I think, there is not major complexities. It depends of high resolution imagery data processed whit GIS.

I would suggest you try with some SAR data as you are looking to identify different materials.

One example below:

Article Automatic Detection of Low-Rise Gable-Roof Building from Sin...

Thanks! I guess that to identify roof areas and shape (gable roof or other shape....) is quite easy. While material type concrete, metal or clay tiles, seems to be a more difficult issue and to organize information in categories that could help us, later,to calculate water harvesting potential is a matter of method. Any direct experience?

Interesting application! LIDAR could get at the slope of roofs as well as solar radiation. High resolution imagery could help you get at the materials of the roofs. Gridded precipitation might help get at precipitation inputs. Of course, you'd want some sort of data from actual rainwater captures to validate your models with.

Thomas Dilts

Thanks Thomas

the idea is to have a fast and reliable idea of current roof water harvesting potential and than to identify methods and systems to increase it by suggesting how to improve quality of roof and quality of systems to collect water like gutters and storage systems without large investments. Many times in rural areal small and simple storage systems could really increase households resilience to droughts period

Have you looked into Open Street Map? There are many buildings properly vectorized. You can derive roof area easily. Data coverage depends on where you want to conduct your research. Also you might want to consider statistical time series analysis for local precipitation patterns to include rainfall intensity in your functions.

Sebastian Herwartz

Thanks both for building issue and rainfall suggestions. The focus is on some small pacific islands where they are facing increasing drought periods, i will check data coverage!

Dependence on the accuracy of space visualization to infer surfaces and identify similar sites for water harvesting using spatial analysis tools and geophysical extractio