Reducing the computational cost and time of environmental flow estimation based on machine learning approaches

Reducing the computational cost and time of environmental flow estimation based on machine learning approaches | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Reducing the computational cost and time of environmental flow estimation based on machine learning approaches Seiran Haghgoo, Jamil Amanollahi, Barzan Bahrami Kamangar This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3939514/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract In recent decades, the reckless exploitation of rivers has caused significant changes in their ecosystems and upstream flow. It is imperative to understand that preservation of river ecosystems solely relies on maintaining the environmental flow (EF). Estimating the EF requires filed sampling, which are both time-consuming and costly. Thus, the purpose of this research is to estimate EF for a river and generalize its result to adjacent rivers using the modelling. To determine the EF, the physical habitat simulation (PHABSIM) model was used. Habitat suitability indexes (HSI) were created based on the filed survey for water velocity, flow depth, channel index and water temperature in a river. To predict the EF for other rivers, the linear regression model (LR) and two different types of neural network models, including Adaptive Neuro-Fuzzy Inference System (ANFIS) and multi-layer perceptron (MLP) were utilized. In this study, 80% and 20% of the data were used for training and testing phases, respectively. Among the models, in the ANFIS model, the date obtained for both training phase and testing phase were as follows respectively. R 2 = 0.98, RMSE = 0.0248 and MAE = 0.0006 as well as R 2 = 0.97, RMSE = 0.0295 and MAE = 0.0008. The accuracy of them were higher compared to MLP and LR models in predicting EF. Therefore, the ANFIS hybrid model can be a suitable alternative method for estimating the EF. Water velocity Flow depth Channel index PHABSIM Habitat Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Population growth and rapid growth of different aspects of cities have increased the demand for water (Pandey 2021 ), which has created challenges and changes in river basins that are necessary to maintain ecosystems (IUCN 2000). In countries where freshwater resources are limited to meet existing needs, they are forced to strike a balance between water extracted for human activities and water stored for a river or watershed management system (Perez-Blanco et al. 2021 ; Senent-Aparicio et al. 2021 ; Zolfagharpour et al. 2021 ). In most developing countries, due to the lack of minimum required information about rivers, there are no comprehensive and flexible methods to determine the environmental flow (EF) (Sedighkia and Datta 2023 ; Aazami et al. 2022 ). Hence, managing and maintaining the ecosystem of such rivers seem very difficult (Wang et al. 2022 ; Shang et al. 2021). EF is the water regime provided for a river, wetland or coastal area in order to protect ecosystems and their benefits (Espinoza et al. 2021 ; Zheng et al. 2020 ). The flow regime, according to the IFC definition, is the hydrological feature of a river and includes four flow levels, including life flow of period of drought (Vivian et al 2014 ), basal flow as a natural flow to maintain river habitat in healthy conditions (Gain et al. 2013 ; Belmar et al. 2011 ), high flow that removes sediment in the environment (Dumitriu 2020 ), and overflow bank that connects the main river to the floodplain, which is morphologically very significant (Bouska 2020 ; Anonymous 2005). The purpose of various methods designed to preserve river ecosystems is to maintain one or more parts of the mentioned flow regimes (Nasiri Khiavi et al. 2024 ; Sedighkia and Abdoli 2023 ). The compromise between economics benefit of rivers such as hydropower potential and environmental protection must be considered in estimate the EF (Yu et al. 2021 ; Mlynski et al. 2020 ; Bejarano et al. 2019 ). There are different divisions in order to determine the EF of rivers. The numerical method of habitat simulation modeling was used to study how water flow affects the conditions of the environment for aquatic life (Bovee 1982 ). The United States Wildlife Service developed the Physical Habitat Simulation (PHABSIM) model during the 1970s ( https://www.usgs.gov/node/279289 ). This model consists of a range of tools that can be used to simulate the appropriateness of hydraulic habitats for aquatic species (Stucchi and Bocchiola 2023 ). The simulation model for aquatic habitats gauges the standard of the habitat by taking into account physical alterations, like the depth, flow velocity, and channel conditions (Mwamila et al. 2008 ; Nagaya et al. 2008 ). The process of simulating a habitat involves two stages, namely habitat simulation and hydraulic simulation (Nikghalb et al. 2016 ; Spence and Hickley 2000 ). PHABSIM model was used for suitable ecological water demand (Miao et al. 2020 ; Dana et al. 2017 ), quantifying the hydrological requirements for fish (Weng et al. 2021 ), and modeling fish habitat condition in ice affected rivers (Knack et al. 2020 ). Nikghalb et al. 2016 conducted a comparison of two hydrologic techniques, namely Tennant and Q95, with a PHABSIM. The comparison was conducted under conditions of data shortage. The results showed that the PHABSIM model was able to provide valid outcomes even when the input data was imprecise. Peng and Sun ( 2016 ) utilized the PHABSIM models to prevent the river ecosystem from hydroelectric exploitation. They clarified that the PHABSIM models establish a relationship between habitat and flow that maximizes instream flow releases while reducing ecological disturbances. The outcome of the research indicates that the PHABSIM model is effective in optimizing the flow release schemes of hydropower stations in the southwestern region of China. Golami et al. (2020) estimated the EF for the wetland's river system using various models, including the PHABSIM model, the Tenant model, the Wetted-Perimeter method, and the flow duration curve. The study concluded that the Wetted-Perimeter and the PHABSIM model were the most effective in estimating the EF for the rivers in the wetland, considering the river's seasonality and hydro-climatic condition of the study area. Using PHABSIM model, Miao et al. ( 2020 ) showed that the land use change can decrease the urban water supply and the quality of fish habitat. Im et al. ( 2018 ) utilized fuzzy neural network models to create habitat-suitable indexes (HSI) which were then used as input data for the PHABSIM model. They showed that fuzzy neural network models are able to consider uncertainties in complex ecosystems. Recently, Adaptive Neuro-Fuzzy Inference System (ANFIS) was used in environmental studies (Zamanzad-Ghavidel et al. 2023 ; Emadi et al. 2022; Amanollahi and Ausati 2020a ) and it seems it is able to accurately predict the environmental variables (Ghasemi and Amanollahi 2019 ; Kaboodvandpour et al. 2015 ). Therefore, it seems that the ANFIS model has capability to predict the EF using HSI data. Thus, the purpose of this study is to simulate PHABSIM model to determine the EF of a river and predict the EF of the other rivers using linear model, Multilayer Perceptron neural network (MLP), as a nonlinear model, and ANFIS as a hybrid model. Materials and methods Study area The Pir Khezran River, located in the west of Kurdistan province, is being studied in this research. It is one of the sub-basins of Azad River (Fig. 1 ) and has an area of 4.70220 hectares with 632 mm rainfall. The basin is of significant importance to Azad Dam and is one of the sub-basins of Sirvan River in western Iran. Figure 1 Data The study's required data was gathered in January of 2018, with measurements taken across the river's cross-section. The hydrology team recorded water velocity, depth, and discharge (Bice et al. 2014 ), while the fisheries studies team gathered data on various fish species, including numbers, ages, sexual maturity, densities, and diversities (Kim et al. 2020 ). To determine channel index variables (Van et al. 2006), substrate materials were sampled from each location, with three samples collected from the bed floor. The channel index was then calculated based on the sizes of sand, gravel, and stones. One of the most important steps in EF modeling is the production of HSI data (Kim et al. 2020 ). The appropriate HSI shows living conditions of the index species at the sampling site (Zhang et al. 2020 ). For example, by analyzing the water depth variable, it is possible to identify the optimal depth range for the index species' living conditions (Sekine et al. 2020 ). This information can then be used to predict how the living conditions will be affected when the depth deviates from the optimal range (Wen et al. 2021 ). Data from all the studies mentioned was used to create habitat suitability curves. Estimation of EF using PHABSIM model The PHABSIM model is essentially a blend of hydraulic and habitat models. It establishes a connection between flow conditions and the appropriateness of physical habitats for index species in the river ecosystem (Miao et al. 2020 ). The method is divided into cells based on diversity. Each section has suitability functions for depth, velocity, and channel index that determine the cell's degree of suitability based on its current state (Kim et al. 2020 ). The suitability index of the composite for each cell in each section is determined by combining the calculated scores for each of the mentioned parameters. The Weight Usable Area (WUA) is the final index used to determine EF, which is calculated by multiplying the HSI for each cell at the relevant available level (Nikghalb et al. 2016 ). A WUA versus discharge map can be generated using the overall WUA amount for a particular section of a river. This map would illustrate the correlation between discharge and the quality and quantity of habitat available. The index is used as a guide to analyze alterations in habitats due to evacuation. It's essential to acknowledge that there have been objections regarding the correlation between fish biomass and the availability of habitat or WUA (Knack et al. 2020 ; Miao et al. 2020 ). Environmental water flow prediction models ANFIS model Jung's ANFIS method is an algorithm that merges fuzzy logic with an artificial neural network (ANN) for hybrid learning. This approach offers the benefits of both ANN and fuzzy models, allowing for the optimization of nonlinear problems and the acquisition of imprecise and ambiguous information (Im et al. 2018 ). The Fuzzy Inference System (FIS) was initially introduced by Mamdani and Assilian. Afterwards, Sugeno presented the Sugeno model which enhanced the computational efficiency of the previous FIS (Jung and Choi 2015 ). The fuzzy approach offers a significant benefit by allowing for the direct incorporation of non-numerical information, including specialized knowledge, into rules. This is particularly useful when there is a lack of field data. Furthermore, membership rules and functions can be defined in such a way as to consider the inherent uncertainty of environmental variables (Fraternali et al. 2012 ; Rinderknecht et al. 2012 ).The ANFIS method was tested using the observed data set