Use of genetic programming and artificial colony optimisation techniques to establish severity-duration-frequency relationships in the Kizilirmak basin

Water is the most important natural resource indispensable for the life of all living things. However, the process of deterioration of the natural functioning of the atmosphere after the industrial revolution has increased even more after the 19th century and its effects are felt today. Changes in water resources, which are one of the most important consequences of global warming, have reached dimensions that threaten sustainable life as well as environmental impacts. This negative situation increases the importance of water and water resources even more. With the negative impact of global warming on water resources, the frequency and severity of many natural disasters such as floods, forest fires and drought have increased. In terms of protection of ecological balance and sustainability, water resources should be used in the most rational way. Knowing the characteristics of precipitation, which is the main source of water resources, such as amount, duration, intensity, spatial and temporal variation is important for planning, design, construction and operation studies of different sectors such as water resources, agriculture, urbanisation, drainage, flood control and transportation. In this context, in order to reduce the negative effects of the pressure on water resources due to global warming and increasing population, effective utilisation of water resources is mandatory. In this study, it is aimed to use Genetic Algorithm (GA) and Artificial Bee Colony (ABC) optimisation methods to obtain intensity-duration-frequency (IDF) relations which play a key role in the design of hydraulic structures. For this purpose, the maximum rainfall amounts of six rainfall stations (Çorum, Kayseri, Nevşehir, Sinop, Sivas) in the Kızılırmak basin for 30 minutes and 1, 2, 3, 6, 12,18 and 24 hours were used as material. Moments, maximum likelihood and L-moments approaches were used for parameter estimation of the theoretical probability distributions that the maximum rainfall series most closely follow. In the study, the weight factors of 11 IDF relationships were estimated based on GA and ABC. Among these relationships, the IDF relationship that gives the closest result to the observed values was determined for each station. In addition, a new IDF relationship was also created by using the rainfall intensity values calculated for each station by Genetic Programming (GP) method. In terms of the difficulty and success of obtaining IDF relationships traditionally, it is possible to make easier and more reliable predictions with the use of optimisation methods.