Probability Distribution of Precipitation Extremes for Weather Index–Based Insurance in the Zhujiang River Basin, South China

Abstract

In a changing climate, understanding the frequency of weather extremes is crucial to improving the management of the associated risks. The concept of weather index-based insurance is introduced as a new approach in weather risk adaptation. It can decrease the vulnerability to precipitation extremes that cause floods and economic losses in the Zhujiang River basin. The probability of precipitation extremes is a key input and the probability distribution of annual precipitation extremes is analyzed with four distribution functions [gamma 3, generalized extreme value (GEV), generalized Pareto, and Wakeby]. Three goodness-of-fit tests (Kolmogorov-Smirnov, Anderson-Darling, and Chi Squared) are applied to the distribution functions for annual time series of 192 meteorological stations. The results show that maximum precipitation and 5-day-maximum precipitation are best described by the Wakeby distribution. On a basin scale, the GEV is the most reliable and robust distribution for estimating precipitation indexes for an index-based insurance program in the Zhujiang River basin. However, each station has to be analyzed individually as GEV is not always the best-fitting distribution function. Based on the distribution functions, spatiotemporal characteristics of return periods for maximum precipitation and 5-day-maximum precipitation are determined. The return levels of the 25-and 50-yr return periods show similar spatial pattern: they are higher in the southeast and lower in the southwest of the basin. This spatial distribution is in line with the annual averages. The statistical distribution of precipitation indexes delivers important information for a theoretical weather index-based insurance program.

Uncertainty analysis of hydrological return period estimation, taking the upper Yangtze River as an example

“…As mentioned in the introduction, many distributions have been used in return period estimations (Park et al, 2001;Öztekin, 2007;Su et al, 2009;Fischer et al, 2012). The selection of only five distributions is one of the primary limitations of the 5 current study.…”

Variations and statistical probability characteristic analysis of extreme precipitation events under climate change in Haihe River Basin, China

“…Fischer et al . () analysed the precipitation extremes in the Pearl River with four distribution functions and showed that on basin scale, the generalized extreme value (GEV) is the most reliable and robust distribution for estimating precipitation indexes for an index‐based insurance programme in the Pearl River Basin. Ding et al .…”

“…Moreover, many scholars apply statistical distributions to make further research on extreme precipitation events. Fischer et al (2012) analysed the precipitation extremes in the Pearl River with four distribution functions and showed that on basin scale, the generalized extreme value (GEV) is the most reliable and robust distribution for estimating precipitation indexes for an index-based insurance programme in the Pearl River Basin. Ding et al (2009) fitted the extreme rainfall in summer in China by generalized Pareto (GP) and found the excellent simulation effects of the statistical model.…”

Probabilistic Distributions for <i>Acacia Mearnsii</i> De Wild Total Height and the Influence of Environmental Factors

“…The Wakeby distribution has been mainly used for modeling extreme events and in the most cases for hydrology studies, for example researches developed by [40][41][42][43][44][45]. The distribution has five parameters, more than the most used ones.…”