Characteristics of Extreme Value Statistics of Annual Maximum Monthly Precipitation in East Asia Calculated Using an Earth System Model of Intermediate Complexity

Tosiyuki Nakaegawa,Hirotaka Kamahori,Takuro Kobashi

doi:10.3390/atmos11121273

Abstract

Extreme precipitation is no longer stationary under a changing climate due to the increase in greenhouse gas emissions. Nonstationarity must be considered when realistically estimating the amount of extreme precipitation for future prevention and mitigation. Extreme precipitation with a certain return level is usually estimated using extreme value analysis under a stationary climate assumption without evidence. In this study, the characteristics of extreme value statistics of annual maximum monthly precipitation in East Asia were evaluated using a nonstationary historical climate simulation with an Earth system model of intermediate complexity, capable of long-term integration over 12,000 years (i.e., the Holocene). The climatological means of the annual maximum monthly precipitation for each 100-year interval had nonstationary time series, and the ratios of the largest annual maximum monthly precipitation to the climatological mean had nonstationary time series with large spike variations. The extreme value analysis revealed that the annual maximum monthly precipitation with a return level of 100 years estimated for each 100-year interval also presented a nonstationary time series which was normally distributed and not autocorrelated, even with the preceding and following 100-year interval (lag 1). Wavelet analysis of this time series showed that significant periodicity was only detected in confined areas of the time–frequency space.

Highlights

Extreme precipitation has the potential to damage human lives and properties through inundation, flooding, landslides, etc
We investigated the characteristics of extreme value statistics of the annual maximum monthly precipitation by simulating climate change on a time scale of thousands to tens of thousands of years over the past 12,000 years with LOVECLIM providing a nonstationary time series of data in order to directly consider nonstationarity in the extreme value analysis instead of proxy data
We focused on a target grid in East Asia and performed extreme value analysis to obtain the estimated annual maximum monthly precipitation with a return level of 100 years

Summary

Introduction

Extreme precipitation has the potential to damage human lives and properties through inundation, flooding, landslides, etc. Multiple countermeasures have been implemented to prevent and mitigate precipitation-induced disasters and damage. Have revealed that heavy precipitation extremes have increased in both frequency and intensity in recent years. Such extremes are partly attributed to recent global warming [1]. As outlined in previous some studies, the occurrence of extreme precipitation is no longer stationary under a changing climate with increasing greenhouse gas emissions. Extreme precipitation with a certain return level is typically estimated using extreme value analysis under a stationary climate assumption without evidence when a countermeasure is quantitatively designed, such as a flood control plan. In formulating a flood control plan, that is, a high-water plan, Atmosphere 2020, 11, 1273; doi:10.3390/atmos11121273 www.mdpi.com/journal/atmosphere

Methods

Results

Conclusion