An Optimal Scheduling Method for Numerical Weather Model Assimilation with Dense Observations

In this paper, the scheduling problem of dense observation data in a numerical weather model assimilation system is studied and an algorithm of "deal-reveal" is proposed. The algorithm identifies the area where all dense observation stations are located, and then obtains a certain number of grid point on which the value represents the stations quantity overlaid on it. The algorithm selects the stations with maximum value corresponding to the dense observation data iteratively and make the amount of each batch of dense observation stations as more as possible to reduce the batch number, enhances the overall performance of the system finally.