This paper reports on generalization and data modeling to create reduced scale versions of the National Hydrographic Dataset (NHD) for dissemination through The National Map, the primary data delivery portal for USGS. Our approach distinguishes local differences in physiographic factors, to demonstrate that knowledge about varying terrain (mountainous, hilly or flat) and varying climate (dry or humid) can support decisions about algorithms, parameters, and processing sequences to create generalized, smaller scale data versions which preserve distinct hydrographic patterns in these regions. We work with multiple subbasins of the NHD that provide a range of terrain and climate characteristics. Specifically tailored generalization sequences are used to create simplified versions of the high resolution data, which was compiled for 1:24,000 scale mapping. Results are evaluated cartographically and metrically against a medium resolution benchmark version compiled for 1:100,000, developing coefficients of linear and areal correspondence.