土壤中铜和镍的植物毒性预测模型的种间外推验证

在基于物种敏感性分布法推导土壤金属生态阈值过程中,利用毒性预测模型对来源于不同土壤的毒理学数据进行归一化处理可消除土壤性质差异的影响,但目前建立的毒性预测模型仅限于少数物种。本研究通过比较土壤中小白菜、西红柿和大麦的铜和镍的毒性预测模型应用于其他高等植物的预测效果,以及归一化前后各物种毒性阈值的种内变异程度,考察了土壤中铜和镍的植物毒性预测模型种间外推的可行性和适用范围,解决了铜和镍土壤生态阈值导出过程中的方法学问题。土壤中镍对小白菜的毒性预测模型能较好地预测芥菜和青椒的镍毒性阈值,利用该模型对芥菜和青椒在不同土壤中的镍毒性阈值进行归一化后亦能显著降低其种内变异,其种内变异系数分别从1.18和1.25降至0.31和0.06;但将镍对小白菜、西红柿和大麦的毒性预测模型应用于莴笋和莴苣的毒性阈值预测时,在pH<6.0的酸性土壤中其预测值均小于实测值,其实测值与预测值的比值在3.2到6.8之间。对小麦、黄瓜和青椒的铜毒性阈值而言,小白菜模型预测效果优于西红柿和大麦模型。利用西红柿模型归一化黄瓜铜毒性阈值,其毒性阈值的种内变异系数从0.83降至0.14。大麦的铜毒性预测模型能较准确地预测水稻、洋葱、芥菜、包菜和萝卜的毒性阈值,且这5个物种的铜毒性阈值经大麦模型归一化后其种内变异均显著降低。本研究结果可为土壤中铜和镍的植物毒性预测模型的种间外推提供科学依据。

Cross-species Extrapolation of Phytotoxicity Prediction Models for Nickel and Copper Added to Soil

During the process of deriving ecological thresholds for metals added to soils based on SSD (species sensitivity distribution) principles, normalization of the ecotoxicity data with metal toxicity prediction models can eliminate the effect of soil property discrepancy. However, the available toxicity prediction models were based on limited species. Applicability of copper and nickel phytotoxicity prediction models of bok choy, tomato and barley to other non-model higher plants was studied and the intra-species variability of the non-model plants' toxicity thresholds before and after normalization was compared in this study. The probability of cross-species extrapolation of phytotoxicity prediction models for nickel and copper added to soil was probed, and the application boundary of the cross-species extrapolation was confirmed. The methodological issues for the development of copper and nickel ecological thresholds in soil were solved. Results showed that the bok choy nickel phytotoxicity model could be used to predict the toxicity thresholds for mustard and green chilli. After normalization the intra-species variability coefficients of their toxicity threshold decreased from 1.18 and 1.25 to 0.31 and 0.06, respectively. However, the nickel phytotoxicity models of bok choy, tomato and barley could not predict the toxicity thresholds for steam lettuce or leaf lettuce accurately in the soil with pH<6.0. The ratio of measured toxicity threshold to the toxicity threshold calculated with nickel phytotoxicity models ranged from 3.2 to 6.8. The bok choy copper phytotoxicity prediction model could predict the toxicity thresholds for wheat, cucumber and green chilli more accurately than the tomato and barley models. The intra-species variability coefficient of cucumber toxicity threshold decreased from 0.83 to 0.14 after normalization with tomato copper phytotoxicity prediction model. The barley copper phoyotoxicity model could be used to predict and normalize the toxicity thresholds for rice, onion, mustard, cabbage and radish which intra-species variability coefficients decreased significantly after normalization. All the results in this study provide quantitative evidence to support cross-species extrapolation of copper and nickel phytotoxicity prediction models in soils.