Wilhelm et al. (2015) employed the widely used Stefan and Kudryavtsev equations to predict the maximum active‐layer thickness (ALT) on Amsler Island, Western Antarctic Peninsula. Their predictions far exceed the observations of ALT reported from other parts of the region. Here, I demonstrate that the values of ALT are significantly overestimated by the predictive equations because the authors incorrectly assumed that little or no latent heat of phase change is absorbed during thawing. Although the area is the warmest in the Antarctic Peninsula region, with a rapid increase in air temperature and permafrost temperatures close to 0 °C, the active layer is likely to be substantially thinner than values predicted by Wilhelm et al. (2015). Copyright © 2016 John Wiley & Sons, Ltd.

Thermal regime and thickness of the active layer respond rapidly to climate variations, and thus they are important measures of cryosphere changes in polar environments. We monitored air temperature and ground temperature at a depth of 5 cm and modeled active‐layer thickness using the Stefan and Kudryavtsev models at the Abernethy Flats site, James Ross Island, Eastern Antarctic Peninsula, in the period March 2006 to February 2016. The decadal average of air and ground temperature was −7.3 and −6.1°C, respectively, and the average modeled active‐layer thickness reached 60 cm. Mean annual air temperature increased by 0.10°C y−1 over the study period, while mean annual ground temperature showed the opposite tendency of −0.05°C y−1. The cooling took place mainly in summer and caused thawing season shortening and active‐layer thinning of 1.6 cm y−1. However, these trends need to be taken carefully because all were non‐significant at p < 0.05. The Stefan and Kudryavtsev models reproduced the active‐layer thickness with mean absolute errors of 2.6 cm (5.0%) and 3.4 cm (5.9%), respectively, which is better than in most previous studies, making them promising tools for active‐layer modeling over Antarctica.

Sorted circles and polygons are widespread features of periglacial landscapes, but the controls on their development remain poorly understood, impeding their use as palaeoenvironmental indicators. We investigate the relationship of sorted circles and polygons to altitude in the northern Billefjorden area, central Svalbard. The patterns occur in two distinct elevation zones, below 200–250 m asl and above 600 m asl. The higher‐elevated patterns have smaller diameters and shallower sorting depths due to a thinner active layer at higher elevations, suggesting that sorted patterns can indicate climate conditions and ground thermal state when the patterns initiated. Geology is believed to be of less importance for pattern morphology in the study area, causing only its fine‐scale variations. The pattern diameter‐to‐sorting depth ratios have a median value of 3.57, consistent with previous studies and theoretical models of patterned‐ground formation involving circulation mechanisms. Large‐scale sorted patterns may develop over centennial timescales in this high‐Arctic environment. They are probably not in equilibrium with present‐day climate conditions and have probably formed throughout the Holocene. Copyright © 2017 John Wiley & Sons, Ltd.