Forcing and variability of the hydroclimate in Central Asia

Abstract

Central Asia is the semiarid region whose territory is roughly defined by the boundaries of Kazakhstan, Uzbekistan, Kyrgyzstan, Turkmenistan, Tajikistan, and the northern part of Afghanistan. It is located in the interior of the Eurasian land mass at about equal distance from the Atlantic and Pacific Oceans, just upstream of the planet’s largest topographic barriers, and in the transition zone between temperate and subtropical continental climates. In addition, it is situated close to the regions known to benefit from summer monsoon precipitation, but is isolated from them by the Karakorum and Hindu Kush mountains. Central Asia has received the attention of a global public regarding the rapid desiccation of the Aral Sea. The combination of these facts highlights that a sound understanding of the Central Asian hydroclimate is imperative both for purely academic interest and potential societal benefit. This thesis addresses a number of key issues in order to contribute to a better understanding of this kind. Chapter 2 complements the introductory first chapter. In terms of widely used techniques of multivariate statistics (principal component analysis and maximum covariance analysis), the dominant modes of the wintertime 500 hPa geopotential height variability over western Eurasia are identified and their covariability with the climate in Central Asia is analysed. It is shown that the Scandinavia Pattern is particularly relevant in this context, since its connexion with both temperature and precipitation in Central Asia is comparatively close. Thereafter, an intercomparison of a wide range of observational and numerical precipitation data sets is presented in chapter 3. Data sets based on re-analyses, direct observations, and numerical downscaling are validated and compared with respect to their ability to represent the Central Asian precipitation climate. The mutual agreement between different observational data sets is used as an indication of whether these data can be used for the validation of data from other sources. In particular, it is shown that the observational data usually agree qualitatively on the occurrence of anomalous years or seasons, while there is considerable disagreement regarding the amplitude of interannual variability. The climate high resolution model (CHRM) is shown to improve the spatial characteristics of the precipitation distribution. At the same time, it strongly underestimates summer precipitation and its variability, while interannual variations are well represented in other seasons, in particular during winter and spring.