Temporal mass variations in the Earth system, which can be detected from the Gravity Recovery and Climate Experiment (GRACE) mission data, cause temporal variations of geoid heights. The main objective of this contribution is to analyze temporal variations of geoid heights over the area of Poland using global geopotential models (GGMs) developed on the basis of GRACE mission data. Time series of geoid height variations were calculated for the chosen subareas of the aforementioned area using those GGMs. Thereafter, these variations were analyzed using two different methods. On the basis of the analysis results, models of temporal geoid height variations were developed and discussed. The possibility of prediction of geoid height variations using GRACE mission data over the area of Poland was also investigated. The main findings reveal that the geoid height over the area of Poland vary within 1.1 cm which should be considered when defining the geoid model of 1 cm accuracy for this area.

Temporal variations of geoid heights are vitally important in geodesy and Earth science. They are essentially needed for dynamic and kinematic updates of the static geoid model. These temporal variations, which substantially differ for different geographic locations, can successfully be determined using the Gravity Recovery And Climate Experiment (GRACE) mission data. So far, statistical decomposition methods, e.g. the Principal Component Analysis/Empirical Orthogonal Function (PCA/EOF) method, have not been implemented for the analysis and modelling of temporal mass variations within the Earth’s system over the area of Poland. The aim of this contribution is to analyse and model temporal variations of geoid heights obtained from GRACE mission data over the area of Poland using the PCA/EOF method. Temporal variations of geoid heights were obtained from the latest release, i.e. release five, of monthly GRACE-based Global Geopotential Models. They can reach the level of 10 mm. The PCA modes and their corresponding EOF loading patterns were estimated using two different algorithms. The results obtained revealed that significant part of the signal of temporal variations of geoid heights over Poland can be obtained from the first three PCA modes and EOF loading patterns. They demonstrate the suitability of the PCA/EOF method for analysing and modelling temporal variations of geoid heights over the area investigated.