Journal of Environmental Management 154 (2015) 351e357
Contents lists available at ScienceDirect
Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman
The impact of exogenous N supply on soluble organic nitrogen dynamics and nitrogen balance in a greenhouse vegetable system Bin Liang a, Lingyun Kang b, Tao Ren c, Li Junliang a, Qing Chen b, *, Jingguo Wang b a
College of Resources and Environmental Sciences, Qingdao Agriculture University, Qingdao 266109, China College of Resources and Environmental Sciences, China Agricultural University, Beijing 100193, China c College of Resources & Environment, Huazhong Agriculture University, Wuhan 430070, China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 September 2014 Received in revised form 22 February 2015 Accepted 27 February 2015 Available online 9 March 2015
A long-term greenhouse experiment (2004e2012) was conducted with continuous tomato (Lycopersicum esculentum Mill.) plantings to understand the influence of an exogenous nitrogen supply from irrigation water, chemical fertilizer, or organic amendment on the N balance and soluble organic nitrogen (SON). The results from 16 tomato growing seasons indicated that the application of organic amendment (manure and straw) alone (Or-N) resulted in the same yield as the conventional chemical N with organic amendment (Co-N) and the reduced chemical N with organic amendment (Re-N) treatments. The annual apparent N loss was >1000 and 438 kg N ha1 in the Co-N and Re-N treatments, respectively. Over the study period, the SON in the 1.8 m soil profile was 1449 and 1978 kg N ha1 in the Re-N and Co-N treatments, respectively, it was 1.7- and 2.3-fold higher than that observed in the Or-N treatment, which indicated that SON increased with the chemical N application. The percentage of SON in the cumulative soluble N (SON plus mineral N) ranged from 28% to 44%, and there were no significant differences across the 0e0.6, 0.6e1.2, and 1.2e1.8 m soil profile, which indicated that the leaching and distribution of SON was similar to those of the mineral N in the 0e1.8 m soil profile. We conclude that the mobility of soluble organic N in the 0e1.8 m of the soil was synchronous with the mineral N under a greenhouse production system, and the risk of soluble organic N leaching increased with inorganic N application rate. Therefore, leaching of SON in the intensive agriculture should not be ignored when evaluating the risk of N leaching. Crown Copyright © 2015 Published by Elsevier Ltd. All rights reserved.
Keywords: N fertilizer Manure Irrigation water Mineral N Plastic greenhouse
1. Introduction Vegetable production in greenhouses has developed rapidly from 15,000 ha in 1983 to 4.67 million ha in 2010, which accounted for >90% of the protected field area in the world (Chang et al., 2013). The average annual application rate of N to improve soil fertility and productivity in Chinese greenhouse vegetable production systems is > 3000 kg ha1 (Ju et al., 2006; Yu et al., 2010). The excessive fertilizer applied with furrow irrigation water resulted in substantial nitrate leaching (Min et al., 2011; Yan et al., 2013). Nitrate leaching has been considered the primary pathway of N loss in intensive vegetable production systems. Investigations into N leaching losses under greenhouse conditions have focused primarily on the measurement of mineral N (Min et al., 2012; Song
* Corresponding author. E-mail address: [email protected] (Q. Chen). http://dx.doi.org/10.1016/j.jenvman.2015.02.045 0301-4797/Crown Copyright © 2015 Published by Elsevier Ltd. All rights reserved.
et al., 2009; Zhou et al., 2010). Ju et al. (2006) reported that > 1000 kg nitrate-N ha1 was leached to a depth of 0.9e1.8 m in a greenhouse system on the North China plains. Serious nitrate-N leaching has been documented in vegetable greenhouse fields because of excessive input of N with furrow irrigation (Song et al., 2009) and the warm and moist conditions in the greenhouses, which are favourable for high N mineralization (Guo et al., 2010). The amount of SON leached from agricultural soils has been measured rarely because it is often assumed to be a negligible portion of the total N budget (van Kessel et al., 2009). Approximately half of the total N input was unaccounted for in the N balance calculations under a vegetable production system in northern China (Song et al., 2009), and the failure to consider dissolved organic N may, in part, explain this discrepancy. Embacher et al. (2008) indicated that the application of chemical N fertilizer increased the SON. This finding was confirmed by Ros et al. (2009), who reported that the application of chemical N fertilizer increased the SON by 17%. However, Liang et al. (2011)
352
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357
found that the addition of chemical N fertilizer alone did not increase the SON content over 17 years in a wheat-maize rotation system. Overall, the effect of chemical N fertilizer on the content of soil SON remains inconclusive. Consequently, it is necessary to evaluate the effects of excessive chemical N fertilizers on the SON content and SON accumulation in soil profile under a greenhouse vegetable production system with higher rates of irrigation. Therefore, the objective of the study was to investigate the effects of excessive inputs of chemical fertilizer N on the distribution of soluble organic N in the soil profile under a greenhouse vegetable production system. 2. Materials and methods
input of N from irrigation water was inevitable in this treatment. (2) Only organic amendment (manure þ straw) application (OrN): Air-dried chicken manure was broadcast and incorporated into the top 0e0.2 m of soil as a base fertilizer before transplanting from 2004 to 2012, and straw was incorporated along with the chicken manure beginning in June 2006. The rates of manure and straw application are shown in Table S1. (3) Reduced chemical N management (Re-N): Reduced chemical N fertilizer was side-dressed based on the Or-N treatment. The added chemical N amount was based on the mineral N content in the root zone (0e0.3 m) and the N target value. The equation used was as follows:
2.1. Experimental site This research was conducted in the Luojia village, Shouguang, Shandong province, China (36 550 N, 118 450 E). The site has a semihumid climate. The mean annual air temperature is 12.4 C with the highest mean monthly temperature of 26.5 C in July and lowest temperature of 3.0 C in January. The mean annual rainfall is 592 mm with 63% in summer and only 4% in winter. The frost-free season during the study was 196 days and lasted from mid-April until late October. The particle size distribution of the soil in the greenhouse was 46, 52, and 2% sand, silt, and clay, respectively, in the 0e0.3 m depth, and 37, 60, and 3% sand, silt, and clay, respectively, in the 0.3e0.6 m depth. Typical solar greenhouses in the region are constructed with clay walls and covered with a sheet of polyethylene film. The greenhouse in the research study was 8.5 m wide and 84 m long and oriented in a northesouth direction. 2.2. Tomato establishment and management Annual double-cropping of tomato (Lycopersicum esculentum Mill.), i.e., winterespring (WS) and autumnewinter (AW), was cultivated. Tomato seedlings were transplanted in mid-February and harvested in mid-June for the WS season. After a short-term summer fallow, tomato seedlings were transplanted for the AW season in early August and harvested the following January. The mean air and soil temperatures were 23.3 and 19.4 C in the WS season and 17.3 and 16.1 C in the AW season, respectively (Fig. S1). The tomato vines and stubble were removed from the plastic greenhouse at the final harvest. There was no rainfall in the greenhouse because of the cover of plastic film; hence, furrow irrigation was employed approximately 9e11 times, and approximately 60 mm of irrigation water was used for each irrigation. Prior to this experiment, which was conducted in 2004, the greenhouse had been cultivated for four years under conventional management. The P and K fertilizers were applied as monophosphate (12% P2O5) and potassium sulphate (50% K2O) at rates of 300 kg P2O5 ha1 and 400 kg K2O ha1 in each season. Half of the P and K fertilizers was applied in basal, and the remaining fertilizers were dissolved and dressed with the furrow irrigation water. Nine to eleven irrigation events took place in each growing season. The entire chemical N was dressed with the furrow irrigation water as urea. 2.3. Experimental design The experiment was designed as follows: (1) No organic and chemical N fertilizer was applied (No-N): No N was applied from organic and chemical fertilizer, but the
