Water is the source of life. Identifying small water bodies in rocky terrain (Guelta) becomes important
especially in desert regions with limited rain throughout the year. Saudi Arabia has many of these gueltas yet
their locations are not entirely known nor the ecosystem surrounding it is sufficiently studied. In this project
we have combined GIS with remote sensing to build an automated supervised classifier to identify these
gueltas. Before that can be used, the full cycle of data preprocessing and data quality checks were
implemented to clean the data. In terms of accuracy of known gueltas, ground truth locations were assessed
which revealed 70% of gueltas were correctly classified. Further more, random sampling technique was used
for 20 random coordinates, out of which five turned out to be gueltas. Also spectral signature plots were used
to study each Guelta independently. In this work we have identified an initial