Logistic regression models including the model parameters of the northern and the western sides.

Classification accuracy was shown with the numbers of pixels that were observed and predicted in the two classes of tree and non-tree. The variables that have considerable effects on the models are marked in bold. Significance of all variables: P < 0.001. PRR represents potential relative radiation which sums daily values over the growing season. SPI represents snow potential index which indicates the snow accumulation in topography. LST represents land surface temperature. Logistic regression models including the model parameters of the northern and the western sides.