Coastal Storms and Shoreline Change : Signal or Noise ?
Coastal Storms and Shoreline Change : Signal or Noise ?
复制标题
沿海风暴和海岸线变化：信号还是噪音？
DOI：
--
复制
发表时间：
2012
期刊：
影响因子：
--
通讯作者：
M. Honeycutt
中科院分区：
文献类型：
--
作者：
B. Douglas;M. Honeycutt
本答复的主题文件的结论是（第718页）“.在排除风暴影响的数据的情况下，海岸线预测的不确定性范围更大。与包含此类数据的案件相比并根据这一结果得出了关于海岸线行为的其他结论。我们不同意这些结论，并将在下文中表明，报告的调查结果来自对调查中使用的一系列海岸线位置的有缺陷的分析。FENSTER、DOLAN和MORTON（2001年）论文（以下简称FDM）中的结果基于对北卡罗来纳州哈特拉斯岛海岸线位置数据系列的分析，他们称之为断面21-13，如图3A和3B所示。第一幅图旨在显示线性回归和误差分析中使用的非风暴影响数据，第二幅图使用所有风暴和非风暴影响数据进行类似的样带分析。比较2010年的预测误差（95%CI），从图中表明，子设置数据，以消除风暴影响的海岸线位置，以确定海岸线位置的潜在长期趋势，导致更差的结果比，如果使用所有的数据。相反，我们对他们的数据的分析表明，相反的情况是正确的（即，消除受风暴影响的海岸线会得到更好的结果），FDM得出结论是因为其将海岸线位置分类为受风暴或不受风暴影响的限制性和不正确的标准。我们开始注意到，作者使用的12个海岸线位置的数据集只是表1所示的17个日期的一个子集。此外，桌子的标签也贴错了。它的标题开始“受风暴影响的海岸线位置和日期......“，但没有给出位置信息，当联系作者时，作者也无法及时提供数据供我们重新分析。幸运的是，扫描图3B并使用数字图像处理专业人员是一件容易的事情。
The subject paper of this reply concludes that (p.718) "... the range of uncertainty for shoreline predictions is greater in cases excluding storm-influenced data ... compared to cases including such data ..." and goes on to draw other conclusions concerning shoreline behavior based on this claimed result. We disagree with these conclusions, and will show below that the reported findings arise from a flawed analysis of the series of shoreline positions used in the investigation. The results in the paper by FENSTER, DOLAN, and MORTON (2001) (hereinafter referred to as FDM) are based on an analysis of a shoreline position data series at Hatteras Island, NC, called by them transect 21-13, and shown in their Figures 3A and 3B. The first of these figures purports to show only nonstorm-influenced data used in a linear regression and error analysis, and the second uses all stormand nonstorm-influenced data for a similar analysis of the transect. Comparison of predicted errors (95% CI) in 2010 from the Figures indicates that sub setting the data to eliminate storminfluenced shoreline positions for determining the underlying long-term trend of shoreline position leads to poorer results than if all of the data are used. In contrast, our analysis of their data shows that the contrary is true (i.e., eliminating storm-influenced shorelines leads to better results) and that FDM reached their conclusion because of their restrictive and incorrect criterion for categorizing shoreline positions as stormor nonstorm-influenced. We begin by noting that the data set of 12 shoreline positions used by the authors is only a subset of the 17 dates shown in Table 1. In addition, the table is mislabeled. Its caption begins "Storm-influenced shoreline positions and dates ...", but no position information is given, nor when contacted were the authors able to supply the data in a timely manner for our reanalysis. Fortunately, it was an easy matter to scan Figure 3B, and use a digital image-processing pro-