Science of the Total Environment 616–617 (2018) 614–621

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Dynamics of multiple elements in fast decomposing vegetable residues

Chun Cao a, Si-Qi Liu b, Zhen-Bang Ma c, Yun Lin d, Qiong Su e, Huan Chen f,⁎, Jun-Jian Wang g,⁎⁎a College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, Chinab Key Laboratory for Heavy Metal Pollution Control and Reutilization, School of Environment and Energy, Peking University Shenzhen Graduate School, Shenzhen 518055, Chinac College of Earth and Environmental Sciences, Lanzhou University, Lanzhou 730000, Gansu, Chinad Department of Atmospheric Sciences, Texas A&M University, College Station, TX 77843, United Statese Water Management & Hydrological Science, Texas A&M University, College Station, TX 77843, United Statesf Belle W. Baruch Institute of Coastal Ecology and Forest Science, Clemson University, SC 29442, United Statesg School of Environmental Science and Engineering, Guangdong Key Laboratory of Soil and Groundwater Pollution Control, Southern University of Science and Technology of China, Shenzhen,Guangdong 518055, China

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Litter decomposed faster in vegetablefarmland than previously studied eco-systems.

• Roots decomposed faster than leaves forstudied vegetables.

• As, Cu, Fe, Hg, Mn, and Pb possibly accu-mulated in the litters after 180 d.

⁎ Correspondence to: H. Chen, PO Box 596, BaruchGeorgetown, SC 29442, United States.⁎⁎ Correspondence to: J.-J. Wang, 1088 Xueyuan Road, N518055, China.

E-mail addresses: huanc@clemson.edu (H. Chen), wan

https://doi.org/10.1016/j.scitotenv.2017.10.2870048-9697/© 2017 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 13 September 2017Received in revised form 27 October 2017Accepted 27 October 2017Available online 31 October 2017

Editor: Jay Gan

Litter decomposition regulates the cycling of nutrients and toxicants but is poorly studied in farmlands. To under-stand the unavoidable in-situ decomposition process, we quantified the dynamics of C, H, N, As, Ca, Cd, Cr, Cu, Fe,Hg, K, Mg, Mn, Na, Ni, Pb, and Zn during a 180-d decomposition study in leafy lettuce (Lactuca sativa var.longifoliaf) and rape (Brassica chinensis) residues in a wastewater-irrigated farmland in northwestern China. Dif-ferent from most studied natural ecosystems, the managed vegetable farmland had a much faster litter decom-position rate (half-life of 18–60 d), and interestingly, faster decomposition of roots relative to leaves for boththe vegetables. Faster root decomposition can be explained by the initial biochemical composition (more O-alkyl C and less alkyl and aromatic C) but not the C/N stoichiometry. Multi-element dynamics varied greatly,with C, H, N, K, and Na being highly released (remaining proportion b 20%), Ca, Cd, Cr, Mg, Ni, and Zn released,and As, Cu, Fe, Hg, Mn, and Pb possibly accumulated. Although vegetable residues serve as temporary sinks ofsome metal(loid)s, their fast decomposition, particularly for the O-alkyl-C-rich leafy-lettuce roots, suggest thattoxic metal(loid)s can be released from residues, which therefore become secondary pollution sources.

© 2017 Elsevier B.V. All rights reserved.

Keywords:Litter decompositionNutrient releaseHeavy metal pollutionAccumulationRootsShootsLitters

Institute, Clemson University,

anshan, Shenzhen, Guangdong

gjj@sustc.edu.cn (J.-J. Wang).

1. Introduction

Litter decomposition is an important environmental process in reg-ulating carbon (C) sequestration, nutrient release, humus formation,

615C. Cao et al. / Science of the Total Environment 616–617 (2018) 614–621

and pollutant cycling in ecosystems (Tsui et al. 2008; Berg andMcClaugherty 2013; Cotrufo et al. 2015). The decomposition rate re-flects the dynamics of element turnover and is believed to be deter-mined by multiple factors, including litter quality (e.g., C/Nstoichiometry and molecular-level biochemical composition) and envi-ronmental conditions (Manzoni et al. 2008; X. Wang et al. 2012;Keiluweit et al. 2015). Studies about litter decomposition have mainlyfocused on natural ecosystems including forests (Paudel et al. 2015;Wang et al. 2016; Yeong et al. 2016), grasslands (Suseela et al. 2014),shrublands (Fioretto et al. 2005), and wetlands (Mackintosh et al.2016; Xie et al. 2017). Although the nutrient and toxicant dynamics inmanaged agroecosystems are critical to sustainable agriculture, foodquality, and human health (Dungait et al. 2012), few studies(Kuzyakov et al. 1997; Chaves et al. 2004) have investigated litter de-composition in the agroecosystems, such as in vegetable farmlands.

It is generally assumed that a large proportion of food crops harvest-ed fromvegetable farmlands leave fewplant residues for in-situfield de-composition and nutrient cycling. However, a recent study showed thataround 1.5 billion tons of vegetables and fruits are produced every yearin the world but as much as 45% of them are discarded as wastes(Mazareli et al. 2016). Vegetable waste is generated at all stages of thesupply chain, including harvest, retailing, and consumption. In China,about 100 million tons of vegetable waste or residues (including theirleaves and roots) are produced per year (Sun et al. 2005), approximate-ly 15–20% of which remains in the vegetable growing farmlands in thestage of vegetable harvest (Gong et al. 2012). A survey also showedthat in some areas like Beijing City, ~70% of the 3.33million tons of veg-etable waste generated in 2007 was not reutilized but directlydecomposed (Gong et al. 2012). Although vegetable waste digestionshave frequently been studied in anoxic reactors (Provenzano et al.2016; Wu et al. 2016), little is known about the in-situ decompositionof vegetable waste or its residues under natural conditions.

