环境科学  2014, Vol. 35 Issue (1): 299-303   PDF    
Biotic Ligand Model的简化模型及预测性能评价
王万宾, 陈莎, 吴敏, 苏德丽, 赵婧    
昆明理工大学环境科学与工程学院, 环境土壤科学重点实验室, 昆明 650500
摘要:通过检索4物种(Fathead minnow、D.magna、D.pulex、Rainbow trout)在地表水中实测的铜半致死浓度(Observed_LC50),及Biotic Ligand Model(BLM)预测其半致死浓度(Predicted_LC50),得到4物种的预测精度依次为0.075、0.52、0.96、0.29,模型对Fathead minnow与Rainbow trout的预测性能较差. 在此基础上,分析显示预测误差值与LA50呈指数关系,表明LA50值并非常数值. 通过对BLM的LA50的校正,Fathead minnow与Rainbow trout的预测精度升为0.59、0.42. 通过分析LA50与硬度的关系,发现BLM在软水环境中预测效果较差. 另外,随机均匀生成500组水质参数组,通过BLM预测,筛选出4项敏感参数为DOC、pH、HCO3-浓度及温度,并建立相应物种的LC50与其的多元线性关系,大大简化了生物配位模型.
关键词BLM     4物种     预测性能     模型简化     LA50    
Simplification of Biotic Ligand Model and Evaluation of Predicted Results
WANG Wan-bin, CHEN Sha, WU Min, SU De-li, ZHAO Jing    
Key Laboratory of Environmental Soil Science, Faculty of Environmental Science and Engineering, Kunming University of Science and Technology, Kunming 650500, China
Abstract: The prediction accuracy of LC50 on four species (Fathead minnow, D.magna, D.pulex, Rainbow trout) was 0.075, 0.52, 0.96 and 0.29 respectively as determined by their onserved values of LC50 in surface water. Predicted results indicated that the correlation between forecast error and LA50 was exponential. The accuracy of Fathead minnow and Rainbow trout became 0.59 and 0.42 after adjusting LA50. The correlation between hardness and LA50 showed that the prediction effectiveness of BLM was poor in soft water. In addition, four important parameters (DOC, pH values, the concentration of HCO3-, temperature) were selected to build the multiple linear relationship with LC50 by applying 500 groups of random uniform water quality parameter in BLM. Biotic ligand model was effectively simplified.
Key words: BLM     four species     predicted results     model simplification     LA50    

水环境中重金属的形态与其生物有效性(毒性)有直接的关系,并不是所有形态的重金属都具有毒性. 因此在金属水质标准制定或金属生态风险评价中,考虑生物有效性问题至关重要. 然而重金属的生物有效性与环境介质中水质参数密切相关,并且物种的生理特性也影响其敏感程度. 生物配位模型(Biotic Ligand Model)用于解释及预测水化学环境中重金属对有机体的急性毒性影响,毒性定义为生物体内的重金属积累超过一个重要阈值(LA50)而对生物体产生的负面影响[1]. 生物配位模型已被广泛用于各类水体中的重金属对生物体毒性的预测[2, 3, 4, 5],其预测精度较高,机制解释清晰. 但大量文献报道[6, 7, 8, 9],生物配位模型也存在明显的缺陷,即对较高或较低pH、 硬度水体、 较多POM水体等预测性能较差. 生物配位模型应用需要考虑众多水环境参数[10](温度、 pH、 DOC、 Ca、 Mg、 Na、 K、 SO2-4、 碱度等),制约了模型更广泛的推广应用. 本研究在Biotic Ligand Model Windows Interface, Version 2.2.3平台下,通过对分布广泛的4物种(Fathead minnow、 D.magna、 D.pulex、 Rainbow trout)文献检索其具体水体应用中的铜LC50数据,评价生物配位模型的预测精度. 在此基础上得出预测效果差的关键因子及影响该因子的水质参数,然后通过因子校正,预测性能明显提高. 另外,本研究通过随机生成水质参数组,利用BLM预测其LC50,然后通过统计分析评价模型各水质参数的相对灵敏度,并在此基础上对BLM加以简化. 模型的预测性能评价将为生物配位模型的进一步完善及新物种的生物配位模型建立提供参考.

