كاربرد مدل‌هاي تلفيق داده در شبيه‌سازي جريان رودخانه با استفاده از سيگنال‌هاي بزرگ‌ مقياس اقليمي، مطالعه موردي: حوزه آبخيز سد جيرفت

Application of data fusion models in river flow simulation using signals of large-scale climate, case study: Jiroft Dam Basin

ﭘﯿﺶ ﺑﯿﻨﯽ آﺑﺪﻫﯽ رودﺧﺎﻧﻪ در ﺣﻮزه ﻫﺎي آﺑﺨﯿﺰ از ﺟﺎﯾﮕﺎه وﯾﮋه اي در ﻣﺪﯾﺮﯾﺖ و ﺑﺮﻧﺎﻣﻪ رﯾﺰي ﻣﻨﺎﺑﻊ آب ﺑﻪ ﻣﻨﻈﻮر ﻃﺮاﺣﯽ ﺗﺄﺳﯿﺴﺎت آﺑﯽ، آﺑﮕﯿﺮي از رودﺧﺎﻧﻪ ﻫﺎ، ﻣﺪﯾﺮﯾﺖ ﻣﺼﺮف و ﻣﻮاردي از اﯾﻦ ﻗﺒﯿﻞ ﺑﺮﺧﻮردار اﺳﺖ. در ﭘﮋوﻫﺶ ﺣﺎﺿﺮ، ﻋﻤﻠﮑﺮد ﺑﺮﺧﯽ از ﻣﺪلﻫﺎي ﺗﻠﻔﯿﻖ داده ﺷﺎﻣﻞ ﻣﯿﺎﻧﮕﯿﻦﮔﯿﺮي ﺳﺎده، ﻣﯿﺎﻧﮕﯿﻦﮔﯿﺮي وزندار و ﻣﺪل ﺷﺒﮑﻪ ﻋﺼﺒﯽ ﻣﺼﻨﻮﻋﯽ ﺗﻠﻔﯿﻘﯽ در ﻣﺪل ﺳﺎزي آﺑﺪﻫﯽ ﻣﺎﻫﺎﻧﻪ ﻣﻮرد ارزﯾﺎﺑﯽ و ﻣﻘﺎﯾﺴﻪ ﻗﺮار ﮔﺮﻓﺘﻪ اﺳﺖ. ﺑﻪ ﻫﻤﯿﻦ ﻣﻨﻈﻮر، اﺑﺘﺪا ﺑﺎ اﺳﺘﻔﺎده از ﻣﺪل ﻫﺎي ﺷﺒﮑﻪ ﻋﺼﺒﯽ ﻣﺼﻨﻮﻋﯽ )ANN(، ﺳﺎﻣﺎﻧﻪ اﺳﺘﻨﺘﺎج ﻋﺼﺒﯽ-ﻓﺎزي ﺗﻄﺒﯿﻘﯽ )ANFIS(، ﻣﺪل آرﯾﻤﺎ )ARIMA( و ﻣﺪل رﮔﺮﺳﯿﻮن ﺑﺮدار ﭘﺸﺘﯿﺒﺎن )SVR( ﺑﻪ ﻋﻨﻮان ﻣﺪل ﻫﺎي ﻣﻨﻔﺮد، ﭘﯿﺶ ﺑﯿﻨﯽ آﺑﺪﻫﯽ ﻣﺎﻫﺎﻧﻪ در ﺣﻮزه آﺑﺨﯿﺰ ﺑﺎﻻدﺳﺖ ﺳﺪ ﺟﯿﺮﻓﺖ ﻣﻮرد ﺑﺮرﺳﯽ ﻗﺮار ﮔﺮﻓﺖ. ﺳﭙﺲ، ﻣﺪل ﻫﺎي ﻣﻨﻔﺮد ﺑﺎ اﺳﺘﻔﺎده از ﻣﺘﻐﯿﺮﻫﺎي ﭘﯿﺶ ﺑﯿﻨﯽ ﮐﻨﻨﺪه ﻣﻨﺘﺨﺐ، آﻣﻮزش و ﺻﺤﺖ ﺳﻨﺠﯽ ﺷﺪه، ﻧﺘﺎﯾﺞ آنﻫﺎ ﺑﺮاي اﺳﺘﻔﺎده در ﻓﺮاﯾﻨﺪ ﺗﻠﻔﯿﻖ اﻧﺘﺨﺎب ﺷﺪ. ﻫﻤﭽﻨﯿﻦ، از ﺳﯿﮕﻨﺎلﻫﺎي ﺑﺰرگ ﻣﻘﯿﺎس اﻗﻠﯿﻤﯽ ﺷﺎﻣﻞ ENSO ،NAO و PDO در ﭘﯿﺶ ﺑﯿﻨﯽ ﻫﺎي ﻫﯿﺪروﻟﻮژﯾﮏ ﺟﺮﯾﺎن رودﺧﺎﻧﻪ اﺳﺘﻔﺎده ﺷﺪه، ﻋﻤﻠﮑﺮد ﻣﺪل ﻫﺎي ﻣﻨﻔﺮد و ﺗﻠﻔﯿﻘﯽ در دو ﺣﺎﻟﺖ ﺑﺎ در ﻧﻈﺮ ﮔﺮﻓﺘﻦ اﯾﻦ ﺳﯿﮕﻨﺎلﻫﺎ و ﺑﺪون در ﻧﻈﺮ ﮔﺮﻓﺘﻦ آنﻫﺎ، ﺑﺮ اﺳﺎس ارزﯾﺎﺑﯽ ﺳﻪ ﻣﻌﯿﺎره ﻧﺶ-ﺳﺎﺗﮑﻠﯿﻒ )NSE(، ﺿﺮﯾﺐ ﺗﺒﯿﯿﻦ )R ( و ﻣﯿﺎﻧﮕﯿﻦ ﻣﺮﺑﻌﺎت ﺧﻄﺎ )MSE( ﻣﻮرد ﻣﻘﺎﯾﺴﻪ ﻗﺮار ﮔﺮﻓﺘﻪ اﺳﺖ. ﻧﺘﺎﯾﺞ ﺣﺎﺻﻞ از اﯾﻦ ﭘﮋوﻫﺶ ﺣﺎﮐﯽ از آن ﺑﻮد ﮐﻪ روﯾﮑﺮد ﺗﻠﻔﯿﻖ داده دﻗﺖ ﭘﯿﺶ ﺑﯿﻨﯽ ﻫﺎ را ﺗﺎ ﺣﺪ ﻗﺎﺑﻞ ﻣﻼﺣﻈﻪ اي اﻓﺰاﯾﺶ ﻣﯽ دﻫﺪ. ﻋﻼوه ﺑﺮ اﯾﻦ، ﻣﺸﺨﺺ ﺷﺪ ﮐﻪ ﺳﯿﮕﻨﺎلﻫﺎي ﺑﺰرگ ﻣﻘﯿﺎس اﻗﻠﯿﻤﯽ ﻣﻨﺠﺮ ﺑﻪ ﺑﻬﺒﻮد ﻧﺘﺎﯾﺞ ﺧﺼﻮﺻﺎً در دوره ﺗﺴﺖ ﺷﺪه اﺳﺖ. ﺑﻪﻋﻨﻮان ﻣﺜﺎل، ﻧﺘﺎﯾﺞ ﺣﺎﺻﻞ از ﻣﺪل ﺗﻠﻔﯿﻘﯽ ANN ﺑﻪ ﻫﻤﺮاه ﺳﯿﮕﻨﺎلﻫﺎي ﺑﺰرگ ﻣﻘﯿﺎس اﻗﻠﯿﻤﯽ ﻧﺸﺎن ﻣﯽ دﻫﺪ ﮐﻪ اﯾﻦ ﻣﺪل ﺑﻬﺘﺮﯾﻦ ﻋﻤﻠﮑﺮد را در ﻣﯿﺎن ﻣﺪل ﻫﺎي ﺗﻠﻔﯿﻖ داده دارا ﻣﯽﺑﺎﺷﺪ. ﻫﻤﭽﻨﯿﻦ، ﻣﻌﯿﺎر NSE ﻧﺴﺒﺖ ﺑﻪ ﻣﺪل ﺗﻠﻔﯿﻘﯽ ANN ﺑﺪون ﺳﯿﮕﻨﺎلﻫﺎي ﺑﺰرگ ﻣﻘﯿﺎس در دوره آﻣﻮزش 0/04 ﺑﻬﺒﻮد ﯾﺎﻓﺘﻪ، ﺧﻄﺎي MSE ﺑﻪ ﻣﯿﺰان 0/001 ﮐﺎﻫﺶ ﭘﯿﺪا ﮐﺮده اﺳﺖ.

