برآورد رواناب ماهانه و فصلی با مدل‌های سری زمانی، درخت تصمیم و رگرسیون خطی چندمتغیره

نوع مقاله : مقاله کامل علمی پژوهشی

نویسندگان

1 دانشجوی کارشناسی‌ارشد گروه مهندسی منابع آب، دانشگاه علوم کشاورزی و منابع طبیعی گرگان.

2 نویسنده مسئول، دانشیار گروه مهندسی آب، دانشگاه علوم کشاورزی و منابع طبیعی گرگان.

3 دانشیار گروه مهندسی آب، دانشگاه علوم کشاورزی و منابع طبیعی گرگان.

4 استادیار گروه مهندسی آب، دانشگاه علوم کشاورزی و منابع طبیعی گرگان.

10.22069/jwsc.2022.19921.3533

چکیده

سابقه و هدف: از جمله عوامل ‌حائز اهمیت در مدیریت و برنامه‌ریزی منابع آب پیش‌بینی مقدار رواناب می‌باشد. با افزایش دقت در پیش‌بینی رواناب رودخانه مدیریت و برنامه‌ریزی کارآمدتری صورت می‌گیرد بنابراین بهبود مدلسازی پیش‌بینی رواناب امری ضروری است.اولین هدف از این مطالعه ارزیابی کارایی مدلهای رگرسیون چندمتغیره خطی، درخت تصمیم M5 و سری زمانی در پیشبینی رواناب رودخانه است. هدف دوم بررسی مقیاس زمانی مدلسازی (ماهانه و فصلی) و نیز تاثیر ورودی های مدل (یک متغیر با گام های تاخیر و چند متغیر با گامهای تاخیر) بر دقت مدلهای مورد مطالعه است.
مواد و روش‌ها: در این پژوهش حوضه آبریز ناورود واقع در غرب استان گیلان جهت منطقه مطالعاتی انتخاب گردیده‌است. داده‌های مورد نیاز دو ایستگاه خرجگیل در سال‌های 1398-1368 و خلیان در سال‌های 1397-1375 شامل دبی، بارش و دما در مقیاس زمانی ماهانه از آب منطقه‌ای استان گیلان جمع‌آوری شده‌است. مقدار رواناب توسط داده‌های دریافت شده در بازه زمانی ماهانه و فصلی با استفاده سه مدل رگرسیون چند متغیره خطی، سری زمانی و درخت تصمیم M5 در دو رویکرد متفاوت پیش‌بینی شده‌است. رویکرد اول متغیرهای ورودی به مدل شامل دبی، بارش و دما با 3 گام تاخیر زمانی و در رویکرد دوم تنها متغیر دبی با 3 گام تاخیر زمانی بوده‌است. شاخص‌های ارزیابی در این پژوهش شامل میانگین انحراف خطا (MBE)، ضریب کارایی نش (NSE) و ضریب تعیین (R^2) می‌باشد.
یافته‌ها یافته‌ها: در رویکرد اول و در پنجره زمانی ماهانه مدل درخت تصمیم M5 با شاخص MBE، NSE 04/0-، 80/0 (آموزش) و 01/0، 72/0 (آزمون) در ایستگاه خرجگیل و 01/0-، 79/0 (آموزش) و 00/0، 82/0 (آزمون) در ایستگاه خلیان بعنوان مدل منتخب انتخاب می گردد. در گام زمانی فصلی نیز مقادیر شاخص ها برای مدل درخت تصمیم M5 در ایستگاه خرجگیل برابر . برابر 02/0، 78/0 (آموزش) 02/0-، 86/0 (آزمون) و در ایستگاه خلیان نیز 01/0-، 79/0 (آموزش ) و 00/0، 86/0 (آزمون) می باشد و این مدل در گام زمانی فصلی در رویکرد اول نیز بهترین مدل مورد مطالعه بوده است. رویکرد دوم در هر دو گام زمانی ماهانه و فصلی منجر به یافته های متفاوتی شده است. در رویکرد دوم در گام زمانی ماهانه مقادیر شاخصها برای مدل سری زمانی در دو مرحله آموزش و آزمون در ایستگاه خرجگیل بهترتیب برابر 05/0- ، 47/0 و 10/0، 52/0 و در ایستگاه خلیان برابر با 02/0-، 63/0 و 02/0، 49/0 بوده است. در گام زمانی فصلی نیز مقادیر شاخص های مدل منتخب در دو مرحله آموزش و آزمون در ایستگاه خرجگیل 42/0-، 58/0 و 06/0، 83/0 و خلیان 09/0، 40/0 و 10/0-، 62/0 می باشد. در گام زمانی فصلی نیز مدل سری زمانی مدل منتخب در رویکرد دوم می باشد

