biochar treatments, although mineral N leaching tended to be lower in the nutrient-enriched biochar treatment than in the regular biochar treatment. The 15N analysis revealed that biochar application increased the plant uptake of added nitrate, but reduced the plant uptake of added ammonium compared to the fertilizer only treatment. Thus, the uptake of total N derived from added NH4NO3 fertilizer was not affected by the biochar addition, and cannot explain the increase in plant biomass in biochar treatments. Instead, the increased plant biomass at the higher biochar application rate was attributed to the enhanced uptake of N derived from soil. This suggests that the interactions between biochar and native soil organic N may be important determinants of the availability of soil N to plant growth.


Introduction
To meet the growing food and energy demand, the application of synthetic nitrogen (N) fertilizers to agricultural soils has dramatically increased over the last century. Unfortunately, this has also caused harmful environmental effects. Specifically, the applied N is not only taken up by plants and retained in soil but is also leached, and hence causes eutrophication, when it reaches water bodies (Isermann 1990). Globally, about 55 Tg N year −1 is leached from agricultural soils (Van Drecht et al. 2003). In addition, part of the N added to soil is emitted into the atmosphere as N 2 O, causing a large impact on climate. The greenhouse gas (GHG) N 2 O has a global warming potential (GWP 100 ) 265 times higher than that of carbon dioxide (CO 2 ) and is regarded as the single most important gas responsible for stratospheric ozone depletion (IPCC 2014;Portmann et al. 2012). The agriculture sector is responsible for 66% of gross anthropogenic N 2 O emission; it has been projected that by 2050, anthropogenic N 2 O emission will be twice as much as today (Davidson and Kanter 2014). Therefore, there is a great need for agricultural practices that can increase efficiency of applied N fertilizer use by crops while minimizing negative environmental effects such as N leaching and N 2 O emissions.
The application of stable C in the form of biochars into agricultural soil has gained popularity because of its potential to sequester atmospheric C in soil. Moreover, the application of biochars to agricultural soil has potential agricultural (Biederman and Harpole 2013;Jeffery et al. 2017a) and environmental benefits such as reduced mineral N leaching and N 2 O emissions (Borchard et al. 2019;Cayuela et al. 2013;Clough et al. 2013;Nguyen et al. 2017). The beneficial effects of biochars are attributed to inherent properties, which commonly include high aromaticity, porosity, specific surface area, negative surface charge, and surface charge density (Downie et al. 2009;Liang et al. 2006). These properties make biochars highly stable in soil (Kuzyakov et al. 2014) and able to retain water and nutrients in soil (Glaser et al. 2002;Karhu et al. 2011;Tammeorg et al. 2014).
A fresh biochar addition may reduce the availability of nutrients to plants, particularly N, which can even reduce the crop yield . However, treatment of biochars with nutrient-rich organic substances, for example through co-composting, can increase the retention and supply of plant-available nutrients, increasing the crop yield (Hagemann et al. 2017;Ye et al. 2016). A promising method of preparing homemade biochar with Kon-Tiki flame curtain pyrolysis, followed by nutrient enrichment has been gaining attention (Schmidt and Taylor 2014). This method follows the principle of pyrolyzing biomass layer after layer in an open, conically built metal kiln (or dug pit) that is easy to operate, fast, and results in high-quality biochar with low greenhouse gas emissions (Cornelissen et al. 2016). Following the pyrolysis, the resulting biochar is steam-activated by quenching with either water or nutrient-rich solutions like urine or cattle slurry, when the biochar is still hot. The activation process increases surface area and porosity, which promote the adsorption of nutrients (Borchard et al. 2012). Schmidt et al. (2015) found that biochar produced with a Kon-Tiki kiln and enriched with cattle urine increased pumpkin yield by 300% compared to urine only treatment, and by 85% compared to same amount of biochar without urine in a silt loam soil. Similarly, Pandit et al. (2017) reported that hot nutrientenriched biochar produced with a similar flame curtain kilns led to significant increases of 153% in aboveground biomass production of maize compared to cold nutrient-enriched biochar, and 209% compared to biochar added separately from the nutrients during a pot trial.
The application of biochars can affect the N cycling processes (Clough et al. 2013;Nguyen et al. 2017). It has been reported that biochars can reduce mineral N leaching and N 2 O emissions, and increase plant N uptake (Borchard et al. 2019;Tan et al. 2018). However, these beneficial effects are unpredictable because the results presented in the literature are contradictory-some report positive effects, while others report negative or no effect. This suggests that the effects are biochar-and soil-specific (Mia et al. 2017). During field aging, the surface properties of a biochar can change with the development of more oxygen-containing carboxylic functional groups (Cheng et al. 2006), which can increase the retention of NH 4 + because of increased cation exchange capacity (Mia et al. 2017). Also, field aging has been reported to enhance the retention of NO 3 − by physical entrapment into biochar pores (Haider et al. 2016;Joseph et al. 2018). Realizing these benefits of field aging, several techniques of preparing biochar with similar properties as that of field-aged biochar are being explored, such as co-composting  and chemical oxidation (Mia et al. 2019). The Kon-Tiki kilnproduced nutrient-enriched biochar may also exhibit promising results because of its higher surface area, porosity, and cation exchange capacity (CEC), resulting from steam activation (Borchard et al. 2012;Cornelissen et al. 2016). However, there is currently no adequate information about the effects of such biochar on the dynamics of applied fertilizer N. Native soil organic N is also an important source of N for plant productivity. Indeed, the effects of biochar on the dynamics of soil organic matter have been identified as one of the priorities in biochar research (Tammeorg et al. 2017). The small fraction of labile C present in biochar can enhance microbial activity, which can lead to increased mineralization of the soil organic matter (priming effect). Though the reported results vary, some biochars have been shown to induce positive priming, i.e., increased mineralization of soil organic matter (Awad et al. 2012;Wardle et al. 2008), which also increase the mineralization of soil organic N (Nelissen et al. 2012). However, the number of studies on the interactions between biochar application and the dynamics of soil native N is conspicuously low (Fiorentino et al. 2019). Such mineralized N is mostly taken up by microbes (Kuzyakov and Xu 2013) and may also get entrapped into biochar pores, which can limit its accessibility for plant uptake. There is limited evidence on whether such additional mineralized N stimulated by biochar addition is available for plant uptake and can thus affect plant productivity in soil-plant system.
Using an innovative 15 N tracing approach with 15 NH 4 NO 3 and NH 4 15 NO 3 fertilizers, we assessed the effects of two types of biochars-a regular biochar and a Kon-Tiki-produced nutrient-enriched biochar-on the dynamics of fertilizer N. We measured plant fertilizer N uptake and leaching losses of added fertilizer N separately for fertilizer-derived NH 4 + -N and NO 3 − -N. We also quantified the importance of fertilizerderived N versus soil-derived N for the plant yield in the control and biochar-amended soils. We hypothesized that soil amendment with biochars will (1) recover more 15 N in plants and soil with both 15 NH 4 + -N and 15 NO 3 − -N fertilization coupled with reduced NH 4 + -N and NO 3 − -N leaching and N 2 O emissions, as well as (2) promote the uptake of soil-derived N to plant biomass.

