Future forecast of global mean surface temperature using machine learning and conventional time series methods
[ X ]
Tarih
2025
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Springer Wien
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
One of the most important indicators of climate change and, consequently, global warming is the rise in the average surface temperature of the entire world. As observed in this light, forecasting the global mean surface temperature is an essential issue that must be addressed to develop adaptation measures for climate change. Many models have been developed to forecast the temperature of the air; however, these models often concentrate on local areas or use a restricted amount of station data. In this study, seasonal autoregressive integrated moving average (SARIMA), long-short-term memory (LSTM), and gated recurrent unit (GRU) models are used to predict global mean surface temperature (GMST) data. The data sets used in the study are GISTEMP and HardCRUT datasets and consist of land surface air temperature and water surface temperature. An evaluation of the performance of the models is carried out using various error measures to guarantee a high level of prediction accuracy. All models' results indicate that the yearly GMST value increase relative to 1961-1990 will be between 0.94 oC and 1.45 oC in 2050. In addition, the yearly GMST value, measured as approximately 14.8-15.00 oC in 2022, will be between 15.15 oC and 15.43 oC in 2050, according to the obtained models.
Açıklama
Anahtar Kelimeler
Kaynak
Theoretical and Applied Climatology
WoS Q Değeri
Scopus Q Değeri
Cilt
156
Sayı
1
