Abstract: Clarifying the distribution of soil nutrients in different habitat types in the Ulanbuh Desert can provide scientific basis for revealing nutrient limitation patterns. The habitats in the Ulanbuh Desert were classified into five types: semi-mobile sand dune, fixed/semi-fixed sand land, Gobi, saline alkali land, and grassland. Soil physical and chemical properties of those five habitat types were measured and stoichiometric ratios were calculated. Results showed that: (1) In the Ulanbuh Desert, soil available nitrogen and total potassium exhibited high stability, with coefficients of variation (CV) being 8.17% and 9.44%, respectively. Habitat type was the principal factor influencing the levels of soil organic matter, available nitrogen, and available potassium. The levels of soil total nitrogen, total phosphorus, and available phosphorus were influenced by the parent material. (2) Soil available nitrogen in the 0-10 cm layer of grassland was 132.90% higher (P<0.05) than the 10-20 cm layer, the available phosphorus was 96.43% higher (P<0.05) compared to 20-40 cm, and the available potassium was 52.11% higher (P<0.05) compared to 40-60 cm, highlighting a pronounced phenomenon of nutrient enrichment in the surface layer. (3) Principal component analysis indicated that soil organic matter contributed the most to soil stoichiometry across different habitat types. Moreover, there were significant positive correlations (P<0.05 or P<0.01) between soil organic matter and total nitrogen, total phosphorus, and available phosphorus. Therefore, soil organic matter can be used as an indicator to evaluate soil nutrient limitations of different habitats in the Ulanbuh Desert. Our results are of great significance for clarifying the differences in soil physical and chemical properties, stoichiometric ratios, and regulatory factors in different habitats of the Ulanbuh Desert, providing scientific basis for the protection and restoration of desert ecosystems.

Key words: ecological stoichiometry, habitat, desert, principal component analysis