River runoff forecasting in watersheds has a special place in the management and planning of water resources for the design of water facilities, water intake from rivers, consumption management and etc. In the present study, the performance of some data integration models including simple averaging, weighted averaging and integrated artificial neural network model in monthly discharge modeling has been evaluated and compared. For this purpose, monthly flow prediction in upstream basin of Jiroft Dam was examined using Artificial Neural Network (ANN) models, Adaptive Neural-Fuzzy Inference System (ANFIS), ARIMA model and Support Vector Regression (SVR) model as an individual model. Then, the individual models were trained and validated using selected predictor variables and their results were selected for use in the integration process. Large-scale climatic signals including NAO, ENSO and PDO are also used in hydrological forecasts of river flow and the performance of single and integrated models in two modes with and without considering these signals has been compared based on the evaluation of three criteria Nash-Sutcliffe (NSE), Coefficient of determination (R2) and Mean Square Error (MSE). Results of this study indicated that the integrated approach significantly increases the accuracy of predictions. In addition, large-scale climatic signals were found to improve results, especially during the test period. For example, the results of the integrated model of artificial neural network with large climatic scale signals show that this model has the best performance among the integrated models. Also, the NSE criterion has improved by 0.04 in training compared to the integrated model of artificial neural network without large-scale signals and the MSE error has been reduced by 0.001.