【作者】 熊春红；

【导师】 谢明勇；

【作者基本信息】 南昌大学 ， 食品科学， 2011， 博士

【摘要】 在野外采样和室内检测分析的基础上,基于GIS(地理信息系统)技术,利用不同数据挖掘途径,如经典统计学方法、模糊数学方法等,结合地统计学Kriging插值预测,分析了江西省主要茶区土壤和茶鲜叶(新叶和老叶)中的铜(Cu)、铅(Pb)、镉(Cd)、砷(As)、汞(Hg)、铬(Cr)六种重金属的整体污染现状、污染来源影响因素、空间变异特征、空间分布预测和污染发展时间预测,并在此过程中首次提出了茶鲜叶限量值应不同于现行商品茶限量标准,并分析了拟定值建议。茶鲜叶是制茶的直接原料,其重金属现状的评价和预测更有利于反映江西茶产业面临的重金属基础风险状况,为保证安全饮茶和吃茶提供第一手基础数据。从茶叶安全监管的角度,防范风险比事后监测具有更重要的意义。首次在全省主要茶区进行茶鲜叶及土壤的野外采样,以及重金属含量调查和评估,并据此进行预测预警,有利于事前防范重金属污染的发展,及时提出相应的控制对策和解决方案,减少污染治理成本,并有助于更好规划和分级利用宜茶园区,可以为江西茶叶发展、政策研究提供一定实际参考。现将本论文主要研究结果归纳如下：(1)采用描述性统计分析、Nemerow综合污染指数(NIPI)法、多变量分析(聚类分析、判别分析、对应分析),分别从原始含量、污染指数、重金属多变量集合特征等方面进行江西省主要茶区土壤和茶鲜叶的重金属污染评价。综合结论表明,江西茶园土壤和茶鲜叶重金属含量,在基于人为划定的5个茶区间有一定差异但自然地域界线不明显,且各区均有污染水平差异显著的各类茶园。显示了茶园管理水平的差异及管理的重要性。从平均值和中值来看,重金属全量、有效态含量、肥力均为上层(0～20cm)土壤大于下层(20-40cm),但有显著差异的只有全氮、有效磷、有效Pb和有效Cd,空间变异也以中等强度为主,说明江西茶园土壤受一定人为活动影响,但目前的污染增加尚并不十分显著和严重。经上层土壤NIPI分析,采样茶区重金属综合污染水平是：分别有47.7%、77.3%和100%的茶园满足有机茶、绿色茶和无公害茶生产环境土壤重金属要求。下层土壤情况更好。各茶区土壤总体污染程度接近,但主要污染元素不同,As污染指数最高,其次为Cr。赣北茶园土壤Cr的高污染可能与肥料等因素有关。赣南以钨矿、稀土、金矿等储存丰富,因此As的高污染更与成土母质关系密切,由于处于中等空间变异,也与人为因素有一定关系。(2)探讨并拟定了比商品茶叶更严格的限量值进行鲜叶重金属评价以提高可信度,江西茶新叶综合污染指数NIPI为0.99,总体未污染,但警戒水平高；种植环境比较平均良好,路边组与高山组无显著差异；老叶有轻度污染,在进一步以粗老茶加工茶叶食品、加工全茶粉时应引起重视。老叶与新叶的Cu、Pb、Hg、Cr含量均存在极显著差异,随生长期的累积明显；分别有11.4%和6.8%的茶园新叶Cu和Cr轻度污染(2≥NIPI>1);全省Cu、Cr轻度污染,Pb为警戒水平(0.6<NIPI≤1),而Cd、As、Hg状况安全(NIPI≤0.6)。(3)运用模糊数学方法评价了江西省主要茶区土壤综合肥力状况。结果表明,采样区上层土壤肥力IFI均值和中值分别为0.64和0.63,属于中上肥力等级。各主要茶区中,只有A区肥力中下等,而C区肥力高,其余均在中上肥力。其中,赣东北B区肥力差别最大,反映该区域施肥及管理水平差异大。从中值和均值来看,采样茶园的土壤pH值适中偏酸,有机质、全钾、碱解氮含量很高,全氮含量较高,有效钾、全磷含量偏低,有效磷含量很低；据此,建议江西茶园①氮肥充足,不需要增施有机肥；②应整体加强磷肥、钾肥施用量及监测管理；③注意监测并着手改良土壤过酸性发展,投入不足和管理粗放的老茶园,尤其要通过适量施用石灰或灰壳等,以中和土壤酸性,使之适宜茶树的生长。根据模糊数学综合评价指标IFI对江西茶园土壤上层肥力的空间分布进行了普通Kriging插值预测,结果显示,中东部整体肥力高于西部,婺源县为高值中心之一,全省西北部茶区应注意加强施肥管理,提高肥力整体水平。(4)结合采样调查和记录,以Pearson相关分析和因子分析为基础,对江西主要茶区土壤和茶鲜叶六种重金属污染的主要来源进行了初步解析,结果表明区内土壤主要污染重金属为Cr、As,来自成土母质以及人为活动的共同作用,如赣北的工业企业排污污灌、废水废气飘尘、农业活动过程中频繁施用的农药、化肥,赣南丰富的矿藏影响等。茶鲜叶以老叶重金属为主要污染物,老叶与土壤重金属有一定显著相关性,新叶虽与土壤重金属相关性微弱,但新叶与老叶同一重金属间均显著正相关,说明茶鲜叶重金属具有随时间积累特性,生长年限越长,积累越多,并且有可能伴随生长代谢从老叶向新叶转移。因此对土壤重金属带来污染的污染源如成土母质、研究区排污、农业活动过程中施用的农药、化肥、以及汽车尾气、煤烟飘尘也间接成为茶新叶重金属的来源。(5)相关分析显示土壤pH值可以影响茶树植物的代谢对鲜叶重金属积累产生影响。改善茶园土壤pH,减缓土壤酸化趋势,不仅有助于改善江西茶园有效磷的低含量现状,也对新叶、老叶Pb、Cd、Hg等重金属含量有显著抑制或控制作用。同一种的重金属全量、有效态和肥力指标,土壤上下层数据间均显著正相关,显示上下层土壤之间元素及肥力存在渗透转移可能；因子分析结果进一步表明土壤上下层同种重金属列于同一公因子中,说明土壤上下层污染同源；主动吸收为主的新老叶Cu与Cr等另外五种重金属不同源。(6)基于GIS的地统计空间分析,采用普通Kriging方法插值预测了江西茶园土壤和茶新叶重金属的空间变异变异状况,预测误差上ME、MSDE均较小,大多接近0,RMSDE接近1,比较理想。不过RMSE和ASDE虽然接近,但数值偏高,显示该预测精度指标尚不够理想,分析原因,可能与赣北与赣中区域之间采样点距离过大有一定关系。补充赣中泰和县、新余等地采样可以进一步提高预测精度。采用NIPI阂值2(即超过绿色茶环境限量要求)构建土壤上层和下层综合污染水平预测,指示Kriging插值显示,赣西南部分地区,土壤上层综合污染水平有65.6%的概率大于2,达到中度污染,不能满足绿色茶生产环境要求。其余地区均符合绿色茶生产要求。土壤下层综合污染水平明显好于上层,说明江西目前人为活动造成的污染尚停留在表层土壤,及时关注重金属控制为时未晚。采用NIPI阈值0.6(安全水平,即新拟茶新叶限量的60%)构建新叶和老叶综合污染水平预测,指示Kriging插值显示,茶新叶和老叶污染超出安全限的概率较高的地区均在南昌周边呈发散性分布,反映了工业化水平、人为活动对茶鲜叶重金属也产生影响。(7)基于时相的污染物含量预测研究。利用情景分析法分别按乐观情景和无突变情景预测了研究区2025年土壤和茶新叶重金属污染情况及全省空间变异分布,同时绘制了相关的Kriging插值预测图。结果表明即使在乐观情景下即所有污染物停止输入,逐步自然净化,到2025年区内上层土壤重金属中值和均值下降也不非常明显,降低率在2.4%～14.5%之间；但茶新叶中值基本无变化。而在无突变情景下即保持现状不变,到2025年区内土壤重金属污染程度更加严重,上层土壤Cr、Hg、As、Cd及茶新叶Cu、Cd、As、Cr超过无公害茶园环境或茶叶最高限量的比例分别增加至2.3%、6.8%、22.7%、2.3%和45.5%、2.3%、4.6%、68.2%。