Abstract
The Aegean region is geologically situated at the western end of the Gediz Graben system, influenced by the Western Anatolian Regime. In addition, the region is characterized by various active fault lines that can generate earthquake activity. Numerous earthquakes have been recorded in the region, causing significant material and moral damage from the past to the present. In this study, earthquake data from three different catalogs are examined. The non-clustered catalog is compiled for the years 1970 to 2020, including earthquakes with a moment magnitude (Mw) greater than 3.0. The monthly average magnitudes of earthquakes in the region are obtained and analyzed using ARIMA, singular spectrum analysis (SSA), and deep learning methods including convolutional neural network (CNN) and long short-term memory (LSTM), as these methods have not been compared for the region previously. Each method has a different benefit. ARIMA analyzes time series trends and seasonal patterns, while SSA focuses on decomposition and feature extraction. LSTM attempts to capture complex relationships using memory mechanisms, while CNN is powerful at pattern recognition and extracting important features. Thanks to this diversity, our study allows for more comprehensive and reliable forecasts of average earthquake magnitudes for the next 36 periods. The estimation capabilities and error rates of each method were analyzed based on earthquake magnitude data, and it was determined that the LSTM method provided the most effective and accurate predictions.
Similar content being viewed by others
Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
Abdalzaher MS, El-Hadidy M, Gaber H, Badawy A (2020) Seismic hazard maps of Egypt based on spatially smoothed seismicity model and recent seismotectonic models. J Afr Earth Sc 170:103894
Abdalzaher MS, Soliman MS, El-Hady SM, Benslimane A, Elwekeil M (2021) A deep learning model for earthquake parameters observation in IoT system-based earthquake early warning. IEEE Internet Things J 9(11):8412–8424
Akın C & Yağmurlu F (2020) Seismicity of Akşehir Graben. J Adv Eng Stud Technol 1(2), 162–170. Retrieved from https://dergipark.org.tr/en/pub/imctd/issue/59372/830343 (in Turkish)
Akkar S, Azak T, Çan T, Çeken U, Demircioğlu Tümsa MB, Duman TY et al (2018) Evolution of seismic hazard maps in Turkey. Bull Earthquake Eng 16:3197–3228
Al Banna MH, Ghosh T, Nahian MJ, Taher KA, Kaiser MS, Mahmud M, Hossain MS, Andersson K (2021) Attention-based bi-directional long-short term memory network for earthquake prediction. IEEE Access 9:56589–56603
Albawi S, Mohammed TA & Al-Zawi S (2017) Understanding of a convolutional neural network. In: 2017 international conference on engineering and technology (ICET), 1–6, IEEE
Altunel E (1999) Geological and geomorphological observations in relation to the 20 September 1899 Menderes earthquake, Western Turkey. J Geol Soc Lond 156:241–246
Amei A, Fu W & Ho CH (2012) Time series analysis for predicting the occurrences of large scale earthquakes. Int J Appl Sci Technol 2(7):64–75
Bayrak Y, Bayrak E (2012) An evaluation of earthquake hazard potential for different regions in Western Anatolia using the historical and instrumental earthquake data. Pure Appl Geophys 169(10):1859–1873
Bayrak E, Yılmaz Ş, Bayrak Y (2017) Temporal and spatial variations of Gutenberg—Richter parameter and fractal dimension in Western Anatolia, Turkey. J Asian Earth Sci 138:1–11
Berhich A, Belouadha F & Kabbaj MI (2021) LSTM-based Models for Earthquake Prediction, Conference: NISS2020: The 3rd International Conference on Networking, Information Systems & Security
Bhandarkar T, Satish N, Sridhar S, Sivakumar R, Ghosh S (2019) Earthquake trend prediction using long short-term memory RNN. Int J Elect Comput Eng 9(2):1304–1312
Bozkurt E (2000) Timing of extension on the Büyük Menderes Graben, Western Turkey and its tectonic implications. Geol Soc Lond 173(1):385–403 (Special Publications)
Cao J, Li Z, Li J (2019) Financial time series forecasting model based on CEEMDAN and LSTM. Physica A 519:127–139
Cekim HO, Tekin S, Ozel G (2021) Prediction of the earthquake magnitude by time series methods along the East Anatolian Fault, Turkey. Earth Sci Inform 14(3):1339–1348
Cho K, Van Merriënboer B, Gulcehre C, Bahdanau D, Bougares F, Schwenk H & Bengio Y (2014) Learning phrase representations using RNN encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078
