Architectural-element analysis was developed to describe and reconstruct fluvial system stacking patterns and understand the controls on fluvial architecture. The architecture of braided fluvial systems at the scale of channel belts is still poorly understood, and upscaling from outcrop observations to basin-scale architecture remains partly speculative. Nonetheless, predictive sedimentary models are necessary to understand the connectivity of sand bodies from small scale (outcrop) to basin scale. This contribution presents a dataset from the Late Triassic sandy braided fluvial system of the Fundy Basin (Nova Scotia, Canada). We describe facies associations, architectural elements, and stacking patterns at different scales, and discuss the controls on fluvial architecture and upscaling issues. The studied section shows, at the base, multilateral stacked channel bodies with little preservation of fine-grained floodplain facies. It changes upwards into multistory stacked channel bodies that form channel-belt complexes intercalated with thicker floodplain deposits. Based on grain-size evolution, large-scale sequences encompass several channel-belt complexes.
Controls on channel and channel-belt migration or avulsion are considered to be related to autocyclic processes. In contrast, large-scale vertical evolution and grain-size variations record allocyclic signals. In particular, gradual grain-size variations suggest a climatic control which influences the bedload transport capacity of the depositing fluvial system. Architectural evolution including an upward increase in the preservation of floodplain units is considered to be due to an increase in generation of accommodation space. We question the existing definition of the order of large-scale (sixth- to seventh- order) bounding surfaces recognized previously and highlight difficulties associated with reconstructing development of a regional fluvial system using these bounding surfaces.