Gully erosion is recognised as a major soil degradation process, with gully growth being a complex phenomenon that is controlled by different processes: incision, gully headcut retreat and gully widening, coupling at different spatial and temporal scales. While gully modelling research has paid considerable attention to headcut retreat processes and incision, our review of the available literature in Web of Science database shows that less attention has been paid to the modelling of gully widening processes. This might hamper progress on gully evolution studies or on gully control and restoration programmes. Our review has found seven models described in the indexed literature: three of them empirical, three based on processes, and a seventh one that can be classified as hybrid. The latter, despite being based on process-based equations, is simplified and requires empirical calibration for implementation. Our review has also noted a limited number of publications using these models, rather for validation or for case studies, with very few examples of models used outside the team that developed it. Of the four process-based models, two of them are the most comprehensive - BSTEM and GULTEM - both incorporating a combination of processes to predict bank failure based on soil, climate and vegetation parameters, coupled with others describing the effect of concentrated channel flow. A third one, the Bank Failure Model (hereafter BFM), is also embedded in the well know landscape evolution model CHILD. Of the seven models identified, only two of them, BSTEM and BFM through CHILD, are directly available to potential users with access to executable versions and instructions, while the others need to be requested from the authors or built from the published equations. The models cover a wide range of spatial and temporal scales, from event to centuries, and from gully stretch to regional scale. Our review suggests that process-based models are the ones providing the most comprehensive approach, especially those including geotechnical approaches, although they would benefit from further development to include wider spatial and temporal scales. This research suggests that more studies and field data are required for calibration, understanding lateral flows, wetting and drying cycles, tension crack prediction, and integrating widening with other gully erosion processes. They should also include further testing and validation of available models beyond the teams that developed them, ideally in model intercomparison exercises.

中文翻译：

---

沟壑加宽的建模，评论。对沟壑演化与恢复研究的启示


沟渠侵蚀被认为是一种主要的土壤退化过程，沟渠生长是一种复杂的现象，受到不同过程的控制：切割、沟渠头后退和沟渠加宽，以及不同时空尺度的耦合。虽然沟壑建模研究对头切后退过程和切口给予了相当多的关注，但我们对 Web of Science 数据库中现有文献的回顾表明，对沟壑加宽过程的建模关注较少。这可能会阻碍沟壑演化研究或沟壑控制和恢复计划的进展。我们的审查发现索引文献中描述了七个模型：其中三个是经验模型，三个基于过程，第七个模型可以归类为混合模型。后者尽管基于基于过程的方程，但被简化并且需要经验校准来实施。我们的评论还指出，使用这些模型的出版物数量有限，而不是用于验证或案例研究，在开发团队之外使用的模型示例很少。在四个基于过程的模型中，其中两个是最全面的 - BSTEM 和 GULTEM - 两者都结合了基于土壤、气候和植被参数来预测河岸溃决的过程组合，以及描述集中河道流量影响的其他模型。第三个模型是银行破产模型（以下简称 BFM），也嵌入到众所周知的景观演化模型 CHILD 中。在确定的七个模型中，只有两个模型（BSTEM 和 BFM through CHILD）可直接提供给有权访问可执行版本和指令的潜在用户，而其他模型则需要向作者请求或根据已发布的方程构建。 这些模型涵盖了广泛的空间和时间尺度，从事件到几个世纪，从沟壑延伸到区域尺度。我们的审查表明，基于过程的模型是提供最全面方法的模型，尤其是包括岩土工程方法的模型，尽管它们将受益于进一步发展以包括更广泛的空间和时间尺度。这项研究表明，需要更多的研究和现场数据来进行校准、了解侧流、湿润和干燥循环、张力裂缝预测以及将加宽与其他沟壑侵蚀过程相结合。他们还应该包括对开发团队之外的可用模型进行进一步的测试和验证，最好是在模型比较练习中。