Here, we designed a 180-d field decomposition experiment in aknown metal-polluted farmland in northwestern China and quantifiedthe residue mass and multi-elements, including C, H, N, As, Ca, Cd, Cr,Cu, Fe, Hg, K, Mg, Mn, Na, Ni, Pb, and Zn. Among these elements, N, Ca,Cu, Fe, K, Mg, Mn, Ni, and Zn are considered as essential elements forplant growth and reproduction (Barker and Pilbeam 2016). As, Cd, Cr,Cu, Fe, Hg, Mn, Ni, Pb, and Zn belong to the category of heavy metal(-loid)s, and As, Cd, Cr, Hg, Mn, Ni, and Pb are commonly considered aspotentially toxic metal(loid)s. We hypothesized that: 1) vegetable resi-dues decomposed faster than plant litters decomposing in many otherecosystems (e.g., forest and wetland) because vegetable residues havevery labile biomass; 2) the decomposition rates of four vegetable resi-dues (including leaves and roots), although highly variable, can be de-termined by the analysis of C/N stoichiometry and biochemicalcomposition of the initial vegetable residues; and 3)metal(loid)s, espe-cially toxic metal(loid)s, are continuously released from thedecomposing vegetable residues, and the changes in their concentra-tions vary greatly among vegetable residues.

2. Materials and methods

2.1. Study area

Baiyin City, or “Tong Cheng” (which means “Copper City” in Chi-nese), is located in the Gansu Province in northwestern China. Thisarea has an annual average temperature of 6–9 °C, an annual averageprecipitation of 180–450 mm, and an annual average evapotranspira-tion of 2101 mm (Cao et al. 2016). It has been an important non-ferrous metal mining and smelting base in China since the 1950s (Nanand Zhao 2000; Li et al. 2006). The suburban farmlands in this cityhave been subjected to wastewater irrigation for decades, and severemetal pollution has been discovered. Therefore, Baiyin City was listedas the first among 30 priority areas that the ChineseMinistry of Financeand Chinese Ministry of Environmental Protection funded for soil metal

remediation in 2015 (http://jjs.mof.gov.cn/zhengwuxinxi/tongzhigonggao/201506/t20150602_1248397.html). The city is dividedinto Xidagou Basin (or “West BigDitch”, 428 km2) andDongdagou Basin(or “East Big Ditch”, 368 km2) (Nan and Zhao 2000). Farmlands inXidagou Basin mainly receive treated municipal (or domestic) waste-water, and those in Dongdagou Basin mainly receive treated industrialwastewater (Nan and Zhao 2000; Cao et al. 2016). Soil properties andconcentrations of heavy metals in these farmlands are shown inTable S1. These large farmlands provide vegetables for local communi-ties and nearby cities (Cao et al. 2016). The cultivated vegetablesmainlyinclude leafy lettuce (Lactuca sativa var. longifoliaf L.), rape (Brassicachinensis L.), carrot (Daucus carota L.), Chinese lettuce (Lactuca sativaL.), Chinese cabbage (Brassica pekinensis L.), tomato (Lycopersiconesculentum L.), zucchini (Cucurbita pepo L.), and eggplant (Solanummelongena L.). High adverse health risks from diets based on vegetablesgrowing in these farmlands have been reported (Cao et al. 2016). Tworepresentative vegetable species, leafy lettuce and rape, growing in agreenhouse in the Dongdagou Basin (36.48°N, 104.31°E; altitude1746m)with the highest degree of pollution (Cao et al. 2016), were se-lected for our litter decomposition experiment.

2.2. Litter bag experiments

Leafy lettuce and rape were planted in the spring of 2014 and har-vested in the summer of 2014. After harvesting, the leafy lettuce andrape samples were gently washed with deionized water, divided intoleaf and root parts, and then dried at 60 °C for 24 h to determine theirmass. In July 2014, litter bags of dimension 20 × 20 cm, with2 mm mesh size, were filled with a 30-g sample of washed and driedvegetable leaf and then placed on the soil surface for in-situfield decom-position. On the same day, litter bags of dimension 10 × 10 cm, with1 mm mesh size, were filled with a 15-g sample of washed and driedvegetable root and then buried in the soil at a depth of 3–5 cm. After60, 120, and 180 d, six bags of each tissue type (leaf or root) of each veg-etable species (leafy lettuce or rape) were collected from the field. Thecollected samples were transported to the lab, gently washed with de-ionized water, dried, and weighed.

2.3. Chemical analyses

The freshly collected (non-decomposed) and decomposed vegetablesamples were all dried, ground, and passed through a 2-mm sieve. Allsamples were analyzed for C, H, and N contents by an elemental analyz-er (Vario EL III, Elementar, Germany) (J.J. Wang et al. 2012). Each one ofthe samples (0.5 g) was mixed with 20 mL concentrated HNO3 in a300 mL Teflon beaker and then evaporated at 240 °C to obtain a finalvolume of 5 mL. The Teflon beaker was again heated at 240 °C for1.5 h after addition of 10 mL concentrated HNO3, 1 mL concentratedHF, and 3 mL concentrated HClO4. After cooling to room temperatures,the digested solution was transferred from the Teflon beaker to a volu-metric flask, which was then filled with Milli-Q water to make a finalvolume of 100 mL. Concentrations of As, Ca, Cd, Cr, Cu, Fe, Hg, K, Mg,Mn, Na, Ni, Pb, and Zn in the digested samples were analyzed by induc-tively coupled plasma atomic emission spectroscopy (using the ICP-AES,IRIS Intrepid XSP equipment, Thermo Fisher Scientific,MA, USA) and in-ductively coupled plasma mass spectrometry (using the ICP-MS, ×7equipment, Thermo Fisher Scientific, MA, USA) following the Chinesestandard methods of JY/T 015-1996 (State Education Commission ofChina 1996) and GB/T 6041-2002 (State Standard of China 2002). Spe-cifically, ICP-MS analyzed the concentrations of As, Cd, Cr, Hg, and Pb,and the concentrations of the other elements were detected by ICP-AES. This detection method showed the recovery of 86.8–116.9% forall the metal(loid)s and standard deviations were lower than 20% inthe repeat tests on standards and procedure blanks.

The organic carbon compositions of the non-decomposed vegetableswere characterized by 13C cross-polarization/total sideband

Fig. 1.Variations in the decomposition half-lives of vegetable residuemass, C, H, andN andthe initial litter quality of vegetable residues. p-values over the bars show the significanceof two-tailed Pearson's correlation between the parameter of litter quality and decompo-sition half-life of vegetable residue mass. p-values b 0.05 are in bold.

