Abstract: Accurate estimation of snowfall is crucial for weather forecasting, climate research, and hydrological management. However, the high variability in snow particles, including size distribution, density, and shape, pose significant challenges for radar-based snowfall estimation. This study evaluates the applicability and accuracy of eight radar reflectivity factor (Z)-snowfall intensity (S) relationships (abbreviated as S1-S8) for snowfall estimation using observation data from the Qitai Station dual-polarization radar and ground meteorological station (Station ID: 51379). Fourteen snowfall events with liquid water equivalent exceeding 1 mm between 2022 and 2024 were selected for analysis. Four statistical metrics—correlation coefficient (CC), mean absolute error (MAE), mean bias error (MBE), and root mean square error (RMSE)—were employed to systematically assess the overall performance of the eight Z-S relationships and their effectiveness in two typical snowfall cases. The results demonstrate that S3 (Z=120S2) and S6 (Z=40S2) exhibit the best overall performance. For the November 9, 2022 event, S3 achieved the smallest errors (MAE=8.1347, MBE=-1.3023, RMSE=10.7004) and the highest correlation with ground observations (CC=0.9936). For the January 11, 2023 event, S6 showed the lowest overall errors (MAE=9.3978, RMSE=10.7111) despite slight overestimation of heavy snowfall (MBE=0.2489). Consequently, S3 and S6 demonstrate superior applicability for snowfall estimation in the study area. These findings provide valuable insights for optimizing Z-S relationships and improving snowfall quantification accuracy.