Long-term modelling with XBeach: combining stationary and surfbeat mode in an integrated approach
Long-term modelling with XBeach: combining stationary and surfbeat mode in an integrated approach
复制标题
使用 XBeach 进行长期建模：以集成方法结合静止模式和冲浪模式
DOI：
10.22237/jmasm/1556669160
复制
发表时间：
2017
影响因子：
--
通讯作者：
L. Bart
中科院分区：
文献类型：
--
作者：
L. Bart
XBeach 是一种基于过程的形态模型，已用于对海滩和沙丘的短期行为进行建模。在过去的几年中，研究人员和工程师对使用 XBeach 进行更长的模拟周期（多年）表现出了兴趣。 XBeach 有多个模块：静止模式通常在平静条件下使用，冲浪模式通常在暴风雨条件下使用。在本研究中，研究了与单独模型相比，两种模式的耦合是否可以提高长期模型的性能，同时关注跨岸过程。还研究了（准）二维模型的附加值。荷兰沿岸统一海滩 Vlugtenburg 的高频测深数据被用于创建为期 1 年的一维模型。为了比较奇异模型和耦合模型之间的差异，首先通过优化不对称性、偏度和地下水流的设置来分别研究静止模型和海浪模型的潜在性能。结果发现，默认设置下的静止模型和冲浪模型的技能为负。侵蚀被高估（尤其是在冲浪模式下），并且模型没有显示季节性影响。事实证明，不对称和偏度的设置是提高模型性能的有效措施，并且能够引入季节效应。优化的固定模型能够很好地预测剖面的发展。地下水流模块对模型稳定性的影响并不像 Zimmermann 等人的研究中发现的那样大。 （2015）和彭德（2013）。为了验证静止模式和海浪拍打模式之间的耦合是否有利于模型性能，研究了静止模式是否能够恢复海浪拍打模型的沙丘和海滩侵蚀。因此，在一年的前 100 天（在更极端的条件下）运行 surfbeat，在一年中的剩余时间（在更温和的条件下）运行固定模式。静止模式似乎无法恢复被侵蚀的海浪轮廓。耦合模型的性能优于 surfbeat 模型，但在大多数标准的长期建模中，发现平稳模式的性能优于耦合模型。人们发现，XBeach 的缺陷在于无法将潮间带（及其上方）的沉积物进一步输送到海滩和沙丘。静止模式下沙丘的侵蚀量与陆上定向风沙输送量处于同一数量级。因此，预计通过将 XBeach 与风模型耦合可以改善长期模型的结果。由于次重力波在二维中的传播，准二维模型被发现比一维模型具有更好的性能；在单独模型和耦合模型中，侵蚀量都没有像一维模型那样被高估。
XBeach is a process-based morphological model that has been used for modelling short term behaviour of beaches and dunes. In the past years interest has been shown by researchers and engineers to use XBeach for longer simulation periods (multiple years). XBeach has multiple modules: stationary mode is often used during calm conditions and surfbeat mode is often used during storm conditions. In this study was investigated whether the coupling of the two modes can increase the performance of long-term models compared to the separate models, while focusing on the cross-shore processes. Also the added value of (quasi-)2D models was studied. The high frequency bathymetryic data of Vlugtenburg, an alongshore uniform beach in The Netherlands, was used for creating 1D models for a period of 1 year. To compare the difference between singular and coupled models, first the potential performance of the stationary and surfbeat modules was investigated separately by optimising the settings for asymmetry, skewness and groundwater flow. It was found that the skill of the stationary and surfbeat models with default settings was negative. Erosion was overestimated (especially in surfbeat mode) and the models showed no seasonal effects. Settings for asymmetry and skewness proved to be effective measures to improve the model performance and were able to introduce seasonal effects. The optimised stationary model was able to predict the profile development reasonably well. The groundwater flow module did not affect the model stability as much as was found in the studies of Zimmermann et al. (2015) and Pender (2013). In order to validate whether a coupling between stationary and surfbeat mode would benefit the model performance, it was investigated whether stationary mode was able to restore the dune and beach erosion of the surfbeat model. Therefore surfbeat was run during the first 100 days of the year (during more extreme conditions) and stationary mode during the remainder of the year (more mild conditions). Stationary mode appeared not to be able to restore the eroded surfbeat profile. The performance of the coupled model was better than the surfbeat model, but the stationary mode was found to perform better than the coupled model for long-term modelling on the most criteria. It was found that a deficiency in XBeach is the ability to transport sediment from the intertidal zone (and just above) further up the beach and dunes. The erosion volumes in the dunes in stationary mode were in the same order of magnitude as the onshore directed aeolian transport. Therefore is expected that the results of long-term models can be improved by coupling XBeach with a wind model. The quasi-2D model was found to have a better performance than 1D models due to the spreading of infra-gravity waves in 2D; in both the separate and coupled models the erosion volumes were not overestimated as much as in the 1D models.