616 C. Cao et al. / Science of the Total Environment 616–617 (2018) 614–621

suppression (CP/TOSS) using a Bruker AVANCE 400 nuclear magneticresonance (NMR) spectrometer (Bruker Analytik, Rheinstetten,Germany) and 4-mmdiameter sample rotors as described in a previousstudy (Li et al. 2017). The organic carbon composition was divided intofour categories: alkyl C (0–46 ppm; from cutin, suberin, lignin, lipids,and amino acid side chains), O-alkyl C (46–114 ppm; from carbohy-drates, peptides, and methoxyl C in lignin), aromatic C (114–164 ppm;from lignin and amino acids found in peptides), and carboxyl C (164–220 ppm; from fatty acids and peptides) (Baldock et al. 1992; Li et al.2017). The alkyl/O-alkyl ratio serves as a proxy for the degradability ofthe vegetable residuals because O-alkyl C, mainly from carbohydrates,commonly degrades faster than alkyl C (Baldock et al. 1992).

2.4. Data analysis

RStudio Desktop version 1.0.44 (Boston, MA, USA) was used for datacalculation, statistical analysis, and figure plotting. The residuemass andtotal elemental mass in the remaining litters were calculated followingthe formula in Eq. (1). Element loss after the 180-d decomposition peri-odwas estimated following the formula in Eq. (2). The proportion (%) ofmass remaining relative to the initialmass, henceforth referred to as thepercentage remaining mass (also known as the accumulation index(Sun et al. 2012; Sun et al. 2016)), was used to evaluate the release (per-centage remaining b100%) or accumulation (percentage remainingN100%) of residue mass and elements during litter decompositions(Eq. (3)). The percentage remaining values of mass, C, H, and N in thevegetable residues along with the decomposition time were fittedwith the exponential model in Eq. (4) (Ranjbar and Jalali 2012; Li andYe 2014). Also, the decomposition half-lives during the 180-d litter de-composition period were calculated following the formula in Eq. (5).

W j;t ¼ Mt ∙C j;t ð1Þ

ELj;180 ¼ W j;180−W j;0

W j;0ð2Þ

PRj;t ¼ W j;t=W j;0 � 100% ð3Þ

kj ¼ −ln PRj;t=aj

� �

tð4Þ

t0:50; j ¼ lnaj=50%

kjð5Þ

whereWj,t is the drymass in g of j at time t;Mt is the drymass in g of theremaining litter at time t; Cj,t is the concentration in g g−1 of j in the re-maining litter at time t; ELj, 180 is the element loss of j, expressed ing g−1, after 180-d decomposition; PRj,t is the percentage of j remainingat time t; aj is the correction factor of j and the intercept for the best-fitting equation (Sun et al. 2012); kj is the decomposition constant(also considered as the decomposition rate) of j, expressed in d−1; j, ismass or elements; and t0.50, j is the time required in d for decomposinghalf of the initial drymass of j. Decomposition half-life was used to indi-cate the decomposition rate, because inclusion of the correction factor ajin Eq. (4) for better fitting led to a biased decomposition constant com-pared with when fitting was done without the correction factor.

Multiple factor analysis of variance (ANOVA) was used to evaluatewhether the effects of decomposition time (i.e., 0, 60, 120, and 180 d),vegetable parts (i.e., leaves and roots), vegetable species (i.e., leafy let-tuce and rape), and the interaction of vegetable parts and vegetable spe-cies on residuemass and concentration of each element during the 180-d decomposition period were statistically significant (p b 0.05). ANOVAwas carried out using the function of aov in the stats package. Correla-tion matrix was used to study the correlations among residue massand the concentrations of elements during decomposition, as well asthe correlations among the decomposition half-life, initial N content,

initial C/N ratio, initial abundances of alkyl, O-alkyl, aromatic, and car-boxyl C, and alkyl/O-alkyl ratio of the four vegetable residues. Itwas cal-culated and visualized using the stats and corrplot packages. Principalcomponent analysis (PCA) was employed to investigate the changes inthe concentrations of the studied metal(loid)s during decomposition.It was performed using the function prcomp in the stats package afterthe variables were centered and scaled, and then the function ggplotin the package ggplot2 was used to plot the first principal componentas the x axis and the second principal component as the y axis.

3. Results

3.1. Dynamics of residue mass

During the 180-d period of in-situ field decomposition, the mass ofall vegetable residues continuously decreased with decompositiontime (Figs. 1 and 2; Table S2). The decomposition half-lives of massshowed a faster decomposition of roots (half-lives of 18 and 32 d forleafy lettuce and rape, respectively) than leaves (half-lives of 60 and56 d for leafy lettuce and rape, respectively). The decomposition half-life did not correlate with the initial N content or C/N ratio for the fourtypes of vegetable residues (p N 0.1; Fig. 1b). However, it had significantpositive correlations with the abundances of O-alkyl and carboxyl C inthe 13C NMR spectra (Fig. S1), and significant negative correlationswith the abundances of alkyl and aromatic C, as well as with the alkyl/O-alkyl ratio (all p b 0.05; Figs. 1b and S1). After 180 d, decomposedleafy-lettuce leaf, leafy-lettuce root, rape leaf, and rape root accountedfor only 16.1 ± 7.1%, 1.0 ± 0.2%, 14.5 ± 1.5%, and 11.8 ± 9.7% of theirrespective initial masses (mean ± SD; n = 6; Table S3).

Fig. 2. Changes in the percent remaining (%) of residuemass and the studied elements in the remaining litters of leafy-lettuce leaf, leafy-lettuce root, rape leaf, and rape root during 180-din-situ field decompositions. 1 greatly released (all b 20%); 2 released (all b 100%); and 3 possibly accumulated (N100%). Error bars show standard derivation of six replicates.

