مقایسه کارایی رگرسیون بردار پشتیبان در پیش‌بینی خشکسالی با استفاده از پیوند از دور و پارامترهای اقلیمی در نمونه‌های اقلیمی ایران

نوع مقاله : مقاله پژوهشی

نویسندگان

1 استادیار، گروه ریاضی و آمار، دانشکده علوم پایه، دانشگاه هرمزگان، بندرعباس، ایران

2 دانشیار، گروه مهندسی منابع طبیعی، دانشکده کشاورزی و منابع طبیعی، دانشگاه هرمزگان، بندرعباس، ایران

چکیده

هدف از تحقیق حاضر، پیش‌بینی خشکسالی کشاورزی با استفاده از سیگنال‌های بزرگ مقیاس و متغیرهای اقلیمی در هشت نمونه‌ اقلیمی ایران است. براین اساس، با استفاده از رگرسیون لاسو، مهمترین متغیرها در هر اقلیم مشخص و با استفاده از رگرسیون بردار پشتیبان و سه تابع خطی، شعاعی و چندجمله‌ای، خشکسالی پیش‌بینی گردید. نتایج نشان داد، در اقلیم فراخشک معتدل و نیمه خشک سرد تابع خطی و در سایر اقلیم‌ها تابع شعاعی مناسب است. بر اساس نتایج، مقدار توافق بین مقدار پیش‌بینی کننده و پیش‌بینی شونده 912/0 تا 731/0 براوردگردید. در بررسی خطای مدل براساس PE یا فرکانس خطا، بیش از 55% خطا ناچیز و مربوط به دسته 5/0± و 27% مربوط به دسته 5/0± تا 1± است که نشان دهنده کارایی مناسب مدل در برآورد SPEI است. برای بررسی عملکرد مدل در پیش‌بینی وقایع خشکسالی در طول دوره آزمایش، متغیرهای شدت خشکسالی، مدت زمان، شدت اوج و بزرگی مورد بررسی قرار گرفت. در متغیر شدت، بزرگی و پیک خشکسالی عموما مدل دچار کم برآورد شده، به‌جز در اقلیم فراخشک و معتدل و نیمه خشک سرد که تابع خطی، رفتاری متفاوت را نشان داد. بیشترین اختلاف بین مقادیر مشاهد‌ه‌ای و پیش‌بینی شده در متغیر شدت خشکسالی در صدک 75ام در اقلیم خشک و گرم (بوشهر) مشاهده گردید. در نهایت، می‌توان بیان کرد، مدل SVR در پیش‌بینی SPEI برای اکثر اقلیم‌ها بسیار کارآمد است. با این حال، عملکرد آن در مناطق متنوع جغرافیایی متفاوت به نظر می رسد، که شاید نشان‌دهنده نقش متفاوت رگرسیون‌های مورد استفاده در آموزش مدل و متغیرهای مختلف باشد.

کلیدواژه‌ها


عنوان مقاله [English]

Comparing the efficiency of several support vector regressions in predicting agricultural drought using teleconnection and climatic parameters in Iranian climatic samples

نویسندگان [English]

  • Hossein Zamani 1
  • Ommolbanin bazrafshan 2
1 Assistant professor, Department of Mathematics and Statistics, Faculty of Science, University of Hormozgan, Bandarabbas, Iran.
2 university of Hormozgan
چکیده [English]

Introduction

Drought is a complex phenomenon which is difficult to define. It is a creeping phenomenon that slowly sneaks up and impacts many sectors of the economy, and operates on many different time scales. Although low rainfall is the source of any drought however, delaying rainfall has more effects on water resources and causes a great damage. Therefore, forecasting drought, especially in long term drought, is very important in agricultural water management and planning.

Although reduced rainfall is very important on drought occurrence, but increasing or decreasing temperature and evapotranspiration can result in a sever or moderate drought situation. Several indices have been developed based on rainfall and evapotranspiration for drought analysis, however different indices significantly have different results. Because, drought is a multivariate phenomenon and along with the precipitation, the evapotranspiration factor also must be included in drought analysis, especially in the arid and semi-arid regions. The sequence theory, stochastic models, and conceptual models are most popular approaches in forecasting drought events. However, deterministic prediction of drought situation has been more noticeable for researchers in the recent years.

Data-driven methods could aspire to extract relations that may further informand augment the current physical understanding. Support Vector Machine (SVM) is one of the data-driven algorithms which has been successfully applied in classification, regression and forecasting in the field of hydrology .It has been emphasized the need of combining an understanding of physics with data mining, not only to avoid generating misleading insight but also to produce new results. There are a number of studies using SVM in drought forecasting. The advantage of SVR is that it could transfer a non-linear problemto a linear problem using the kernel function, and be effective in solving a high dimension problem.



