Improved monitoring of dust on snow from mining activities in the Arctic and its impact on spring snowmelt is proposed as one component of a dustfall standard for environmental assessment presently being developed by the Government of the NWT, Department of Environment and Natural Resources (ENR). Observations of earlier spring snowmelt along mining haul roads have been locally reported, but not consistently documented. Dust on snow along these roads could affect local snowmelt, vegetation growth and distribution, hydrological process, and permafrost dynamics, and provide positive feedback to climate change globally.
With funding from the NWT Cumulative Impact Monitoring Program (NWT CIMP), our project investigated the feasibility of using satellite remote sensing for quantifying dust on snow and its impact on snowmelt. First, we investigated the temporal dynamics of dust on snow snowmelt along unpaved haul roads around the Ekati Diamond Mine using MODIS data with daily coverage since 2000. The MODIS data fits well with that of the Ekati Diamond Mine development, which started construction in 1998 and operations in 2003. While the spatial extent of dust on snow could be detected using MODIS, its 250-m spatial resolution limits map accuracy. To improve accuracy, we further investigated 30-m Landsat data to quantify the spatial extent of dust on snow and its impact on snowmelt.
Initial results indicated that clean snow and dust-contaminated snow can be clearly separated using satellite data, with a visible spectral reflectance (albedo) ranging from ~0.9 for fresh clean snow to ~0.6 for packed clean snow, against only ~0.2 for dust-contaminated snow. The separability reduces when clean snow starts to melt (albedo = ~ 0.4), especially where bare tundra without snow cover (albedo = ~0.1) also occurs. Dust on snow appears to accumulate in detectable levels starting in March, and usually ends in May or early June along the Ekati Diamond Mine haul roads. In most cases, the spatial extent of dust on snow and resultant observable earlier snowmelt occurred within 1 km from the road. However, in several cases, a continuous field of dust on snow could be detected up to about 2 km from the haul roads. Other sources of dust on snow were also detected further away from roads, potentially related to exposed natural sources such as eskers.
The spatial extent of dust on snow detected in this study compare well with direct dustfall measurements (~ 1 km ), and the gradients of dust on leaves measured in situ (and supported by spectral satellite observations) and soil pH measured along the distance from the road (~ 1.5 km). The snow and lichen chemistry analyses also showed that “many of the highest concentrations occurred within 1 km of roads.” However, the snow and lichen chemistry analyses also found “many elements from the various sampling locations were above background levels within 10 to 30 km from the mine site”. Further investigation is needed for better understanding the reason for the enhancement of element levels within 10-30 km from the mine.