Assessing the impacts of biochar‐blended urea on nitrogen use efficiency and soil retention in wheat production

Improving nitrogen (N) use efficiency (NUE) in crop plants is important to reduce the negative impact of excessive N on the environment. Although biochar‐blended fertilizer had been increasingly tested in crop production, the fate of fertilized N in soil and plant had not been elucidated in field conditions. In this study, a novel biochar‐blended urea (BU) was prepared by pelleting maize straw biochar, bentonite, sepiolite, carboxymethylcellulose sodium, and chitosan with urea (commercial urea without biochar [CU]). N fertilization in a winter wheat field was treated with BU and CU at both 265 kg N ha−1 (HL) and 186 kg N ha−1 (LN). Within a treatment plot, a microplot was fertilized with 15N‐labeled urea at a relevant N level. We investigated the influence of fertilizer management on biomass, grain yield, bioaccumulation of nutrient, soil properties, 15N isotopic abundance, and greenhouse gas emissions. Microscopic and spectroscopic analysis showed that micro/nanonetwork of biochar could bind N to form a loss control agglomerated particle, and organo‐mineral coatings on BU may protect N from quick release. Compared with CU, BU significantly increased grain yield by 13% and 38%, and grain N allocation by 19% and 55%, respectively, at HN and LN level. The total recovery of urea 15N in wheat plant (15N based NUE) was 32.8% under CU regardless of N rates but increased to 41.7% (HN rate) and 56.3% (LN rate) under BU. Whereas, the soil proportion (soil residual 15N) was 20.1% and 13.4% under CU but 32.5% and 18.8% under BU, in 0‐20cm topsoil, respectively, at HN and LN rate. Compared with the CU, BU had no effect on CO2 and CH4 emissions but significantly reduced the total N2O emission by 23%–28%. These important findings suggested that BU can be beneficial to uplift plant NUE to reduce reactive N loading but boost crop production.


| INTRODUCTION
N fertilization in agriculture, often in excess, has raised serious concerns about adverse environmental effects (Steffen et al., 2015;Tilman et al., 2001). As much as 50%-70% of fertilizer N is lost in intensive crop production via ammonia volatilization, ammonium and nitrate leaching, and nitrous oxide emissions (Ladha et al., 2005). All these N losses can be addressed under the umbrella of reactive N (N r ; Ju et al., 2006;Larry et al., 2011), contributing to the ongoing global environmental changes of soil and water acidification, eutrophication of waters, global warming, and groundwater nitrate pollution (Cui et al., 2013;Ju et al., 2009;. Unfortunately, the N use efficiency (NUE) of conventional mineral N fertilizers is in the narrow range of 30%-40%. "The Nutrient Nexus," a new global effort by 2020, called for the reduction of N losses and improvement of NUE by 20% to boost food and energy production while reducing potential environmental risks (Sutton et al., 2013). Addressing the triple challenges of food security, environmental quality and climate change (Zhang et al., 2015), improvements in NUE are critical to future world crop production.
Technologies developed thus far to increase fertilizer NUE include nitrification inhibitors (dicyandiamide and 3,4-dimethylepyrazole phosphate) and the urease inhibitor N-(n-butyl) thiophosphoric triamide (Abalos et al., 2014), slow releasing fertilizer technology, fertilization management (Herrera et al., 2016), and the use of nanomaterials as nanofertilizers (DeRosa et al., 2010;Gogos et al., 2012) to enhance slow release and avoid quick dissolution and leaching (Lupwayi et al., 2010). However, these technologies are often debated due to the potential translocation of synthetic inhibitors into the food web and the disturbance of soil microbial communities, in addition to the high cost of their adoption by farmers Coskun et al., 2017). As an increasingly recommended approach, biochar soil amendment (BSA) could help to reduce soil loss of N through leaching and gas emissions (Hagemann et al., 2017;Widowati et al., 2011;Xu et al., 2019;Zhang et al., 2017;Zheng et al., 2013) while providing multiple co-benefits through organic carbon (OC) sequestration Werner et al., 2018;Woolf et al., 2010Woolf et al., , 2016) and improved crop production (Biederman & Harpole, 2013;Jeffery et al., 2011;Liu et al., 2014). Indeed, significantly higher crop yield and, in turn, higher N agronomic use efficiency were often observed in agricultural soils under BSA (Blackwell et al., 2010;Zheng et al., 2013). Although one single application could bring about such a change over years (Liu et al., 2014(Liu et al., , 2019Wang et al., 2018), BSA with a massive amount of biochar was often beyond farmer adoption owing to the high cost, urgent to exploit further potential agricultural value of biochar (Clare et al., 2014(Clare et al., , 2015Dandamudi et al., 2021;Shan et al., 2020).
As a shifting paradigm for chemical fertilizers, the use of biochar combined (compound) fertilizer (BCF), biochar in small amounts blended with mineral nutrients, was proposed (Joseph et al., 2013;Pan et al., 2017). In these novel BCFs, urea was blended with biochar as clay-aggregated granules, where N was potentially bound to functional groups on the biochar surface (Joesph et al., 2013) and thus was slowly released and minimally accessible to microbes. In field experiments with paddy rice and maize, these BCFs (mainly with urea as the N source) could boost crop yield while reducing potent N emissions, greatly improving agronomic NUE in comparison with mineral fertilizers (Qian et al., 2014;Zheng et al., 2017). Thanks to the development of engineered biochar technology, BCFs have been put into massive production and wide use for crop production as a countermeasure to reduce chemical fertilizer input in China (Ministry of Agriculture & Rural Affairs, Ministry of Finance, China [MOAC-MFC], 2017; Pan et al., 2017).
There have been increasing efforts to improve the functionality of biochar-blended N fertilizers to enhance N retention and plant use in soil-crop systems. For example, bentonite and sepiolite clay (Shi et al., 2019) were preferably used as mineral binders alone or in combination with biodegradable organic binders such as carboxymethylcellulose sodium and chitosan with more information in the Supporting Information (Chen et al., 2018;Dimin et al., 2014;Gao, 2013;Jiang et al., 2014;Kim et al., 2014;Manikandan & Subramanian, 2015;Wu et al., 2008;Wu, 2014). Biocharcomposite blended urea using green waste biochar blended with a bentonite-sepiolite composite mineral binder was shown to have positive effects in prolonging the N supply for plant growth and diminishing leaching loss (Shi et al., 2019). However, the effects of such novel biochar-composite blended urea on N use by crops and residual N retention in soils have not yet been assessed quantitatively.  N labeling, biochar, nitrogen use efficiency, soil-plant system, urea, wheat production Wheat is one of the most important staple crops, meeting human requirements for carbohydrates and partially for protein for over half of the global population (Blakeney et al., 2009). In the world, China is the largest wheat producer next to the European Union (Foreign Agricultural Service, United States Department of Agriculture, & Office of Global Analysis [FAS-USDA-OGA], 2016), producing more than 130 million metric tons from a total of 24 million ha of cultivated land in 2015. However, wheat is a cereal crop with a high plant nutritional demand for N, particularly for the reproductive stage, and large quantities of N fertilizers have been required to enable high grain yield and protein content for wheat (Cao et al., 2001;Zebarth et al., 2007). N fertilizers, over two-thirds as urea, have been applied at rates of up to 300 kg N ha −1 in China's wheat production, constituting 20% of the state's total N fertilizer (25 million metric tons per annum) used in agriculture (Ye et al., 2007). Such high N application, often in excess of plant uptake, has already caused intensive pollution of soil and water as well as N 2 O emissions via nitrification and denitrification (Nosengo, 2003). To minimize environmental loading while boosting yield and grain quality in wheat production, the N application rate has been the primary management option in addition to the application timing both in China and globally (Walsh et al., 2018;Wu & Ma, 2015;Wu et al., 2019). Although BSA could help increase crop productivity and reduce N 2 O emissions (Liu et al., 2014;Zhang, Bian, et al., 2013), how biochar-blended urea (BU) fertilizer could reduce N loading to the environment while boosting wheat production is not clearly understood.
Therefore, it is hypothesized that urea blended with selected biochar and clay binder could improve NUE by wheat by improving plant uptake and enhancing soil retention through preservation with biochar in soil. This could be assessed with the recovery of N from fertilization with 15 N-labeled urea in soil-plant systems with different N applications. This study quantified the effects of biochar-mineral blended urea on wheat NUE, soil retention and greenhouse gas (GHG) emissions in a soil-wheat system field experiment with 15 N-labeled urea. We aimed to address how the use of BU could reduce N loading and to recommend shifting chemical N fertilizers to novel BU for sustainable wheat agriculture.

