Spatial and Temporal Variations of Extreme Precipitation in Central Asia during 1982–2020

Zhen Li,Zhongwei Yan,Yalin Tian

doi:10.3390/atmos13010060

Abstract

As one of the largest arid and semi-arid regions in the world, central Asia (CA) is very sensitive to changes in regional climate. However, because of the poor continuity of daily observational precipitation records in CA, the spatial and temporal variations of extreme precipitation in recent decades remain unclear. Considering their good spatial and temporal continuity, gridded data, such as Climate Prediction Center (CPC) global precipitation, and reanalysis data, such as ERA-Interim (ERA), are helpful for exploring the spatial–temporal variations of extreme precipitation. This study evaluates how well CPC and ERA can represent observed precipitation extremes by comparing the differences in eight extreme precipitation indices and observation data at 84 meteorological stations. The results indicate that the CPC (except for 1979–1981) is more suitable for depicting changes in precipitation extremes. Based on the CPC data for the period 1982–2020, we found that seven indices of precipitation extremes, including consecutive wet days (CWD), max1-day precipitation amount (Rx1day), max5-day precipitation amount (Rx5day), number of heavy precipitation days (R10), very wet days (R95p), annual total precipitation in wet days (PRCPTOT), and simple precipitation intensity index (SDII) have increased by 0.2 d/10a, 0.9 mm/10a, 1.8 mm/10a, 0.3 d/10, 8.4 mm/10a, 14.3 mm/10a and 0.1 mm/d/10a, respectively, and the consecutive dry days (CDDs) have decreased by −3.10 d/10a. It is notable that CDDs decreased significantly in the north of Xinjiang (XJ) but increased in Kyrgyzstan (KG), Tajikistan (TI), and eastern Turkmenistan (TX). The other indices increased clearly in the west of XJ, north of Kazakhstan (KZ), and east of KG but decreased in the south of KG, TI, and parts of XJ. For most indices, the largest change occurred in spring, the main season of precipitation in CA. Therefore, the large-scale atmospheric circulation in April is analyzed to contrast between the most and least precipitation years for the region. A typical circulation pattern in April for those extremely wet years includes an abnormal low-pressure center at 850 hpa to the east of the Caspian Sea, which enhances the southerly winds from the Indian Ocean and hence the transportation of water vapor required for precipitation into CA. This abnormal circulation pattern occurred more frequently after 2001 than before, thus partly explaining the recent increasing trends of precipitation extremes in CA.

Highlights

The combined effects of human activities such as emissions of atmospheric greenhouse gases and land uses have changed water circulation, leading to uneven spatial and temporal distribution of water resources, especially with changing precipitation extremes [1,2,3,4,5,6]
Donat et al [13] found that increases in total and extreme precipitation in dry regions were linearly related to the model-specific global warming, with implications for increased risk of flooding, for the world’s dry regions
The four steps included in the Ensemble empirical mode decomposition (EEMD) calculation are as follows: (1) a noise series is added to the target data; (2) the data with the added noise were decomposed into intrinsic mode functions (IMFs); (3) repeat (1) and (2) calculations 1000 times; (4) the final result can be obtained as the ensemble means of the corresponding IMFs of the decompositions

Summary

Introduction

The combined effects of human activities such as emissions of atmospheric greenhouse gases and land uses have changed water circulation, leading to uneven spatial and temporal distribution of water resources, especially with changing precipitation extremes [1,2,3,4,5,6]. According to the IPCC AR6 [15], anthropogenic global warming has caused an increase in the frequency, intensity, and/or amount of heavy precipitation events at a global scale, especially over most land regions with good observational coverage. We firstly analyzed the temporal–spatial distribution of eight precipitation indices (CDDs, CWDs, Rx1day, Rx5day, R10, R95p, PRCPTOT, and SDII, detailed in Table 1) in CA from 1979 to 2005 based on station observations, evaluated the applicability of ERA and CPC gridded data for characterizing climate extremes, and demonstrate the spatial–temporal variations of precipitation extremes in CA during 1982–2020 based on CPC, which was proven to be the more suitable dataset. Annual total precipitation in wet days with daily precipitation ≥ 1 mm Annual total precipitation in wet days divided by number of these days with daily precipitation ≥1 mm mm mm/d

Data and Methodology

Evaluating Methods

Pettitt Test

Trend Estimation Methods

Changes of Precipitation Indices during 1979–2005