Authored by Qizhong (George) Guo
Original version: July 8, 2026; revised: July 25, 2026

Overview

This technical brief was prepared to support professional discussion of aging earth-dam safety, overtopping risk, the actual performance and reliability of flood-release facilities, adaptive reassessment of design floods and engineering standards under nonstationary climate conditions, reservoir operation during floods exceeding the design standard, and risk-based management of reservoir systems.

On July 6, 2026, Liulan Reservoir in Hengzhou, Nanning, Guangxi, China, experienced overtopping, local breach initiation, and rapid breach enlargement during Typhoon “Maysak”-related extreme rainfall and rapidly increasing reservoir inflow. On July 24, the State Council established an investigation and assessment team for the Liulan Reservoir “7·6” dam failure disaster. This brief incorporates publicly available information through July 25, including official announcements, later expert hydrologic interpretations, more detailed reporting on reservoir operations and flood-release facilities, pre- and post-event imagery, and professional discussions. It supersedes the initial version published on July 8.

The combined evidence indicates that extreme hydrologic loading was the primary external driver of the event. Available information supports the view that erosion of the homogeneous earthfill dam under sustained overtopping was likely one of the key mechanisms responsible for local breach initiation and rapid enlargement. The event, however, should not be attributed to a single factor before completion of the formal investigation.

Public information indicates that staged gate releases began on the evening of July 5. When operators planned to fully open all five gates early on July 6, some gates and hoisting equipment did not reach their intended fully open positions. The fuse-plug embankment at the entrance to the emergency spillway on the east side did not breach as intended and could not be activated manually in time; its precise design conditions and the reasons it did not activate remain to be determined. Reservoir operations were also constrained by downstream safe-release considerations, evacuation progress, communications interruptions, and incomplete real-time hydrologic information.

The available evidence more strongly supports the following provisional sequence: rapidly increasing inflow caused the reservoir water level to reach the dam crest and overtopping to begin; concentrated erosion and an initial breach developed at or shortly after the east abutment; and a larger breach subsequently formed and expanded in the central portion of the dam. It is not yet possible to determine precisely whether a clearly distinguishable time interval existed between the beginning of overtopping and initial breaching near the east abutment.

The brief examines rainfall and inflow reconstruction, pre-event reservoir storage, the operating timeline, actual gate performance, the service and emergency spillways, monitoring and communications, downstream warning and evacuation, breach development and flood routing, sediment and water-quality effects, and post-disaster functional recovery. It also considers possible local flow concentration associated with reservoir and dam geometry, while treating this as a hypothesis requiring hydrodynamic modeling and field verification rather than as an established cause.

Beyond the immediate event, the brief emphasizes two complementary priorities. First, design floods, effective discharge capacity, dam-crest freeboard, overtopping protection, and related engineering standards should be adaptively reassessed using updated hydrologic, climate, land-use, and downstream-exposure information. Second, because no design standard can eliminate all risk, reservoir owners and public agencies must also manage the residual risk associated with floods exceeding the design standard, including equipment failure, nonactivation of emergency facilities, communications loss, evacuation difficulties, and simultaneous emergencies at multiple reservoirs.

The brief further proposes evaluating the “effective age” of the entire dam system rather than relying on the year of construction alone. Embankment materials, foundations, spillways, gates and hoisting equipment, monitoring systems, the hydrologic design basis, maintenance practices, and institutional arrangements may deteriorate or become inadequate at different rates. Dam safety should therefore be assessed according to the present condition, functional reliability, and remaining safety margin of the whole system.

This technical review is based on public information that could be obtained and cross-checked through July 25, 2026. Original hydrologic records, gate-operation logs, design documents, monitoring data, and complete field-investigation findings have not yet been publicly released. The numerical values and mechanistic interpretations presented in the brief therefore remain provisional and should ultimately be confirmed, corrected, or superseded by the formal investigation.

The English and Chinese technical review files each include the revised version followed by the original version and are available for download below.

