“…Ice-sheet nucleation may, in addition, depend on chaotic aspects of the weather/climate system; for example, successive winters with heavy snowfall may e almost randomly e cause some locations to receive an initial snow cover with enough volume and albedo feedback to ensure its survival and subsequent growth potential (e.g., Oglesby, 1990). Finally, modelling studies (e.g., Abe-Ouchi et al, 2013) indicate that glacial culminations like the PGM and LGM reflect the outcome of temporal developments in forcings and feedbacks through the preceding glacial cycle that include insolation (e.g., Laskar et al, 2004;Colleoni et al, 2011), CO 2 and CH 4 concentrations (Monnin et al, 2001;Loulergue et al, 2008; Waelbroeck et al, 2002;Rohling et al, 2009Rohling et al, , 2014Elderfield et al, 2012;Grant et al, 2014), and also in state variables such as surface and deep-sea temperature (e.g., Stenni et al, 2010;Elderfield et al, 2012;Parrenin et al, 2013;Rohling et al, 2012Rohling et al, , 2014Martínez-Botí et al, 2015;Snyder, 2016a,b). Climate simulations by Colleoni et al (2014) suggest that orbital and greenhouse-gas changes for the penultimate glacial cycle were more favourable for glacial inception over Eurasia than over North America, relative to the last glacial cycle.…”