A global high-resolution atmospheric general circulation model with grid size about 20 km is used to project future changes in rainfall extremes associated with El Niño at the end of the 21st century. In the future climate projections, hypothetical sea surface temperature (SST) is assumed where the interannual variability of SST remains same as in the present climate. The annual maximum 1-day precipitation total (R1d) over the western North Pacific is largely associated with tropical cyclone activity and positively correlated to the Niño3.4 SST anomalies. It is found that climatological mean R1d will only modestly increase in the western North Pacific in the future, but interannual variability of R1d will largely increase compared to the present due to enhanced association with El Niño. This implies an increasing risk of heavier rainfall events by global warming around the western North Pacific countries.
This paper evaluates the ability of a regional-scale climate model to simulate precipitation over the South Asian tropical region. Experiments were conducted using three different spatial resolutions, with and without cumulus parameterization (CP), to assess the influence of horizontal mesh size and the CP on regional-scale precipitation. The experiments that used a finer mesh size but no CP improved the spatial distribution of monthly precipitation relative to that in the experiments based on a coarser spatial resolution. Meanwhile, the impact of horizontal mesh size was much less in the experiments that included the CP because an overestimation of precipitation caused by the CP strongly affected the simulation accuracy in these experiments. Regional differences in diurnal variations in precipitation intensity and frequency were captured reasonably well in the pre-monsoon season regardless of the spatial resolution, and both with and without the CP. In contrast, the diurnal characteristics of precipitation were difficult to simulate during the mature monsoon season. During both seasons, those experiments that incorporated the CP tended to predicted a continuous weak precipitation due to the excessive release of convective instability; accordingly, precipitation intensity was weaker, and precipitation frequency greater than in those experiments that did not use the CP.