The features of coefficients of determination and coefficients of leave-one-out method for spatial vegetation model and spatial models of squared deviations are discussed. The properties of models are illustrated in key area for spatial model of Cladonia stellaris projective cover.Spatial modeling of the vegetation and its different features can be realised as spatial extrapolation of classification units of vegetation [1] or spatial extrapolation of unclassified variables [2]. The last way makes opportunity to create a spatial model of unclassified information [3]. In both cases, spatially distributed variables e.g. the spectral responses of the surface are used as predictors of extrapolation. The quality of modeling is estimated by the coefficient of determination (R 2 , r-squared), which shows the ratio of variation (the sum of the squared deviations) explained by the adopted model [4]. This estimation of the statistical dependence of modified variables and predictors can be increased unlimitedly by increasing the flexibility of the regression functions and by including more predictors. The validity of such a model improvement is evaluated by validation methods, such as the coefficient of determination based on the leave-one-out method, the successive exclusion of each sample from the training data and the use of its deviation from the predicted value [5]. Depending on the completeness of the data and the features of the regression model, its deviations from the observed values in different areas of the model can vary significantly. Spatial modeling of these deviations can provide a detailed assessment of the quality of the model and the quality of area survey. To investigate deviations of spatial models we used the key area with geobotanical relevés. Spatial models of the squared deviations of the vegetation model and its individual variables from the observed data, and the squared deviations of the leave-one-out method are considered.25 geobotanical relevés of the key area (N 63.379, E 75.865 / N 63.105, E 76.451), Yamal-Nenets Autonomous Region were used as observed data. 88 rare species were excluded. The key area is within the subzone of the northern taiga [6]. Here the good drained loamy habitats are occupied by zonal larch shrub-moss forests, drained sands are occupied by intrazonal pine shrub-lichen forests. Flat watersheds are occupied by frozen and thawed mires. Oligomesotrophic bogs occupy hydrological net of swamps. Floodplain series of vegetation are common near rivers and streams. Forests and frozen mires on the key area have been partly burned by recent fires. The spatial modeling of species abundance was based on the variables of the spectral responses of the composite of 6
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.