唐凤娇, 祁生文, 郭松峰, 等. 2022. 金沙江溪洛渡库区水库诱发滑坡时空分布规律及易发性研究[J]. 工程地质学报, 30(3): 609-620. doi: 10.13544/j.cnki.jeg.2022-0119.
    引用本文: 唐凤娇, 祁生文, 郭松峰, 等. 2022. 金沙江溪洛渡库区水库诱发滑坡时空分布规律及易发性研究[J]. 工程地质学报, 30(3): 609-620. doi: 10.13544/j.cnki.jeg.2022-0119.
    Tang Fengjiao, Qi Shengwen, Guo Songfeng, et al. 2022. Spatio-temporal distribution pattern and susceptibility of reservoir-induced landslides in Xiluodu Hydropower Station[J]. Journal of Engineering Geology, 30(3): 609-620. doi: 10.13544/j.cnki.jeg.2022-0119.
    Citation: Tang Fengjiao, Qi Shengwen, Guo Songfeng, et al. 2022. Spatio-temporal distribution pattern and susceptibility of reservoir-induced landslides in Xiluodu Hydropower Station[J]. Journal of Engineering Geology, 30(3): 609-620. doi: 10.13544/j.cnki.jeg.2022-0119.

    金沙江溪洛渡库区水库诱发滑坡时空分布规律及易发性研究

    SPATIO-TEMPORAL DISTRIBUTION PATTERN AND SUSCEPTIBILITY OF RESERVOIR-INDUCED LANDSLIDES IN XILUODU HYDROPOWER STATION

    • 摘要: 水库诱发滑坡作为重大工程对地质环境影响的一种重要形式,是国内外工程地质学科研究的前沿和热点。金沙江流域地处青藏高原东缘高山峡谷区,地质环境脆弱,水能资源丰富,是世界上水电站建设最密集的地区之一,规划了25级梯级电站,但是水库蓄水对岸坡改造的时空效应尚不清楚。本文以溪洛渡水电站为例,对金沙江流域水库诱发滑坡的分布规律进行研究,利用2013~2020年多期遥感影像解译溪洛渡库区范围内的水库诱发滑坡,共解译滑坡433处。在此基础上,对水库诱发滑坡数量和面积随蓄水时间的变化趋势进行了分析,随后利用频率比法对水库诱发滑坡的分布与高程、坡度、坡向、工程地质岩组、断裂、距死水位距离6个因素的关系进行了统计分析,同时进行了各个单因子的滑坡易发性评价,利用曲线下面积AUC法验证了评价结果的可靠性,并基于评价结果选取了高程、距断裂距离、坡度、距死水位距离4个因素进行了水库诱发滑坡易发性评价。研究认为:(1)水库诱发滑坡主要发生在蓄水初期3~4年内,之后滑坡数量和面积逐渐减少,岸坡表生演化逐渐趋于稳定。(2)库区内水库诱发滑坡主要分布在高程1 km以内,2 km以上无诱发滑坡分布,优势坡度为30°~60°,发育滑坡的斜坡坡向以SE、W和NW向为主;在距断裂400~3200 m范围内更有利于滑坡发育;距死水位100~200 m范围内灾害发育数量最多。(3)易发性评价验证AUC高达0.912,评价结果可信度较好。(4)水库诱发滑坡的主控因素为距死水位距离和高程,极高易发区与高易发区主要分布在距死水位400 m以内、高程1 km以下的范围内。本文首次利用多期遥感影像建立了溪洛渡水电站水库诱发滑坡数据库,研究结果能够为已建水电站正常运营、未建及在建水电站的规划建设和防灾减灾提供借鉴。

       

      Abstract: The reservoir-induced landslide is an important form of the impact of major engineering on geological environment and one of the frontier hot spots of research in engineering geology disciplines. Jinsha River Basin is located in the alpine gorge region at the eastern margin of the Tibetan Plateau, and has fragile geological environment and abundant hydro energy resources. It is one of the most densely built areas for hydropower plants and has 25 levels of cascade hydropower stations planned. However, the spatio-temporal effect of reservoir storage on reservoir bank reconstruction is still unclear. We took Xiluodu hydropower station as an example and study the distribution pattern of reservoir-induced landslides in the Jinsha River Basin. We use the multi-period remote sensing images from 2013 to 2020 to interpret the reservoir-induced landslides within the Xiluodu reservoir area. We finally acquire a total of 433 landslides. On this basis, we analyse the trend of the number and area of reservoir-induced landslides with impounding time. Subsequently, we statistically analyse the distribution of reservoir-induced landslides in relation to using the frequency ratio method. The six factors are elevation, slope, slope direction, engineering geological rock group, fracture and distance from dead water level. The assessment on susceptibility of each single factor was also carried out, and the reliability of the assessment results was verified by using the AUC method. And based on the evaluation results, four factors were selected to evaluate the susceptibility of reservoir-induced landslides. The four factors are elevation, distance from the fracture, slope and distance from the dead water level. The study concludes that:(1)Reservoir-induced landslides mainly occur during the initial 3-4 years of impounding, after which the number and area of landslides gradually decrease and the bank slope exogenetic reconstruction gradually stabilizes. (2)The range of slope and aspect that are liable to induce landslides are 30°~60°and SE, W, NW direction respectively. Landslide development is more favorable in the range of 400~3200 m from the fracture and 100~200 m from the reservoir area. (3)The AUC is 0.912, which proved that the evaluation result is highly reliable. (4)The main control factors of reservoir-induced landslide are the elevation and distance from the dead water level. The extremely high susceptibility area and high susceptibility area are mainly located within 400 m from the reservoir area and below 1 km in elevation. In this paper, a database of reservoir-induced landslides is established for the first time using multi-period remote sensing images, and the research results can provide support for the planning and construction of hydropower plants and disaster prevention and mitigation.

       

    /

    返回文章
    返回