| The northeastern Tibetan Plateau (NETP) is an important unit climatologically and hydrologically, for it functions as a transitional area between Asian Monsoon and westerly winds. Although clustered robust Qilian Juniper (Sabina przewalskii) dendrochronologies have been developed to reconstruct past climate variability over this region, associated reports on nonlinear relationships between tree growth and climate dataset, as well as potential driving forces on regional climate change are seldom.The process-based Vaganov-Shashkin (VS) model was used to simulate regional patterns of climate-tree growth relationships linking daily length, temperature and precipitation from meteorological data (AD1957-2000) over the NETP. The results exhibit that the leading principle component of the hypothetical growth curves is broadly consistent with that of the actual tree-ring chronologies, demonstrating the interpretability of the simulations as an accurate representation of the climatic controls on tree growth of Qilian Juniper. Output from this model both agrees well with the statistical relationships between tree-ring growth and climate factors as well as observational physiological behavior, i.e. precipitation in June acts as the most contributing role in annual ring formation of Qilian Juniper over the northeastern TP. The non-stationary and nonlinear response of tree growth to climate variability has important implications for calibration of tree-ring records for paleoclimate reconstructions and prediction for forest carbon sequestration. Modeling the complex relationships between climate and tree growth is of vital significance in dendroclimatic studies, but impossible to implement using the traditional regression approach. The technique of the artificial neural network (ANN) is employed in this study to explore the potential nonlinear relationships between climatic dataset and Qilian Juniper tree growth over the northeastern NETP. We observe that trees tend to integrate precipitation signals of consecutive months (May-June) and could "compensate" short-term water shortages or surpluses to maintain annual growth, but this association could only be realized when May-June precipitation totals are not extremely high. Speaking of the temperature-growth relationship, the increase in June average temperature accompanies a climb in evapotranspiration that could limit regional tree growth. High winter temperature plays its role by boosting tree’s growth-potential for the following year, while winter precipitation has no association with the next year’s tree growth. In addition, comparison between the linear and ANN-based nonlinear reconstructions reveals that the latter seems to be more reliable for recovering May-June precipitation variability over the NEQB. The linear reconstruction model could slightly underestimate the moisture changes due to its restricted ability in recording extreme pluvial conditions. However, small disparities between these two reconstruction models reveal that both of them are able to well place current May-June moisture conditions in the context of the past millennium.Additionally, the availability of five individual millennium-long width chronologies from Qilian Juniper allowed us to investigate long-term variability in the May-June precipitation over the NETP. The first principal component of the five chronologies (PC1) could well capture the contemporary climate changes and thus place the period in the context of the past millennium. Based on the empirical mode decomposition (EMD) method, the PC1was decomposed into eight intrinsic mode functions (IMF) and one trend (Res), which exhibited the moisture variations at inter-annual to centennial timescales around the northeastern TP. The dry epochs of the1140s-1150s,1290s-1300s, and1710s-1720s occurred during the cold phases of Pacific Decadal Oscillation (PDO) that often gave rise to the weakened southwest Asian summer monsoon. The severe drought during1450s-1500s may be related to the reduced land-sea thermal gradient as well as the intensified westerly winds. Comparison of the500hPa geopotential height fields in the years with anomalous May-June rainfall over the twentieth century revealed that positive mid-troposphere pressure variations over the Greenland and part of North Atlantic cohered with the1920s-1930s drought extremes, whereas negative geopotential height anomalies over the northeast Eurasian continent triggered the wetting tendency in the latter half of the twentieth century. In this study, we attempted to investigate the long-term correlation characteristics of the regional extreme pluvial and dry conditions by using the R/S analysis method. The results revealed that the Hurst exponents were0.84and0.92respectively, revealing that a long-enduring characteristic was implied for extreme conditions in the last millennium. Due to both greater than0.5, we conclude that the extremely pluvial conditions is likely to show a remarkable increase in the future, while quite the reverse was true for the case of extremely drought conditions. |
PDF Full Text Request