Saman Javanroodi; Mohammad Mahdavi; Baharak Motamedvaziri
Abstract
In order to determine the role of hydrological statistic data periods on flood forecasting in Sefidrood Watershed, maximum daily flows were collected from hydrometric stations, thus seventeen valid stations with long-term statistical data period was chosen and in some cases which data were corrupted, ...
Read More
In order to determine the role of hydrological statistic data periods on flood forecasting in Sefidrood Watershed, maximum daily flows were collected from hydrometric stations, thus seventeen valid stations with long-term statistical data period was chosen and in some cases which data were corrupted, data has been reproduced by considering common time base.. At the second step, maximum daily flow data have been divided in to statistic series with 10, 15, 20, 25, 30 and more than 30 years data periods. Fitting statistic distribution of Normal, Normal log, Normal log 2 parameters, Normal log 3 parameters, Pearson log 2, Pearson log 3 and Gamble were determined by “Smada” software, also flood with two, three, five, 10, 25, 50, 100 and 200 return periods were calculated either. Statistic distribution for all of seventeen hydrometric stations were determined separately with RMSE method in Excel and SPSS for maximum daily flow.. Regarding zero-error in forecasted floods for long statistic data period (more than 30 years) in related return periods, compared with short statistic data period (10, 15, 20, 25 and 30 years), flood’s estimated error was calculated for all stations. In general, results indicated that increasing of statistic period, will enhance the accuracy of forecasting in hydrometric stations, and applying short-term data period will result in high levels of error.