N ¼ N target value NO 3 eN amount in the top 0.3 m of the soil profile before the fertilization NO 3 eN from irrigation water The N target value used was at a rate of 300 kg N ha1 for the side-dressing at each stage of fruit cluster development in 2004. The N target values from transplanting to the third and fourth cluster growth stages were changed to 250 and 200 kg N ha1 in the WS season, and 200 and 250 kg N ha1 in the AW season from 2005. The N recommendation was simplified from AW season of 2007 based on the experiences of preceding years. Three or four side-dressing events with an interval of 7e10 days at a rate of 50 kg N ha1 were required in April and October (Ren et al., 2010). (4) Conventional N management (CoeN): Organic materials were applied at the same rate as in the Or-N treatment, and chemical N fertilizer was side-dressed in the tomato growing period based on conventional practice in the region with an N side-dressing rate of 120 kg N ha1 occasionally depending on the weather conditions and the tomato cultivar used in the different seasons. A completely randomized design with three replications was used. The plot size was 54.6 (7.0 m 7.8 m) for the Co-N and Re-N treatments, 32.8 (5.4 m 7.8 m) for the Or-N, and 21.8 m2 (7.8 m 2.8 m) for the No-N treatment. The base fertilizer was broadcast on the soil surface and incorporated into the soil by ploughing at the beginning of every growing season. The N input amount from manure, straw, chemical fertilizer, and irrigation is shown in Table S1. The timing of the top-dressing in the Re-N and Co-N treatments was based on the tomato cultivar of the season and the weather conditions. The chemical fertilizers for top-dressing were dissolved and applied with furrow irrigation water, and 9e11 irrigations occurred in each growing season. 2.4. Plant and soil sampling and analyses Tomato fruits were picked by hand and weighed at each harvest. Plant shoot samples were divided into leaf and stem and were collected at the end of the final harvest. All plant samples were dried in the oven at 70 C for over 48 h and were weighed before and after drying to determine water content. The total N in fruits, leaf, and stem was analysed using the Kjeldahl method after digestion with H2SO4eH2O2 (Thomas et al., 1967). Composite soil samples were obtained by mixing three soil cores from each plot and were collected in 0.3 m increments from 0 to 1.8 m to measure the mineral and soluble organic N after uprooting of the plants in June 2010, 2011, and 2012. After sieving the soil through a 2 or 5 mm sieve, subsamples were extracted with 1 M KCl at a ratio of 1:10 (dry soil: solution) for 1 h. Mineral N
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357 (NHþ 4 eN and NO3 eN) was determined using a continuous flow analyser (AA3, Bran þ Luebbe, Germany). The total soluble N (TSN) in the filtrate was measured by dual-wavelength (220 and 275 nm) ultraviolet spectrophotometry after alkaline persulfate oxidation (Norman et al., 1985). The soil soluble organic N (SON) was determined as the difference between TSN and mineral N. The annual apparent N loss (Nloss), including N leaching, N immobilization, ammonia volatilization and denitrification, was calculated according to Eq. (1). The calculation of N-use efficiency (NUE) refers to the ratio of the N absorbed by plants to the applied N from the chemical fertilizer, manure, straw, and irrigation water. The equation for the NUE calculation is shown in Eq. (2).
Nloss ¼ Norg þ Nirri þ Nfert Ncrop (Nmin
2012
Nmin
2004)/8
NUE ¼ Ncrop / (Norg þ Nfert þ Nirri),
(1) (2)
where Norg is average annual total N input from organic material, Nirri is the average annual total N input from irrigation water, Nfert is the average annual chemical fertilizer N input, Ncrop is the average annual total N uptake by aboveground parts and fruit of tomato, Nmin 2004 is the soil mineral N at 0e1.8 m in June 2004, and Nmin 2012 is the soil mineral N in the top 1.8 m of soil in June 2012. Based on the assumption that the accumulation of mineral N, SON, and TSN was slower with increasing applied N, we fitted an exponential model (Eq. (3)) to the measured accumulated mineral N, SON, and TSN. Cumulative mineral N, SON or TSN ¼ a * (1 eb*x),
(3)
where x is the annual rate of added N (kg N ha1), b is the cumulative rate constant, and a is the maximum accumulation of N in the 0e1.8 m of the soil. 2.5. Data analyses The data were analysed as a completely randomized design; the significance of treatment effects was determined using the analysis of variance. Multiple comparisons of the mean values were performed using the Fisher Least Significant Difference (LSD) test at P < 0.05. Statistical analyses were performed using version 8.0 of the SAS software package (SAS Institute Inc., Cary, NC, USA). Correlation analysis was applied to reveal the relationships between cumulative TSN, mineral N, or SON and the rates of N fertilizer applications using Sigmaplots (Systat Software Inc. Version 12.2). 3. Results and discussions 3.1. Proportions and sources of N input over 17 seasons During the 17 seasons of tomato production, the total N from manure, chemical fertilizer and irrigation water in the No-N, Or-N, Re-N, and Co-N treatments was 3286, 7418, 10,415, and 17,598 kg N ha1, respectively, with an average of 387, 873, 1225, and 2070 kg N ha1 yr1, respectively (Table 1). The average N input from irrigation was as high as 387 kg N ha1 yr1, accounting for 44, 32, and 19% of the total N input amount in the Or-N, Re-N, and Co-N treatments, respectively (Table S1). Previous research in this region showed that the N input from irrigation water varied from 59 to 603 kg ha1, with an average of 298 kg N ha1 during a growing season. The irrigation input accounted for 11%e47%, with an average of 32% of the total N input (Song et al., 2009). These findings were in agreement with our results for the Re-N and Co-N treatments. Thus, irrigation with
353
