Tidal networks 1. Automatic network extraction and preliminary scaling features from digital terrain maps

We propose a method of automatic extraction of the tidal channel network fr om topographic data of marsh and tidal flat lands that uses a combination o f a threshold elevation and threshold curvature. Not only the location but also the area of the channel bed is identified. This method differs substan tially from that used to identify terrestrial channel networks, and it succ essfully predicts all of the main channels and nearly all of the smaller tr ibutaries of the channel networks derived from SPOT imagery of the northern Venice Lagoon. Channel network maps of Venice and other sites (Petaluma Ma rsh in the San Francisco Bay and Barnstable Marsh in Massachusetts) were ex amined for scaling properties. Because of the large width of the channels r elative to a characteristic length of their drainage area, we had to develo p procedures for automatically delineating channel width and then for ident ifying the skeleton of the channel network (the pattern connecting the loci of the channel centerlines) for box-counting analysis. Box-counting dimens ions of the network skeleton proved site-dependent and showed finite-size e ffects. Because of the large widths we also performed a scaling analysis ba sed on the proportions of the total channel bed area occupied by the tidal networks (i.e., a "fat" fractal analysis). This analysis showed a strong br eak in scaling between large and small channels. These analyses suggest tha t tidal channels differ significantly in their scaling relationships from t errestrial systems. In subsequent papers [Rinaldo et al., this issue (a), ( b)] we pursue this point much further.