Showing 20 results for Sensitivity Analysis
A. Montazar, S. Kouchakzadeh, A. Liaghat, M. H. Omid,
Volume 8, Issue 3 (10-2004)
Abstract
The operation of an irrigation network is the result of a decision-making system in which three elements, i. e. the physical condition of the structures, control capacity, and hydraulic behavior of canal system, have important roles. The impact of these components is incorporated in the hydraulic sensitivity concept. For this purpose, the sensitivity of structures could be considered as the most important factor in the hydraulic characteristics of a system. The sensitivity analysis approach is one of the flow analysis methods that is used to forecast the flow behavior in several irrigation networks. Baffle Modules are one of the most suitable structures for flow regulating and delivery. In this paper, hydraulic sensitivity equations are presented for this type of structure. Also, the quantitative variations of the sensitivity index have been evaluated under operating conditions and compared with those of theoretical conditions. Then the influence of the sensitivity index variations on the performance of structures has been studied. This study was carried out on some modules of the Qazvin network.
The results indicated that the hydraulic sensitivity of these off-takes varies under operating conditions. In some cases, variations were estimated to be more than 100%. The range of discharge variations was twice the predefined acceptable value (20%) in some offtakes.
M. Hamidpour, A. Jalalian, M. Afyuni, B. Ghorbani,
Volume 16, Issue 62 (3-2013)
Abstract
Models are helpful tools to predict runoff, sediment and soil erosion in watershed conservation practices. The objectives of this research were to investigate sensitivity analysis, calibration and validation of EUROSEM model in estimation of runoff in Tangh-e-Ravagh sub-basin of Karoun watershed. The model was tested in a one hectare experimental test site. The area was divided into nine elements according to EUROSEM user's manual. A triangular weir was installed at the outlet of the area to collect runoff in specified time periods for six rainfall events. Sensitivity analysis of the model was performed by a ±10% change in the dynamic parameters of the model and examining the outputs for a rainstorm. Sensitivity analysis showed that total runoff was sensitive to saturated hydraulic conductivity and insensitive to soil cohesion. Sensitivity analysis indicated that the model sensitivity depends on evaluation conditions and it is site-specific in nature. Calibration and validation of the model was performed on input parameters. Calibration of hydrographs was performed by decreasing saturated hydraulic conductivity and capillary drive and increasing initial soil moisture. Validation results showed that EUROSEM model simulated well the total runoff and peak of runoff discharge, but it could not simulate well the time of runoff, time to peak discharge
H. Akbari Mejdar, A. Bahremand, A. Najafinejad, V. Sheikh,
Volume 18, Issue 67 (6-2014)
Abstract
Over-parameterization is a well-known and often described problem in hydrological models, especially in distributed models. Therefore, using special methods to reduce the number of parameters via sensitivity analysis is important to achieve efficiency. This paper describes a sensitivity analysis strategy that graphically assigns for each parameter a relative sensitivity index and relationship of the parameter and the outputs of the model. The method is illustrated with an application of SWAT model in the Chehelchai catchment, Golestan province. In this study, total water yield, along with four major parts of water budget including surface runoff, lateral flow, groundwater and evapotranspiration was selected as objective function. SWAT is a river basin model that can be used to predict the impact of land management practices on water, sediment and agricultural chemical yield in watersheds. A relative sensitivity index was used for ranking the sensitivity of parameters. The results showed that soil evaporation compensation facto (ESCO), CN, soil available water capacity (SOL-AWC), deep aquifer percolation fraction (RCHRG-DP) and soil bulk density (SOL-BD) have the most influence on river flow. These parameters are generally stated as the most sensitive parameters of SWAT model in most of the same researches worldwide
F. Moradi, B. Khalilimoghadam, S. Jafari, S. Ghorbani Dashtaki,
Volume 18, Issue 69 (12-2014)
Abstract
