气候变化领域集成服务门户

The stable carbon isotopic composition of bulk organic matter (delta C-13(org)) in paleosols has been widely used as a proxy indicator for reconstructing past vegetation and climate. Previous studies generally show a negative correlation between the delta C-13(org) values of modern C-3 plants and surface soils under C-3 plant-dominated ecosystems and mean annual precipitation (MAP). However, the relationship between delta(13)Corg and MAP varies among different climatic regimes, resulting in uncertainties in paleoprecipitation reconstructions. In this study, we analyzed the delta C-13(org) values of surface soil samples collected along a north-south climatic gradient in northern Iran, in the western part of arid central Asia (ACA). Our aims were to explore the relationship between surface soil delta C-13(org) values and climatic factors; to use it to develop a transfer function for ACA; and then to apply the transfer function to a well-dated loess-paleosol sequence (AB1) to reconstruct early Pleistocene (2.4-1.8 Ma) MAP in northern Iran. The results show that: (1) the delta C-13(org) values of 44 surface soil samples range from -27.56%, to -23.61%,, with an average of -25.77%,, indicating that the modern natural ecosystem in northern Iran is dominated by C-3 vegetation; (2) The delta C-13(org) values of the surface soil samples are strongly negatively correlated with MAP (y = -0.0079x - 22.8418, R-2 = 0.4419, p < 0.001), with a coefficient of -0.79%/100 mm. We infer that the effects of mean annual temperature (MAT) and elevation on the relationship are minimal; (3) Reconstructed early Pleistocene MAP in northern Iran was similar to 649 mm, much wetter than the present day (similar to 331 mm); (4) The loess delta C-13(org) values from ACA exhibit similar trends to records from monsoonal Asia, suggesting a similar history of Quaternary climate change between ACA and monsoonal Asia, which is likely related to the growth and decay of Northern Hemisphere ice sheets. (C) 2018 Elsevier Ltd. All rights reserved.