Calibrating fluvial erosion laws and quantifying river response to faulting in Sardinia, Italy

Abstract

It is now widely accepted that rivers modify their erosion rates in response to variable rock uplift rates, resulting in changes in channel slope that propagate upstream through time. Therefore, present-day river morphology may contain a record of tectonic history. The simple stream power incision model can, in principle, be used to quantify past uplift rates over a variety of spatial and temporal scales. Nonetheless, the erosional model's exponents of area and slope (m and n respectively) and ‘bedrock erodibility’ (k) remain poorly constrained. In this paper, we will use a geologically and geomorphically well constrained Plio-Pleistocene volcanic landscape in central Sardinia, Italy, to calibrate the stream power erosion equation and to investigate the slip rate of faults that have been seismically quiescent in the historic past. By analysing digital elevation models, geological maps and Landsat imagery, we have identified the geomorphic expression of several volcanic features (eruption centres and basaltic lava flows) and three normal faults with 6 to 8 km fault traces within the outcrop. Downstream, river longitudinal profiles show a similar transient response to relative base level fall, probably as a result of relief inversion at the edge of the volcanic outcrop. From measurements of incision, local slope and upstream catchment area across eight different rivers, we calculate n ≈ 1, m = 0.50 ± 0.02 and, using a landscape age from literature of 2.7 Ma, bedrock erodibility k = 0.10 ± 0.04 m(1−2m) Myr−1. There are also knickpoints on rivers upstream of two normal faults, and we used numerical inverse modelling of the longitudinal profiles to predict the slip rate of these faults since 2.7 Ma. The results from the inverse model show that the erosional parameter values derived in this study can produce theoretical longitudinal profiles that closely resemble observed river profiles upstream of the faults. The lowest misfit theoretical longitudinal profiles were generated by a model of temporally discontinuous footwall uplift with consistently low throw rates (<0.1 mm yr−1). The predicted footwall uplift history is similar for the two faults, both showing periods of fault slip and no fault movement since 2.7 Ma.