Materials and methods

This study investigates the effect of climatic variables (precipitation, maximum and minimum temperature, evapotranspiration) and large-scale climatic signals arising from temperature fluctuations in the Pacific and Atlantic levels (SST) on the SPEI variable. The present study has been conducted in different climatic types of Iran. Accordingly, using the extended De Martone classification climate method, climatic types were identified and sample synoptic stations were selected from 8 climatic types. We applied SVR machine learning algorithm for prediction the SPEI, based on several meteorological predictors during the 1966–2014. In order to learn the SVR model we considered 80% of dataset for training and 20% of the rest as the testing dataset. We also applied the Lasso regression approach to select the important variables affecting the SPEI. In this regard, three different kernel functions were used for training the support vector regression (SVR) including linear, radial and polynomial kernels and the results were evaluated using different criterions.



Results and discussion

Lasso regression results showed that in all climates, the most effective drought variable is precipitation. Among the teleconnection signals in all climates in Iran (except semi-arid cold and extra-arid warm), the effect of SST in the North Atlantic and South Pacific is effective on agricultural drought. The results showed that in all climates, the radial function showed the best results (except in the semi-arid cold, extra-arid climates) where the linear function had the best performance. The agreement index between the predictive and predictable values in the best model is 0.912 to 0.731, which is very humid and arid-warm climate, respectively. Therefore, the highest accuracy is related to Ramsar station and the lowest accuracy is related to Bushehr station. To evaluate the performance of the SVR model in predicting drought in training phase, the variables of drought including severity, duration and peak and magnitude were examined. The model is generally underestimated in estimating the severity, magnitude, and peak of drought (except at Tabas and Urmia stations, where the linear function showed different behaviors in the extra-arid warm and semi-arid cold climates). The largest difference between the observed and predicted values of drought severity was observed in the 75th percentile (Bushehr station).

Conclusion

Finally, it can be said that SVR is very efficient in predicting SPEI for most climates. However, its performance in different climates seems to be different, which may indicate the different role of regressions used in training SVM model and different variables. Finally, the results show that SVR is an attractive machine learning tool for drought prediction and can provide useful information for water management in agriculture.

کلیدواژه‌ها [English]