groundwater is an important source of N in vegetable production systems, and should be considered when estimating the added N rate in an area that is irrigated with groundwater. The N from manure accounted for 56, 40, and 23% of the total N input in the Or-N, Re-N, and Co-N treatments, respectively. In the Re-N and Co-N treatments, the chemical N represented 29 and 58% of the total N input, respectively. The total and chemical N inputs in the Re-N treatment were 41 and 72% lower, respectively, than the Co-N treatment (Table S1). 3.2. Tomato yield and N balance The tomato yield ranged from 53.0 to 130.0 and 54.2 to 115.4 t ha1 season1 over the 8.5 years study period, with an average of 85.5 t ha1 in the WS and 82.3 t ha1 in the AW seasons, respectively (Fig. 1). Fertilization significantly increased the tomato yield by 8.7%e12.0% in the WS season and by 11.3%e15.9% in the AW season relative to the No-N treatment. Although the N application rate was 40% higher in the Re-N and 137% higher in the Co-N treatments compared with the Or-N treatment, the tomato yield had no significant difference among the three treatments. This result clearly demonstrates that the excessive N application in the Co-N treatments had no benefit on tomato production and the high N inputs for greenhouse vegetables could be reduced effectively without yield loss (He et al., 2007). Total N removal in the plants occurred in the following order: Co-N ¼ Re-N > Or-N > No-N treatments in the WS season and CoN ¼ Re-N ¼ Or-N > No-N in the AW season (Fig. 1). The average N uptake in the No-N treatment was 201 and 286 kg N ha1 in the WS and AW seasons, respectively. Over the 17 seasons study period, the average total N uptake in the tomato plants in the Or-N, Re-N, and Co-N treatments was 326, 330, 345 kg N ha1 season1. The N-use efficiency was substantially different for the Or-N, Re-N, and Co-N treatments, with values of 64%, 48%, and 29%, respectively (Table 1). After 16 growing seasons from June 2004 to June 2012, the accumulated mineral N in the 0e1.8 m of soil significantly increased by 1059, 1577, 1268 kg N ha1 in the Or-N, Re-N, and Co-N treatments, respectively (Table 1). The annual surplus N (the gap between input and output) in the Co-N treatment was 1446 kg N ha1, of which 1287 kg N ha1 was lost. Using the same greenhouse that was used in the current study, Song (2012) reported that the annual N2O emissions ranged from 10.2 to 19.9 kg N ha1 yr1, accounting for 0.9%e1.5% of the annual N input. The NH3 volatilization was only 3.0e11.0 kg N ha1 yr1 in a greenhouse production system in the same area (Zhu et al., 2005). Thus, we suspected that leaching of N below 1.8 m in the soil profile may have accounted for about 1000 kg N ha1 yr1 under conventional greenhouse cultivation. This was agreed with Guo et al. (2010) who reported that annual N leaching was approximately 1000 kg N ha1 under conventional greenhouse vegetable planting. The annual apparent loss of N in the Re-N and Or-N treatments was significantly lower compared to that in the Co-N treatment. Compare with the Co-N treatment, the decreases in added and apparent N loss were similar (845 and 849 kg N ha1, respectively) in the ReeN treatment. This finding indicated that the entire additional N added in the Co-N treatment was lost compared with the Re-N treatment. 3.3. Mineral N in the soil profile After WS tomato harvesting in 2010, 2011, and 2012, the mineral N in the No-N treatment ranged from 18.7 to 65.2 mg N kg1 in the 0e1.8 m soil profile (Fig. 2a). Compared with the No-N treatment, the application of organic materials alone increased the mineral N content by 96% in 2011 and by 77% in 2012. The combined
354
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357
Table 1 Nitrogen balance in the long-term fertilization treatments including no N addition (No-N), only organic amendment treatment (Or-N), reduced chemical N management (ReN), and conventional N management (Co-N) during June 2004 and June 2012 in a greenhouse vegetable system. Treatments
N input (kg N ha1 yr1)
No-N Or-N Re-N Co-N
399 883 1217 2047
Accumulated mineral N in 0e1.8 m soil layer (kg N ha1) Jun 2004
a
554 730 756 1192
± ± ± ±
24 ba 124 b 136 b 182 a
N output by plants (kg N ha1 yr1)
N loss (kg N ha1 yr1)
N use efficiency (%)
101 190 438 1287
e 64 ± 3 a 48 ± 2 b 29 ± 1 c
Jun 2012 696 1789 2333 2460
± ± ± ±
117 c 90 b 153 a 133 a
483 561 582 601
± ± ± ±
23 12 15 29
b a a a
± ± ± ±
17 27 32 36
d c b a
Different lowercase letters indicates the significant (P < 0.05) difference of the average value between treatments (P < 0.05).
Fig. 1. Tomato fruit yields and total N uptake in plant (2004e2012) influenced by different N management practices (no N addition, No-N; only organic amendment (manure þ straw) application,Or-N; reduced chemical N management, Re-N; and conventional N management, Co-N) in greenhouse vegetable production system between 2004 and 2012. Different lowercase letters indicate the significant (P < 0.05) difference in the mean value (the dotted lines) between treatments. The short dash lines represent the mean values.
application of organic materials and chemical fertilizer significantly increased the mineral N content in the 0e1.8 m soil profile compared with the No-N and Or-N treatments. On average, the mineral N in the 0e1.8 m soil profile of the Re-N treatment was 1.7, 1.5, and 1.3 fold higher than that of the Or-N treatment in 2010, 2011, and 2012, respectively. The mineral N content was significantly higher in the Co-N treatment than in the Re-N treatment in 2010. The average cumulative mineral N in the 0e1.8 m of the soil in 2010e2012 was as high as 2068 kg N ha1 in the Co-N treatment (Fig. 3), similar to the findings of Ju et al. (2006) for a greenhouse vegetable production system. 3.4. Soluble organic N in the soil profile Soil soluble organic N in the 0e0.3 m soil depth ranged from 50 to 94 mg N kg1 in the Co-N treatment (Fig. 4). It was significantly higher than that in a cropland reported by Liang et al. (2011) and in a forest reported by Xing et al. (2010). Soluble organic N content for the No-N treatment in the 0e1.8 m soil profile ranged from 6.6 to 29.4 mg N kg1, which was significantly lower than the Or-N treatments in 2011 and 2012 (Fig. 2b). The application of organic materials in the Or-N treatment
significantly increased the SON in 0e0.3 cm layer by 32%e189% over the study period compared with the No-N treatment (Fig. 4). The Re-N treatment significantly increased the SON content in the 0e1.8 m soil profile relative to the Or-N treatment by 216, 48, and 33% in 2010, 2011, and 2012, respectively (Fig 2b). Overall, the application of additional chemical N fertilizer in the Co-N treatment significantly increased the SON content by 31e42% compared with the Re-N treatment for all three years. Generally, the cumulative SON amounts for each treatment in the 0e0.6, 0.6e1.2 and 1.2e1.8 m soil profiles were in the order of Co-N > Re-N > Or-N > No-N (Table 2). Over the study period the average cumulative SON levels in the 0e1.8 m soil profile of the NoN, Or-N, Re-N and Co-N treatments were 373, 849, 1449 and 1977 kg N ha1, respectively. Over the study period, the application of chemical N fertilizer in the Re-N and Co-N treatments significantly increased the cumulative SON by 33%e202%, and 82%e326%, respectively, compared with the Or-N treatment. The cumulative SON in the 0e1.8 m soil layers of the Co-N treatment was significantly higher by 41%, 30%, and 40% in 2010, 2011, and 2012, respectively, relative to the cumulative SON in the Re-N treatment. The cumulative TSN, mineral N and SON in the 0e1.8 m of the soil profile increased consistently with the rate of added N (Fig. 3),
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357
355
Fig. 2. Mineral N (a) and soluble organic N (b) extracted by 1 M KCl solution in the soil samples collected from 0 to 1.8 m soil profile after harvest of winterespring season tomato in 2010e2012 under no N fertilization (C), only organic amendment (manure þ straw) application (B), reduced chemical N management (;), and conventional N management ( ) in greenhouse vegetable production system.
