Predicted Temperature of Deep Soil Layers Using Time Series Models

Soil temperature regime has a direct effect on plant growth، biodiversity، and soil biological activity. However، available soil temperature data are insufficient، while their reconstruction for agro-climatic and ecological studies is important. Soil temperature at different depths has different spatial and temporal behavior. In this study، we tried to predict monthly data using the time series of soil temperature at depths of 5، 20، and 30 cm in Hamadan Airport Station for the period 1996 to 2008. The results showed that soil temperature followed seasonal SARIMA model. The best time-series structure for 5، 20، and 30 cm depths were SARIMA (2،0،0) (2،0،0)، SARIMA (1،0،0) (1،0،0) and SARIMA (1،1،0) (1،1،1)، respectively. The calculated R2 and RMSE for the selected depths were 0. 96 and 1. 63 ºC for 5 cm depth، 0. 98 and 1. 49 ºC for 20 cm depth، and 0. 97 and 1. 56 ºC for 30 cm depth. Considering the suitability of the estimated results، the model was used for prediction of the soil temperature for 2009-13، which revealed negligible variation in soil temperature trend during this period.