  • Drought
  • Large climatic Signals
  • Support Vector Regression
  • Lasso Regression
  1. Abbasi, A., K. Khalili, J. Behmanesh, A. Shirzad, 2020, Application of support vector machine and Bayesian network for agricultural drought prediction. J. Watershed engineering and management, NO. 12(1), PP. 107-124.
  2. Abbasi, A., S. Nabavi, 2010, The effect of remote linking patterns on Iranian rains. J. Geography.
  3. Azhdari Moghadam, M., Khosravi, M., Hosseinpour Niknam. H., Jaffari Nadoushan, E. 2011. Prediction of drought using Fuzzy-Neural network model, climatic indices, precipitation and drought index. Kournal of Geography and Development, 10(26): 61-72.
  4. Azizi, gh., 2000, El nino and drought- wet periods in the climate of iran. J. Geographical reseach quarterly, NO. 32(38), PP. 71-84.
  5. Bazrafshan, O., Mahmoudzadeh, A.  Asgarinezhad, J. Bazrafshan,  2019, Adaptive Evalution of SPI, RDI, and SPEI indices in analyzing the trend of intensity, duration, and frequency of drought in arid and semi- arid regions of iran. J. Agriculture, NO. 42(3), PP. 117-131.
  6. Blenkinsop, S., H. J. Fowler, 2007, Changes in drought frequency, severity and duration for the British Isles projected by the PRUDENCE regional climate models. J. Hydrology, NO. 342, PP. 50-71.
  7. Boser, B., I. Guyon, V. Vapnik, 1992, A training algorithm for optimal margin classi_ers. In Proceedings of the Fifth Annual Workshop on Computational Learning Theory.
  8. Cordery, I., M. McCall, 2000, A Model For Forecasting Drought From Teleconnections. J. Water Resources Research, NO. 36(3), PP. 763–76.
  9. Cortes, C., V. Vapnik, 1995, Support-vector network. Machine Learning, NO. 20, PP. 273-297.
  10. Deo, R. C., S. Salcedo-Sanz, L. Carro-Calvo, B. Saavedra-Moreno, 2018, Drought prediction with standardized precipitation and evapotranspiration index and support vector regression models, J. In Integrating disaster science and management , pp. 151-174.
  11. Fatehi Marj, A., A. Borhani darian, M. H. Mahdian, 2006, Seasonal rainfall using teleconnection pattern case stady: orumiyeh lake basin. J. Soil and water sciences, NO. 10(3), PP. 45-58.
  12. Friedman, J., T. Hastie, R. Tibshirani, 2001, The elements of statistical learning, NO. 10.
  13. Gandomkar, A., Y. Toloei, K. Esmaeili, R. Deghani, M. pour Gholami, 2014, Investigation of the effect of NAO fluctuations on precipitation Ardabil station. J. Intrnational conference on sustainable development, strategies and challenges.
  14. Gerkaninezhad moshozi, Z., O. Bazrafshan, 2018, Impact of climatic signals on the wet and dry season precipitation (Case study: Persian gulf and oman sea watersheds). J. The earth and space physics, NO. 44(2), PP. 333-349.
  15. Ghasemieh, H., O. Bazrafshan, k. Bakhshayesh manesh, 2017, Artificial neural network for monthly rainfall forecasting using teleconnection patterns (Case study: central plateau basin of iran). J. The earth space physics, NO. 43(2), PP. 405-418.
  16. Ghazipour, F., & Mahjouri, N. (2022). A multi-model data fusion methodology for seasonal drought forecasting under uncertainty: Application of Bayesian maximum entropy. Journal of Environmental Management, 304, 114245.
  17. Iraqinejad, Sh., M. Karamooz, 2018, Advanced hydrology, Amirkabir university. pp. 468.
  18. James, G., D. Witten, T. Hastie, R.  Tibshirani, 2013,  An introduction to statistical learning,  18.
  19. Julio Canon, J., J. Gonzalez, J. Valde’s, 2007, Precipitation in the Colorado River Basin an Its Low Frequency Associations With PDO and ENSO Signals, J. Hydrology, NO. 344, PP. 252-264.
  20. Kao, H. Y., J. Y. Yu, 2009,  Contrasting eastern-Pacific and central-Pacific types of ENSO, J. Climate, NO. 22(3), PP. 615-632.
  21. Kushnir, Y., W. A. Robinson, I. Bladé, N. M. J.  Hall, S.  Peng, R. Sutton, 2002, Atmospheric GCM response to extratropical SST anomalies: Synthesis and evaluation, J. Climate, NO. 15(16), PP. 2233-2256.
  22. Lin, H. T., C. J. Lin, 2003,  A study on sigmoid kernels for SVM and the training of non-PSD kernels by SMO-type methods. J. submitted to Neural Computation, NO. 3(1-32), PP. 16.
  23. Malik, A., Kumar, A., Rai, P., & Kuriqi, A. (2021). Prediction of Multi-Scalar Standardized Precipitation Index by Using Artificial Intelligence and Regression Models. Climate, 9(2), 28.
  24. Mendicino, G., A. Senatore, P. Versace, 2008, Groundwater resource index (GRI) for drought monitoring and forecasting in a Mediterranean climate, J. Hydrology, NO. 357, PP. 282-302.
  25. Morid, S., V. Smakhtin, K. Bagherzadeh, 2007, Drought forecasting using artificial neural networks and time series of drought indices, J. Climatology, NO.  27(15), PP. 2103-2111.
  26. Nezamolsadat, S. 1999, Investigation of the effect of El nino phenomenon- southern fluctuations (ENSO) on autumn rainfall in iran. J. Regional conference on climate change, meteorological organization.
  27. Sebghati, M., H. Ahmadi Pirgani, A. Moghaddam, 2016, The calculation of continuity and intensity of droughts using modified SPEI index (Case study: Tabriz and Urmia cities). J. Environment and water engineering, NO. 2(2), PP. 188-195.
  28. Somali, R., O. Bazrafshan, M. Biniaz, H. Moslemi, 2019, Investigation of the relationship between the phenomenon of enso on the drought and wet year in the douthern coastal provinces of iran. J. Irrigation and drainage, NO. 13(1), PP. 217-231.
  29. Tian, Y., Y. P. Xu, G. Wang, 2018, Agricultural drought prediction using climate indices based on Support Vector Regression in Xiangjiang River basin, J. Science of The Total Environment, NO.  622, PP. 710-720.
  30. Vicente-Serrano, S. M., S. Beguería, J. I. López-Moreno,  2010, A multiscalar drought index sensitive to global warming: the standardized precipitation evapotranspiration index, J.  climate, NO. 23(7), PP. 1696-1718.
  31. Zhang, Y., H. Yang, H. Cui, Q. Chen,         2020, Comparison of the ability of ARIMA, WNN and SVM models for drought forecasting in the Sanjiang Plain, China. J. Natural Resources Research, NO. 29(2), PP. 1447-1464.