کاربرد مدل تلفیقی تبدیل موجک و شبکه عصبی مصنوعی در پایش خشک سالی هواشناسی (مطالعه موردی: حوضه آبریز کوهرنگ)

نوع مقاله : مقاله پژوهشی

نویسندگان

1 دانش آموخته کارشناسی ارشد مهندسی منابع آب، دانشگاه شهرکرد.

2 دانشیار گروه مهندسی آب، دانشگاه شهرکرد

3 استادیارگروه مهندسی آب، دانشگاه شهرکرد.

4 دکترای مهندسی منابع آب، دانشگاه شهید چمران اهواز.

چکیده

خشک­سالی یکی از بلایای طبیعی است که سالانه خسارات فراوانی را برای جوامع مختلف به دنبال دارد. ارزیابی و پیش­بینی خشکسالی­ها می­تواند اطلاعات با ارزشی جهت تدوین برنامه­های مقابله با خشک­سالی و کاهش خسارات مربوط به آن در اختیار مدیران و برنامه­ریزان منابع آب بگذارد. در تحقیق حاضر، خشک­سالی­های هواشناسی ایستگاه کوهرنگ در استان چهارمحال و بختیاری با استفاده از شاخص خشک­سالی احیایی (RDI) در دوره آماری 2016-1987 مورد بررسی و تحلیل قرار گرفت. سپس با استفاده از مدل شبکه عصبی مصنوعی (ANN) و مدل تلفیقی موجک با شبکه عصبی مصنوعی (WANN) داده­های ماهانه بارش و تبخیر­تعرق پتانسیل برای سال 2016 پیش­بینی شد. نتایج نشان داد که هرچند مدل ANN در پیش­بینی داده­های تبخیر­تعرق پتانسیل از دقت قابل قبولی برخوردار بود، اما دقت آن در پیش­بینی داده­های بارش مناسب نبوده است. در حالی­که مدل WANN از دقت خوبی در پیش­بینی داده­های بارش ماهانه و تبخیر­تعرق پتانسیل برخوردار بود، به­طوری­که میزان  مرحله تست، در پیش­بینی داده­های ماهانه بارش معادل 69/0 و برای داده­های ماهانه تبخیر­تعرق پتانسیل معادل 99/0 بود که نتایج مطلوب­تری نسبت به مدل شبکه عصبی داشت (میزان  مدل شبکه عصبی مصنوعی برای بارش 52/0 بود). بنابراین، از مدل WANN برای پیش­بینی داده­های بارش و تبخیر­تعرق پتانسیل استفاده شد. در مرحله بعد با استفاده از داده­های پیش­بینی­شده، مقادیر شاخص RDI محاسبه و با مقادیر متناظر این شاخص که با داده­های مشاهداتی محاسبه شده بودند، مقایسه گردید. نتایج نشان داد که مدل WANN عملکرد خوبی در پیش­بینی خشک­سالی کوهرنگ داشته است.

کلیدواژه‌ها

موضوعات


عنوان مقاله [English]

Prediction of Meteorological Droughts in Kuhrang Using the Hybrid Model of Wavelet and Artificial Neural Network

نویسندگان [English]

  • Marziyeh Bahrami Samani 1
  • Rasoul Mirabbasi Najafabadi 2
  • Ahmad Reza Ghasemi Dastgerdi 3
  • Sajjad Abdollahi AsadAbadi 4
1 MSc. Student of Water Resources Engineering, Shahrekord University, Shahrekord, Iran
2 Associate Professor, Department of Water Engineering, Shahrekord University Address: Water Eng. department, Faculty of Agriculture, Shahrekord University, Shahrekord, Iran.
3 3- Assistant Professor, Department of Water Engineering, Shahrekord University Shahrekord, Iran
4 PhD. of Water Resources Engineering, Department of Water Engineering, Shahid Chamran University od Ahvaz, Ahvaz, Iran
چکیده [English]

