Evaluation of bulk methods in simulating turbulent heat fluxes in katabatic flows at a glacier surface

Abstract:

Physics-based models of glacier surface melting are most lacking in their ability to accurately simulate turbulent heat fluxes. These fluxes are often parameterized using bulk aerodynamic methods, whose performance cannot be adequately evaluated, as direct flux measurements are rare at glacier surfaces. In this study, we obtain 57 days of continuous measurements of turbulent fluxes and meteorological variables from a set of stations installed in the summer of 2019 on Kaskawulsh Glacier, situated in the Yukon, Canada. Our goal is to explore the daily evolution of katabatic flow and associated fluxes, and evaluate the performance of bulk aerodynamic methods in simulating the fluxes at different stages of katabatic flow development. Eddy covariance and meteorological measurements (temperature, wind speed, and relative humidity) are collected at three heights (1m, 2m, and 3m) above the glacier surface. In order to identify different stages of katabatic flow within a day, we perform principal component analysis and clustering on the measurements of temperature and wind profiles. Results show that seven out of 57 days displayed a typical diurnal cycle of katabatic wind development, with a low-level wind speed maximum in the morning that deepens into the late afternoon and subsides again in the evening. Five days were associated with non-katabatic conditions due to the passage of low-pressure weather systems, while the remaining days displayed a range of katabatic flows mixed with wind development on a synoptic scale. During typical katabatic days, bulk methods overestimate fluxes during weak nighttime winds and underestimate fluxes during strong afternoon winds. Fluxes are overestimated at all wind speeds during high pressure synoptic systems. These findings indicate that a 'one-size-fits-all' bulk method may not be suitable for parameterizing turbulent heat fluxes at glacier surfaces.