Author(s): Tung Chiung Chang; Yu Ting Chen
Linked Author(s):
Keywords: Step-pool system; Development model; Neural network
Abstract: Step-pool systems present the significantly ecological role and can enlarge flow resistance and river-bed anti-scouring capability to stabilize river-bed and bank slope. Cobble stone-piled steps and fine sediment-deposited silt seam at the pools in buffer stagnant areas form the habitat suitable for diverse organisms. Mountain streams in Kaohsiung City are selected for the fieldwork investigation, measuring the river length, average slope, catchment area, and shape factor. The research results show that most developed step-pools, about 86% , present the length of 1~2 km, while undeveloped step-pools appear the length of 0~5 km, with wider range, where the length of 1~2 km is about 22% . The slope of most developed step-pools is in 0~20 degrees, about 66% , while it of undeveloped step-pools shows in 0~25 degrees, about 82% . The watershed area of most developed step-pools reveals in 0~2 km2, about 86% , while it of undeveloped step-pools is about 72. 5% . The shape factor of developed step-pools mostly appears in 0. 2~0. 3, about 58% , while it of undeveloped step-pools shows in 0. 1~0. 3, about 61% . In consideration of the characteristics of step-pools, the development model is proposed for the prediction with artificial neural networks. The correctness rate achieves 0. 79% , proving that ANN is a suitable prediction model of step-pool development.
Year: 2013