| Materials
Commercial urea without biochar (CU) was purchased from Luxi Chemical Group Co., Ltd, with white pellets 2-3 mm in size and contained 46% N. The biochar used for the BU preparation was produced via pyrolysis of maize straw at 450 °C in a rotatory kiln pyrolyzer operated by the Nanjing Qinfeng Straw Biomass Technology Co Ltd. The basic properties of the biochar are presented in the Supporting Information. Both carboxymethylcellulose sodium (C 6 H 7 O 2 (OH) 2 CH 2 COONa, viscosity > 1500 mPa s) and chitosan (C 6 H 11 NO 4 X 2 , viscosity > 400 mPa s) were purchased from Shanghai Macklin Biochemical Co., Ltd. The properties of the bentonite, sepiolite and maize wood vinegar were reported in a previous study and provided in the Supporting Information (Shi et al., 2019). Wood vinegar was provided by Nanjing Qinfeng Straw Biomass Technology Co Ltd., and its properties were reported previously (Shi et al., 2019). The urea used in the BU was purchased from Xilong Scientific Co., Ltd. and contained 46% N and 20% OC. All these materials were ground to pass through a 1.0-mm nylon sieve and homogenized before use. Both the CU and BU fertilizers were stored in doublesealed plastic bags prior to analysis and field experiments.

| Preparation of biochar-blended urea
BU was prepared from urea blended with maize biochar and organo-mineral binders. As described previously (Shi et al., 2019), in this study, the proportions of the raw materials in the BU were maize straw-derived biochar at 20%, urea at 40%, bentonite at 25%, sepiolite at 9%, carboxymethylcellulose sodium at 2%, and chitosan at 1% (w/w). The details of the preparation of the BU are presented in the Supporting Information. The prepared BU fertilizer was black pellets 2-3 mm in diameter with a Brinell hardness of 19.15 N mm −2 (Figure 1a,b). The prepared BU fertilizer contained 16.9% OC, 18.0% total N, 0.34% total K, 0.08% total P, and 1.4% and 0.9% Ca and Mg, respectively.

| Experiment site and soil condition
The soil-wheat system field experiment was carried out on a farm located in Kangbo village (31°35′N, 120°55′E), Guli township, Changshu municipality, Jiangsu, China. Lying in the center of the Taihu Lake plain in the lower Yangtze River valley, the area is a traditionally important region of rice, wheat, and rape seed production in China. Under a prevailing subtropical monsoon climate, the mean annual temperature was 15.4°C, and precipitation was 1054 mm over the period of 2015-2018. The soil, with the local name Wushantu, was classified as a Gleyic Stagnic Anthrosol developed through long-term rice cultivation and paddy management. Derived from lacustrine deposits, the soil texture was clay loam (33.8% sand, 38.6% silt and 27.6% clay). The basic properties of the topsoil (0-20 cm) before the experiment were as follows: pH (H 2 O) of 7.05, bulk density of 1.09 g cm −3 , soil OC (SOC) of 14.24 g kg −1 , and total nitrogen, phosphorus and potassium of 1.83, 1.05 and 1.58 g kg −1 , respectively.