Biochar and soil
The biochars used in this study were a commercially available regular biochar (BC1) and a nutrient-enriched biochar (BC2). Both biochars were produced from wood. The regular biochar (product of RPK Hiili Oy, Mikkeli, Finland) was selected as a representative of commercially available biochars. It was produced by pyrolyzing mixed deciduous wood (hardwood) at 400°C in a retort kiln. The nutrient-enriched biochar was obtained by pyrolyzing hardwood branches and split logs (approximately 80-90% Willow (Salix spp.), 5-10% Birch (Betula spp.), and 5-10% other hardwood species: Alder (Alnus), Bird cherry (Prunus padus L.), and Norway maple (Acer platanoides L.)) in a 0.3 m 3 Kon-Tiki kiln (http://terramagica.info/index.php/terra-magica-kontiki/kon-tiki-garten). After pyrolysis, the biochar was soaked with a mixture of tap water:cattle slurry (in a ratio of 7:3). Due to technical limitations, the pyrolysis temperature was not measured for the production of BC2. However, it has been reported that the temperature just below flame curtain can reach 680 to 750°C (Schmidt and Taylor 2014). BC2 has a higher specific surface area and C:N ratio, but lower total N content compared to BC1 (Table 1). The biochar particle size used in the study was 0.2 to 2 mm. Sandy loam soil (55% sand, 35% silt, and 10% clay, measured with Coulter LS 230 Laser Diffraction Particle Size Analyzer, Beckman Coulter Inc.) was collected from an agricultural field in southwestern Finland (61°00′ 18.5″ N, 22°4 6′ 01.7″ E) from a depth of 0-30 cm. The field had been previously planted with barley, peas, and winter wheat for the past 4 years, and fertilized annually with 15 m 3 ha −1 pig slurry. Prior to the experiment, the soil was stored at 5°C and later sieved through a 4-mm sieve to remove large plant and gravel particles. The soil had pH of 6.9 and electrical conductivity 85 μS cm −1 (1:2.5 w/v), and contained 1.06% and 0.10% of total C and N, respectively (measured with Variomax CN analyzer, Elementar Analysensysteme GmbH, Germany).

