RESEARCH HIGHLIGHT

Northern Alaska Connections III: Winter observations of in-situ snow depth during 2005-2008

April 29, 2008

It is very difficult to measure snow cover in the Arctic from satellite data. However, understanding snowmelt and coverage in the Arctic is of great importance for climate research and global warming studies in the Arctic.

The advanced microwave sensors aboard AMSR-E satellites have recently returned new data sets with snow coverage information from the Arctic region. The AMSR-E has provided temperature data with more channels, higher spatial resolution and frequent coverage than previously collected. This information is being analyzed for snow depth and snow-melt using new techniques.

The winter observations of in-situ snow depth project were conducted as part of a IJIS (IARC-JAXA Information System, formerly known as INIS, IARC-NASDA Information System) project during the winter from 2005 to 2008. It was an observational component of the build-up of a new algorithm on the wide-range snow depth using remote sensing techniques. Passive microwave radiometer AMSR and AMSR-E on AQUA provides comprehensive snow and ice data in the Arctic. This project used AMSR/AMSR-E microwave brightness temperature (Tb) data and also AMSR/AMSR-E products through EORC.

For the observational component, permanent sites were fixed along the Trans-Alaska pipeline along the Dalton highway, at 20 miles. Eleven sites are in the boreal forest and 12 are on areas of tundra. (Figure 1)

In order to better understand temporal snow depth along a latitudinal transect, the observational periods were scheduled for mid-January for 2005, early February for 2006, mid-February to early March for 2007, and mid-March for 2008.

Results of annual snow depth observations in boreal forest and tundra areas can be seen in Figure 2. In the boreal forest region, snow depth remarkably decreased in 2007 but increased in 2008, whereas the tundra region measurements showed more consistency between observations suggesting little accumulation of snow by strong wind.

Kimura et al. (2007) built a new algorithm for analysis of the wide-range snow depth of satellite data (Figures 3 and 4). The algorithm was modified with snow pit-wall observations and takes into account the effect of tree-coverage. Their findings were similar to the in-situ snow depth observations but indicated that snow depth in the boreal forest was somewhat overestimated while snow depth on the tundra was underestimated.

Figure 1. Site locations along the Trans-Alaska pipeline.

Figure 2. Annual variation of snow depth in boreal forest and tundra.

Figure 3. Comparison of algorithms: Kimura (Enomoto's group) and by Chang et al. (1987) with observations along the Trans-Alaska pipeline during three winters of 2005-2007. The empty and solid triangles denote Chang and Kimura's algorithms.

Figure 4. Algorithm visualization by Kimura (right) and Chang et al., 1987, (left) from observations along the Trans-Alaska pipeline during the winters of 2005-2007.

Snow depth measurements on the tundra (left) and in the boreal forest (right) along the Trans-Alaska pipeline.