Simulating the impact of wind and radiation on snow dynamics across linear disturbances in boreal forests
Milodowski, Rosa L.
Boreal forests are Earth‘s second largest forest biome, covering an area of 12.0–14.7 million km2. Winters are typically long, cold and dry, creating ideal conditions for sustaining snowpacks throughout this period. The spatial and temporal distribution of snow cover in boreal forest environments plays a crucial role in hydrological and ecological processes at local and regional scales. The dynamics of snow accumulation and melt reflect the interplay between such processes as the wind-driven redistribution of snow and the net energy balance at the snowpack surface. The presence of a forest canopy exerts a modifying effect on these processes; snow on the forest floor is typically sheltered from wind and direct solar radiation, whilst receiving enhanced longwave radiation from the surrounding canopy. However, the balance between these effects can be complex, particularly in the case of discontinuous forest canopies where clearings allow wind and light to penetrate down to the underlying snowpack. Understanding how the interplay between environmental factors drives spatially and temporally varying patterns of snow cover across forest edges is of particular importance and relevance in boreal regions where rates of climate change are high and forest fragmentation is increasing. In this thesis I explore how linear clearings, such as roads and tracks, may alter patterns of wind flow and incoming radiation, and consequently modify the dynamics of snow accumulation and melt across discontinuous forest canopies. This investigation uses field data collected during this research project and observations from long-running monitoring at the Arctic Research Centre of the Finnish Meteorological Institute (FMI-ARC), in northern Finland. Using a Met Office wind flow model (BLASIUS) I simulate patterns of wind flow across forest discontinuities and show that the clearing width is a key influence on these dynamics. There is less drag on the wind flow within the clearing relative to the forest canopy. Sufficient distance (approx. 100 m) is required for the wind flowing across the gap to adjust to this change in the boundary conditions. A region of reversed flow as the wind enters the gap was found for all gap widths. Within the 100 m gap, the wind speed then increases with distance across the gap until it is slowed by the presence of the downwind canopy edge. Narrow gaps (<30 m wide) have less impact on sub-canopy wind speeds as there is insufficient distance for the flow to fully adjust to the change in conditions. The influence of a forest gap on sub-canopy wind dynamics becomes negligible for very narrow gaps (approx. 3 m wide). Canopy height and density have a second-order effect on the wind flow dynamics across the gap. Increasing the canopy height accentuates the region of reversed flow, and faster flowing air above the canopy is not drawn down as deeply down into the gap. A denser forest canopy results in greater vertical velocities at the canopy edges. Reducing the canopy density results in greater overall wind speeds across the model domain. The wind flow model was coupled to a forest snow model (a simplified version of FSM2) using a linear scaling relationship observed between above- and below- canopy wind speeds, and similarly for surface friction velocity. This coupled model was used to explore the interactions between forest canopy, wind, and the surface energy balance on snow accumulation and melt across a range of forest gap scenarios. The simulated snow mass accumulates sooner and at a greater rate within the gap compared to under the forest canopy. In wider gaps (>50 m) there is an asymmetric pattern of snowmelt, with melt occurring sooner towards the exposed downwind edge of the gap and persisting for up to a month longer towards the sheltered upwind edge. Snow melts more evenly across narrower gaps. In the simulated scenarios I show that turbulent heat fluxes drive the spatial pattern of snowmelt across a gap. Simulated snowmelt patterns in the wider gaps correlate with sub-canopy wind speeds across the gap; higher wind speeds lead to greater fluxes of sensible heat and therefore earlier onset of melting and higher melt rates. Radiative fluxes provide a secondary influence on snow melt and have most impact on the melt dynamics towards the upwind edge where wind speeds are lowest. The canopy density influences the amount of sub-canopy snow accumulation and modulates the snowmelt patterns set by the energy fluxes across the gap. In the widest gap (100 m), increasing the LAI leads to later snow disappearance. The findings from this thesis demonstrate that introducing clearings into boreal forests produces a significant change in the local wind flow dynamics and snow hydrology. The width of the clearing is important, with canopy characteristics providing a secondary, modulating effect. The modifications to wind and snow induced by the presence of a gap in the canopy are greatest in the widest gaps. However, even narrow canopy gaps may have a significant impact if their orientation aligns closely with the prevailing wind direction. While the effects of an individual gap may be localised, they could become regionally significant in areas of boreal forest undergoing extensive fragmentation.