| Field experiment design
In this study, CU and BU was tested at two levels of N supply for wheat. The soil-wheat system field experiment was designed and carried out in this study with the urea form (CU vs. BU) as the primary factor, while the N application rate (conventionally high level vs. reduced level) was considered a secondary factor. The five treatments included a null (CK) without N fertilization, 265 kg N ha −1 of CU and BU (CU HN and BU HN , respectively, hereafter), and 186 kg N ha −1 (LN, 70% of the conventionally high rate by Guo et al., 2019) of CU and BU (CU LN and BU LN , respectively, hereafter). Each plot was 20 m 2 (5.0 m × 4.0 m) in area was separated with surrounding protection rows 30 cm in width between the plots. Each treatment was performed in four replicates, and the 20 treatment plots were arranged in a randomized block design ( Figure S1).
As per the nutrient requirements for wheat (Cao et al., 2001), N fertilization was divided, with 30% as basal fertilizer before sowing, another 30% at the elongating stage and the remaining 40% at the boot stage. Following local practices for wheat production, both phosphorus and potassium were used only for basal application as Ca(H 2 PO 4 ) 2 at 75 kg P 2 O 5 ha −1 and as KCl at 30 kg K 2 O ha −1 . The basal fertilizer was hand sprayed and incorporated into the topsoil by ploughing to the depth of 10 cm before planting, but others were surface broadcast at elongating stage and elongating stage. The total nutrient inputs of the different treatments are listed in Table 1. F I G U R E 1 Form and physical structure of the prepared novel biochar-blended urea (BU). (a), physical appearance of the prepared BU as firmly aggregated granules; (b), the size of individual BU granules; (c), the surface of a randomly selected BU granule under scanning electron microscopy; (d), distribution of N throughout the surface of (c); (e), an energy dispersive spectrometer (EDS) spectrum of (c), showing N in association with C, O, Si, Al, K and Ca/Mg from biochar and the minerals; (f), a secondary electron image of the internal structure of the yellow-dashed surface portion of (c), showing needle-shaped minerals, with filamentous organic matter (brown dashed area), coating/intercalating biochar (right side) and urea; (g), N distribution across the internal structure of (f), showing the concentration of the mineral/biochar intercalates; (h), an EDS spectrum of (f), showing N in close association with the mineral elements of C, Si, Al, Ca, and Mg T A B L E 1 Total nutrient input (kg ha −1 ) with the fertilizer treatments Following the protocol for the 15 N-labeled fertilizer experiment (Ruisi et al., 2016;Yao et al., 2018), a microplot equipped with a static chamber 50 cm (length) × 50 cm (width) × 100 cm (height) was set up within each treatment plot ( Figure S1). Commercial 15 N-enriched urea ( 15 N isotope abundance enriched by 10.11%) was purchased from Shanghai Wusheng Biotechnology Co., Ltd. and used for the preparation of 15 N-enriched CU or BU as per the protocol described above. The prepared 15 N-labeled CU or BU was fertilized at 0, 4.6 and 6.6 g N in the microplots for CK, LN and HN, respectively, as per the urea treatment design. Other farm practices in the microplots were kept consistent with the relevant treatment plots.
On November 19, 2018, after the treatment plots received basal fertilizers, seeds of a local winter wheat cultivar (Triticum aestivum L. cv. of Yangmai 14) were directly sown at a density of 262.5 kg ha −1 . As the wheat grew, urea treated with the required N amount was top-dressed in the treatment plots at the elongating and boot stages as per the treatment design and N division mentioned above. Throughout the growing period, wheat production was managed without irrigation despite soil temperature and moisture being highly dynamic ( Figure S2). All farm management practices, including plant protection and weed control, followed the local practices and the details are given in the Supporting Information.

| Plant and soil sampling
At wheat harvest on May 25, 2019, an area of 1.0 m 2 in the center was selected in each treatment plot, and all the wheat plants therein were collected. Grains from the collected plants were manually threshed and then homogenized to estimate grain yield. Meanwhile, 10 wheat bushes (including grains, shoots and roots) were randomly collected in each treatment plot, washed, and then air-dried in the field. For 15 N tracing in the microplots, all the wheat plants were carefully removed, and plant tissues were carefully separated to obtain grain, shoot and root samples for N quantification. After cleaning with water in the field, all the separated samples were sealed in paper bags and shipped to the lab within 24 h following sampling.
A composite topsoil sample was collected from three random soil cores at depths of 0-20 cm with an Eijkelkamp soil core sampler in each treatment plot. Additionally, within each 15 N microplot following plant sampling, a composite soil sample was obtained of three random soil cores taken at three depths: 0-20 cm (topsoil), 20-40 cm (subsoil) and 40-60 cm (bottom soil). All the collected soil samples were sealed in plastic bags, shipped to the lab, and stored at 4°C in a refrigerator prior to analysis. Following soil core sampling, soil bulk density was measured using a 100 cm 3 cylinder within depth intervals of 0-20, 20-40 and 40-60 cm in each microplot under each treatment.

| Chemical analyses
All the grains, shoots, and root samples were defoliated at 105°C for 30 min followed by oven-drying at 60°C to estimate grain yield and/or biomass production, for both treatment plots and the microplots. Subsequently, a portion of each dried sample of grain, shoot, and root was crushed, further ground to pass through a 0.15-mm sieve and homogenized prior to chemical analysis of P and K following the protocols (Lu, 2000). For grain, the contents of free amino acids were also determined with the ninhydrin colorimetry method (Lu, 2000).
After removing root detritus and gravel, if any, all the soil samples were air-dried and ground to pass through a 2-mm sieve, and a random portion was further ground to pass through a 0.15-mm sieve. Basic soil properties (pH, soil bulk density, SOC, dissolved OC [DOC] and nitrogen, alkaline-extractable nitrogen, total nitrogen, soil microbial biomass carbon [MBC] and nitrogen) were determined as per the procedures (Lu, 2000;Wu et al., 1990). The MBC of soil samples were measured using the fumigationextraction method. The details of these measurements are provided in the Supporting Information.
For soil and plant samples, the total N content and isotope abundance of 15 N were determined with an isotope ratio mass spectrometer (Isoprime100) at the Institute of Environment and Sustainable Development of Agriculture, China Academy of Agricultural Sciences, Beijing, China. The results were in atom % of 15 N to the total N of a sample.
For the applied fertilizers, total OC and N were measured with the methods for soil mentioned above, whereas the other nutrients including P, K, Ca, and Mg were measured with inductively coupled plasma mass spectrometry (iCAP Q ICP-MS; Thermo Fisher Scientific; Lu, 2000). In addition, the prepared BU was tested for granule strength with a granulometer (KQ-2, particle strength tester; Nanjing Kehuan Analytical Instrument Co. Ltd., 2017).

| Greenhouse gas emissions
GHG emissions were sampled with the static closed chamber method . The plastic collar (50 cm × 50 cm) of the microplot was set up in each plot before wheat sowing, allowing it to be filled with water to avoid gas leakage when gas sampling. A 50 cm (length) × 50 cm (width) × 00 cm (height) chamber wrapped with a layer of sponge and aluminum foil was mounted on the top edge to minimize air temperature variability inside the chamber during gas collection. Sampling and quantification of GHG emissions are given in the Supporting Information Zou et al., 2005).