1 材料与方法
1.1 毒性数据的获取及水质参数组的随机生成

本研究的LC50数据检索涵盖了不同水体(湖泊、 河流)和较广泛的水质参数范围,具有较好代表性. 指标除了模型本身所需的输入参数外,还包括硬度,LC50[11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22]. 水质参数范围及毒性数据数量见表 1.

表 1 铜对4物种的毒性数据汇总 1)

Table 1 Statistics of cupreous toxicity data about the four species

随机均匀生成500组水质参数,各水质参数的范围为中国地表水体的主要含量范围,见表 2.

表 2 水质参数范围表 Table 2 Range of water quality parameters

1.2 BLM应用简介

本研究应用软件平台为Biotic Ligand Model Windows Interface, Version 2.2.3(HydroQual,Inc.). 该软件为模型应用平台的最新版本,已被广泛采用. 该软件只需注重模型的输入输出,并且方便修改模型参数,所以使用较为简便. 利用4物种的Observed_LC50相对应的水质参数,输入软件中可预测其Predicted_ LC50. 比较Observed_LC50及Predicted_ LC50,可得到模型对4物种的预测精度. 在软件形态分布平台下,利用4物种的Observed_LC50及其相对应的水质参数,输入软件中可得到其LA50. 软件主要默认参数见表 3.

表 3 BLM关键参数表

Table 3 Key parameters of BLM
1.3 模型简化及精度评价方法

采用多元线性统计中的逐步回归法进行模型简化,筛选出灵敏度较高且所占影响比重较大的因子组. 预测误差设定为Predicted/Observed Value(P/O), 0.5<P/O<2为可接受范围内. 预测精度定义为N(0.5<P/O<2)/N(total),其中N代表数量.

2 结果与讨论
2.1 BLM预测值与实测值比较

图 1可计算出,4物种(Fathead minnow、 D.magna、 Rainbow trout、 D.pulex)的预测精度值分别为0.075、 0.52、 0.29、 0.96. D.magna、 D.pulex的预测效果较好,而Fathead minnow、 Rainbow trout的预测值普遍高于实测值,这说明BLM中存在某个参数的值与实际不相符合.

图 1 4物种的预测值与实测值比较图

Fig. 1 Log-log plots of predicted versus observed copper concentration associated with the 50% of mortality (LC50) for the four species
2.2 半致死累积量(LA50)与预测精度的关系

利用毒性数据,通过BLM软件铜的形态分布分析获取4物种的半致死体内积累量(LA50). LA50并非模型描述的为一个常量,其随着水环境参数的变化会相应变化. 图 2利用预测误差与LA50作图,发现预测误差的好坏与LA50值密切相关,其统计分析见表 4.

图 2 4物种的预测误差与LA50的关系

Fig. 2 plots of P/O versus LA50 for the four species

表 4 预测误差与LA50的拟合关系式 1)

Table 4 Functional expression between forecast error and LA50

表 4分析可知,LA50与P/O之间呈指数关系,拟合效果较好. LA50值是否合理直接关系到预测精度的好坏,所以找到一个合理的LA50值或者建立其与影响因子之间的关系式显得较为重要. 物种的本身生理特征(年龄,大小等)对LA50值有所影响,水环境化学参数(如pH、 硬度等)对LA50值的影响更不容忽视. 目前水质参数对LA50的影响机制研究较少,在BLM里已把LA50假设为一个常数值.

2.3 BLM的 LA50的校正

由上述叙述可知,提高预测精度的重要方法为修正模型中的LA50值. 4物种的Cu形态分布LA50统计显示如表 5.

表 5 4物种LA50统计

Table 5 Statistics of LA50 for four species

表 5分析可知,物种Fathead minnow和Rainbow trout的均值远远小于模型默认值,即该两种物种模型的LA50值明显偏高. 利用均值对物种Fathead minnow和Rainbow trout的LA50进行校正,校正后结果见图 3.