نتیجه‌گیری: نتایج حاصل از این پژوهش حاکی از آن است که در رویکرد اول در هر دو ایستگاه و در هر دو گام زمانی مدل درخت تصمیم M5 دقت بالاتری در پیش‌بینی نسبت به دو مدل دیگر از خود نشان داده‌است در حالیکه در رویکرد دوم مدل درخت تصمیم نتایج با دقت بالا از خود نشان نمی‌دهد و در مقابل مدل سری زمانی دقت بالاتری نسبت به دو مدل دیگر در هر دو ایستگاه و هر دو گام زمانی داشته‌است. یافته های این مطالعه بر این موضوع تاکید دارد که رویکرد مورد استفاده در انتخاب ورودی های مدل می تواند به شکل کامل موثری دقت مدلسازی و مدل منتخب را تحت تاثیر قرار دهد.

کلیدواژه‌ها


عنوان مقاله [English]

Monthly and seasonal runoff estimation using time series, decision tree, and multivariable linear regression

نویسندگان [English]

  • Hedieh Khodakhah 1
  • khalil ghorbani 2
  • Meysam Salarijazi 3
  • Mohammad Abdolhosseini 4
1 M.Sc. Student, Dept. of Water Resources Engineering, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran.
2 Corresponding Author, Associate Prof., Dept. of Water Engineering, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran.
3 Associate Prof., Dept. of Water Engineering, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran.
4 Assistant Prof., Dept. of Water Engineering, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran.
چکیده [English]

Background and Objectives: One of the essential factors in the programming and management of water resources is predicting the amount of runoff. Increasing the accuracy in predicting runoff will increase the efficiency of programming and management; therefore, improving the modeling of discharge prediction is a requisite issue. The first aim of this study is to evaluate the efficiency of the multivariable linear regression, M5 decision tree, and time series in predicting the river runoff. The second aim is to analyze the modeling time step (monthly or seasonal) and the effects of model inputs (one delay steps variable against several delay steps variable) on the accuracy of the studied models.
Material and Methods: Navrood watershed located in the west part of Gilan province is chosen for the study area in this research. Required data is collected from Kharjgil (1368-1398) and Kholian (1375-1397), including monthly river flow, rainfall, and temperature from Gilan regional water company. The amount of runoff is predicted in two approaches by the received data in monthly and seasonal time steps sing three models of multivariable linear regression, time series, and M5 decision tree. In the first approach, input variables to the model were river flow, rainfall, and temperature with three steps delay. In the second approach, the only variable was river flow with three steps delay. The model evaluation criteria in this research are the mean bias error (MBE), Nash-Sutcliffe efficiency (NSE), and coefficient of determination (R^2).
Results: In the first approach and in monthly timestep, M5 decision tree is selected model with MBE-NSE equal to -0.04,0.80 (train) and 0.01,0.72 (test) in Kharjgil station, and -0.01,0.79 (train) and 0.00,0.86 (test) in Kholian station. In the seasonal time step, the criteria for the M5 decision tree in Kholian station are equal to 0.02,0.78 (train), -0.02,0.86 (test), and in Kholian station are -0.01,0.79 (train), 0.00,0.86 (test). This model was the best in this study for the first approach in the seasonal time step. The second approach has led to different findings considering both monthly and seasonal time steps. In the second approach, the criteria in monthly time step for time series model during train and test in Kharjgil station are respectively -0.05,0.47 and 0.10,0.52 and in Kholian are -0.02,0.63 and 0.2,0.49. The selected model criteria for seasonal time step considering train and test are -0.42,0.58 and 0.06,0.83 in Kharjgil station, and 0.09,0.40 and -0.10,0.62 in Kholian station. The time series model is selected in the second approach in the seasonal time step.
Conclusion: The findings of this research have shown that in both stations and time steps, the M5 decision tree model has shown a higher accuracy in prediction than the two other models in the first approach. Meanwhile, the decision tree model does not show accurate results in the second approach. Alternatively, compared to two other models in both stations and both time steps, the time series model had a higher accuracy. Findings of this research have emphatically shown that specific approaches in choosing the model's inputs can effectively influence the selected model and the accuracy of modeling.

