从生态环境角度论鄱阳湖区农业可持续发展

鄱阳湖区农业资源十分丰富，具有发展大农业的优越条件。本文通过对湖区农业生态环境的脆弱性分析，指出湖区农业的非良性循环特性，提出防洪、抗旱、治涝、血吸虫病治理等对策措施，为湖区农业可持续发展提供决策参考。     

生物反应器填埋场渗滤液及其处理对策

根据垃圾的降解和稳定机理分析了生物反应器填埋场渗滤液的动态变化规律，并对其处理对策进行了探讨。生物反应器填埋场前期产生的渗滤液可全部回灌，后期再根据实际情况排放处理部分过量的渗滤液。在对渗滤液水量和水质全面直观认识后，再对生物反应器填埋场渗滤液处理工艺进行针对性的选择和设计，并宜采用物化法处理方式。     

总量控制下环境容量产权交易及政府行为分析

利用经济学的原理，本文分析了环境容量的经济特点(不是环境容量产权交易的特征)，提出了环境容量产权交易的可行性和必要性。同时针对政府在环境容量产权交易中的作用进行了必要的分析。     

生态小区特征和指标体系探讨

保护和建设生态小区，推行生态细胞工程是自然生态工作的有力补充。对于目前出现的许多生态小区，中国尚无一套正式通用的生态小区考评指标体系。本文从生态学的角度出发，探讨了生态小区的概念、内涵与其特征，并以此为基础，提出了生态小区的考评指标体系，希望以此引导四川省生态小区的创建活动。     

五日生化需氧量(BOD5)测定方法探讨

本文针对GB／T7488-1987《水质五日生化需氧量(BOD5)稀释与接种法》中BOD5的测定和计算时接种稀释水的空白扣除方法，提出了直接测接种液的BOD5值，按接种液的比例在水样中进行扣除的方法，提高了BOD5值的准确性。更能反应BOD5的真实性。     

数据包络分析(DEA)模型在清洁生产评价中的应用

清洁生产技术方案评价对于指导企业实施清洁生产有重要意义。将污染物产生量数据作为一种输入引入数据包络分析(DEA)模型，利用该模型对清洁生产技术方案中的投入、产出和污染物发生量进行有效性评价。分析不同方案相对效率值、目标改进值，发现最有效的清洁生产技术方案，达到经济效益和环境效益的统一。本研究为企业筛选和决策可实施清洁生产技术方案提供技术理论支持。     

Genetically Modified (GM) Crops: Precautionary Science and Conflicts of Interests

Risk governance of GM plants and GMfood products is presently subject to heatedscientific and public controversies. Scientistsand representatives of the biotechnologyindustry have dominated debates concerningsafety issues. The public is suspicious withregard to the motives of scientists, companies,and political institutions involved. Thedilemmas posed are nested, embracing valuequestions, scientific uncertainty, andcontextual issues. The obvious lack of data andinsufficient information concerning ecologicaleffects call for application of thePrecautionary Principle (PP). There are,however, divergent opinions among scientistsabout the relevance of putative hazards,definition of potential ``adverse effects,' andwhether actions should be taken to preventharm. The reliance on the concept ofsubstantial equivalence in safety evaluation ofGM food is equally controversial. Consequently,value assumptions embedded in a scientificframework may be a barrier for employment ofthe PP. One of our major conclusions is thatprecautionary GMP usage requires riskassessment criteria yet undeveloped, as well asbroader and more long-term conceptions of risk,uncertainty, and ignorance. Conflicts ofinterest and public participation are otherissues that need to be taken intoconsideration. GMP governance regimes that arejustifiable from a precautionary and ethicalpoint of view must transcend traditionalscientific boundaries to include alternativescientific perspectives as well as publicinvolvement.     

论国家生态安全

国家生态安全是一种全新的现代国家安全观，要重视和维护国家生态安全。     

BUILDUP,WASHOFF, AND EVENT MEAN CONCENTRATIONS1

ABSTRACT: Most watershed water quality simulation models require the user to specify pollutant buildup and washoff rate parameters for pollutants, by land use. Buildup and washoff rates are difficult to measure directly, and only limited guidance and few observed data are available from the literature. Many studies, however, report storm event mean concentrations (EMCs). These EMCs must arise as a result of the buildup and washoff processes, but typically represent the net contribution from a variety of pervious and impervious surfaces. This paper explores the relationship between EMCs and buildup/washoff parameters. An assumption of the mathematical form of the buildup/washoff relationship gives an algebraic expression for the EMC consistent with model assumptions. This yields techniques to separate observed EMCs into contributions from different land uses and from pervious and impervious surfaces. Given this relationship, numerical optimization may be used to estimate site specific values of buildup and washoff parameters from observed storm EMCs for use in modeling. Use of this approach helps ensure that model parameters are consistent with observed data, providing a rational starting point for final model calibration. Several site examples demonstrate use of the method.     

入侵检测的1类支持向量机模型

计算机网络尤其是互联网的迅速发展与普及 ,使得信息安全已经成为一个全球瞩目的重要研究课题。随着攻击技术的不断进步与更新 ,迫切需要一种有效的入侵检测技术来保护信息系统的安全。由于几乎所有的攻击与滥用都被记录在系统的网络数据中 ,因而可以基于计算机系统的网络数据构造入侵检测系统。在对网络数据进行深刻的分析和研究的基础上 ,提出了入侵检测的 1类支持向量机模型。第一 ,构造适于异常点检测的1类支持向量机模型 ;第二利用抽象化的网络数据对该模型进行训练以确定其中各个参数的值。实验表明 ,该方法是行之有效的