for validation (Jung and Choi 2015 ). The construction of the HSI model has also been achieved by utilizing fuzzy neural network models (Fukuda 2009 ; Fukuda and Hiramatsub 2008) and the results indicate that the evaluation of ecosystems demands the use of fuzzy neural network models, particularly when dealing with intricate ecosystems (Jung and Choi 2015 ). The least squares method and multiplication algorithm are used to predict output values and optimize the parameters of membership functions in this approach. The Sugeno inference system is employed to enhance the efficiency of optimization. The ANFIS network is composed of five layers, which include the fuzzy layer, product layer, normalized layer, de-fuzzy layer, and output layer. Each layer's output variables are determined by the input variables from the previous layers and the parameters in each node (Im et al. 2018 ). Multilayer perceptron neural network MLP The most famous, adaptable, and uncomplicated form of artificial neural network is MLP. This approach is extensively employed to articulate the nonlinear correlation between anticipated and recorded information (Dawson and Wilby 2001 ). The primary objective of the neural network is to enhance the efficiency of the predicted and observed values. MLP is specifically engineered to excel in modeling nonlinear phenomena. An MLP network that is forward-facing comprises an input layer as well as an output layer, both of which are flanked by one or more hidden layers in the middle. Within each layer, there are a predetermined number of artificial neurons. linear regression model LR Regression analysis is a statistical technique used to investigate the connections between variables. It involves an independent variable and a dependent variable, with the researchers generally seeking to determine the impact of the independent variable Y on the dependent variable xi. All of these components, including the dependent and independent variables and the error, are part of the regression analysis process, and the resulting equation for prediction is commonly referred to as a regression model (Sykes 1993 ). It should be kept in mind that prior to analyzing the correlation between the dependent and independent variables, it is recommended to use a distribution diagram to assess the noteworthy correlation between them. Furthermore, it is important to determine whether there exists a significant association between these variables. The statistical technique of linear regression holds a crucial position in the realm of statistical methods. Results and discussion Estimating EF of Pir Khezran river Habitat suitability curves in January for Pir Khezran River and surrounding rivers are shown in Fig. 2 . Figure 2 As Fig. 2 shows, the appropriate flow velocity for the Pir Khezran River and the surrounding rivers in January is between 0.2 and 0.6 (m/s), i.e. in this range, the maximum rate of fishing and benthos and the highest rate of diversity were obtained. At the flow velocity higher or less than this range, habitat suitability for the index species gradually tends to zero. The appropriate flow depth according to the Fig. 2 is between 0.22 to 0.51 m, which, as stated for the flow velocity, in this depth range, it was the highest catch of index species. Habitat suitability decreases at very high depth of river but will not be zero because the index species studied in this research has also been seen in the Azad dam which has a depth of nearly 85 m, so the highest depth has 0.1 value in flow depth suitability index. Habitat suitability for the channel index is highest at 5 and tends to zero at lower and higher than this rate. Also, the appropriate temperature in the best biological condition was 7 to 10 ° C. In order to obtain the EF of the rivers of the study area in January, the PHABSIM model developed HIS indexes. As Fig. 3 a shows, the amount of EF in this river in January is equal to 2.2 m 3 /s. since then, the habitat suitability does not change and remains the same, but reducing the volume of water from this amount causes reduction in the habitat suitability. Figure 3 The desirability of habitats in different parts of Pir Khezran River for flow velocity, flow depth and channel index are shown separately below. Habitat suitability for the flow velocity variable is shown in Fig. 3 b. The legend section is stated and different numerical ranges are given for each color. As this figure shows, the habitat suitability in the left part of the river has the highest rate and moving from left to right reduces the habitat suitability. Figure 4 a shows the desirability of habitat at different flow depths in the Pir Khezran River in January. As this figure shows, most parts of the river are suitable for fish life. Figure 4 Figure 4 b indicates the habitat suitability for the indicator species based on a combination of the flow velocity and flow depth variables. In this figure, the left side of the river has good desirability and only in two parts of the left side, the desirability has decreased. It can be said that this situation is also observed in Figs. 3 b and 4 a. Figure 5 a shows the extent of habitat desirability based on channel index, in which the expression of the degree of desirability for the channel index is equal to 5, which is the highest. Then, this rate gradually tends to zero. As Fig. 5 a shows, all the different parts of the river have the highest habitat quality at this range, and therefore it can be said that the river does not limit the life of the indicator species at this range of channel index. Figure 5 Finally, the habitat suitability of the index species is modeled by PHABSIM model based on all the variables mentioned shown in Fig. 5 b. As this figure shows, Pir Khezran River has different ranges of habitat suitability for the indicator species. Predicting environmental water requirement using ANFIS, MLP and LR models In this study, the data were divided into 80% and 20% for training and testing phases, respectively. In order to eliminate the alignment effect on the modeling results, the VIF and Tolerance tests were performed and the results are shown in Table 1 . Table 1 Tolerance and VIF results of independent variables for estimation EFR Test V min V max D min D max Ch min Ch max T min T max VIF 4.418 4.638 4.328 4.863 4.856 4.871 5.500 5.700 Tolerance 0.208 0.216 0.202 0.206 0.010 0.211 0.185 0.171 Table 1 The results of Table 1 show that the appropriate water temperature for indicator species in both minimum and maximum conditions has a high correlation with other parameters, so these two variables were not used in the modeling. Mean, minimum, maximum and standard deviation of independent and dependent variables used in modeling are shown in Table 2 . Table 2 Descriptive statistics of independent and EFR data No. Min Max Mean SD V min 100 0.20 0.80 0.48 0.13 V max 100 0.30 0.97 0.60 0.15 D min 100 0.10 0.95 0.57 0.18 D max 100 0.22 1.35 0.70 0.22 Ch min 100 5.00 5.50 5.25 0.02 Ch max 100 6.00 6.50 6.25 0.03 EFR 100 7.00 3.20 2.10 0.49 Table 2 As Table 2 shows, by increasing the independent variables value to its maximum value, the minimum amount of EF increased. Rresults can indicate that there is a strong relationship between increasing EF and the increase of independent variables. However, the linearity or nonlinearity of this relationship is not yet clear which were tested using the linear model LR and nonlinear model such as MLP and hybrid model like ANFIS. Predicting the minimum environmental water requirement using the LR model The LR model is a linear regression model that has many applications in environmental studies (Kayes et al. 2019 ; Ceylan and Bulkan 2018 ; Karatzas et al. 2018 ). The obtained LR model is shown using independent and dependent variables in Eq. 1. Equation (1) EWF = 0.115–0.13 * (V min) -0.597 * (V max) + 0.155 * (D min) + 0.197 *(D max) + 0.01 * (Ch min) + 0.02 (Ch max) As Eq. 1 shows, the minimum EF has an inverse relationship with the flow velocity at both the lowest and highest levels and also has a direct relationship with the flow depth and channel index at both the lowest and highest levels. The results of the LR model training phase are shown in Fig. 6 a. Figure 6 The correlation between the observed and simulated data is equal to R 2 = 0.67 and also RMSE = 0.1301 and MAE = 0.0191. As Fig. 6 a shows, there is not high correlation between the observed and simulated data. Figure 6 b shows the results of the test phase of the LR model in which R 2 = 0.81, RMSE = 0.1233 and MAE = 0.0152. As Fig. 6 b shows, the correlation between the observed and predicted data at this phase is greater than at the training phase. Predicting the minimum environmental water requirement using the MLP model The MLP model is one of the most common artificial neural network models that is widely used in natural resource studies (Mishra and Goyal 2015 ; Mishra and Goyal 2016 ; Rahimi 2017 ). The results of MLP model in the training stage of determining the EF of Pir Khezran River are shown in Fig. 7 a, where R 2 = 0.96, RMSE = 0.0370 and MAE = 0.0014. As Fig. 7 a shows, the correlation coefficient between the observed and simulated data in the MLP model is greater than the results of the LR model in training phase. Figure 7 The results of the testing phase of the MLP model (R 2 = 0.92, RMSE = 0.0026 and RME = 0.0027) in predicting the EF of Pir Khezran River are shown in Fig. 7 b. According to the Fig. 7 b, the correlation coefficient between the observed and predicted data at this phase is more than the results of the LR model, which are consistent with the results (Abba et al. 2017 ; Abba and Elkirn 2017 ; Mohammadi et al. 2019 ). Prediction of minimum environmental water requirement using ANFIS model As stated in the methodology section, the ANFIS model is a hybrid nonlinear model that has more advantages than the nonlinear models (Amanollahi and Ausati 2020b ; Ghasemi and Amanollahi 2018, Kaboodvandpour et al. 2015 ). Figure 8 a indicates the results of the ANFIS model in training phase (R 2 = 0.98, RMSE = 0.0248 and MAE = 0.0006) to determine the EF. Figure 8 Comparison of Fig. 8 a with the previous figures related to the models training phases (LR and MLP models) shows that the correlation coefficient of ANFIS model is the highest. The results of the testing phase of the ANFIS model (R 2 = 0.97, RMSE = 0.0295 and MAE = 0.0008) in predicting the EF of the Pir Khezran River are shown in Fig. 8 b. Comparison of the results of ANFIS model in the testing phase compared to the results of the other two models shows that the ANFIS model has a higher accuracy in predicting the amount of EF in the Pir Khezran River. Conclusion In the present study, the accuracy of linear, nonlinear and hybrid models in determining the EF of Pir Khezran River was evaluated. The PHABSIM model was used to estimate EF based on HSI and discharge of the Pir Khezran river. The PHABSIM model needs to provide accurate data which is time consuming and costly. In this study, EF data obtained by PHABSIM model were used as an input of ANFIS, MLP and LR models. Results indicate that nonlinear models have acceptable accuracy compared to the linear model. Results showed that the ANFIS model as a hybrid model has better performance and accuracy than the MLP model in predicting EF. According to the ANFIS model result, it can be concluded that in the basin with the same ecological rivers condition, the PHABSIM model can be used to estimate the EF of one river and the ANFIS model can be utilized to predict the EF of other rivers. Declarations Funding statement The authors declare that the funding is not available. Conflicts of interest/Competing interests The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Availability of data and material All of the data and material are owned by the authors and/or no permissions are required. All authors (Seiran Haghgoo, Jamil Amanollahi, Barzan Bahrami Kamangar) have read, understood, and have complied as applicable with the statement on “Ethical responsibilities of Authors” as found in the Instructions for Authors and are aware that with minor exceptions, no changes can be made to authorship once the paper is submitted. Code availability Not applicable Author contributions Seiran Haghgoo: Conceptualization, Formal analysis, Writing - original draft. Jamil Amanollahi: Conceptualization, Supervision, Methodology, Resources, Project administration, Validation, Writing - review & editing; Barzan Bahrami Kamangar: Conceptualization, Methodology, Investigation, Formal analysis; Ethics approval statement Not applicable Consent to participate Not applicable Consent for publication Not applicable References Aazami J, Motevalli A, Savabieasfahani M (2022) Correction to: Evaluation of three environmental flow techniques in Shoor wetland of Golpayegan, Iran. Int. J. Environ. Sci. Technol. 