It just came to my full realization that the Chinese authorities are actually trying to release the water and remove the landslide debris/dam simultaneously at the Tangjiashan "quake lake." Apparently, keeping the dam in place was not an option. The dam removal and water release operation will start within several days!
A drainage/sluice channel is being dug through top of the dam, and is almost completed (see photo to the left). Once the water starts to flow in the channel, water current will scour the channel bottom and the channel sides. The channel will become wider and deeper while water level in the "quake lake" will be lower and lower. The channel bottom and sides will scour because of steep slope of the channel being dug.
If it works perfectly, a good mixture of debris/soil and water will be flowing gradually downstream, water in the lake will be draining down slowly, and the dam itself will disappear steadily. In this scenario, the flooding downstream will be minimal, although some debris/soil will deposit along the way.
However, if scour of the channel is not gradual and initial release of the water triggers a catastrophic collapse of the dam, entire volume of the lake water will rush downstream rapidly, leading to devastating downstream flooding and destruction.
Therefore, this daring engineering has to be designed and executed extremely carefully. A detailed hydrodynamic and sediment transport analysis (including computer modeling and physical modeling which Chinese engineers are fully capable of conducting) would help properly design and construct the channel. A reinforcement at downstream end of the channel and/or diversion of the released water far away from toe of the dam may also be necessary.
The best wishes to the Chinese endeavors!!!
The channel dimensions are: Channel length = 300 m, bottom width = 13 m, top width = 50 m, inlet bottom elevation = 742 m, outlet elevation = 740 m (i.e., an elevation drop of 2 m over the length of 300 m, a steep slope that will definitely result in soil scour once the water flow starts), upstream depth = 8 m, downstream = 12 m. Total volume of the debris/dam = 20,370,000 cubic meters. At 8:00 a.m. of May 29, 2008 (Beijing time), lake water level = 730.13 m (rising), water depth = 62.58 m, lake water volume = 161,000,000 cubic meters.
Three downstream evacuation plans were planned: 1/3 of the dam opened up instantly, 1/2 of the dam opened up instantly, and the entire dam opened up instantly (the catastrophic collapse). Nearly 200,000 downstream people are being evacuated based on the scenario of 1/3 dam collapse. More than one million people are on alert, and they would have to be evacuated if the total collapse, in addition to inconceivable property and natural environment damages. A lot is at stake, and it is obviously highly desirable to have a well engineered slow dam removal and water release.
The Chinese Short Version:
从媒体报道得知，唐家山堰塞湖的泄洪槽(导流槽)已基本上修完。 湖水面在近日就会上升至槽（渠）底开始排水。
因为新修的渠比较陡 （300 米距离内底部高程降2米），放水一开始，渠道就会冲刷。我想这是中国工程师预期的效果，既边排放湖水，边冲走堰塞土堆/坝。
我相信中国的水利专家已经为决策高层设计了最佳的渠道高程，断面，和坡度。修造最佳的渠道可以达到缓慢放水，平稳冲涮（除坝）的目的。最佳设计可以先通过水动力学与泥沙传输原理简单计算得出，然后最好再用复杂的数学模型及/或大型物理模型验证。万一这一步还没做，需马上请国内专家算， 以作渠道调整。我在美国也可以帮
算，但需要土质资料。
我还想提醒中国工程师的是要用当地石块加固坝的下游底面和河床，以避免坝趾掏空而溃堤。
祝唐家山堰塞湖排水成功！
Photo credit and info source: CCTV.com
By 22:00 May 31 (Beijing time), the entire channel was finalized and completely constructed (see photo to the left for a view of the channel inlet segment). The final channel dimensions are: Total length = 475 m; bottom elevation of inlet segment = 740 m, bottom width > 7 m; bottom elvation of the middel segment = 740 m, bottom width > 7 m; bottom elevation of outlet segment = 739 m, bottom width > 10 m. 
