This paper compares the ability of two channel network simulation models to simulate the channel network properties of the Ashley River in the foothills of the Southern Alps of the South Island of New Zealand. The basin was chosen because it contains a large-scale topographic feature, a central ridge, the simulation of which would provide a measure of each model's ability to handle spatial nonuniformities. The two models assessed were the optimal channel network (OCN) model and a catchment evolution model (SIBERIA) catchment evolution model. The ability of these models to replicate observed geomorphic statistics and relationships was assessed. The models were also compared against each other to assess their relative suitability for simulating the observed geomorphology. Model performance was gauged using the single-valued measures of catchment convergence, hypsometric integral, and energy expenditure by the network; and plots of the width function, the slope of the cumulative area diagram, and the hypsometric curve of basin topography. The effects of different forcing were examined. The basic forcing was one in which the climate and the geological and tectonic properties of the basin were assumed to be unchanging in both space and time. The first variant on the basic forcing looked at what happens when there is a permanent spatial gradient in the rainfall over a basin with spatially uniform geological and tectonic properties. The second variant considered the effects of constant, spatially uniform rainfall on a basin in which spatially variable tectonic uplift is occurring. Neither model adequately simulated the observed geomorphology when the spatially nonuniform tectonic forcing was ignored. When spatially nonuniform tectonic effects at length scales of tens of kilometers were simulated, SIBERIA performed more satisfactorily. The effect of nonuniform rainfall was found to be small for both models. The performance of the OCN as gauged by single-valued measures improved markedly when energy expenditure calculations were consistent with the geometric length of flow paths. ¿ 1999 American Geophysical Union