| dc.description.abstract | The Snow Albedo Feedback (SAF) is an important contributor to Arctic warming,
however models disagree significantly on the strength of this effect. Previous work has
investigated the influence of vegetation on surface albedo, however the accuracy has been
limited by the resolution of model output. In this work, we perform a pan-Arctic survey
using Moderate Resolution Imaging Spectroradiometer (MODIS) remote sensing data and
Coupled Model Intercomparison Project (CMIP6) model output, to perform a comprehensive
analysis of the effects of vegetation on SAF. We computed pan-Arctic composites of
MODIS observational data at the 500m scale and compared the results with the CMIP6
ensemble.
Using MODIS data, we found a mean SAF of -2.17 % ⋅ K−1 from April-July over the
climatological period 2001-2019, which is stronger than predicted by the CMIP6 intermodel
mean. Additionally, we identified the source of this discrepancy - models currently do not
adequately capture the dynamics of late-season melt-off in the high Arctic grassland and
barren regions, which results in an underestimate of SAF. In this work, we demonstrate
that land cover changes have a small but nonzero (≤ 10%) contribution to overall changes
in SAF on the timescale of decades, indicating the importance of dynamic vegetation
models. Furthermore, we identify upscaling resolution as a major source of local error in
SAF, however due to cancellation of errors this has minimal impact on estimates of pan-
Arctic mean SAF. Finally, we identify a logarithmic relationship between LAI (Leaf Area
Index) and SAF. This work can benefit modelling groups seeking to better capture SAF
dynamics, by explaining SAF errors in terms of vegetation dynamics, and by demonstrating
the existence of spatial structure in SAF fields at sub-model grid scales. | en |