Soft computing techniques have been extensively studied and applied in the last three decades for scientific research and engineering computing. The purpose of this study was to investigate the abilities of multilayer perceptron neural network (MLP) and neuro-fuzzy (NF) techniques to estimate the soil-water retention curve (SWRC) from Khozestan sugarcane Agro-Industries data. Sensitivity analysis was used for determining the model inputs and appropriate data subset. Also, in this paper, the van Genuchten and Fredlund and xing models were used to predict SWRC. Measured soil variables included particle size distribution, organic matter, bulk density, calcium carbonate, sodium adsorption ratio, electrical conductivity, acidity, mean weight diameter, plastic and liquid limit, resistance of soil penetration, water saturation percentage and water content for matric potentials -33, -100, -500 and -1500 kPa. The results of this study in terms of various statistical indices indicated that both MLP and NF provide good predictions but the neural network provides better predictions than neuro-fuzzy model. For example, using MLP and NF models values of NMSE at prediction θs, θr, α, n and m in Fredlund and Xing equation corresponded to (0.059, 0.065), (0.154, 0.162), (0.109, 0.117), (0.129, 0.135) and (0.129, 0.145), respectively. Furthermore, α and n parameters at the first depth, and θr and α parameters at the second depth in Fredlund and Xing equation were estimated with higher accuracy compared with equivalent parameters in van Genuchten equation
S. S. Heshmati, H. Beigi Harchegani,
Volume 18, Issue 69 (12-2014)
Abstract
The aim of this study was to assess the drinking quality of Shahrekord aquifer based on a GWQI (groundwater quality index) within a GIS framework. To do this, samples from 97 wells were analyzed for pH, Electrical Conductance (EC), Total Dissolved Solids (TDS), Total Suspended Solids (TSS), Turbidity, Ca2+, Mg2+, Na+, K+, Cl-, HCO3- and SO42-, and total hardness was also calculated. These water quality parameters were geostatistically mapped. Maps showed that maximum quality of water occurs in the northwest while the lowest quality occurs in the south of aquifer. To calculate GWQI index, each map was difference-normalized and converted to a rank map. Assuming the mean value of each rank map to be the weight of corresponding parameter, a GWQI map was created with values varying from 0 (lowest) to 99 (highest quality). Mean GWQI of 84 indicates a relatively good drinking quality of water in the aquifer. However, based on the GWQI map the quality of water declines from very good (GWQI=87) in northwest to a lower quality (GWQI= 80) in southern part of the aquifer. The lower quality of water in the southern part may have been caused by industrial activities, intensive animal husbandry, presence of wastewater plant, irrigation with treated municipal effluent and also by the inward hydraulic gradient. Map removal sensitivity analysis indicated that TSS and to some extent Na+ were important water parameters in this aquifer, which must be monitored with greater accuracy and frequency.
R. Lalehzari, S. Boroomand-Nasa, M. Bahrami,
Volume 18, Issue 69 (12-2014)
Abstract
Advance velocity is an important factor in surface irrigation system design and simulation. Volume balance is a simple model based on continuity equation used in surface irrigation design and management. In the past volume balance models, it is generally assumed that the upstream depth of surface water is constant and equal to normal depth. This initial assumption may cause significant errors in computing advance flow. In this paper, a modified volume balance (MVB) model is developed to predict the advance curve in furrow irrigation. In the suggested method the upstream surface, water depth is actual depth and variable in time. Predicted advance distance of VB, VB-ZI and MVB was compared to the observed data obtained for the three furrow lengths of 60, 80 and 90m. Evaluation indexes indicated that the modified volume balance equation is more accurate than the previous equations by RMSE 9.26, 7.37 and 6.76 respectively. Sensitivity analysis showed that the inlet discharge has the greatest effect on the model and the model is more sensitive to decreasing the discharge amount than to increasing it
S. Dowlatabadi, S. M. A. Zomorodian,
Volume 19, Issue 71 (6-2015)
Abstract
One of the most essential and appropriate groundwater model components is accurate information of the recharge values among input data often introduced to the model as the percentage of rainfall of aquifers. The recharge values are influenced by many temporal and spatial factors. Firoozabad plain is one of the suitable plains for agriculture in the Fars province in which utilization of groundwater resources has been banned since 23 September 2002, due to the declining water level and negative balance. The main purpose of this study was to estimate the recharge values of groundwater aquifer by using SWAT in the MODFLOW model. Firstly, surface water was simulated via SWAT model, and sensitivity analysis, calibration, validation and uncertainty analysis of results were performed by SWAT-CUP software. After extraction of aquifer recharge values from the calibrated model, the groundwater of basin was simulated via MODFLOW model in both steady and unsteady conditions. Following the model calibration, the hydrodynamic coefficients of plain were determined and sensitivity of model was checked in terms of hydraulic conductivity and discharge rate of pumping wells. As for the confidence, the model was revalidated, which proved in simulating the behavior of the aquifer very well.