| Data analysis
The partial factor productivity of applied N (PFP N ) was calculated using the following equation: where, Y f and N f represents the grain yield (kg ha −1 ) of wheat and the total N fertilized (kg N ha −1 ), respectively, under a fertilizer treatment.
Subsequently, N agronomic efficiency (AE N ; kg grain per kg N input) was estimated using the equation as: where, Y c is the wheat grain yield (kg ha −1 ) under CK (no N fertilized) while Y f and N f are the same as in Equation (1).
In addition, a harvest index (HI) was estimated using the following equation: where, Y and AGB represents, respectively, the grain yield and aboveground biomass of wheat, in kg ha −1 , under a fertilizer treatment.
Furthermore, for the fertilizer-NUE by wheat and residue storage in soil layers, 15 N recovery on percentage basis of labeled-fertilizer were calculated using data of 15 N enrichment of a sample (soil, grain, shoot, and root). Firstly, total N (g) recovered in a pool sample (TNs) either of soil or of a plant tissue of the wheat-soil system could be estimated as: where, W and N is, respectively, the dry mass (g) and nitrogen content (g kg −1 ) of a plant pool or soil layer.
Secondly, N derived from the 15 N labeled fertilizer in a pool could be estimated using the following equation: where, Ndff S is the percentage of N derived from the fertilizer to the total N in an analyzed sample; SL and NL denotes labeled and unlabeled samples, respectively; FL represents the labeled fertilizer. Finally, total N recovery of the fertilized 15 N could be derived with the equation as follows: where, 15 N REC is the labeled-fertilizer N recovery (%) in a pool of wheat plan and soil layer; AN F is the amount of N fertilized under a urea treatment.
Following, NUE (%) by wheat plant could be predicted by summing all the 15 N recovery values of the plant pools, known as 15 N-based NUE, calculated using the equations as follows: where, 15 N REC (g) ,

15
N REC (s), and 15 N REC (r) is the analyzed 15 N recovery (%) of labeled urea 15 N, respectively, in grain, shoot and root of harvested wheat in the microplot under a fertilizer treatment.
Alternatively, apparent plant NUE (NUE A ) under a treatment could be also estimated as the proportion of net total plant N uptake to total N fertilized in a treatment plot, known as CK-based NUE, using an equation as: where, TNU TR and TNU CK represents the total plant N uptake by wheat, as the sum of N uptake (kg N ha −1 ) by grain, shoot, and root of wheat, respectively, under a urea treatment and under CK; N f is the total N fertilized, as defined in Equation (1).
In addition, a total GWP value with non-CO 2 gases of CH 4 and N 2 O was estimated following the equation: where, E(CH 4 ) and E(N 2 O)represents the total seasonal emissions of CH 4 and N 2 O (kg ha −1 ) over the wheat growing season, respectively. The GWP of CH 4 and N 2 O is 28 and 265 times than that of CO 2 over a 100-year horizon.
A carbon (greenhouse gas emission) intensity (GHGI) was calculated following the equation: where, GWP is the sum of CH 4 and N 2 O (kg CO 2 -e ha −1 ) emissions in CO 2 equivalence, Y represents the wheat grain yield (kg ha −1 ).
Finally, a N fertilizer-induced emission factor (EF) of N 2 O was estimated with the total N 2 O emission measured over the wheat-growing season, using an equation: where, E f and E c is the total seasonal emission of N 2 O-N (kg ha −1 ), respectively, under a N fertilization treatment and CK.
Sample data were all expressed as the mean plus/minus one standard deviation of four replicates. All data were processed with Excel 2016 software. Statistical analysis was performed using SPSS 19.0 and Origin 9.0 software. The differences among the treatments were examined by one-way ANOVA followed by Duncan's post-hoc test, whereas the interaction between N fertilizer type and application rate was examined by two-way ANOVA. The significance of a difference was defined at p < 0.05.

| Physical characterization of biochar-blended urea
Designed to be analogous to soil aggregates, the prepared urea with biochar (BU) was firm granules approximately 2-3 mm in size, as shown in Figure 1a,b. Under scanning electron microscopy (SEM) (Figure 1c), the BU granule was clearly embedded with organic and mineral binders coatings on the agglomerated particles. As shown in Figure  1d N was evenly distributed in the internal structure, in which the combination of urea with clay minerals (Si, Al, Mg, Ca, etc.) and biochar (C, Fe, S, P, K, etc.) was reflected in the energy dispersive spectrometer (EDS) (Figure 1e). The added organic binder was found partly in the form of a filamentous coating and interacted with biochar and/ or mineral particles on the surface of granule for the BU (Figure 1f). In more detail (Figure 1f), needle-shaped minerals of sepiolites and bentonites coated or intercalated the biochar and urea, forming mineral-biochar combined urea clusters and N-enriched microgranules, with plenty of micropores between them (Figure 1g). This combination was supported by the high abundance of Si and Ca as well as C, detected by EDS in the clusters (Figure 1h).

| Biomass production, grain yield, and quality of wheat
Total biomass, plant growth, and grain production data under the treatments are presented in Table 2. Wheat growth on the farm was stressed by low soil moisture over the reproductive growth period ( Figure S2). Plant growth was very weak under CK without N fertilization but similar among the treatments with N application ( Figure S3). The total biomass and grain yield were both higher under the CU and BU treatments compared with CK. Spikes per unit area, 1000-grain weight, and total biomass were all not significantly different among the N treatments (Table  2). However, grains per spike was significantly higher under BU than under CU by 26.3% at the HN level but by 27.8% at the LN level, respectively.
Among the treatments, there were also variations in grain quality related to food nutrition, although no difference was observed in the grain contents of P or K (Table  S1). Compared with CK, grain N content was significantly increased under CU and BU, with the exception of CU at LN. Under BU, the grain N content was 32% higher at the HN rate and 52% higher at the LN rate than CK. Furthermore, the grain content of free amino acids increased markedly under N fertilization compared with CK and was significantly higher under BU than CU by 20% at the HN rate and by 10% at the LN rate.

