Assessment of long-term coordinate time series using hydrostatic-season-time model for rock-fill embankment dam

The safety control of dams is based on monitoring activities and modelling of registered observations. The statistical hydrostatic-season-time model was originally developed and proposed for analyzing of monitoring data on concrete dams. In some later works, the model was implemented for earth-fill embankment dams. The model admits a simultaneous estimation of hydrostatic load, temperature influences and irreversible deformations. In our study, we analyze long-term coordinate time series of a geodetic point on the crest of a rock-fill embankment dam. Coordinate time series are result of an adjustment of the observations in a permanent geodetic network for different epochs. An optimal model is defined using a multiple linear regression by combined process of exclusion and inclusion of individual parameters. In the process, different statistical parameters are observed. The analyses confirmed that not all parameters are significant. The most interesting and important conclusion of computations can be stated as follows: after inclusion of significant coefficients of the hydrostatic load and long-term trend, the residual time series still expose underlying periodicities. They can be removed by inclusion of at least one parameter of the seasonal term, where the temperature influences (i.e. air, water and soil) are modelled. The influences of the temperature on the dam are not significant in any direction, but the inclusion of the parameters improves the statistical performance of the used model.