Constraining the Surface Energy Balance of Snow in Complex Terrain
Lapo, Karl Eric
MetadataShow full item record
Physically-based snow models form the basis of our understanding of current and future water and energy cycles, especially in mountainous terrain. These models are poorly constrained and widely diverge from each other, demonstrating a poor understanding of the surface energy balance. This research aims to improve our understanding of the surface energy balance in regions of complex terrain by improving our confidence in existing observations and improving our knowledge of remotely sensed irradiances (Chapter 1), critically analyzing the representation of boundary layer physics within land models (Chapter 2), and utilizing relatively novel observations to in the diagnoses of model performance (Chapter 3). This research has improved the understanding of the literal and metaphorical boundary between the atmosphere and land surface. Solar irradiances are difficult to observe in regions of complex terrain, as observations are subject to harsh conditions not found in other environments. Quality control methods were developed to handle these unique conditions. These quality control methods facilitated an analysis of estimated solar irradiances over mountainous environments. Errors in the estimated solar irradiance are caused by misrepresenting the effect of clouds over regions of topography and regularly exceed the range of observational uncertainty (up to 80Wm-2) in all regions examined. Uncertainty in the solar irradiance estimates were especially pronounced when averaging over high-elevation basins, with monthly differences between estimates up to 80Wm-2. These findings can inform the selection of a method for estimating the solar irradiance and suggest several avenues of future research for improving existing methods. Further research probed the relationship between the land surface and atmosphere as it pertains to the stable boundary layers that commonly form over snow-covered surfaces. Stable conditions are difficult to represent, especially for low wind speed values and coupled land-atmosphere models have difficulty representing these processes. We developed a new method analyzing turbulent fluxes at the land surface that relies on using the observed surface temperature, which we called the offline turbulence method. We used this method to test a number of stability schemes as they are implemented within land models. Stability schemes can cause small biases in the simulated sensible heat flux, but these are caused by compensating errors, as no single method was able to accurately reproduce the observed distribution of the sensible heat flux. We described how these turbulence schemes perform within different turbulence regimes, particularly noting the difficulty representing turbulence during conditions with faster wind speeds and the transition between weak and strong wind turbulence regimes. Heterogeneity in the horizontal distribution of surface temperature associated with different land surface types likely explains some of the missing physics within land models and is manifested as counter-gradient fluxes in observations. The coupling of land and atmospheric models needs further attention, as we highlight processes that are missing. Expanding on the utility of surface temperature, Ts, in model evaluations, we demonstrated the utility of using surface temperature in snow models evaluations. Ts is the diagnostic variable of the modeled surface energy balance within physically-based models and is an ideal supplement to traditional evaluation techniques. We demonstrated how modeling decisions affect Ts, specifically testing the impact of vertical layer structure, thermal conductivity, and stability corrections in addition to the effect of uncertainty in forcing data on simulated Ts. The internal modeling decisions had minimal impacts relative to uncertainty in the forcing data. Uncertainty in downwelling longwave was found to have the largest impact on simulated Ts. Using Ts, we demonstrated how various errors in the forcing data can be identified, noting that uncertainty in downwelling longwave and wind are the easiest to identify due to their effect on night time minimum Ts.
- Atmospheric sciences