Download:

English PDF Version of the Technical Brief →
技术简报中文 PDF 版本 →

广西横州六蓝水库“7·6”溃坝灾害技术复盘

作者：郭祺忠
初版：2026年7月8日；修订版：2026年7月25日

概述

本技术简报旨在支持有关老旧土坝安全、漫顶风险、泄洪设施实际性能与运行可靠性、非平稳气候条件下设计洪水与工程标准的适应性复核、超标准洪水下的水库运行调度，以及水库系统风险管理的专业讨论。

2026年7月6日，广西南宁市横州市六蓝水库在台风“美莎克”相关极端降雨和入库洪水快速增加的过程中发生漫顶、局部缺口形成及缺口迅速扩大。7月24日，国务院成立广西南宁横州市六蓝水库“7·6”溃坝灾害调查评估组。本版纳入了截至7月25日能够获得的公开资料，包括官方通报、后续专家水文分析、有关水库运行和泄洪设施的进一步报道、事件前后影像及专业讨论，并取代7月8日发布的初版简报。

现有综合资料表明，极端水文荷载是本次事件的主要外部驱动因素。持续漫顶水流对均质土坝的冲刷，很可能是局部缺口形成和迅速扩展的关键机制之一。不过，在正式调查完成前，仍不宜将事件简单归因于任何单一因素。

公开资料显示，水库于7月5日晚开始分阶段开闸泄洪。7月6日清晨计划将五孔闸门全部开启时，部分闸门及启闭设备未能达到预定的全开状态。位于大坝东侧非常溢洪道入口处的自溃式土埂也未发生预期自溃，并且未能通过人工方式及时启用；其具体设计启用条件及未能发挥作用的原因，仍有待正式调查。水库运行同时受到下游安全泄量、人员转移进度、通信中断以及实时水文信息不完整等因素的制约。

现有证据较为支持以下初步过程：快速增加的入库洪水推动库水位持续上涨，直至达到坝顶并发生漫顶；东侧坝肩附近在漫顶开始时或其后不久出现集中冲刷和初始缺口；随后坝体中部又形成较大的缺口并迅速扩展。目前尚无法准确判断东侧坝肩初始漫顶与局部缺口形成之间是否存在清晰可辨的时间间隔，也无法仅依据现有影像确定各处缺口形成的精确顺序和机制。

简报重点讨论了降雨与入库洪水过程重建、事发前水库蓄水状况、运行调度时间线、闸门实际运行状态、正常溢洪道与非常溢洪道的功能、监测与通信条件、下游预警和人员转移、缺口演变与洪水演进、泥沙和水质影响，以及灾后水库功能恢复等问题。简报还讨论了水库平面形态、坝体几何条件及东侧坝肩局部地形可能造成水流集中这一假设，但将其作为需要通过水动力模拟、现场测量和原始记录加以验证的研究问题，而不是已经确定的事故原因。

本次事件提出了两个相互补充的行业重点。第一，应根据更新后的水文、气候、土地利用和下游暴露资料，对设计洪水、有效泄洪能力、坝顶安全超高、防漫顶措施及相关工程标准开展适应性复核，并在必要时进行提升。第二，由于任何设计标准都不能完全消除风险，水库管理部门还必须系统管理超标准洪水下的剩余风险，包括泄洪设备故障、非常泄洪设施未能启用、监测和通信中断、人员转移困难以及多座水库同时出险等情形。

简报进一步提出，不应仅依据建成年份判断大坝安全状况，而应评估整个大坝系统的“有效年龄”。坝体材料、地基、溢洪道、闸门及启闭设备、监测系统、水文设计依据、维护管理和组织制度可能以不同速度老化或变得不再适用。因此，大坝安全评价应更加重视整个系统当前的实际状态、功能可靠性和剩余安全裕度。“有效年龄”在本文中是一种用于综合认识系统老化和功能退化的概念性框架，并非现行规范中的定量指标。

公开材料尚未显示2026年5月广西地震是本次缺口形成的直接原因。震后排查和汛前检查仍有必要，正式调查也应核查坝体变形、裂缝、渗流、坝肩接口和闸门启闭系统是否在地震后出现异常；但就目前证据而言，本次事件更明确的直接过程仍是极端降雨导致入库洪水迅速增加、库水位持续上涨并发生漫顶，随后漫顶水流冲刷坝体，促使局部缺口形成并迅速扩大。

本技术复盘基于截至2026年7月25日能够获得并进行交叉核对的公开资料。原始降雨和水位记录、完整闸门运行日志、工程设计文件、监测资料及系统性现场调查成果尚未全部公开。因此，简报中的部分数值、时间节点和机理分析仍属于阶段性判断，简报中的部分数值、时间节点，最终应由正式调查结果加以确认、修正或取代。本简报不替代政府部门的正式水文复核、大坝安全鉴定或溃坝灾害调查评估。

中英文技术复盘文件均包含修订版及附于其后的初版，可通过以下链接下载。

下载：

技术简报中文 PDF 版本 →
English PDF Version of the Technical Brief →