19 :7899. Abba S, Hadi SJ, Abdullahi J (2017) River water modelling prediction using multi linear regression, artificial neural network, and adaptive neuro-fuzzy inference system techniques. . Procedia Comput. Sci. 120:75-82. Abba SI, Elkirn G (2017) Effluent prediction of chemical oxygen demand from the astewater treatment plant using artificial neural network application. Procedia Comput. Sci. 120:156-163. Amanollahi J, Ausati Sh (2020a) PM 2.5 concentration forecasting using ANFIS, EEMD-GRNN, MLP, and MLR models: A case study of Tehran, Iran. Air Qual. Atmos. Health. 13:161-17. Amanollahi J, Ausati Sh (2020b). Validation of linear, nonlinear, and hybrid models for predicting particulate matter concentration in Tehran. Iran. Theor. Appl. Climatol. 140:709-717. Bejarano MD, Sord-Ward A, Gabriel-Martin I, Garrote L (2019) Tradeoff between economic and environmental costs and benefits of hydropower production at run-of-river-diversion schemes under different environmental flows scenarios. J. Hydrol. 572:790-804. Belmar O, Velasco J, Martinez-Capel F (2011) Hydrological Classification of Natural Flow Regimes to Support Environmental Flow Assessments in Intensively Regulated Mediterranean Rivers, Segura River Basin (Spain). Environ. Manage. 47 :992–1004. Bice CM, Gehrig SL, Zampatti BP et al. (2014) Flow-induced alterations to fish assemblages, habitat and fish–habitat associations in a regulated lowland river. Hydrobiologia. 722 :205–222. Bouska K (2020) Regime change in a large-floodplain river ecosystem: patterns in body-size and functional biomass indicate a shift in fish communities. Biol. Invasions. 22 :3371–3389. Bovee KD (1982) Instream flow methodology. US Fish and Wildlife Service. FWS/OBS, 82, 26 Caudill, M. and Butler, C. 1992. Understanding Neural Networks. Basic Networks, vol. 1. MIT Press, Cambridge, MA. Ceylan Ζ, Bulkan S (2018) Forecasting PM 10 levels using ANN and MLR: A case study for Sakarya City. Global Nest J. 20:281-290. Dana K, Jana C, Tomáš V (2017) Analysis of environmental flow requirements for macroinvertebrates in a creek affected by urban drainage (Prague metropolitan area, Czech Republic). Urban Ecosyst 20 :785–797. Dawson,CW, Wilby R L (2001) Hydrological modelling using artificial neural networks, Progr Phys, Geogr. 25:80–108. Dumitriu D (2020) Sediment flux during flood events along the Trotuș River channel: hydrogeomorphological approach. J. Soils Sediments . 20 :4083–4102. Emadi A, Sobhani R, Ahmadi H et al. Multivariate modeling of agricultural river water abstraction via novel integrated-wavelet methods in various climatic conditions. Environ. Dev. Sustain. 24 :4845–4871. Espinoza T, Burke CL, Carpenter-Bundhoo L, Marshal SM, McDougall AJ, Roberts DT, Campbell HA, Kennard MJ (2021) Quantifying movement of multiple threatened species to inform adaptive management of environmental flows. J. Environ. Managet. 295:113067. Fraternali P, Castelletti A, Soncini-Sessa R, Vaca Ruiz C, Rizzoli AE (2012) Putting humans in the loop: social computing for Water Resources Management. Environ. Modell. Softw. 37:68-77 Fukuda S (2009) Consideration of fuzziness: Is it necessary in modelling fish habitat preference of Japanese medaka (Oryzias latipes)? Ecol. Model. 220:2877-2884. Fukuda S, Hiramatsu B (2008) Prediction ability and sensitivity of artificial intelligence-based habitat preference models for predicting spatial distribution of Japanese medaka (Oryzias latipes). Ecol. Model. 215:301-313. Gain AK, Apel H, Renaud FG et al. (2013) Thresholds of hydrologic flow regime of a river and investigation of climate change impact—the case of the Lower Brahmaputra river Basin. Climatic Change. 120 :463–475. Ghasemi A, Amanollahi J (2019) Integration of ANFIS model and forward selection method for air quality forecasting. Air Qual. Atmos. Health. 12:59-72. Gholami V, Khalili A, Sahour H et al. (2020) Assessment of environmental water requirement for rivers of the Miankaleh wetland drainage basin. Appl. Water Sci. 10 :233 Im D, Choi SU, Choi B (2018) Physical habitat simulation for a fish community using the ANFIS method. Ecol. Inform. 43:73-83. IUCN (World Conservation Union) (2000). Vision for Water and Nature. A World Strategy for conservation and sustainable management of water resources in the 21st Century. IUCN, Gland, Switzerland and Cambridge, UK, p 52. Jung SH, Choi S-Uk (2015) Prediction of composite suitability index for physical habitatsimulations using the ANFIS method. Appl. Soft Comput. 34:502–512. Kaboodvandpour S, Amanollahi J, Qhavami S, Mohammadi B (2015) Assessing the accuracy of multiple regressions, ANFIS, and ANN models in predicting dust storm occurrences in Sanandaj, Iran. Nat. Hazards. 78:879-893. Karatzas K, Katsifarakis N, Orlowski C, Sarzynski A (2018) Revisiting urban air quality forecasting: a regression approach. Vietnam J. Comput. Sci. 5:177-184. Kayes I, Shahriar SA, Hasan K, Akhter M, Kabir MM, Salam MA (2019) The relationships between meteorological parameters and air pollutants in an urban environment. Global J. Environ. Sci. Manag. 5:265-278. Kim HJ, Kim JH, Ji U et al. (2020) Effect of Probability Distribution-Based Physical Habitat Suitability Index on Environmental-Flow Estimation. KSCE J. Civ. Eng. 24 :2393–2402. Knack IM, Huang F, Shen HT (2020) Modeling fish habitat condition in ice affected rivers. Cold Reg. Sci. Technol. 176:103086. Miao Y, Li J, Feng P, Dong L, Zhang T, Wu J, Katwal R (2020) Effects of land use changes on the ecological operation of the Panjiakou-Daheiting Reservoir system, China. Ecol. Eng. 152:10585 Mishra D, Goyal P (2015) Development of artificial intelligence based NO 2 forecasting models at Taj Mahal, Agra. Atmos. Pollut. Res. 6:99-106. Mishra D, Goyal P (2016) Neuro-Fuzzy approach to forecasting ozone episodes over the urban area of Delhi, India. Environ. Technol. Inno. 5:83-94. Mlynski D, Operacz A, Walega A (2020) Sensitivity of methods for calculating environmental flows based on hydrological characteristics of watercourses regarding te hydropower potential of rivers. J. Clean. Pro. 250:119527. Mohammadi S, Naseri F, Abri R (2019) Simulating soil loss rate in Ekbatan Dam watershed using experimental and statistical approaches. Int. J. Sediment. Res. 34:226-239. Mwamila TB, Kimwaga RJ, Mtalo FW (2008) Eco-hydrology of the Pangani River downstream of Nyumba ya Mungu reservoir, Tanzania. Phys. Chem. Earth. 33:695-700. Nagaya T, Shiraishi Y, Onitsuka K, Higashino M, Takami T, Otsuka N, Akiyama J, Ozeki H (2008) Evaluation of suitable hydraulic conditions for spawning of ayu with horizontal 2D numerical simulation and PHABSIM. Ecol. Modell. 215:133-143. Nasiri Khiavi A, Mostafazadeh R Ghanbari Talouki F (2024) Using game theory algorithm to identify critical watersheds based on environmental flow components and hydrological indicators. Environ. Dev. Sustain. https://doi-org.access.semantak.com/10.1007/s10668-023-04390-8 Nikghalb S, Shokoohi A, Singh VP et al (2016) Ecological Regime versus Minimum Environmental Flow:Comparison of Results for a River in a Semi Mediterranean Region. Water Resour. Manage. 30 :4969–4984. Pandey CL (2021) Managing urban water security: challenges and prospects in Nepal. Environ. Dev. Sustain. 23 :241–257. Peng L, Sun L (2016) Minimum instream flow requirement for the water-reduction section of diversion-type hydropower station: a case study of the Zagunao River, China. Environ. Earth Sci. 75 :1210 Perez-Blanco CD, Gil-Garcia L, Saiz-Santiago (2021) An actionable hydroeconomic descision support system for the assessment of water reallocations in irrigated agriculture. Astudy of minimum environmental flows in the Douro River Basin, Spain. J. Environ. Managet. 298:113432. Rahimi A (2017) Short-term prediction of NO 2 and NO x concentrations using multilayer perceptron neural network: a case study of Tabriz, Iran. Ecol. Process. 6: 4. Rinderknecht SL, Borsuk ME, Reichert P (2012) Bridging uncertain and ambiguous knowledge with imprecise probabilities. Environ. Modell. Softw. 36:122-130. Sedighkia M, Abdoli A (2023) Design of optimal environmental flow regime at downstream of multireservoir systems by a coupled SWAT-reservoir operation optimization method. Environ Dev Sustain. 25 :834–854. Sedighkia M, Datta B (2023) Analyzing environmental flow supply in the semi-arid area through integrating drought analysis and optimal operation of reservoir. J. Arid Land 15 :1439–1454. Sekine M, Wang J, Yamamoto K et al. (2020) Fish habitat evaluation based on width-to-depth ratio and eco-environmental diversity index in small rivers. Environ Sci. Pollut. Res. 27 :34781–34795. Senent-Aparicio J, George C, Srinivasan R (2021) Introducing a new post-processing tool for the SWAT+model to evaluate environmental flows. Environ. Modell. Softw. 136:104944. Spence R, Hickley P (2000) The use of PHABSIM in the management of water resources and fisheries in England and Wales. Ecol. Eng. 16:153-158. Stucchi L. Bocchiola D (2023) Environmental Flow Assessment using multiple criteria: A case study in the Kumbih river, West Sumatra (Indonesia). Sci. Total Environ. 901:166516. Sykes AO (1993) An Introduction to Regression Analysis. Am. Stat, 61, 101. Van der Lee GEM, Van der Molen DT, Van den Boogaard HFP et al. (2006) Uncertainty analysis of a spatial habitat suitability model and implications for ecological management of water bodies. Landscape Ecol. 21 :1019–1032. Vivian LM, Godfree RC, Colloff MJ et al. (2014) Wetland plant growth under contrasting water regimes associated with river regulation and drought: implications for environmental water management. Plant Ecol. 215 :997–1011. Wang H, Cong P, Zhu Z, Zhang W, Ai Y, Huai W (2022) Analysis of environmental dispersion in wetland flows with floating vegetation islands. J. Hydrol. 