جستجوی مقالات مرتبط با کلیدواژه « Galikesh Watershed » در نشریات گروه « آب و خاک »

One of the most critical issues that have always been the concern of researchers in water engineering and hydrology is the simulation of runoff or river discharge to plan, prevent damages, and solve the water shortage problem, and soil erosion. In addition, due to the ever-increasing limitations of extractable freshwater resources, it is very important to predict the river discharge and its changes as accurately as possible. The prediction of runoff, this important hydrological variable, significantly impacts the sustainable management of water resources, engineering designs, environmental protection, water supply planning, water quality management, irrigation systems, and electricity generation worldwide. Besides, for data accuracy and to modify and complete the data, the results of these models can be used. The rainfall-runoff process is one of the most complex, dynamic, and non-linear hydrological phenomena, which is influenced by various factors such as temporal and spatial changes, geomorphology, and climatic characteristics of the catchment area. Knowing the connection between precipitation and runoff is one of the important issues of hydrology because precipitation data are used in flood prediction, and a good prediction is made when a suitable relationship is defined. By increasing the accuracy of river runoff prediction, more efficient management and planning are done. Therefore, improving runoff prediction modeling seems essential.

So far, complex and diverse relationships have been presented to predict the extent of river floodings, such as conceptual rainfall-runoff models, time series linear models, and hybrid models. However, due to the lack of accurate knowledge and the complexity of factors affecting river flooding, the values ​​calculated from various relationships have significantly differed in many cases. In the meantime, hydrological models, with their potential, are considered efficient tools, especially in climate change conditions. One of the models that researchers use to model rainfall and runoff is the RRL (Rainfall Runoff Library) hydrological model. This software package includes integrated and conceptual models (such as AWBM, Sacramento, SMAR, SimHyd, and Tank). In this research, the comparative evaluation of the Sacramento, SimHyd, and SMAR models of this tool has been done in predicting the long-term runoff of the Galikesh watershed and also investigating the effect of parameters on the performance of each model.

In this research, after preparing the input data, the models were calibrated and validated for 1989-2010 and 2010-2019. The simulated and observation runoff results were analyzed to check the potential of the models. Furthermore, after evaluating the model using the optimized parameters, the sensitivity of each of the parameters of the three Sacramento, SimHyd, and SMAR models was investigated in the simulation of runoff from the Galikesh watershed so that the sensitivity of the models to the change of parameters and the effect of each parameter in the simulation makes it clear. To be the results of the evaluations indicate the optimal performance of all models in runoff simulation. However, the Sacramento model with a Nash Sutcliffe coefficient of 0.82 for the calibration period and 0.7 for the validation period, and then the SimHyd model with a Nash Sutcliffe coefficient of 0.71 and 0.76 for the calibration and validation period has the best performance in the runoff simulation of the Galikesh watershed. The SMAR model has shown weaker results than other models in runoff simulation. Finally, the sensitivity of all parameters was checked. The results showed some parameters such as LZTWM (Lower zone tension water capacity) and Zperc (Maximum percolation rate coefficient) in the Sacramento model, impervious threshold parameters and infiltration coefficient in the SimHyd model, and the evaporation conversion (T) parameter in the SMAR model was the most sensitive to reducing their values. Also, increasing the value of the Rexp (Percolation equation exponent) parameter in the Sacramento model and the proportion direct runoff (H) parameter in the SMAR model has the greatest impact on the simulation runoff compared to other parameters.

Different models have been proposed to explain these complexities, considering the importance of runoff forecasting and the non-linearity of converting precipitation into the runoff. Different structures and approaches of studied models have led to their different predictions, which has led to the importance of the comparative evaluation of the models for various purposes. For this purpose, in this research, three hydrological models, Sacramento, SimHyd, and SMAR have been used to simulate the runoff of the Galikash catchment area. The investigations showed that all three models could simulate the outflow of the watershed, and all the models have successfully simulated high amounts of runoff. However, the Sacramento model has performed better than the others. Models parameters sensitivity analysis has been investigated considering the importance and effect on runoff simulation. Finally, The sensitivity analysis showed that some parameters are more sensitive than others in the runoff simulation. The optimal amount of these parameters should be considered during the simulation due to their high sensitivity.

Artificial neural networks (ANN) are one of the data mining methods applied by many researchers in different fields of studies such as rainfall runoff modeling. To improve the performance of these networks, deep learning neural networks were developed to increase modeling accuracy. This study evaluated deep learning networks to improve the performance of artificial neural networks in Galikesh watershed and to predict discharge for 1, 3, 6 and 12-month time scale based on 1 to 5 month time scale lags made in precipitation and temperature data. Based on 70% and 30% of the data used for training and test respectively the results demonstrated that in all time steps, the deep learning neural network improved the performance of artificial neural network and on average RMSE decreased in both training and test from 0.68 to 0.65 and 0.84 to 0.73 respectively. Moreover, R-square was increased on average from 0.57 to 0.62 and 0.51 to 0.67 respectively in training and test. We can also denote the effect of temperature on the increase of accuracy of rainfall-runoff modeling.

Climate change causes changes in temperature and rainfall and consequently affects river discharge. Changes in rainfall can be simulated by global circulation models under different climatic scenarios but investigations of changes in river discharge require rainfall-runoff models. The Galikesh basin as one of most flood-prone basins in Gloestan Province is considered for determination of changes in river discharge under climate change effect. Temperature and rainfall is produced for future climatic period (2011-2030) based on global circulation model HADCM3, using LARS-WG data generator model. The produced data under different climatic scenario are used as inputs of calibrated IHACRES model to simulate river discharge for future climatic periods. The climate change analysis shows that under different scenarios in the study area air temperature in different months increases, but the annual precipitation decreases. The Mann-Kendall test is used to detect monotonic trend of seasonal and semiannual river discharge series. The results show no trend for spring and first half of the year and a negative trend for other seasons and second half series in 5 percent significance level.