The NCEP reanalyses are widely used to study pan-Arctic climate including the evaluation of global climate model simulations of present day climate and the assessment of the arctic hydrological cycle and energy budget. While these analyses provide a best guess of the atmospheric variability over polar regions for the past 40+ years, the horizontal resolution of this dataset and others like it are inadequate for assessing the local climate in regions of complex surface characteristics such as those found in coastal regions. Profiles from the 2.5 degree NCEP reanalyses obtained at the “Barrow” grid-point are compared with results from a mesoscale model with much higher resolution (10 km)to ascertain the amplification of climate variability with increasing model resolution. The mesoscale model analyses are obtained by performing a series of concatenated 12-36 hour forecasts. Simulations are performed for time periods of minimum and maximum horizontal variability in surface conditions to assess the importance of resolving surface conditions in coastal regions. Detailed observations available from UAV flights and DOE ARM’s surface-based remote sensors and soundings are used to evaluate the mesoscale model analyses. Key variables from NCEP reanalyses and the mesoscale model analyses (e.g., temperature, cloud cover, precipitation, surface energy budget) are compared to quantify the amplification of climatic variability associated with model resolution and surface heterogeneity. Implications for arctic coastal studies requiring a long history of atmospheric data are given.