Recent observations show an accelerated ice loss of the Greenland ice sheet during the past few decades which makes it a major contributor to sea-level rise. The SeaRISE experiments have shown the sensitivity of current ice sheet models to external forcing to gain insight into the potential future contribution to sea level from the Greenland ice sheet. Although the model results show a similar trend of mass loss, the model responses are not always homogeneous. This may an effect of different employed e.g. numerics, approximation to the Stokes flow, spatial resolution, and initialization methods. Here, we use the thermomechanical coupled Ice Sheet System model (ISSM) to create an initial condition for projections of the ice sheet. For the initialization, we use a hybrid procedure that combines assimilation and a temperature spin-up over longer periods. The temperature spin-up is performed with two different climatic forcings: (1) present-day climatic and (2) palaeoclimatic conditions. As we employ the higer-order approximation to the Stokes flow grid refinements are made during the whole initialization procedure. The presented procedure is part of the initMIP-Greenland intercomparison exercise (Model: AWI-ISSM). In order to examine the difference of mass loss for both different initializations, we re-run the SeaRISE experiments.