Coban KH, Sayil N (2019) Evaluation of earthquake recurrences with different distribution models in western Anatolia. J Seismolog 23:1405–1422
Coban KH, Sayil N (2022) Magnitude type conversion models for earthquakes in turkey and its vicinity with machine learning algorithms. J Earthquake Eng 27(9):2533–2554
Cohen HA, Dart CJ, Akyüz HS, Barka A (1995) Syn-rift sedimentation and structural development of the Gediz and Büyük Menderes graben, Western Turkey. J Earth Soc Lond 152:629–638
Coussin M (2022) Singular spectrum analysis for real-time financial cycles measurement. J Int Money Financ 120:102532
Danese M, Lazzari M & Murgante B (2009) Integrated Geological, Geomorphological and Geostatistical analysis to study macroseismic effects of 1980 Irpinian earthquake in urban areas (southern Italy). In International conference on computational science and ıts applications (pp 50–65). Springer, Berlin, Heidelberg
Dobilas S (2022) LSTM Recurrent Neural Networks—How to Teach a Network to Remember the Past. Medium. https://towardsdatascience.com/lstm-recurrent-neural-networks-how-to-teach-a-network-to-remember-the-past-55e54c2ff22e. Accessed 28 Jun 2022
Durak S (2008) The masonry structures commonly used in the Aegean Region and the earthquake safety of these structures (Master's thesis. Pamukkale University Institute of Science and Technology) (in Turkish)
Elhadidy M, Abdalzaher MS, Gaber H (2021) Up-to-date PSHA along the Gulf of Aqaba-Dead Sea transform fault. Soil Dyn Earthq Eng 148:106835
Elman JL (1990) Finding structure in time. Cogn Sci 14(2):179–211
Emre Ö, Duman TY, Özalp S, Şaroğlu F, Olgun Ş, Elmacı H, Çan T (2018) Active fault database of Turkey. Bull Earthq Eng 16(8):3229–3275
Erdik M, Demircioglu M, Sesetyan K, Durukal E, Siyahi B (2004) Earthquake hazard in Marmara region, Turkey. Soil Dyn Earthq Eng 24(8):605–631
Fuentes AG, Nicolis O, Peralta B, Chiodi M (2022) Spatio-temporal seismicity prediction in Chile using a multi-column ConvLSTM. IEEE Access 10:107402–107415
Gao W, Guo J, Zhou M, Yu H, Chen X & Ji B (2020) Gravity tides extracted from SSA-denoised superconducting gravity data with the harmonic analysis: a case study at Wuhan station. China. Acta Geodaetica et Geophysica 1–17
Gardner JK, Knopoff L (1974) Is the sequence of earthquakes in Southern California, with aftershocks removed, Poissonian? Bull Seismol Soc Am 64(5):1363–1367
Gers FA & Schmidhuber J (2000) Recurrent nets that time and count. In IJCNN, Vol. 3 (pp 189–194). IEEE
Ghil M, Vautard R (1991) Interdecadal oscillations and the warming trend in global temperature time series. Nature 350(6316):324–327
Ghil M, Allen MR, Dettinger MD, Ide K, Kondrashov D, Mann ME et al (2002) Advanced spectral methods for climatic time series. Rev Geophys 40(1):1–3
Gioncu V, Mazzolani F (2011) Earthquake engineering for structural design. CRC Press, London
Golyandina N, Zhigljavsky A (2013) Singular Spectrum Analysis for time series. Springer Science & Business Media
Golyandina N, Nekrutkin V, Zhigljavsky AA (2001) Analysis of time series structure: SSA and related techniques. CRC Press
Graves A, Schmidhuber J (2005) Framewise phoneme classification with bidirectional lstm and other neural network architectures. Neural Netw 18(5–6):602–610
Gu J, Wang Z, Kuen J, Ma L, Shahroudy A, Shuai B et al (2018) Recent advances in convolutional neural networks. Pattern Recogn 77:354–377
Hasan Al Banna M, Ghosh T, Taher KA, Kaiser MS & Mahmud M (2021) An earthquake prediction system for bangladesh using deep long short-term memory architecture. In: Intelligent Systems: Proceedings of ICMIB 2020 (pp 465–476). Springer Singapore
Hassani H (2007) Singular spectrum analysis: methodology and comparison, MPRA Paper No: 4991
Hochreiter S, Schmidhuber J (1997) Long short-term memory. Neural Comput 9(8):1735–1780
Huang JP, Wang XA, Zhao Y, Xin C, Xiang H (2018) Large earthquake magnitude prediction in Taiwan based on deep learning neural network. Neural Network World 28(2):149–160
Kadılar C, Öncel Çekim H (2020) Introduction to SPSS and R applied time series analysis. Seçkin Publication (in Turkish)
Kadirioğlu FT, Kartal RF (2016) The new empirical magnitude conversion relations using an improved earthquake catalogue for Turkey and its near vicinity (1900–2012). Turkish J Earth Sci 25(4):300–310
