Runoff cascades, channel network and computation hierarchy determination on a structured semi-irregular triangular grid

When developing a physically based mathematical description of hydrological processes over natural watersheds, the modeler faces the requirement of complete spatial specification of the various hydro-geological parameters as well as the ground surface characteristics. The usual approach, i.e. the use of representative average values in simplified configurations of the actual conditions, often leads to discrepancies. A detailed description of the topography is therefore needed but, on a complex topography, the accurate determination of overland cascades and streams often poses difficult problems. The proposed method aims to provide some more insight into the above-mentioned problems. In order to simulate overland trajectories of runoff water and channel flow paths, the method represents the topography using a network of non-overlapping triangular surfaces of known properties (i.e. slope, orientation, area, soil texture, land-use), created with an appropriate digital elevation model. The simulated overland cascades may feed channels, terminate to local minima of altitude or reach the boundaries of the domain under consideration. The model allows for the use of variable resolution and has nesting capabilities. The method has been applied on the Lucky Hills watersheds at Walnut Gulch, Arizona, using four different resolution schemes. The examination of the model's output shows that the various overland trajectories and channel flow paths are accurately computed in accordance with the model's specifications. Additionally, each simulated river network ended up at an outlet that precisely coincided (within the resolution) with the fixed position of the measurement station. No tuning efforts were undertaken and the model did not need calibration. This method seems to be advantageous compared to many of those in use, since it is fully automatic and time saving while proven to accurately simulate the drainage network. In addition, the method creates a detailed database easily accessible for the forthcoming assessment of the hydrologic response of the watershed in the frame of a distributed model.