What is the issue or question you have?My road data is a linear shp file. How can I convert it into a binary raster?
I set a buffer for the linear shp file according to the general width of the road (about 30m), and then converted the surface to raster to generate a binary raster, 1 represents the road, 0 represents no road. But my LULC resolution is 1km, which makes the map with the shape of the road discontinuous and becomes a few scattered grids. (Square “points”) Does it make sense to do this?
Also, the maximum impact distance in the threat table is the reference for invest calculation, right? When I created the binary raster, this maximum impact distance has nothing to do with me? Because if the buffer width of my linear road is set according to this maximum impact distance. Then in the end, invest still calculates the same thing again, which is equivalent to repeating.
Thank you for your reply! ! ! !
If your LULC resolution is 1km, then that will be the resolution used by the model. The User Guide says this about the threat rasters:
The extent and resolution of these raster datasets does not need to be identical to that of the input LULC maps. In cases where the threats and LULC map resolutions vary, the model will use the resolution and extent of the LULC map.
It is unfortunate if the LULC map is so coarse, since the smallest unit that the model can work at is 1 pixel, or 1km. As you’re seeing, converting 30m to 1km can create unrealistic results. One thing that you could try is converting the LULC map to 30m resolution. You’ll still have large (1km) blocks of data with the same value, but you could use threat rasters at 30m more easily.
The maximum impact distance is used to essentially create a buffer around your threat features. So if your road threat distance is 1km and your road raster is 1km in resolution, the threat will be applied for 1km/1 pixel around the roads in the threat raster, even though the roads are already 1km. If you convert your LULC raster to 30m, then you can make a 30m roads threat raster that has a 1km buffer around it, which might be more realistic.