基于ARCH效应的黄河高寒区水资源预报分析

Hydrological forecast for the upper Yellow River high cold region based on conditional heteroskedasticity

黄河高寒区水资源受气候变化影响,自演化过程中表现出持续性、波动性等非平稳特征。水文条件异方差即ARCH效应的存在,增加了寒旱区水资源管理的风险和不确定性。本文以黄河源区出口控制站-吉迈水文站1958-2011年,共计54年水文资料为研究基础,分析水文条件异方差性存在特性、来源及对于区域水资源管理的影响;建立能够反映ARCH效应的水文过程预报模型,进一步提高预报精度、为区域水资源管理提供可靠基础。

Affected by climate changing,the hydrological time series in the upper Yellow River high cold region demonstrate persistent,volatility and other non-stationary features.Conditional heteroskedasticity(ARCH effect) exists in these hydrological time series,which leads to an increase in the risk and uncertainty of modern water resources management.In this paper,we use 54-years time series of monthly runoff data recorded at the Jimai hydrological station to study the characteristics and origin of such heteroscedasticity and its effect on regional water resources management.And we develop a hydrological process prediction model that can reduce the ARCH effects and improve prediction accuracy.This model would provide a reliable basis for regional water resources management.