江西茶园急需注意控制重金属发展,重点是As、Cr,不产生恶化,2025年茶新叶才可以保持以绿色茶、有机茶产品为主。(8)运用Plackett-Burman设计和响应面法,探讨并首次建立了超声波辅助消解—原子光谱法快速测定茶叶Hg和Pb的方法。茶叶Hg最优提取条件为：浸泡时间6 min、超声时间8.1 min,超声温度70.5℃,HN03：H202(1：1,V/V)4.4 mL,称样量300 mg,超声频率40kHz。茶叶Pb最优提取条件为：200mg样品中加入HN03：H202(1：1,V/V)5mL浸泡6min、78℃超声8.4min,超声频率40kHz。经鲜叶及商品茶测定检验,超声辅助消解法与国标法测定结果无显著差异,但大大缩短消解时间,减少试剂消耗和污染。经过几年的摸索研究工作,在下列三个方面有所创新：(1)本研究是首次针对江西全省主要茶区野外自然生长的茶鲜叶进行采样调查。不同于盆栽试验,能更直接反映江西商品茶生产的直接原料目前的重金属含量状况。为评价本省茶叶原料质量,反映茶产业管理水平、寻找重金属污染发展控制因素及措施,提供第一手资料和参考。(2)通过GIS技术,首次插值计算描绘了江西茶新叶、茶园土壤Cu、Pb、Cd、As、Hg、Cr含量的空间分布特征图,以及基于六种重金属综合污染指数(NIPI)的污染概率图。利用空间预测图和污染概率图联合界定了研究区各污染物的污染范围或程度,为污染修复和补充采样提供了决策信息,在研究手段上具有集成创新。同时,将经典统计分析、多变量分析(如聚类、对应分析、判别分析)和模糊数学等数据挖掘方法应用于茶园重金属和肥力的评价,更有利于从集合角度获取宏观评价结果,在研究思路上有一定集成创新。(3)首次提出了茶鲜叶重金属限量评价标准的问题并拟定了新的三级评价指标。提出应在现有有机茶、绿色茶、无公害茶限标的基础上考虑扣除后期加工污染的影响程度,采用更严格的限量值。有利于解决目前茶鲜叶重金属与商品茶重金属评价时标准混用的不严格性。更多还原

【Abstract】 After sampling in natural tea plantation and chemical analysis indoors, with the support of GIS (Geographic Information System) technology especially like geostatistical Kriging interpolation, several data mining technologies, such as classical statistics and fuzzy mathematics were chosen to research the main sources of pollution, spatial variability characters and their predition and temporal prediction for six types of heavy metals(Cu, Pb, Cd, As, Hg and Cr) in soils and fresh tea leaves in major tea regions of Jiangxi provice. Firstly, more tightening limited values for fresh tea leaves were suggested and discussed in this study other than present national limited values for branded tea. Fresh tea leaves were the first-hand raw materials for branded tea, so we could get the basic risk assessment data we were facing while tea were drinked or eaten by evaluating the pollution stadus and predicting for them. It was more significative to keep watch risk than survey afterwords for monitoring safty of tea. It’s the first time to sample outdoors in Jiangxi major tea regions with soils and fresh tea leaves followed by examing, estimating and predicting.The main results obtained in this dissertation were summarized as follows:1. Descriptive statistics, Nemerow integrated pollution index (NIPI) and multivariate analysis (cluster analysis, discriminant analysis and correspondence analysis) were used to get assessment results for soils and fresh tea leaves in Jiangxi major tea regions with respect to initial concentrations, pollution index and integrated characteristics of multivariance. In conclusion, there was a slight difference among heavy metals in soils and fresh leaves from the five tea producing districts divided artificially in Jiangxi but without clear natural separatrix. However, there were different tea gardens at different pollution level in each district, which indicated their various management levels and the importance for better running. Considering mean and medium value, whole concentrations or avalible concentrations of heavy metals and fertility level value in upper soils (0-20cm) were higher than it in sublayers (20～40cm) but without significant differences except the whole nitrogen, avalible phosphorus, avalible Pb and avalible Cd mainly