Akademik Veri Yönetim Sistemi
9. INTERNATIONAL ANKARA MULTIDISCIPLINARY STUDIES CONGRESS, Ankara, Türkiye, 13 - 14 Kasım 2025, cilt.1, ss.905-912, (Tam Metin Bildiri)
The selection of appropriate probability distribution functions for modeling IDF (Intensity–Duration– Frequency) curves, which are crucial for determining standard-duration rainfall intensities and flood flow discharge and volume for engineering design projects such as infrastructure drainage. Kolmogorov–Smirnov, Chi-square, and Anderson–Darling statistical goodness-of-fit tests are used to identify the best models. An examination of the existing applications reveals that several probability distribution functions (PDFs) often provide acceptable fits to models at certain significance levels. So, the models that pass these tests are selected based on experience or engineering practice. This is a significant problem when differences between model results are at a level that affects designs. This study addresses the problem and evaluates the importance of objective criteria for its solution. In this research, IDF model curves were generated for five different PDFs (2- and 3-parameter Log Normal (LN-2, LN3), 3-parameter Gamma (G3P), 3-parameter Log-Pearson Type-III (LP3), and Gumbel (G)) using monthly precipitation data from 1957 to 2020 for Trabzon. The results of the extreme rainfall with fourteen different standard durations in the literature revealed significant differences among the models, particularly for long return periods. A difference of up to 28% was identified between the models in predicting the extreme rainfall with a 30-minute standard duration for a 500-year return period. Therefore, these results cause uncertainty in determining the appropriate model for engineering designs. In conclusion, the findings of this study clearly demonstrate the critical importance of selecting an appropriate probability distribution function model for ensuring safety and economic efficiency in the design of flood control structures and urban drainage infrastructure systems. Finally, selecting design models based solely on statistical fit tests introduces certain uncertainties. Therefore, to manage the uncertainties encountered in design studies, relevant model selections should be made in consideration of climatic conditions, hydrological realities, and engineering principles