| Plant N uptake, grain partitioning, and nitrogen use efficiency
According to the N content data in Table S1, the wheat plant tissue biomass data and the estimated total plant N uptake of the treatment plots are provided in Table S2. Specifically, the calculated plant N uptake and N allocation in wheat tissues are shown in Figure 2. The N contents of both roots and shoots were not significantly different between the N treatments. The change in biomass with N fertilization was much higher for grains and roots than for shoots. Plant total N uptake was only 33.23 kg N ha −1 in CK but varied in the range of 115.44-221.73 kg N ha −1  (Figure 2b), the percentage of total N fertilized that was used in total plant N uptake for BU compared with CU was 20.6% higher significantly at the HN level and 38.9% higher significantly at the LN level. However, this value was unchanged by the N levels under CU but significantly increased at the LN level compared with the HN level under BU. As shown in Figure 2b, grain N allocation was increased by 18.7% and 54.8% under BU compared with CU at the HN and LN level. This increase was greater (by 57%) than the increase in the total N uptake (by 38%; Figure 2a). Relevantly, the ratio of grain N uptake to shoot and root N uptake, referring to as the grain partitioning coefficient of N (PPR N ), was 1.35 under CK but in the range of 1.93-3.92 under the urea treatments (Table S2). Similar to grain allocation, PPR N increased nonsignificantly with BU compared with CU at HN but significantly at the LN level. While grain allocation was higher at the LN level than the HN level, the difference was not significant under CU but significant under BU, and PPR N was not significantly different between the N levels. In addition, total grain N uptake accounted for 35.3% and 37.2% of the total urea N input under CU and 44.2% and 54.6% under BU, respectively, at the HN and LN rates. Estimated from the above data, plant NUE showed a wide range of variation with the treatments. First, the estimated partial productivity (PFP N ) was in the range of 24.0-36.5 kg grain kg −1 N, and the agronomic efficiency of N (AE N ) was in the range of 19.5-30.0 kg grain kg −1 N fertilized across the treatments (Figure 3). There was an increase in PFP N of 2.4 and 8.3 kg grain kg −1 N under BU compared with CU at the HN and LN rates, respectively. Similarly, a 13% and 38% increase in AE N was observed under BU compared with CU at the HN and LN levels, respectively. Furthermore, the wheat HI reflecting the distribution ratio of crop assimilates in grains and vegetative organs was significantly elevated under BU compared with CU at the LN level, with 0.54 under BU and 0.38 under CU ( Figure S4).

| Soil physicochemical properties
The physicochemical properties of the topsoil (0-20 cm) samples collected at wheat harvest under the treatments are shown in Table 3. Soil pH (H 2 O) ranged from 7.09 to 7.21, and bulk density ranged from 1.02 to 1.11 g cm −3 , with no significant differences among the treatments. Compared with CK, the SOC content increased nonsignificantly under CU but significantly under BU, by over 20% at both N rates. There was no change in DOC content among the treatments except for a 40% increase under BU at the HN rate compared with CK.
Despite no change in total N, alkaline extractable N increased significantly, but moderately (by 25%), under BU at the HN rate and slightly (by 11%-16%) under the other fertilization treatments compared with CK. Furthermore, F I G U R E 2 Total N uptake (kg N ha −1 ) by wheat plant (a) and allocation (%) of fertilized urea N in wheat tissues (b), estimated of the N content and biomass of plant tissues under the treatments. CK, no N fertilizer as control; CU HN and BU HN , and CU LN and BU LN , urea and biochar-blended urea, respectively, at N rate of 265 kg ha −1 and of 186 kg ha −1 . Different letters above the bars (in a) represent a significant difference in total uptake while in the yellow blocks (in b) in grain N allocation, among the treatments at p < 0.05 dissolved organic nitrogen (DON) significantly increased by over 80% under the BU treatments compared with CK at both the HN and LN rates. However, there was no change in either NH + 4 -N or NO − 3 -N across the urea treatments, although both were significantly higher in the treatments than in CK. In addition, the soil moisture content was more or less different among the treatments. Although the soil moisture in CK was close to that monitored in the microstation field ( Figure S2), it was nonsignificant lower under CU at both the HN and LN levels. Under BU compared with CU, the soil moisture was significantly higher at the HN level but not significantly higher at the LN level.
As shown in Figure 4, topsoil MBC increased significantly under all urea treatments, by 21%-31% compared with CK. Although there were no differences either between CU and BU or between N rates across the urea treatments, soil microbial biomass nitrogen was enhanced significantly by 29%-55% compared with CK. This enhancement was significantly higher under BU than under CU at the HN level but nonsignificant at the LN level.  Evidently, this change was related to that in available N and DON in the soils. In addition, compared with CK, the soil microbial C/N ratio was significantly lower under BU at HN but not significantly lower under the other urea treatments.

| Recovery of 15 N in wheatsoil system
Data on the relative enrichment of 15 N determined from plant and soil samples collected in the microplots are provided in Table S3 and S4. Using data on N content, biomass or soil bulk density (Table S1, S3, and S4) and the amount of 15 N quantified (Table S5) from plant tissues and soil layers, the calculated recovery percentage of labeled 15 N in the wheat-soil system of the microplots is plotted in Figure 5 in a systematic diagram. The N of the plant tissues derived from fertilized urea 15 N was in the range of 47%-56% ( Figure S5) and was relatively unchanged among the fertilizer treatments. This portion was not significantly changed between CU and BU at the HN rate but was significantly higher under BU than under CU at the LN rate. Total plant recovery of labeled 15 N ranged from 32.8% to 56.3% and was contributed predominately by wheat grains (58%-67%), followed by wheat shoots (27%-35%) and only slightly by wheat roots (5%-7%), generally in correspondence to their biomasses ( Figure S4). The total plant recovery of the labeled 15 N, equivalent to the N fertilizer use efficiency, was 32.8% under CU regardless of the N application rate but 41.7% and 56.3% under BU at the HN and LN application rates, respectively. The plant use efficiency of N was significantly and greatly increased under BU by 30% and 72% compared with CU at the HN and LN rates, respectively. As per the two-way ANOVA, biochar blended urea had a stronger impact on N uptake by grains (p < 0.01) and roots (p < 0.05) than the N rate or the interplay (Table S6).
The data in Table S4 show 15 N enrichment in the 0-60 cm soil. As shown in Figure 5, the residual fertilized 15 N in the rooted topsoil (0-20 cm) was 13.4%-32.5% of the total urea 15 N, varying by treatments. Both fertilizer type and N rate had a significant impact on the recovery of labeled 15 N in the 0-20 cm topsoil (p < 0.01; Table S6). In detail, the topsoil residual recovery was 62% higher under BU (32.5%) than under CU (20.1%) at the HN rate and 40% higher under BU (18.8%) than under CU (13.4%) at the LN rate. In contrast, 23.2% and 21.1% of the 15 N was recovered in the deep soil (20-40 and 40-60 cm) under CU, compared with 10.1% and 8.4% under BU, respectively, at the HN and LN rates. Apparently, residual 15 N was 12% and 5% greater in the topsoil under BU than under CU at the HN and LN rates, respectively. In contrast, the residual 15 N in the deep soil was 13% less regardless of the N application rate. Overall, the total recovery of 15 N-labeled fertilizer N down to the rooted topsoil across the entire wheat-soil ecosystem was 52.9% and 46.2% under CU but 74.2% and 76.1% under BU, respectively, at the HN and LN rates, resulting in a 20% increase in the ecosystem use efficiency of N for BU compared with CU.