Experiment
A pot experiment was conducted in a greenhouse at the Viikki Campus, University of Helsinki. The treatments consist of control, fertilizer only, 1% (w/w dry weight basis) BC1 + fertilizer, 1% BC2 + fertilizer, 5% BC1 + fertilizer, and 5% BC2 + fertilizer in a randomized complete block design. 15 Nenriched fertilizers were applied to all five treatments except the control. Each of those five fertilized treatments contained two groups, receiving fertilizers as either 15 NH 4 NO 3 or NH 4 15 NO 3 . Five replicates were included for each group, for a total of 55 pots. The pots (6 × 6 × 6 cm 3 ) with holes at the bottom were filled with moist soil containing approximately 100 g soil on dry weight basis at 50% water holding capacity. The pots were lined with a nylon mesh (50 μm mesh size) at the bottom-this prevented loss of soil through holes while allowing a free flow of water. In the biochar treatments, the soil and corresponding amounts of biochars were mixed properly before filling into the pots. Italian ryegrass (Lolium multiflorum) seeds (1 g pot −1 , approximately 250 seeds) were spread and gently hand-pressed on top of the soil/soil biochar mixture in every pot. The pots were covered with thin plastic film until germination. For fertilization, 2 mL of 2.5 mg N mL −1 10 atom% (at%) 15 NH 4 NO 3 or NH 4 15 NO 3 solution was pipetted over the pots after germination. Each pot was placed inside another, larger pot (9 cm diameter and 6 cm height). This was done to allow watering the plants from below, and for leachate collection during the leaching tests, when water was added on top of the inner pots to simulate rain and allowed to drain through the soil to the bottom of the  Hellstedt et al. (2018) outer pot. The plants were watered by pouring about 10-40 mL of N-free reverse osmosis water frequently (usually every 2 days) in the bottom of the outer pot. Usually, after 2 days, there was no water remaining in the outer pots, although some pots occasionally had some water. The volumes of water used for watering differed due to the varying amount of water that remained from previous watering. Also, during later phases of the experiment, more water was required to supplement the growing plants. For the measurement of soil pH, soil and biochar were mixed separately in the same ratios as in the planted pots, but fertilizer was not added. The pH of those soil and soil biochar mixtures was measured by a standard combination electrode in a 1:2.5 (w/v) suspension in Milli-Q water. The average daily temperature inside the greenhouse was 18.4°C and average daily relative humidity was 71%.

Sampling and measurements
Leaching test The leaching test was carried out on days 4, 12, 17, and 24 after the application of 15 N-labelled fertilizers. We aimed at collecting about 40 mL of leachate for later analysis of NH 4 + -N and NO 3 − -N concentrations and their 15 N content; thus, the amounts of water used for the leaching test were adjusted during the experiment. The leaching test was also carried out 2 days after adding the fertilizer, but since a very small amount of leachate was collected from the 5% biochar treatments, the leachate collected on day 2 was combined with that of day 4 for analysis. Before conducting the leaching test, any remaining water (from the previous watering) on the bottom of the outer pot was discarded, to allow us to quantify the volume of water that leached through the soil. For leaching, about 45-60 mL of reverse osmosis water was poured on top of the soil in the inner pot. The volume of water used was the same for all the pots each time a leaching test was conducted, but was adjusted during the experiment to allow ca. 40 mL of leachate to be collected at all times, as explained above. The exact volume of water added was recorded. The added water was allowed to leach through the soil for ca. 30 min and then collected from the bottom of the outer pot. The volume of this leachate was measured, and the leachate was frozen at − 20°C until further analyses. To determine 15 NH 4 + -N and 15 NO 3 − -N in the leachate, the leached NH 4 + -N and NO 3 − -N were concentrated on acidified filter paper prior to analysis (Sørensen and Jensen 1991). The 15 N contents on the filter papers were measured by elemental analysis (CE 1110, Thermo Electron, Milan, Italy) coupled in continuous flow mode to a Finnigan MAT Delta PLUS isotope ratio mass spectrometer (IRMS; Thermo Scientific, Bremen, Germany). The mineral N concentration in the leachate was determined with an automated colorimetric method ["Ammonia (DIC)" for NH 4 + -N and "TON-V" for NO 3 − -N] using Gallery™ Plus Discrete Analyzer (Thermo Scientific™, Vantaa, Finland). The initial plan was to measure 15 N in the leachate throughout the experiment. However, the concentration of N in leachates was very low, apart from the first leaching test conducted on days 2 and 4 after fertilization. Hence, only the leachate collected on days 2 and 4 after fertilization (combined to allow sufficient volume) was analyzed for 15 N; in the remaining sampling periods, the amounts of N in the leachate were under the detection limit for 15 N analysis. Accordingly, we present the 15 N results for the first leaching test (day 2 + day 4) only; however, the concentrations of NH 4 + -N and NO 3 − -N are presented for all the leaching tests, and all time points were used to calculate the cumulative NH 4 + -N and NO 3 − -N leaching.