کلیدواژه‌ها [English]

  • Runoff
  • Time series
  • Multivariable linear regression
  • Decision tree M5
1.Abdolahnezhad, K. 2015. Forecasting of Monthly Sum-raining by Stochastic Models in Time Series. Geographical Planning of Space, 5: 17. 15-25.(In Persian)
2.Adnan, R.M., Yuan, X., Kisi, O., and Curtef, V. 2017. Application of time series models for streamflow forecasting. Civil and Environmental Research,9: 3. 56-63.‏
3.Ahmad, S., and Simonovic, S.P. 2005. An artificial neural network model for generating hydrograph from hydro-meteorological parameters. Journal of Hydrology, 315: 1-4. 236-251.
4.Akhtar, M.K., Corzo, G.A., Van Andel, S.J., and Jonoski, A. 2009. River flow forecasting with artificial neural networks using satellite observed precipitation pre-processed with flow length and travel time information: case study of the Ganges river basin. Hydrology and Earth System Sciences, 13: 9. 1607-1618.
5.Arab, S., Khashei, S.A., Pourreza, B.M., and Hashemi, S.R. 2018. Comparison of Two Nonparametric Models, K-nearest neighbor and M5 Decision Tree in Forecasting the River Discharge in the Karaj Catchment. Watershed Management Research Journal. 30: 4. 47-58. (In Persian)
6.Asadi, H., Honarmand, M., Vazifedoust, M., and Mousavi, A. 2017. Assessment of Changes in Soil Erosion Risk Using RUSLE in Navrood Watershed, Iran. J. Agric. Sci. Tech. 19: 231-244.
7.Bashari, M., and Vatankhah, M. 2011. Comparison of different time series analysis methods for forecasting monthly discharge in Karkheh watershed. Irrigation and Water Engineering,1: 2. 75-86. (In Persian)
8.Bhattacharya, B., and Solomatine, D.P. 2005. Neural networks and M5 model trees in modelling water level–discharge relationship. Neurocomputing, 63: 381-396.
9.Cheng, C.T., Niu, W.J., Feng, Z.K., Shen, J.J., and Chau, K.W. 2015. Daily reservoir runoff forecasting methodusing artificial neural network basedon quantum-behaved particle swarm optimization. Water, 7: 8. 4232-4246.
10.Cryer, J.D. 1992. Time series Analysis. Translated by Niroomand, H.A., Mashhad University Publication, 404p. (In Persian)
11.Eshghi, P., Farzadmehr, J., Dastorani, M.T., and Arabasadi, Z. 2016. The effectiveness of intelligent models in estimating the river suspended sediments (Case Study: Babaaman Basin, Northern Khorasan. JWMR. 2017; 7: 14. 95-88. (In Persian)
12.Esmaeili, H., Akhond Ali, A.M., Zarei, H., and Taghian, M. 2018. Regional Flood Analysis Via Comparison of
The M5 Decision Tree Algorithmand Regression Models. Irrigation Sciences and Engineering, 40: 4. 183-195. (In Persian)
13.Fathi, P., Mohammadi, Y., and Homaei, M. 2009. Intelligent modeling of monthly flow time series into vahdat dam in sanandaj city. Journal of Water and soil, 21: 1. 209-220. (In Persian)
14.Ghorbani, K., Naeimi Kalourazi, Z., Salarijazi, M., and Dehghani, A.A. 2016. Estimation of monthly discharge using climatological and physiographic parametesr of ungauged basin. Journal of Water and Soil Conservation,23: 3. 207-224. (In Persian)