Karadaş A, Öner E (2021) The effects of alluvial geomorphology of the Bornova plain on the damage caused by the 30 October 2020 Samos earthquake in İzmir-Bayraklı. J Geogr 42:139–153 (in Turkish)
Kaynar O, Taştan S (2009) Comparison of Mlp artificial neural networks and purification model in time series analysis. J Erciyes Univ Faculty Econ Admin Sci 33:161–172 (in Turkish)
Keskin S & Külahcı F (2023) ARIMA model simulation for total electron content, earthquake and radon relationship identification. Nat Hazards 115(3):1955–1976
Ketin İ (1968) Relationships between general tectonic situation of Turkey and major earthquake zones. J Min Res Explor 71(71):129–134 ((in Turkish))
Lakshmi SS, Tiwari RK (2009) Model dissection from earthquake time series: a comparative analysis using modern non-linear forecasting and artificial neural network approaches. Comput Geosci 35(2):191–204
LeCun Y, Bottou L, Bengio Y, Haffner P (1998) Gradient-based learning applied to document recognition. Proc IEEE 86(11):2278–2324
LeCun Y, Boser B, Denker J, Henderson D, Howard R, Hubbard W & Jackel L (1989) Handwritten digit recognition with a back-propagation network. Adv Neural İnf Process Syst 2
Li T, Zhang Y, Wang T (2021) SRPM–CNN: a combined model based on slide relative position matrix and CNN for time series classification. Complex Intell Syst 7:1619–1631
Lin Y, Ling BWK, Xu N, Zhou X (2022) Two dimensional quaternion valued singular spectrum analysis with application to image denoising. J Franklin Inst 359(8):3808–3830
Liu CL, Hsaio WH, Tu YC (2018) Time series classification with multivariate convolutional neural network. IEEE Trans Industr Electron 66(6):4788–4797
Livieris IE, Pintelas E, Pintelas P (2020) A CNN–LSTM model for gold price time-series forecasting. Neural Comput Appl 32:17351–17360
Mousavi SM, Beroza GC (2020) A machine-learning approach for earthquake magnitude estimation. Geophys Res Lett 47(1):e2019GL085976
Mouslopoulou V & Hristopulos DT (2011) Patterns of tectonic fault interactions captured through geostatistical analysis of microearthquakes. J Geophys Res 116(B7):1–18
Nicolis O, Plaza F, Salas R (2021) Prediction of intensity and location of seismic events using deep learning. Spatial Stat 42:100442
O'Shea K & Nash R (2015) An introduction to convolutional neural networks. arXiv preprint arXiv:1511.08458.
Paton S (1992) Active normal faulting. Drainage patterns ve sedimentation in Southwestern Turkey. J Earth Soc Lond. 149:1031–1044
Polat O, Gok E, Yilmaz D (2008) Earthquake hazard of the Aegean extension region (West Turkey). Turkish J Earth Sci 17(3):593–614
Putriasari N, Nugroho S, Rachmawati R, Agwil W, Sitohang YO (2022) Forecasting a weekly red Chilli Price in Bengkulu City using autoregressive integrated moving average (ARIMA) and Singular Spectrum Analysis (SSA) methods. J Stat Data Sci 1(1):1–6
Ramanjaneyulu K, Swamy KV & Rao CS (2018) Novel CBIR system using CNN architecture. In: 2018 3rd International conference on inventive computation technologies (ICICT) (pp 379–383). IEEE
Saad OM, Hafez AG, Soliman MS (2020) Deep learning approach for earthquake parameters classification in earthquake early warning system. IEEE Geosci Remote Sens Lett 18(7):1293–1297
Schmidhuber J, Wierstra D, Gagliolo M, Gomez F (2007) Training recurrent networks by evolino. Neural Comput 19(3):757–779
Schoellhamer DH (2001) Singular spectrum analysis for time series with missing data. Geophys Res Lett 28(16):3187–3190
Şengör AMC (1979a) Mid-Mesozoic closure of Permo-Triassic Tethys and its implications. Nature 279:590–593
Şengör AMC (1979b) The north Anatolian transform fault: its age, offset and tectonic significance. J Geol Soc Lond 136:269–282
Şengör AMC, Yilmaz Y (1981) Tethyan evolution of Turkey: a plate tectonic approach. Tectonophysics 75(3–4):181–241
Şengör AMC (1982) Factors governing the neotectonic evolution of the Aegean. Young Tectonics and Volcanism of Western Anatolia Panel. 59–71. (in Turkish)
Seyitoğlu G, Scott B (1992) The age of the Büyük Menderes Graben (West Turkey) and its tectonic implications. Geol Mag 129:239–242
Shishegaran A, Taghavizade H, Bigdeli A, Shishegaran A (2019) Predicting the earthquake magnitude along Zagros fault using time series and ensemble model. J Soft Comput Civ Eng 3(4):67–77
Siami-Namini S, Tavakoli N & Namin AS (2019) The performance of LSTM and BiLSTM in forecasting time series. In 2019 IEEE International Conference on Big Data, 3285–3292
