Weathering instability and landscape evolution

The argument in this paper is that the fundamental control on landscape evolution in erosional landscapes is weathering. The possibility of and evidence for instability in weathering at four scales is examined. The four scales are concerned with weathering processes, allocation of weathered products, the interrelations of weathering and denudation, and the topographic and isostatic responses to weathering-limited denudation (the regolith, hillslope, landscape unit, and landscape scales, respectively). The stability conditions for each model, and the circumstances under which the models themselves are relevant, are used to identify scale-related domains of stability and instability. At the regolith scale, the interactions among weathering rates, resistance, and moisture are unstable, but there are circumstances-over long timescales and where weathering is well advanced-under which the instability is irrelevant. At the hillslope scale, the system is stable when denudation is transport rather than weathering limited and where no renewal of exposure via regolith stripping occurs. At the level of landscape units, the stability model is based entirely on the mutual reinforcements of weathering and erosion. While this should generally lead to instability, the model would be stable where other, external controls of both weathering and erosion rates are stronger than the weathering-erosion feedbacks. At the broadest landscape scale, the inclusion of isostatic responses destabilizes erosion-topography-uplift relationships. Thus, if the spatial or temporal scale is such that isostatic responses are not relevant, the system may be stable. Essentially, instability is prevalent at local spatial scales at all but the longest timescales. Stability at intermediate spatial scales is contingent on whether weathering-erosion feedbacks are strong or weak, with stability being more likely at shorter and less likely at longer timescales. At the broadest spatial scales, instability is likely; although stability may be present at intermediate temporal scales if weathering-erosion feedbacks are weak. The distinction is important because stability is associated with convergent evolution whereby the effects of initial variations or disturbances are reduced over time as the landscape converges toward a stable equilibrium state. Instability, by contrast, indicates divergent evolution, increasing differentiation over time, and the persistence and growth of disturbance effects and initial variations.