Premium
Physical properties of Arctic versus subarctic snow: Implications for high latitude passive microwave snow water equivalent retrievals
Author(s) -
Derksen C.,
Lemmetyinen J.,
Toose P.,
Silis A.,
Pulliainen J.,
Sturm M.
Publication year - 2014
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
eISSN - 2169-8996
pISSN - 2169-897X
DOI - 10.1002/2013jd021264
Subject(s) - snow , water equivalent , subarctic climate , arctic , environmental science , microwave , high latitude , latitude , atmospheric sciences , climatology , meteorology , geology , geography , oceanography , engineering , geodesy , telecommunications
Two unique observational data sets are used to evaluate the ability of multi‐layer snow emission models to simulate passive microwave brightness temperatures (T B ) in high latitude, observation sparse, snow‐covered environments. Data were utilized from a coordinated series of 18 sites measured across the subarctic Northwest Territories and Nunavut, Canada in April 2007 during a 1000 km segment of a 4200 km snowmobile traverse from Fairbanks, Alaska to Baker Lake, Nunavut (~64°N). In April 2011, a network of 22 high Arctic sites was sampled across a 60 × 60 km study area on the Fosheim Peninsula, Ellesmere Island (~80°N). In comparison to sites across the subarctic, high Arctic snow was more spatially variable, thinner (site averages between 15 and 25 cm versus 30 to 40 cm), colder (−25°C versus −10°C), composed of fewer layers, had a proportionally higher fraction of wind slabs (storing 57% of the snow water equivalent (SWE) versus 15%), with these slabs comparatively denser (often exceeding 450 g/cm 3 , compared to 350 g/cm 3 in the subarctic). The physical snow measurements were used as inputs to snow emission model simulations. The radiometric difference between simulations of “typical” arctic and subarctic snow reached 30 K at 37 GHz. Sensitivity analysis showed that this T B difference could be partitioned between the effects of physical temperature (~5 K between −25°C and −10°C), wind slab density (~5 K between 0.40and 0.35 g/cm 3 ), and vertical depth hoar fraction (~20 K between 70% and 30% vertical fraction of total snow depth). Model simulations at the satellite scale (625 km 2 ) were produced using the observational spread for snow depth and snow stratigraphy. The range of T B from simulations with varied stratigraphy extended unrealistically far below the magnitude of satellite measured T B , illustrating that the snow depth first guess is very important for SWE retrieval schemes that are based on forward emission model simulations.