Current Status and Future Forecast of Global CO2 Concentration Using Statistical and Deep Learning Time Series Methods

Öz

In this study, four models were developed and assessed, including Autoregressive Integrated Moving Average (ARIMA), Feedforward Neural Network (FNN), Gated Recurrent Unit (GRU), and Long Short-Term Memory (LSTM) to evaluate current status and future forecast of global CO2 concentration. A total of 554 global monthly datasets were employed to train and test the developed models, aimed at estimating future CO₂ concentrations. Then, each developed model was also utilized to estimate CO₂ concentrations for the future 110 months, from March 2025 to February 2035. Among all generated techniques, the LSTM model showed the highest estimation accuracy with an MAPE of 0.05%, an MAE of 0.2028 ppm, and an RMSE of 0.3216 ppm. Whereas GRU and FNN techniques also obtained good results with the same MAPE of 0.05%, their MAE and RMSE values were slightly higher. The four developed models (ARIMA, FNN, GRU, and LSTM) agree on a continuous rise in atmospheric CO2 level within the range between March 2025 and early 2035, and they typically show CO2 concentrations starting from approximately 425 ppm in early 2025 to 442-443 ppm by the end of 2034.

Anahtar Kelimeler

• CO2 Concentration
• Deep Learning
• Global Warming
• Time Series Methods

Etik Beyan

The authors declare that they have no competing interests.

İstatistiksel ve Derin Öğrenme Zaman Serisi Yöntemleri Kullanılarak Küresel CO2 Yoğunluğunun Mevcut Durumu ve Gelecekteki Tahmini

Öz

Bu çalışmada global CO2 yoğunluğunun güncel durumunu ve gelecek tahminini değerlendirebilmek amacıyla, Otoregresif Entegre Hareketli Ortalama (ARIMA), İleri Beslemeli Sinir Ağı (FNN), Kapılı Yinelemeli Birim (GRU) ve Uzun Kısa Süreli Bellek (LSTM) olmak üzere dört model geliştirilmiştir. Gelecekteki CO₂ yoğunluklarını tahmin etmeyi amaçlayan bu modelleri eğitmek ve test etmek için toplam 554 aylık veri kullanılmıştır. Daha sonra geliştirilen her model, Mart 2025'ten Şubat 2035'e kadar olan 110 aylık süre için CO₂ yoğunluklarını tahmin etmek için de kullanılmıştır. Oluşturulan tüm teknikler arasında LSTM modeli, %0,05'lik bir MAPE, 0,2028 ppm'lik bir MAE ve 0,3216 ppm'lik bir RMSE ile en iyi tahmin doğruluğunu göstermiştir. GRU ve FNN teknikleri de aynı %0,05'lik MAPE ile iyi sonuçlar elde ederken, MAE ve RMSE değerleri biraz daha yüksek çıkmıştır. Geliştirilen dört model (ARIMA, FNN, GRU ve LSTM), Mart 2025 ile 2035 başı aralığında atmosferik CO2 seviyesinde sürekli bir artış ve tipik olarak CO2 yoğunluğunun 2025 başı yaklaşık 425 ppm'den başlayarak 2034 sonu itibarıyla 442-443 ppm arasına çıkması konusunda hemfikirdir.

Anahtar Kelimeler

• Karbondioksit Emisyonu
• Derin Öğrenme
• Küresel Isınma
• Zaman Serisi Yöntemleri

Etik Beyan

Yazarlar herhangi bir çıkar çatışması olmadığını beyan etmektedirler.

Kaynakça

1. 1. Liu, Z., Deng, Z., Davis, S. & Ciais, P. (2023). Monitoring Global Carbon Emissions in 2022. Nature Reviews Earth & Environment, 4(4), 205-206.
2. 2. Liu, B., Chang, H., Li, Y. & Zhao, Y. (2023). Carbon emissions predicting and decoupling analysis based on the PSO-Elm combined prediction model: Evidence from chongqing municipality, China. Environmental Science and Pollution Research, 30(32), 78849-78864.
3. 3. Wang, B. & Liu, J. (2024). Impact of climate change on green technology innovation-an examination based on microfirm data. Sustainability, 16(24), 11206.
4. 4. Arora, N.K. (2019). Impact of climate change on agriculture production and its Sustainable Solutions. Environmental Sustainability, 2(2), 95-96.
5. 5. Fan, Z., Yan, Z. & Wen, S. (2023). Deep Learning and Artificial Intelligence in sustainability: A review of sdgs, renewable energy, and environmental health. Sustainability, 15(18), 13493.
6. 6. Li, X. & Zhang, X. (2023). A comparative study of statistical and machine learning models on carbon dioxide emissions prediction of China. Environmental Science and Pollution Research, 30(55), 117485-117502.
7. 7. Yildirim, M., Bingol, H., Cengil, E., Aslan, S. & Baykara, M. (2023). Automatic classification of particles in the urine sediment test with the developed artificial intelligence-based hybrid model. Diagnostics, 13(7), 1299.
8. 8. Al Nuaimi, H.S., Acquaye, A. & Mayyas, A. (2025). Machine learning applications for carbon emission estimation. Resources, Conservation & Recycling Advances, 27, 200263.