Greenhouse gas measurement
For measuring the GHG emissions, each pot was placed inside a glass jar (3.1 L volume) with an air-tight nozzle fixed in its lid for gas sampling. After over-pressurizing the glass jar with 80 mL of ambient air, 20-mL gas samples were taken into 12 mL helium flushed evacuated Exetainers® (Labco Scientific, High Wycombe, UK) at 0, 4, 20, and 24 h after closing the jars. The concentrations of GHGs (CO 2 , CH 4 , and N 2 O) in the gas samples were measured using a gas chromatograph (7890A, Agilent Technologies, California, USA) equipped with a flame ionization detector (FID) and a methanizer for CO 2 and CH 4 , and an electron capture detector (ECD) for N 2 O (Pihlatie et al. 2013). The GHG emission rate was calculated by fitting linear regression to the measured gas concentrations over measurement time. The GHG measurements were carried out from only three replicates. The presented GHG results were measured 1 day after fertilizer application. Since the ryegrass seeds were just germinated at that time, the contributions of photosynthesis and plant respiration to the CO 2 flux measurement were considered negligible and hence the measured CO 2 emissions represent only the respiration of soil or soil biochar mixture.

Plant and soil sampling
At the end of the experiment (33 days after sowing), plants and soils were destructively sampled. The aboveground biomass was separated by cutting the ryegrass at the soil surface, and the roots were separated from the growing media by washing with water. The plant samples were dried at 60°C for 48 h to calculate aboveground biomass (AGB) and root biomass. The plant samples (aboveground and root) and soil samples were finely ground using a ball-mill and analyzed for 15 N content with an elemental analyzer coupled in continuous flow mode to a Finnigan MAT Delta PLUS isotope ratio mass spectrometer (Thermo Scientific, Bremen, Germany).

Calculation and data analysis
The 15 N content in plant, soil, and leachate derived from added labelled fertilizer was calculated using the so-called N derived from fertilizer (Ndff) equation (Powlson and Barraclough 1993): Nds where T is the total N content in the samples ( NO 3 ) and contained the same weight of N, with the only difference of 15 N being linked in different moieties. Hence, while analyzing the data when there is no relevance of 15 N, both fertilizers were regarded as the same. This resulted in 10 replicates for a given treatment except for the control (which has 5 replicates). The effects of the treatments were tested with one-way analysis of variance (ANOVA) followed by Tukey's post hoc test to determine the differences among the level of treatments. The residual normal distribution and homogeneity of variances were ascertained by plotting residuals of the model against theoretical quantiles and fitted values, respectively. Whenever relevant, a t-test was carried out to compare the means between two specific categories for better contrast, for example, biochar vs. without biochar treatments, 1% biochar vs. 5% biochar treatments, and BC1 vs. BC2 treatments. The data was analyzed at 5% level of significance. For Ndff and Nds in plant biomass, the standard deviations were calculated using propagation of uncertainty. The means of Ndff and Nds among the treatments were compared using a pairwise t-test. The statistical analysis was carried out using R version 3.5.3 (R Core Team 2018).

Plant biomass and soil pH
Compared to the fertilized control, the 1% biochar treatments had no effect, while the 5% biochar treatments significantly increased aboveground biomass (Fig. 1a). The 5% BC1 and 5% BC2 treatments increased the aboveground biomass by 22% and 23% respectively compared to the fertilizer only treatment. At the corresponding biochar application rates, there was no difference in aboveground biomass between BC1 and BC2. No differences across treatments were found in terms of root biomass yield (Fig. 1b). The addition of both biochars at 1% application rate significantly increased soil pH from 6.9 to 7.2 (p < 0.05). The addition of both biochars at 5% application rate significantly increased soil pH further to 7.5 (p < 0.05). Soil pH was similar in BC1 and BC2 at respective application rates.