with middle level of spatial variability. It indicated that soils pollution was just slightly affected by human activities without serious change. Respectively, there were 47.7%,77.3% and 100% of tea gardens which could be suitable for planting organic tea, green tea or nuisanceless tea conformed to its related environmental standard when considering NIPI in upper soil. The pollution level in sublayer was slighter. The pollution level in various districts was close with different main heavy metal. The highest NIPI was for As, then the Cr. The higher polution of Cr in northern district in Jiangxi was assumed to be related to fertilizer, while the higher polution of As in southern district in Jiangxi was assumed to be mainly related to soil parent material combined with humanity.2. More tightening limit values for fresh tea leaves other than branded tea are more reliable when used to assess metal contamination in fresh tea leaves. Based on it, NIPI of new tea leaves was caculated to be 0.99 indicating high precautious level while old leaves reach to slightly polluted level which reminded us to be carefull to use it as leaf food or powder additives. It’s good and uniform or similar in natural producing condition in Jiangxi and there are no great differences between the mountain group and the road one. The results revealed that concentrations of Cu, Pb, Hg and Cr showed highly significant differences within new leaves and old leaves which indicated obvious accumulation accompanying with years. Respectively, Cu and Cr of new leaves in 11.4% and 6.8% tea garden came to slightly polluted level (2≥NIPI>1). Appraisal results of Jiangxi province were slightly polluted for Cu and Cr, precautious for Pb (0.6<NIPI≤1), while safe for Cd, As and Hg (NIPI≤0.6).3. A fuzzy mathematics method was adopted to estimate the stadus of soils fertility level in Jiangxi major tea regions. The results reveal that integrated fertility index (IFI) of upper soil sampled was 0.64 for mean while 0.63 for medium which indicated to be better than average expect region of number A in north-western Jiangxi where was worse than average. Among these regions, it’s the region of number B which showed the most diversity of IF I related to the most difference on fertilizer management. For mean and medium value of soils, pH was on the little low side with higher for organic matter, whole potassium and alkali-hydrolyzable nitrogen, a little rich for whole nitrogen, and a little low side for avalible potassium and whole phosphorous, while worst for avalible phosphorous. So it was suggested that fertilizer with potassium and phosphorous should be applied more than organic fertilizer in Jiangxi tea regions combined with better or more regular management to prevent the deterioration of pH especially in extensive planting tea gardens. Lime or plant ash was a good choice for neutralization to get proper soil pH for tea plant. Based on IFI, ordinary Kriging interpolation was used to predeterminating the spatial variability of upper soil. It showed that IFIs were higher in middle and east regions with a high value center