| Greenhouse gas emissions
The dynamics of the N 2 O, CH 4 , and CO 2 fluxes under different treatments across wheat seasons are presented in Figure S6. The peaks in N 2 O emissions were found after adding N fertilizer in the boot stage, with 49.79 ± 17.96 μg N 2 O-N m −2 h −1 for BU and F I G U R E 4 Soil microbial biomass carbon (SMBC; a) and nitrogen (SMBN; b) contents under the treatments. CK, no N fertilizer as control; CU HN and BU HN , and CU LN and BU LN , urea and biochar-blended urea, respectively, at N rate of 265 kg ha −1 and of 186 kg ha −1 . Different letters above the bars represent a significant difference among the treatments at p < 0.05 59.51 ± 22.46 μg N 2 O-N m −2 h −1 for CU, while CH 4 and CO 2 emissions depended primarily on soil moisture and temperature but not on N fertilizer. For wheat cultivated soil, BU significantly reduced the total emissions of N 2 O by 23% and 28% compared with CU fertilization for HN and LN, respectively, but there was no change in CH 4 or CO 2 among the N application treatments (Table  4). Furthermore, the BU in HN-fertilized soil with wheat plants had the lowest N 2 O emissions of 0.43 kg ha −1 over the entire wheat growing season. Similarly, varying in a range of 0.10-0.25, EF (%) was lower by 29% and 39% under BU compared with CU at HN and LN, respectively, whereas the estimated GWP and GHGI under BU at HN and LN were both significantly lower than those under CU at HN and LN.

| Wheat grain yield and quality under commercial urea without biochar and biochar-blended urea
Wheat yield change under the urea treatments (Table  2) demonstrated a significant but great effect by urea N fertilization on wheat production, being higher under BF than under CF. Wheat grain yield of 5.25-7.06 t ha −1 under N fertilization in this study was close to the grand mean value of 5.6 t ha −1 with 206 kg N ha −1 fertilization reported for wheat in this geographical region in a national network fertilization experiment over -2008(Wang et al., 2010. However, the grain yield in this study  18.00 ± 3.31c 0.10 ± 0.03c Note: Values are mean ± SD (n = 4). Different letters in a single column denote a significant difference among the treatments at p < 0.05.
CK, no N fertilizer as control; CU HN and BU HN , and CU LN and BU LN , urea and biochar-blended urea, respectively, at N rate of 265 kg ha −1 and of 186 kg ha −1 .
Abbreviations: EF, N fertilizer-induced emission factor of N 2 O; GHGI, greenhouse gas intensity; GWP, global warming potential. a Not applicable. seemed higher than the statistical mean of approximately 4.0 t ha −1 with 220 kg N ha −1 fertilization reported for a conventional wheat farming system in Jiangsu during 1998-2004(Ye et al., 2007. In contrast with the low yield (1.2 t ha −1 ) in CK, probably due to stress from soil conditions, urea N fertilizer induced a much greater increase in wheat grain yield than in rice, with yields <30% in the same soil cultivated with rice (Pan et al., 2003) and <15% in a clay loam rice paddy in an adjacent rice area . This N fertilization dependence was further demonstrated by the sharp decrease of 18% in grain yield with a 30% reduction in urea N application (from the conventional 265 to 186 kg N ha −1 ) compared with an insignificant change in rice when the conventional rate was cut by 30% in an adjacent rice area (Pan et al., 2003).
On the other hand, the yield increases for BU compared with CU were 10.8% and 29.3%, respectively, for the HN (rate of 265 kg N ha −1 ) and LN (rate of 186 kg N ha −1 ) treatments. These increases could be related to changes in plant growth traits (Table 2). Meanwhile, this positive yield change was in general agreement with the previously observed grain yield increases of 11%-31% for maize and rice under biochar compound fertilizer compared with conventional mineral fertilizers (Puga et al., 2020;Qian et al., 2014;Zheng et al., 2017). The wheat grain yield gains for BU compared with CU at the HN level in the present study were comparable with those under surfaceenhanced slow-releasing urea compared with urea at an application rate of 210 kg N ha −1 , reported in a field study in a sandy loam Entisol from central North China (Wang et al., 2007). In a field trial in North China with nanocarbon-enhanced urea compared with pure urea, a wide range of winter wheat grain yield increases (4.5%-21.8%) were reported at N applications of 220-240 kg ha −1 (Ma et al., 2009). In a meta-analysis, the use of N fertilizers with urease as nitrification inhibitors showed a positive yield change of 2%-5% for cereal crops despite an overall mean increase of 7.5% for all crops (Abalos et al., 2014). In comparison, the BU prepared in this study could have a great potential to boost wheat grain yield at a rational N application. In addition, a 10%-20% increase in total free amino acids of wheat grain confirmed the significantly improved grain nutrition quality, in addition to the grain yield gain. This finding was in contrast to the decline in grain protein of wheat with the use of nanocarbon slow-release fertilizer (Ma et al., 2009). In comparison, both the grain yield and grain protein of wheat increased by less than 5% at a rate of 200 kg ha −1 N in an experiment examining improved management of N fertilizers in wheat production . Although wheat has been a staple crop with particular market value for human carbohydrates and protein requirements, not only grain yield but also quality is subject to climate change (Blakeney et al., 2009;Fernando et al., 2014;Wang et al., 2018). The synergistic improvement of grain production and grain protein observed here could provide a hint for using biochar blended fertilizer as a potential measure in farming systems to tackle climate change impacts on food production and safety while reducing reactive N loading (Coskun et al., 2017;Tirado et al., 2010). However, how biochar in urea promotes plant N translocation and grain protein build-up is not yet understood, although biochar is known to be beneficial for plant functional gene expression (Viger et al., 2015). Contributed by yield improvement, both partial productivity (PFP N ), agronomic efficiency of N (AE N ) and wheat HI fertilized with urea was significantly elevated under BU over CU both at high and low N application levels ( Figure 3). This result suggests greater economic efficiency of grain production under BU than under CU with reduced N application. Moreover, in line with the improved N plant uptake, a significant increase in grain N allocation (Figure 2b) suggested a benefit of BU for N translocation from roots/shoots to grains, helping protein synthesis in wheat grains.