H. Beigi Harchegani, S. S. Heshmati,
Volume 19, Issue 72 (8-2015)
Abstract
The aim of this paper is to adapt a water quality index for individual samples and to compare the results with that of the original GIS-based approach. Thirteen water quality parameters observed in 97 wells from the Shahrekord aquifer were used. In GIS-based method, quality parameters maps are difference-normalized, ranked and GWQI map is drawn. In derived method, observations from individual wells were separately and similarly treated to obtain WQI for each well. Both GWQI maps displayed similar trends and were highly correlated (R=0.91). While the minimum and mean GWQI for both methods were identical (respectively 81 and 84) the derived method estimated the maximum GWQI slightly lower (7%) and showed up to 6% difference in water quality class coverage. Overall, the derived method GWQI is more correlated with observations and performs better than the GIS-based method, and therefore, can be used for determining the overall quality of individual water samples and without the requirement of samples being spatially distributed.
M. Goodarzi, J. Abedi Koupai, M. Heidarpour, H. R. Safavi,
Volume 19, Issue 73 (11-2015)
Abstract
Due to the time and space changes of hydrological events in the arid and semi-arid regions, recharge measurement in these areas is very difficult. Hence, groundwater recharge is a complicated phenomenon for which there is not a fixed method to determine. The aim of this research was to develop a method for estimation of groundwater recharge based on a hybrid method. In this study, a hybrid method for calculating recharge was presented by combining empirical methods with a mathematical model, MODFLOW, and AHP analysis. The results showed that the most important parameters affecting groundwater recharge are soil properties, unsaturated thickness, land cover, land slope, irrigation and precipitation, from which the soil properties and precipitation are most important. The results showed that the overall impact of small changes in precipitation and temperature significantly affect the groundwater recharge, and heavy soils are much more sensitive to these changes than light soils. By changing 10% precipitation, the recharge rate is changed between 16% and 77% and by changing 1ºC temperature, the recharge rate is changed between 6% and 42%. Also, results showed that precipitation and evapotranspiration changes in four months including December, January, February and March had significant effects on annual recharge rate. Using the results of this research, the vulnerable areas of the plain, appropriate places and time for artificial recharge could be identified. Overall, the results of this study can be useful in various aspects of groundwater management.
A. H. Boali, H. Bashari, R. Jafari, M. Soleimani,
Volume 21, Issue 2 (8-2017)
Abstract
Appropriate criteria and methods are required to assess desertification potential in various ecosystems. This paper aimed to assess desertification levels in Segzi plain located in east part of Isfahan, with a focus on soil quality criteria used in MEDALUS model. Bayesian Belief Networks (BBNs) were also used to convert MEDALUS model into a predictive, cause and effects model. Soil samples were collected from 17 soil profiles in all land units and some of their characteristics such as texture, soluble sodium and chlorine, organic material, Sodium Absorption Ratio (SAR), Electrical Conductivity (EC) and CaSo4 of all soil samples were determined in soil laboratory. The effects of measured soil quality indicators on desertification intensity levels were assessed using sensitivity and scenario analysis in BBNs. Results showed that the used integrated method can appropriately accommodate uncertainty in the desertification assessments approaches created as a result of the influence of different soil characteristics on desertification. According to the results of MEDALUS model, 28.28 % and 71.72 % of the study area were classified as poor and moderate areas in terms of soil quality respectively. Sensitivity analysis by both models showed that soil organic matter, SAR and EC were identified as the most important edaphic variables responsible for desertification in the study area. Evaluating the effects of various management practices on these variables can assist managers to achieve sound management strategies for controlling desertification.
