The effect of bathymetric filtering on nearshore process model results
The effect of bathymetric filtering on nearshore process model results
复制标题
测深滤波对近岸过程模型结果的影响
DOI：
10.1016/j.coastaleng.2008.10.010
复制
发表时间：
2009
影响因子：
4.4
通讯作者：
K. Holland
中科院分区：
文献类型：
--
作者：
N. Plant;K. Edwards;J. Kaihatu;J. Veeramony;L. Hsu;K. Holland
近岸波和流模型的结果显示出很强的灵敏度的分辨率的输入测深。在这一分析中，通过对高分辨率勘测数据应用平滑滤波器以产生若干测深网格面，从而改变了测深分辨率。我们表明，模型预测的波高和流量的变化在水深分辨率的敏感性有不同的特点。波高的预测是最敏感的分辨率与近岸沙洲的结构相关的跨岸变异。流量预测是最敏感的解决方案的中间尺度沿岸变化与突出的沙洲韵律。在沙洲更靠近海岸和更浅的情况下，流动敏感性增加。也许这些结果最令人惊讶的含义是，水深数据的插值和平滑可以优化不同的波和流模型。我们发现，观测和建模的流量和波高之间的误差预测模型模拟结果进行比较，逐步过滤测深的结果，从最高分辨率的模拟。因此，可以使用模型模拟来估计过度平滑或不充分采样所造成的损害。我们的结论是，量化预测误差的能力将有助于支持未来的数据同化工作，需要这些信息。
Nearshore wave and flow model results are shown to exhibit a strong sensitivity to the resolution of the input bathymetry. In this analysis, bathymetric resolution was varied by applying smoothing filters to high-resolution survey data to produce a number of bathymetric grid surfaces. We demonstrate that the sensitivity of model-predicted wave height and flow to variations in bathymetric resolution had different characteristics. Wave height predictions were most sensitive to resolution of cross-shore variability associated with the structure of nearshore sandbars. Flow predictions were most sensitive to the resolution of intermediate scale alongshore variability associated with the prominent sandbar rhythmicity. Flow sensitivity increased in cases where a sandbar was closer to shore and shallower. Perhaps the most surprising implication of these results is that the interpolation and smoothing of bathymetric data could be optimized differently for the wave and flow models. We show that errors between observed and modeled flow and wave heights are well predicted by comparing model simulation results using progressively filtered bathymetry to results from the highest resolution simulation. The damage done by over smoothing or inadequate sampling can therefore be estimated using model simulations. We conclude that the ability to quantify prediction errors will be useful for supporting future data assimilation efforts that require this information.