| Soil properties under commercial urea without biochar and biochar-blended urea
As wheat being highly N dependent, N supply could be the predominant driver for wheat biomass and grain build-up (Cao et al., 2001). In this study, loss of grain yield and less grain N under CU but not under BU particularly when N rate reduced by 30% suggested a potential prolonged N supply for wheat growth till ripening. Compared with urea, BU was already shown significantly slower in N release, probably owing to N protected in biochar pores (Manikandan & Subramanian, 2015). This was tested and proved in a previous study that the BU stabled N release for 22 days in comparison with 3 days for urea (Shi et al., 2019). For the BU in this study, the organic and mineral coating provided further protection from quick release and leaching with water flow (Chen et al., 2018).
A significant but moderate (by over 20%) increase in SOC under BU over CK at both HN and LN rates (Table  3) could not be attributable to the direct input of OC from the biochar in the fertilizers (0.93-0.14 t C ha −1 ; Table 1). Zheng et al. (2013) indicated biochar addition stimulated root biomass of maize, while Xiang et al. (2017) observed a general promotion of root biomass by 20% in biocharamended soils. A study of grass reported root access to N in biochar pores promoted biomass production in prolonged growth (Wen et al., 2017). As an important driver of ecosystem carbon and nutrient cycling, SOC content could be increased via carbon input through root exudation due to promoted root growth (Bardgett et al., 2014). The higher DOC under BU could be due to the direct release of vinegar from BU and indirect pathway through wheat root exudation. In this study, topsoil microbial biomass N (Figure 4b) was greatly enhanced under BU (by over 30%) compared with CU, and was related to DON and partly to available N in soil, implicating the potential involvement of biological transformation and turnover N in improving the N supply to plants (Pan et al., 2009). Furthermore, the increase in the soil DOC and DON contents (Table 3) reflecting the enhancement of organic N in the system in BU compared with CU. All these results suggest potential improvement of soil N storage and availability together with microbial growth due to SOC enhancement, likely contributed by improved wheat production following BU application. Biochar added with fertilizer could provide habitat for plant roots and microbial communities (Lehmann et al., 2011;Liu et al., 2020;Xiang et al., 2017), which could have the best response to available N, particularly DON, in the rooted zone with active soil-rootbiochar interactions (Lehmann et al., 2015).
In addition, in the experiment site, wheat growth and ripening were stressed by decreased moisture content in spring ( Figure S2), as a spring drought has been increasingly impacting wheat production under climate change in this area (Pan et al., 2011). Soil moisture at harvest tended more or less high under BU than under CU at the HN (Table 3). Enhanced soil water retention was indeed observed with biochar-assisted slow-releasing fertilizer (Wen et al., 2017). Biochar addition could potentially improve soil moisture retention and thus enhance plant drought resistance for plant growth (Artiola et al., 2012;Liang et al., 2014;Omondi et al., 2016). The use of the BU with high inner porosity ( Figure 1) could potentially enhance moisture retention and, thus, partly contribute to improved grain production if ripening and grain filling were restricted by low moisture in the field.

| Wheat nitrogen use efficiency and ecosystem N fate under commercial urea without biochar and biochar-blended urea
The percentage of N derived from 15 N labeled urea (Ndff) with 47%-56% under the urea treatments, at the high end of the range of 32%-60% in a similar microplot experiment in an adjacent area , across grains, shoots and roots of wheat. In a meta-analysis, Yan et al. (2020) reported a grand mean Ndff (N derived from the 15 N labeled fertilizer) of 37% for small grain crops, including wheat. Compared with CU treatments, a 7% increase of Ndff was observed consistently across grain, shoot and root under BU at LN rate ( Figure S5). In a previous study, decreases in Ndff of 2%-4% and 6%-8% were seen with band application compared with broadcast application and with high N (at 240 kg ha −1 ) and low N (at 160 kg ha −1 ), respectively . Thus, the observed increase in Ndff with BU compared with CU at the LN rate could highlight the improved plant N uptake under urea with biochar when the N supply was reduced by 30% of the local practical dose.
The wheat NUE values (32. 8%-56.3%, 15 N method) in this study were comparable with the range of 34%-56% previously found under treatments with release-controlled urea (Zheng et al., 2016). Although higher plant NUE was recognized for organic fertilizers (manure, green manure, and compost) than for mineral fertilizer in plant-soil systems (Yan et al., 2020), BU, compared with pure urea, did exhibit a higher recovery of fertilizer N in crops derived from current-season N fertilizer and a potentially higher N supply for subsequent crops in the longer term. In a similar experiment in an adjacent province of China , wheat plant NUE ( 15 N method) of urea fell in the range of 31%-39%, with a nonsignificant decrease with band application compared with broadcast application. Varying with N application rates, the global mean of fertilized N use by wheat plants (CK-based) was almost 50% (Coskun et al., 2017). Respectively, compared with the global mean NUE of small grain crops including wheat (48% by the CK method and 42% by the 15 N method; Yan et al., 2020), wheat NUE at the two N levels hereby was less under urea (42%-43% for CK based and 32.8% for 15 N based) but higher under urea with biochar (52%-63% for CK based and 42%-56% for 15 N based). Thus, plant recovery of labeled 15 N showed a greater (by 30%-72%) uplift of wheat NUE of fertilized BU than CU.
The topsoil residual 15 N recovery was higher under BU than under CU, but lower in deep soil (20-60 cm depth). This indicated the residual fertilizer-N in topsoil that was stabilized or recycled as organic N such as DON in the soil (Table 3) could persist in the ecosystem, continuing to supply N for subsequent crops (Sebilo et al., 2013;Yan et al., 2014), while N loss from the system was approximately half under CF but was reduced to one-fourth only under BU. As clarified recently, residual N in topsoil (0-20 cm depth) could be potentially plant-available for the subsequent crop following wheat, but N in unrooted deep soil (20-60 cm depth) could be less beneficial as environmental loading for wheat, which is a shallow root crop, mostly rooted within 15 cm depth (Cao et al., 2001;Wu & Ma, 2015;Yan et al., 2020). By using the tested BU instead of urea, the potential avoidance of N loss was estimated at 66.2 and 46.5 kg ha −1 in the tested wheat-soil system for N applications of 265 and 186 kg N ha −1 , respectively. The greater ecosystem functions but smaller down-soil migration highlighted the benefits of using urea with biochar to improve the N cycle in croplands and to reduce N environmental loading (Clough et al., 2013;Spokas et al., 2012).
Based on the data presented in Figure 5, the portioning of the recovered urea 15 N in the plant to the soil, in the grains to the shoots and roots, and in the rooted topsoil to the deep soil were all generally higher under BU than CU and greater at the LN rate than HN rate. More evidently, with the improved total plant NUE and topsoil residual N derived from fertilizer, the loss of fertilizer-N through leaching to the subsoil (20-60 cm) was reduced by over 25%, with the potential to mitigate the N pollution to water. This could be considered a synergy for building soil N and managing soil N turnover with crop yield improvement by enhancing NUE and retention (Yan et al., 2020;Zheng et al., 2013). For dryland crops, residual urea N remained in the topsoil after harvest at approximately 7%-20%, and 25%-40% of the total residual urea N leached to the subsoil or deep soil (Sui & Zhang, 2014;Yuan et al., 2014). Therefore, the replacement of urea with BU fertilizer would enhance the turnover of soil organic nitrogen but reduce the leaching loss of urea N, meeting a significant N deficit of 30-55 kg ha −1 year −1 for wheat (Yan et al., 2020). In Ye et al. (2007), total urea of over 12.5 Mt per annum was used for wheat production. The method with the lower ecological impacts also had better performance in increasing grain production and reducing Nr losses from wheat croplands (Ying et al., 2017). Therefore, using BU as in this study to replace CU in wheat cultivation could potentially save over 1.2 Mt urea at the HN rate but over 3.0 Mt urea at the LN rate, thus avoiding a large amount of Nr loading to the environment. Therefore, urea blended with biochar, as an environmentally friendly fertilizer, should be urgently considered for use for green agriculture in China (Ministry of Agriculture & the People's Republic of China [MoAC], 2015). Depending on the N application rate and soil texture, the use of synthetic inhibitors improved the NUE by 2%-5% for cereal crops, even though the mean was 12.9% (Abalos et al., 2014). In an early pot experiment with wheat and urea in a loam Alfisol, plant 15 N recovery was increased by 12% with a urease inhibitor but by 35% with both urease and nitrification inhibitors (Xu et al., 2000). In a pot experiment with rice, the recovery of labeled 15 N from mineral N fertilizer was elevated by 40% with the application of inorganic (70%) N combined with organic (30%) N fertilizer compared with mineral N fertilizer alone (Peng et al., 2011). Although organic N tended to be much less efficiently used by plants (by 28%) than mineral N (by 42%; Yan et al., 2020), urea blended with biochar as an organic amendment to fertilizer in this study indeed helped improve wheat use of urea N, particularly when less N was provided. Nonetheless, Quemada et al. (2013) argued that improved fertilizer technology was not sufficiently effective in reducing N leaching from mineral fertilizers.
As predicted with the 15 N method and CK-based method (Yan et al., 2020), the plant use efficiency (NUE) of urea N fertilized by wheat significantly increased under BU compared with CU at both the HN and LN levels ( Figure 6). Therefore, the net increase in wheat NUE under BU compared with CU was approximately 10% at HN and 20% at the LN level and was mostly consistent between the 15 N method and the CK-based method. However, relative increases of 19% and 50% were found with the CK-based method and of 30% and 72% with the 15 N method at the HN and LN levels, respectively. In the present case with winter wheat, the use of BU provided a promising improvement of more than 20% as the target of "The Nutrient Nexus 2020" (Sutton et al., 2013).

