摘要
水位变化直接关系鄱阳湖流域防汛抗旱的科学决策、水资源的高效利用、水生态健康以及湖区经济的可持续发展。利用鄱阳湖各站点的平均日降水、气温和水位资料,建立了三层反向传播神经网络模型(BP- NNM)。基于统计降尺度方法把全球气候模式BCC-CSM1-1输出结果降解到流域尺度的气温和降水数据,输入到BP-NNM得到未来气候变化情景下鄱阳湖水位系列,对未来鄱阳湖的平均水位进行分析。研究结果表明:在RCP4.5情节下,鄱阳湖未来平均水位的多年均值变幅小于2.5%,但是平均水位的年际变化在各月份之间差异较大。
The water level change directly relates to scientific decision-making of flood control and drought relief, efficient utilization of water resources, healthy of water ecology and sustainable development of economy in the Poyang Lake basin. Back Propagation Neural Networks Model (BP-NNM) was established based on observed daily temperature, precipitation and water level data series in this study region. The statistical downscale method was coupled with BCC-CSM1-1 global climate model to obtain basin scale air temperature and precipitation data series, and then as inputs of BP-NNM to simulate future water levels of the Poyang Lake. The future average water levels of the Poyang Lake were analyzed and discussed. The results show that the future annual mean water level change will less than 2.5% compared to the baseline period, but the inter-annual variability of water levels among months is large under RCP4.5 scenarios.
出处
《水资源研究》
2014年第6期532-538,共7页
Journal of Water Resources Research
