The central focus of our paper, published in Environmental Research Letters (ERL), was to determine whether model uncertainty in estimates of low Arctic NEE could be reduced by using satellite remote sensing observations to represent the influence of snow on NEE.

These investigations were informed primarily by observations of meteorology, NEE, and fractional snow cover area (SCA) collected within four locations at a low Arctic site (Daring Lake, NT, Canada) in May and June 2010. NEE was measured using the eddy covariance technique, and SCA was observed by classifying thrice daily time-lapse camera observations.

At all four locations, the timing of snow melt marked an important transition in both photosynthesis and respiration, and satellite remote sensing observations accurately estimated SCA. Remote sensing estimates of SCA were therefore incorporated into an existing model – the Vegetation Photosynthesis Respiration Model (VPRM) – using six different formulations that simulated the effects of SCA on photosynthesis and/or respiration. All model formulations were then calibrated using 2005 observations of NEE and meteorological conditions at Daring Lake, and all were run over Daring Lake and its paired validation site (Ivotuk, AK, US) for 2004–2007. Model errors were quantified for each formulation by examining model outputs against the corresponding eddy covariance observations of NEE.

The findings indicate that model uncertainties were reduced when remote sensing observations of SCA were used to delineate the snow and growing seasons, allowing growing season respiration to be estimated according to air temperature, and snow season respiration to be estimated according to soil temperature. This model formulation captures the insulating influence of snow, which maintains warm soil temperatures conducive to soil respiration in spite of sub-zero air temperatures.

We therefore suggest that future estimates of Arctic NEE consider representing the influence of snow on NEE using remote sensing, both to improve model fit against observations, and to allow insights into how changing snow and growing season conditions are affecting the Arctic carbon cycle.

The team’s paper was part of the ERL Focus on Cryospheric Ecosystems.

Related links