The definition of the percentile-based extreme rainfall index takes regional variances into account and is consistent with earlier studies [26]–[28]. If the daily rainfall for a certain gridded period (1981–2020) exceeds the threshold of 90th percentile every rainy day records (daily rainfall > 0.1 mm), it is considered exceptional rainfall. Every grid establishes its threshold in a same way.
The 90th percentile criterion for daily rainfall across Myanmar is shown in Fig. 2a. With the exception of the steady northern area, the thresholds progressively rise from Myanmar's central to southern regions, ranging from 10 to 100 mm at two costal core regions. It is agreeded with the spatial spread of June-September (JJAS) mean rainfall of the previous study [6], since the percentile threshold rainfall is mainly decided by southwest summer monsoon rainfall. ERDs are defined as the number of days where the daily rainfall surpasses the relevant threshold. It should be noted that the ERDs defined here have strong variability that is not affected by the selected criteria.
The southwest monsoon core region, which includes the southern and western coastal regions, experiences more annual mean rainfall than the northern portion of Myanmar, with the exception of the centre dry region (Fig. 2b). Here, we neglect the central dry region in this study due to topography climatology features. The ratio of the 90-percentile threshold to the annual mean rainfall is exhibited in (Fig. 2a). The ratio is excessive in northern Myanmar, especially in northern Kachin state, while the ratio is decreased in central dry and southern Myanmar (Fig. 2c). This is because the variation in daily rainfall in dry regions is significantly greater than that at wet area. Siginificantly, over the centeral and southern delta regions of the Ayeyarwady River, the annual north-south mean rainfall has variation, but the threshold values in the north is ecxeed than in the south, indicating that the Ayeyarwady River basin's north has more variation in its daily rainfall than its south (Fig. 2d).
The current study only targets on the ERDs in Myanmar since the country is mostly the only place where heavy rainfall events occur. May through October is the major rainy monsoon season in Myanmar [29]–[31]. The climatology seasonal variation of the climatology monthly mean ERDs from May to October is depicted in Fig. 3. Without a doubt, the monsoon onset season, which runs from May to July, is when the significant values of climatological ERDs primarily appear, gradually moving northward from the south.
The monthly trend of days with excessive rainfall (daily rainfall > 50 mm) exhibits a similar northward trend, but the monsoon regions tend to experience heavier rainfall (Fig. 4). Therefore, the JJAS months might be thought of as Myanmar's season with the heaviest rainfall. Observe that the maximum numbers of ERDs were mainly occurs in two monsoon core regions (MC, western coastal 92°E-95°E 16°N-23°N and southern coastal 96°E-99°E and 11°N-19°N ) during June-July (JJ), while the large value of ERDs concentrates in northern Myanmar (NM, north of 92°E-99°E and 24°N-30°N) during August-September (AS). This characteristic of the seasonal variation of ERDs divides all of Myanmar's regions into two local domains, MC and NM.
The annual cycle count of ERDs is depicted in Fig. 5 as an average over MC and NM (orange bar). The ERDs averaged over MC has a peak in JJ, while ERDs over NM has a peak in AS for summer. The green bar in Fig. 5 illustrates this same characteristic, which corresponds to annual fluctuations in the monthly mean rainfall. Here NM has an annual distribution of ERDs have occurred one to two days every month of years. For the annual time scale across NM, the real peak ERD months are February, March, and April. However, because MC showed a temporal homogeneous pattern, we limited the investigation to the summer months. Therefore, JJ and AS could be taken into account as the mainland Indochina southwest monsoon rainfall (MSWM) season for MC and NM. The two area-averaged ERDs indices, namely ERDs-MC and ERDs-NM, are specified at each local domain during their summer southwest monsoon rainfall season, to support the well understanding and accurate estimation of the ERD variability over Myanmar.
The time series analysis of ERDs-MC and ERDs-NM are shown in Fig. 6. It is understand that the ERDs-MC (ERDs-NM) and the mean seasonal rainfall averaged over MC (NM) exhibit extremely annual variation. The TCC for MC and NM is 0.99 and 0.96, respectively, suggesting that the mean monsoon rainfall over Myanmar and seasonal ERDs may have somewhat comparable sources of predictability. Intriguingly, the area average of mean temperature over MC exhibits a substantial linear up trend of 0.005°C/year at the 99 percent confidence level, which is consistent with a considerable increasing trend of the ERDs-MC (0.04 days/year) during 1981–2020. On the other hand, during the same time, the mean temperature averaged over NM shows a substantial (at 95% confidence level) increasing trend of 0.07°C/year, however the ERDs-NM show an essentially up trend (0.02 days/year). This may point to a global warming-related trend in intense rainfall that is independent of the area.