For the analyses of the excavation at HLO1 I created a digital terrain model of the site. A problem I encountered was caused by the wadi edge west of the site (see picture).
For the surface interpolation I have measured hundreds of points all over the site and down towards the wadi. I have also measured the edge of the wadi terrace. As software I used GRASS-GIS (Version 6.4). Using all the points in one step yielded a Digital Terrain Model (DTM), where the edge of the Wadi is smoothed. Therefore I decided to do the interpolation in two steps.
In the next step I used QGIS do digitest a close vector polygon (blue in the next picture) which includes all the points west of the faultline. This polygon I imported back in GRASS. You can also do this step directly in GRASS.
In the next step I wanted to use this polygon to create an vector mask. Before I could do this I had to run v.category on the vector layer with the polygon.
Now I could run v.to.rast to get a raster out of the close polygon. I used this raster as a mask with the command r.mask. Because of the mask the calculation in the next step is limited to the mask.
Last but not least I could calculate the surface west of the fault line. I used for example v.surf.rest to create an DTM from the west part.
In the next step I had to use again the r.mask command but now with the -i flag to get an inverse Mask; Then we get a mask from the rest of the map and can calculate the DTM again.
To combine the two raster files we have just created, we can use r.patch to get one big raster file from the whole site.
And here is the result: