黑河流域非一致性极端高温频率特征分析

Non-stationary frequency analysis of extreme high temperature in the Heihe River Basin

气候变化影响下很多水文气象时间序列的一致性条件遭到破坏。以黑河流域9个气象站1960-2010年逐日气温资料为基础,选取年平均日最高气温和高温强度2个极端高温指数,采用多种统计检验方法和概率分布模型,探讨非一致性条件下研究区极端高温指数的频率特征。结果表明:黑河流域极端高温指数呈现显著的非一致性;基于还原途径对极端高温指数进行一致性修正后,通过对比修正前、后不同重现期水平下极端高温指数的估算值,可以发现,气候变化条件下黑河流域极端高温指数呈现强度增强、重现期缩短、发生频率增加的趋势,与中国西北地区极端气温总体变化趋势一致。

Studying the characteristics of extreme high temperature events under the changing environment is important for the mitigation and adaptation of climate change, as it provides theoretical basis for local disaster prevention and mitigation. How to quantify the non- stationary extreme high temperature and its changes has not well established so far. In this paper, two extreme high temperature indices, i.e. annual mean maximum temperature （AMMaxT） and high temperature intensity （HI）, are proposed to describe the extreme high temperature events in the Heihe River Basin. Daily temperature observations from 1960 to 2010 of nine meteorological stations in the Heihe River Basin are collected. Four different statistical tests methods （including Mann-Kendall test, Spearman rank correlation test, Rank sum test and Pettitt test） are employed to detect the non-stationary characteristics of the two extreme high temperature indices. Eight theoretical probability distribution models （including Bate, Gamma, GEV, GPD, Log-Logistic, Lognormal, Wakeby and Weibull） are used to fit the frequency characteristics of the two indices. Trend analysis and change point detection show that nearly all the nine stations have experienced significant trends and obvious change points in both AMMaxT and HI series, and the main variation type is change point variation. Since the theoretical probability distribution models are commonly used to fit the stationary series, the non-stationary AMMaxT and HI series in this study are modified to be stationary by means of the backward restore for consistency. All of the eight probability distribution models can give good fittings to the modified AMMaxT series, while only three of the eight models, i.e. the GEV, GPD and Wakeby models give satisfactory fittings to the modified HI series. According to the ranking of goodness of fit, the GEV and Wakeby models perform the best for both AMMaxT and HI series. Considering its wide applications in other related researches, the GEV model is finally selected as the optimum theoretical one for fitting the extreme high temperature indices in the study area. Based on the GEV model, we calculate the estimated return levels for both the modified and non-modified extreme series at different return periods, and assess the changes of the extreme series at three different return periods （i.e. 10-year, 20- year and 50-year）. Overall, the estimated return levels for non-modified extreme series are greater than those for the modified series. This means that the extreme high temperature indices in the study area present trends of increased intensity, shortened return period and increased frequency, which is consistent with the changes of temperature in Northwest China.