DHyDAMO: consider refactor use of `zonal_stats` in `drreader`

[DanielTollenaar]

@RuudHurkmans; documentation suggests any rasterio raster will work for land-use: https://hkvconfluence.atlassian.net/wiki/spaces/DHYD/pages/598507527/Landgebruikskaart

This is not true as the land-use needs to be of signed integer type and -999 as nodata value:

HYDROLIB/hydrolib/dhydamo/io/drrreader.py

Lines 54 to 60 in 97f99b4

	lu_rast, lu_affine = self.unpaved.drrmodel.read_raster(landuse, static=True)
	lu_counts = zonal_stats(
	gpd.GeoDataFrame(catchments),
	lu_rast.astype(int),
	affine=lu_affine,
	categorical=True,
	all_touched=all_touched,

Possible fixes:

1. Include constraints in documentation: het formaat moet een raster zijn met -999 en met een "signed integer" waarde
2. Fix code, e.g. feeding the raster to the zonal_stats module in stead of the numpy array and the affine as by example in the rasterio documentation: https://pythonhosted.org/rasterstats/manual.html#zonal-statistics