Sublimation is a key component of land–atmosphere interactions and represents an important pathway of mass loss from snow/ice surfaces. Accurate quantification of sublimation is crucial for understanding these processes in which sublimation is involved. However, conventional approaches, including the eddy covariance and bulk aerodynamics, suffer from considerable uncertainties, making the observation and simulation of sublimation highly challenging. To solve this problem, we established an empirical formula based on the daily maximum temperature (Tmax), relative humidity (RH), wind speed (U), and D20 pan sublimation (Span) in the Hulugou small watershed of Qilian Mountains to simulate the daily ice surface sublimation (Sice). At the same time, bulk aerodynamics, Penman-Monteith equation, gravimetric method, and two other empirical formulas were also utilized to estimate sublimation. A comparative analysis demonstrated that our proposed formula outperformed all other methods, with correlation coefficients (CC), Kling-Gupta coefficients (KGE), relative deviations (Bias), and mean absolute errors (MAE) of 0.84, 0.76, 0.04 mm d −1 , and 0.23 mm d −1 , respectively. Owing to its simple and easily obtainable parameters and reliable simulation capability, this formula is highly suitable for quantifying the sublimation and snow/ice material balance research in the Qilian Mountains. This provides important assistance for local water resource management and climate research. • We proposed a parameterization to estimate the ice sublimation in the Qilian Mountains in winter. • Sublimation estimates derived from six commonly used methods are systematically evaluated against in situ observations in the Hulugou watershed of the Qilian Mountains. • The performance of different sublimation estimation methods is discussed based on previous studies.
Ma et al. (2026) studied this question.