606:127359. Wen X, Lv Y, Liu Z, Ding Z, Lei X, Tan Q, S un Y (2021) Operation chart optimization of multi-hydropower system incorporating the long- and short-term fish habitat requirements. J. Clen. Pro. 121:107053. Weng X, Jiang C, Yuan M, Zhang M, Zeng T, Jin C (2021) An ecologically dispatch strategy using environmental flows for a cascade multi-sluice system: A case study of the Yongjiang River Basin, China. Ecol. Indic. 121:107053. Yu L, Wu X, Wu S, Jia B, Han G, Xu P, Dai J, Zhang Y, Wang F, Yang Q, Zhou Z (2021) Multi-objective optimal operation of cascade hydropower plants considering ecological flow under different ecological conditions. J. Hydrol. 601:126599. Zamanzad-Ghavidel S, Fazeli S, Mozaffari S et al . (2023) Estimating of aqueduct water withdrawal via a wavelet-hybrid soft-computing approach under uniform and non-uniform climatic conditions. Environ. Dev. Sustain . 25 :5283–5314. Zhang Y, Yu H, Yu H. et al. (2020) Optimization of environmental variables in habitat suitability modeling for mantis shrimp Oratosquilla oratoria in the Haizhou Bay and adjacent waters. Acta Oceanol. Sin. 39 :36–47. Zheng Y, Tian Y, Du E, Han F, Wu Y, Zheng C, Li X (2020) Addressing the water conflict between agriculture and ecosystems under environmental flow regulation: An integrated modeling study. Environ. Modell. Softw. 134:104874. Zolfagharpour F, Saghafian B, Delavar M (2021) Adapting reservoir operation rules to hydrological drought state and environmental flow requirements. J. Hydrol. 660:126581. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3939514","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":272736532,"identity":"41332182-6eda-4b2b-a04a-8db9838f6782","order_by":0,"name":"Seiran Haghgoo","email":"","orcid":"","institution":"University of Kurdistan","correspondingAuthor":false,"prefix":"","firstName":"Seiran","middleName":"","lastName":"Haghgoo","suffix":""},{"id":272736533,"identity":"b18c52da-c6a3-4347-be50-2244b59b9b8f","order_by":1,"name":"Jamil Amanollahi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7ElEQVRIiWNgGAWjYBAC9gYILcPHwMNw4AMDgwGYm4BHCyNUCw8bUMvBGSRrYeaBacEHGBuYn274ucOGh4397MHDtm02xgzshx8wPNyDTwub2c3eM2k8bDx5CYdz29LMGHjSDBgSnuF1mNkN3rbDQIflGAC1HLZhYMgB+uUAPi3s327+BWnhf2Nw2BKkhf8Nfi2CDTxmt8G2SABtYWw7bMYgQcAWaWaestuybUC/SLwxONhzLs2YTeKZwQF8WvjY27fdfNtmI8fPn2P84UeZjWE/f/LDhz/waGFgRhdgA2J8GkbBKBgFo2AUEAEApF1KGemRLVwAAAAASUVORK5CYII=","orcid":"","institution":"University of Kurdistan","correspondingAuthor":true,"prefix":"","firstName":"Jamil","middleName":"","lastName":"Amanollahi","suffix":""},{"id":272736534,"identity":"c5a3d484-e072-48cd-996c-39fdfeb49194","order_by":2,"name":"Barzan Bahrami Kamangar","email":"","orcid":"","institution":"University of Kurdistan","correspondingAuthor":false,"prefix":"","firstName":"Barzan","middleName":"Bahrami","lastName":"Kamangar","suffix":""}],"badges":[],"createdAt":"2024-02-08 09:55:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3939514/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3939514/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":51176917,"identity":"9fb40374-ede8-4b49-81cf-e81185ce916d","added_by":"auto","created_at":"2024-02-15 13:11:33","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":712555,"visible":true,"origin":"","legend":"\u003cp\u003eGeographical location of Pir Khezran river\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-3939514/v1/8f51ef7fbe281a7490b40dfb.png"},{"id":51176920,"identity":"0f026b2c-c211-457f-8939-c995767a67c0","added_by":"auto","created_at":"2024-02-15 13:11:33","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":331924,"visible":true,"origin":"","legend":"\u003cp\u003eFlow Depth, flow velocity, channel index and suitable temperature for the life of index species in Pir Khezran River in January\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-3939514/v1/403fb2733013e9a233da2d9b.png"},{"id":51176918,"identity":"9110924e-5e05-47b1-a3aa-934d92c827b3","added_by":"auto","created_at":"2024-02-15 13:11:33","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":144457,"visible":true,"origin":"","legend":"\u003cp\u003e(a) EFR in Pirkhezran river in January, (b) Habitat suitability at the flow velocity of 2.2 m\u003csup\u003e3\u003c/sup\u003e/s in Pir Khezran river in January\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-3939514/v1/43cdb17e0eff884cca7c0e60.png"},{"id":51176916,"identity":"f38895b9-a5de-4a32-880e-aabb441d2f42","added_by":"auto","created_at":"2024-02-15 13:11:32","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":556277,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Habitat suitability at flow depths between 0.22 to 0.51 m of Pirkhezran river in January, (b) Habitat suitability based on a combination of flow velocity and flow depth variables in Pir Khezran river in January.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-3939514/v1/d6f8574f71d087e370eeec23.png"},{"id":51176915,"identity":"fccc4a2b-0771-4f56-9cb2-726a9b97b604","added_by":"auto","created_at":"2024-02-15 13:11:32","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":158441,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Habitat suitability in Pir Khezran river based on substrate index, (b) Different desirability of Pir Khezran river at the 2.2 m\u003csup\u003e3\u003c/sup\u003e/s of EFR\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-3939514/v1/de6dc04efd3a8d4432dc6b9c.png"},{"id":51176921,"identity":"df111725-8d95-429d-b02a-db36ee22d539","added_by":"auto","created_at":"2024-02-15 13:11:34","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":134122,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Observed and simulated data of EFR in LR model, (b) Observed and predicted data of EFR in LR model.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-3939514/v1/1e1da1cdc9aee10fc7a60353.png"},{"id":51176914,"identity":"b72a0c16-bf97-4883-87e6-4c7d15ed3e97","added_by":"auto","created_at":"2024-02-15 13:11:32","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":156102,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Observed and simulated data of EFR in the MLP model, (b) Observed and predicted data of EFR in MLP model.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-3939514/v1/65255be5704156c804115246.png"},{"id":51176919,"identity":"45ffb251-7419-41d3-89e6-e468839a6577","added_by":"auto","created_at":"2024-02-15 13:11:33","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":126992,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Observed and modeled data on environmental water demand in ANFIS model, (b) Observed and predicted data of EFR in the testing phase of ANFIS model.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-3939514/v1/995ad1efe241cd12b3b63d9d.png"},{"id":52817491,"identity":"5076821c-0685-46c4-844b-360deb480cec","added_by":"auto","created_at":"2024-03-16 13:59:48","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2071222,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3939514/v1/b4544d72-1673-40bb-896a-1b86bec9cfb2.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Reducing the computational cost and time of environmental flow estimation based on machine learning approaches","fulltext":[{"header":"Introduction","content":"\u003cp\u003ePopulation growth and rapid growth of different aspects of cities have increased the demand for water (Pandey \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), which has created challenges and changes in river basins that are necessary to maintain ecosystems (IUCN 2000). In countries where freshwater resources are limited to meet existing needs, they are forced to strike a balance between water extracted for human activities and water stored for a river or watershed management system (Perez-Blanco et al. \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Senent-Aparicio et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Zolfagharpour et al. \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2021\u003c/span\u003e ). In most developing countries, due to the lack of minimum required information about rivers, there are no comprehensive and flexible methods to determine the environmental flow (EF) (Sedighkia and Datta \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Aazami et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Hence, managing and maintaining the ecosystem of such rivers seem very difficult (Wang et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Shang et al. 2021). EF is the water regime provided for a river, wetland or coastal area in order to protect ecosystems and their benefits (Espinoza et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Zheng et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The flow regime, according to the IFC definition, is the hydrological feature of a river and includes four flow levels, including life flow of period of drought (Vivian et al \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2014\u003c/span\u003e), basal flow as a natural flow to maintain river habitat in healthy conditions (Gain et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Belmar et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), high flow that removes sediment in the environment (Dumitriu \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and overflow bank that connects the main river to the floodplain, which is morphologically very significant (Bouska \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Anonymous 2005). The purpose of various methods designed to preserve river ecosystems is to maintain one or more parts of the mentioned flow regimes (Nasiri Khiavi et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Sedighkia and Abdoli \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The compromise between economics benefit of rivers such as hydropower potential and environmental protection must be considered in estimate the EF (Yu et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Mlynski et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Bejarano et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). There are different divisions in order to determine the EF of rivers. The numerical method of habitat simulation modeling was used to study how water flow affects the conditions of the environment for aquatic life (Bovee \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e1982\u003c/span\u003e). The United States Wildlife Service developed the Physical Habitat Simulation (PHABSIM) model during the 1970s (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.usgs.gov/node/279289\u003c/span\u003e\u003cspan address=\"https://www.usgs.gov/node/279289\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). This model consists of a range of tools that can be used to simulate the appropriateness of hydraulic habitats for aquatic species (Stucchi and Bocchiola \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The simulation model for aquatic habitats gauges the standard of the habitat by taking into account physical alterations, like the depth, flow velocity, and channel conditions (Mwamila et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Nagaya et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2008\u003c/span\u003e ). The process of simulating a habitat involves two stages, namely habitat simulation and hydraulic simulation (Nikghalb et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Spence and Hickley \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). PHABSIM model was used