▵
and there were significant (P < 0.05) correlations between the model and the measured data. The fitted a values were 7373, 5955, and 3353 kg N ha1 for the TSN, mineral N, and SON, respectively, indicating that the maximum amount of cumulative TSN, mineral N, and SON in the 0e1.8 m soil layer was 7373, 5955, and 3353 kg N ha1, respectively. These indicated that application of chemical N fertilizer not only increased mineral N, but also increased the accumulation of SON. Pellerin et al. (2006) stated in a review that the dissolved organic N concentrations in forested watersheds were correlated significantly with the N rates. Increases in the loss of dissolved organic N have been observed in forests (Fang et al., 2009; McDowell et al., 2004),
Fig. 3. Relationship between the N input amount and cumulated mineral N (B), soluble organic N (SON, ), and total soluble N (TSN, C) in 0e1.8 m soil profile (mean of 2010, 2011, and 2012) and the distribution of soluble organic N in 0e0.6 ( ), 0.6e1.2 ( ), and 1.2e1.8 m ( )soil layers after harvesting the winterespring tomato under no N fertilization (No-N), only organic amendment (manure þ straw) application (Or-N), reduced chemical N management (Re-N), and conventional N management (Co-N) in greenhouse vegetable production system.
▵
Fig. 4. Soluble organic N extracted by 1 M KCl solution in 0e0.3 m soil layer in greenhouse tomato double cropping system under no N addition (No-N), only organic amendment (manure þ straw) application (Or-N), reduced chemical N management (Re-N), and conventional N management (Co-N) from Aug. 2011 to Jul. 2012.
356
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357
Table 2 Cumulative soluble organic N (kg N ha1) in 0e0.6, 0.6e1.2, 1.2e1.8, and 0e1.8-m soil profile in a greenhouse vegetable system under no N addition (No-N), only organic amendment (Or-N), reduced chemical N management (Re-N), and conventional N management (Co-N) during 2010e2012. Sampling Soil layers No-N time (m) 2010
2011
2012
0e0.6 0.6e1.2 1.2e1.8 0e1.8 0e0.6 0.6e1.2 1.2e1.8 0e1.8 0e0.6 0.6e1.2 1.2e1.8 0e1.8
203 113 75 391 118 78 146 341 93 172 123 387
Or-N
Re-N
Sampling time
Co-N
± 5 ca 153 ± 22 c 468 ± 44 b 713 ± 73 a ± 19 c 163 ± 25 c 525 ± 54 b 688 ± 31 a ±8c 147 ± 25 c 407 ± 43 b 572 ± 138 a ± 20 c 463 ± 70 c 1401 ± 89 b 1973 ± 83 a ± 38 c 292 ± 6 b 448 ± 57 a 556 ± 139 a ± 23 d 476 ± 44 c 648 ± 19 b 754 ± 20 a ± 30 d 319 ± 51 c 522 ± 20 b 795 ± 7 a ± 71 d 1087 ± 95 c 1618 ± 64 b 2105 ± 115 a ±7d 266 ± 29 c 355 ± 55 b 562 ± 50 a ± 42 d 347 ± 31 c 463 ± 63 b 574 ± 31 a ±5d 336 ± 38 c 510 ± 90 b 718 ± 57 a ± 44 d 997 ± 142 c 1328 ± 190 b 1855 ± 65 a
a Different lowercase letters indicates the significant (P < 0.05) difference between the means between different N management treatments.
grasslands (Dijkstra et al., 2007), and croplands (Huang et al., 2011), where the rate of chemical N addition was increased. Dijkstra et al. (2007) and Embacher et al. (2008) attributed the increased dissolved organic N loss to increased plant residue and rhizodeposition because of the increased N addition. However, the similar tomato yields in our study indicated the plant residue and rhizodeposition were similar among the fertilization treatments. There was also a possibility that the N additions increased enzyme activities (West et al., 2006; Zak et al., 2011), suggesting greater gross N mineralization rates and potentially greater SON production. McDowell et al. (2004) suggested that because of the change of microbial community structure and an increase in the production of N-rich fractions, the addition of chemical N fertilizers increased the hydrophilic character of organic N. In addition to the biotic process, applied chemical N may be converted to SON through abiotic processes (Davidson et al., 2003; Huygens et al., 2008). Additionally, excessive chemical N fertilizer leads to the rapid acidification of soils (Song et al., 2012) and solubilizes soil organic N ka si (2011) because of the pH effect (Chantigny, 2003). Filep and Re also reported that the dissolved organic N concentration increased with decreasing soil pH. The cumulative SON in the 0e0.6, 0.6e1.2, and 1.2e1.8 m soil layers accounted for 31%, 36%, and 33% of the cumulative SON in the 0e1.8 m soil profile across the different types of N management, respectively (Fig. 3), and there were no significant (P > 0.05) differences in SON/TSN among the 0e0.6, 0.6e1.2, and 1.2e1.8 m soil layers (Table 3). The similar SON/TSN values in the 0e0.6, 0.6e1.2, and 1.2e1.8 m soil layers showed that the distribution of SON in the 0e1.8 m soil profile under this greenhouse vegetable production system was uniform, and the SON leaching in the 0e1.8 m soil
Table 3 F-statistics from three-way ANOVA of the effect of years (2010, 2011, and 2012), soil layers (0e0.6, 0.6e1.2, and 1.2e1.8 cm), and N fertilization (no N fertilization, only organic amendment treatment, reducing N management, and conventional N management) on the percentage of cumulative soluble organic N in the total soluble N.
Years Soil layers N fertilization Years Soil layers Years N fertilization Soil layers N fertilization Years Soil layers N fertilization
Table 4 The cumulative soil soluble organic N as percentage of total soluble N in 0e1.8-m soil profile in a greenhouse vegetable system under no N addition (No-N), only organic amendment (Or-N), reduced chemical N management (Re-N), and conventional N management (Co-N) at Jun of 2010, 2011, and 2012.
