Tests of the random network model, and its application to basin hydrology

TLDR

The random network model assumes all topologically distinct channel networks are equally likely, implying geomorphic processes cannot explain network structure; thus testing this model is crucial for future research. The study aims to apply network topology, via random walk theory and simulated random networks, to basin hydrology. Tests are based on basin size frequencies, maximum network widths, and random walk simulations of network topology. Network topology improves predictions of basin hydrographs, peak discharge (with width varying by a factor of two affecting discharge proportionally), and lag‑to‑peak.

Abstract

Abstract If the random model, in which all topologically distinct channel networks are equally likely, is assumed valid, then general explanations of network structure from basin geomorphic processes cannot be expected. Tests for the random model are therefore critical to the direction of future work. Proposed tests are based on frequencies of basins of different magnitudes or diameters, and on network maximum widths. Network topology is also shown to be potentially significant in the prediction of basin hydrographs. Network width commonly varies by a factor of 2 × for a given drainage area and drainage density, and is shown to influence peak discharge in proportion. Lag‐to‐peak can also be predicted better, using network topology. The paper pursues these applications of network topology using random walk theory and simulated random networks.