Links between annual surface temperature variation and land cover heterogeneity for a boreal forest as characterized by continuous, fibre-optic DTS monitoring

A fibre-optic DTS (distributed temperature sensing) system using Raman-scattering optical time domain reflectometry was deployed to monitor a boreal forest research site in the interior of Alaska. Surface temperatures range between −40 ∘C in winter and 30 ∘C in summer at this site. In parallel experiments, a fibre-optic cable sensor system (multi-mode, GI50/125, dual core; 3.4 mm), monitored at high resolution, (0.5 m intervals at every 30 min) ground surface temperatures across the landscape. In addition, a high-resolution vertical profile was acquired at one-metre height above the upper subsurface. The total cable ran 2.7 km with about 2.0 km monitoring a horizontal surface path. Sections of the cable sensor were deployed in vertical coil configurations (1.2 m high) to measure temperature profiles from the ground up at 5 mm intervals. Measurements were made continuously over a 2-year interval from October 2012 to October 2014. Vegetation at the site (Poker Flat Research Range) consists primarily of black spruce underlain by permafrost. Land cover types within the study area were classified into six descriptive categories: relict thermokarst lake, open moss, shrub, deciduous forest, sparse conifer forest, and dense conifer forest. The horizontal temperature data exhibited spatial and temporal changes within the observed diurnal and seasonal variations. Differences in snow pack evolution and insulation effects co-varied with the land cover types. The apparatus used to monitor vertical temperature profiles generated high-resolution (ca. 5 mm) data for air column, snow cover, and ground surface. This research also identified several technical challenges in deploying and maintaining a DTS system under subarctic environments.