Long‐term landscape evolution: linking tectonics and surface processes

Abstract Research in landscape evolution over millions to tens of millions of years slowed considerably in the mid‐20th century, when Davisian and other approaches to geomorphology were replaced by functional, morphometric and ultimately process‐based approaches. Hack's scheme of dynamic equilibrium in landscape evolution was perhaps the major theoretical contribution to long‐term landscape evolution between the 1950s and about 1990, but it essentially ‘looked back’ to Davis for its springboard to a viewpoint contrary to that of Davis, as did less widely known schemes, such as Crickmay's hypothesis of unequal activity. Since about 1990, the field of long‐term landscape evolution has blossomed again, stimulated by the plate tectonics revolution and its re‐forging of the link between tectonics and topography, and by the development of numerical models that explore the links between tectonic processes and surface processes. This numerical modelling of landscape evolution has been built around formulation of bedrock river processes and slope processes, and has mostly focused on high‐elevation passive continental margins and convergent zones; these models now routinely include flexural and denudational isostasy. Major breakthroughs in analytical and geochronological techniques have been of profound relevance to all of the above. Low‐temperature thermochronology, and in particular apatite fission track analysis and (U–Th)/He analysis in apatite, have enabled rates of rock uplift and denudational exhumation from relatively shallow crustal depths (up to about 4 km) to be determined directly from, in effect, rock hand specimens. In a few situations, (U–Th)/He analysis has been used to determine the antiquity of major, long‐wavelength topography. Cosmogenic isotope analysis has enabled the determination of the ‘ages’ of bedrock and sedimentary surfaces, and/or the rates of denudation of these surfaces. These latter advances represent in some ways a ‘holy grail’ in geomorphology in that they enable determination of ‘dates and rates’ of geomorphological processes directly from rock surfaces. The increasing availability of analytical techniques such as cosmogenic isotope analysis should mean that much larger data sets become possible and lead to more sophisticated analyses, such as probability density functions (PDFs) of cosmogenic ages and even of cosmogenic isotope concentrations (CICs). PDFs of isotope concentrations must be a function of catchment area geomorphology (including tectonics) and it is at least theoretically possible to infer aspects of source area geomorphology and geomorphological processes from PDFs of CICs in sediments (‘detrital CICs’). Thus it may be possible to use PDFs of detrital CICs in basin sediments as a tool to infer aspects of the sediments' source area geomorphology and tectonics, complementing the standard sedimentological textural and compositional approaches to such issues. One of the most stimulating of recent conceptual advances has followed the considerations of the relationships between tectonics, climate and surface processes and especially the recognition of the importance of denudational isostasy in driving rock uplift (i.e. in driving tectonics and crustal processes). Attention has been focused very directly on surface processes and on the ways in which they may ‘drive’ rock uplift and thus even influence sub‐surface crustal conditions, such as pressure and temperature. Consequently, the broader geoscience communities are looking to geomorphologists to provide more detailed information on rates and processes of bedrock channel incision, as well as on catchment responses to such bedrock channel processes. More sophisticated numerical models of processes in bedrock channels and on their flanking hillslopes are required. In current numerical models of long‐term evolution of hillslopes and interfluves, for example, the simple dependency on slope of both the fluvial and hillslope components of these models means that a Davisian‐type of landscape evolution characterized by slope lowering is inevitably ‘confirmed’ by the models. In numerical modelling, the next advances will require better parameterized algorithms for hillslope processes, and more sophisticated formulations of bedrock channel incision processes, incorporating, for example, the effects of sediment shielding of the bed. Such increasing sophistication must be matched by careful assessment and testing of model outputs using pre‐established criteria and tests. Confirmation by these more sophisticated Davisian‐type numerical models of slope lowering under conditions of tectonic stability (no active rock uplift), and of constant slope angle and steady‐state landscape under conditions of ongoing rock uplift, will indicate that the Davis and Hack models are not mutually exclusive. A Hack‐type model (or a variant of it, incorporating slope adjustment to rock strength rather than to regolith strength) will apply to active settings where there is sufficient stream power and/or sediment flux for channels to incise at the rate of rock uplift. Post‐orogenic settings of decreased (or zero) active rock uplift would be characterized by a Davisian scheme of declining slope angles and non‐steady‐state (or transient) landscapes. Such post‐orogenic landscapes deserve much more attention than they have received of late, not least because the intriguing questions they pose about the preservation of ancient landscapes were hinted at in passing in the 1960s and have recently re‐surfaced. As we begin to ask again some of the grand questions that lay at the heart of geomorphology in its earliest days, large‐scale geomorphology is on the threshold of another ‘golden’ era to match that of the first half of the 20th century, when cyclical approaches underpinned virtually all geomorphological work. Copyright © 2007 John Wiley & Sons, Ltd.