From chronology to frequency pairing: a paradigm shift in the science of forests and floods

森林与洪水科学卷入了冲突，并迫切需要根据气候变化、虫害流行、伐木和毁林进行重新评估。我们对森林和洪水关系的科学认识是通过几十年的成对流域研究进一步确认的。然而，这些研究的结果最近受到了挑战，因为一个不适当的实验设计因素是配对流域研究方法的基础。几十年的成对流域研究集中于调查在相同的气象或风暴输入的情况下，收获前和收获后洪水的幅度变化。这种按时间顺序排列的事件配对不能解释因采伐引起的峰值流量频率的变化，因此从森林和洪水科学中剥离了物理学的关键部分。森林、森林采伐或森林砍伐改变了产生洪水的过程，不仅改变了洪水的规模，还改变了洪水的频率。由于洪水事件的这两个属性之间的高度非线性联系，它们应该同时进行评估。这最好通过峰值流量频率分布的框架来完成，而不是方差和协方差的分析，传统上在几十年的森林水文文献中进行。两个密不可分的问题应该指导森林和洪水关系的研究调查：对频率（震级）感兴趣的洪水事件的震级（频率）变化是什么？这种调查需要频率配对事件分析，其中流域的水文响应是整个洪水频率分布。我们将实地调查与数值模拟相结合，利用新的等频率配对洪水事件的实验设计，重新评估我们对森林与洪水关系的科学认知。这项分析是在广泛的流域规模、水文气候制度和造林系统范围内进行的。

The science of forests and floods is embroiled in conflict and is in urgent need of reevaluation in light of changing climates, insect epidemics, logging, and deforestation. Our scientific perception of the forests and floods relation is preconceived by further reaffirmed by decades of paired watershed studies. However, the outcomes of these studies had recently been challenged because of an inappropriate experimental design factor fundamental to the paired watershed study approach. Decades of paired watershed studies focused on investigating a change in magnitude between pre- and post-harvesting floods when paired by equal meteorological or storm input. This type of chronological event pairing fails to account for the changing frequency of peak flows caused by harvesting and as a result strips a critical part of the physics from the science of forests and floods. Forests, forest harvesting or deforestation alters processes generating floods in ways that change not only the magnitude of a flood but also its frequency. Due to the highly nonlinear linkage between these two attributes of a flood event they should be evaluated simultaneously. This is best done via the framework of a peak flow frequency distribution and not the analysis of variance and covariance, as traditionally conducted in decades of forest hydrology literature. The two inextricably linked questions that should guide a research investigation of the forests and floods relation are: What is the change in magnitude (frequency) for a flood event with a frequency (magnitude) of interest? Such investigation requires a frequency-paired event analysis, where the watershed's hydrologic response of interest is the entire flood frequency distribution. We combine fieldwork with numerical modelling to re-evaluate our scientific perception of the forests and flood relation using the new experimental design of pairing flood events by equal frequency. The analysis is being conducted across a wide range of watershed sizes, hydro-climate regimes, and silvicultural systems.