Introduction: River discharge as one of the most important hydrology factors has avital role in physical, ecological, social and economic processes. So, accurate and reliable prediction and estimation of river discharge have been widely considered by many researchers in different fields such as surface water management, design of hydraulic structures, flood control and ecological studies in spetialand temporal scale. Therefore, in last decades different techniques for short-term and long-term estimation of hourly, daily, monthly and annual discharge have been developed for many years. However, short-term estimation models are less sophisticated and more accurate. Various global and local algorithms have been widely used to estimate hydrologic variables. The current study effort to use Lazy Learning approach to evaluate the adequacy of input data in order to follow the variation of discharge and also simulate next-day discharge in Talar River in Kasilian Basinwhere is located in north of Iran with an area of 66.75 km2. Lazy learning is a local linear modelling approach in which generalization beyond the training data is delayed until a query is made to the system, as opposed to in eager learning, where the system tries to generalize the training data before receiving queries.

Drought is a national calamity that has bad effects on ecological, social and economic divisions. Growing of crowd and increasing of water demand aggravate the effects of drought. Analyzing of drought and attention to it in development programs has much importance, especially in Iran that is an arid and semiarid country. “Hydrological drought” means time periods that discharge of river is not sufficient to supply planned demands. In this research the flow duration curve (FDC) established using Pole-Shaloo station daily discharges for 1957-2001 years. Next, the hydrological drought periods extracted by truncation level method and analyzed. The “interior criterion” method (IC) used to eliminating minor droughts and pooling dependence droughts. Concluded that Johnson and Generalized Pareto distributions was best cases for annual maximum series (AMS) include “duration” and “deficit volume” respectively, based on Q70 truncation level so, type 3 Pearson and Generalized Pareto distributions best cases for annual maximum series (AMS) include “duration” and “deficit volume” respectively, based on Q90 truncation level. Therefore the return periods calculated for historical droughts. The most drought’s return periods was less than 20 years except 1963-1964 drought. For example The 2001 drought’s return period, that enveloped most of regions in Iran, was 20 years based on analyzing of deficit volume. Also drought risk calculated for return periods and water resource projects life. The maximum drought deficit volume is related to primary 11 years and 5 years at the end of time, so that in primary 11 years 57 and 50 percent of droughts was happened in lieu of Q90 and Q70 respectively. Also in 5 years of end of studying period 40 and 20 percent of droughts was happened in lieu of Q90 and Q70 respectively. The minimum amount of drought deficit volume was related to 1966-1974 period, so that equal to zero and 20 percent in lieu of Q90 and Q70. In final part of research, the annual maximum series of deficit volume extracted with respect to 10, 20, 30, 60, 90, 120, 150 and180 days durations, and the best probability distributions appointed for these. Consequently, the severity-durationfrequency curve furnished.

Accurate prediction of daily discharge in ungauged watersheds is an important issue in hydrology. Several models have been developed for predicting discharge. The IHACRES is a friendly use model with limited input data which is being used for simulation of watershed outputs. The present study was conducted to assess the applicability of the IHACRES model to simulate daily mean discharge in seven sub-watersheds of Gorganrood in Golestan Province, Iran. Daily time series data of rainfall, stream flow and temperature were applied for the duration of 1986 to 2007. The model was calibrated and consequently validated for each study watershed. The results of the evaluation showed that the model performance in prediction of high flows was fairly good while its performance for flows with frequency of more than 60% was weak. The results also indicated that the simulated peak flows were lower than observed values in almost all stations and in both the calibration and validation periods. The efficiency of IHACRES model in predicting daily discharge was found to be fairly good with maximum and respective coefficients of determination and efficiency of 0.67 and 0.70 (P<0.05).

Introduction River flow prediction is one of the most important key issues in the management and planning of water resources, in particular the adoption of proper decisions in the event of floods and the occurrence of droughts. In order to predict the flow rate of rivers, various approaches have been introduced in hydrology, in which intelligent models are the most important ones. The application of artificial neural networks (ANNs) to various aspects of hydrological modeling has undergone much investigation in recent years. This interest has been motivated by the complex nature of hydrological systems and the ability of ANNs to model non-linear relationships. ANNs are essentially semiparametric regression estimators and well suited for hydrological modeling, as they can approximate virtually any (measurable) function up to an arbitrary degree of accuracy (Hornik et al., 1989). A significant advantage of the ANN approach in system modeling is that one need not have a welldefined process for algorithmically converting an input to an output...

