【摘要】 历史洪水资料起着延长经验频率曲线,减少成果误差的作用。处理历史洪水的关键,在于正确审定其流量数值及重现期.后者可根据历史文献,考证更远时期中洪水发生的情况来估定。实测系列的代表性,对曲线的外延也有影响.可与邻近的长期站的观测资料,进行比较分析。应用本逊(M.A.Benson)公式 P′=(a+[(N-a)/n]m)/(N+1)来调整实测期洪水的频率,具有很大偶然性.因 P′恒大于 P=m/(n+1).而实际上,P′的真值也可小于 P.不连续系列统计参数的计算,可以按 P=M/(N+1)在频率线上取样,并分组计算。也可以历史洪水及实测洪水的各自机率为权数,进行加权计算.应用克里茨基-明克里公式时,需注意 n 年系列的代表性,及 N,a,n 间应有一定的关系.更多还原

【Abstract】 Use of historical flood data makes it possible to extend more accurately the fre- quency curve beyond the observed portion.Careful examination of discharges and return periods is the key to the treatment of historical floods.For return periods,reasonable answer may be obtained by studying the historical literature. The representativeness of the observed series in a longer periodmay also affect either the extension of the frequency curve or the computed results of the statistical parameters. Comparing the observed series with the longer records of adjacent stations may throw light on the answer.Adjusting the frequency of observed data by Benson’s formula P′=（a+[（N-a）/n]m）/（N+1）gives uncertain results.Actually the true P may lie on either side of P=m/（n+1）,while Benson’s formula always gives a larger P′.It is also not true even when the P of the smallest observation in n years is changed from n/（n+1）to N/（N+1）. The statistical parameter of the discontinuous series may be computed by the tradi- tional method Q₀′=sum from 1 to N Q_i′/N after a new series（Q₁,Q₂,…,Q_M,…,Q_N）has been picked out from the cmperical frequency curve by making the frequency of each item follows the order of P=M/（N+1）.The parameters can also be computed by the general formula Q₀′=ΣQ△p/Σ△P.While using the formula of Kritski and Menkel,atten- tion should be paid in examining the representativeness of the recorded series and as to whether a reasonable relation between N,n,and a exists.更多还原