617C. Cao et al. / Science of the Total Environment 616–617 (2018) 614–621

3.2. Dynamics of C, H, and N

Both the remaining percentages and concentrations of C, H, and N inall the vegetable residues decreased with increasing decompositiontime (Figs. 1 and 2; Fig. S2; Table S2). Consistent with the half-life of lit-termass, the half-lives of C, H, andNwere 18–53 d andmuch shorter forroots than leaves during the 180-d decomposition period (Figs. 1 and 2;Table S2). These half-lives also had significant correlations with theinitial biochemical components (i.e., alkyl, O-alkyl, aromatic, andcarboxyl C; all p b 0.05) but not with the initial N content or C/N ratio(all p N 0.05), across all the four types of vegetable residues (Table S4).For all the 180-d decomposed vegetable residues, the percentageremaining values of C, H, and N were 0.6–9.3%, 0.3–8.0%, and 0.4–7.1%,respectively (Fig. 2; Table S5). Specifically, whereas the percentageremaining values of C, H, and N were comparable among leafy-lettuceleaf, rape root, and rape leaf, those for leafy lettuce root were aboutone order of magnitude lower (all percentage remaining values b 1%;Fig. 2; Table S5). The C/N ratio was 7.3–11.0 in the non-decomposedvegetable residues, and it increased to 11.3–18.1 in the 180-ddecomposed vegetable residues, which is consistent with the shorterhalf-life (faster decomposition) of N relative to that of C (Fig. 1).

3.3. Dynamics of metal(loid)s

The percentage remaining values and concentrations of the alkalimetals K andNadecreased continuouslywith increasing decompositiontime and were greatly released after the 180-d decomposition period(Fig. 2), similar to that of C, H, and N. However, for other metals, theirremaining percentages and concentrations in litter highly fluctuatedwith increasing decomposition time (Figs. 2 and S3). The initial concen-trations of metal(loid)s were higher in leaf than in root (except for Asand Na) for leafy lettuce, but higher in root than in leaf (except for Cd,Cr, Cu, Fe, and Zn) for rape (Table S3). After the 180-d decompositionperiod, concentrations of metal(loid)s in all the vegetable residues in-creased, except for K, Na Ca, Cr, and Hg in leafy-lettuce leaf, K and Crin leafy-lettuce root, and K and Na in both rape leaf and root (Table S3).

Contrary to our hypothesis that the metal(loid)s would be continu-ously released, the metal(loid)s were found to be accumulated in the180-d decomposed vegetable residues (Fig. 2; Table S5). Elementswere operationally divided into three categories: greatly released (allpercentage remaining values b 20%), released (all percentage remainingvalues b 100%), and possibly accumulated (at least one of the percent-age remaining value N 100%). Based on this classification, K and Nawere considered to be greatly released, Ca, Cd, Cr, Mg, Ni, and Zn wereconsidered to be released, and As, Cu, Fe, Hg, Mn, and Pb were consid-ered to be possibly accumulated (Fig. 2; Table S5). The release and accu-mulation trends also highly depended on the vegetable parts and thespecies (Table S6). For example, the percentage remaining value waslower than 10% for Cr, Hg, K, and Na in leafy-lettuce leaf, for As, Ca, Cd,Cr, Hg, K, Na, and Ni in leafy-lettuce root, but only for K in rape leafand for K and Na in rape root. Almost all metal(loid) species werecompletely released from the 180-d decomposed leafy-lettuce root,due to its low percentage remaining mass of 1.0 ± 0.2% (Fig. 2;Table S5). On the other hand, the percentage remaining value washigher than 100% for Fe, Mn, and Pb in leafy-lettuce leaf, for As, Fe, Hg,Mn, and Pb in rape leaf, as well as for Cu, Fe, Mn, and Pb in rape root(Fig. 2; Table S5).

3.4. Relations among multiple elements

The correlationmatrix among residuemass and concentrations of allthe studied elements during the decomposition period varied across thefour different types of vegetable residues (Fig. 3). In general, residuemass was significantly and positively correlated with the contents ofC, H, and N in all the vegetable residues except rape root, showingthat the decomposition of residue mass was mainly contributed to bythe decomposition of plant organic matter. The concentration K wasalso consistently correlated with residue mass. Conversely, most othermetal(loid) concentrations had significant negative correlation withresidue mass in all the vegetable residues (except rape root). Besides,significant positive correlations among the concentrations of metal(-loid)s commonly occurred in all the vegetable residues. In addition,

Fig. 3. Correlation matrix among residue mass and concentrations of all studied elements during 180-d in-situ field decomposition in the vegetable residues of (a) leafy-lettuce leaf,(b) lettuce root, (c) rape leaf, and (d) rape root. Correlation coefficients either on the principal diagonal or with p-value greater than or equal to 0.05 are shown as blank.

618 C. Cao et al. / Science of the Total Environment 616–617 (2018) 614–621

Ca, Fe, K, Mg, and Na concentrations showed different temporal varia-tions with other metal(loid)s (Fig. S4).

4. Discussion

4.1. Factors influencing fast decomposition

Consistent with our first hypothesis, the vegetable farmland investi-gated here showed amuch faster litter decomposition rate (half-lives of18–60 d) compared to those reported from awide range of natural eco-systems including forests, grasslands, shrublands, and wetlands (half-lives usually N60 d) (Suseela et al. 2014; Mackintosh et al. 2016; Raiet al. 2016; Sun et al. 2016) (Table S7). The faster degradation of vege-tables here relative to that of plant litter in other ecosystems is support-ed by the stoichiometric decomposition theories (Manzoni et al. 2008;Manzoni et al. 2010). Specifically, a lower initial C/N ratio andhigher ini-tial N content of litter usually reflects higher abundances of labile pro-teins, and is generally considered as a predictor of fasterdecomposition rate, particularly in the early stages of decomposition(Mooshammer et al. 2012; Freitas et al. 2014; Sun et al. 2016; Xieet al. 2017). The C/N ratio was lower (7.3–11.0) than those reported inthe studies in other ecosystems (usually N15; e.g., 24.0–39.5 for Jatrophacurcas L. (Negussie et al. 2015); 52.5±0.1, 53.7±0.1, and 67.5±0.2 forSalix atrocinerea, Quercus robur, and Pinus pinaster, respectively (Freitas

et al. 2014)). Also, the high initial N contents (~4%) observed here arenot common in other types of ecosystems (usually b3%; e.g., 1.4–2.5%in the Sonbhadra forest (Rai et al. 2016); b1% in the roots of Artemisiahalodendron (Luo et al. 2016)). The lower C/N ratios and high N concen-trations observed in our study are consistent with a shorter decomposi-tion half-life. More importantly, in addition to the possibility ofstoichiometric controls, high decomposition rates for vegetables are di-rectly supported by the fact that 50–71% of the vegetable C compositionis fromO-alkyl C (mainly fromeasily biodegradable carbohydrates), and8–10% is from aromatic C that mainly originates from recalcitrant ligninand aromatic peptides (Figs. 1 and S1). This is similar to a previous re-port that shows vegetable and fruit waste can contain ~75% sugarsand hemicellulose, 9% cellulose, and 5% lignin (Bouallagui et al. 2005).Such high carbohydrate and low lignin contents in plant residues arenot common in other ecosystems, which explains the faster litter de-composition rate observed in the vegetable farmland than those report-ed from other ecosystems.