DF
F
P
2 2 3 4 6 6 12
19.26 1.57 70.96 25.19 6.96 5.51 5.59
Contents lists available at ScienceDirect
Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman
The impact of exogenous N supply on soluble organic nitrogen dynamics and nitrogen balance in a greenhouse vegetable system Bin Liang a, Lingyun Kang b, Tao Ren c, Li Junliang a, Qing Chen b, *, Jingguo Wang b a
College of Resources and Environmental Sciences, Qingdao Agriculture University, Qingdao 266109, China College of Resources and Environmental Sciences, China Agricultural University, Beijing 100193, China c College of Resources & Environment, Huazhong Agriculture University, Wuhan 430070, China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 September 2014 Received in revised form 22 February 2015 Accepted 27 February 2015 Available online 9 March 2015
A long-term greenhouse experiment (2004e2012) was conducted with continuous tomato (Lycopersicum esculentum Mill.) plantings to understand the influence of an exogenous nitrogen supply from irrigation water, chemical fertilizer, or organic amendment on the N balance and soluble organic nitrogen (SON). The results from 16 tomato growing seasons indicated that the application of organic amendment (manure and straw) alone (Or-N) resulted in the same yield as the conventional chemical N with organic amendment (Co-N) and the reduced chemical N with organic amendment (Re-N) treatments. The annual apparent N loss was >1000 and 438 kg N ha1 in the Co-N and Re-N treatments, respectively. Over the study period, the SON in the 1.8 m soil profile was 1449 and 1978 kg N ha1 in the Re-N and Co-N treatments, respectively, it was 1.7- and 2.3-fold higher than that observed in the Or-N treatment, which indicated that SON increased with the chemical N application. The percentage of SON in the cumulative soluble N (SON plus mineral N) ranged from 28% to 44%, and there were no significant differences across the 0e0.6, 0.6e1.2, and 1.2e1.8 m soil profile, which indicated that the leaching and distribution of SON was similar to those of the mineral N in the 0e1.8 m soil profile. We conclude that the mobility of soluble organic N in the 0e1.8 m of the soil was synchronous with the mineral N under a greenhouse production system, and the risk of soluble organic N leaching increased with inorganic N application rate. Therefore, leaching of SON in the intensive agriculture should not be ignored when evaluating the risk of N leaching. Crown Copyright © 2015 Published by Elsevier Ltd. All rights reserved.
Keywords: N fertilizer Manure Irrigation water Mineral N Plastic greenhouse
1. Introduction Vegetable production in greenhouses has developed rapidly from 15,000 ha in 1983 to 4.67 million ha in 2010, which accounted for >90% of the protected field area in the world (Chang et al., 2013). The average annual application rate of N to improve soil fertility and productivity in Chinese greenhouse vegetable production systems is > 3000 kg ha1 (Ju et al., 2006; Yu et al., 2010). The excessive fertilizer applied with furrow irrigation water resulted in substantial nitrate leaching (Min et al., 2011; Yan et al., 2013). Nitrate leaching has been considered the primary pathway of N loss in intensive vegetable production systems. Investigations into N leaching losses under greenhouse conditions have focused primarily on the measurement of mineral N (Min et al., 2012; Song
* Corresponding author. E-mail address: [email protected] (Q. Chen). http://dx.doi.org/10.1016/j.jenvman.2015.02.045 0301-4797/Crown Copyright © 2015 Published by Elsevier Ltd. All rights reserved.
et al., 2009; Zhou et al., 2010). Ju et al. (2006) reported that > 1000 kg nitrate-N ha1 was leached to a depth of 0.9e1.8 m in a greenhouse system on the North China plains. Serious nitrate-N leaching has been documented in vegetable greenhouse fields because of excessive input of N with furrow irrigation (Song et al., 2009) and the warm and moist conditions in the greenhouses, which are favourable for high N mineralization (Guo et al., 2010). The amount of SON leached from agricultural soils has been measured rarely because it is often assumed to be a negligible portion of the total N budget (van Kessel et al., 2009). Approximately half of the total N input was unaccounted for in the N balance calculations under a vegetable production system in northern China (Song et al., 2009), and the failure to consider dissolved organic N may, in part, explain this discrepancy. Embacher et al. (2008) indicated that the application of chemical N fertilizer increased the SON. This finding was confirmed by Ros et al. (2009), who reported that the application of chemical N fertilizer increased the SON by 17%. However, Liang et al. (2011)
352
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357
found that the addition of chemical N fertilizer alone did not increase the SON content over 17 years in a wheat-maize rotation system. Overall, the effect of chemical N fertilizer on the content of soil SON remains inconclusive. Consequently, it is necessary to evaluate the effects of excessive chemical N fertilizers on the SON content and SON accumulation in soil profile under a greenhouse vegetable production system with higher rates of irrigation. Therefore, the objective of the study was to investigate the effects of excessive inputs of chemical fertilizer N on the distribution of soluble organic N in the soil profile under a greenhouse vegetable production system. 2. Materials and methods
input of N from irrigation water was inevitable in this treatment. (2) Only organic amendment (manure þ straw) application (OrN): Air-dried chicken manure was broadcast and incorporated into the top 0e0.2 m of soil as a base fertilizer before transplanting from 2004 to 2012, and straw was incorporated along with the chicken manure beginning in June 2006. The rates of manure and straw application are shown in Table S1. (3) Reduced chemical N management (Re-N): Reduced chemical N fertilizer was side-dressed based on the Or-N treatment. The added chemical N amount was based on the mineral N content in the root zone (0e0.3 m) and the N target value. The equation used was as follows:
2.1. Experimental site This research was conducted in the Luojia village, Shouguang, Shandong province, China (36 550 N, 118 450 E). The site has a semihumid climate. The mean annual air temperature is 12.4 C with the highest mean monthly temperature of 26.5 C in July and lowest temperature of 3.0 C in January. The mean annual rainfall is 592 mm with 63% in summer and only 4% in winter. The frost-free season during the study was 196 days and lasted from mid-April until late October. The particle size distribution of the soil in the greenhouse was 46, 52, and 2% sand, silt, and clay, respectively, in the 0e0.3 m depth, and 37, 60, and 3% sand, silt, and clay, respectively, in the 0.3e0.6 m depth. Typical solar greenhouses in the region are constructed with clay walls and covered with a sheet of polyethylene film. The greenhouse in the research study was 8.5 m wide and 84 m long and oriented in a northesouth direction. 2.2. Tomato establishment and management Annual double-cropping of tomato (Lycopersicum esculentum Mill.), i.e., winterespring (WS) and autumnewinter (AW), was cultivated. Tomato seedlings were transplanted in mid-February and harvested in mid-June for the WS season. After a short-term summer fallow, tomato seedlings were transplanted for the AW season in early August and harvested the following January. The mean air and soil temperatures were 23.3 and 19.4 C in the WS season and 17.3 and 16.1 C in the AW season, respectively (Fig. S1). The tomato vines and stubble were removed from the plastic greenhouse at the final harvest. There was no rainfall in the greenhouse because of the cover of plastic film; hence, furrow irrigation was employed approximately 9e11 times, and approximately 60 mm of irrigation water was used for each irrigation. Prior to this experiment, which was conducted in 2004, the greenhouse had been cultivated for four years under conventional management. The P and K fertilizers were applied as monophosphate (12% P2O5) and potassium sulphate (50% K2O) at rates of 300 kg P2O5 ha1 and 400 kg K2O ha1 in each season. Half of the P and K fertilizers was applied in basal, and the remaining fertilizers were dissolved and dressed with the furrow irrigation water. Nine to eleven irrigation events took place in each growing season. The entire chemical N was dressed with the furrow irrigation water as urea. 2.3. Experimental design The experiment was designed as follows: (1) No organic and chemical N fertilizer was applied (No-N): No N was applied from organic and chemical fertilizer, but the
