Purpose Surface forest litter serves both as an important carbon substrate and as a physical barrier regulating the soil microenvironment. Although its general role in controlling soil carbon emissions has been widely recognized, it remains unclear, how litter removal alters the sensitivity of soil respiration to hydrothermal factors in water-limited semi-arid regions, and how these hydrothermal responses differ among forest stand types. This study aimed to quantitatively assess the effects of litter removal on soil carbon fluxes and their hydrothermal sensitivity in two typical forest stands of Xinglong Mountain, thereby providing essential empirical parameters for the accurate assessment and management of regional forest carbon sinks.

Methods Typical Pinus tabuliformis plantations and natural Populus davidiana–Betula platyphylla stands were selected in Xinglong Mountain, Gansu Province. A randomized complete block design was employed to establish control and litter removal treatments. Soil respiration rates during the growing-season were monitored in situ over two consecutive years from 2024 to 2025 Using an LI-8100 automated open-path soil CO₂ flux system. Soil temperature and moisture at a depth of 5 cm were measured simultaneously.

Results (1) The mean growing-season respiration rate in the control plots of the natural P. davidiana–B. platyphylla stands (4.31 μmol·m⁻²·s⁻¹) was significantly higher than that in the P. tabuliformis plantations (3.41 μmol·m⁻²·s⁻¹). Litter removal significantly suppressed soil respiration in both stand types, reducing the cumulative carbon fluxes of the natural P. davidiana-B. platyphylla stands and P. tabuliformis plantations to 4.98~5.07 t C·hm⁻² and 3.92~4.02 t C·hm⁻², respectively, with an average reduction in the P. tabuliformis plantations increased significantly, as indicated by an increase in the explanatory power of soil temperature from 37.34% to 57.62% and an increase in Q₁₀ from 1.42 to 2.05. In contrast, Q₁₀ in the natural P. davidiana-B. platyphylla stands remained relatively stable, ranging from 1.02 to 1.03 after litter removal. However, the coupling between soil respiration dynamics and soil moisture variability was substantially strengthened, with the explanatory power of soil moisture increasing from 59.01% to 75.81%.

Conclusion Litter removal not only reduced soil respiration fluxes but also significantly altered their dependence on hydrothermal factors, with distinct stand-specific response patterns. Soil respiration in the P. tabuliformis plantations became more sensitive to temperature, whereas that in the natural P. davidiana-B. platyphylla stands showed enhanced sensitivity to soil moisture variation. These findings suggest that future assessments and simulations of regional forest carbon fluxes should fully account for forest stand type and stand-specific hydrothermal response pathways, thereby improving the accuracy of carbon sink potential estimation in semi-arid forest ecosystems.