论文部分内容阅读
我们应用层次贝叶斯模型模拟大气二氧化碳浓度加倍可能对美国北卡罗来纳州西部威塔(Coweeta)流域水文的影响。这个模型整合了多重数据来源并且同时考虑了数据,参数和模型结构的不确定性。贝叶斯分析的预测分布显示流量和土壤含水量在秋季和夏季将明显下降,这将造成这两个季节更严重的干旱。同时我们用通用极值分布(Generalized Extreme Value distribution)和通用普拉托分布(Generalized Pareto distribution)分析预测流量,结果显示洪水频率也会增减。层次贝叶斯模型,和许多只能得到最佳参数估计的水文模型相比,能提供更丰富的信息,包括预测的不确定性。这将有助于可持续水资源管理的大前提下发展应对气候变化的措施。
We apply a hierarchical Bayesian model to simulate the effects of double atmospheric concentrations of atmospheric carbon dioxide on the hydrology of the Coweeta basin in western North Carolina. This model incorporates multiple data sources and takes into account the uncertainty of the data, parameters and model structure. The predicted distribution of the Bayesian analysis shows that the flow and soil moisture content will decrease significantly in autumn and summer, which will result in more severe drought in both seasons. At the same time, we use the Generalized Extreme Value distribution and the Generalized Pareto distribution to analyze the predicted flows, and the results show that the flood frequency also increases and decreases. Hierarchical Bayesian models provide richer information, including the uncertainty of forecasts, than many hydrological models that receive only the best estimates of the parameters. This will help to develop measures to combat climate change on the premise of sustainable water management.