for suitable ecological water demand (Miao et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Dana et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), quantifying the hydrological requirements for fish (Weng et al. \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and modeling fish habitat condition in ice affected rivers (Knack et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Nikghalb et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2016\u003c/span\u003e conducted a comparison of two hydrologic techniques, namely Tennant and Q95, with a PHABSIM. The comparison was conducted under conditions of data shortage. The results showed that the PHABSIM model was able to provide valid outcomes even when the input data was imprecise. Peng and Sun (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) utilized the PHABSIM models to prevent the river ecosystem from hydroelectric exploitation. They clarified that the PHABSIM models establish a relationship between habitat and flow that maximizes instream flow releases while reducing ecological disturbances. The outcome of the research indicates that the PHABSIM model is effective in optimizing the flow release schemes of hydropower stations in the southwestern region of China. Golami et al. (2020) estimated the EF for the wetland's river system using various models, including the PHABSIM model, the Tenant model, the Wetted-Perimeter method, and the flow duration curve. The study concluded that the Wetted-Perimeter and the PHABSIM model were the most effective in estimating the EF for the rivers in the wetland, considering the river's seasonality and hydro-climatic condition of the study area. Using PHABSIM model, Miao et al. (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) showed that the land use change can decrease the urban water supply and the quality of fish habitat. Im et al. (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) utilized fuzzy neural network models to create habitat-suitable indexes (HSI) which were then used as input data for the PHABSIM model. They showed that fuzzy neural network models are able to consider uncertainties in complex ecosystems. Recently, Adaptive Neuro-Fuzzy Inference System (ANFIS) was used in environmental studies (Zamanzad-Ghavidel et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Emadi et al. 2022; Amanollahi and Ausati \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2020a\u003c/span\u003e) and it seems it is able to accurately predict the environmental variables (Ghasemi and Amanollahi \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Kaboodvandpour et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Therefore, it seems that the ANFIS model has capability to predict the EF using HSI data. Thus, the purpose of this study is to simulate PHABSIM model to determine the EF of a river and predict the EF of the other rivers using linear model, Multilayer Perceptron neural network (MLP), as a nonlinear model, and ANFIS as a hybrid model.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy area\u003c/h2\u003e \u003cp\u003eThe Pir Khezran River, located in the west of Kurdistan province, is being studied in this research. It is one of the sub-basins of Azad River (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and has an area of 4.70220 hectares with 632 mm rainfall. The basin is of significant importance to Azad Dam and is one of the sub-basins of Sirvan River in western Iran.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eData\u003c/h2\u003e \u003cp\u003eThe study's required data was gathered in January of 2018, with measurements taken across the river's cross-section. The hydrology team recorded water velocity, depth, and discharge (Bice et al. \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2014\u003c/span\u003e), while the fisheries studies team gathered data on various fish species, including numbers, ages, sexual maturity, densities, and diversities (Kim et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). To determine channel index variables (Van et al. 2006), substrate materials were sampled from each location, with three samples collected from the bed floor. The channel index was then calculated based on the sizes of sand, gravel, and stones. One of the most important steps in EF modeling is the production of HSI data (Kim et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The appropriate HSI shows living conditions of the index species at the sampling site (Zhang et al. \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). For example, by analyzing the water depth variable, it is possible to identify the optimal depth range for the index species' living conditions (Sekine et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). This information can then be used to predict how the living conditions will be affected when the depth deviates from the optimal range (Wen et al. \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Data from all the studies mentioned was used to create habitat suitability curves.\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section3\"\u003e \u003ch2\u003eEstimation of EF using PHABSIM model\u003c/h2\u003e \u003cp\u003eThe PHABSIM model is essentially a blend of hydraulic and habitat models. It establishes a connection between flow conditions and the appropriateness of physical habitats for index species in the river ecosystem (Miao et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The method is divided into cells based on diversity. Each section has suitability functions for depth, velocity, and channel index that determine the cell's degree of suitability based on its current state (Kim et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The suitability index of the composite for each cell in each section is determined by combining the calculated scores for each of the mentioned parameters. The Weight Usable Area (WUA) is the final index used to determine EF, which is calculated by multiplying the HSI for each cell at the relevant available level (Nikghalb et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). A WUA versus discharge map can be generated using the overall WUA amount for a particular section of a river. This map would illustrate the correlation between discharge and the quality and quantity of habitat available. The index is used as a guide to analyze alterations in habitats due to evacuation. It's essential to acknowledge that there have been objections regarding the correlation between fish biomass and the availability of habitat or WUA (Knack et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Miao et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003eEnvironmental water flow prediction models\u003c/h2\u003e \u003cdiv id=\"Sec7\" class=\"Section4\"\u003e \u003ch2\u003eANFIS model\u003c/h2\u003e \u003cp\u003eJung's ANFIS method is an algorithm that merges fuzzy logic with an artificial neural network (ANN) for hybrid learning. This approach offers the benefits of both ANN and fuzzy models, allowing for the optimization of nonlinear problems and the acquisition of imprecise and ambiguous information (Im et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The Fuzzy Inference System (FIS) was initially introduced by Mamdani and Assilian. Afterwards, Sugeno presented the Sugeno model which enhanced the computational efficiency of the previous FIS (Jung and Choi \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The fuzzy approach offers a significant benefit by allowing for the direct incorporation of non-numerical information, including specialized knowledge, into rules. This is particularly useful when there is a lack of field data. Furthermore, membership rules and functions can be defined in such a way as to consider the inherent uncertainty of environmental variables (Fraternali et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Rinderknecht et al. \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2012\u003c/span\u003e).The ANFIS method was tested using the observed data set for validation (Jung and Choi \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The construction of the HSI model has also been achieved by utilizing fuzzy neural network models (Fukuda \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Fukuda and Hiramatsub 2008) and the results indicate that the evaluation of ecosystems demands the use of fuzzy neural network models, particularly when dealing with intricate ecosystems (Jung and Choi \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The least squares method and multiplication algorithm are used to predict output values and optimize the parameters of membership functions in this approach. The Sugeno inference system is employed to enhance the efficiency of optimization. The ANFIS network is composed of five layers, which include the fuzzy layer, product layer, normalized layer, de-fuzzy layer, and output layer. Each layer's output variables are determined by the input variables from the previous layers and the parameters in each node (Im et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003eMultilayer perceptron neural network MLP\u003c/h2\u003e \u003cp\u003eThe most famous, adaptable, and uncomplicated form of artificial neural network is MLP. This approach is extensively employed to articulate the nonlinear correlation between anticipated and recorded information (Dawson and Wilby \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). The primary objective of the neural network is to enhance the efficiency of the predicted and observed values. MLP is specifically engineered to excel in modeling nonlinear phenomena. An MLP network that is forward-facing comprises an input layer as well as an output layer, both of which are flanked by one or more hidden layers in the middle. Within each layer, there are a predetermined number of artificial neurons.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003elinear regression model LR\u003c/h2\u003e \u003cp\u003eRegression analysis is a statistical technique used to investigate the connections between variables. It involves an independent variable and a dependent variable, with the researchers generally seeking to determine the impact of the independent variable Y on the dependent variable xi. All of these components, including the dependent and independent variables and the error, are part of the regression analysis process, and the resulting equation for prediction is commonly referred to as a regression model (Sykes \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e1993\u003c/span\u003e). It should be kept in mind that prior to analyzing the correlation between the dependent and independent variables, it is recommended to use a distribution diagram to assess the noteworthy correlation between them. Furthermore, it is important to determine whether there exists a significant association between these variables. The statistical technique of linear regression holds a crucial position in the realm of statistical methods.