图 3 2物种的预测值与实测值比较

Fig. 3 Log-log plots of predicted versus observed copper concentration associated with the 50% of mortality (LC50) for two species

物种 Fathead minnow和Rainbow trout校正前预测精度分别为 0.075和0.29,而校正后依次为 0.59和0.42,预测性能明显提高. 因此,在生物配位模型应用中,选择一个合理的 LA50值较为重要.

2.4 半致死累积量(LA50)与水体硬度的关系

由于LA50会随水质参数的变化而变化,LA50与水体硬度的关系见图 4. 从中可见,随着硬度的增加,LA50值呈现上升的趋势,并逐渐接近于模型默认值. 图 4也显示,在软水环境[23]中, 3物种(Fathead minnow、 D.magna、 Rainbow trout)的LA50值远小于模型默认值,即生物配位模型不适合于软水环境中的铜对物种的毒性预测. 实线代表模型默认LA50值,虚线代表校正LA50

图 4 4物种的LA50与硬度关系 Fig. 4 Plots of hardness versus LA50 for the four species
2.5 BLM的简化

生物配位模型能较好诠释水环境参数对重金属生物有效性的影响,但其需较多水质参数,而很多水质参数较难检测或者代价昂贵,这制约着BLM在经济科技不发达地区的推广应用. 本研究基于该模型,筛选出4个重要参数(pH、 DOC、 碱度、 温度),建立了其与LC50之间的多元线性关系,见表 6.

表 6 4物种的模型简化线性式

Table 6 Simplified linear expression of the four species

在实际应用中,DOC、 pH、 碱度、 温度这4个参数对毒性的影响非常敏感,其解释了生物有效性的很大一部分. 对于特定水环境,温度在不同季节会有较大差异,所以在生态风险评价或者基准建立中,温度指标应给与足够重视.

3 结论

(1)生物配位模型能较好预测D.magna、 D.pulex,而对Fathead minnow、 Rainbow trout的预测效果较差.

(2)半致死累积量(LA50)并非一个常数值,其直接影响模型预测的好坏. 建立新物种的BLM时,应重视LA50的修改或计算.

(3)软水环境中,BLM预测结果较差,其原因为硬度对LA50值的影响. 模型应该考虑水体的总硬度,而不只是单纯的Ca、 Mg的影响.

(4)4物种简化的线性模型较好,能为相对不发达地区提供生态风险评价依据.