Leaching
The biochar treatments reduced the volume of leachate collected. The 5% biochar treatments were more efficient than the 1% biochar treatments in reducing the leaching of water through soil. The 5% BC2 significantly reduced leachate volume in day 2 + day 4 after fertilizer N application compared to 5% BC1. Similarly, in day 17 after fertilizer N application, 1% BC2 had significantly lower leachate volume compared to 1% BC1 ( Supplementary Fig. 1).
The concentrations of NH 4 + -N and NO 3 − -N in leachate were significantly lower in biochar treatments relative to the fertilizer only treatment during the first (day 2 + day 4) leaching test (Table 2). However, no differences were observed between biochar treatments and the fertilizer only treatment in the later leaching tests (days 12, 17, and 24 after fertilization). When comparing the biochar treatments at 1% application rate, BC1 had significantly higher NH 4 + -N concentration in the leachate during the first leaching test compared to BC2. When looking at the effects of biochar application rates, we found that 5% BC1 had significantly lower NH 4 + -N and NO 3 − -N concentration in the leachate compared to 1% BC1 on day 24.
The biochar treatments significantly reduced cumulative NH 4 + -N and NO 3 − -N leaching compared to fertilizer only treatment ( Table 2). The biochar treatments reduced total NH 4 + -N leaching by 21-53% and total NO 3 − -N leaching by 47-68%. At 1% biochar application rate, BC1 had significantly higher cumulative NH 4 + -N leaching compared to BC2, while no differences were observed at 5% biochar application rate. There were no statistical differences in cumulative NO 3 − -N leaching between the two biochar treatments at either of the application rates.

Greenhouse gas emissions
Immediately following fertilizer application, the biochar treatments reduced N 2 O emissions compared to the fertilizer only treatment by 57-81% (average 69%). However, no differences in N 2 O emissions were observed between the biochar treatments (Fig. 2a). N 2 O emissions were negligible from the non-fertilized control pots. On the other hand, 1% BC1 treatment significantly increased CO 2 emission rate by 34% (Fig.   2b) and overall, biochar treatments had higher average CO 2 emissions compared to the treatments without biochar (t = 3.23, p < 0.01). There were almost no CH 4 emissions from any of the treatments (data not shown).

Distribution of added 15 N
When the fertilizer was applied in the form of 15 NH 4 NO 3 , biochar treatments significantly reduced the recovery of 15 N in ryegrass biomass compared to the fertilizer only treatment (Fig. 3a). In contrast, when the fertilizer was applied in the form of NH 4 15 NO 3 , the biochar treatments significantly increased the recovery of 15 N in ryegrass biomass compared to the fertilizer only treatment (Fig. 3b). There were no statistically significant differences in the 15 N amounts retained in the soil ( Fig. 3c and d), even though with 15 NH 4 NO 3 fertilization, there was a tendency for less 15 N to be retained in the soil of   Fig. 1 Ryegrass aboveground (a) and root (b) biomass yield (Mean ± SE) across the treatments. Different letters represent significant differences between treatments (p < 0.05) indicated by Tukey's test. The percentages above the bars represent the relative change compared to the fertilizer only treatment when statistically significant. The biochar treatments (1% BC1, 1% BC2, 5% BC1, and 5% BC2) received the same fertilization as the fertilizer only treatment. Control (n = 5); other treatments (n = 10) Different letters across treatments indicate significant differences between treatments (p < 0.05). The biochar treatments (1% BC1, 1% BC2, 5% BC1, and 5% BC2) received the same fertilization as the fertilizer only treatment the biochar treatments compared to the fertilizer only treatment, especially at the higher 5% biochar application rate (Fig. 3c).
No differences were found in 15 NH 4 + -N leaching across all treatments in both fertilization types ( Fig. 3e and f). However, there was high variation in 15 NH 4 + -N leaching within the fertilizer only treatment. The average 15 NH 4 + -N leaching was ca. 9 times higher than that of the average of the biochar treatments when fertilizer was applied as 15 NH 4 NO 3 (Fig. 3e). The 5% BC1, 5% BC2, and 1% BC2 treatments had significantly lower 15 NO 3 − -N leaching compared to the fertilizer only treatment, when fertilizer was applied as 15 NH 4 NO 3 (Fig. 3g). When fertilizer was applied as NH 4 15 NO 3 , the observed pattern was the same, but only the 5% BC2 treatment had significantly reduced 15 NO 3 − -N leaching compared to the fertilizer only treatment, due to higher variability within treatments (Fig. 3h). When compared to the fertilizer only treatment, 15 NO 3 − -N leaching tended to be reduced in BC2 more than the BC1 treatments, but the differences between the two biochar treatments were not statistically significant. It is notable that the amounts of 15  The total plant N uptake (mg N pot −1 ) was significantly higher in the 5% biochar treatments compared to the fertilizer only treatment (Fig. 4). The 15 N analysis revealed that there was no difference in the total amount of N derived from added NH 4 NO 3 fertilizer in plant biomass between the different treatments, because the higher plant uptake of 15 NO 3 − -N in the biochar treatments compared to the fertilizer only treatment was offset by the lower 15 NH 4 + -N uptake. However, the 5% biochar treatments significantly increased the amount of N derived from soil (or soil + biochar mixture) by about 24% compared to the fertilizer only treatment (Fig. 4). Both the average amount of soil-derived N in the plant biomass as well as plant biomass yield increased in the ordercontrol < fertilizer only < 1% biochar treatments < 5% biochar treatments. There was a strong positive relationship between the N derived from soil and plant biomass yield (R 2 = 0.85, p < 0.05, Fig. 5), but there was no relationship between total N derived from fertilizer NH 4 NO 3 and plant biomass yield (R 2 = 0.01, p > 0.05).
The biochar treatments increased the loss of added 15 NH 4 + -N, but reduced the loss of added 15 NO 3 − -N (Supplementary Table 1). Most of the total applied 15 N was recovered in plant biomass and soil (84-86%) while a rather small fraction was lost (14-16%) in all treatments. The amount of 15 N lost via leaching in the first leaching test (day 2 + day 4) was less than 0.1% of the total applied 15 N in all fertilized treatments.