Interestingly, the decomposition half-lives for residue mass, C, H,and N were lower for roots than leaves for both the vegetable species,following the order of leafy-lettuce leaf ≈ rape leaf N rape root N

leafy-lettuce root (Table 1). The vegetable roots decomposed almosttwice as fast as the leaves. This is in contrast with the results of manystudies in forest, grassland, or cropland ecosystems showing slower de-composition of roots than leaves (Abiven et al. 2005; Freschet et al.

Table 1Element loss for each kg of dried vegetable residue after 180-d in-situ field decomposition (mean ± standard deviation; n = 6).

Elements Units Leafy-lettuce leaf Leafy-lettuce root Rape leaf Rape root Root vs leaf

C g kg−1 331.2 ± 15.7 348.4 ± 0.5 270.2 ± 3.1 275.5 ± 24.1 N

H g kg−1 48.3 ± 2.1 50.6 ± 0.1 42.3 ± 0.6 41.4 ± 3.1N g kg−1 40.2 ± 1.7 31.8 ± 0.1 37.9 ± 0.4 34.2 ± 1.8K g kg−1 70.9 ± 0.5 71.5 ± 0.0 67.0 ± 0.3 68.5 ± 1.2 N

Na g kg−1 8.8 ± 0.4 10.5 ± 0.1 6.2 ± 1.0 9.4 ± 0.8 N

Ca g kg−1 9.3 ± 0.7 7.0 ± 0.0 3.2 ± 1.1 2.8 ± 3.6 b

Cd⁎ mg kg−1 0.4 ± 0.3 0.7 ± 0.0 1.0 ± 0.0 0.6 ± 0.3Cr⁎ mg kg−1 97.3 ± 0.3 93.4 ± 0.0 9.9 ± 27.0 51.8 ± 36.1Mg g kg−1 2.1 ± 0.7 2.7 ± 0.1 1.5 ± 0.3 2.4 ± 0.9 N

Ni⁎ mg kg−1 32.1 ± 8.9 30.5 ± 0.3 16.1 ± 7.4 37.2 ± 9.8Zn mg kg−1 22.3 ± 23.5 46.5 ± 1.6 22.6 ± 5.5 16.9 ± 37.2As⁎ mg kg−1 0.1 ± 0.6 1.1 ± 0.0 −0.1 ± 0.0 0.6 ± 0.7 N

Cu mg kg−1 1.6 ± 2.4 2.8 ± 0.5 1.0 ± 1.2 −1.7 ± 7.4Fe g kg−1 −2.0 ± 1.2 0.3 ± 0.1 −2.6 ± 0.6 −1.5 ± 1.6 N

Hg⁎ mg kg−1 0.5 ± 0.0 0.1 ± 0.0 −0.0 ± 0.0 0.4 ± 0.0Mn⁎ mg kg−1 −51.1 ± 42.4 20.2 ± 2.2 −51.6 ± 19.1 −44.3 ± 76.3 N

Pb⁎ mg kg−1 −4.3 ± 5.0 0.9 ± 0.3 −6.8 ± 1.4 −3.5 ± 6.5 N

N and b show significantly higher and lower loss from root residues compared to that from leaf residues, respectively.Negative values that indicate the accumulation of elements in the 180-d decomposed vegetable residues are in bold.⁎ indicates potentially toxic metal(loid).

619C. Cao et al. / Science of the Total Environment 616–617 (2018) 614–621

2013). The faster mass loss and C release of roots over leaves in ourstudy suggests that the vegetable root materials are not likely to be amore persistent C pool than leaves. However, the faster decompositionof roots than leaves and faster nitrogen release from roots than leaveshighlight that roots are more relevant for nitrogen release than leaves.Note that faster decomposition of roots than leaves could not be attrib-uted to stoichiometric controls because roots had even higher C/N ratiosthan leaves (Fig. 1) and therewas no correlation between the decompo-sition half-life and initial C/N ratio. Although the C/N ratio has been longused to indicate litter decomposability and to guide solid-wastecomposting, our findings support the inability of using only the C/Nratio to predict the decomposition rate (Wang et al. 2015). Here, weclearly showed that faster root decomposition is caused by very highO-alkyl C abundance (mainly from labile carbohydrates) and lower re-calcitrant alkyl and aromatic C abundances (mainly from plant lipidsand lignin, respectively) in the vegetable roots compared to those inthe leaves (Fig. 1). Furthermore, the pattern of faster root decomposi-tion compared to that of the leaves is more remarkable for leafy lettucethan for rape (Table 1), which is also in agreementwith the NMR resultsthat show a larger leaves-vs-root difference for leafy lettuce comparedto that for rape (Fig. 1). The inter-vegetable-species difference high-lights the weakness of the C/N ratio in predicting litter decomposabilityand future research on the cycling of multi-elements in agroecosystemsmust fully consider the critical role of biochemical compositions acrossplant parts and species.

