<p>Misclassification of convective precipitation pixels as stratiform precipitation over the Tibetan Plateau (TP) in summer by the precipitation type classification algorithm (CSF algorithm) of the Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR) has been reported in several studies. To investigate the cause of this misclassification, we validated the CSF algorithm using Doppler radar observations of mesoscale wind divergence. The case studies revealed that the algorithm performs well for typical isolated deep convection, whereas embedded convection within stratiform precipitation was misclassified; the divergence profile clearly indicated convective characteristics, although the CSF algorithm classified it as stratiform. This provides the first evidence of misclassification of precipitation type over the TP in summer based on mesoscale wind divergence. We found that the misclassification is primarily caused by the thresholds of the Peakedness function in the H-method, and that reclassification aimed at enhancing the detection of embedded convection and weak convective precipitation led to a more reasonable convective precipitation fraction.</p>