2:15 PM - 2:30 PM
[ACG46-08] Forecast skill of Arctic atmosphere and its relation to cyclone development by multi-resolution atmosphere–ocean coupled models
Keywords:cyclone, atmosphere-ocean coupled simulation, model resolution
This study investigated the influences of the atmosphere–ocean coupled processes on the weather forecasts over the Arctic using multi-resolution atmospheric and oceanic coupled systems. We analyzed a series of 11-day forecasts by the Meteorological Research Institute (MRI) Earth System Model (MRI-ESM2) initialized at 00 UTC every day in September 2018. Four types of coupled experiments were conducted by the high-resolution atmospheric model coupled with high- and low-resolution oceanic models (referred to as HiHi and HiLo, respectively) and the low-resolution atmospheric model coupled with high- and low-resolution oceanic model (referred to as LoHi and LoLo, respectively).
Although the high-resolution oceanic model influenced positively in some cases and negatively in others, the difference in forecast skill of sea level pressure (SLP) over the Arctic (north of 65ºN) was larger between HiHi and HiLo than between LoHi and LoLo (Fig. 1a-c, e-g). These results suggest that the low-resolution atmospheric model cannot receive the effect of the increase in the oceanic resolution. We focused on the HiHi and HiLo experiments initialized on 17th September 2018, when a cyclone generated over the North Atlantic and traveled into the Arctic during the 7-day prediction (Fig. 1f, h). The difference in skill between HiHi and HiLo was associated with the cyclone position prediction. The origin of the difference between HiHi and HiLo can be traced to surface wind and temperature differences over the North Atlantic generated through the vertical mixing and pressure adjustment mechanisms around the Gulf Stream extension. The fine-scale SST gradient in the HiHi experiment generated higher potential vorticity (PV) at the lower-level troposphere around the extension region at lead times of 24 hours and 48 hours. Then, the PV merged with the existing cyclone northwest of the UK, leading to the difference in the position of the cyclone and associated warm and cold fronts. This study showed the detailed processes for the impact of the fine-scale SST structures on the cyclone, which was statistically shown in previous studies. Besides, although the high-resolution oceanic model influenced positively in some cases and negatively in others, the results indicate the importance of atmosphere–ocean coupled prediction because the mid-latitude ocean influences the Arctic ocean and sea ice through the cyclone on medium-range timescales.
Although the high-resolution oceanic model influenced positively in some cases and negatively in others, the difference in forecast skill of sea level pressure (SLP) over the Arctic (north of 65ºN) was larger between HiHi and HiLo than between LoHi and LoLo (Fig. 1a-c, e-g). These results suggest that the low-resolution atmospheric model cannot receive the effect of the increase in the oceanic resolution. We focused on the HiHi and HiLo experiments initialized on 17th September 2018, when a cyclone generated over the North Atlantic and traveled into the Arctic during the 7-day prediction (Fig. 1f, h). The difference in skill between HiHi and HiLo was associated with the cyclone position prediction. The origin of the difference between HiHi and HiLo can be traced to surface wind and temperature differences over the North Atlantic generated through the vertical mixing and pressure adjustment mechanisms around the Gulf Stream extension. The fine-scale SST gradient in the HiHi experiment generated higher potential vorticity (PV) at the lower-level troposphere around the extension region at lead times of 24 hours and 48 hours. Then, the PV merged with the existing cyclone northwest of the UK, leading to the difference in the position of the cyclone and associated warm and cold fronts. This study showed the detailed processes for the impact of the fine-scale SST structures on the cyclone, which was statistically shown in previous studies. Besides, although the high-resolution oceanic model influenced positively in some cases and negatively in others, the results indicate the importance of atmosphere–ocean coupled prediction because the mid-latitude ocean influences the Arctic ocean and sea ice through the cyclone on medium-range timescales.