Extensive, detailed information on the spatial distribution of active layer thickness (ALT) in northern Alaska and how it evolves over time could greatly aid efforts to assess the effects of climate change on the region and also help to quantify greenhouse gas emissions generated due to permafrost thaw. For this reason, we have been developing high-resolution maps of ALT throughout northern Alaska. The maps are produced by upscaling from high-resolution swaths of estimated ALT retrieved from airborne P-band synthetic aperture radar (SAR) images collected for three different years. The upscaling was accomplished by using hundreds of thousands of randomly selected samples from the SAR-derived swaths of ALT to train a machine learning regression algorithm supported by numerous spatial data layers. In order to validate the maps, thousands of randomly selected samples of SAR-derived ALT were excluded from the training in order to serve as validation pixels; error performance calculations relative to these samples yielded root-mean-square errors (RMSEs) of 7.5-9.1 cm, with bias errors of magnitude under 0.1 cm. The maps were also compared to ALT measurements collected at a number of in situ test sites; error performance relative to the site measurements yielded RMSEs of approximately 11-12 cm and bias of 2.7-6.5 cm. These data are being used to investigate regional patterns and underlying physical controls affecting permafrost degradation in the tundra biome.

Knowledge of the spatial and temporal distribution of active layer thickness (ALT) throughout northern Alaska would help to understand the effects of climate change in the region, as well as to quantify how much the permafrost degradation manifestly in progress there is contributing to the accumulation of greenhouse gases in the atmosphere. For this reason, we are developing extensive high-resolution maps of ALT in northern Alaska. We use machine learning along with an extensive set of spatial data layers to upscale ALT from thousands of training pixels taken from high resolution swaths of estimated ALT derived from airborne polarimetric P-band synthetic aperture radar (SAR). The resulting maps of up-scaled ALT have been compared to thousands of validation samples set aside from the PolSAR-derived swaths and to in situ ALT measurements. The maps have achieved root-mean-square errors (RMSEs) of 5-7 cm relative to validation samples, and RMSEs of approximately 10-12 cm relative to in situ ALT measurements.

In this study we assess the total storage, landscape distribution, and vertical partitioning of soil organic carbon (SOC) stocks on the Brogger Peninsula, Svalbard. This type of high Arctic area is underrepresented in SOC databases for the northern permafrost region. Physico-chemical, elemental, and radiocarbon (C-14) dating analyses were carried out on thirty-two soil profiles. Results were upscaled using both a land cover classification (LCC) and a landform classification (LFC). Both LCC and LFC approaches provide weighted mean SOC 0-100 cm estimates for the study area of 1.0 +/- 0.3 kg C m(-2) (95% confidence interval) and indicate that about 68 percent of the total SOC storage occurs in the upper 30 cm of the soil, and about 10 percent occurs in the surface organic layer. Furthermore, LCC and LFC upscaling approaches provide similar spatial SOC allocation estimates and emphasize the dominant role of vegetated area (4.2 +/- 1.6 kg C m(-2)) and solifluction slopes (6.7 +/- 3.6 kg C m(-2)) in SOC 0-100 cm storage. LCC and LFC approaches report different and complementary information on the dominant processes controlling the spatial and vertical distribution of SOC in the landscape. There is no evidence for any significant SOC storage in the permafrost layer. We hypothesize, therefore, that the Brogger Peninsula and similar areas of the high Arctic will become net carbon sinks, providing negative feedback on global warming in the future. The surface area that will have vegetation cover and incipient soil development will expand, whereas only small amounts of organic matter will experience increased decomposition due to active-layer deepening.

Rapid and extensive snowmelt occurred during 2 days in March 2013 at a low-Arctic study site in the ice-free part of southwest Greenland. Meteorology, snowmelt, and snow-property observations were used to identify the meteorological conditions associated with this episodic snowmelt event (ESE) occurring prior to the spring snowmelt season. In addition, outputs from the SnowModel snowpack-evolution tool were used to quantify the snow-related consequences of ESEs on ecosystem-relevant snow properties. We estimated a 50-80% meltwater loss of the pre-melt snowpack water content, a 40-100% loss of snow thermal resistance, and a 4-day earlier spring snowmelt snow-free date due to this March 2013 ESE. Furthermore, the accumulated meltwater loss from all ESEs in a hydrological year represented 25-52% of the annual precipitation and may potentially have advanced spring snowmelt by 6-12 days. Guided by the knowledge gained from the March 2013 ESE, we investigated the origin, past occurrences, frequency, and abundance of ESEs at spatial scales ranging from local (using 2008-2013 meteorological station data) to all of Greenland (using 1979-2013 atmospheric reanalysis data). The frequency of ESEs showed large interannual variation, and a maximum number of ESEs was found in southwest Greenland. The investigations suggested that ESEs are driven by foehn winds that are typical of coastal regions near the Greenland Ice Sheet margin. Therefore, ESEs are a common part of snow-cover dynamics in Greenland and, because of their substantial impact on ecosystem processes, they should be accounted for in snow-related ecosystem and climate-change studies.

We scale a model of net ecosystem CO2 exchange (NEE) for tundra ecosystems and assess model performance using eddy covariance measurements at three tundra sites. The model, initially developed using instantaneous (seconds-minutes) chamber flux (similar to m(2)) observations, independently represents ecosystem respiration (ER) and gross primary production (GPP), and requires only temperature (T), photosynthetic photon flux density (I-0), and leaf area index (L) as inputs. We used a synthetic data set to parameterize the model so that available in situ observations could be used to assess the model. The model was then scaled temporally to daily resolution and spatially to about 1 km(2) resolution, and predicted values of NEE, and associated input variables, were compared to observations obtained from eddy covariance measurements at three flux tower sites over several growing seasons. We compared observations to modeled NEE calculated using T and I-0 measured at the towers, and L derived from MODIS data. Cumulative NEE estimates were within 17 and 11% of instrumentation period and growing season observations, respectively. Predictions improved when one site-year experiencing anomalously dry conditions was excluded, indicating the potential importance of stomatal control on GPP and/or soil moisture on ER. Notable differences in model performance resulted from ER model formulations and differences in how L was estimated. Additional work is needed to gain better predictive ability in terms of ER and L. However, our results demonstrate the potential of this model to permit landscape scale estimates of NEE using relatively few and simple driving variables that are easily obtained via satellite remote sensing.