15.Ghorbani, K., Sohrabian, E., and Salarijazi, M. 2016. Evaluation of hydrological and data mining models in monthly river discharge simulation and prediction (Case study: Araz-Kouseh watershed). Journal of Water and Soil Conservation, 23: 1. 203-217. (In Persian)
16.Hadizadeh, R., Eslamian, S., and Chinipardaz, R., 2013, Investigation of long-memory properties in streamflow time series in Gamasiab River, Iran’, Int. J. Hydrology Science and Technology, 3: 4. 319-350.
17.Haghizadeh, A., Mohammadlou, M., and Noori, F. 2015. Simulation of rainfall-runoff process using multilayer perceptron and adaptive neuro-fuzzy interface system and multiple regression (Case Study: Khorramabd Watershed. Iranian journal of Ecohydrology,2: 2. 233-243. (In Persian)
18.Jandaghi, N., Azimmohseni, M., and Ghareh Mahmoodlu, M. 2021. Rainfall-runoff process modeling using time series transfer function. Environmental Erosion Research Journal, 11: 2. 111-128. (In Persian)19.Kang, K.W., Kim, J.H., Park, C.Y.,
and Ham, K.J. 1993. Evaluation of hydrologic forecasting system based on neural network model. In Proceedings of the Congress-international Association for research. 1: 257-257.
20.Khodakhah, H., Aghelpour, P., and Hamedi, Z. 2022. Comparing linear and non-linear data-driven approaches in monthly river flow prediction, based on the models SARIMA, LSSVM, ANFIS, and GMDH. Environmental Science and Pollution Research, 29: 15. 21935-21954.
21.Kia, I., Emadi, A.R., and Gholami, M. 2019. Rainfall-Runoff Modeling by Adaptive Neuro-Fuzzy Inference System (ANFIS) and Multi-Variable Linear Regression (MLR). Irrigation and Water Engineering, 9: 4. 39-51. (In Persian)
22.Kratzert, F., Klotz, D., Brenner, C., Schulz, K., and Herrnegger, M. 2018. Rainfall–runoff modelling using long short-term memory (LSTM) networks. Hydrology and Earth System Sciences, 22: 11. 6005-6022.
23.Krstanovic, P.F., and Singh, V.P. 1991. A univariate model for long-term streamflow forecasting. Stochastic hydrology and hydraulics, 5: 3. 189-205.
24.Laux, P., Vogl, S., Qiu, W., Knoche, H. R., and Kunstmann, H. 2011. Copula-based statistical refinement of precipitation in RCM simulations over complex terrain. Hydrology and Earth System Sciences, 15: 7. 2401-2419.
25.Mantua, N.J., Hare, S.R., Zhang, Y., Wallace, J.M., and Francis, R.C.1997. A Pacific interdecadalclimate oscillation with impacts on salmon production. Bulletin of the american Meteorological Society,78: 6. 1069-1080.
26.Masoumpour Samakosh, J., Jalilian, A., and Yari, E. 2017. The analysis of seasonal precipitation time series in Iran. Physical Geography Research Quarterly, 49: 3. 457-475. (In Persian)
27.Mirzapour, H., Haghizadeh, A., and Alijani, R. 2018. The Evaluation of the Performance of SARIMA Time Series Model in the Simulation of the Average Monthly Discharge of the Kashkan Afrineh and Kakareza Rivers (Lorestan province). Hydrogeomorphology,4: 55. 153-169. (In Persian)
28.Naeimi Kalourazi, Z., Kh, G.,Salarijazi, M., and Dehghani, A.A. 2016. Estimation of monthly discharge using climatic and physiographic parameters of ungauged basins. Journal of Water and Soil Conservation,23: 3. 207-224. (In Persian)