A. H. Boali, R. Jafari, H. Bashari,
Volume 21, Issue 3 (11-2017)
Abstract
This paper aimed to assess the severity of desertification in Segzi plain located in the eastern part of Isfahan city, focusing on groundwater quality criteria used in MEDALUS model. Bayesian Belief networks (BBNs) were also used to convert MEDALUS model into a predictive, cause and effects model. Different techniques such as Kriging and IDW were applied to water quality data of 12 groundwater wells to map continuous variations of the CL, SAR, EC, TDS, pH and decline in water table indices in GIS environment. The effects of measured water quality indicators on desertification severity levels were assessed using sensitivity and scenario analysis in BBNs model. According to the results of the MEDALUS, the desertification of the study area was classified as severe class due to its low quality of groundwater. Sensitivity analysis by the both models showed that decline in waater table, water chloride content and electrical conductivity were the most important parameters responsible for desertification in the region from ground water condition standpoint. The determination coefficient between the outputs of the MEDALUS and BBNs models (R2>0.63) indicated that the results of both models were significantly correlated (α=5 %). These results indicate that the application of BBNs model in desertification assessment can appropriately accommodate the uncertainty of desertification methods and can help managers to make better decision for upcoming land management projects.
S. Esmailian, A. Talebi, M. Esmailian,
Volume 22, Issue 1 (6-2018)
Abstract
This research was aimed to simulate and prioritize the effective factors on water erosion using USLE-M in the system dynamic model. In this integrated model, by using the system dynamic simulation software (Vensim), all variables and factors involved in erosion and soil loss were considered according to the USLE-M model. After model implementation, the estimated values and observations were compared and then sensitivity analysis was done to determine the sensitive parameters. Then, calibration was performed on the sensitive parameters. This study found that that the results of the model were acceptable for soil erosion simulation due to considering all the effective factors in soil erosion. The results of the sensitivity analysis also indicated high model sensitivity to the slope and vegetation cover in high and low slopes, respectively. By investigating the changes in various parameters such as vegetation cover and slope on erosion, the optimal vegetation cover with 67 and 40% slope, was estimated to be 20 and 60%, respectively.
Z. Abbasi, H. Azimzadeh, A. Talebi, A. Sotoudeh,
Volume 22, Issue 4 (3-2019)
Abstract
Groundwater quality evaluation is very necessary to provide drinking water. Groundwater excessive consumption can cause subsidence and penetration of saline groundwater into freshwater aquifers in Ajabshir Plain, on the Urmia lake margin. The main goal of the current project was to evaluate the groundwater quality by employing the qualitative indices of groundwater and GIS. Ten parameters of 15 wells including EC, TDS, total hardness as well as the concentration of Ca++, Na+, Mg++, K+, SO4--, HCO3- and Cl- were analyzed. At first, the maps of parameters concentration were prepared by the kiriging method. Then based on WHO drinking water standards, the maps were standardized and ranked for drawing the maps of quality indices. The results showed that quality index changes were in the range of moderate (61) to acceptable (81). Removing the single map method of sensitivity analysis detected the quality index was more sensitive to the K+ parameter. Finally, the quality index from the eastern north to the western south of Ajabshir Plain and the other areas was ranked in the acceptable and moderate classes, respectively.
F. Amirimijan, H. Shirani, I. Esfandiarpour, A. Besalatpour, H. Shekofteh,
Volume 23, Issue 3 (12-2019)
Abstract
Use of the curve gradient of the Soil Water Retention Curves (SWRC) in the inflection point (S Index) is one of the main indices for assessing the soil quality for management objectives in agricultural and garden lands. In this study Anneling Simulated – artificial neural network (SA-ANN) hybrid algorithm was used to identify the most effective soil features on estimation of S Index in Jiroft plain. For this purpose, 350 disturbed and undisturbed soils samples were collected from the agricultural and garden lands and then some physical and chemical soil properties including Sand, Silt, Clay percent, Electrical Conductivity at saturation, Bulk Density, total porosity, Organic Mater, and percent of equal Calcium Carbonate were measured. Moreover, the soil moisture amount was determined within the suctions of 0, 10, 30, 50, 100, 300, 500, 1000, 1500 KP using pressure plate. Then, the determinant features influencing the modeling of S Index were derived using SA-ANN hybrid algorithm. The results indicated that modeling precision increased by reducing the input variables. According to the sensitivity analysis, the Bulk Density had the highest sensitivity coefficient (sensitivity coefficient=0.5) and was identified as the determinant feature for modeling the S Index. So, since increasing the number of features does not necessarily increase the accuracy of modeling, reducing input features is due to cost reduction and time-consuming research.
