Zhonglin Xu ^1* Yao Wang ²

• ¹ Xinjiang University, Urumqi, China
• ² China Meteorological Administration, Urumqi, China

Ecological stoichiometry of terrestrial ecosystems has been a hot issue in current research, with intense focus on the proportional relationships of nutritional elements within plants and between plants and their environment. To clarifying these relationships along continuous environmental gradients is essential for a more comprehensive understanding how plants adapt to changing environment. In arid regions, the varying plant and soil types along altitude gradient offer a unique opportunity to examine the vertical spectrum of plant and soil ecological stoichiometry. In this study, the northern slope of the Tianshan Mountains was selected as the study area to explore the carbon (C), nitrogen (N) and phosphorus (P) ecological stoichiometric characteristics of herbaceous plants along 900-3500 m altitude gradients. We also investigated the variation of ecological stoichiometric characteristics among different grassland types. The results indicated that the mean C, N, P in leaf of grassland were 342.95-557.73 g• kg -1 , 6.02-20.97 g• kg -1 , and 0.71-3.14 g• kg -1 , respectively. There was no significant change in leaf carbon content along the elevation gradient and the highest and lowest leaf C concentration was in the upland meadow and the semidesert grasslands.. Both N and P concentrations obtained their highest value in the meadow steppe. The P concentration gradually increased in desert, semidesert grasslands and reached the highest value in meadow steppe, then decreased to lowest value in upland meadow and subsequently increased in alpine meadow. The ranges of the C:N ratio, C:P ratio and N:P ratios were 16.36-155.53, 109.36-786.52 and 2.58-17.34, respectively. Due to fluctuations in the P concentration, C:P ratio and N: P ratio reached the lowest value in meadow steppe, and obtained their highest value in upland meadow. Redundancy analysis showed that temperature were the dominant factors affecting the C, N, and P ecological stoichiometry of herbaceous plants, followed by soil organic carbon, mean annual precipitation, soil pH and soil electrical conductivity. Corresponding results could enhance predictive models of nutrient cycling and ecosystem responses to climate change, particularly in arid and semi-arid regions.