Sözbilir H (2001) Young-tectonics of Nazilli and its surroundings (Büyük Menderes Graben). Great Menderes Earthquakes Geophysical Meeting. 54–61 (in Turkish)
Stein RS, Barka AA, Dieterich JH (1997) Progressive failure on the North Anatolian fault since 1939 by earthquake stress triggering. Geophys J Int 128(3):594–604
Tağil Ş (2004) Neotectonic characteristics and seismicity of Balikesir plain and its neighborhood. J Geog Sci 2(1):73–92 ((in Turkish))
Tağil Ş, Alevkayali C (2013) Spatial Distribution of Earthquake in the Aegean Region: A Geostatistical Approach, Social. Int J Stud 6(28):69–379 (in Turkish)
Tan O (2021) Turkish Homogenized Earthquake Catalogue (TURHEC). Natural Hazards and Earth System Sciences (NHESS). Zenodo. https://doi.org/10.5281/zenodo.5056801
Todelo TL & Chris Jordan GA (2019) Predictability of earthquake occurrence using auto regressive integrated moving average (ARIMA) model. In: Lecture notes in engineering and computer science: proceedings of the ınternational multiconference of engineers and computer scientists, pp 13–15
Uhrhammer RA (1986) Characteristics of northern and central California seismicity. Earthquake Notes 57(1):21
Wang Q, Guo Y, Yu L, Li P (2017) Earthquake prediction based on spatio-temporal data mining: an LSTM network approach. IEEE Trans Emerg Top Comput 8(1):148–158
Yadav A, Gahalaut K, Mallika K, Purnachandra Rao N (2015) Annual periodicity in the seismicity and water levels of the Koyna and Warna reservoirs, western India: a singular spectrum analysis. Bull Seismol Soc Am 105(1):464–472
Yadav A, Jha CK, Sharan A (2020) Optimizing LSTM for time series prediction in Indian stock market. Procedia Comput Sci 167:2091–2100
Yamashita R, Nishio M, Do RKG, Togashi K (2018) Convolutional neural networks: an overview and application in radiology. Insights Imaging 9:611–629
Yang Q, Deng C, Chang X (2022) Ultra-short-term/short-term wind speed prediction based on improved singular spectrum analysis. Renew Energy 184:36–44
Yang J & Li J (2017) Application of deep convolution neural network. In: 2017 14th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP) (pp 229–232). IEEE
Yu R, Gao J, Yu M, Lu W, Xu T, Zhao M et al (2019) LSTM-EFG for wind power forecasting based on sequential correlation features. Fut Gener Comput Syst 93:33–42
Yuan X, Dan H, Qiuyin Y, Wenjun Z, Min R & Jing Y (2022) ARIMA model analysis of the regularities of earthquake origin times in the Longmen mountain fault zone. (Preprint in Research Square)
Zaremba W, Sutskever I & Vinyals O (2014) Recurrent neural network regularization. arXiv preprint arXiv:1409.2329.
Zha W, Liu Y, Wan Y, Luo R, Li D, Yang S, Xu Y (2022) Forecasting monthly gas field production based on the CNN-LSTM model. Energy 260:124889
Zhai D, Zhang X, Xiong P (2020) Detecting thermal anomalies of earthquake process within outgoing longwave radiation using time series forecasting models. Ann Geophys 63(5):PA48–PA548
Zhang H, Melgar D, Sahakian V, Searcy J, Lin JT (2022) Learning source, path and site effects: CNN-based on-site intensity prediction for earthquake early warning. Geophys J Int 231(3):2186–2204
Zhao B, Lu H, Chen S, Liu J, Wu D (2017) Convolutional neural networks for time series classification. J Syst Eng Electron 28(1):162–169
Acknowledgements
We thank the support of The Scientific and Technological Research Council of Turkey (TUBITAK) ARDEB 1001 [Project number: 121F208] program. The authors also thank Assistant Prof. Dr. Tuba Eroğlu Azak for preparing dataset, Prof. Dr. Tolga Çan for his valuable advice and reviewers for their constructive comments.
Author information
Authors and Affiliations
Contributions
H. O. Cekim and H. N. Karakavak composed methods and results, G. Ozel wrote the main manuscript text, and S. Tekin prepared the data and Figures 1 and 2. All authors reviewed the manuscript.
Corresponding author
Ethics declarations
Conflict of interest
The authors report that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Öncel Çekim, H., Karakavak, H.N., Özel, G. et al. Earthquake magnitude prediction in Turkey: a comparative study of deep learning methods, ARIMA and singular spectrum analysis. Environ Earth Sci 82, 387 (2023). https://doi.org/10.1007/s12665-023-11072-1
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s12665-023-11072-1