29.Panahi, A., and AIijani, B. 2013. Forcasting Peek Flood In The Madarsoo Basin Using Neural Network And Variable Several Regressions Method (Case Study: Madar Soo Basin). Geography, 11: 38. 113-132. (In Persian)
30.Poul, A.K., Shourian, M., and Ebrahimi, H. 2019. A comparative study of MLR, KNN, ANN and ANFIS models with wavelet transform in monthly stream flow prediction. Water Resources Management, 33: 8. 2907-2923.
31.Salahi, B., and Sarmasti, T. 2014. Simulation of Runoff-runoff presses in southern Sub-basin of Gharesou by Artificial neural networks mode (ANNS). Geography and environmental planning, 24: 4. 119-134. (In Persian)
32.Salarijazi, M., Ghorbani, K., Sohrabian, E., and Abdolhosseini, M. 2016. Prediction of Daily Stream-flow Using Data Driven Models. Iranian Journal of Irrigation & Drainage, 10: 4. 479-488. (In Persian)
33.Samadi, M., Bahremand, A., and Fathabadi, A. 2020. The Boustan Dam monthly inflow forecasting usingdata-driven and ensemble models inthe Golestan Province, Watershed Engineering and Management,11: 4. 1044-1058. (In Persian)
34.Solomatine, D.P., and Dulal, K.N. 2003. Model trees as an alternative to neural networks in rainfall-runoff modelling. Hydrological Sciences Journal,48: 3. 399-411.
35.Sudheer, K.P., Gosain, A.K., and Ramasastri, K.S. 2002. A data‐driven algorithm for constructing artificial neural network rainfall‐runoff models. Hydrological processes, 16: 6. 1325-1330.
36.Talebi, A., and Akbari, Z. 2013. Investigation of ability of decision trees model to estimate river suspended sediment (case study: Ilam Dam basin). Journal of Science and Technology of Agriculture and Natural Resources,17: 63. 109-121. (In Persian)
37.Valipour, M., Banihabib, M.E., and Behbahani, S.M.R. 2013. Comparison of the ARMA, ARIMA, and the autoregressive artificial neural network models in forecasting the monthly inflow of Dez dam reservoir. Journal of hydrology, 476: 433-441.
38.Vyas, S.K., Mathur, Y.P., Sharma, G., and Chandwani, V. 2016. December. Rainfall-Runoff Modelling: Conventional regression and Artificial Neural Networks approach. In 2016 International Conference on Recent Advances and Innovations in Engineering (ICRAIE) (pp. 1-6). IEEE.
39.Wang, W.C., Chau, K.W., Xu, D.M., and Chen, X.Y. 2015. Improving forecasting accuracy of annual runoff time series using ARIMA based on EEMD decomposition. Water Resources Management, 29: 8. 2655-2675.
40.Witten, I.H., and Frank, E. 2002. Data mining: practical machine learning tools and techniques with Java implementations. Acm Sigmod Record, 31: 1. 76-77.
41.Wood, E.F., and Rodríguez‐Iturbe, I. 1975. Bayesian inference and decision making for extreme hydrologic
events. Water Resources Research,11: 4. 533-542.
42.Wu, C.L., and Chau, K.W. 2010. Data-driven models for monthly streamflow time series prediction. Engineering Applications of Artificial Intelligence, 23: 8. 1350-1367.
43.Zahiri, A.R., and Ghorbani, K.H. 2013. Flow discharge prediction in compound channels by using decision model
tree M5. Journal of Water andSoil Conservation, 20: 3. 113-132.(In Persian)