N ¼ N target value NO 3 eN amount in the top 0.3 m of the soil profile before the fertilization NO 3 eN from irrigation water The N target value used was at a rate of 300 kg N ha1 for the side-dressing at each stage of fruit cluster development in 2004. The N target values from transplanting to the third and fourth cluster growth stages were changed to 250 and 200 kg N ha1 in the WS season, and 200 and 250 kg N ha1 in the AW season from 2005. The N recommendation was simplified from AW season of 2007 based on the experiences of preceding years. Three or four side-dressing events with an interval of 7e10 days at a rate of 50 kg N ha1 were required in April and October (Ren et al., 2010). (4) Conventional N management (CoeN): Organic materials were applied at the same rate as in the Or-N treatment, and chemical N fertilizer was side-dressed in the tomato growing period based on conventional practice in the region with an N side-dressing rate of 120 kg N ha1 occasionally depending on the weather conditions and the tomato cultivar used in the different seasons. A completely randomized design with three replications was used. The plot size was 54.6 (7.0 m 7.8 m) for the Co-N and Re-N treatments, 32.8 (5.4 m 7.8 m) for the Or-N, and 21.8 m2 (7.8 m 2.8 m) for the No-N treatment. The base fertilizer was broadcast on the soil surface and incorporated into the soil by ploughing at the beginning of every growing season. The N input amount from manure, straw, chemical fertilizer, and irrigation is shown in Table S1. The timing of the top-dressing in the Re-N and Co-N treatments was based on the tomato cultivar of the season and the weather conditions. The chemical fertilizers for top-dressing were dissolved and applied with furrow irrigation water, and 9e11 irrigations occurred in each growing season. 2.4. Plant and soil sampling and analyses Tomato fruits were picked by hand and weighed at each harvest. Plant shoot samples were divided into leaf and stem and were collected at the end of the final harvest. All plant samples were dried in the oven at 70 C for over 48 h and were weighed before and after drying to determine water content. The total N in fruits, leaf, and stem was analysed using the Kjeldahl method after digestion with H2SO4eH2O2 (Thomas et al., 1967). Composite soil samples were obtained by mixing three soil cores from each plot and were collected in 0.3 m increments from 0 to 1.8 m to measure the mineral and soluble organic N after uprooting of the plants in June 2010, 2011, and 2012. After sieving the soil through a 2 or 5 mm sieve, subsamples were extracted with 1 M KCl at a ratio of 1:10 (dry soil: solution) for 1 h. Mineral N
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357 (NHþ 4 eN and NO3 eN) was determined using a continuous flow analyser (AA3, Bran þ Luebbe, Germany). The total soluble N (TSN) in the filtrate was measured by dual-wavelength (220 and 275 nm) ultraviolet spectrophotometry after alkaline persulfate oxidation (Norman et al., 1985). The soil soluble organic N (SON) was determined as the difference between TSN and mineral N. The annual apparent N loss (Nloss), including N leaching, N immobilization, ammonia volatilization and denitrification, was calculated according to Eq. (1). The calculation of N-use efficiency (NUE) refers to the ratio of the N absorbed by plants to the applied N from the chemical fertilizer, manure, straw, and irrigation water. The equation for the NUE calculation is shown in Eq. (2).
Nloss ¼ Norg þ Nirri þ Nfert Ncrop (Nmin
2012
Nmin
2004)/8
NUE ¼ Ncrop / (Norg þ Nfert þ Nirri),
(1) (2)
where Norg is average annual total N input from organic material, Nirri is the average annual total N input from irrigation water, Nfert is the average annual chemical fertilizer N input, Ncrop is the average annual total N uptake by aboveground parts and fruit of tomato, Nmin 2004 is the soil mineral N at 0e1.8 m in June 2004, and Nmin 2012 is the soil mineral N in the top 1.8 m of soil in June 2012. Based on the assumption that the accumulation of mineral N, SON, and TSN was slower with increasing applied N, we fitted an exponential model (Eq. (3)) to the measured accumulated mineral N, SON, and TSN. Cumulative mineral N, SON or TSN ¼ a * (1 eb*x),
(3)
where x is the annual rate of added N (kg N ha1), b is the cumulative rate constant, and a is the maximum accumulation of N in the 0e1.8 m of the soil. 2.5. Data analyses The data were analysed as a completely randomized design; the significance of treatment effects was determined using the analysis of variance. Multiple comparisons of the mean values were performed using the Fisher Least Significant Difference (LSD) test at P < 0.05. Statistical analyses were performed using version 8.0 of the SAS software package (SAS Institute Inc., Cary, NC, USA). Correlation analysis was applied to reveal the relationships between cumulative TSN, mineral N, or SON and the rates of N fertilizer applications using Sigmaplots (Systat Software Inc. Version 12.2). 3. Results and discussions 3.1. Proportions and sources of N input over 17 seasons During the 17 seasons of tomato production, the total N from manure, chemical fertilizer and irrigation water in the No-N, Or-N, Re-N, and Co-N treatments was 3286, 7418, 10,415, and 17,598 kg N ha1, respectively, with an average of 387, 873, 1225, and 2070 kg N ha1 yr1, respectively (Table 1). The average N input from irrigation was as high as 387 kg N ha1 yr1, accounting for 44, 32, and 19% of the total N input amount in the Or-N, Re-N, and Co-N treatments, respectively (Table S1). Previous research in this region showed that the N input from irrigation water varied from 59 to 603 kg ha1, with an average of 298 kg N ha1 during a growing season. The irrigation input accounted for 11%e47%, with an average of 32% of the total N input (Song et al., 2009). These findings were in agreement with our results for the Re-N and Co-N treatments. Thus, irrigation with
353
