4.2 Implementation of incremental analysis updates (IAU) in 1-hr updated cycling forecasting system.
Chen, Min, Institute of Urban Meteorology, CMA, Beijing, China, and Xiang-yu Huang, Centre for Climate Research Singapore, Singapore
Rather than the
widely used DFI technique, an incremental analysis updates (IAU) scheme is
successfully implemented into a WRF/WRFDA-based 1-hr updated cycling forecast
system with the goal to reduce the imbalance introduced by the high-frequency
intermittent data assimilation. With the application of IAU, the analysis
increments are smoothly introduced into the integration with time-weighted
optimal digital filter coefficients during the [-1h,1h] time window centered at
the analysis time.
As only a tiny part of analysis increments of all prognostic variables is
incorporated at each integration step, it makes the model away from the great
shock at the initial time due to the intermittent data assimilation. Therefore,
IAU leads to reasonable and desirable results in terms of the initial noise and
hydrometeor spin-up for the 1-hr updated cycling system. In addition, the
precipitation forecasts of the continuous cycles are much more stable, and
precipitation forecast skills are significantly improved (around 10+% for CSI
scores). On the performance of data assimilation, another advantage of IAU can
be found that the better balanced 1-st guess generated by the previous cycle
forecasts potentially give rise to more data can be assimilated and faster
minimization.