4.2. Implications for metal nutrient releases

Most of the nutrient elements were gradually released from thedecomposing vegetable residues. These nutrient elements are consid-ered essential to sustain soil fertility and crop growth. For example, Kis critical for maintaining the osmotic potential and activating enzymesfor respiration and photosynthesis, Ca is critical for cell wall synthesis,Mg for enzyme activations for DNA and RNA synthesis, and Fe for acti-vating redox-associated enzymes (Taiz and Zeiger 2010). However,the inter relationship of element concentrations (Fig. 3), as well as thevarying percent remaining values of different elements (Fig. 2) clearlyshowed that the release of nutrients such as Ca, Fe, K, Mg, and Na didnot follow consistent temporal variation during decomposition. K andNa in the remaining litters of the four vegetable parts were releasedfastest among the metal(loid)s, and almost completely disappeared inthe 180-d decomposed vegetable residues. Similar results were previ-ously observed in some other ecosystems (Bragazza et al. 2007;Polechonska and Samecka-Cymerman 2015; Jing et al. 2016), which

were attributed to these elements' high leachability from thedecomposed litters, because they are highly soluble and loosely boundin plant tissues (Azeez et al. 2009). After the 180-d decomposition peri-od, the release of K and Na from the vegetable residues to the soil wereat the rates of 67.02–71.53 g kg−1mass and 6.19–10.50 g kg−1mass, re-spectively (Table 1). Similarly, but to lesser extents, Ca, Mg, Ni, and Znwere released from all the vegetable residues (Fig. 2), indicating thatin-situ field decomposition of vegetable residues can partially releasethese plant nutrients back to the soils and potentially improve soil fertil-ity. Although root residues decomposed faster than leaf residues andthus released more C, H, N, K, Na, andMg per kg residue, these root res-idues released less Ca than leaf residues (Table 1). This example demon-strates that a faster overall litter mass decomposition does notnecessarily suggest faster release of all elements.

Interestingly, despite the very fast decomposition of vegetable resi-dues, this process did not release all kinds of nutrients, and possiblyeven accumulated essential nutrients such as Cu, Fe, and Mn from theenvironment. A previous litter decomposition study in wetlands foundthat decomposing plant litter can act as a short-term sink for nutrientmetals, such as Fe and Zn from mine water (Batty and Younger 2007).Our study demonstrates that even in semiarid agroecosystem, wherethe vegetable residues decompose very fast, plant residues can also bea short-term sink of essential metallic nutrients. This finding appearsto contradict the belief that fast litter decomposition would rapidly in-crease the available nutrients in soils. One possible reason responsiblefor this phenomenon could be the strong affinity of Cu, Fe, and Mn forplant organic residue, and the low solubility of these elements at highpH in this site (Table S1) (McCauley et al. 2009). Because the temporalvariations in the percentage remaining values of these nutrients fluctu-ated greatly (Fig. 2), it is currently unpredictablewhether long-termde-composition (N180 d) would eventually release these nutrients in abioavailable form from the residues, or turn the residues into parts ofthe soil thereby stabilizing these metal nutrients.

4.3. Implications for metal(loid) pollution

Manymetal(loid)s elements were enriched in the fast decomposingresiduemass. Several previous studies have also shown increasedmetalconcentrations during litter decomposition (Du Laing et al. 2006; Denget al. 2016; Pourhassan et al. 2016; Sun et al. 2016). Increased metalconcentrations in decomposing litters may be attributed to metal en-richment during organic matter mineralization, contamination by sedi-ments or soils, passive sorption on the surface of recalcitrant organiccompounds, or accumulation of active microbial colonizers (Du Laing

620 C. Cao et al. / Science of the Total Environment 616–617 (2018) 614–621

et al. 2006; Sun et al. 2016). Because the increasedmetal concentrationsin decomposing litter seems common even for very fast decomposingvegetable residues, the resulting higher toxicity of these decomposingvegetable residues may cause increased health risks if these residuesare unintentionally consumed by domestic animals.

Although vegetable residues have very fast decomposition rates,we hypothesized that vegetable residues would be net sources ofmetal(loid)s, and the potentially toxic metals As, Hg, Mn, and Pbwere found to be accumulated in vegetable residues after the 180-ddecomposition period. Results showed that 44.3–51.1 mg kg−1 of Mnand 3.5–6.8 mg kg−1 of Pb were accumulated in all the vegetableresidues except in leafy-lettuce root (Table 1). Although the vegetableresidues can be used to improve soil fertility through in-situ fielddecomposition, the accumulation of toxic metal(loid)s can potentiallymake these vegetable residues more toxic to local animals after short-term decomposition. It seems beneficial that decomposing vegetableresidues accumulate some toxic metals from the environment andserve as temporary metal pools, thereby reducing the bioavailablemetal levels in the soil for the crop to uptake in a short term (Bakeret al. 1994). However, the very fast litter decomposition in vegetablefarmlands compared to that in the other ecosystem types suggeststhat the metals associated with plant organic matter would be releasedmore efficiently to soils, with plant organic matter serving as secondarypollution source of bioavailable metals in the vegetable farmlands thanin the other ecosystems.With the same drymass, the root residues thatdecompose faster than leaf residues could retain less or releasemore As,Mn, and Pb back to the soils (Table 1). In fact, root residues are not theedible parts of leafy vegetables and thus usually account for a largerfraction of vegetable wastes during harvesting in farmlands. Therefore,appropriate disposal of decomposing residues, particularly for the fastdecomposing residues such as O-alkyl-C-rich leafy-lettuce roots, aresuggested to minimize their ecological risks to the environments andconsequently to human health.

Acknowledgments

The project was supported by National Natural Science Foundationof China (41761074 and 41401204) and China Scholarship Council(CSC). Thanks to Dr. Alexander Ruecker from Belle W. Baruch Instituteof Coastal Ecology and Forest Science at Clemson University, SC, forreviewing this manuscript.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.scitotenv.2017.10.287.

References

Abiven, S., Recous, S., Reyes, V., Oliver, R., 2005. Mineralisation of C and N from root, stemand leaf residues in soil and role of their biochemical quality. Biol. Fertil. Soils 42,119–128.

Azeez, P.A., Prusty, B.A.K., Jagadeesh, E.P., 2009. Alkali and alkaline earth metals indecomposing macrophytes in a wetland system. Acta Ecol. Sin. 29, 13–19.

Baker, A.J.M., Mcgrath, S.P., Sidoli, C.M.D., Reeves, R.D., 1994. The possibility of in-situheavy-metal decontamination of polluted soils using crops of metal-accumulatingplants. Resour. Conserv. Recycl. 11, 41–49.

Baldock, J., Oades, J., Waters, A., Peng, X., Vassallo, A., Wilson, M., 1992. Aspects of thechemical structure of soil organic materials as revealed by solid-state 13C NMR spec-troscopy. Biogeochemistry 16, 1–42.

Barker, A.V., Pilbeam, D.J., 2016. Handbook of Plant Nutrition. CRC Press, Boca Raton, FL.Batty, L.C., Younger, P.L., 2007. The effect of pH on plant litter decomposition and metal

cycling in wetland mesocosms supplied with mine drainage. Chemosphere 66,158–164.

