The panel data model accounts for the relationships and the heterogeneities among points, therefore, it was introduced in this study to construct a zoned safety monitoring model for super high arch dams. Firstly, the principal component analysis(PCA)was used to select the typical variables of measured temperatures. Secondly, the hierarchical clustering method based on PCA was applied to propose a data-based zoning method for deformation monitoring points. Then the panel data of each zone was established, and its type was determined by statistical tests, thus the corresponding regression model zone was constructed. Finally, through the comparative analyses of the spatial distribution of zones and the fitted parameters(e. g. , intercept terms and variable coefficients of various zones), both the deformation characteristics of zones and the spatial deformation characteristics of the dam can be identified. A super high arch dam during its initial operation period was investigated, and the focused plumb line points were divided into four zones according to the observation data. After the statistical tests, fixed effect regression models of zones were established. The results show that the spatial deformation is in accordance with the actual condition, and the factors such as the unstable temperature field and the nonlinear time varying effect during the initial operation period have obvious influences on the dam deformation.