Study On The Influence Of Sample Size On Flood Pseak Frequency Calculation

Hydrological frequency calculation is an important content in hydrology research.Its main purpose is to reveal the relationship between the value of hydrological events and their occurrence frequency by using probability distribution,so as to provide basis for water conservancy project planning and optimal allocation of water resources.Sufficient samples and suitable probability distribution types are the necessary premise to ensure the accuracy of hydrological frequency calculation results.In order to better reveal the influence of sample size on the selection of probability distribution types and parameter estimation in hydrological frequency calculation,this paper takes the flood peak frequency calculation as the research object,and takes the Tangwang River and Hulan River Basins in Heilongjiang Province as the typical areas to carry out the flood peak sample size variation The main research contents and conclusions are as follows:(1)This paper reviews the literature content of hydrological frequency research in recent years,reviews the existing methods of hydrological frequency calculation,introduces the generalized extreme value distribution,Pearson three curve distribution,the parameter statistics method of moment method,and estimates the probability distribution parameters based on the moment method,and introduces the sum of the cumulative error square(δ),OLS criterion and AIC information criterion are used to evaluate the fit degree。(2)Based on the statistical calculation of the mean value and histogram of the annual maximum flood peak of 15 hydrological stations in typical area with different sample sizes,the histogram and mean value of the annual maximum flood peak with different sample sizes and the cumulative probability of the maximum flood peak with different sample sizes are compared and analyzed.It is found that when the sample size is more than 30,the annual maximum flood peak with different sample sizes can be obtained The similarity degree of statistical histogram of flood peak volume is the closest.The mean value of annual maximum flood peak volume of different sample sizes and the cumulative probability of maximum flood peak in different sampling intervals no longer fluctuate significantly with the increase of sample size.That is to say,when the annual maximum flood peak sample of each station reaches 30,its statistical characteristics tend to be stable。(3)The generalized extreme value distribution and Pearson type three curve distribution are used as the probability distribution types of annual flood peak.The moment method is used to estimate the initial values of parameters and optimize them.The theoretical frequency curve of annual maximum flood peak is determined,and the cumulative error square sum criterion is adopted(δ)The results show that the fitting effects of the two theoretical frequency distribution curves are similar,which indicates that the generalized extreme value distribution and Pearson distribution are suitable for the calculation of the annual maximum flood peak frequency in the study area。(4)In order to study the influence of sample size on parameter estimation of generalized extreme value distribution and Pearson three distribution,the annual maximum flood peak samples of each station are calculated from the 25 th year,and then increase year by year until the latest year of the station.The results show that the skewness coefficient(CS)changes greatly,while the coefficient of variation(CV)and expected value(Ex)change little.In order to further reveal the influence of different starting and ending time on skewness coefficient(CS),variation coefficient(CV)and expected value(Ex)under different sample sizes,equal length sliding sampling was carried out under different sample sizes,and the corresponding change curves of various parameters were drawn.After analysis,it is found that when the sample size reaches about 30,the corresponding skewness coefficient(CS)of different start and end time nodes fluctuates significantly,while the variation coefficient(CV)and expected value(Ex)fluctuate slightly and tend to be stable。...