Toward refined tropical cyclone modeling via balancing multi-scale constraints in nudging-based downscaling
Cansheng He, Jiyang Fu, Yujie Liu, Yuncheng He
Abstract
Regional climate models (RCMs) are widely used for tropical cyclone (TC) simulation and forecast. However, RCM-based results of large-scale environments often deviate from those represented in the driving data, which inevitably results in modeling errors, especially for TC tracks. For downscaling research, a practical strategy is to use large-scale nudging in the outer domain (OD) and meanwhile avoid over-constraints for the dynamics in the inner domain (ID). Nevertheless, the effectiveness of the above strategy, particularly from the perspective of avoiding large-scale inconsistencies between RCM-based results and the driving data, remains less explored. This study extends the concept of domain-size sensitivity, previously framed in the context of large-scale or seasonal-to-interannual variability in regional climate modeling, to TC modeling under the OD-sited nudging framework, with a focus on the role of the unnudged ID size. These results show that oversized IDs reduce large-scale constraint, amplifying internal RCM variability and increasing track errors. Conversely, while reducing ID size improves the track accuracy, excessively small IDs degrade intensity representation. Based on multi-case analysis, the study offers preliminary practical guidance for ID configuration to balance track and intensity performance in TC downscaling applications.