Parallel calculation of flow directions on hexagonal discrete global grid systems

Global small-scale hydrological modeling datasets are in demand in many geosciences. The results of corresponding simulations include flow directions, catchment area, watershed delineation and stream networks. Regardless of the method of obtaining global product, one will face the task of processing at least continent-scale datasets, which may be time-consuming or hardware-expensive. The recent research on hydrological modeling shows that calculations performed on a hexagonal mesh grid outperformed those on a rectangular one. Therefore, there is increasing interest in globalscale hydrological simulations on hexagonal grids. Discrete global grid systems (DGGS) which are spatial reference systems that use a hierarchy of equal area tessellations to partition the surface of the spherical Earth into grid cells; it seems to be an efficient way to manage big geospatial data. There are studies where hydrological algorithms are successfully applied on hexagonal DGGS, although locally. This research presents an algorithm for parallel computation of flow directions and upslope area on the hexagonal DGGS using the example of the African hydrological system. Referring to the hierarchical nature of DGGS, we cut the study area into tiles along the cells’ boundaries of one of the small-scale levels. Hydrological modeling is then performed on the desired level child cells of each tile. Afterwards the results are stitched into a single coverage. This study is practical not only for hydrological analysis, but also for combining the results of raster algebra analysis in any other areas