groundwater is an important source of N in vegetable production systems, and should be considered when estimating the added N rate in an area that is irrigated with groundwater. The N from manure accounted for 56, 40, and 23% of the total N input in the Or-N, Re-N, and Co-N treatments, respectively. In the Re-N and Co-N treatments, the chemical N represented 29 and 58% of the total N input, respectively. The total and chemical N inputs in the Re-N treatment were 41 and 72% lower, respectively, than the Co-N treatment (Table S1). 3.2. Tomato yield and N balance The tomato yield ranged from 53.0 to 130.0 and 54.2 to 115.4 t ha1 season1 over the 8.5 years study period, with an average of 85.5 t ha1 in the WS and 82.3 t ha1 in the AW seasons, respectively (Fig. 1). Fertilization significantly increased the tomato yield by 8.7%e12.0% in the WS season and by 11.3%e15.9% in the AW season relative to the No-N treatment. Although the N application rate was 40% higher in the Re-N and 137% higher in the Co-N treatments compared with the Or-N treatment, the tomato yield had no significant difference among the three treatments. This result clearly demonstrates that the excessive N application in the Co-N treatments had no benefit on tomato production and the high N inputs for greenhouse vegetables could be reduced effectively without yield loss (He et al., 2007). Total N removal in the plants occurred in the following order: Co-N ¼ Re-N > Or-N > No-N treatments in the WS season and CoN ¼ Re-N ¼ Or-N > No-N in the AW season (Fig. 1). The average N uptake in the No-N treatment was 201 and 286 kg N ha1 in the WS and AW seasons, respectively. Over the 17 seasons study period, the average total N uptake in the tomato plants in the Or-N, Re-N, and Co-N treatments was 326, 330, 345 kg N ha1 season1. The N-use efficiency was substantially different for the Or-N, Re-N, and Co-N treatments, with values of 64%, 48%, and 29%, respectively (Table 1). After 16 growing seasons from June 2004 to June 2012, the accumulated mineral N in the 0e1.8 m of soil significantly increased by 1059, 1577, 1268 kg N ha1 in the Or-N, Re-N, and Co-N treatments, respectively (Table 1). The annual surplus N (the gap between input and output) in the Co-N treatment was 1446 kg N ha1, of which 1287 kg N ha1 was lost. Using the same greenhouse that was used in the current study, Song (2012) reported that the annual N2O emissions ranged from 10.2 to 19.9 kg N ha1 yr1, accounting for 0.9%e1.5% of the annual N input. The NH3 volatilization was only 3.0e11.0 kg N ha1 yr1 in a greenhouse production system in the same area (Zhu et al., 2005). Thus, we suspected that leaching of N below 1.8 m in the soil profile may have accounted for about 1000 kg N ha1 yr1 under conventional greenhouse cultivation. This was agreed with Guo et al. (2010) who reported that annual N leaching was approximately 1000 kg N ha1 under conventional greenhouse vegetable planting. The annual apparent loss of N in the Re-N and Or-N treatments was significantly lower compared to that in the Co-N treatment. Compare with the Co-N treatment, the decreases in added and apparent N loss were similar (845 and 849 kg N ha1, respectively) in the ReeN treatment. This finding indicated that the entire additional N added in the Co-N treatment was lost compared with the Re-N treatment. 3.3. Mineral N in the soil profile After WS tomato harvesting in 2010, 2011, and 2012, the mineral N in the No-N treatment ranged from 18.7 to 65.2 mg N kg1 in the 0e1.8 m soil profile (Fig. 2a). Compared with the No-N treatment, the application of organic materials alone increased the mineral N content by 96% in 2011 and by 77% in 2012. The combined
354
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357
Table 1 Nitrogen balance in the long-term fertilization treatments including no N addition (No-N), only organic amendment treatment (Or-N), reduced chemical N management (ReN), and conventional N management (Co-N) during June 2004 and June 2012 in a greenhouse vegetable system. Treatments
N input (kg N ha1 yr1)
No-N Or-N Re-N Co-N
399 883 1217 2047
Accumulated mineral N in 0e1.8 m soil layer (kg N ha1) Jun 2004
a
554 730 756 1192
± ± ± ±
24 ba 124 b 136 b 182 a
N output by plants (kg N ha1 yr1)
N loss (kg N ha1 yr1)
N use efficiency (%)
101 190 438 1287
e 64 ± 3 a 48 ± 2 b 29 ± 1 c
Jun 2012 696 1789 2333 2460
± ± ± ±
117 c 90 b 153 a 133 a
483 561 582 601
± ± ± ±
23 12 15 29
b a a a
± ± ± ±
17 27 32 36
d c b a
Different lowercase letters indicates the significant (P < 0.05) difference of the average value between treatments (P < 0.05).
Fig. 1. Tomato fruit yields and total N uptake in plant (2004e2012) influenced by different N management practices (no N addition, No-N; only organic amendment (manure þ straw) application,Or-N; reduced chemical N management, Re-N; and conventional N management, Co-N) in greenhouse vegetable production system between 2004 and 2012. Different lowercase letters indicate the significant (P < 0.05) difference in the mean value (the dotted lines) between treatments. The short dash lines represent the mean values.
application of organic materials and chemical fertilizer significantly increased the mineral N content in the 0e1.8 m soil profile compared with the No-N and Or-N treatments. On average, the mineral N in the 0e1.8 m soil profile of the Re-N treatment was 1.7, 1.5, and 1.3 fold higher than that of the Or-N treatment in 2010, 2011, and 2012, respectively. The mineral N content was significantly higher in the Co-N treatment than in the Re-N treatment in 2010. The average cumulative mineral N in the 0e1.8 m of the soil in 2010e2012 was as high as 2068 kg N ha1 in the Co-N treatment (Fig. 3), similar to the findings of Ju et al. (2006) for a greenhouse vegetable production system. 3.4. Soluble organic N in the soil profile Soil soluble organic N in the 0e0.3 m soil depth ranged from 50 to 94 mg N kg1 in the Co-N treatment (Fig. 4). It was significantly higher than that in a cropland reported by Liang et al. (2011) and in a forest reported by Xing et al. (2010). Soluble organic N content for the No-N treatment in the 0e1.8 m soil profile ranged from 6.6 to 29.4 mg N kg1, which was significantly lower than the Or-N treatments in 2011 and 2012 (Fig. 2b). The application of organic materials in the Or-N treatment
significantly increased the SON in 0e0.3 cm layer by 32%e189% over the study period compared with the No-N treatment (Fig. 4). The Re-N treatment significantly increased the SON content in the 0e1.8 m soil profile relative to the Or-N treatment by 216, 48, and 33% in 2010, 2011, and 2012, respectively (Fig 2b). Overall, the application of additional chemical N fertilizer in the Co-N treatment significantly increased the SON content by 31e42% compared with the Re-N treatment for all three years. Generally, the cumulative SON amounts for each treatment in the 0e0.6, 0.6e1.2 and 1.2e1.8 m soil profiles were in the order of Co-N > Re-N > Or-N > No-N (Table 2). Over the study period the average cumulative SON levels in the 0e1.8 m soil profile of the NoN, Or-N, Re-N and Co-N treatments were 373, 849, 1449 and 1977 kg N ha1, respectively. Over the study period, the application of chemical N fertilizer in the Re-N and Co-N treatments significantly increased the cumulative SON by 33%e202%, and 82%e326%, respectively, compared with the Or-N treatment. The cumulative SON in the 0e1.8 m soil layers of the Co-N treatment was significantly higher by 41%, 30%, and 40% in 2010, 2011, and 2012, respectively, relative to the cumulative SON in the Re-N treatment. The cumulative TSN, mineral N and SON in the 0e1.8 m of the soil profile increased consistently with the rate of added N (Fig. 3),
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357
355
Fig. 2. Mineral N (a) and soluble organic N (b) extracted by 1 M KCl solution in the soil samples collected from 0 to 1.8 m soil profile after harvest of winterespring season tomato in 2010e2012 under no N fertilization (C), only organic amendment (manure þ straw) application (B), reduced chemical N management (;), and conventional N management ( ) in greenhouse vegetable production system.