9. 9. Ma, Y., He, P.J., Lü, F., Zhang, H., Yan, S., Cao, D., Mao, H. & Jiang, D. (2024). Machine learning-based prediction of the CO2 concentration in the flue gas and carbon emissions from a waste incineration plant. ACS ES&T Engineering, 4(3), 737-747.
10. 10. Begum, A.M. & Mobin, M.A. (2025). A machine learning approach to carbon emissions prediction of the top eleven emitters by 2030 and their prospects for meeting Paris agreement targets. Scientific Reports, 15(1), 1-40.
11. 11. Alhussan, A.A., Metwally, M. & Towfek, S.K. (2025). Predicting CO2 emissions with advanced deep learning models and a hybrid greylag goose optimization algorithm. Mathematics, 13(9), 1481.
12. 12. Ayaz, İ. (2024). Forecasting co₂ emissions with machine learning methods: Türkiye example and future trends. NATURENGS MTU Journal of Engineering and Natural Sciences Malatya Turgut Ozal University, 5(2), 82-87.
13. 13. Uluocak, İ. (2025). Comparative study of emission prediction using deep learning models. Çukurova Üniversitesi Mühendislik Fakültesi Dergisi, 40(2), 337-346.
14. 14. Aslan, E. (2024). Araçlarda CO2 emisyonlarının farklı yapay sinir ağı modelleri kullanılarak tahminlerinin karşılaştırılması. Çukurova Üniversitesi Mühendislik Fakültesi Dergisi, 39(2), 309-324.
15. 15. What Is ARIMA Modeling, (2023). Master's in Data Science, edX.org, [Online]. Available: https://www.mastersindatascience.org/learning/statistics-data-science/what-is-arima-modeling. Access date: 20.10.2025.
16. 16. Koçak, H. (2024). Time series prediction of temperature using Seasonal Arima and LSTM models. Gazi Journal of Engineering Sciences, 9(3), 574-584.
17. 17. Kaur, J., Parmar, K.S. & Singh, S. (2023). Autoregressive models in environmental forecasting time series: A theoretical and application review. Environmental Science and Pollution Research, 30(8), 19617-19641.
18. 18. Box, G.E.P., Jenkins, G.M. & Reinsel, G.C. (1994). Time series analysis: forecasting and control. 3rd edn. Prentice-Hall, Englewood Cliffs, NJ.
19. 19. Hornik, K., Stinchcombe, M. & White, H. (1989). Multilayer feedforward networks are universal approximators. Neural Networks, 2(5), 359-366.
20. 20. Engelbrecht, A.P. (2007). Computational intelligence: an introduction. Wiley, New York.
21. 21. Abdulkarim, S.A. & Engelbrecht, A.P. (2020). Time series forecasting with feedforward neural networks trained using particle swarm optimizers for dynamic environments. Neural Computing and Applications, 33(7), 2667-2683.
22. 22. Milano, P. & Elettrotecnica, D. (2004). PSO as an effective learning algorithm for neural network applications. Proceedings. ICCEA 2004. 3rd International Conference on Computational Electromagnetics and Its Applications, , 557-560.
23. 23. Tumse, S., Bilgili, M. & Sahin, B. (2022). Estimation of aerodynamic coefficients of a non-slender delta wing under ground effect using artificial intelligence techniques. Neural Computing and Applications, 34(13), 10823-10844.
24. 24. Arora, I., Gambhir, J. & Kaur, T. (2020). Data normalisation-based solar irradiance forecasting using artificial neural networks. Arabian Journal for Science and Engineering, 46(2), 1333-1343.
25. 25. Hochreiter, S. & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735-1780.
26. 26. Mathworks, (2020). Long short-term memory networks. https://www.mathworks.com/help/ deeplearning/ug/longshort-term-memory-networks.html. Access date: 20.10.2025.
27. 27. Mateus, B.C., Mendes, M., Farinha, J.T., Assis, R. & Cardoso, A.M. (2021a). Comparing LSTM and GRU models to predict the condition of a pulp paper press. Energies, 14(21), 6958.
28. 28. Lynn, H.M., Pan, S.B. & Kim, P. (2019). A deep bidirectional GRU network model for biometric electrocardiogram classification based on recurrent neural networks. IEEE Access, 7, 145395-145405.