Biochar reduced soil mineral N leaching and N 2 O emission
In agreement with earlier lysimeter studies (Lehmann et al. 2003;Xu et al. 2016), we found that biochar treatments reduced the concentrations and total cumulative amounts of both NH 4 + -N and NO 3 − -N in leachate compared to the fertilizer only treatment (Table 2). In addition, biochar treatments reduced the fertilizer-derived 15 NH 4 + -N (though not statistically significant) and 15 NO 3 − -N in leachate compared to the fertilizer only treatment (Fig. 3e-h)  FerƟlizer only 1% BC1 1% BC2 5% BC1 5% BC2 -N leached (µg N pot ) (h) by biochar addition (Gai et al. 2014;Kizito et al. 2015;Liang et al. 2006). Although there were no differences in the leaching of fertilizer 15 NH 4 + -N between the two biochars ( Fig. 3e), BC2 was more efficient in reducing the total cumulative NH 4 + -N leaching compared to BC1 at 1% application rate (Table 2). This was likely because of the higher adsorption of NH 4 + -N associated with the higher specific surface area in BC2 (about 8-fold higher than BC1, Table 1), which might have resulted from the steam activation after the pyrolysis.

NO
Since biochars usually have limited affinity to adsorb NO 3 − -N (Gai et al. 2014;Yao et al. 2012), the reduced NO 3 − -N leaching may be due to its entrapment inside the biochar pores (Haider et al. 2016;Kammann et al. 2015), because of strong capillary action. Such entrapment is mostly because of physical rather than chemical processes, and is thus affected by the surface area of biochar , as a higher surface area provides more micropores. Accordingly, during the first leaching test (day 2 + day 4), there was an indication that BC2 reduced 15 NO 3 − -N leaching more than BC1 (Fig. 3g-h), even though this difference was not statistically significant. The reduction in mineral N leaching may have also resulted from the reduction in hydraulic conductivity and thus enhanced retention of water ( Supplementary Fig. 1) in biochar micropores, and changes in pore-size distribution (Glaser et al. 2002). The reduced leaching of NH 4 + -N and NO 3 − -N in biochar treatments was observed only during the first leaching test when the ryegrass plants were still small. In later leaching tests, the differences in NH 4 + -N and NO 3 − -N concentrations of leachates between fertilizer only and biochar treatments became smaller, most likely because the otherwise leachable, plant-available mineral N had already been taken up by the ryegrass.
The N 2 O emission from the control treatment was almost non-existent, and therefore, we can conclude that the observed N 2 O emissions from fertilizer only and biochar treatments originated from the added fertilizer N. The biochar treatments reduced N 2 O emission by an average of 69% immediately after fertilization-a much greater N 2 O emission reduction with biochar than reported in the recent meta-analyses (6-38%) (Borchard et al. 2019;Liu et al. 2019). Our experimental design did not allow detailed studies on the mechanisms underlying the reduction of N 2 O emission. However, based on previous studies, it is possible that the reduced N 2 O emissions in biochar treatments might be a result of the limited accessibility of NO 3 − -N retained in the pores of biochar for denitrification (Cayuela et al. 2013;Haider et al. 2016;Van Zwieten et al. 2014). In addition, through the increased soil pH, biochar could have promoted complete denitrification from NO 3 − to N 2 instead of N 2 O as the end product of denitrification (Cayuela et al. 2013;Dannenmann et al. 2018;Harter et al. 2013;Sánchez-García et al. 2014). The biochar treatments at both application rates had a pH higher than 7, which favors the synthesis and assembly of N 2 O reductase, promoting N 2 as the final product of denitrification over N 2 O (Bergaust et al. 2010). The lower N 2 O emission might also result from   (Baggs et al. 2006;Lan et al. 2017). However, we found no difference in the N 2 O emissions between the two types of biochars even though the C:N ratio of BC2 was much higher than that of BC1, with a higher potential for N immobilization. Our finding is similar to the result obtained by Cayuela et al. (2013), who reported no relationship between C:N ratios of biochars and N 2 O mitigation, suggesting that microbial N immobilization may not be the main driving mechanism behind reduced N 2 O emission.