▵
and there were significant (P < 0.05) correlations between the model and the measured data. The fitted a values were 7373, 5955, and 3353 kg N ha1 for the TSN, mineral N, and SON, respectively, indicating that the maximum amount of cumulative TSN, mineral N, and SON in the 0e1.8 m soil layer was 7373, 5955, and 3353 kg N ha1, respectively. These indicated that application of chemical N fertilizer not only increased mineral N, but also increased the accumulation of SON. Pellerin et al. (2006) stated in a review that the dissolved organic N concentrations in forested watersheds were correlated significantly with the N rates. Increases in the loss of dissolved organic N have been observed in forests (Fang et al., 2009; McDowell et al., 2004),
Fig. 3. Relationship between the N input amount and cumulated mineral N (B), soluble organic N (SON, ), and total soluble N (TSN, C) in 0e1.8 m soil profile (mean of 2010, 2011, and 2012) and the distribution of soluble organic N in 0e0.6 ( ), 0.6e1.2 ( ), and 1.2e1.8 m ( )soil layers after harvesting the winterespring tomato under no N fertilization (No-N), only organic amendment (manure þ straw) application (Or-N), reduced chemical N management (Re-N), and conventional N management (Co-N) in greenhouse vegetable production system.
▵
Fig. 4. Soluble organic N extracted by 1 M KCl solution in 0e0.3 m soil layer in greenhouse tomato double cropping system under no N addition (No-N), only organic amendment (manure þ straw) application (Or-N), reduced chemical N management (Re-N), and conventional N management (Co-N) from Aug. 2011 to Jul. 2012.
356
B. Liang et al. / Journal of Environmental Management 154 (2015) 351e357
Table 2 Cumulative soluble organic N (kg N ha1) in 0e0.6, 0.6e1.2, 1.2e1.8, and 0e1.8-m soil profile in a greenhouse vegetable system under no N addition (No-N), only organic amendment (Or-N), reduced chemical N management (Re-N), and conventional N management (Co-N) during 2010e2012. Sampling Soil layers No-N time (m) 2010
2011
2012
0e0.6 0.6e1.2 1.2e1.8 0e1.8 0e0.6 0.6e1.2 1.2e1.8 0e1.8 0e0.6 0.6e1.2 1.2e1.8 0e1.8
203 113 75 391 118 78 146 341 93 172 123 387
Or-N
Re-N
Sampling time
Co-N
± 5 ca 153 ± 22 c 468 ± 44 b 713 ± 73 a ± 19 c 163 ± 25 c 525 ± 54 b 688 ± 31 a ±8c 147 ± 25 c 407 ± 43 b 572 ± 138 a ± 20 c 463 ± 70 c 1401 ± 89 b 1973 ± 83 a ± 38 c 292 ± 6 b 448 ± 57 a 556 ± 139 a ± 23 d 476 ± 44 c 648 ± 19 b 754 ± 20 a ± 30 d 319 ± 51 c 522 ± 20 b 795 ± 7 a ± 71 d 1087 ± 95 c 1618 ± 64 b 2105 ± 115 a ±7d 266 ± 29 c 355 ± 55 b 562 ± 50 a ± 42 d 347 ± 31 c 463 ± 63 b 574 ± 31 a ±5d 336 ± 38 c 510 ± 90 b 718 ± 57 a ± 44 d 997 ± 142 c 1328 ± 190 b 1855 ± 65 a
a Different lowercase letters indicates the significant (P < 0.05) difference between the means between different N management treatments.
grasslands (Dijkstra et al., 2007), and croplands (Huang et al., 2011), where the rate of chemical N addition was increased. Dijkstra et al. (2007) and Embacher et al. (2008) attributed the increased dissolved organic N loss to increased plant residue and rhizodeposition because of the increased N addition. However, the similar tomato yields in our study indicated the plant residue and rhizodeposition were similar among the fertilization treatments. There was also a possibility that the N additions increased enzyme activities (West et al., 2006; Zak et al., 2011), suggesting greater gross N mineralization rates and potentially greater SON production. McDowell et al. (2004) suggested that because of the change of microbial community structure and an increase in the production of N-rich fractions, the addition of chemical N fertilizers increased the hydrophilic character of organic N. In addition to the biotic process, applied chemical N may be converted to SON through abiotic processes (Davidson et al., 2003; Huygens et al., 2008). Additionally, excessive chemical N fertilizer leads to the rapid acidification of soils (Song et al., 2012) and solubilizes soil organic N ka si (2011) because of the pH effect (Chantigny, 2003). Filep and Re also reported that the dissolved organic N concentration increased with decreasing soil pH. The cumulative SON in the 0e0.6, 0.6e1.2, and 1.2e1.8 m soil layers accounted for 31%, 36%, and 33% of the cumulative SON in the 0e1.8 m soil profile across the different types of N management, respectively (Fig. 3), and there were no significant (P > 0.05) differences in SON/TSN among the 0e0.6, 0.6e1.2, and 1.2e1.8 m soil layers (Table 3). The similar SON/TSN values in the 0e0.6, 0.6e1.2, and 1.2e1.8 m soil layers showed that the distribution of SON in the 0e1.8 m soil profile under this greenhouse vegetable production system was uniform, and the SON leaching in the 0e1.8 m soil
Table 3 F-statistics from three-way ANOVA of the effect of years (2010, 2011, and 2012), soil layers (0e0.6, 0.6e1.2, and 1.2e1.8 cm), and N fertilization (no N fertilization, only organic amendment treatment, reducing N management, and conventional N management) on the percentage of cumulative soluble organic N in the total soluble N.
Years Soil layers N fertilization Years Soil layers Years N fertilization Soil layers N fertilization Years Soil layers N fertilization
Table 4 The cumulative soil soluble organic N as percentage of total soluble N in 0e1.8-m soil profile in a greenhouse vegetable system under no N addition (No-N), only organic amendment (Or-N), reduced chemical N management (Re-N), and conventional N management (Co-N) at Jun of 2010, 2011, and 2012.
DF
F
P
2 2 3 4 6 6 12
19.26 1.57 70.96 25.19 6.96 5.51 5.59