Accurate prediction of river flow is one of the most important factors in surface water recourses management especially during floods and drought periods. In fact deriving a proper method for flow forecasting is an important challenge in water resources management and engineering. Although, during recent decades, some black box models based on artificial neural networks (ANN), have been developed to overcome this problem and the accuracy privilege to common statistical methods (such as auto regression and moving average time series method) have been shown. However these types of models are implicit and complex in proper network design and can not be simply used by other investigators. In this research the genetic programming (GP) model has been developed as an explicit method for river flow prediction and has been used for investigation the effect of daily discharge trend in Absardeh river flow forecasting. The results have been compared with artificial neural network technique. The results indicated that the proposed GP method performed quite well compared to artificial neural network method and is applicable for river flow prediction.

Forecasting of river discharge is a key aspect of efficient water resources planning and management. In this study, two models based on Wavelet Analysis and Artificial Neural networks (ANNs) were developed for forecasting discharge of Behesht-Abad River. For this purpose, mean daily discharge data of mentioned river as well as precipitation data of 17 meteorological stations were used in the period 1999-2008. In the first method, called Cross Wavelet (CW), complex Morlet wavelet was used as analyzer function. Wavelet analyzing was performed for every daily rainfall and average discharge time series, separately. The model equation derived for 1, 2, 3 and 7 days ahead forecasting horizon. In the second method, called Wavelet Neural Networks conjunction (WNN), a preprocessing was done on the initial input matrix using Meyer wavelet. Then the elements of the initial input matrix were normalized and the second input matrix was created. A three layer Feed Forward Back Propagation (FFBP) was formed based on the second input matrix and target matrix. After training the model using Levenberg-Marquardt (LM) algorithm, the river discharges were predicted for short term time horizons.The results showed that the WNN method had higher accuracy in short-term forecasting of river discharge in comparison with CW and ANN methods. In testing stage, in CW model when forecasting horizon reduced from 7-days to 1day, the R2 value increased from 0.5113 to 0.9388, and RMSE decreased from 17.9171 to 8.3226 m3/s. In ANN model when forecasting horizon reduced from 7-days to 1day, the R2 value increased from 0.6705 to 0.9166, and RMSE decreased from 5.9828 to 2.5600 m3/s. Whereas, in WNN model, when forecasting horizon reduced from 7-days to 1day, the R2 value increased from 0.8424 to 0.9927, and RMSE decreased from 3.4678 to 0.8145 m3/s.

Prediction of river flow is one of the main issues in the field of water resources management. Because of the complexity of the rainfall-runoff process, data-driven methods have gained increased importance. In the current study, two newly developed models called Least Square Support Vector Regression (LSSVR) and Regression Tree (RT) are used. The LSSVR model is based on the constrained optimization method and applies structural risk minimization in order to yield a general optimized result. Also, in the RT, data movement is based on laws discovered in the tree. Both models have been applied to the data in the Kashkan watershed. Variables include (a) recorded precipitation values in the Kashkan watershed stations, and (b) outlet discharge values of one and two previous days. Present discharge is considered as output of the two models. Following that, a sensitivity analysis has been carried out on the input features and less important features have been diminished, so that both models have provided better prediction on the data. The final results of both models have been compared. It was found that the LSSVR model has better performance. Finally, the results present these models as suitable models in river flow forecasting.

