@论文综述国外近一个世纪以来在配对集水区研究方面所取得的结论,从水的自然属性出发,从森林变化对径流(年径流量、洪峰与枯水径流)的影响,径流响应的干扰临界值及水文恢复各方面探讨森林变化与径流关系的一致性与复杂性。森林变化与径流关系的一致性主要表现在由较长时间尺度表达的年径流量上。绝大多数的配对集水区的试验研究表明,采伐森林就会增加年径流量,而在荒地上造林就会减少年径流量。而由较短时间尺度表达的洪峰径流与枯水流量则呈现较大的复杂性和难预估性。综述表明,对径流特别是洪峰与枯水径流的定义及分析方法的不同也是造成森林与径流关系复杂性的重要原因。森林与径流关系的复杂性要求人们在研究及应用其关系时就必须有系统观,必须考虑植被、径流与其它过程(土壤变化、气候变化等)的相互作用。论文还认为尽管配对集水区试验作为一种研究方法为研究者提供了许多可靠的结论,但由于许多研究者只把集水区看作是“黑箱”Q从而对认识森林与径流关系的复杂性有一定的局限性。未来的研究应把配对集水区的试验与其它对过程的研究技术(同位素、GIS等技术)结合起来。

Evaluation of the relationship between forest changes and streamflow has an important role in forest planning and water resource management.Although numerous studies on the subject have been conducted in various countries over the last century,controversies still exist.This is mainly due to the complexities of the relationship and the variations between the studied forest ecosystems.This paper examines both consistence and complexity of the forest-streamflow relations by reviewing the topics including the impacts of forest changes on streamflow,harvesting thresholds and hydrological recovery.The literatures used for this review are all from the paired watershed experiments.Our review shows that a consistent or predictable relationship generally exists when examining annual mean flow.The majority of the paired watershed studies demon-strated that harvesting increases annual mean flow,and reforestation from open land decreases it. The relationship between forest change and the short-duration hydrological variables such as peak and low flows appears more complex and less predictable.This is because many factors including changes in forest vegetation,soil condition,harvesting location,etc.interactively affect peak and low flows.Application of various definitions and analytical methods in streamflow also contributes to the complexities.We further conclude that a systematic approach considering interactions between streamflow and other processes or components(e.g.forest vegetation,soil and climate)is needed for understanding the forest-streamflow relation.The paired watershed approach is generally thought as a reliable method.However,it may offer limited use in evaluating complexity of the forest-streamflow relation if the paired watersheds are treated as the"black boxes".We suggest that combination of the paired watersheds with other process-related approaches should be used for studying the relationship between forest changes and streamflow in the future.