Berg, B., McClaugherty, C., 2013. Plant litter: decomposition, humus formation, carbon se-questration. 3rd ed. Springer Press, New York.

Bouallagui, H., Touhami, Y., Cheikh, R.B., Hamdi, M., 2005. Bioreactor performance in an-aerobic digestion of fruit and vegetable wastes. Process Biochem. 40, 989–995.

Bragazza, L., Siffi, C., Iacumin, P., Gerdol, R., 2007. Mass loss and nutrient release duringlitter decay in peatland: the role of microbial adaptability to litter chemistry. SoilBiol. Biochem. 39, 257–267.

Cao, C., Chen, X.P., Ma, Z.B., Jia, H.H., Wang, J.J., 2016. Greenhouse cultivationmitigatesmetal-ingestion-associated health risks from vegetables in wastewater-irrigatedagroecosystems. Sci. Total Environ. 560, 204–211.

Chaves, B., De Neve, S., Hofman, G., Boeckx, P., Van Cleemput, O., 2004. Nitrogen mineral-ization of vegetable root residues and green manures as related to their (bio)chemi-cal composition. Eur. J. Agron. 21, 161–170.

Cotrufo, M.F., Soong, J.L., Horton, A.J., Campbell, E.E., l Haddix, M., Wall, D.H., Parton, W.J.,2015. Formation of soil organic matter via biochemical and physical pathways of lit-ter mass loss. Nat. Geosci. 8, 776.

Deng, H.G., Zhang, J., Chen, S.Y., Yang, L.W., Wang, D.Q., Yu, S.Y., 2016. Metal release/accu-mulation during the decomposition of Potamogeton crispus in a shallow macrophyticlake. J. Environ. Sci. (China) 42, 71–78.

Du Laing, G., Van Ryckegem, G., Tack, F.M.G., Verloo, M.G., 2006. Metal accumulation in in-tertidal litter through decomposing leaf blades, sheaths and stems of Phragmitesaustralis. Chemosphere 63, 1815–1823.

Dungait, J.A.J., Cardenas, L.M., Blackwell, M.S.A., Wu, L., Withers, P.J.A., Chadwick, D.R., Bol,R., Murray, P.J., Macdonald, A.J., Whitmore, A.P., Goulding, K.W.T., 2012. Advances inthe understanding of nutrient dynamics and management in UK agriculture. Sci.Total Environ. 434, 39–50.

Fioretto, A., Di Nardo, C., Papa, S., Fuggi, A., 2005. Lignin and cellulose degradation and ni-trogen dynamics during decomposition of three leaf litter species in a Mediterraneanecosystem. Soil Biol. Biochem. 37, 1083–1091.

Freitas, A.C., Rodrigues, D., Rocha-Santos, T.A.P., Goncalves, F., Duarte, A.C., Pereira, R.,2014. The impact of uranium mine contamination of soils on plant litter decomposi-tion. Arch. Environ. Contam. Toxicol. 67, 601–616.

Freschet, G.T., Cornwell, W.K., Wardle, D.A., Elumeeva, T.G., Liu, W.D., Jackson, B.G.,Onipchenko, V.G., Soudzilovskaia, N.A., Tao, J.P., Cornelissen, J.H.C., 2013. Linking litterdecomposition of above- and below-ground organs to plant-soil feedbacks world-wide. J. Ecol. 101, 943–952.

Gong, J.J., Tian, S.X., Wang, Z.Z., Li, G.X., Li, Y.M., 2012. Effect of inoculation and poultrydung on composting of vegetable residues. Chin. J. Environ. Eng. 06, 2813–2817 (inChinese).

Jing, F., Kang, Y.L., Tan, J.X., Tian, B.G., Ma, F.X., Liu, J.G., 2016. Decomposition characteris-tics of cotton stalks from fall to spring as affected by continuous cropping. Acta Agric.Scand. Sect. B Soil Plant Sci. 66, 510–515.

Keiluweit, M., Nico, P., Harmon,M.E., Mao, J., Pett-Ridge, J., Kleber, M., 2015. Long-term lit-ter decomposition controlled by manganese redox cycling. Proc. Natl. Acad. Sci. 112,E5253–E5260.

Kuzyakov, Y., Ruhlmann, J., Geyer, B., 1997. Kinetic and parameters of the decompositionof vegetable crop residues in soil. Gartenbauwissenschaft 62, 151–157.

Li, T., Ye, Y., 2014. Dynamics of decomposition and nutrient release of leaf litter in Kandeliaobovatamangrove forests with different ages in Jiulongjiang Estuary, China. Ecol. Eng.73, 454–460.

Li, Y., Wang, Y.B., Gou, X., Su, Y.B., Wang, G., 2006. Risk assessment of heavymetals in soilsand vegetables around non-ferrous metals mining and smelting sites, Baiyin, China.J. Environ. Sci. (China) 18, 1124–1134.

Li, Y., Li, Y., Chang, S.X., Liang, X., Qin, H., Chen, J., Xu, Q., 2017. Linking soil fungal commu-nity structure and function to soil organic carbon chemical composition in intensivelymanaged subtropical bamboo forests. Soil Biol. Biochem. 107, 19–31.

Luo, Y.Q., Zhao, X.Y., Li, Y.Q., Zuo, X.A., Lian, J., Wang, T., 2016. Root decomposition of Ar-temisia halodendron and its effect on soil nitrogen and soil organic carbon in theHorqin Sandy Land, northeastern China. Ecol. Res. 31, 535–545.

Mackintosh, T.J., Davis, J.A., Thompson, R.M., 2016. Impacts of multiple stressors on eco-system function: leaf decomposition in constructed urban wetlands. Environ. Pollut.208, 221–232.

Manzoni, S., Jackson, R.B., Trofymow, J.A., Porporato, A., 2008. The global stoichiometry oflitter nitrogen mineralization. Science 321, 684–686.

Manzoni, S., Trofymow, J.A., Jackson, R.B., Porporato, A., 2010. Stoichiometric controls oncarbon, nitrogen, and phosphorus dynamics in decomposing litter. Ecol. Monogr.80, 89–106.

