Remote sensing versus surveyed object-based cadastre data: Comparing the advantages of COPERNICUS imperviousness density and ATKIS data in Bavaria/Germany using GIS analysis

As the built-up areas extend and become denser with time, high-resolution land use and especially imperviousness data is of increasing importance, e.g., for detailed studies on settlement and landscape water retention to provide for extremes such as (flash) floods, heatwaves, and droughts.

The open-source COPERNICUS imperviousness density dataset covers 2006 until 2018 with a three-year timestep with incremental raster resolution (2018: ten meters). On the other hand, there is an accurate object-oriented cadastre dataset of German authorities called ATKIS. It describes the geometries of buildings and all types of traffic routes from airports to dirt roads with an increasing amount of attributes like shape, area, width, or material. While both datasets are very detailed, they have specific (dis-)advantages due to their very different data and surveying type.

We used the information stored in the ATKIS database of 2020 to create an alternative object-based imperviousness map of the German state Bavaria with roughly 70,000 km² to contrast and compare it with the COPERNICUS imperviousness density dataset of 2018 in several different (urban and rural) areas. We found that COPERNICUS indicates much higher imperviousness for densely settled areas as city centres and commercial and industrial zones and can describe even complex types of significant extent (such as golf courses, transhipment stations, and allot settlements) better. Vice versa, ATKIS could resolve even linear traffic elements in rural areas and detached house settlements to an outstanding level of detail, which COPERNICUS cannot afford due to its limited resolution. Both products cannot distinguish clearly between sealed concrete and loose stony material (e.g., from construction or mining sites), nor give indications on sub-surface water retention or sewage infrastructure.

Overall, we can name both methods' (dis-)advantages, relate them to surprisingly distinct land use classes, and give guidance, where additional object-oriented information can significantly improve COPERNICUS imperviousness data. Finally, the resulting maps highlight the hotspots of extreme building-independent imperviousness. They can serve as a tool to prioritize areal and building-centred measures to prevent large amounts of fast runoff.