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Results and discussion","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eEstimating EF of Pir Khezran river\u003c/h2\u003e \u003cp\u003eHabitat suitability curves in January for Pir Khezran River and surrounding rivers are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003c/p\u003e \u003cp\u003eAs Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows, the appropriate flow velocity for the Pir Khezran River and the surrounding rivers in January is between 0.2 and 0.6 (m/s), i.e. in this range, the maximum rate of fishing and benthos and the highest rate of diversity were obtained. At the flow velocity higher or less than this range, habitat suitability for the index species gradually tends to zero. The appropriate flow depth according to the Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e is between 0.22 to 0.51 m, which, as stated for the flow velocity, in this depth range, it was the highest catch of index species.\u003c/p\u003e \u003cp\u003eHabitat suitability decreases at very high depth of river but will not be zero because the index species studied in this research has also been seen in the Azad dam which has a depth of nearly 85 m, so the highest depth has 0.1 value in flow depth suitability index. Habitat suitability for the channel index is highest at 5 and tends to zero at lower and higher than this rate. Also, the appropriate temperature in the best biological condition was 7 to 10 \u0026deg; C. In order to obtain the EF of the rivers of the study area in January, the PHABSIM model developed HIS indexes. As Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea shows, the amount of EF in this river in January is equal to 2.2 m\u003csup\u003e3\u003c/sup\u003e/s. since then, the habitat suitability does not change and remains the same, but reducing the volume of water from this amount causes reduction in the habitat suitability.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e\u003c/p\u003e \u003cp\u003eThe desirability of habitats in different parts of Pir Khezran River for flow velocity, flow depth and channel index are shown separately below. Habitat suitability for the flow velocity variable is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb. The legend section is stated and different numerical ranges are given for each color. As this figure shows, the habitat suitability in the left part of the river has the highest rate and moving from left to right reduces the habitat suitability. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea shows the desirability of habitat at different flow depths in the Pir Khezran River in January. As this figure shows, most parts of the river are suitable for fish life.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb indicates the habitat suitability for the indicator species based on a combination of the flow velocity and flow depth variables. In this figure, the left side of the river has good desirability and only in two parts of the left side, the desirability has decreased. It can be said that this situation is also observed in Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea shows the extent of habitat desirability based on channel index, in which the expression of the degree of desirability for the channel index is equal to 5, which is the highest. Then, this rate gradually tends to zero. As Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea shows, all the different parts of the river have the highest habitat quality at this range, and therefore it can be said that the river does not limit the life of the indicator species at this range of channel index.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e\u003c/p\u003e \u003cp\u003eFinally, the habitat suitability of the index species is modeled by PHABSIM model based on all the variables mentioned shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb. As this figure shows, Pir Khezran River has different ranges of habitat suitability for the indicator species.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003ePredicting environmental water requirement using ANFIS, MLP and LR models\u003c/h2\u003e \u003cp\u003eIn this study, the data were divided into 80% and 20% for training and testing phases, respectively. In order to eliminate the alignment effect on the modeling results, the VIF and Tolerance tests were performed and the results are shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eTolerance and VIF results of independent variables for estimation EFR\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTest\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eV min\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eV max\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eD min\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eD max\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCh min\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCh max\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eT min\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eT max\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVIF\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4.418\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.638\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.328\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e4.863\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e4.856\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e4.871\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e5.500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e5.700\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTolerance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.208\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.216\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.202\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.206\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.010\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.211\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.185\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.171\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u003c/p\u003e \u003cp\u003eThe results of Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e show that the appropriate water temperature for indicator species in both minimum and maximum conditions has a high correlation with other parameters, so these two variables were not used in the modeling. Mean, minimum, maximum and standard deviation of independent and dependent variables used in modeling are shown in Table \u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDescriptive statistics of independent and EFR data\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMin\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMax\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSD\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eV min\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.13\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eV max\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.15\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eD min\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.18\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eD max\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.22\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCh min\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e5.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCh max\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e6.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEFR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.49\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003c/p\u003e \u003cp\u003eAs Table \u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows, by increasing the independent variables value to its maximum value, the minimum amount of EF increased. Rresults can indicate that there is a strong relationship between increasing EF and the increase of independent variables. However, the linearity or nonlinearity of this relationship is not yet clear which were tested using the linear model LR and nonlinear model such as MLP and hybrid model like ANFIS.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003ePredicting the minimum environmental water requirement using the LR model\u003c/h2\u003e \u003cp\u003eThe LR model is a linear regression model that has many applications in environmental studies (Kayes et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Ceylan and Bulkan \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Karatzas et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The obtained LR model is shown using independent and dependent variables in Eq.\u0026nbsp;1.\u003c/p\u003e \u003cp\u003eEquation (1) EWF\u0026thinsp;=\u0026thinsp;0.115\u0026ndash;0.13 * (V min) -0.597 * (V max)\u0026thinsp;+\u0026thinsp;0.155 * (D min)\u0026thinsp;+\u0026thinsp;0.197\u003c/p\u003e \u003cp\u003e*(D max)\u0026thinsp;+\u0026thinsp;0.01 * (Ch min)\u0026thinsp;+\u0026thinsp;0.02 (Ch max)\u003c/p\u003e \u003cp\u003eAs Eq.\u0026nbsp;1 shows, the minimum EF has an inverse relationship with the flow velocity at both the lowest and highest levels and also has a direct relationship with the flow depth and channel index at both the lowest and highest levels. The results of the LR model training phase are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e\u003c/p\u003e \u003cp\u003eThe correlation between the observed and simulated data is equal to R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.67 and also RMSE\u0026thinsp;=\u0026thinsp;0.1301 and MAE\u0026thinsp;=\u0026thinsp;0.0191. As Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea shows, there is not high correlation between the observed and simulated data. Figure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb shows the results of the test phase of the LR model in which R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.81, RMSE\u0026thinsp;=\u0026thinsp;0.1233 and MAE\u0026thinsp;=\u0026thinsp;0.0152. As Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb shows, the correlation between the observed and predicted data at this phase is greater than at the training phase.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003ePredicting the minimum environmental water requirement using the MLP model\u003c/h2\u003e \u003cp\u003eThe MLP model is one of the most common artificial neural network models that is widely used in natural resource studies (Mishra and Goyal \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Mishra and Goyal \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Rahimi \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). The results of MLP model in the training stage of determining the EF of Pir Khezran River are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ea, where R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.96, RMSE\u0026thinsp;=\u0026thinsp;0.0370 and MAE\u0026thinsp;=\u0026thinsp;0.0014. As Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ea shows, the correlation coefficient between the observed and simulated data in the MLP model is greater than the results of the LR model in training phase.