The projected occurrence changes of the lower atmospheric ducts accompanied with sea fog under different CMIP6 scenarios across the Hangzhou Bay

In this study, diagnostic schemes for sea fog and lower atmospheric ducts were established based on the WRF model. Using these schemes, four groups of ten-year continuous simulations were conducted for the Hangzhou Bay area, covering the historical period and three future scenarios, to investigate the spatio-temporal characteristics of lower atmospheric ducts associated with sea fog and their response to changes under different emission scenarios. The simulation results showed that in the Hangzhou Bay area, the occurrence rates of elevated ducts and simple surface ducts were 5.81% and 5.65% respectively, while the occurrence rates of surface-based ducts and composite ducts were relatively low, at 3.15% and 0.25%. The duct rate exhibited a bimodal monthly pattern in summer and autumn with a trough in winter, which was consistent with the onset and withdrawal of the East Asian summer monsoon. Among sea fog-associated ducts, elevated ducts accounted for the highest proportion (27.43%), while the proportion of simple surface ducts was extremely low. Their peak occurrence coincided with the high-incidence season of sea fog, slightly earlier than the overall ducts’ average. Under future scenarios, the occurrence rates of the four types of atmospheric ducts increased with rising carbon emissions, with the largest increases in autumn and winter and a weaker increase in spring. For sea fog-associated ducts, due to the decrease in sea fog events during the scenario periods, the changes in different duct types varied. The occurrence rates of simple surface ducts and surface-based ducts both decreased significantly compared to the historical period. The average rates of elevated ducts and composite ducts increased slightly, with the increasing areas concentrated in the eastern offshore areas. However, the amplitude of seasonal oscillations in these two types of ducts was strengthened during the scenario period, and this amplitude showed a negative correlation with carbon emissions.