of Wuyuan county than north-west parts in Jiangxi.4. Pearson correlation analysis and factor analysis were executed to analysis the major resources of heavy metals pollution in Jiangxi main tea regions consulting records as sampling. The results indicated that As and Cr were the major pollution matters brought from soil parent material combined with humanity, such as polluted water, sewage, exhaust gas and floating dust arised from industry in northern Jiangxi, rebundant fertilizer in agriculture and rich mineral reserves in southern Jiangxi. It was the old leaves that were mainly polluted for fresh leaves as to show a little significant correlation with soil heavy metals, while new ones were weakly correlated with soil. However, new leaves were significant to old ones considering heavy metals which indicated the possibility of heavy metals thansfer during the leaves of two types. Thus, major resources of heavy metals pollution for soils also were the major resources for new tea leaves.5. Correlation analysis revealed that pH of soils could effect accumulation of heavy metals in fresh leaves through influence metabolization of plant. Not only for improving the level of avalible phosphrous, but also for controling the concentrations of Pb、Cd、Hg in old leaves, it was important to prevent deterioration of pH. There were significant correlations for the same heavy metal concentrations whole or avalible between two layers of soils, the same as the fertility level. It indicated a possibility of thansport in heavy metals within them. Furthermore, factor analysis showed the same metal in two layers was caculated into the same factor representing the same pollution resources. However, there were no same resources for Cu and apart from the other five metals of fresh leaves.6. With the support of GIS geostatistical ordinary Kriging interpolation, spatial variability characters and their predition were caculated. For prediction error, it was almost satisfied with ME and MSDE almost zero and RMSDE near to one, while RMSE and ASDE was a little higher than zero. Prediction precision could be better after additional sampling in middle region in Jiangxi like Taihe county or Xinyu city to shorten the sampling distance. Threshold value of NIP] of two was used to make a indicator Kriging interpolation for integrated pollution level of soils. The results revealed that probability of 65.6% exsited in south-western Jiangxi when NIPI exceeding the value of two which referred to moderate pollution and not suitable to plant green tea. Other regions were better for producing green tea and the sublayer was better than the upper one, which indicated just upper pollution presented that shoule be modified earlier. Threshold value of NIPI of 0.6 was used to make a indicator Kriging interpolation for integrated pollution level of leaves. The results showed that regions with higher probability of NIPI exceeding 0.6 spreaded around the provincial capital of Nanchang city, which also indicated the effect of industy and humanity activities on leaves pollution.7. Temporal prediction divided into optimistic