Excessive fertilization leads to high nitrogen (N) leaching under intensive plastic-shed vegetable production systems, and thereby results in the contaminations of ground or surface water. Therefore, it is urgent to develop cost-effective strategies of nitrogen management to overcome these obstacles. A 15-year experiment in annual double-cropping systems was conducted to explore impacts of N application rate and straw amendment on mineral N leaching loss in plastic-shed greenhouse. The results showed that seasonal mineral N leaching was up to 103.4–603.4 kg N ha⁻¹, accounting for 12%–41% of total N input under conventional N fertilization management. However, optimized N application rates by 47% and straw addition obviously decreased mineral N leaching by 4%–86%, while had no negative impacts on N uptake and tomato yields. These large decreases of N leaching loss were mainly due to the reduced leachate amount and followed by N concentration in leachate, which was supported by improved soil water holding capacity after optimizing N application rates and straw addition. On average, 52% of water leachate and 55% of mineral N leaching simultaneously occurred within 40 days after planting, further indicating the dominant role of water leakage in regulating mineral N leaching loss. Moreover, decreasing mineral N leaching was beneficial for reducing leaching loss of base cations. Therefore, optimized N application rates and straw amendment effectively alleviates mineral N leaching losses mainly by controlling the water leakage without yield loss in plastic-shed greenhouse, making this strategy promising and interesting from environmental and economical viewpoints.

Overuse of phosphorus (P) fertilizer and the resulting soil P accumulation in vegetable production increases the risk of P runoff and leaching. However, P transformations under continuous fertilization and their effects on environmental risk are unclear. The current study examined the effects of long-term P fertilizer application on P fractions in different soil layers, and assessed the correlations between P fractions and environmental risks in intensive vegetable production in a subtropical region. A total of 32 fields were studied, including 8 uncultivated fields and 24 fields continuously used for vegetable production for 1–3, 4–9, or 10–15 years. The results showed that excessive P fertilizer input caused soil P surpluses ranging from 204.6 to 252.4 kg ha⁻¹ yr⁻¹. Compared to uncultivated fields, vegetable fields contained higher levels of labile P, moderately labile P, sparingly labile P, and non-labile P. The combined percentage of labile P and moderately labile P increased from 55.2% in fields cultivated for 0–3 year to 65.5% in fields cultivated for 10–15 years. The concentrations of soil P fractions were higher at 0–20 cm soil depth than at 20–40 and 40–60 cm soil depth. Soil available P was positively correlated with all soil P fractions except diluted HCl-Pᵢ or concentrated HCl-Pₒ. Long-term vegetable production increased CaCl₂–P downward movement, which was positively correlated with levels of labile and moderately labile P. The P index indicated a high risk of P losses from the vegetable fields. The P index was on average 3.27-fold higher in the vegetable fields than in uncultivated fields, and was significantly correlated with soil available P and organic and inorganic P fertilizer input. The environmental risk caused by P in vegetable production should be reduced by reducing P fertilizer input so as to maintain soil available P within an optimal range for vegetable production.

China accounts for more than half of the world's vegetable production, and identifying the contribution of vegetable production to nitrous oxide (N₂O) emissions in China is therefore important. We performed a meta-analysis that included 153 field measurements of N₂O emissions from 21 field studies in China. Our goal was to quantify N₂O emissions and fertilizer nitrogen (N) based-emission factors (EFs) in Chinese vegetable systems and to clarify the effects of rates and types of N fertilizer in both open-field and greenhouse systems. The results indicated that the intensive vegetable systems in China had an average N₂O emission of 3.91 kg N₂O-N ha⁻¹ and an EF of 0.69%. Although the EF was lower than the IPCC default value of 1.0%, the average N₂O emission was generally greater than in other cropping systems due to greater input of N fertilizers. The EFs were similar in greenhouse vs. open-field systems but N₂O emissions were about 1.4 times greater in greenhouses. The EFs were not affected by N rate, but N₂O emissions for both open-field and greenhouse systems increased with N rate. The total and fertilizer-induced N₂O emissions, as well as EFs, were unaffected by the type of fertilizers in greenhouse system under same N rates. In addition to providing basic information about N₂O emissions from Chinese vegetable systems, the results suggest that N₂O emissions could be reduced without reducing yields by treating vegetable systems in China with a combination of synthetic N fertilizer and manure at optimized economic rates.

