جستجوی مقالات مرتبط با کلیدواژه « WG model » در نشریات گروه « مهندسی آب »

Due to global warming and the occurrence of climate change and its effect on meteorological variables, especially rainfall, which is a key parameter in urban infrastructure design and water resources management, studies in this area are very important. It is now possible to predict future rainfall using the output of climate change models; this is however accompanied by many problems including lack of easy access, being large scale and the need for spatial and temporal micro scaling, increasing error, and imposing risk to urban flood management structures. Therefore, the application of a weather generator model with the ability of climatic models in simulating future precipitation using observational and recorded data has been considered in this research. In the present study, the non-parametric K-NN WG model was used which produces data outside the observational range by using two mechanisms of shuffling and perturbing. In this model, observational short-term precipitation data of Mehrabad station with a time duration of 10 minutes to 3 hours were used to predict short-term precipitation in the future. The results were compared with the output of the PRECIS regional climate model for the years 2070-2100 under emission scenarios A2 and B2. The results showed an increase in the intensity of future precipitation compared to the present time and moreover proved the capability of the WG model used in the study compared to the output of the PRECIS regional model. These results were evaluated using appropriate statistical tests.

The main objective of this research is evaluating IDF curves variation in the future using the observed rainfall and the forecasted one by PRECIS, while focused on yielding the uncertainties of the forecast via employing the K-NN WG model under two historical and wet scenarios. The historical climate change scenario and the B2 emission scenario were used to generate the lower and upper-reliability bounds of the forecast, respectively. By applying the shuffling and perturbation mechanisms, those random data, which are not recorded in the observing period but will be probably happened in the future, were generated and the results were employed for developing the IDF curves of 2, 5, 10, 20, 50, and 100-year return periods. Comparing the two climate change scenarios and the base period IDFs reveal the increase in the intensity of the extreme short-duration rainfalls. Moreover, it was found that the maximum amount of rainfall for the duration of 10 minutes to 3 hours on average would have a difference of 26 to 31 percent under the historical and wet scenarios concerning the base period, respectively. This difference between the historical and wet scenario is limited to 4.8%. The small difference between the historical scenario, as the lower bound, and the wet one, as the upper bound of the probable storms in the future, firstly shows the accuracy of PRECIS in simulating rainfall in the study region; secondly, it shows the certain increase in the intensity of the design storms which matter in urban flood management.