Westerly jet stream and past millennium climate change in Arid Central Asia simulated by COSMO-CLM model
- First Online:
- 1.1k Downloads
This study tackles one of the most debated questions around the evolution of Central Asian climate: the “Puzzle” of moisture changes in Arid Central Asia (ACA) throughout the past millennium. A state-of-the-art Regional Climate Model (RCM) is subsequently employed to investigate four different 31-year time slices of extreme dry and wet spells, chosen according to changes in the driving data, in order to analyse the spatio-temporal evolution of the moisture variability in two different climatological epochs: Medieval Climate Anomaly (MCA) and Little Ice Age (LIA). There is a clear regime behavior and bimodality in the westerly Jet phase space throughout the past millennium in ACA. The results indicate that the regime changes during LIA show a moist ACA and a dry East China. During the MCA, the Kazakhstan region shows a stronger response to the westerly jet equatorward shift than during the LIA. The out-of-phase pattern of moisture changes between India and ACA exists during both the LIA and the MCA. However, the pattern is more pronounced during the LIA.
1.1 The prerequisite for studying the climate in ACA
Previous studies suggest that the recent warming trend is very likely to be human-induced (Santer et al. 1996; Hegerl et al. 1996; Stocker et al. 2013; Ramaswamy et al. 2006; Santer et al. 2003). The evolution of future global climate change has become one of the main concerns during the twenty first century. Central Asia, one of the largest deserts of the globe, is likely to be extremely vulnerable to the future warming (Chen et al. 2010). Recurring climatic phenomena can have a large influence on societies, economies, and human health, with extreme events potentially leading to crises of some kind. Extreme drought events, which affect extended areas and persist over a prolonged period, are defined as “exceptional drought events” (Shen et al. 2007). They may affect the economy, environment, and society of densely populated areas in Asia, hence there is an urgent need to address this issue for planning and securing the future mitigation and adaptation strategies under global warming scenarios.
An investigation of moisture variability of the past millennium will help us to understand how the climate system responded to the natural forcings over a period of time longer than the available observational data. This will lead to the improvement of the future climate simulations. Various proxy data make the past millennium the best documented historical climate period. Despite the existence of long-term climate reconstructions, the regional hydro-climatic change in ACA over the past millennium is poorly understood. Our knowledge about the sensitivity of the severe climate conditions (e.g., droughts, floods, etc.) to changes in climate forcing is mostly limited to the modern instrumental records (Easterling et al. 2000). The published climate reconstructions like the Monsoon Asia Drought Atlas (Cook et al. 2010) are based on proxy data which are not homogeneously distributed over the Asia. The analysis of local effects is a challenging approach when using such data, as the time resolution of proxies does not allow the investigation of inter-annual and seasonal changes.
Previous studies based on different proxy data show that there is an out-of-phase behavior in moisture changes between the ACA and monsoon Asia (Chen et al. 2008; Chen et al. 2010; Cook et al. 2010; Fallah and Cubasch 2015; Polanski et al. 2014). According to Chen et al. (2010), this out-of-phase relation is more evident during the LIA. They proposed that the strengthening and southward shift of the westerly jet stream may have contributed to the moist LIA in the ACA. Sato et al. (2007) have studied the origins of water vapor over ACA in the recent climate. They concluded that the westerly circulations are major drivers of moisture changes in ACA. Lioubimtseva et al. (2005) concluded that the cyclones which originate from Mediterranean are transported by westerly jet stream into the ACA.
1.2 Regime behavior in the climate system
According to Lorenz (1969), the large-scale dynamics make the atmospheric long-term predictions possible. His fundamental work introduced application of the chaos theory in the atmospheric studies. He suggested that the atmosphere, which is a nonlinear system, may present deterministic regime behavior that is conditional to random changes (Slingo and Palmer 2011). Palmer (1993) have introduced the concept of extended-range atmospheric predictions by using the Lorenz convection model (Slingo and Palmer 2011). He showed that the climate response is predictable, even when a weak forcing in the Lorenz equation was included. Following this concept, Ruti et al. (2006) used NCEP-NCAR reanalysis and 40-year European Centre for Meduim-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) to show that although the extra-tropical troposphere is extremely chaotic, the subtropical jet strength shows a bimodal regime behavior. In this paper, we use a RCM to study the high resolution regional response of the climate to the forcings of the past millennium.
