4.1 The correlation between β and extreme characteristics
of precipitation
In order to further explore the parameters α and β of PDC
on a national scale, we analyzed the Spearman’s rank correlation
coefficient (Spearman, 2010) between them and the physical control
factors to obtain their possible influencing factors. As shown inFigure 6 , β is inversely proportional to the Aridity
index (AI ), and has a high correlation with P and
precipitation percentile, with a correlation of 0.9816 with 99 wet day
percentile precipitation. It further confirms the conclusion of Geng, S
to some extent that catchments with higher precipitation tend to have
larger values of β which are estimated by gamma distribution
(Geng et al., 1986). By comparing Figure 6 (c) (d) (e) and (f),
it can be seen that the larger the values of β , the more
dispersed the probability distribution of precipitation at different
levels is, with the greater likelihood of extreme precipitation
occurrence.
Schär C et al. pointed out that the precipitation percentile (or
quantile) was used to evaluate the trend and prediction of heavy
precipitation events, and analyzed the temporal and spatial distribution
and change characteristics of extreme precipitation threshold (Schär et
al., 2016). In evaluating extreme events such as rainstorm, the use of
events such as drizzle (precipitation less than 0.1 mm or 1 mm) seems
illogical, so the wet day precipitation threshold has a better
performance on the extreme precipitation threshold of heavy
precipitation events.