Biochar enhanced the recovery of added 15 NO 3 − -N, but reduced that of added 15 NH 4 + -N The recovery of 15 N in plant biomass was decreased by biochar treatments when the added 15 N was in the form of 15 NH 4 + -N. On the other hand, the 15 N recovery increased when it was in 15 NO 3 − -N form. This was partly contradictory to our hypothesis 1, since we had expected that biochar would increase the recovery of both mineral N forms in plant biomass. Similarly, the recovery of 15 N in soils, when added in 15 NH 4 + -N form, tended to be lower in the biochar treatments than in the fertilizer only treatment, especially at the higher 5% biochar application rates, but these differences were not statistically significant. This is contradictory to the assumption that increasing the CEC of soil with the addition of biochar would help to retain more 15 NH 4 + -N in plant available form, and thus increase its plant uptake. Instead, these results suggest that biochar stimulated the loss of 15  -N (through nitrification), which is in line with the findings that biochar can stimulate the nitrification process (Berglund et al. 2004;Dannenmann et al. 2018). Nevertheless, we found that the application of biochars helped to reduce NO 3 − -N leaching and N 2 O emission. It is also unlikely that significant amounts of added 15 NH 4 + -N were lost during denitrification as N 2 , because if that had been the case, we would have observed the loss of added 15 NO 3 − -N in the presence of biochar in the 15 NO 3 − -N fertilized treatments too. As such, the loss of added 15 NH 4 + -N through biochar-induced nitrification-denitrification pathway seems implausible. Rather, the loss of added 15 NH 4 + -N might be due to ammonia (NH 3 ) volatilization. The increased soil pH with the addition of biochar with high pH (> 9) and/or combination of biochar with NH 4 + based N fertilizer might have increased NH 3 volatilization (Mandal et al. 2018;Sha et al. 2019) due to conversion of NH 4 + -N to NH 3 under high soil pH. Application of 5% biochar increased soil pH more than application of 1% biochar, stimulating higher loss of added 15 NH 4 + -N and thus less recovery in soil and plant biomass. As we did not measure NH 3 volatilization in this study, we could not confirm it as the reason behind the loss of added 15 NH 4 + -N. Some previous studies have also shown that biochar-induced NH 3 volatilization had led to significant N losses from soil (Schomberg et al. 2012;Sun et al. 2019). Contrary to this, others have outlined the potential of biochar to enhance the adsorbtion of NH 4 + and NH 3 , and hence decrease NH 3 volatilization (Mandal et al. 2016;Taghizadeh-Toosi et al. 2012;Sun et al. 2020). Therefore, the effect of a biochar on NH 3 volatilization seems to depend on soil and biochar properties. Thus, further studies measuring NH 3 volatilization in different soil conditions combined with various biochars with different properties would be required before tailored solutions for reducing NH 3 volatilization could be recommended. When fertilizer was applied in 15 NO 3 − -N form, the recovery of 15 N in plant biomass and soil was higher in the biochar treatments compared to the fertilizer only treatment as hypothesized. Our findings of the coupling between 15 NO 3 − -N loss and its recovery in plant biomass indicate that the 15 NO 3 − -N retained in the biochar was safe from leaching and denitrification, but was accessible for plant uptake. Though the values seem comparatively small, 15 NH 4 + -N leaching from all treatments with NH 4 15 NO 3 fertilization (Fig. 3f) suggests a possible re-mineralization of immobilized 15 NO 3 − -N by microorganisms. Nevertheless, a very small amount (< 0.1%) of added 15 N was lost through leaching in the first leaching test, and the amount of 15 N in the leachates was under the detection limit for 15 N analysis in the subsequent leaching tests. This result is similar to the findings of a field 15 N tracing experiment, where only about 0.05% of added 15 N was recovered in the sub-soil layer suggesting that a very small proportion of the added 15 N was lost via leaching (Mia et al. 2017). Such a small amount of added 15 N leached indicates that increased plant uptake of added 15 NO 3 − -N by biochars was more likely achieved due to reduced loss of added 15 NO 3 − -N via gaseous emissions than the reduced leaching of NO 3 − -N in our experiment. In addition, even though biochars reduced the average 15 NH 4 + -N leaching, the recovery of added 15 NH 4 + -N in soil and plant biomass was decreased because the magnitude of added 15 NH 4 + -N lost (most probably via NH 3 volatilization) was much higher than the reduced 15 NH 4 + -N leaching.