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e\u003c/p\u003e \u003cp\u003eThe results of the testing phase of the MLP model (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.92, RMSE\u0026thinsp;=\u0026thinsp;0.0026 and RME\u0026thinsp;=\u0026thinsp;0.0027) in predicting the EF of Pir Khezran River are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eb. According to the Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eb, the correlation coefficient between the observed and predicted data at this phase is more than the results of the LR model, which are consistent with the results (Abba et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Abba and Elkirn \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Mohammadi et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003ePrediction of minimum environmental water requirement using ANFIS model\u003c/h2\u003e \u003cp\u003eAs stated in the methodology section, the ANFIS model is a hybrid nonlinear model that has more advantages than the nonlinear models (Amanollahi and Ausati \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2020b\u003c/span\u003e; Ghasemi and Amanollahi 2018, Kaboodvandpour et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Figure\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ea indicates the results of the ANFIS model in training phase (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.98, RMSE\u0026thinsp;=\u0026thinsp;0.0248 and MAE\u0026thinsp;=\u0026thinsp;0.0006) to determine the EF.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e\u003c/p\u003e \u003cp\u003eComparison of Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ea with the previous figures related to the models training phases (LR and MLP models) shows that the correlation coefficient of ANFIS model is the highest. The results of the testing phase of the ANFIS model (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.97, RMSE\u0026thinsp;=\u0026thinsp;0.0295 and MAE\u0026thinsp;=\u0026thinsp;0.0008) in predicting the EF of the Pir Khezran River are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eb. Comparison of the results of ANFIS model in the testing phase compared to the results of the other two models shows that the ANFIS model has a higher accuracy in predicting the amount of EF in the Pir Khezran River.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn the present study, the accuracy of linear, nonlinear and hybrid models in determining the EF of Pir Khezran River was evaluated. The PHABSIM model was used to estimate EF based on HSI and discharge of the Pir Khezran river. The PHABSIM model needs to provide accurate data which is time consuming and costly. In this study, EF data obtained by PHABSIM model were used as an input of ANFIS, MLP and LR models. Results indicate that nonlinear models have acceptable accuracy compared to the linear model. Results showed that the ANFIS model as a hybrid model has better performance and accuracy than the MLP model in predicting EF. According to the ANFIS model result, it can be concluded that in the basin with the same ecological rivers condition, the PHABSIM model can be used to estimate the EF of one river and the ANFIS model can be utilized to predict the EF of other rivers.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003ethe funding is not available.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of interest/Competing interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll of the data and material are owned by the authors and/or no permissions are required. All authors (Seiran Haghgoo, Jamil Amanollahi, Barzan Bahrami Kamangar) have read, understood, and have complied as applicable with the statement on \u0026ldquo;Ethical responsibilities of Authors\u0026rdquo; as found in the Instructions for Authors and are aware that with minor exceptions, no changes can be made to authorship once the paper is submitted.\u003cstrong\u003eCode availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSeiran Haghgoo:\u0026nbsp;\u003c/strong\u003eConceptualization, Formal analysis, Writing - original draft. \u003cstrong\u003eJamil Amanollahi:\u0026nbsp;\u003c/strong\u003eConceptualization, Supervision, Methodology, Resources, Project administration, Validation, Writing - review \u0026amp; editing; \u003cstrong\u003eBarzan Bahrami Kamangar:\u0026nbsp;\u003c/strong\u003eConceptualization, Methodology, Investigation, Formal analysis;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAazami J, Motevalli A, Savabieasfahani M (2022) Correction to: Evaluation of three environmental flow techniques in Shoor wetland of Golpayegan, Iran. Int. J. Environ. Sci. Technol. \u003cstrong\u003e19\u003c/strong\u003e:7899.\u003c/li\u003e\n \u003cli\u003eAbba S, Hadi SJ, Abdullahi J (2017) River water modelling prediction using multi linear regression, artificial neural network, and adaptive neuro-fuzzy inference system techniques. . Procedia Comput. Sci. 120:75-82.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eAbba SI, Elkirn G (2017) Effluent prediction of chemical oxygen demand from the astewater treatment plant using artificial neural network application. Procedia Comput. Sci. 120:156-163.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eAmanollahi J, Ausati Sh (2020a) PM\u003csub\u003e2.5\u0026nbsp;\u003c/sub\u003econcentration forecasting using ANFIS, EEMD-GRNN, MLP, and MLR models: A case study of Tehran, Iran. Air Qual. Atmos. Health. 13:161-17.\u003c/li\u003e\n \u003cli\u003e\u0026nbsp;Amanollahi J, Ausati Sh (2020b).\u0026nbsp;Validation of linear, nonlinear, and hybrid models for predicting particulate matter concentration in Tehran. Iran.\u0026nbsp;Theor. Appl. Climatol. 140:709-717.\u003c/li\u003e\n \u003cli\u003eBejarano MD, Sord-Ward A, Gabriel-Martin I, Garrote L (2019) Tradeoff between economic and environmental costs and benefits of hydropower production at run-of-river-diversion schemes under different environmental flows scenarios. J. Hydrol. 572:790-804.\u003cu\u003e\u0026nbsp;\u003c/u\u003e\u003c/li\u003e\n \u003cli\u003eBelmar O, Velasco J, Martinez-Capel F (2011) Hydrological Classification of Natural Flow Regimes to Support Environmental Flow Assessments in Intensively Regulated Mediterranean Rivers, Segura River Basin (Spain). Environ. Manage. \u003cstrong\u003e47\u003c/strong\u003e:992\u0026ndash;1004.\u003c/li\u003e\n \u003cli\u003eBice CM, Gehrig SL, Zampatti BP et al. (2014) Flow-induced alterations to fish assemblages, habitat and fish\u0026ndash;habitat associations in a regulated lowland river. Hydrobiologia. \u003cstrong\u003e722\u003c/strong\u003e:205\u0026ndash;222.\u003c/li\u003e\n \u003cli\u003eBouska K (2020) Regime change in a large-floodplain river ecosystem: patterns in body-size and functional biomass indicate a shift in fish communities. Biol. Invasions. \u003cstrong\u003e22\u003c/strong\u003e:3371\u0026ndash;3389.\u003c/li\u003e\n \u003cli\u003eBovee KD (1982) Instream flow methodology. US Fish and Wildlife Service. FWS/OBS, 82, 26 Caudill, M. and Butler, C. 1992. Understanding Neural Networks. Basic Networks, vol. 1. MIT Press, Cambridge, MA.\u003c/li\u003e\n \u003cli\u003eCeylan \u0026Zeta;, Bulkan S (2018) Forecasting PM\u003csub\u003e10\u003c/sub\u003e levels using ANN and MLR: A case study for Sakarya City. Global Nest J. 20:281-290.\u003c/li\u003e\n \u003cli\u003eDana K, Jana C, Tom\u0026aacute;\u0026scaron; V (2017) Analysis of environmental flow requirements for macroinvertebrates in a creek affected by urban drainage (Prague metropolitan area, Czech Republic). Urban Ecosyst \u003cstrong\u003e20\u003c/strong\u003e:785\u0026ndash;797.\u003c/li\u003e\n \u003cli\u003eDawson,CW, Wilby R L (2001) Hydrological modelling using artificial neural networks, Progr Phys, Geogr. 25:80\u0026ndash;108.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eDumitriu D (2020) Sediment flux during flood events along the Trotuș\u0026nbsp;River channel: hydrogeomorphological approach. J. Soils Sediments\u003cstrong\u003e. 20\u003c/strong\u003e:4083\u0026ndash;4102.\u003c/li\u003e\n \u003cli\u003eEmadi A, Sobhani R, Ahmadi H \u003cem\u003eet al.\u003c/em\u003e Multivariate modeling of agricultural river water abstraction via novel integrated-wavelet methods in various climatic conditions. Environ. Dev. Sustain. \u003cstrong\u003e24\u003c/strong\u003e:4845\u0026ndash;4871.\u003c/li\u003e\n \u003cli\u003eEspinoza T, Burke CL, Carpenter-Bundhoo L, Marshal SM, McDougall AJ, Roberts DT, Campbell HA, Kennard MJ (2021) Quantifying movement of multiple threatened species to inform adaptive management of environmental flows. J. Environ. Managet. 295:113067.\u003c/li\u003e\n \u003cli\u003eFraternali P, Castelletti A, Soncini-Sessa R, Vaca Ruiz C, Rizzoli AE (2012) Putting humans in the loop: social computing for Water Resources Management. Environ. Modell. Softw. 37:68-77\u003c/li\u003e\n \u003cli\u003eFukuda S (2009) Consideration of fuzziness: Is it necessary in modelling fish habitat preference of Japanese medaka (Oryzias latipes)? Ecol. Model. 220:2877-2884.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eFukuda S, Hiramatsu B (2008) Prediction ability and sensitivity of artificial intelligence-based habitat preference models for predicting spatial distribution of Japanese medaka (Oryzias latipes). Ecol. Model. 215:301-313.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eGain AK, Apel H, Renaud FG \u003cem\u003eet al. (2013)\u003c/em\u003e Thresholds of hydrologic flow regime of a river and investigation of climate change impact\u0026mdash;the case of the Lower Brahmaputra river Basin. Climatic Change. \u003cstrong\u003e120\u003c/strong\u003e:463\u0026ndash;475.\u003c/li\u003e\n \u003cli\u003eGhasemi A, Amanollahi J (2019) Integration of ANFIS model and forward selection method for air quality forecasting. Air Qual. Atmos. Health. 12:59-72.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eGholami V, Khalili A, Sahour H et al. (2020) Assessment of environmental water requirement for rivers of the Miankaleh wetland drainage basin. Appl. Water Sci. \u003cstrong\u003e10\u003c/strong\u003e:233\u003c/li\u003e\n \u003cli\u003eIm\u0026nbsp;D, Choi SU, Choi B (2018) Physical habitat simulation for a fish community using the ANFIS method. Ecol. Inform. 43:73-83.\u003c/li\u003e\n \u003cli\u003eIUCN (World Conservation Union) (2000). Vision for Water and Nature. A World Strategy for conservation and sustainable management of water resources in the 21st Century. IUCN, Gland, Switzerland and Cambridge, UK, p 52.\u003c/li\u003e\n \u003cli\u003eJung SH, Choi S-Uk (2015) Prediction of composite suitability index for physical habitatsimulations using the ANFIS method. Appl. Soft Comput. 34:502\u0026ndash;512.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eKaboodvandpour S, Amanollahi J, Qhavami S, Mohammadi B (2015) Assessing the accuracy of multiple regressions, ANFIS, and ANN models in predicting dust storm occurrences in Sanandaj, Iran. Nat. Hazards. 78:879-893.\u003c/li\u003e\n \u003cli\u003eKaratzas K, Katsifarakis N, Orlowski C, Sarzynski A (2018) Revisiting urban air quality forecasting: a regression approach. Vietnam J. Comput. Sci. 5:177-184.\u003c/li\u003e\n \u003cli\u003eKayes I, Shahriar SA, Hasan K, Akhter M, Kabir MM, Salam MA (2019) The relationships between meteorological parameters and air pollutants in an urban environment. Global J. Environ. Sci. Manag. 5:265-278.\u003c/li\u003e\n \u003cli\u003eKim HJ, Kim JH, Ji U et al. (2020) Effect of Probability Distribution-Based Physical Habitat Suitability Index on Environmental-Flow Estimation. KSCE J. Civ. Eng. \u003cstrong\u003e24\u003c/strong\u003e:2393\u0026ndash;2402.\u003c/li\u003e\n \u003cli\u003eKnack IM, Huang F, Shen HT (2020) Modeling fish habitat condition in ice affected rivers. Cold Reg. Sci. Technol. 176:103086.