نوع مقاله : مقاله پژوهشی
2 استادیار، گروه مهندسی عمران، دانشگاه شهید باهنر، کرمان، ایران،
3 دانشگاه آزاد اسلامی، واحد رودسر و املش، گروه مهندسی عمران، رودسر، ایران،
عنوان مقاله [English]
Inattention to flood frequency analysis (FFA) can cause irreparable damages. Thus, increasing accuracy in this area has been considered by various researchers. In this paper, data observed over a period of 52 years in Polrood River are used in frequency analysis. FFA has been performed using two different wavelet functions (Haar and daubechies3) and results are presented. In two stages, data less than 80% and 95% of the smallest annual maximum (AM) data are removed to reduce the impact of very small data in wavelet model. Then FFA using partial duration (PD) series and AM series has been performed with fitting different statistical distributions to data and results are compared with wavelet analysis. Results comparison between wavelet, AM and PD methods indicates this very important point that the accuracy of analysis and the selection of method are mostly based on the nature of data and study area. In Polrood basin, AM series method overestimates the results because just 0.06 percent of data are more than 200 m3/sec and 10 data are between 200 to 537 m3/sec.