1.3 Regional climate modeling
Along with the Global Circulation Models (GCMs), RCM simulations can contribute to a better understanding of the impact of extreme climate regimes on past environmental changes (Diffenbaugh et al. 2006). Due to the highly complex topography of Central Asia, which is not well represented in coarse-resolution GCMs, RCM simulations provide more information about the small scale moisture changes under different boundary conditions. Additionally, proxy reconstructions present regional climatic variability which is not resolved by GCM simulations (Thompson and Anderson 2000).
This study tackles important subjects for which there is up to now mainly controversial evidences. Here, the sensitivity of extreme moisture events to climate of the past millennium is tested with a focus over Central Asia. First, the model set up is presented in Section 2 followed by results in Section 3. Finally, conclusions and discussions are presented in Section 4.
2 Data and methods
RCM’s time slices
Here, the AGCM runs of the past millennium from step (ii) are used as the driving model for RCM time-slice simulations. For the higher resolution simulations, the COSMO-CLM (CCLM) model version 4.8 clm17 (Steppeler et al. 2003; Dobler and Ahrens 2008; Asharaf et al. 2012) is applied with a horizontal resolution of 0.5∘×0.5∘ and with 32 vertical levels. The CCLM model is a non-hydrostatic RCM which uses terrain following height coordinates (Rockel et al. 2008), developed from the COSMO model, the current weather forecast model used by the German weather service (DWD).
COSMO-CLM model configuration parameters
Lateral Relaxation Layer (rlwidth)
Ritter and Geleyn
Rayleigh Damping Layer (rdheight)
Robert-Asselin time filter (alphaass)
Kinetic Energy (TKE)
The capability of the model of identifying the “exceptional” droughts during historical period was tested prior to applying the model for simulating the past millennium. The recent climate (1979–2005) simulation was validated against the global data set of monthly PDSI (Dai-PDSI, hereafter) (Dai 2011a, b, c). Dai-PDSI is calculated from the Climate Research Unit (CRU) monthly surface air temperature (Jones et al. 2001) and precipitation data from National Centers for Environmental Prediction (NCEP). The soil texture-based water-holding-capacity map from Webb et al. (1993) is applied as available water capacity (AWC) in the Dai-PDSI calculation.
We have additionally calculated the PDSI from the output of CCLM, driven by ECMWF ERA-Inrerim reanalysis data at a resolution of approximately 0.7∘. The results reveal that the COSMO-CLM is able to capture the PDSI patterns of the recent climate (not shown). The classical PDSI calculation method, following the one of Palmer (1965), is used to estimate the drought index from monthly surface air temperature and precipitation data from climate simulations. Dry area index is calculated based on the percentage of the grid points under dry conditions in each time-step. According to Dai (2013), the threshold of PDSI <−2 is suitable for such a consideration.
Following Zhao et al. (2014), the first principle component of 200 hPa zonal wind over 30∘– 50∘ N and 60∘– 100∘ E is used as the West Asian Subtropical Westerly Jet (WASWJ) displacement and the second one as the WASWJ strength. Expanding this domain up to 5∘ does not affect the EOF analysis results. The EOF patterns from AGCM and RCM simulations were similar as in the study of Zhao et al. (2014) (not shown). The positive values of the standardized Principle Component (PC1) time-series (with unit standard deviation and zero mean) indicate an equator-ward shift of the WASWJ. The positive values of the standardized PC2 time-series reveal the strength of the WASWJ. As in the study of Zhao et al. (2014), the summer (JJA) season is considered in our analysis. The AGCM simulations were previously validated against the reconstructions of summer monsoon failure index (Fallah and Cubasch 2015; Polanski et al. 2014).
3.1 Time-slice selection
The combined model-proxy comparison over the monsoon region with new paleo records revealed that the AOGCM was capable of capturing the majority of historical Asian droughts (Polanski et al. 2014; Fallah and Cubasch 2015).
3.2 AGCM simulations
3.3 RCM simulations
The probability of the PDF’s mode for different regimes of West Asian Subtropical Westerly Jet derived from RCM simulations
Probability of PDF’s mode
1 = dry MCA (960–990 AD)
2 = wet MCA (1060–1090 AD)
3 = wet LIA (1615–1645 AD)
4 = dry LIA (1645–1675 AD)
During the LIA (Fig. 8), most of the Asian domain shows near normal conditions (−2<PDSI<2). The region around the Bohai Sea in East China, however, shows a dry spell for all the preferred regimes. The Kazakhstan region does not show the alternating pattern which existed throughout the MCA under changing regimes in westerly jet.