参考文献
[1] Di Toro D M, Allen H E, Bergman H L, et al. Biotic ligand model of the acute toxicity of metals. 1. Technical basis[J]. Environmental Toxicology and Chemistry, 2001, 20 (10): 2383-2396.
[2] Constantino C, Scrimshaw M, Comber S, et al. An evaluation of biotic ligand models predicting acute copper toxicity to Daphnia magna in wastewater effluent[J]. Environmental Toxicology and Chemistry, 2011, 30 (4): 852-860.
[3] Villavicencio G, Urrestarazu P, Carvajal C, et al. Biotic ligand model prediction of copper toxicity to daphnids in a range of natural waters in Chile[J]. Environmental Toxicology and Chemistry, 2005, 24 (5): 1287-1299.
[4] Natale O E, Leis M V. Biotic Ligand Model estimation of copper bioavailability and toxicity in the Yacyretá Reservoir on the Paraná River[J]. Lakes & Reservoirs: Research & Management, 2008, 13 (3): 231-244.
[5] Santore R C, Di Toro D M, Paquin P R, et al. Biotic ligand model of the acute toxicity of metals. 2. Application to acute copper toxicity in freshwater fish and Daphnia[J]. Environmental Toxicology and Chemistry, 2001, 20 (10): 2397-2402.
[6] Ng T Y T, Chowdhury M J, Wood C M. Can the biotic ligand model predict Cu toxicity across a range of pHs in softwater-acclimated rainbow trout?[J]. Environmental Science & Technology, 2010, 44 (16): 6263-6268.
[7] Ryan A C, Tomasso J R, Klaine S J. Influence of pH, hardness, dissolved organic carbon concentration, and dissolved organic matter source on the acute toxicity of copper to Daphnia magna in soft waters: implications for the biotic ligand model[J]. Environmental Toxicology and Chemistry, 2009, 28 (8): 1663-1670.
[8] Taylor L N, Wood C M, McDonald D G. An evaluation of sodium loss and gill metal binding properties in rainbow trout and yellow perch to explain species differences in copper tolerance[J]. Environmental Toxicology and Chemistry, 2003, 22 (9): 2159-2166.
[9] Meyer J S, Boese C J, Collyard S A. Whole-body accumulation of copper predicts acute toxicity to an aquatic oligochaete (Lumbriculus variegatus) as pH and calcium are varied[J]. Comparative Biochemistry and Physiology Part C: Toxicology & Pharmacology, 2002, 133 (1): 99-109.
[10] HydroQual Inc, Mahwah N J. The Biotic Ligand Model windows interface[M]. Version 2.2.3: User's Guide and Reference Manual, 2005.
[11] Sciera K L, Isely J J, Tomasso J R, et al. Influence of multiple water-quality characteristics on copper toxicity to fathead minnows (Pimephales promelas)[J]. Environmental Toxicology and Chemistry, 2004, 23 (12): 2900-2905.
[12] Welsh P G, Skidmore J F, Spry D J, et al. Effect of pH and dissolved organic carbon on the toxicity of copper to larval fathead minnow (Pimephales promelas) in natural lake waters of low alkalinity[J]. Canadian Journal of Fisheries and Aquatic Sciences, 1993, 50 (7): 1356-1362.
[13] Welsh P G, Parrott J L, Dixon D G, et al. Estimating acute copper toxicity to larval fathead minnow (Pimephales promelas) in soft water from measurements of dissolved organic carbon, calcium, and pH[J]. Canadian Journal of Fisheries and Aquatic Sciences, 1996, 53 (6): 1263-1271.
[14] Vardy D W, Oellers J, Doering J A, et al. Sensitivity of early life stages of white sturgeon, rainbow trout, and fathead minnow to copper[J]. Ecotoxicology, 2013, 22 (1): 139-147.
[15] Van Genderen E, Gensemer R, Smith C, et al. Evaluation of the Biotic Ligand Model relative to other site-specific criteria derivation methods for copper in surface waters with elevated hardness[J]. Aquatic Toxicology, 2007, 84 (2): 279-291.
[16] Kramer K J M, Jak R G, Van Hattum B, et al. Copper toxicity in relation to surface water-dissolved organic matter: Biological effects to Daphnia magna[J]. Environmental Toxicology and Chemistry, 2004, 23 (12): 2971-2980.
[17] De Schamphelaere K A C, Heijerick D G, Janssen C R. Refinement and field validation of a biotic ligand model predicting acute copper toxicity to Daphnia magna[J]. Comparative Biochemistry and Physiology Part C: Toxicology & Pharmacology, 2002, 133 (1): 243-258.
[18] Long K E, Van Genderen E J, Klaine S J. The effects of low hardness and pH on copper toxicity to Daphnia magna[J]. Environmental Toxicology and Chemistry, 2004, 23 (1): 72-75.
[19] Welsh P G, Lipton J, Mebane C A, et al. Influence of flow-through and renewal exposures on the toxicity of copper to rainbow trout[J]. Ecotoxicology and Environmental Safety, 2008, 69 (2): 199-208.
[20] Hansen J A, Lipton J, Welsh P G. Relative sensitivity of bull trout (Salvelinus confluentus) and rainbow trout (Oncorhynchus mykiss) to acute copper toxicity[J]. Environmental Toxicology and Chemistry, 2002, 21 (3): 633-639.
[21] Little E E, Calfee R D, Linder G. Toxicity of copper to early-life stage Kootenai River white sturgeon, Columbia River white sturgeon, and rainbow trout[J]. Archives of Environmental Contamination and Toxicology, 2012, 63 (3): 400-408.
[22] Besser J M, Mebane C A, Mount D R, et al. Sensitivity of mottled sculpins (Cottus bairdi) and rainbow trout (Onchorhynchus mykiss) to acute and chronic toxicity of cadmium, copper, and zinc[J]. Environmental Toxicology and Chemistry, 2007, 26 (8): 1657-1665.
[23] SL 395-2007, 地表水资源质量评价技术规程[S].