The diffusive gradients in thin films (DGT) technique is recognized to have advantages over traditional techniques. For example, the passive measurement generally follows the principle of metal uptake by plants, and its result incorporates the influences of soil properties, which may make DGT a good protocol for improving soil quality guidelines (SQGs). However, DGT has rarely been applied to assess Cd phytoavailability in soils under greenhouse vegetable production (GVP) systems. In this study, 29 turnips (Raphanussativus L.), 21 eggplants (Solanum melongena L.) and their corresponding soils were collected from GVP systems in Dongtai and Shouguang, eastern China. Simple linear regression and stepwise regression were performed using the soil Cd content and soil properties to predict the vegetable Cd content. Soil thresholds were derived based on both total and available Cd concentrations. The results showed that total Cd, DGT-measured Cd (DGT-Cd), soil-solution Cd (Soln-Cd) and CaCl2-extractable Cd (CaCl2-Cd) were all significantly correlated with vegetable Cd. DGT-Cd had the best correlation with turnip Cd. The total Cd threshold values ranged from 4.87 (pH 6.5) to 5.18 (pH 7.5) mg kg−1 for turnips and 14.60 (pH 6.5) to 14.90 (pH 7.5) mg kg−1 for eggplants. These Cd thresholds were higher than the current SQGs. The predicted of turnip Cd by DGT-Cd was not improved significantly by further considering the soil properties. The calculated soil threshold of DGT-Cd was 5.35 μg L−1 for turnips. However, the predicted soil threshold of DGT-Cd for eggplant was improved by including SOM, with R2 values from 0.53 to 0.70. The DGT-Cd threshold was calculated as 1.81 μg L−1 for eggplant (30.0 g kg−1 SOM). In conclusion, whether DGT measurements are independent of soil properties and preferable for the evaluation of Cd phytoavailability and the generation of soil thresholds remains to be clarified in future research.

The change in land-use from woodland to crop production leads to increased nitrous oxide (N2O) emissions. An understanding of the main N2O sources in soils under a particular land can be a useful tool in developing mitigation strategies. To better understand the effect of land-use on N2O emissions, soils were collected from 5 different land-uses in southeast China: shrub land (SB), eucalyptus plantation (ET), sweet potato farmland (SP), citrus orchard (CO) and vegetable growing farmland (VE). A stable isotope experiment was conducted incubating soils from the different land use types at 60% water holding capacity (WHC), using 15NH4NO3 and NH415NO3 to determine the dominant N2O production pathway for the different land-uses. The average N2O emission rates for VE, CO and SP were 5.30, 4.23 and 3.36 μg N kg−1 dry soil d−1, greater than for SB and ET at 0.98 and 1.10 μg N kg−1 dry soil d−1, respectively. N2O production was dominated by heterotrophic nitrification for SB and ET, accounting for 51 and 50% of N2O emissions, respectively. However, heterotrophic nitrification was negligible (<8%) in SP, CO and VE, where autotrophic nitrification was a primary driver of N2O production, accounting for 44, 45 and 66% for SP, CO and VE, respectively. Denitrification was also an important pathway of N2O production across all land-uses, accounting for 35, 35, 49, 52 and 32% for SB, ET, SP, CO and VE respectively. Average N2O emission rates via autotrophic nitrification, denitrification and heterotrophic nitrification increased significantly with gross nitrification rates, NO3− contents and C:N ratios respectively, indicating that these were important factors in the N2O production pathways for these soils. These results contribute to our understanding and ability to predict N2O emissions from different land-uses in subtropical acidic soils and in developing potential mitigation strategies.