Meteorological drought is defined as a lack of rainfall over long periods, which reduces soil moisture and river flow. One of the critical drought assessment tools is drought indices (Tsakiris & Vangelis, 2005). So far, many drought indicators have been developed by researchers, for example, the RDI )Reconnaissance Drought Index) (Tsakiris & Vangelis, 2005). The difference between this index and other drought indices is that it is estimated based on two variables of precipitation and potential evapotranspiration. For this reason, it is more accurate than indices that are calculated only based on precipitation. So far, some studies have been used the RDI for drought assessment. Zarei et al. (2016) studied the spatial pattern of drought using the RDI index in southern Iran. The results showed that the area with dry conditions had an increasing trend. Asadi Zarch (2017) investigated the drought trend in Yazd province between 1966-2009 using the RDI index. The results showed that drought occurrence in Yazd increased during the studied period. Because, unlike other natural disasters, it is difficult to accurately determine the onset and the end of the drought period (Moried et al., 2005). Accordingly, it is difficult to diagnose and evaluate the drought phenomenon. Therefore, monitoring and predicting drought is very important in water resources management. The use of wavelets is a new and very effective way of analyzing signals and time series. Application of Wavelet in Wavelet- Artificial Neural Network (WANN) models as a function for training has recently been used as an alternative for Artificial Neural Network (ANN) models. In recent years, the combination of wavelet theory and artificial neural networks has led to the development of wavelet neural networks (Thuillard, 2000). Zhang et al. (2017) applied the ARIMA, ANN, WANN, and Support Vector Regression (SVR) models to predict droughts in China's northern Haihe River basin using the SPI index. The results showed that the WANN model performed better than other considered models for predicting the SPI values at 6 and 12 months time scales. This study aimed to predict the meteorological droughts in the Kuhrang region using ANN and WANN models. To this end, the efficiency of ANN and WANN models in predicting precipitation and potential evapotranspiration will be evaluated. Then, the Resilience Drought Index (RDI) will be calculated based on the predicted values by ANN and WANN to describe and prediction of Kuhrang wetness conditions.

کلیدواژه‌ها [English]

  • Meteorological Drought
  • WANN Model
  • ANN Model
  • SPI Index
  • RDI Index
  • Kuhrang
  • Adamowski, J. and Sun, K., 2010. Development of a coupled wavelet transform and neural network method for flow forecasting of non-perennial rivers in semi-arid watersheds. Journal of Hydrology, 390, pp.85-91.

 

  • Anshuka, A. Floris, F. and Rutger, W., 2019. Drought forecasting through statistical models using standardised precipitation index: a systematic review and meta‑regression analysis. Journal of the International Society for the Prevention and Mitigation of Natural Hazards, 97, pp.955-977.

 

  • AsadiZarch, M.A., 2017. Analyzing climate change effects on drought occurrence in Yazd province, Iran. Scientific Association of Desert Management and Control, 9, pp.74-90. (In Persian).

 

  • Cannas, B., Fanni, A., See, L. and Sias, G., 2006. Data preprocessing for river flow forecasting using neural network, wavelet transforms and data partitioning. Physics and Chemistry of the Earth, 31(18), pp.1164-1171.

 

  • Daubechies, I., 1992. Ten lectures on wavelets. In CBMS-NSF Regional Conference Series in Applied Mathematics Philadelphia, PA. USA.

 

6- Djerbouai, S. and Souag-Gamane, D., 2016. Drought forecasting using neural networks, Wavelet neural networks, and stochastic nodels. Water Resources Management, 30, pp.2445–2464.

  • Ghamarnia, H., Rezvani, V., Khodaei, E. and Mirzaei, H., 2012. Time and place calibration of the Hargreaves equation for estimating monthly reference evapotranspiration under different climatic conditions. Journal of Agricultural Science, 4(3), pp.111-122. (In Persian).

 

  • Hargreaves, G.H. and Samani, Z.A., 1982. Estimating potential evapotranspiration. Journal of Irrigation and Drainage Engineering, 108, pp.223-230.

