Quadtree and Sub-Grid Sampling FAQ: Difference between revisions

You can get accurate results with less 2D cells across the width of flow if you use SGS with a high resolution DEM or preferably a TIN that is based on accurate field surveys. The quality of the channel’s bathymetry is critical to the quality of the modelling, ideally the TIN would extend to dry ground if the data are available to avoid bumpy terrain data from the DEM along the edge of the TIN within the channel (this is even more so for concrete trapezoidal channels).<br>

Having less 2D cells is however not recommended when using gully lines or wide Z shape gully lines, as they only modify the ZU, ZV, ZC and ZH elevations – there is no high resolution SGS sampling so the resulting modification produces a zig-zag channel perpendicular to the cell faces. For wide gully lines this zig-zag channel can be modifying different number of cell faces so the flow or conveyance of the channel can be continually constricted and expanded as a consequence. They also won’t be a good representation of the true shape of the channel. The constant change in flow area and the channel not having a smooth conveyance could also cause the flow to repeatedly transition from subcritical to supercritical causing the water level to be erratic (Froude number to be jumping around 1). When the channel runs in supercritical flow regime, it is more important to check that the model is conserving the water energy level well (include "E" into <font color="blue"><tt>Map Output Data Types</tt></font> ). SupercriticalFlow flowsclose canto bea veryFroude challengingnumber of 1 is always likely to be jumpy as so much energy is hidden in the velocity of the water, but as long as the energy line is not bumpy, the water surface profile can be regarded as a sound result.<br>