Detection of transience in eroding landscapes

Past variations in climate and tectonics have led to spatially and temporally varying erosion rates across manylandscapes. In this contribution I examine methods for detecting and quantifying the nature and timing of transience in erodinglandscapes. At a single location, cosmogenic nuclides can detect the instantaneous removal of material or acceleration of erosionrates over millennial timescales using paired nuclides. Detection is possible only if one of the nuclides has a significantly shorterhalf-life than the other. Currently, the only practical way of doing this is to use cosmogenic in situ carbon-14 (14C) alongside a longerlived nuclide, such as beryllium-10 (10Be). Hillslope information can complement or be used in lieu of cosmogenic information: insoil mantled landscapes, increased erosion rates can be detected for millennia after the increase by comparing relief and ridgetopcurvature. This technique will work as long as the final erosion rate is greater than twice the initial rate. On a landscape scale,transience may be detected based upon disequilibria in channel profiles or ridgetops, but transience can be sensitive to the natureof transient forcing. Where forcing is periodic, landscapes display differing behavior if forcing is driven by changes in base levellowering rates versus changes in the efficiency of either channel or hillslope erosion (e.g. driven by climate change). Oscillationsin base level lowering lead to basin averaged erosion rates that reflect a long term average erosion rate despite strong spatialheterogeneity in local erosion rates. This averaging is reflected in10Be concentrations in stream sediments. Changes in hillslopesediment transport coefficients can lead to large fluctuations in basin averaged erosion rates, which again are reflected in10Beconcentrations. The variability of erosion rates in landscapes where both the sediment transport and channel erodibility coefficientsvary is dominated by changes to the hillslope transport coefficient.