H. Siasar, T. Honar, M. Abdolahipour,
Volume 23, Issue 4 (2-2020)
Abstract
The estimation of reference crop evapotranspiration (ETo) is one the important factors in hydrological studies, irrigation planning, and water resources management. This study attempts to explore the possibility of predicting this key component using three different methods in the Sistan plain: Generalized Linear Models (GLM), Random Forest (RF) and Gradient Boosting Trees (GBT). The maximum and minimum temperature, mean temperature, maximum and minimum humidity, mean humidity, rainfall, sunshine hours, wind speed, and pan evaporation data were applied for years between 2009 to 2018. Using various networks, the ETo as output parameter was estimated for different scenarios including the combination of daily scale meteorological parameters. In order to evaluate the capabilities of different models, results were compared with the ETo calculated by FAO Penman-Monteith as the standard method. Among studied scenarios, M1 covering the maximum number of input parameters (10 parameters) showed the highest accuracy for GBT model, with the lowest RMSE (0.633) and MAE (0.451) and the maximum coefficient of regression (R = 0.993). Air temperature was found as the most sensitive parameters during sensitivity analysis of studied models. It indicated that accuracy and precision of temperature data can improve the results. Application of the GBT model could decrease the time consumed to run the model by 70%. Therefore, the GBT model is recommended for estimation of ETo in the Sistan plain.
H. Mahmoudpour, S. Janatrostami, A. Ashrafzadeh,
Volume 24, Issue 3 (11-2020)
Abstract
Given the fact that the DRASTIC index is ineffective in addressing the saltwater uprising issue in coastal plains, in the present study, three factors including land use, distance to shoreline, and differences between groundwater and sea level were added to the DRASTIC index. The proposed modification to DRASTIC was validated using the measured electrical conductivity (EC) data gathered from groundwater monitoring wells throughout the Talesh Plain. The results showed that the coefficient of correlation between the map of EC over the region and the modified DRASTIC was 0.52, while for the original DRASTIC, the coefficient was 0.45, thereby implying a stronger relationship between EC and the modified DRASTIC in the Talesh Plain. Sensitivity analysis also showed that DRASTIC and the modified DRASTIC were the most sensitive to, respectively, depth to groundwater (D) and land use (Lu). According to the single-parameter sensitivity analysis results, depth to water table and net recharge were the most effective parameters in DRASTIC, whereas the modified DRASTIC was the most sensitive to land use and depth to groundwater. It could be concluded that modifying the DRASTIC index would result in decreasing the area of very high and high vulnerable classes, and the area classified as low and moderate vulnerable could be increased.
E. Yarmohammadi, S. Shabanlou, A. Rajabi,
Volume 25, Issue 1 (5-2021)
Abstract
Optimization of artificial intelligence (AI) models is a significant issue because it enhances the performance and flexibility of the numerical models. In this study, scour depth around bridge abutments with different shapes was estimated by means of ANFIS and ANFIS-Genetic Algorithm. In other words, the membership functions of the ANFIS model were optimized using the genetic algorithm, finding that the performance of ANFIS model was increased. Firstly, effective input parameters on the scour depth around bridge abutments were defined. Then, by using the input parameters, eleven ANFIS and ANFIS-GA models were produced. Next, the superior ANFIS and ANFIS-GA models were introduced by analyzing the numerical results. For example, the correlation coefficient and scatter index for ANFIS model were calculated to be 0.979 and 0.070; for ANFIS-GA, these were 0.986 and 0.056, respectively. In addition, the average discrepancy ratio (DRave) for ANFIS and ANFIS-GA models was 0.984 and 0.988, respectively. Also, it was shown that the ANFIS-GA models had more accuracy, as compared to the ANFIS models. Moreover, a sensitivity analysis showed that Froude number (Fr) and ratio of flow depth to radius of scour hole (h/L) were the most influential input parameters for simulating the scour depth around bridge abutments.
