The Impact of Horizontal Resolution on the CNOP and on Its Identified Sensitive Areas for Tropical Cyclone Predictions

In this study, the impacts of horizontal resolution on the conditional nonlinear optimal perturbation (CNOP) and on its identified sensitive areas were investigated for tropical cyclone predictions. Three resolutions, 30 km, 60 km, and 120 km, were studied for three tropical cyclones, TC Mindulle (2004), TC Meari (2004), and TC Matsa (2005). Results show that CNOP may present different structures with different resolutions, and the major parts of CNOP become increasingly localized with increased horizontal resolution. CNOP produces spiral and baroclinic structures, which partially account for its rapid amplification. The differences in CNOP structures result in different sensitive areas, but there are common areas for the CNOP-identified sensitive areas at various resolutions, and the size of the common areas is different from case to case. Generally, the forecasts benefit more from the reduction of the initial errors in the sensitive areas identified using higher resolutions than those using lower resolutions. However, the largest improvement of the forecast can be obtained at the resolution that is not the highest for some cases. In addition, the sensitive areas identified at lower resolutions are also helpful for improving the forecast with a finer resolution, but the sensitive areas identified at the same resolution as the forecast would be the most beneficial.