 

  • Hassanzadeh, Y., Abdi Kordani, A. and Fakheri Fard, A., 2013. Drought forecasting using genetic algorithm and conjoined model of neural network-wavelet. Journal of Water and Wastewater, 23, pp.48-59. (In Persian).

 

  • Labat, D., Ronchail, J. and Guyot, J.L., 2005. Recent advances in wavelet analyses: Part 2-Amazon, Parana, Orinoco and Congo discharges time scale variability. Journal of Hydrology, 314, pp.289-311.

 

  • Mallat, S.G., 1989. A theory for multiresolution signal decomposition: the wavelet representation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 11, pp.674-693.

 

  • Mason, J.C., Price, R.K. and Tem, M.E., 1996. A neural network model of rainfall-runoff using radial basis functions. Journal of Hydraulic Research, 34, pp.537-548.

 

  • Mckee, T.B., Doesken, N.J. and Kleist, J., 1993. The relationship of drought frequency and duration on time scales. In 8th Conference on Applied Climatology, Anaheim university, California.

 

  • Mendicino, G., Senatore, A. and Versace, P., 2008. A groundwater resource index (GRI) for drought monitoring and forecasting in a Mediterranean Journal of Hydrology, 357, pp.282-302.

 

  • Moried, S., Moghaddasi, M., Paemozd, S.H. and Ghaemi, H., 2005. Designing drought monitoring system of Tehran province. Applied Research Report Ministry of Energy.

 

  • Nourani, V., Komasi, M. and Mano, A., 2009a. A multivariate ANN-Wavelet approach for rainfall-runoff modeling. Water Resources Management, 23, pp.2877-2894.

 

  • Nourani, V., Alami, M.T. and Aminfar, M.H., 2009b. A combined neural-wavelet model for prediction of Ligvanchai watershed precipitation. Journal of Engineering Application of Artificial Intelligence, 22, pp.466-472. (In Persian).

 

  • Palmer, W.C., 1965. Meteorological drought. Research Paper No. 45. US Department of Commerce Weather Bureau. Washington DC.

 

  • Polikar, R., 1996. Fundamental concept and an overview of the wavelet theory wavelet tutorial. Rowan University. Glassboro.

 

20- Raziei, T., Daneshkar, A., Akhtari, R. and Saghafian, B., 2007. Investigation of meteorological droughts in the Sistan and Balouchestan province using the standardized precipitation index and Markov chain model. Water Resources Research, 3(1), pp.25-35. (In Persian).

 

21- Sifuzzaman, M., Islam, M.R. and Ali, M.Z., 2009. Application of wavelet transform and its advantages compared to fourier transform. Journal of Physical Sciences, 13, pp.121-134.

 

22- Thuillard, M., 2000. A review of wavelet networks, wavelet, fuzzy wavelet and their application. ESIT 2000 Aachen. Germany.

 

23- Tigkas, D., Vangelis, H. and Tsakiris, G., 2016. Introducing a modified reconnaissance drought index (RDIe) incorporating effective precipitation. Procedia Engineering, 162, pp.332-339.

 

24- Tsakiris, G. and Vangelis, H., 2005. Establishing a drought index incorporating evapotranspiration. European Water, 10, pp.3-11.

 

25- Wang, W. and Ding, J., 2003. Wavelet network model and its application to the prediction of hydrology. Nature and Science, 1, pp.67-71.

 

26- Zarei, A.R., Moghimi, M.M. and Mahmoudi, M.R., 2016. Analysis of changes in spatial pattern of drought using RDI index in south of Iran. Water Resources Management, 11, pp.3723-3743. (In Persian).

 

27- Zarei, M.A., Tabatabaei, S.H., Babazadeh, H. and Sedghi, H., 2013. Determining the best radiation model for Hargreaves-Samani equation in Shahrekord plain under Lysimeter condition. Journal of Water Research, 3(9), pp.47-56.

 

28- Zhang, Y., Li, W., Chen, Q., Pu, X. and Xiang, L., 2017. Multi-models for SPI drought forecasting in the north of Haihe River Basin, China. Stochastic Environmental Research and Risk Assessment, 31, pp.2471-2481.