Quantitative Analysis of River Profiles and Fluvial Landscapes
The ability to invert large numbers of longitudinal river profiles to obtain testable cumulative uplift histories implies that channels respond predictably to regional tectonic forcing. However, these models make simplifying assumptions about drainage networks that require careful justification. For example, precipitation rate is assumed to be constant through space and time. Since catchments are assumed to be invariant, the effects of stream piracy, divide migration and interfluvial erosion are also ignored. Here, I investigate the general validity of these assumptions. First, I jointly invert >3200 Asian river profiles for uplift. When erosional parameters are calibrated, the recovered uplift history supports models of stepwise growth of the Tibetan Plateau during Cenozoic times. The predicted history of cumulative uplift and sediment flux are tested against independent observations. The ability of river profile inverse modelling to address geodynamical problems is assessed. Secondly, I invert > 1500 African river profiles to determine a continent-wide uplift history, calibrated using geological observations. This uplift forcing is then used to drive landscape simulations that explicitly allow catchment dynamicism. These simulations reproduce the large-scale features of the African landscape with average misfits of O(10)s of metres, yielding sediment flux histories with misfits of <20%. The effect of varying precipitation rate is investigated using increasingly severe tests. Inverse modelling of synthetic fluvial profiles from the resultant landscapes demonstrates the original forcing can be reliably recovered. This result implies that tectonic forcing plays a primary role in sculpting landscapes. Precipitation rate only influences large-scale landscape evolution when it is varied at periods & 10 Ma. Thirdly, I build the naturalistic landscape simulation into an inverse scheme. Synthetic landscapes generated by known forcing are inverted as a function of regional uplift. Temporally varying forcing can be recovered with an uncertainty of < 1%, even subject to noisy starting topographic seeds. Complex uplift histories are less reliably resolved, but episodes of uplift spaced > 10 Ma apart are generally recoverable. Temporal damping reduces local minima in the misfit function. In principle, landscapes can also inverted for uplift as a function of space and time. Spatial damping permits the recovery of smooth input uplift forcing. I compile a global inventory of >18,000 river profiles. Scaling analyses reveal significant commonalities over a range of scales. Spectral analysis of river profiles demonstrates different scaling regimes exist. African river profiles contain self-similar scaling at long wavelengths, with an increase in power spectral slope at wavelengths >100 km. Synthetic river profiles generated by tectonic forcing obtained from inverse modelling of observed rivers do not contain changes in spectral slope. Neither addition of topographic nor precipitational seeding produces a change in scaling regime. This result suggests that uplift forcing recovered from inverse modelling is self-similar and that at short wavelengths discontinuities caused by flow instabilites or lithological variation may generate different scaling. Universal scaling appears to be a global characteristic, which justifies river profile inverse modelling as a method for obtaining meaningful uplift histories.