Tectonically shattered subduction trench lithologies, rapid uplift, and intense monsoon and typhoon rains combine to produce rapid erosion rates that make Taiwan’s sediment yield to be among the highest in the world. Chijiawan Creek, a tributary of the Tachia River in central Taiwan, is the only habitat for the endangered, landlocked Formosan salmon. The habitat of the salmon has been reduced significantly since the 1960s when 12 check dams were built in Chijiawan Creek watershed to reduce sedimentation in the downstream Techi Reservoir. In 1999, Shei-Pa National Park began a river restoration project that removed four 4 m to 6 m dams in Kaoshan Creek, a tributary of Chijiawan Creek. While most of the post-dam-removal efforts were focused on fish monitoring, little was learned on how channel would respond under conditions of high sediment supply and high transport capacity. Twelve years later in May 2011, Shei-Pa National Park further removed the largest and lowermost check dam in Chijiawan Creek due to both safety and ecological concerns. The 13-m high dam was built in 1972. After nearly 40-years, the dam was backfilled with an estimated 0.2 million m3 of sediment and the dam foundation had been scoured about 4 m deep, raising the risk of dam failure. Prior to removal, we conducted field surveys, flume experiments and numerical simulations to suggest appropriate dam removal strategies. After removal, we repeated field surveys to monitor the channel responses over time.
The removal allowed built up sediment to be transported downstream. Following a 20-year return period discharge event in August 2012, the knickpoint migrated 800 m upstream of the dam, and about 75% of the estimated impounded sediment (150,000 m3) was transported downstream. In light of the high sediment yields and highly seasonal hydrology, long-term monitoring is particularly important to enhance our understanding of dam removal in Taiwan. We further proposed a channel evolution model and a decision-making procedure for dam removal in Taiwan based on documenting the channel responses in Chijiawan Creek, in an effort to draw lessons learned from these past efforts and discuss the potential of future practices.