Scale dependency in quantifications of the avulsion frequency of coastal rivers

Quantification of the frequency with which coastal-plain rivers avulse is important for elucidating autogenic dynamics and their role as controls on landscape change and stratigraphic architectures. An outstanding question exists, however, as to whether measures of avulsion frequency are inherently affected by the spatial and temporal scales at which they are evaluated, which has implications regarding our ability to make direct comparisons between different river systems or deltas. To address this problem, a quantitative analysis of the avulsion histories of 57 coastal-plain river systems is undertaken. Nine alternative measures of avulsion frequency are extracted. These are based on numbers of (i) avulsion events, (ii) active or abandoned channel courses, and (iii) delta lobes, all considered per unit time. Additional sets of avulsion-frequency proxies are established based on normalization of these numbers relative to the size of the area being studied, and to the number of distinct river systems that drain into that area. The sensitivity of these quantities to the spatial and temporal extent of study areas and time intervals, and their relationships with quantities describing the scale of the river systems, are assessed. All avulsion-frequency estimates demonstrate apparent negative relationships with the timespan over which they are determined; this may reflect global Holocene trends, or variations in resolution with the time window. Avulsion metrics that are not normalized by the planform extent of the study area do not show proportionality with the size of the study areas themselves, nor with the scale of the river systems; correspondingly, the spatio-temporal density of avulsion events tends to be higher for smaller rivers and associated study areas. This may be due to systematic variations in data resolution, to the influence of external controls that relate to the scale of deltas, or to inherent non-stationarity in the avulsion dynamics of lowland rivers, in association with forms of self-organization that do not vary with scale. Although non-normalized avulsion-frequency estimates do not scale with measures of river-system size, they are seen to correlate with progradation rates, which are themselves scaled to sediment discharge and catchment size. Practical considerations can be drawn on how avulsion frequency may be appropriately quantified to enable meaningful comparisons of the autogenic dynamics of coastal-plain rivers