A Modeling Study on the Extreme Rainfall Event along the Northern Coast of Taiwan on 2 June 2017

Abstract

<p>In the present study, two experiments at 1-km grid size are compared on the extreme-rainfall event (up to 635 mm in 12 h) along the northern coast of Taiwan on 2 June 2017. The first is a forecast (F1) that was driven by the best forecast member in a previous study at 3-km grid size (360 mm) and produced a peak amount of 541 mm in 24 h along the northern shore (618 mm at 15 km offshore), and the second is a simulation (S1), driven by a 3-km simulation (S3), that produced a maximum of 393 mm over land. The frontal moving speed are both slow in the two experiments, and the front spends slightly more time moving across the area in S1 (15 h) than in F1 (12 h). So, under similar conditions in most of the important factors previously identified, the present comparison allows us to examine and isolate the reasons for the main differences between the peak rainfall amounts.</p><p>The results show that in S1, multiple rainbands moved slowly across northern Taiwan to produce the rainfall, with a more widespread heavy rainfall area, but the peak amount was lower due to the trainsient nature of these rainbands (however slow they are). In F1, on the contrary, linked to the presence of a low-pressure disturbance along the front to the north-northwest of Taiwan, the low-level westerlies was enhanced to produce a persistent convergence zone with the cold northeasterly air behind the front, right across the northern coast of Taiwan without movement. Thus, one rainband remained at more or less the same location over several hours, and produced the localized peak rainfall reaching 541 mm on land. The results indicate that it is possible to produce a higher peak amount from km-scale models with a finer grid size, but the predictability of such extreme amount might be relatively low and requires further study.</p>