Mazareli, R.C.D., Duda, R.M., Leite, V.D., de Oliveira, R.A., 2016. Anaerobic co-digestion ofvegetable waste and swine wastewater in high-rate horizontal reactors with fixedbed. Waste Manag. 52, 112–121.

McCauley, A., Jones, C., Jacobsen, J., 2009. Soil pH and organic matter. 8 pp. 1–12.Mooshammer, M., Wanek, W., Schnecker, J., Wild, B., Leitner, S., Hofhansl, F., Blochl, A.,

Hammerle, I., Frank, A.H., Fuchslueger, L., Keiblinger, K.M., Zechmeister-Boltenstern,S., Richter, A., 2012. Stoichiometric controls of nitrogen and phosphorus cycling indecomposing beech leaf litter. Ecology 93, 770–782.

Nan, Z.R., Zhao, C.Y., 2000. Heavy metal concentrations in gray calcareous soils of Baiyinregion, Gansu Province, PR China. Water Air Soil Pollut. 118, 131–141.

Negussie, A., Degerickx, J., Norgrove, L., Achten, W.M.J., Hadgu, K.M., Aynekulu, E., Muys,B., 2015. Effects of accession, spacing and pruning management on in-situ leaf litterdecomposition of Jatropha curcas L. in Zambia. Biomass Bioenergy 81, 505–513.

Paudel, E., Dossa, G.G.O., de Blecourt, M., Beckschafer, P., Xu, J.C., Harrison, R.D., 2015.Quantifying the factors affecting leaf litter decomposition across a tropical forest dis-turbance gradient. Ecosphere 6.

Polechonska, L., Samecka-Cymerman, A., 2015. The effect of environmental contamina-tion on the decomposition of European frog-bit (Hydrocharis morsus-ranae L.) in nat-ural conditions. Aquat. Bot. 127, 35–43.

Pourhassan, N., Bruno, S., Jewell, M.D., Shipley, B., Roy, S., Bellenger, J.P., 2016. Phosphorusandmicronutrient dynamics during gymnosperm and angiosperm litters decomposi-tion in temperate cold forest from Eastern Canada. Geoderma 273, 25–31.

Provenzano, M.R., Cavallo, O., Malerba, A.D., Di Maria, F., Cucina, M., Massaccesi, L.,Gigliotti, G., 2016. Co-treatment of fruit and vegetable waste in sludge digesters:chemical and spectroscopic investigation by fluorescence and fourier transform infra-red spectroscopy. Waste. Manage 50, 283–289.

621C. Cao et al. / Science of the Total Environment 616–617 (2018) 614–621

Rai, A., Singh, A.K., Ghosal, N., Singh, N., 2016. Understanding the effectiveness of litterfrom tropical dry forests for the restoration of degraded lands. Ecol. Eng. 93, 76–81.

Ranjbar, F., Jalali, M., 2012. Calcium, magnesium, sodium, and potassium release duringdecomposition of some organic residues. Commun. Soil Sci. Plant Anal. 43, 645–659.

State Education Commission of China, 1996. General Rules for Inductivety Coupled Plas-ma-Atomic Emission Spectromety (JY/T 015-1996) (Beijing, China).

State Standard of China, 2002. General Rules Formass Spectrometric Analysis (GB/T 6041-2002) (Beijing, China).

Sun, Y.M., Li, G.X., Zhang, F.D., Shi, C.L., Sun, Z.J., 2005. Status Quo and DevelopmentalStrategy of Agricultural Residues Resources in China. 21 pp. 69–173.

Sun, Z.G., Mou, X.J., Liu, J.S., 2012. Effects of flooding regimes on the decomposition andnutrient dynamics of Calamagrostis angustifolia litter in the Sanjiang Plain of China.Environ. Earth Sci. 66, 2235–2246.

Sun, Z.G., Mou, X.J., Sun,W.L., 2016. Decomposition and heavymetal variations of the typ-ical halophyte litters in coastal marshes of the Yellow River estuary, China.Chemosphere 147, 163–172.

Suseela, V., Tharayil, N., Xing, B., Dukes, J.S., 2014. Warming alters potential enzyme activ-ity but precipitation regulates chemical transformations in grass litter exposed tosimulated climatic changes. Soil Biol. Biochem. 75, 102–112.

Taiz, L., Zeiger, E., 2010. Plant Physiology. Sinauer Associates, Sunderland, Massachusetts.Tsui, M.T.K., Finlay, J.C., Nater, E.A., 2008. Effects of stream water chemistry and tree spe-

cies on release and methylation of mercury during litter decomposition. Environ. Sci.Technol. 42, 8692–8697.

Wang, J.J., Guo, Y.Y., Guo, D.L., Yin, S.L., Kong, D.L., Liu, Y.S., Zeng, H., 2012. Fine root mer-cury heterogeneity: metabolism of lower-order roots as an effective route for mercu-ry removal. Environ. Sci. Technol. 46, 769–777.

Wang, X., Sun, B., Mao, J., Sui, Y., Cao, X., 2012. Structural convergence of maize and wheatstraw during two-year decomposition under different climate conditions. Environ.Sci. Technol. 46, 7159–7165.

Wang, J.J., Tharayil, N., Chow, A.T., Suseela, V., Zeng, H., 2015. Phenolic profile of fine rootshighlights a disconnect in root tissue quality predicted by elemental- and molecular-level carbon composition. New Phytol. 206, 1261–1273.

Wang, F.C., Fang, X.M., Ding, Z.Q., Wan, S.Z., Chen, F.S., 2016. Effects of understory plantroot growth into the litter layer on the leaf litter decomposition of twowoody speciesin a subtropical forest. For. Ecol. Manag. 364, 39–45.

Wu, Y.Y., Wang, C.P., Liu, X.J., Ma, H.L., Wu, J., Zuo, J.N., Wang, K.J., 2016. A newmethod oftwo-phase anaerobic digestion for fruit and vegetable waste treatment. Bioresour.Technol. 211, 16–23.

Xie, Y., Xie, Y., Xiao, H., Chen, X., Li, F., 2017. Controls on litter decomposition of emergentmacrophyte in Dongting lake wetlands. Ecosystems 1–7.

Yeong, K.L., Reynolds, G., Hill, J.K., 2016. Leaf litter decomposition rates in degraded andfragmented tropical rain forests of Borneo. Biotropica 48, 443–452.