Vegetable production in China is associated with high inputs of nitrogen, posing a risk of losses to the environment. Organic matter mineralisation is a considerable source of nitrogen (N) which is hard to quantify. In a two-year greenhouse cucumber experiment with different N treatments in North China, non-observed pathways of the N cycle were estimated using the EU-Rotate_N simulation model. EU-Rotate_N was calibrated against crop dry matter and soil moisture data to predict crop N uptake, soil mineral N contents, N mineralisation and N loss. Crop N uptake (Modelling Efficiencies (ME) between 0.80 and 0.92) and soil mineral N contents in different soil layers (ME between 0.24 and 0.74) were satisfactorily simulated by the model for all N treatments except for the traditional N management. The model predicted high N mineralisation rates and N leaching losses, suggesting that previously published estimates of N leaching for these production systems strongly underestimated the mineralisation of N from organic matter.

Herbicide application is a practice commonly used in agricultural systems because it is an efficient method of weed control. An inherent characteristic of some herbicides used in mountain agriculture, such as oxyfluorfen, is high adsorption to soil organic matter (SOM). Thus, intensive management that changes the quantity and quality of SOM, such as soil tillage and the massive application of organic fertilizers such as poultry litter, may favor soil contamination by this herbicide and alter its dynamics in the environment. Therefore, this study aimed to characterize the structures of humic substances (HSs) in the soil of forest areas and areas with intensive production of vegetables, relating them to the accumulation of the herbicide oxyfluorfen in tropical mountain agroecosystems. Organic carbon content was quantified in HSs, humic acid (HAs) were structurally characterized by CP/MAS ¹³C-NMR spectroscopy, and the oxyfluorfen molecule was detected and quantified using the QuEChERS residue detection method with subsequent analysis by LC-MS/MS. Oxyfluorfen was not detected in the forest areas, but it was detected in the vegetable growing areas at points with the lowest slope and high contents of organic matter and clay, with values of up to 0.13 mg kg⁻¹. The intensification in the SOM mineralization process, promoted by the intensive management adopted in the vegetable growing areas, resulted in a 16.46% reduction in COT, a 58.84% reduction in the carbon content in the form of SH and a reduction in the structures that give recalcitrance to the HA molecule (CAₗₖyₗ-H,R, CCOO–H,R, CAᵣₒₘₐₜᵢc-H,R, and CAᵣₒₘₐₜᵢc-O) when compared to those values in the forest area, presenting HAs with more aliphatic and labile properties. Thus, due to the structural characteristics of the HAs in the vegetable production areas, the herbicide oxyfluorfen showed a close relationship with the more aliphatic oxygenated structures, namely, CAₗₖyₗ-O,N, CAₗₖyₗ-O and CAₗₖyₗ-di-O.

Vegetable production in greenhouses is often associated with the use of excessive amounts of nitrogen (N) fertilizers, low NUE (15–35%), and high N losses along gaseous and hydrological pathways. In this meta-analysis, we assess the effects of application rate, fertilizer type, irrigation, and soil properties on soil N₂O emissions and nitrogen leaching from greenhouse vegetable systems on the basis of 75 studies. Mean ± standard error (SE) N₂O emissions from unfertilized control plots (N₂Ocₒₙₜᵣₒₗ) and N leaching (NLcₒₙₜᵣₒₗ) of greenhouse vegetable systems were 3.2 ± 0.4 and 91 ± 20 kg N ha⁻¹ yr⁻¹, respectively, indicating legacy effects due to fertilization in preceding crop seasons. Soil organic carbon concentrations (SOC) and irrigation were significantly positively correlated with NLcₒₙₜᵣₒₗ losses, while other soil properties did not significantly affect N₂Ocₒₙₜᵣₒₗ or NLcₒₙₜᵣₒₗ. The annual mean soil N₂O emission from fertilized greenhouse vegetable systems was 12.0 ± 1.0 kg N₂O–N ha⁻¹ yr⁻¹ (global: 0.067 Tg N₂O–N yr⁻¹), with N₂O emissions increasing exponentially with fertilization. The mean EFN₂O was 0.85%. The mean annual nitrogen leaching (NL) was 297 ± 22 kg N ha⁻¹ yr⁻¹ (global: 1.66 Tg N yr⁻¹), with fertilization, irrigation, and SOC explaining 65% of the observed variation. The mean leaching factor across all fertilizer types was 11.9%, but 18.7% for chemical fertilizer. Crop NUE was highest, while N₂O emissions and N leaching were lowest, at fertilizer rates <500 kg N ha⁻¹ year⁻¹. Yield-scaled N₂O emissions (0.05 ± 0.01 kg N₂O–N Mg⁻¹ yr⁻¹) and nitrogen leaching (0.79 ± 0.08 kg N Mg⁻¹ yr⁻¹) were lowest at fertilizer rates <1000 kg N ha⁻¹ yr⁻¹. Vegetables are increasingly produced in greenhouses, often under management schemes of extreme fertilization (>1500 kg N ha⁻¹ yr⁻¹) and irrigation (>1200 mm yr⁻¹). Our study indicates that high environmental N₂O and N leaching losses can be mitigated by reducing fertilization rates to 500–1000 kg N ha⁻¹ yr⁻¹ (mean: ∼762 kg N ha⁻¹ yr⁻¹) without jeopardizing yields.