Biochar increased the uptake of N derived from soil
The addition of biochar did not benefit plant growth through uptake of added NH 4 NO 3 because increased plant uptake of the added 15 NO 3 − -N was offset by the loss of added 15 NH 4 + -N. Instead, addition of biochar at the higher application rate increased the uptake of N derived from soil (or soil + biochar mixture) into plant biomass as expected according to our hypothesis 2. There was a strong positive relationship between soil-derived N in plant biomass and plant biomass yield. Both, the N derived from soil and plant biomass yield increased in the following order: control < fertilized control < 1% biochar treatments < 5% biochar treatments (Fig. 5). This suggests that biochar increased plant biomass yields by increasing the availability of soil-derived (or soil + biochar-derived) N, either by increasing mineralization of native soil organic matter (positive priming effect), and thus releasing organic N into mineral form (Fiorentino et al. 2019;Nelissen et al. 2012;Singh and Cowie 2014), or by acting as a source of N itself. The latter is less likely to be an important mechanism, as it is known that biochar is highly recalcitrant to microbial decomposition, and the amount of N mineralized from biochar itself is usually very small (Fiorentino et al. 2019;Jeffery et al. 2017b).
There was no difference in the soil-derived N uptake into plant biomass between BC1 and BC2 treatments, contrary to our expectation that BC2 would have increased the availability of plant available N. This suggests that the soaking of biochar with cattle slurry did not load the biochar pores with plant-available N. Although BC2 was enriched with cattle slurry, the total N content of BC2 was lower than that of the BC1, likely because of the difference in feedstock and pyrolyzing temperature. BC2 was produced at a higher pyrolysis temperature than the BC1, and biochars produced at higher temperatures have relatively lower N content (Gai et al. 2014;Mandal et al. 2018). In addition, the amount of N in the cattle slurry mixture used may not have been high enough to increase the N content of BC2 because of high dilution-three parts of cattle slurry were mixed with 7 parts of water to maintain optimum consistency for homogeneous activation of biochar.

Conclusion
Our results suggest that the effect of biochars on the fate of fertilizer N depends on the type of fertilizer N applied. Biochars increased the plant uptake of added NO 3 − -N, which was coupled with reduced NO 3 − -N leaching and N 2 O emissions. On the other hand, biochars stimulated the loss of added NH 4 + -N, which lead to decreased plant uptake. The mechanism behind the loss of added NH 4 + -N was suspected to be NH 3 volatilization, which needs further confirmation. Thus, our results indicate that, at least in the short-term, the incorporation of NO 3 − -based N fertilizer in combination with wood-based biochars increases soil retention and plant uptake of added fertilizer N. Adding NH 4 + -based N fertilizer together with such biochars with high pH can increase the risk of fertilizer N loss from soil with neutral or higher initial soil pH. Biochars did not improve the total plant N uptake from fertilizer NH 4 NO 3 because increased uptake of NO 3 − -N was offset by reduced uptake of added NH 4 + -N. Instead, biochars increased the plant biomass yield as a result of increased uptake of soil-derived N. This suggests that biochars could have stimulated the mineralization of soil organic N, and kept the mineralized N accessible for plant uptake. Although the two types of hardwood biochars tested in this study were produced using different techniques and constitute varying physico-chemical properties, they did not differ in their effects on plant productivity and dynamics of fertilizer and soil-derived N.