ORIGINAL PAPER Soil temperature estimation using an artificial neural network and co-active neuro-fuzzy inference system in two different climates Hamid Zare Abyaneh 1 & Maryam Bayat Varkeshi 2 & Golmar Golmohammadi 3 & Kourosh Mohammadi 3 Received: 28 June 2015 /Accepted: 12 February 2016 # Saudi Society for Geosciences 2016 Abstract Soil temperature is an important meteorological pa- rameter which determines the rates of physical, chemical, and biological reactions in the soil. However, measured values are very sparse in space and time and often not available for a given site. In this study, two intelligent neural models includ- ing artificial neural networks (ANNs) and co-active neuro- fuzzy inference system (CANFIS) were used for the estima- tion of soil temperatures at six depths (5, 10, 20, 30, 50, and 100 cm) with minimum input data (mean air temperature). For this purpose, use was made of the 14-year meteorological data obtained for the two regions of Gorgan in northern Iran with a humid climate and Zabol in southeastern Iran with a dry cli- mate. Comparisons of the model performances in arid and humid regions showed that both ANNs and CANFIS models performed better in arid regions. The accuracy of the soil temperature predictions by both ANNs and CANFIS models gradually decreased from the surface down to the various depths. The results also indicated the capabilities of the ANNs in predicting soil temperature in arid and humid regions. Keywords ANN . CANFIS . Soil temperature . Air temperature . Climate Introduction Soil temperature is an important agricultural and environmen- tal factor which plays a critical role in different ecosystems, ranging from deserts to forests (Yin and Arp 1993). Soil tem- perature also plays an important role in a wide range of agri- cultural management practices and engineering designs. It al- so determines the rates of physical, chemical, and biological reactions in soil and has strong influences on plant growth and soil formation (Tenge et al. 1998; Brooks et al. 2004). For example, seed germination is most rapid when soil tempera- ture is optimal (Krishnan and Rao 1979). Gongalsky et al. (2008) studied the effects of soil temperature and moisture on the feeding activity of soil organisms and showed that the effect of soil temperature was greater than that of soil moisture. Therefore, the knowledge of soil temperature is a key factor for agronomists and engineers to make proper decisions re- garding planting dates, appropriate design of water-table man- agement systems, application of pesticides and fertilizers, and the control of chemical pollution in soil and groundwater (Yang et al. 1997). Despite the importance of soil temperature, measured values are very sparse in space and time and often not avail- able for a given site. When measured data are not available, the usual practice is to determine the soil temperature from meteorological parameters using theoretical or empirical models (Ozturk et al. 2011). Therefore, it is important to de- velop models that are capable of accurate and quick prediction of soil temperature fluctuations. In recent years, there have been several studies on estimat- ing soil temperatures using various analytical, numerical, and experimental methods, including Fourier techniques and arti- ficial neural network models (e.g., Carson 1963; Tenge et al. 1998; Droulia et al. 2009; Prangnell and McGowan 2009; * Hamid Zare Abyaneh zare@basu.ac.ir 1 Department of Irrigation and Drainage Engineering, Bu-Ali Sina University, Hamedan, Iran 2 Department of Water Engineering, Malayer University, Malayer, Iran 3 School of Engineering, University of Guelph, Guelph, Canada Arab J Geosci (2016) 9:377 DOI 10.1007/s12517-016-2388-8