4 Conclusions and discussion
This study focuses on the dynamical drivers of the moisture changes in central Asia during the past millennium by using the ECHAM5 AGCM and COSMO-CLM RCM simulations. After evaluating the performance of models in detecting the wet and dry regimes of the observational and reanalysis data, we applied them for the climatic moisture extremes of the past millennium. Our model experiments covered different possible climate behaviors throughout the past millennium which could be clustered into wet and dry spells. By comparing the dynamical behavior of westerly jet in the selected time-slices, we studied the differences of internal variations of the climate system between extreme dry and wet spells.
Following Lorenz’s idea of predictable climate response to different forcing, we analyzed the existence of a regime behavior in the westerly jet stream from model output. The evolution of the westerly jet showed a clear bimodal behavior. This regime behavior existed in both RCM and GCM simulations. The existence of the bimodality is mostly linked to the subtropical westerly jet displacement. The analysis based on the regional response of the hydro-climate of ACA to the large-scale climate forcing of the past millennium revealed that, during the MCA, this region was as sensitive as in the recent climate to the westerly jet stream. During the MCA the dipole pattern between India and ACA was not as pronounced as during the LIA. The sensitivity of moisture changes in Kazakhstan to westerly variability was stronger during the MCA. During the LIA East China showed dry patterns and Kazakhstan remained unaffected by the regimes of westerly jet changes.
We note that our simulations were based on a single driving model and the timing in the model may be uncertain, preventing us to make any conclusion about a specific year in the simulations. However, for 30-year time periods within MCA and LIA, the results depicted the averaged internally produced climate variability within these epochs. Using different driving GCMs for dynamical downscaling with RCMs will largely improve the certainty of the results. Regarding this, we suggest that considering more realizations using ensemble of driving GCMs and nested RCMs will produce a lot of added value in the results. This will lead to a larger coverage of the sample space. However, the computational costs will significantly increase in such approaches.
Finally, it should be emphasized that the main goal of this study was to investigate the possible drivers of regional hydro-climate changes during different periods of the past millennium and the speculations are based on simplified model set-ups. Any direct comparison of such climate simulations with proxy reconstructions may be misleading and would need further improvements.
This research was supported and funded by German Federal Ministry of Education and Research (BMBF) research project Central Asian Climate Dynamics (CADY) as part of the joint research program “CAME” Central Asia: Monsoon Dynamics and Geo-Ecosystems” and the DFG research group FOR 1380 “HIMPAC”. The authors thank the individual CADY/CAME and HIMPAC teams for permanent support and fruitful discussions. We acknowledge Edoardo Mazza for his precious critique and proofreading. Further, we thank Walter Acevedo, Emmanuele Russo, and Alexander Walter for their interesting discussions. The computational resources were made available by the German Climate Computing Center (DKRZ) through support from the BMBF. Finally, we acknowledge support by the German Research Foundation and the Open-Access Publication Funds of the Freie Universität Berlin.
- Ljung L (1999) System Identification: Theory for the users. Printice-Hall, Inc.Google Scholar
- Lorenz EN (1969) How complicated is circulation of earths atmosphere. Ann N Y Acad Sci 163(A1)Google Scholar
- Palmer WC (1965) Meteorological drought. U.S. Weather Bureau, Washington, D.C. 20852 3(3):1–10Google Scholar
- Sato T, Tsujimura M, Yamanaka T, Iwasaki H, Sugimoto A, Sugita M, Kimura F, Davaa G, Oyunbaatar D (2007) Water sources in semiarid northeast Asia as revealed by field observations and isotope transport model. J Geophys Res 112(D17)Google Scholar
- Stocker T, Qin D, Plattner G-K, Tignor M, Allen S, Boschung J, Nauels A, Xia Y, Bex V, Midgley P (2013) Ipcc, 2013: Climate change 2013: The physical science basis. contribution of working group i to the fifth assessment report of the intergovernmental panel on climate change. Technical report. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USAGoogle Scholar