F. Hayati, A. Rajabi, M. Izadbakhsh, . S. Shabanlou,
Volume 25, Issue 1 (5-2021)
Abstract
Due to drought and climate change, estimation and prediction of rainfall is quite important in various areas all over the world. In this study, a novel artificial intelligence (AI) technique (WGEP) was developed to model long-term rainfall (67 years period) in Anzali city for the first time. This model was combined using Wavelet Transform (WT) and Gene Expression Programming (GEP) model. Firstly, the most optimized member of wavelet families was chosen. Then, by analyzing the numerical models, the most accurate linking function and fitness function were selected for the GEP model. Next, using the autocorrelation function (ACF), the partial autocorrelation function (PACF) and different lags, 15 WGEP models were introduced. The GEP models were trained, tested and validated in 37, 20- and 10-years periods, respectively. Also, using sensitivity analysis, the superior model and the most effective lags for estimating long-term rainfall were identified. The superior model estimated the target function with high accuracy. For instance, correlation coefficient and scatter index for this model were 0.946 and 0.310, respectively. Additionally, lags 1, 2, 4 and 12 were proposed as the most effective lags for simulating rainfall using hybrid model. Furthermore, results of the superior hybrid model were compared with GEP model that the hybrid model had more accuracy.
A.h. Azimi, S Shabanlou, F. Yosefvand, A. Rajabi, B. Yaghoubi,
Volume 25, Issue 4 (3-2022)
Abstract
In this research, the scour hole depth at the downstream of cross-vane structures with different shapes (i.e., J, I, U, and W) was simulated utilizing a modern artificial intelligence method entitled "Outlier Robust Extreme Learning Machine (ORELM)". The observational data were divided into two groups: training (70%) and test (30%). Then, using the input parameters including the ratio of the structure length to the channel width (b/B), the densimetric Froude number (Fd), the ratio of the difference between the downstream and upstream depths to the structure height (Δy/hst), and the structure shape factor (φ), eleven different ORELM models were developed for estimating the scour depth. Subsequently, the superior model and also the most effective input parameters were identified through the conduction of uncertainty analysis. The superior model simulated the scour values by the dimensionless parameters b/B, Fd, Δy/hst. For this model, the values of the correlation coefficient (R), the variance accounted for (VAF), and the Nash-Sutcliffe efficiency (NSC) for the superior model in the test mode were obtained 0.956, 91.378, and 0.908, respectively. Also, the dimensionless parameters b/B and Δy/hst were detected as the most effective input parameters. Furthermore, the results of the superior model were compared with the extreme learning machine model and it was concluded that the ORELM model was more accurate. Moreover, an uncertainty analysis exhibited that the ORELM model had an overestimated performance. Besides, a partial derivative sensitivity analysis (PDSA) model was performed for the superior model.
M.j Amiri, M. Bahrami, M. Mousavi Poor, A. Shabani,
Volume 26, Issue 4 (3-2023)
Abstract
Class A pan evaporation method as one of the most common methods for reference evapotranspiration (ET0) estimation has been widely used in the world due to its simplicity, relatively low cost, and ability to estimate daily ET. In this study, the performance of 8 empirical methods consisting of Allen and Pruitt (1991), Cuenca (1989), Snyder (1992), modified Snyder, Pereira, et al. (1995), Orang (1998), Raghuwanshi and Wallender (1998), and FAO/56 were analyzed to estimate class A pan coefficient and ET0 at Fasa synoptic station located in Fars province. The calculated pan evaporation coefficients from the above equations were compared with measured pan evaporation coefficients which were obtained from the ratio of evapotranspiration calculated by the FAO-Penman-Monteith method to the rate of evaporation from the pan. The results showed that all empirical methods did not predict pan coefficient values well (R2 < 0.3 and NRMSE > 0.25). The comparison results between ET0 from empirical methods and ET0 obtained from FAO-Penman–Monteith indicated that the FAO/56 method had the best performance (R2 = 0.72 and NRMSE = 0.3). To increase the accuracy of empirical pan coefficient equations, these equations were modified with eight years (2007-2015) of meteorological data from the Fasa synoptic station and validated using two years of independent data (2015-2017). The results showed that the accuracy of all empirical models was improved and the Cuenca equation with NRMSE = 0.16 and R2= 0.63 was selected as the best equation for pan coefficient estimation and ET0 (R2 =0.85; NRMSE =0.18) in Fasa region. The sensitivity analysis revealed that the estimated pan coefficient is more sensitive to wind speed, followed by relative humidity, fetch distance, the slope of the saturation vapor pressure curve, sunshine hours, and air pressure. According to statistical results and sensitivity analysis, an equation was expanded for the Fasa region and other areas with the same climate.