scene and unmutation scene were used to predict the heavy metals concentration of studied area in 2025 as to the Kriging interpolation maps. The results showed that the concentration will just decrease a little in optimistic scene with 2.4%～4.5%, and medium value of new leaves nearly decreased; Furthermore, it will be getting worse in unmutation scene. Cr, Hg, As and Cd in upper soil and Cu, Cd, As, Cr in new leaves will exceed the top threshold value with percentage of 2.3%,6.8%,22.7%,2.3%,45.5%,2.3%,4.6% and 68.2%, respectively. It’s necessary to pay attention to heavy metals pollution in tea gardens for we can still produce organic or green tea in future. 8. Rapid ultrasound-assisted extraction (UAE) procedures were firstly developed for the determination of Hg and Pb in tea leaves followed by atomic spectrometry. A Plackett-Burman design followed by response surface methodology was used to get the optimized results:for Hg, with 6 min of presonication time,8.1 min of sonication time,70.5℃of ultrasonic bath temperature,4.4 mL of HNO3: H2O2(1:1, V/V) and 300 mg of sample mass; for Pb,200mg sample added HNO3: H2O2 (1:1, V/V) 5mL should be treated with UAE for 8.4min at 78.0℃after soaking 6min at 40 kHz of ultrasound frequency. No significant differences were established between the analytical results of UAE method and national standard method.Exploring for several years, three fields of innovations or new developments in this study were showed as follows:1. It’s the first time for us to sample fresh tea leaves in natural tea plantation outdoors of Jiangxi at home and abroad. Other than pot incubation experiment, it’s more representative of current situation of heavy metals pollution for it’s raw materials for branded tea. Thus the study of fresh tea leaves could be better for assessing the quality of tea in Jiangxi, finding the problem in tea plantation management, drawing up countermeasures and resolving methods in advance to hold back the deterioration of heavy metals pollution in tea. Also, it offered a first-hand practical reference for tea industry development.2. Prediction maps, as the same as probability maps of integrated pollution level, were caculated for the first time for spatial variability of heavy metals in fresh tea leaves and soils in Jiangxi province based on GIS Kriging interpolation. The range and the degree of pollution were defined for offering information to make up policy about recovering and additional sapmling. It’s an integrated innovation on study methods. Also, there was an integrated innovation on study ideas for classic statistics, multivariate analysis (cluster analysis, discriminant analysis and correspondence analysis) and fuzzy mathmatics were fistly integrated to get more macroscopical assessment results for soils and fresh tea leaves in Jiangxi major tea regions. 3. A new standard with three degree of limited values was announced and discussed firstly for fresh tea leaves. Processing pollution levels following plucking should be eliminated from the standard values of branded tea to assess the fresh leaves more properly and tightening.更多还原