Braided stream channels, described and discussed by many writers, still pose problems of origin, geometry, and persistence through time. Recent advances in the classification of deltaic channel networks (of which braids are a subset) and computer simulation of braided networks provide guides to further field and simulation research. We selected small but complete braided units that could be mapped sufficiently quickly to record the minor islands and channels, as well as to measure relative discharge in the channels. Some islands are less than 0.5 m in length, and minor channels may be as narrow as 20 cm, with flowing water less than 1 cm in depth. Topological analysis showed good agreement with expected proportions of the four possible kinds of links generated by combinations of bifurcations (forks) and junctions.
A simulation model developed here, based on the number of channels in equally spaced cross sections through the braid, yields output in fair topological agreement with short braided segments of measured real-world input. Enclosure-length distributions in nature and simulation output differ markedly, however, especially in the number of very short enclosures produced by simulation.
Our general conclusion is that neither aerial photographs nor published maps are wholly satisfactory for detailed braided stream research. Rapid field mapping is essential for defining the initial stages of braid generation, inasmuch as the smallest components, especially those less than 1 m long, appear to change shortly after new bifurcations occur.