| N 2 O emission under commercial urea without biochar and biochar-blended urea
The rate of reduction of N 2 O under BU compared with CU could result from the N-urea in the BU was protected by the biochar and mineral from soil microbe to N 2 O as shown in the SEM of Figure 1 and is consistent with previous studies. Qian et al. (2014) indicated that biochar compound fertilizer application with a lower rate of N input reduced N 2 O emissions by approximately 30% in rice production F I G U R E 6 Urea N use efficiency (%) estimated based on the control method (blue bars) and on plant recovery of labeled 15 N, respectively, obtained in this study. CU HN and BU HN , and CU LN and BU LN represent urea and biochar-blended urea, respectively, at N application rates of 265 and 186 kg ha −1 , respectively. Different letters above the same colored bars represent a significant difference among the treatments at p < 0.05 compared with conventional chemical fertilizer. Li (2013) reported that biochar compound fertilizer reduced N 2 O emissions by 27% on average over a 2-year experiment in maize croplands. The findings in the present study confirm the significant potential of using the BU to replace conventional urea fertilizer to boost wheat production while greatly reducing N loss to the environment in global wheat production, largely avoiding N losses such as N 2 O emissions and nitrate leaching.

| CONCLUSIONS
The BU showed excellent capacity to bind N as a loss control agglomerated pellet. Furthermore, BU demonstrated a synergistic improvement in wheat productivity, plant N use, soil N retention and wheat grain quality as well as GHG emission mitigation. Thus, we recommend the use of crop residue BU as a strategy for sustaining grain production and for reducing excessive input of fertilizers in wheat production. More field studies should be conducted to examine the N dynamics in plant soil-ecosystems and the long-term feedbacks to soil and ecosystem functioning.

ACKNOWLEDGMENTS
This study was supported by Ministry of Science and Technology of China under the National Key Research and Development Program of China with a grant number of 2017YFD0200802 and by National Natural Science Foundation of China under a grant number of 41771332, 41877096 and 41877097. The research team was also financially funded by the National "Double First-rank" Construction Plan by the Ministry of Education of China and the Priority Academic Program Development of Jiangsu Higher Education Institutions. The work was also a contribution to the provincial consortium of "Jiangsu Collaborative Innovation Center for Solid Organic Waste Resource Utilization, supported by the Jiangsu Government. The authors are grateful to Mr Qiang Wang, owner of the farm, for his assistance in field works with the experiment.

CONFLICT OF INTEREST
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

AUTHOR CONTRIBUTION
Wei Shi: design and conduct the field experiment, analysis and MS drafting; Rongjun Bian: methodology, project administration, funding acquisition, experiment design; Lianqing Li: conceptualization and data inspection; Xiaoyu Liu and Wanli Lian: experiment management, sampling and analysis; Jufeng Zheng, Xuhui Zhang, Marios Drosos: data analysis and discussion; Kun Cheng: statistics and discussion; Stephen Joseph: microscopic analysis and data interpretation; Genxing Pan: conceptualization, experiment design, analysis and MS editing.

DATA AVAILABILITY STATEMENT
Data are available on request from the authors.