This discussion aims to highlight the underlying cause of several aspects of the greenfield settlement data presented by Boonsiri and Takemura (Geotech Geol Eng 33(3):621–640, 2015). The discussion considers, for the geotechnical centrifuge tests that were reported, the effects of the boundary conditions imposed at the model tunnel on resulting settlements. Data obtained using the rigid boundary model tunnel in Boonsiri and Takemura (Geotech Geol Eng 33(3):621–640, 2015) are compared against other available data from tests using a fluid-filled flexible membrane model tunnel. It is demonstrated that the boundary conditions used to simulate tunnel ground loss have an important impact on the settlement mechanism; compared to a fluid-filled flexible membrane, a rigid boundary model tunnel results in wider settlement troughs, which do not vary in shape considerably with changes in relative tunnel depth, and can result in higher ratios between the area of the settlement troughs and the tunnel ground loss. The appropriateness of the different tunnel boundary conditions is also discussed.