Comment on ‘The inference of mantle viscosity from an inversion of the Fennoscandian relaxation spectrum’ by J. X. Mitrovica and W. R. Peltier

Abstract

Mitrovica & Peltier (1993; hereafter MP) have recently analysed the Fennoscandian uplift data using inversion techniques. They concluded that, although observations support the presence of a weak asthenosphere under Fennoscandia, the decay spectrum for the best-fit model (1.3 x lo1’ Pa s asthenosphere overlying a 10zlPas mantle) found by previous workers (Fjeldskaar & Cathles 1991a,b Cathles 1975, 1980, hereafter FC) lies outside the decay spectrum observed in Fennoscandia and is thus excluded by the observed data. We believe this conclusion results from the authors inadvertantly using a different asthenosphere thickness in their calculations than the one that they report. The purpose of this communication is to argue this case and to show that when proper asthenosphere thicknesses are used, the models of MP and FC are both compatible with the observed uplift data and that each in fact defines one edge of the observed uplift spectrum. The essential difference in the two end-member models is the elected tradeoff between lithosphere thickness and asthenosphere viscosity. MP choose a stiff lithosphere and a minor asthenosphere, whereas FC choose the opposite. We believe lithosphere rigidity data (not discussed in MP) argue for a weak lithosphere in Fennoscandia and suggest that the FC model is to be preferred. The relaxation spectrum for the best-fit model of FC, as presented in M P s Fig. 11, is shown by the dashed curve labelled LVZ in Fig. 1. It clearly lies below and outside the observed Fennoscandian relaxation data (shaded area). The decay spectrum for the LVZ model is, according to MP, based on loading of a 70 km lithosphere on a 90 km asthenosphere of viscosity 2 x lO”Pa s overlying a 10” P a s mantle. The maximum relaxation time for the LVZ model exceeds 6000 years. However, the maximum relaxation time for FC’s bestfit asthenosphere model that the LVZ curve is supposed to