Comparative Study of Soil Respiration Characteristics and Influencing Factors of Four Land-Use Types in the Aksu Desert–Oasis Ecotone

Aims: Land-use types and soil salinization in arid desert–oasis ecotones are associated with differences in soil CO2 efflux, yet the combined effects of multiple environmental factors remain poorly understood. This study aims to characterise the seasonal patterns, temperature sensitivity, and major correlates of soil respiration across four typical land-use types, thereby supporting regional carbon budget assessments. Methods: We conducted in situ soil respiration monitoring during the 2024 growing season (April–October) in four representative habitats (Populus euphratica forest, jujube orchard, grassland, and wetland) in the Aksu desert–oasis ecotone, Xinjiang. Concurrent measurements of topsoil (0–20 cm) physicochemical properties and extracellular enzyme activities were taken. Pearson correlation, variation partitioning, and the Van’t Hoff temperature-response model were employed. Results: Soil respiration exhibited a unimodal seasonal pattern peaking in July, with growing-season mean fluxes ranked as grassland > wetland > jujube orchard > Populus euphratica forest. Temperature sensitivity (Q10) followed wetland (3.23) > grassland (2.44) > orchard (2.08) > forest (1.57), a pattern consistent with differences in soil moisture and salinity among the sites. Respiration correlated positively with pH, electrical conductivity, total salt, organic matter, urease, and sucrase, and negatively with available N, P, K, and peroxidase (p < 0.05). Physicochemical properties and enzyme activities jointly explained 70.7% of variation, with total soil salt (12.5%), sucrase (12.2%), and available N (11.1%) as the top contributors. Root and heterotrophic respiration accounted for 37.2% and 62.8% of total soil respiration, respectively, with heterotrophic respiration dominating in all seasons. Conclusions: Soil CO2 emissions differed markedly among the four land-use types, and these differences were associated with water–salt–thermal–nutrient–enzyme conditions. Carbon cycling mechanisms in wetlands and grasslands differ markedly from those in woody habitats, highlighting the need for differentiated land-use planning and salinity-adapted carbon management strategies in arid regions.