Introduction Due to the heterogeneity in watersheds and the non-linearity of hydrological behaviors, it is very complicated and difficult to fully understand the relationships within watersheds. Therefore, in evaluating these systems, a modeling process is necessary. Over the last few decades, hydrological/hydraulic models have become essential in hydrology studies due to the development of programming languages and the provision of optimal and efficient algorithms for solving differential problems. The application of rainfall-runoff simulation models for flood events has been extensively studied by researchers in the field of water and soil protection, leading to the development of various models to simulate rainfall-runoff processes. One of the successful models in this field is the TOPKAPI-X model. This model was created in the 1990s at the University of Bologna by Professor Todini as a distributed rainfall-runoff model in watersheds. An important feature of distributed models is their ability to simulate components at any point of the watershed, allowing results to be extracted at any required point. Unlike lumped models that consider the entire watershed as a single unit, distributed models allow spatial distribution at any point in the watershed. Therefore, in this research, after calibrating and validating the TOPKAPI-X physical-distributed model in the studied basin, the model was optimized for flood estimation.     Materials and methods The Gamasiab basin is located in the west of Iran, in the northern region of the Zagros mountain ranges, to the north of the Karkheh dam basin, and primarily within the territories of Hamadan and Kermanshah provinces. The mountainous regions of this basin are mainly concentrated in the northern and southern parts, while its lowlands and plains are mostly located in the middle and southeastern parts of the basin (Ministry of Energy, 2014). In this research, the TOPKAPI-X model was used to simulate floods in the Gamasiab watershed. First, the watershed boundary was delineated using a digital elevation model (DEM) with a resolution of 30 meters. Land use maps, soil texture, watershed network, and climatic components were entered into the TOPKAPI-X model. The outlet location of the basin (hydrometric station) was used to simulate the flow using the TOPKAPI-X distributed hydrological model. Continuous time series data on a daily time step were used in this rainfall-runoff model. Specifically, daily rainfall data from 13 rain gauge stations and temperature data from 4 synoptic stations during the statistical period (1999 to 2020) were used to simulate the flow. After running the model several times, the general parameters were manually adjusted each time until the optimal values of the general parameters were obtained by considering the appropriate values of the evaluation criteria (NS and Bias) for the basin.     Results and discussion This research was conducted to analyze the flood discharge of one of the main sub-basins of the Karkheh dam basin using the TOPKAPI-X model on a daily time scale. In the TOPKAPI-X software environment, simulations were performed during the calibration period using input maps and observational rainfall, temperature, and discharge data. A visual comparison of the observed and simulated hydrographs allows for a general and quick evaluation of the model's accuracy. The graphical results of the comparison between the discharge generated by the TOPKAPI-X model with the calibrated parameters and the measured discharge in the Gamasiab basin were presented. The TOPKAPI-X model has the ability to estimate the maximum daily flow rates of the Gamasiab basin, however, some of the simulated flow rates are higher than the observed flow rates. Four criteria—NSE, R, BIAS, and RMSE—were used to evaluate the model. The evaluation results of the TOPKAPI-X model indicate the accuracy of flow simulation, with a Nash-Sutcliffe criterion of 0. 697 during the calibration period (1999-2014) and 0. 660 during the validation period (2015-2020) for the Gamasiab basin. Therefore, it can be concluded that this model has good performance for flow simulation.     Conclusions The importance and usefulness of hydrological models for water resources management, understanding hydrological processes, and conducting impact assessment studies is clear. Hydrological models are crucial tools that enable scientists and policymakers to make informed decisions based on simulations of watershed behavior. Considering the increasing demand for water and the impact of climate change, hydrological simulation will be one of the essential methods for future water management. The results of this study showed that the TOPKAPI-X model has potential in simulating runoff in the selected basin. Due to the capabilities of the TOPKAPI-X distributed hydrological model, this software is recommended as a modeling tool for other basins.

One of the most important of hydrological computing in ecosystem is estimation of the relationship between rainfall and runoff. So that investigation occurred processes in it and the estimate of important outputs such as flood and sediment is considered one the most important mission of a watershed project. Because of variable spatial and temporal characteristics of incident in the water cycle and the nonlinear relationship and uncertainties, none of the statistical and conceptual models are able to be a better and capable model for that. But today using nonlinear networks as intelligent system for forecasting such complicated event can be efficient and effective in many problems of ecology. For this aim it is used variables such as precipitation, temperature, evartanspiration, relative humidity and discharges in daily scale over 42 years period and assessment 62 different suggested structures for surveying river flow in Amame representative watershed. For comparison it used Multi Layer Perceptoron (MLP) and Radial Basis Function (RBF). The results show that out of 6000 available models for estimation river flow model number 54 with 8-9-8-1 network structure and 8 types of input variable such as precipitation (Pt), precipitation with two lags (Pt-1 and Pt-2), temperature (Tt), evartanspiration (ETt), relative humidity (Rht), and discharge with two lags (Qt-1 and Qt-2) with Multi Layer Perceptoron method has the best function. The error of model was 0. 03, 0. 18 and 0. 04 in training and 0. 02, 0. 14 and 0. 02 for testing stage for MSE, RMSE and MAE, respectively.