This study investigates the spatiotemporal characteristics of dust devil activities and their contributions to dust emissions across different underlying surfaces in the Xiaotang desert-oasis transition zone (northern edge of the Taklimakan Desert) and the Tazhong shifting sand terrain (desert hinterland). By integrating meteorological observations and intensive experimental data, it reveals that Xiaotang exhibits significantly higher annual dust devil frequency (127 occurrence) than Tazhong (51 occurrence), with a pronounced seasonal concentration in summer (55.6% occurrence). Both regions predominantly feature short-lived dust devils lasting 1-4 min (72.5% in Xiaotang vs. 82.3% in Tazhong), reflecting the transient nature of thermally driven turbulence. The formation mechanisms are governed by thermo-dynamic coupling, where Tazhong's shifting sand terrain requires stronger thermal forcing: critical land-air temperature gradients (the surface and near- surface 2 m air temperature) increase to 16.6-18.3 degrees C (12.4-16.2 degrees C in Xiaotang), and initial wind speed thresholds rise to 3.3-4.6 m/s, attributed to lower thermal capacity and reduced energy transfer efficiency of sandy substrates. Spatiotemporal divergence in dust emission contributions is evident-Xiaotang's summer (July) dust devil contribution peaks at 57.85% (30.1% annually), while Tazhong's declines to 36.72% (19.2% annually) due to wind-dominated suppression of thermal processes. During non-summer periods (March), dusty weather dominate emissions with contributions exceeding 98% in both regions, highlighting synergistic regulation between thermal dynamics and climatic conditions. The study quantifies surface heterogeneity's role in modulating dust aerosol emissions across arid zones, establishing critical parameters for multiscale dust modeling. These findings advance the precision of aerosol-climate effect assessments and provide theoretical foundations for desertification control and dust storm early-warning systems, bridging critical gaps in global dust source strength simulations.