Precipitation characteristics in high-mountain regions with complex terrain remain poorly understood because observational networks are sparse and robust instrumentation is rarely deployed. This study evaluates precipitation type measurements from a PARSIVEL2 optical disdrometer installed at 4709 m a.s.l., approximately 2.5 km from the Huaytapallana tropical glacier in the Peruvian Andes. The instrument measures the equivalent diameter and fall velocity from particle shadows crossing its laser beam; it then computes precipitation intensity (mm h^−1) and classifies hydrometeor types at 1 min resolution. Based on one year of observations, we identified seven precipitation types: rain, drizzle, drizzle with rain, snow, hail, soft hail, and mixed rain drizzle with snow. The original PARSIVEL2 classification indicated that drizzle with rain was the most frequent type (30.57 %), followed by snow (26.15 %). We identified 70 precipitation events (duration ≥10 min) and compared the corresponding accumulations against a Pluvio2 weighing rain gauge (threshold ≥0.25 mm). The PARSIVEL2 systematically overestimated precipitation, especially during mixed-phase events (98.5 %, 3.92 mm bias) and solid precipitation events (84.1 %, 7.09 mm bias), whereas liquid precipitation events showed minimal bias (15.7 %, 0 mm bias). The largest discrepancies occurred during extreme events (>10.6 mm and >1 h) dominated by snow, soft hail, and hail, which we attribute to misclassification of coexisting particle types and systematic deviation of the instrument’s internally calculated density values. We developed a correction methodology that combines velocity diameter reclassification criteria based on established empirical relationships with site specific density optimization. For solid precipitation, RMSE decreased from 13.05 to 4.4 mm; for mixed precipitation, the scope value improved from 1.936 to 1.004. The corrected classification identified wet snow (29.3 %) and graupel (15 %) as dominant precipitation types, whereas pure snow represented only 3.7 %. These results demonstrate the need for post processing corrections of disdrometer measurements in tropical glacier environments and provide an improved characterization of mixed-phase precipitation processes relevant to glacier mass balance assessments.