CNOP based on ACPW for Identifying Sensitive Regions of Typhoon Target Observations with WRF Model

Abstract. In this paper, we rewrite the ACPW (adaptive cooperation co-evolution of parallel particle swarm optimization and wolf search algorithm based on principal component analysis) and applied it to solve conditional nonlinear optimal perturbation (CNOP) in the WRF-ARW for identifying sensitive areas of typhoon target observations, which is proposed by us in the study of Zhang et al. (2018), to investigate its feasibility and effectiveness in the WRF-ARW model. Fitow (2013) and Matmo (2014) are taken as two typhoon cases, and simulated with the 60 km horizontal resolution. The total dry energy is adopted as the objective function. The CNOP is also calculated by the method based on the adjoint model (ADJ-method) as a benchmark. To evaluate the ACPW-CNOP, five aspects are analysed, such as the pattern, energy, similarity, benefits from the CNOPs reduced in the whole domain and the sensitive regions identified, and the simulated typhoon tracks. The experimental results show that the temperature and wind patterns of ACPW-CNOP is similar to those of the ADJ-CNOP in all typhoons. And the similarity values of ADJ-CNOP and ACPW-CNOP of two typhoon cases are more than 0.5. When reducing CNOPs in the sensitive regions, the forecast income of ACPW-CNOP is greater than that of ADJ-CNOP in all typhoons. Moreover, the sensitive regions identified by the ACPW-CNOP has the similar influence with the ADJ-CNOP on the simulation of typhoon tracks, sometimes the ACPW-CNOP has more positive impact on the simulation of typhoon tracks. The ACPW is more efficient than the ADJ-method in this paper.