A simple, fast parallel implementation of GRAPES 3DVAR

The three-dimensional variational (3DVAR) is an objective analysis approach developed in the late 1980s. Currently, it is one of the most common data assimilation methods used in numerical weather prediction (NWP). The global and regional assimilation prediction system (GRAPES) is the Chinese new generation NWP system, in which 3DVAR is one of the main components and plays an important role in direct assimilation for non-conventional observations. After giving a brief description of the GRAPES 3DVAR, this paper puts forward a simply, fast parallel scheme of the GRAPES 3DVAR. The testing results on an 8-node cluster show that the parallel implementation achieves considerable load balancing and parallel efficiency