Heavy metal contamination in protected-field vegetable production has aroused widespread concern and manure is considered to be one of the contamination sources. Little is known about its long-term effects on heavy metal pollution in uncontaminated soils. A 15-year protected-field vegetable production experiment was carried out with three manure treatments (chicken manure: cattle manure = 3:1) with high (HMAR), medium (MMAR) and low (LMAR) application rates to evaluate the long-term risks of heavy metal pollution. It was found that continuous and high manure application rates significantly increased the total concentrations of soil Cd, Zn, Cr, and Cu rather than Pb, Ni or As. The high application rate of manure also increased soil available heavy metals although the soil organic matter was increased as well. Though total soil Cd under the HMAR exceeded the threshold of national soil standard, Cd content in tomato and fennel still complied with the food safety requirements of vegetables. Generally, the accumulation rates of soil Zn, Cu, and Cr with 1 t⋅ha⁻¹ of manure application in three treatments were ranked by HMAR < MMAR < LMAR. Based on the results of the ratio of heavy metal accumulation risk (RAR), Zn, Cu, and Cr under HMAR and Cd and Zn under MMAR would exceed their soil threshold values within 100 years and RAR could be a useful indicator for monitoring the long-term risk of soil heavy metal pollution. Recommended manure application rates to guarantee a 100-year period of clean production were 44, 74, and 63 t⋅ha⁻¹⋅yr⁻¹ for Zn, Cu, and Cr, respectively. Measurements should be taken to minimize the risk of heavy metals (Cd, Zn, Cr, and Cu) pollution sourced from manure to ensure food safety and ‘cleaner’ protected-field vegetable production.

In an attempt to evaluate the behavior of Dechlorane plus (DP) in soil-vegetable systems, this work investigated the uptake and translocation of DP by vegetables and the dissipation of DP in soil under greenhouse and conventional conditions. To address human dietary exposure to DP, estimated dietary intake via vegetable consumption was calculated. The uptake potential indexes of DP from soil into root for tomato and cucumber cultivated under different conditions ranged from 0.089 to 0.71. The ranges of uptake potential indexes of DP from resuspended soil particles into stem, leaf and fruit were 0.68–0.78, 0.27–0.42 and 0.39–0.75, respectively. The uptake potential indexes in greenhouse vegetables were generally higher than those in conventional vegetables when the vegetables had been planted in contaminated soil, indicating that greenhouse enhanced the uptake of DP with a high soil concentration by vegetables. The translocation factor (TF) values of DP in vegetables were in the range of 0.022–0.17, indicating that DP can be transported from root to fruit even though it has a high octanol water partition coefficient (KOW). The half-lives of DP dissipation in soil ranged from 70 to 102 days. The dissipation of DP in greenhouse soil was slightly slower than that in conventional soil. Higher estimated dietary intake (EDI) values of DP via greenhouse vegetables were observed due to the higher concentration of DP in greenhouse vegetables than conventional vegetables. These results suggested that greenhouses should not be adopted for vegetable production in contaminated regions.