Permafrost soils are especially vulnerable to global climate change and warming temperatures can turn them from carbon sinks into carbon sources. Estimates of Arctic carbon stocks are still highly uncertain, despite their importance to predict the magnitude of CO2 and CH4 release to the atmosphere. Because most of the Arctic is difficult to access, remote sensing techniques are particularly important to monitor the changing landscape. Recent studies have attempted to use spectral images, like Landsat, to estimate soil organic carbon (SOC) and nitrogen (TN). Most studies worked on a regional to global scale and use relatively coarse landscape classes. However, good, high resolution estimates of SOC and TN are crucial to estimates for permafrost related uncertainties in storage and spatial heterogeneity needed for Earth System Models. Furthermore, they are an invaluable step from data collection toward a process oriented understanding of the landscape. This project is one of the first to use high resolution images (1.65m GeoEye (4 spectral bands: blue-infrared), 2m DEM) to predict SOC and TN within different Tundra vegetation classes in a small twin watershed (4 km²) on Herschel Island, Yukon, Canada. Vegetation classes were based on indicator species and geomorphic disturbance levels. Remote sensing detection accuracy varied strongly between classes. Field based moisture measurements were most strongly correlated with the carbon to nitrogen (CN) ratio (r²=0.80, p<0.05). However, slope and the normalized difference vegetation index (NDVI) which were extracted from remote sensing images have a statistically significant relationship to CN (r²=-0.56, p<0.05, r²=0.48, p<0.05). This suggests that fine scale estimates of carbon and nitrogen stocks are possible using few spectral bands from high resolution images. Given the high correlation of soil moisture with CN ratios we encourage further research to improve validation of satellite radar moisture information with field data.