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMiao Y, Li J, Feng P, Dong L, Zhang T, Wu J, Katwal R (2020) Effects of land use changes on the ecological operation of the Panjiakou-Daheiting Reservoir system, China. Ecol. Eng. 152:10585\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMishra D, Goyal P (2015) Development of artificial intelligence based NO\u003csub\u003e2\u003c/sub\u003e forecasting models at Taj Mahal, Agra. Atmos. Pollut. Res. 6:99-106.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMishra D, Goyal P (2016) Neuro-Fuzzy approach to forecasting ozone episodes over the urban area of Delhi, India. Environ. Technol. Inno. 5:83-94.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMlynski D, Operacz A, Walega A (2020) Sensitivity of methods for calculating environmental flows based on hydrological characteristics of watercourses regarding te hydropower potential of rivers. J. Clean. Pro. 250:119527.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMohammadi S, Naseri F, Abri R (2019) Simulating soil loss rate in Ekbatan Dam watershed using experimental and statistical approaches. Int. J. Sediment. Res. 34:226-239.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMwamila TB, Kimwaga RJ, Mtalo FW (2008) Eco-hydrology of the Pangani River downstream of Nyumba ya Mungu reservoir, Tanzania. Phys. Chem. Earth. 33:695-700.\u003c/li\u003e\n \u003cli\u003eNagaya T, \u0026nbsp;Shiraishi Y, Onitsuka K, Higashino M, Takami T, Otsuka N, Akiyama J, Ozeki H (2008) Evaluation of suitable hydraulic conditions for spawning of ayu with horizontal 2D numerical simulation and PHABSIM. Ecol. Modell. 215:133-143.\u003c/li\u003e\n \u003cli\u003eNasiri Khiavi A, Mostafazadeh R Ghanbari Talouki F (2024) Using game theory algorithm to identify critical watersheds based on environmental flow components and hydrological indicators. Environ. Dev. Sustain.\u0026nbsp;\u003ca href=\"https://doi-org.access.semantak.com/10.1007/s10668-023-04390-8\"\u003ehttps://doi-org.access.semantak.com/10.1007/s10668-023-04390-8\u003c/a\u003e\u003c/li\u003e\n \u003cli\u003eNikghalb S, Shokoohi A, Singh VP \u003cem\u003eet al\u003c/em\u003e (2016) Ecological Regime versus Minimum Environmental Flow:Comparison of Results for a River in a Semi Mediterranean Region. Water Resour. Manage. \u003cstrong\u003e30\u003c/strong\u003e:4969\u0026ndash;4984.\u003c/li\u003e\n \u003cli\u003ePandey CL (2021) Managing urban water security: challenges and prospects in Nepal. Environ. Dev. Sustain. \u003cstrong\u003e23\u003c/strong\u003e:241\u0026ndash;257.\u003c/li\u003e\n \u003cli\u003ePeng L, Sun L (2016) Minimum instream flow requirement for the water-reduction section of diversion-type hydropower station: a case study of the Zagunao River, China. Environ. Earth Sci. \u003cstrong\u003e75\u003c/strong\u003e:1210\u003c/li\u003e\n \u003cli\u003ePerez-Blanco CD, Gil-Garcia L, Saiz-Santiago (2021) An actionable hydroeconomic descision support system for the assessment of water reallocations in irrigated agriculture. Astudy of minimum environmental flows in the Douro River Basin, Spain. J. Environ. Managet. 298:113432.\u003c/li\u003e\n \u003cli\u003eRahimi A (2017) Short-term prediction of NO 2 and NO x concentrations using multilayer perceptron neural network: a case study of Tabriz, Iran. Ecol. Process. 6: 4.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eRinderknecht SL, Borsuk ME, Reichert P (2012) Bridging uncertain and ambiguous knowledge with imprecise probabilities. Environ. Modell. Softw.\u0026nbsp;36:122-130.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eSedighkia M, Abdoli A (2023) Design of optimal environmental flow regime at downstream of multireservoir systems by a coupled SWAT-reservoir operation optimization method. Environ Dev Sustain. \u003cstrong\u003e25\u003c/strong\u003e:834\u0026ndash;854.\u003c/li\u003e\n \u003cli\u003eSedighkia M, Datta B (2023) Analyzing environmental flow supply in the semi-arid area through integrating drought analysis and optimal operation of reservoir. \u003cem\u003eJ. Arid Land\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e:1439\u0026ndash;1454.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eSekine M, Wang J, Yamamoto K \u003cem\u003eet al.\u003c/em\u003e (2020) Fish habitat evaluation based on width-to-depth ratio and eco-environmental diversity index in small rivers. Environ Sci. Pollut. Res. \u003cstrong\u003e27\u003c/strong\u003e:34781\u0026ndash;34795.\u003c/li\u003e\n \u003cli\u003eSenent-Aparicio J, George C, Srinivasan R (2021) Introducing a new post-processing tool for the SWAT+model to evaluate environmental flows.\u0026nbsp;Environ. Modell. Softw. 136:104944.\u003c/li\u003e\n \u003cli\u003eSpence R, Hickley P (2000) The use of PHABSIM in the management of water resources and fisheries in England and Wales. Ecol. Eng. 16:153-158.\u003c/li\u003e\n \u003cli\u003eStucchi L. Bocchiola D (2023) Environmental Flow Assessment using multiple criteria: A case study in the Kumbih river, West Sumatra (Indonesia). Sci. Total Environ. 901:166516.\u003c/li\u003e\n \u003cli\u003eSykes AO (1993) An Introduction to Regression Analysis. Am. Stat, 61, 101.\u003c/li\u003e\n \u003cli\u003eVan der Lee GEM, Van der Molen DT, Van den Boogaard HFP et al. (2006) Uncertainty analysis of a spatial habitat suitability model and implications for ecological management of water bodies. Landscape Ecol. \u003cstrong\u003e21\u003c/strong\u003e:1019\u0026ndash;1032.\u003c/li\u003e\n \u003cli\u003eVivian LM, Godfree RC, Colloff MJ \u003cem\u003eet al. (2014)\u003c/em\u003e Wetland plant growth under contrasting water regimes associated with river regulation and drought: implications for environmental water management. Plant Ecol. \u003cstrong\u003e215\u003c/strong\u003e:997\u0026ndash;1011.\u003c/li\u003e\n \u003cli\u003eWang H, Cong P, Zhu Z, Zhang W, Ai Y, Huai W (2022) Analysis of environmental dispersion in wetland flows with floating vegetation islands.\u0026nbsp;J. Hydrol.\u0026nbsp;606:127359.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003e\u003ca href=\"https://www.sciencedirect.com/science/article/abs/pii/S0959652620353373#!\"\u003eWen X,\u003c/a\u003e \u003ca href=\"https://www.sciencedirect.com/science/article/abs/pii/S0959652620353373#!\"\u003eLv Y,\u0026nbsp;\u003c/a\u003e\u003ca href=\"https://www.sciencedirect.com/science/article/abs/pii/S0959652620353373#!\"\u003eLiu Z,\u0026nbsp;\u003c/a\u003e\u003ca href=\"https://www.sciencedirect.com/science/article/abs/pii/S0959652620353373#!\"\u003eDing Z,\u003c/a\u003e \u003ca href=\"https://www.sciencedirect.com/science/article/abs/pii/S0959652620353373#!\"\u003eLei X,\u003c/a\u003e \u003ca href=\"https://www.sciencedirect.com/science/article/abs/pii/S0959652620353373#!\"\u003eTan Q,\u003c/a\u003e \u003ca href=\"https://www.sciencedirect.com/science/article/abs/pii/S0959652620353373#!\"\u003eS\u003c/a\u003eun Y (2021) Operation chart optimization of multi-hydropower system incorporating the long- and short-term fish habitat requirements.\u0026nbsp;J. Clen. Pro.\u0026nbsp;121:107053.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eWeng X, Jiang C, Yuan M, Zhang M, Zeng T, Jin C (2021) An ecologically dispatch strategy using environmental flows for a cascade multi-sluice system: A case study of the Yongjiang River Basin, China. Ecol. Indic. 121:107053.\u003c/li\u003e\n \u003cli\u003eYu L, Wu X, Wu S, Jia B, Han G, Xu P, Dai J, Zhang Y, Wang F, Yang Q, Zhou Z (2021) Multi-objective optimal operation of cascade hydropower plants considering ecological flow under different ecological conditions.\u0026nbsp;J. Hydrol. 601:126599.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eZamanzad-Ghavidel S, Fazeli S, Mozaffari S \u003cem\u003eet al\u003c/em\u003e. (2023) Estimating of aqueduct water withdrawal via a wavelet-hybrid soft-computing approach under uniform and non-uniform climatic conditions. Environ. Dev. Sustain\u003cem\u003e.\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e:5283\u0026ndash;5314.\u003c/li\u003e\n \u003cli\u003eZhang Y, Yu H, Yu H. et al. (2020) Optimization of environmental variables in habitat suitability modeling for mantis shrimp Oratosquilla oratoria in the Haizhou Bay and adjacent waters. Acta Oceanol. Sin. \u003cstrong\u003e39\u003c/strong\u003e:36\u0026ndash;47.\u003c/li\u003e\n \u003cli\u003eZheng Y, Tian Y, Du E, Han F, Wu Y, Zheng C, Li X (2020) Addressing the water conflict between agriculture and ecosystems under environmental flow regulation: An integrated modeling study.\u0026nbsp;Environ. Modell. Softw. 134:104874.\u003c/li\u003e\n \u003cli\u003eZolfagharpour F, Saghafian B, Delavar M (2021) Adapting reservoir operation rules to hydrological drought state and environmental flow requirements. J. Hydrol. 660:126581.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Water velocity, Flow depth, Channel index, PHABSIM, Habitat","lastPublishedDoi":"10.21203/rs.3.rs-3939514/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3939514/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn recent decades, the reckless exploitation of rivers has caused significant changes in their ecosystems and upstream flow. It is imperative to understand that preservation of river ecosystems solely relies on maintaining the environmental flow (EF). Estimating the EF requires filed sampling, which are both time-consuming and costly. Thus, the purpose of this research is to estimate EF for a river and generalize its result to adjacent rivers using the modelling. To determine the EF, the physical habitat simulation (PHABSIM) model was used. Habitat suitability indexes (HSI) were created based on the filed survey for water velocity, flow depth, channel index and water temperature in a river. To predict the EF for other rivers, the linear regression model (LR) and two different types of neural network models, including Adaptive Neuro-Fuzzy Inference System (ANFIS) and multi-layer perceptron (MLP) were utilized. In this study, 80% and 20% of the data were used for training and testing phases, respectively. Among the models, in the ANFIS model, the date obtained for both training phase and testing phase were as follows respectively. R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.98, RMSE\u0026thinsp;=\u0026thinsp;0.0248 and MAE\u0026thinsp;=\u0026thinsp;0.0006 as well as R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.97, RMSE\u0026thinsp;=\u0026thinsp;0.0295 and MAE\u0026thinsp;=\u0026thinsp;0.0008. The accuracy of them were higher compared to MLP and LR models in predicting EF. Therefore, the ANFIS hybrid model can be a suitable alternative method for estimating the EF.\u003c/p\u003e","manuscriptTitle":"Reducing the computational cost and time of environmental flow estimation based on machine learning approaches","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-02-15 13:11:19","doi":"10.21203/rs.3.rs-3939514/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4f0b931f-78fb-4ac3-8fe8-3961a51ae841","owner":[],"postedDate":"February 15th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-03-16T13:59:24+00:00","versionOfRecord":[],"versionCreatedAt":"2024-02-15 13:11:19","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3939514","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3939514","identity":"rs-3939514","version":["v1"]},"buildId":"J0_U0BvcaRcwD8yVFaRlm","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}